Measurement

of Geometric Tolerances in Manufacturing

title:
Manufacturing Engineering and Materials Processing ; 52
author: Meadows, James D.
publisher: CRC Press
isbn10 | asin: 0824701631
print isbn13: 9780824701635
ebook isbn13: 9780585157191
language: English
subject Tolerance (Engineering)
publication date: 1998
lcc: TS172.M43 1998eb
ddc: 620/.0045
subject: Tolerance (Engineering)
 Page aa

Measurement of Geometric Tolerances in Manufacturing

 Page ab
MANUFACTURING ENGINEERING AND MATERIALS PROCESSING
A Series of Reference Books and Textbooks
FOUNDING EDITOR
Geoffrey Boothroyd
University of Rhode Island
Kingston, Rhode Island
1. Computers in Manufacturing, U. Rembold, M. Seth, and J. S. Weinstein
2. Cold Rolling of Steel, William L. Roberts
3. Strengthening of Ceramics: Treatments, Tests, and Design Applications, Harry P.
Kirchner
4. Metal Forming: The Application of Limit Analysis, Betzalel Avitzur
5. Improving Productivity by Classification, Coding, and Data Base Standardization: The
Key to Maximizing CAD/CAM and Group Technology, William F. Hyde
6. Automatic Assembly, Geoffrey Boothroyd, Corrado Poli, and Laurence E. Murch
7. Manufacturing Engineering Processes, Leo Alting
8. Modern Ceramic Engineering: Properties, Processing, and Use in Design, David W.
Richerson
9. Interface Technology for Computer-Controlled Manufacturing Processes, Ulrich
Rembold, Karl Armbruster, and Wolfgang lzmann
10. Hot Rolling of Steel, William L. Roberts
11. Adhesives in Manufacturing, edited by Gerald L. Schneberger
12. Understanding the Manufacturing Process: Key to Successful CAD/ CAM
Implementation, Joseph Harrington, Jr.
13. Industrial Materials Science and Engineering, edited by Lawrence E. Murr
14. Lubricants and Lubrication in Metalworking Operations, Elliot S. Nachtman and
Serope Kalpakjian
15. Manufacturing Engineering: An Introduction to the Basic Functions, John P. Tanner
16. Computer-Integrated Manufacturing Technology and Systems, Ulrich Rembold,
Christian Blume, and Ruediger Dillman
17. Connections in Electronic Assemblies, Anthony J. Bilotta
18. Automation for Press Feed Operations: Applications and Economics, Edward Walker
19. Nontraditional Manufacturing Processes, Gary F. Benedict
20. Programmable Controllers for Factory Automation, David G. Johnson
21. Printed Circuit Assembly Manufacturing, Fred W. Kear
22. Manufacturing High Technology Handbook, edited by Donatas Tijunelis and Keith E.
McKee
23. Factory Information Systems: Design and Implementation for CIM Management and
Control, John Gaylord
24. Flat Processing of Steel, William L. Roberts

 Page ac
25. Soldering for Electronic Assemblies, Leo P. Lambert
26. Flexible Manufacturing Systems in Practice: Applications, Design, and Simulation,
Joseph Talavage and Roger G. Hannam
27. Flexible Manufacturing Systems: Benefits for the Low Inventory Factory, John E.
Lenz
28. Fundamentals of Machining and Machine Tools: Second Edition, Geoffrey Boothroyd
and Winston A. Knight
29. Computer-Automated Process Planning for World-Class Manufacturing, James Nolen
30. Steel-Rolling Technology: Theory and Practice, Vladimir B. Ginzburg
31. Computer Integrated Electronics Manufacturing and Testing, Jack Arabian
32. In-Process Measurement and Control, Stephan D. Murphy
33. Assembly Line Design: Methodology and Applications, We-Min Chow
34. Robot Technology and Applications, edited by Ulrich Rembold
35. Mechanical Deburring and Surface Finishing Technology, Alfred F. Scheider
36. Manufacturing Engineering: An Introduction to the Basic Functions, Second Edition,
Revised and Expanded, John P. Tanner
37. Assembly Automation and Product Design, Geoffrey Boothroyd
38. Hybrid Assemblies and Multichip Modules, Fred W. Kear
39. High-Quality Steel Rolling: Theory and Practice, Vladimir B. Ginzburg
40. Manufacturing Engineering Processes: Second Edition, Revised and Expanded, Leo
Alting
41. Metalworking Fluids, edited by Jerry P. Byers
42. Coordinate Measuring Machines and Systems, edited by John A. Bosch
43. Arc Welding Automation, Howard B. Cary
44. Facilities Planning and Materials Handling: Methods and Requirements, Vijay S. Sheth
45. Continuous Flow Manufacturing: Quality in Design and Processes, Pierre C.
Guerindon
46. Laser Materials Processing, edited by Leonard Migliore
47. Re-Engineering the Manufacturing System: Applying the Theory of Constraints,
Robert E. Stein
48. Handbook of Manufacturing Engineering, edited by Jack M. Walker
49. Metal Cutting Theory and Practice, David A. Stephenson and John S. Agapiou
50. Manufacturing Process Design and Optimization, Robert F. Rhyder
51. Statistical Process Control in Manufacturing Practice, Fred W. Kear
52. Measurement of Geometric Tolerances in Manufacturing, James D. Meadows
Additional Volumes in Preparation

 Page i

Measurement of Geometric Tolerances in Manufacturing

JAMES D. MEADOWS
Institute for Engineering & Design, Inc.
Hendersonville, Tennessee
http://www.geotolmeadows.com

MARCEL DEKKER, INC.

NEW YORK BASEL HONG KONG

 Page ii
Library of Congress Cataloging-in-Publication Data
Meadows, James D.
Measurement of geometric tolerances in manufacturing / James D.
Meadows.
p. cm. -- (Manufacturing engineering and materials processing ; 52)
Includes index.
ISBN: 0-8247-0163-1 (alk. paper)
1. Tolerance (Engineering). I. Title. II. Series.
TS172.M43 1998
620' .0045dc21 98-17836
CIP
This book is printed on acid-free paper.
Headquarters
Marcel Dekker, Inc.
270 Madison Avenue, New York, NY 10016
tel: 212-696-9000; fax: 212-685-4540
Eastern Hemisphere Distribution
Marcel Dekker AG
Hutgasse 4, Postfach 812, CH-4001 Basel, Switzerland
tel: 44-61-261-8482; fax: 44-61-261-8896
World Wide Web
http://www.dekker.com
The publisher offers discounts on this book when ordered in bulk quantities. For more
information, write to Special Sales/Professional Marketing at the headquarters address
above.
Copyright 1998 by Marcel Dekker, Inc. All Rights Reserved.
Neither this book nor any part may be reproduced or transmitted in any form or by any
means, electronic or mechanical, including photocopying, microfilming, and recording, or
by any information storage and retrieval system, without permission in writing from the
publisher.
Current printing (last digit):
10 9 8 7 6 5 4 3 2 1
PRINTED IN THE UNITED STATES OF AMERICA

 Page iii

Preface
Sometimes we learn best by seeing, reading, or hearing about the tribulations of others.
We analyze what they have faced, how they dealt with it, and whether they were
successful or failed miserably. As a consultant, I have visited hundreds of companies and
listened to thousands of problems, and then logged the best and the worst of the problem-
solving approaches used. This book relates many of the most interesting situations and
discusses relevant ethical concerns.
This is a technical book about measurement of geometric characteristics and how they are
practiced on the shop floor. But it is also a book about quality, company politics, and how
otherwise logical professionals can perform in the most illogical of ways when placed
under the pressures of too little time, too little input, and too much misinformation. It
shows how ego, bias, and vested interests can destroy a product and compromise a
person's principles.
In my travels I have seen that common sense is anything but common. Many of the
stories in this book are amusing and some downright unbelievable. But take my word for
it: I couldn't make this stuff up. Much of it is just too weird. Just be glad it isn't happening
to you. But if it isand it very well may beplease tell me about it. Someday, I may want to
write another book.
This is the third book I've had published by Marcel Dekker, Inc., and I must admit I
never thought it would come together this easily. As always, I remain indebted to Michael
Gay of Nashville CAD, Inc. for his great illustrations, Kay Du Vall for her help in putting
this book together, Jeannie Winchell for overseeing and guiding all of my projects since
the beginning, and my editors at Marcel Dekker, Inc.

 Page iv
I would also like to thank Mike Groszko of Chrysler Quality Institute for his assistance in
putting together the SPC unit. Finally, I would like to thank Carl Lance for all of the
GD&T insight he has given me over the last 20 years. He is truly one of the most
knowledgeable people I have ever met.
JAMES D. MEADOWS

 Page v

Contents
Preface iii
1. Introduction to Measurement Principles: Once Upon a
1
Timein the Land of Measurement
2. Geometric Product Definition and Measurement,
7
Inspection, and Gaging Principles
2.1
11
Virtual Condition Illustration
2.2
14
Verification of Position with Open Set-up
3. Reading the Feature Control Frame
3.1
19
Just Talk the Talk
3.2
21
How to Read a Feature Control Frame
3.3
How to Read a Feature Control Frame for Purposes of 23
Measurement
4. Functional Product Definition Creates Functional
33
Inspection
4.1
33
Inspecting Parts as They Are Used
4.2
34
Functionality Used to Suggest Manufacturing Procedures
4.3
Multiple Possibilities of Dimensioning and Tolerancing 36
for Functionality and Inspectability
4.4
38
I Can't Measure That!
 4.5 41
Functionality vs. Producibility
4.6
42
The Things I've Seen (Where Tolerances Come From)
4.7
Hog-Tied and Cow-Chipped (Importance of Geometric 44
Definition)

 Page vi
4.8
Correct and Complete (Tips on Functional Product 46
Definition)
4.9
Miles to Go Before We Sleep (Importance of Geometric 50
Definition)
4.10
52
Plastic Parts and Stretch Tolerances
4.11
Design, Inspection, Production, and Prototype Needs
54
and Capabilities Regarding Dimensioning and
Tolerancing
4.12
Implying a Manufacturing Sequence on Complex Part 57
Configurations
5. Principles of Measurement for Geometric
63
Characteristics
5.1
63
Geometric Characteristic Symbols
5.2
64
StraightnessForm Category
5.3
76
Flatness Form Category
5.4
79
Circularity Form Category
5.5
89
Cylindricity Form Category
5.6
93
Parallelism Orientation Category
5.7
97
Perpendicularity Orientation Category
 5.8 102
Angularity Orientation Category
5.9
106
Profile of a Surface Profile Category
5.10
Composite Profile vs. Two Single-Segment Profile 109
Controls
5.11
Coplanarity (Profile for Coplanar Surfaces) Profile 113
Category
5.12
117
Conicity (Profile for Cones) Profile Category
5.13
119
Profile of a Line Profile Category
5.14
126
Circular Runout Runout Category
5.15
128
Total Runout Runout Category
5.16
135
Concentricity Location Category
5.17
139
Symmetry Location Category
5.18
144
Inspecting Threaded Holes
5.19
150
The Boundary Concept
5.20
Composite Positional Tolerancing vs. Two Single- 154
Segment Positional Tolerancing Controls
5.21
160
Rules Sheet
6. Comparison of Controls 161
6.1
 Food Chain of Symbology 161

6.2
162
2-D vs. 3-D Controls
6.3
Comparison of Coaxiality Controls (Position, 162
Concentricity, Circular Runout, Total Runout)
6.4
166
Concentricity vs. Runout

 Page vii
7. Datums 169
7.1
When Sweet Seeds Grow Rancid Fruit (How to Select 169
Datum Features)
7.2
172
Part Stabilization (Datums for Flexible Parts)
7.3
Sequential Part Definition vs. Simultaneous 176
Requirement (The Game of Leaders and Followers)
7.4
178
Datum Features
7.5
185
Simultaneous and Separate Requirements
7.6
189
Datum Target Assignment
7.7
Oddly Configured and Curved Surfaces as Datum 196
Features
7.8
198
Secondary and Tertiary Datum Features of Size
7.9
Free-State Variation in Sheet Metal Parts/Fixturing 203
Sheet Metal
8. Material Conditions (Modifiers) Meaning and Uses 211
8.1
211
Maximum Material Condition
8.2
The Maximum Material Condition Symbol and Its 214
Ramifications
8.3
220
Allowed vs. Actual Deviation From True Position
 8.4 230
Least Material Condition (Meaning and Uses)
8.5
231
Regardless of Feature Size
9. Virtual and Resultant Condition Boundaries 235
9.1
235
Virtual ConditionA Functional Boundary
9.2
239
Surface vs. Axis Verification
9.3
241
Zero Positional Tolerancing at MMC for a Greater Yield
10. GagingPrinciples, Terms, and Examples 245
10.1
245
Gage General Principles
10.2
250
Gaging Size Limits
10.3
254
Terms
10.4
257
A Few Major Measurement Concepts
10.5
295
Gaging a Complex Sheet Metal Part
10.6
300
General Principles and Terms of Tolerance Compliance
11. Merging Theory with Reality 303
11.1
Merging Theory with Reality (Having the Right 303
Measurement Tool)
11.2
304
Theory vs. Reality (Uncertainties)
11.3
306
Reactions to a Non-Perfect Datum Establishment
12. How to Improve Quality 309
12.1
309
The Zen of Quality
12.2
311
Tying Measurements to the Machine
12.3
311
How Many Points Should I Probe?

 Page viii
12.4
The Search for Common Sense (Statistics Out of 311
Control)
12.5
Quality Issues from the Afterlife (Slogans vs. Programs 315
and Planning)
12.6
317
Focus and Excel (Maximizing Your Abilities)
12.7
Competing in a Global Economy (Great Products at an 319
Affordable Cost)
13. Inspection Planning 321
13.1
321
Dimensional Inspection Planning
13.2
331
Methodical Measurement
14. Source of Measurement Error 343
14.1
What We Know Is Often Untrue and Other Sources of 343
Measurement Error
14.2
344
CMM Error Compensation
14.3
347
Certification vs. Calibration
15. Graphical Inspection Techniques and Examples 351
15.1
351
Graphical Techniques and Inspection Reporting
15.2
Comparison of Measurement Techniques and 375
Equipment (Advantages and Disadvantages)
16. Teamwork Issues 381
 16.1
381
Unity of Purpose Affects the Bottom Line
16.2
385
Troubled Times and Technical Leaps
16.3
387
Concurrent EngineeringKey to Our Future
16.4
Magical, Mystical Stuff (Doing Something vs. Doing 389
Nothing)
16.5
Avoiding Tiger PitsWhile Helping a Leopard Change Its 391
Spots
17. Training Issues 395
17.1
395
I Wonder as I Wander
17.2
396
Training: An Ongoing Process
18. Statistical Process Control 399
18.1
399
Faulty ImplementationWonderful Concepts
18.2
405
Concepts of Statistical Process Control
18.3
414
Calculating Percent of Nonconforming
18.4
421
Statistical Tolerancing
18.5
423
Glossary of Terms
18.6
426
Symbology for SPC Formulas
19. Letters: Common Problems and Their Solutions 429
Index 451

 Page 1

1
Introduction to Measurement Principes
Once Upon a Time in the Land of Measurement
In the history of the world of manufacturing, the shop floor has often been tasked with
doing quick, rough checks on the products undergoing production. These measurement
procedures sometimes were visual checks by the manufacturer to determine flaws in the
parts. When this was insufficient, instrumentation was employed, but usually of a less
sophisticated nature than would perhaps later be utilized in the inspection lab. Still, some
of the earliest and most effective gages were mating parts that had already been produced
and were plucked from a barrel to serve as a test of functionality. The part just finished or
nearing completion was tried on the potential mating part. If it assembled, it was assumed
the newly formed part was functional and, therefore, acceptable.
After a time, it was realized that this type of test did not assure that the part would
assemble with any of the mating parts plucked from the barrel, but perhaps only with
someor even one. This lack of interchangeability was sometimes unacceptable. We were,
in these instances, creating matched sets of parts. What was often desirable, instead, was
the creation of a part that would assemble with any of the mating parts produced so that,
for example, a person could take a broken part in his automobile or tractor into the parts
store and buy an off-the-shelf replacement.
At some point, it was decided a gage that represented the worst condition of mating parts
would be needed to gage the capability of newly produced parts to fit into the assembly
and be compatible with any part already in place. These gages are commonly called
receiver-type or functional gages. Some of the initial gages were meant to inspect only
size requirements. They are still used to this day. They are called GO gages. GO gages
check maximum material conditions and, in doing so, contribute to checks of geometric
form. They are designed to go over the largest shafts or into the smallest holes allowed by
size require-

 Page 2
ments. A NO-GO gage is often used or simulated to inspect least material conditions.
Their job would be not to go over the smallest cross section of the shaft or into the largest
cross section of a hole. If they go, it is assumed size at least material condition has been
violated. The same as if the GO gage refuses to go over the entire shaft or into the entire
hole, it is assumed size at maximum material condition has been violated.
But, as we all know, there is often more to a feature reacting functionally than just its size.
Functional gages check geometric characteristic requirements, such as perpendicularity,
position of features and patterns of features, such as holes, slots, shafts, tabs, and other
features of size. They check to determine whether the boundary generated by the
collective effect of worst case size and applicable geometric tolerances has been violated.
These boundaries are often constant boundaries that exist between mating features,
confining the shaft within and keeping the surface of the hole outside of it. It is assumed
that if the hole never moves inside of this boundary, it never occupies space meant for the
shaft. And it is also assumed if the shaft never moves outside of the boundary, it never
occupies space reserved for the hole. Never crossing this line of the worst case mating
boundary by either assures us these features will assemble without interference.
The problem with these gages has always been that because they employ no probes or
indicators, just hard planar rails, shafts, holes, and other physical configurations meant to
be the physical embodiment of a worst case scenario, they give little or no information
except whether or not the part will fit into the assembly. They don't say how good or bad
a feature is, or in what direction it might be straying. This information is often considered
vital to improving a manufacturing procedure by charting or graphing deviations from
geometric perfection. These pieces of variables data are regularly employed to statistically
control a manufacturing process. So, other mechanisms of inspection are sometimes
needed to collect this information, not to replace, but to augment, the attribute (good/bad)
information given by functional gages.
These collectors of variables data, such as dial indicators and electronic probes, tell the
story of a feature's deterioration from geometric perfection. From this, it is possible to
determine that within a large run of parts critical features are getting worse as the run
continues and, if allowed to resume in that manner, will eventually exceed the boundaries
of functionality and no longer assemble or operate.
With the proliferation of computers, we have connected many of our variables data
collectors directly to them or fed the information collected into computerized programs
written to analyze the data. But, with these came another source of error. The algorithms
that are written aren't always correct. This allows data to be improperly analyzed,
sometimes causing functional parts to be rejected and nonfunctional parts to be accepted.

 Page 3
Controversy also ensues as to the use of touch probes. Optimal probe sizes, the number
of points to take to represent a surface, and the distribution of these points on the part are
areas of controversy. Then, scrutiny of what the software is doing with the collected
points becomes acute. Many times, the software is supposed to duplicate what the parts
rest on or clear in an assembly but, instead, computes best fit surfaces. This means that
high points that are critical to determine assembly capability are treated as no more
important than low surface points. It means that origins of measurement, often called
datums, are formed, not from the high points of surfaces but, rather, as best fit planes or
axes from all points collected. Since high points protrude farther than these best fit
simulations, parts are often accepted that have no chance of assembling, hanging up
instead on the high points.
Fixturing would help. Placing parts on fixtures as simple as surface plates and angle plates
would allow inspectors to probe them instead of datum surfaces. And, since they rest
against high points on the parts, software problems would be lessened. But many are told
that because a computer is involved, all possible problems will be solved inside the
mysterious world of software algorithms, without fixturing, and to just trust the
computerwhich many do, and they suffer for their trusting nature.
In the late 1980s, a group of quality engineers from Westinghouse Corporation sent out a
cry of alarm to the U.S. military and later to others who would listen. They chronicled
their finds of unreliable and nonrepeatable measurements taken with computer-aided
measurement machines, pointing out errors in a variety of areas. It was discovered that
software was written incorrectly by programmers well schooled in writing software but
either unaware of or unconcerned about technical problems found in their software. It
was also discovered that measurement machine manufacturers were giving instruction to
users of their mechanisms that exacerbated the problems. Too few points taken to
establish representative datum planes and axes were suggested and fixturing was often
discouraged as expensive, time-consuming, and unnecessary. Too few points and/or an
unwise distribution of points taken to represent surfaces under test drove the problem to
the brink of, in many cases, disaster.
When the military took the problem very seriously, even offering to shut down every
such machine doing work for them, another cry, equally desperate, was heard. This came
from those with a vested interest in the status quo. Greed and pride prompted many to
attack the findings, downplay the problems, and denigrate the alarmed. Many of those
crying the loudest were powerful and tried to destroy all efforts to cure the problems,
even to the point of suggesting we change the definitions of datum establishment and
product inspection to simply say what the flawed software and measurement machines
were doing was correct.
To a large part, both sides of this battle were successful. Many wanted to

 Page 4
believe all things were right with the world of measurement, and those with a prideful or
monetary interest in perpetuating this falsehood were ready to tell them what they wanted
to hear. On the other side, ANSI committees were formed to resolve the problems.
Committees on correct mathematics, measurement methods, sources of common errors,
functional gaging, and fixturing were formed and began to try to make things better.
In the end, it was discovered that those who wanted to listen did and those who found it
easier to belittle the efforts or divert them from their intended path set about doing exactly
that. Perhaps it was because some of the problems seemed unsolvable, or because some
of those involved in collecting points could not accept a merging of strengths between
gaging disciplines (for example, some fixturing with some probing), that this story has no
happy ending. Maybe it was just that once started down the path of ill-conceived
procedures some were unwilling to consider another tactic.
As far back as 1988, Russ Shelton and Klaus Ulbrich of EMD performed tests whose
results were shocking to many and ignored by others. The response to their findings
seemed to be that no action was taken to correct the problems raised. Touch-trigger
probes for data collection are used by about 90% of the world's CMMs. These probes are
slow and plodding in their efforts to collect enough points to represent a surface. They
lower, hit, raise, move, lower, and hit again. Analog probes or drag probes (as they are
sometimes referred to), which are faster at collecting a large sampling of surface points,
are expensive and most CMM users can't afford them. Therefore, the touch probe users
continue on in the mistaken belief that they are getting a representative sample of a
surface by taking relatively few points, which is all they have time to take given the
technology with which they are stuck. Shelton and Ulbrich tested six sample surfaces to
construct a plane. Each was about one square inch. All were lapped aluminum. They
represented many different deformities such as gouges, concave and convex surfaces, etc.
Each was scanned in a grid that ended up with about 2500 evenly spaced data points. I
quote from them: Depending on the specific sample, it took between 80 and 357 points
(average 191) per square inch to achieve a 90 percent certainty of form, and between 48
and 227 points (average 102) per square inch to achieve a 75 percent certainty of form.
Compare this to the number of points commonly probed by the average inspector to
validate or represent a surface and the horror of our current situation should begin to sink
in.
Although there is more information available now on these problems than ever before
and more solutions are available, it is also true that the spreaders of repackaged, badly
flawed software procedures and expedient, poorly thought out recommendations on
general set-ups and tool and equipment usage still abound, even flourish, to this day.
Wonderful equipment and software, capable of evaluating data correctly, are often
ignored in favor of a flashier name and a better sales pitch. Buyers are commonly
uneducated as to what measurement mecha-

 Page 5
nisms and their accompanying software should be doing, so they fall prey to a smooth
line and a knife-bearing pat on the back by a slick salesperson who says Buddy and Pal a
lot.
We are still in a tub of trouble out here. Our systems are full of holes and we're looking
left when we should be bailing with an unrestrained fervor. But it is hard to create zealots
for change when the general consensus is that we're sliding along in greased grooves.
Those who experience the world of politics and cover-up with an acute feeling of nausea
are seen as relatives of Chicken Little shrieking that the sky is still falling!
Many give up, battered and bruised on the outside, with a pure heart still beatingbut only
faintly audible. Perhaps it is true that they shoot the messengers of bad tidings and that it
is easier to entertain the troops with a sweet song and a fast shuffle than to end the war.
One thing is certain in the land of the blindthe one-eyed man is not king; he is shunned
because he is different.

 Page 6

 Page 7

2
Geometric Product Definition and Measurement, Inspection, and Gaging
Principles
Let us start with a comparison of the plus and minus system for location and what has
often been called true positioning.
In Figure 2-1, this simple part uses plus and minus tolerancing to locate the

FIGURE 2-1

 Page 8
hole. In reality, we cannot be certain whether the hole is to be measured from the edges of
the part, or the edges of the part are to be measured from the hole. This ambiguity could
cause difficulties in determining set-ups for both manufacturing and inspection. It also
makes it difficult to discern the exact configuration and location of the tolerance zone. If
the origin of measurement were to be the axis of the hole, then tolerance zones could be
assumed to contain the edges of the part and perhaps even be perceived to control the
location and profile/form of the two edges.
For this example, however, even though that interpretation could be easily argued as
valid, let us assume the hole is to be measured from existing edges of the part. If so, the
drawing says we would like the hole to be 25 mm from one edge and 90 mm from
another. Knowing we are imperfect beings and need a tolerance on that, we are allowed a
plus and minus tolerance in each direction of 0.2. This generates a 0.4 square tolerance
zone in which the actual axis of the hole must lie. The tolerance zone is centered around
the mean dimensional location of the 25 and 90 mm specifications.
We can see the tolerance zone is square, but the hole (in two-dimensional terms) is round.
The tolerance zone is not reflective of the shape of the hole it protects. Curious. However,
we know this is often done because of part function, so let us examine the mating part. It
is a rectangular plate with a shaft diameter mounted on it (see Fig. 2-2).
The mating situation is simply that we would like the shaft to fit into the hole with the
seating perpendicular surfaces firmly against one another and the edges of the parts (from
which the 25 and 90 mm dimensions emanate) to line up flush. That would not dictate a
square tolerance zone. Sometimes a tolerance zone is created as nonreflective of hole
shape to protect wall thickness where a too thin wall is a possibility from the hole surface
to the outside of the part. But with only a 27.5 to 28.5 diameter hole and a 0.4 square
location tolerance zone,

FIGURE 2-2
Mating part.

 Page 9
one can easily see the shortest distance of 25 millimeters of an inch is not capable of
creating so thin a wall that the hole weakens the part to a point of danger.
So, there is no logic in having a round hole with a square tolerance zone considering the
difficulties that zone places on us in tolerance analysis, worst case boundary calculations,
and functional gage design.
Let us examine some of the decisions that would have to be made if the hole were drilled.
If a machinist were asked to drill the hole, one would have to choose what surface to lay
the part down on first to come in contact with the machine (drill press) table. The
drawing allows us to choose one of at least two. The primary surface strikes a minimum
of three high points of contact and eliminates part movement in two rotational and one
linear direction.
Since the part can still move, the machinist slides a rail or angle plate against another
surface chosen from at least two surfaces and slides into two point minimum high points
of contact to eliminate two more degrees of freedom (one rotational and one linear).
Whatever surface is left is used as a tertiary feature requiring a one point minimum high
point of contact to eliminate the last degree of linear movement. The machinist then
clamps the part into that orientation, measures from those rails and/or angle plates 25 and
90 mm and drills the hole. The problem occurs when the part must be checked. There is
no repeatability factor here.
When the parts inspector gets this part and must set it up to inspect it, he or she must
make the same choices as did the machinist. Suppose he or she chooses differently. How
many times and in how many different part orientations must one set the part to
determine that the part does not meet the drawing requirements? If the inspector sets it up
the first time and it checks good, the part will most likely be accepted as within tolerance.
Still, since the drawing does not indicate how this part seats in the assembly, there is no
assurance that it has been inspected in a functional manner. It may check within tolerance
but not assemble since it may not have been inspected in the same manner that it fits into
the assembly.
But, if the first set-up allows the part to check bad, the inspector might have chosen
different primary, secondary, and tertiary surfaces than the part uses in the assembly. The
part might check within specifications with a different set-up (part orientation). So, the
inspector resets it and resets it, knowing each different set-up might be the one that allows
the part to be bought off as good. Each set up takes time, and time and money are being
wasted.
If the designer had used some simple criteria to choose the set-up surfaces, the machinist
and inspector would have had no choices to make. The drawing could have reflected the
designer's knowledge of datum feature selection criteria (for example, part function and
representation of the mating situation).
The same part using geometric dimensioning and tolerancing might appear as in Figure 2-
3. With this drawing, the primary datum feature is A (3 point high point minimum
contact), the secondary B (2 point high point minimum contact

 Page 10

FIGURE 2-3
while maintaining A), the tertiary C (1 point high point minimum contact while
maintaining A and B).
We know this is not because that is how they appear in the alphabet, but rather because
this is how they appear in the information block (called a feature control frame) read
from left to right. It is located directly below the size limits of the hole so as to localize
much of the information about the hole.
It tells us we want to position the hole so the position symbol is the first piece of
information given. The tolerance zone is described as a diameter so as to reflect the hole
shape. The 0.4 square has been replaced by a 0.4 diameter tolerance zone in which the
actual axis of the hole must lie. This is a design change surelybut one that will work
because the mating shaft will be dimensioned and toleranced in a way to allow it to mate
with this new requirement.
Although initially a smaller tolerance, the M in a circle will allow more as the hole grows
from 27.5 toward the 28.5 size limit. The datum features assure more repeatability and,
together with the potentially larger tolerance, make the part more cost effective to produce
and inspect. A 28.5 hole is allowed 1.4 position tolerance.
Datum feature A locks in perpendicularity of the tolerance zone to it and, therefore,
protects perpendicularity of the hole. This is a multipurpose control because, besides
datum A controlling perpendicularity, B and C locate the toler-

 Page 11
ance zone axis (25 exactly from B and 90 exactly from C) and, consequently, specify the
hole axis's perfect location. The dimensions from those surfaces are boxed to indicate an
exact specification.
Now we know that if we were able to make the hole perfect, it would be perfectly
perpendicular to the plane formed by the three highest points of surface A and 25 mm
from the plane formed by the two highest points of B and 90 mm from the plane formed
by one point of high point contact from C. Each plane is dependent upon and mutually
perpendicular to the other two, forming what is commonly called a datum reference
frame. This datum reference frame eliminates the 6 degrees of part freedom (previously
described), orients the part, and positions the hole for repeatability.
Since the hole can't be made exactly where the datums and basic dimensions indicate, the
tolerance zone of 0.4 is used to allow hole axis deviation from perfection, or more than
0.4 as the hole grows. The reason the circled M was used was a recognition of reality in
this situation. The reality is simply that if the hole moves and/or leans 0.4 at a diameter
of 27.5, the size shaft that can fit into that hole (in a manner described in the mating part
drawing and mating situational requirements) is 27.5 minus 0.4 or a diameter of 27.1.
So 27.1 is the boundary we are protecting, and it isn't just important some of the time on
some of the parts, but rather all of the time and on all of the parts. A combination of the
smallest hole size and the allowed out-of-perfect perpendicularity and location on this
part reduces its effective mating size. That's why we must design the parts so the hole
never infringes on a boundary of 27.1 that belongs to the mating shaft, and the mating
shaft never infringes on a boundary of 27.1 belonging to the hole. This complied with,
material never interferes.
So, if the hole was produced at a diameter of 28, we could allow the positional tolerance
on the hole location for that produced part to grow (by virtue of the circled M) to 0.9,
because 28 minus 0.9 still protects a 27.1 boundary. At 28.5, we could allow 1.4, because
28.5 minus 1.4 is still 27.1. The part is allowed more hole positional tolerance based on its
actual produced size. It is all functional and very cost-effective, because a hole allowed to
move or lean away from its perfect location has the effect of reducing the area available
for a shaft to insert into at that location.
2.1 Virtual Condition Illustration
This virtual condition boundary is generated by the hole's size and geometric
imperfection. As the hole moves away from its perfect position or leans out of perfect
perpendicularity, the area available for a mating pin to go into the hole at that location and
angle is diminished, creating a virtual condition boundary. This boundary is smaller than
the smallest hole by the amount of the allowed positional tolerance. Therefore, the pin
must be smaller than or equal to the virtual condition of the hole, if it is to be inserted
without interference under all circum-

 Page 12

FIGURE 2-4

FIGURE 2-5
stances. This example (Fig. 2-4) shows the collective effect of the actual size and the
geometric tolerance applicable at that size for a hole controlled for position at maximum
material condition (MMC). This is the worst mating boundary.
By the same token, a mating shaft made at 26.7 diameter would only have available to it a
positional tolerance of 0.4 before it approached the worst mating boundary of 27.1,
but at 25.7 it would be allowed 1.4 (provided it is made within its size limits) because
25.7 plus 1.4 is 27.1. Shafts, and for that matter all external features of size that move or
lean away from their perfect locations, have the effect of increasing the area needed for
the shaft to insert into at those locations.

 Page 13

FIGURE 2-6
One possible mating part design appears in Figure 2-5. This is just one example. This text
shows many techniques to assist you in the usage of the ASME Y14.5 standard (geometric
dimensioning and tolerancing).
Figure 2-6 is a possible inspection procedure performed on a part produced for the
specification shown in Figure 2-3.
With the information collected as shown above we can deduce in two different ways that
the produced part meets its positional tolerance.

 Page 14

2.2 Verification of Position With Open Set-Up

This section will explore one often-used method of verifying the location of positionally
toleranced features on parts, such as shown in Figure 2-7.
Before the part is inspected for the shaft's compliance with its positional control, size
requirements should be verified. Least material condition (LMCsmallest shaft size) should
be checked at every two opposing points, such as a micrometer-type measurement would
accomplish. Maximum material condition (MMClargest shaft size) should be checked for
violations of a perfect cylindrical envelope, such as a GO gage would accomplish. Then,
if functional gaging is not available or variable data are required, many measurement
techniques may be employed to inspect the position of the shaft. A computerized
coordinate measuring machine (CMM) is a tool commonly used for this purpose and will
be discussed later in this book.
A surface plate type set-up (often referred to as an open set-up) may also be used.
Because of the MMC symbol next to the geometric tolerance, the feature mating
sizesmallest cylindrical hole that will fit over the shaft without regard to orientation or
locationmust be determined. This can usually be accomplished quite simply through the
use of fine or adjustable gages. This minimum circumscribed cylinder size will be used as
the factor in determining bonus positional tolerance to be added to the original 0.4
diameter positional tolerance. Let us assume the feature mating size is 12.2.

FIGURE 2-7

 Page 15
The part can then be positioned into the datum reference frame. For example, it would be
nice if datum feature A could be pushed against an angle plate of sufficient size to allow
feature A to establish the minimum of 3 high point contact to simulate datum plane A.
But, of course, if the datum feature is on the same side of the part as the shaft being
inspected, that would not be an easy thing to do. So, if that is the case, the primary datum
feature may be established by clamping the opposing side of the part to the angle plate
and, through the use of shim stock or leveling screws and a dial indicator, the datum
feature can be indicated in. In other words, if the datum feature is inaccessible to mount
on, the shim stock (by placing varying thicknesses of it between the part surface and the
angle plate at appropriate locations) can be used to make the datum feature parallel to the
surface of the angle plate. This will have the effect of negating (within a range) the out-
of-parallelism between the surface that is accessible (and on which you will mount the
part to the angle plate) and the datum feature.
While maintaining this contact and orientation, the angle plate could be placed on a
surface plate or machine table and datum feature B brought into a minimum 2 high point
contact with this surface to simulate datum plane B. Another plate could then be brought
into contact with 1 high point of datum feature C to position the part so that datum plane
C is not only established, but established 90 to an edge of the first angle plate and 90 to
the surface plate (or machine table) on which the part was mounted. This allows us to
rotate the angle plate 90 and have datum planes B or C simulated by the surface plate (or
machine table) on which the angle plate (which holds our part) is resting.
Let us assume we have done this and our part is oriented in its simulated datum reference
frame. A height gage with an affixed dial indicator can now be used to determine the
location of the shaft from simulated datums B and C.

StepRotate the angle plate so that datum plane B is simulated

1: by the surface plate.
StepZero out the dial indicator on the surface plate and
2: record the vernier reading from the height gage.
Raise the height gage indicator to the top of the shaft
and run the indicator over the shaft as close to the part
as possible. Keep adjusting the gage height until the
same zero is recorded on the top of the shaft by the
Step
indicator. (We are looking for the deviation from the 38
3:
basic dimension but must laterin Step 5take into
consideration the pin diameter.) Take the reading of the
 height of the top of the pin from the vernier scale on the
height gage.
StepSubtract the reading taken in Step 2 from the reading
4: taken in Step 3.
StepLet's assume the difference in those readings equals
5: 44.4. Sub-

 Page 16

FIGURE 2-8

tract from this number one-half of the diameter of the

shaft (minimum circumscribed cylinder).
12.2 divided by 2 = 6.1 Thus, 44.4 minus 6.1 = 38.3
We know we wanted to be at 38 but ended up at 38.3, a
deviation of 0.3. Record this 0.3 deviation.
Rotate the angle plate 90 so that datum C is now
simulated by the surface plate. Repeat Steps 25 to
Stepdetermine the distance of the maximum deviation of the
6: shaft axis from datum C. Let's assume the worst distance
of the axis from C is 64.4a maximum deviation from 64
of 0.4
We now know the bonus tolerancethe difference
between MMC (12.3) and its actual feature mating size
Step(12.2)is 0.1. Since we started with 0.4, this gives us a
7: total of a 0.5 diameter positional tolerance zone allowed.
We also know the actual deviation from true position is
0.3 from datum B and 0.4 from datum C.

 Page 17

FIGURE 2-9
Since we are allowed to be out of position a diameter of 0.5 and are actually out a little
more than a diameter of 1, the feature is rejected as produced. If it had checked good, this
process would also be performed at the other end of the shaft (farthest from the primary
datum) and, if time permits, at various places in between.
When the indicator is brought into actual contact with the feature surface, rather than in
contact with a tool representing mating size, surface variations must be taken into
consideration. This direct contact type of check is sometimes made more valid if
differential readings (180 from one another) are taken down the shaft and compared to
determine the actual derived axial location. If any surface variations on the shafts or holes
being checked are complex, this direct contact type of inspection technique can be time-
consuming and difficult.
A similar procedure (such as described) will work for holes as well as shafts, and with
slight modification is also accurate for other features of size such as elongated holes,
slots, tabs, bosses, etc. Coordinate measuring machines (CMM) make this procedure
easier and less time-consuming while using the same or similar basic steps. A functional
gage may be used to collect attribute (ac-

 Page 18
cept or reject) data and is even faster than a CMM, but this type of receiver gage will not
collect variables data, as will the CMM and the surface plate set-up.
However, soft functional gages, which use data collected by (for example) a CMM, can
give variables data concerning the direction and amount a feature has deviated from
geometric perfection. These soft gages, which are computer generated, will be discussed
more later in this text, as will the CMM and other measurement machines.
[Inspection Note for Holes:
The maximum inscribed cylinder minus the maximum inscribed cylinder with perfect
perpendicularity equals the out-of-perpendicularity of a produced hole axis. Also, the
maximum inscribed cylinder minus the maximum inscribed cylinder with perfect position
equals the out-of-position of a produced hole axis.]

 Page 19

3
Reading the Feature Control Frame
(For a Functional Understanding and to Guide Product Measurement)
3.1 Just Talk the Talk
Many people read only the symbols and not their implications. Without reading in the
implications, it takes longer to understand what the control is trying to convey. I have
some guidelines that I use when creating or interpreting a drawing, and generally they
hold up if others involved in the process use similar techniques.
1. I read the diameter symbol as of the axis or the axis may be out of; therefore, a control
that starts out is read position of the axis or if you start with the diameter symbol and
read backward, it would read: The axis may be out of position.
2. The next thing one must understand is that the feature control frame is tolerancing. The
dimensioning is somewhere else, usually also on the field of the drawing, but a
completely separate component of the drawing. The dimensioning states the desired
relationships, but the tolerancing states the allowed deviation from those relationships.
More specifically, the feature control frame is there to state the allowed deviation from
what would be considered geometrically perfect. The next itemthe tolerancedoes this. For
example: says: The axis may be out of position a diameter of .030.
3. Now it is time to state or imply at what size the feature or features may be out of
geometric perfection (in this case, out of perfect position). could be read The axis
may be out of position a diameter of .030 if the controlled feature or features are
produced at Maximum Material Condition. Of course, this implies that if the features of
size being controlled depart from maximum material condition (MMC), staying within
size limits, they

 Page 20
would be allowed to be out of their perfect position (move away from their true position)
more.
Use of the is an indicator that the feature being controlled mates with something. It
states that an alteration in size away from the MMC indicates an additional tolerance is
called for, an excellent signal that the feature mates. That is when that statement would be
true. If an , or least material condition (LMC) symbol, had been used, that would have
been an excellent indication of a need to preserve material as in situations where the
designer is concerned primarily with wall thickness problems that might occur. It is also
used on casting drawings when it is a concern that there be enough material on the casting
to be able to clean up in subsequent machining operations that may be called for on other
drawings.
In both of these situations, the (which may be implied by omission), or regardless of
feature size (RFS) symbol, may have worked, but would, by virtue of the fact that it
allows no additional bonus tolerance to be drawn from size limits, have made the product
potentially cost more. In fact, the circled M and L don't normally improve the function of
a product. They are used in place of the circled S almost exclusively to reduce product
cost. And as long as functionality is preserved, the use of the bonus tolerance symbols is
a great idea.
There are times when the use of anything other than the RFS concept would endanger the
ability of the product to function. For example, if balance is a concern, the use of the
circled S symbol (especially after any datum features of size) is a good place from which
to begin your final decision. This decision about whether or not to use the circled S is
often based on how fast the features spin, how many are being made (because in a large
run of parts it is often desirable to use functional gaging and the use of the RFS
symbology makes that very difficult), and the material being used, among other factors.
For me, the use of is an indication that this feature either does not mate with
anything or is a spinning part or both. The circled S after the .001 says that this feature
probably does not mate with anything, unless what it mates with expands into or onto it.
The circled S after the datum feature (B) indicates that (a) the part spins and I am
concerned with balance, or that (b) I do not want inspection to buy parts that can only be
assembled off center (again maybe a balance item), or that (c) datum feature B does not
mate with anything, or that (4) maybe B does mate with something, but that mating item is
a completely separate component that would have no effect on the location of the feature
being toleranced by this control.
4. Next comes the datum feature(s). First, it is important to understand the nature of the
position symbol. If there is more than one feature in the pattern being controlled (a
pattern of two or more holes, for example), then the position symbol itself begins by
controlling the relationship between those holes. applied to such a pattern
already has the ability to control the posi-

 Page 21
tion between the holes to within that tolerance per hole. In other words, a basic
dimension of 1.500 between two holes within a pattern would have a range of deviation
similar to 1.500 plus or minus .030, to put the hole axes anywhere from 1.470 to 1.530
from one another.
If looked at in this context, datums are an extra added attraction, to lend orientation or
additional location from part edge planes or other datum axes or centerplanes generated,
for example, by holes or slots. So we say, in addition to being positioned to one another,
to what else are the holes positioned? could be read: The hole axes may be
out of position to one another and datums A, B, and C a diameter of .030 if they are made
at MMC. This also implies that any hole not made at MMC but within size limits may be
out of perfect position more than .030, and that the holes probably mate with shafts. If the
hole axes are meant to be (drawn) perpendicular to datum plane A, then the same .030
tolerance zones that were controlling only the hole-to-hole tolerance are now oriented
perfectly perpendicular to datum plane A. If the hole axes are within those dimensions
given from each to the hole pattern, then the .030 tolerance zones are further restricted as
stationary with their centers at precisely that location. If the hole axes stay within their
respective tolerance zones, then they are not only in the allowed positional tolerance from
one another (and the allowed perpendicularity tolerance to datum plane A) but now
finally within the allowed positional tolerance from datum planes B and C.
This really is a universal language of symbols that, if learned properly, can be used to
communicate with other interested parties, in other departments, other companies, and
even other countries. It also will open up to you a whole array of other augmenting
documents about product measurement, mathematics, software writing, manufacturing,
and quality techniques that have either been written around this Dimensioning and
Tolerancing standard or use it as an integral component of recommended procedures
therein.
3.2 How to Read a Feature Control Frame
Symbology contained in a feature control frame can convey information vital to the
functionality and measurement of a part. To understand this symbology, one must not
only be able to read the individual name of a symbol but also understand the implications
and ramifications of the symbols when strung together in sentence or paragraph form.
This section begins to explain some of the vital nuances in reading part functionality from
a feature control frame.
Just as we have replaced many of the words on signs with symbols, we are replacing
many of the notes on our engineering drawings with symbols. These symbols can,
indeed, be read as one would read a sentence to describe how a part functions or is to be
measured or made. For example:

 Page 22

FIGURE 3-1
This feature control frame can be read, starting with the diameter sign:
= The axis may be out of
= position
Then, reading the diameter sign again:
= a diameter of
.010 = .010
= if produced at maximum material condition (a diameter of .250)
A = to A for perpendicularity
B = and B for location (holding the .875 dimension)
C = and C for location (holding the .750 dimension)
So, as a complete sentence, it may read, The axis of this hole may be out of position a
diameter of .010 if the hole is produced at a diameter of .250 holding perpendicularity to
datum A, and location to datums B and C. It may also be read, Position of the axis of this
hole must be held to A for perpendicularity and to B and C for location to within a
diameter of .010 if the hole is produced at .250. It also implies some important
information. It implies that this hole may be further out of its perfect position if the hole
grows. For example, .011 would be allowed for a hole made at a diameter of .251 and
.012 for a .252 hole.
This is so because the hole uses the MMC symbol. The MMC symbol allows the
geometric tolerance of location to grow as the hole grows and thereby

 Page 23
creates a reduction in part cost by the acceptance of more of the parts that have been
produced. This increase in geometric positional tolerance is only allowed because the
functional requirements of the part are not endangered by this tolerance zone growth.
This is based on the information that using the MMC symbol implies. It says that a larger
hole has to be less perfectly located than a smaller hole. This is most often true when a
hole mates. A shaft will more easily fit into a larger hole; therefore, the location of that
larger hole can afford to be less perfect.
In fact, we may choose to read the control in an entirely different manner because of the
circled M in use. For example, we might read it as: Position of a diameter that mates while
the part is seated on A and the hole is located from B and C. And, if it does indeed mate,
we have to assume we are trying to preserve a mating boundary. That mating boundary is
the collective effect of the actual size of the hole and the positional imperfection allowed
to it.
For example, when a hole moves or leans out of its perfect location in space, its available
area for mating at that location and angle closes up by the amount that it has moved or
leaned. Since we cannot allow it to infringe on the boundary in which the mating shaft
will reside or interference may occur, we protect that mating boundary (virtual condition-
MMC concept).
So, a hole produced at .250 may move or lean only .010, while a hole produced at
.255 may move or lean out of its perfect location .015 because we have determined the
mating boundary to be protected to be .240 and:

.250 - .010 = .251 - .011 = .252 - .012 =

.240 .240 .240
.255 - .015 = .258 - .018 = .260 - .020 =
.240 .240 .240

Therefore, we have determined that if the holes produced at these sizes are out of their
perfect location by no more than an amount that would allow them to mate with a .240
or smaller shaft, they will function and should be accepted by inspection as being within
their functional limits of hole location.
This is only a bare-bones sketch of all the information that can be packed into one of
these controls.
3.3 How to Read a Feature Control Frame for Purposes of Measurement
Figure 3-2 is a good one to discuss the way to measure a part for functionality. The
feature control frames listed each serve a purpose to preserve this functionality. Each can
be read as an integral part of how the part works. If one control is not met, we can
determine its negative effect on how the part interacts with other parts or the next part it
interacts with. And, we can be reasonably certain that it does interact with other parts in
an assembly by reading the progression of controls given to this figure.

 Page 24

FIGURE 3-2
For example, a flatness control has been assigned to the primary datum feature. It can be
speculated that this feature is a seating surface and that the flatness control is assigned to
stop this part from rocking too much in the assembly. It may also be speculated that the
surface it seats on has been assigned a flatness control of a similar value for the same
reason.
If this is true, we know the two parts may rock on one another a maximum of the sum of
their flatness tolerances. This sum would optimally provide no rock at all; but, as one can
see, a flatness tolerance of zero would not be attainable, so a tolerance of .001 has been
assigned to this part. This leads us to surmise that the sum of flatness allowable between
the two seating surfaces may be about .002. When tolerancing a part, a balance must be
struck between the perfect geometry desired and the costs associated with any tolerances
assigned. When a tolerancing engineer assigns tolerances, it should be with the overall
product cost in mind and how much each control will add to that cost. Too much
tolerance and the part may not function. Too little tolerance and the part perhaps cannot
be produced to comply or, if it can, the cost to do so may be so prohibitive that no
customer is willing to pay for it. So, to be practical, we settle for what will function
within acceptable parameters.
The datum plane for this primary datum A is struck ideally from the surfaces' three
highest points of contact. In this case, when inspecting the flatness control, the
measurement is to confirm that the datum feature surface low points and the datum plane
struck from its three highest points of contact do not depart

 Page 25
from one another by more than .001. If they do not, it will tell us that this surface resting
on that plane is stabilized to within .001, doesn't rock more than that, and is in compliance
with the flatness tolerance.
The second geometric control in the sequence of controls that define this part is the
perpendicularity control. This control is an axial control with a cylindrical tolerance zone
that is to enclose the axis of datum feature B. Datum axis B will then be struck from the
datum feature, but the axis will be struck as perfectly perpendicular to datum plane A.
Then this axis will act as the origin of measurement for the five-hole pattern.
We could read this perpendicularity control as the axis may be out of perpendicularity a
diameter of .004 if the feature is produced at maximum material condition to datum plane
A. The implication is that if the feature is produced within size limits but at a size less than
2.990 (MMC), the tolerance zone would grow in size.
The tolerance stated in the control is a minimum. It isn't a maximum or even an average
geometric tolerance for a produced feature. The feature, if produced at least material
condition, would receive a tolerance of a diameter of .034 for perpendicularity of feature
B to datum plane A. This is based on the functional reason for such a tolerance. This
reason is easy to discover if the control is read in a different way. The second way is
based on the concept that the MMC symbol is used after a geometric tolerance, most
commonly if that feature mates. That being the case, the control could be read that we are
holding perpendicularity on a diameter that mates while the part seats on datum feature A.
If we were to read that into the control, we could extend the logic even more. If datum
feature B mates while the part seats on A, then what does it mate with? Datum feature B is
a shaft. A shaft mates with a hole. The hole that accommodates B and has a bottom that
interfaces with A is a cavity of sufficient size to accept B in its worst case allowed
relationship to A. That case might be considered to be B produced at its largest size and at
its most allowed out-of-perpendicularity to A at that size. That size is 2.990, and that
perpendicularity tolerance is a .004. The worst case condition is the sum of those
numbers, or 2.994.
However, the same worst case condition would occur even if datum feature B was
produced at its smallest allowable size of 2.960, but experienced its full allowed
perpendicularity tolerance to A at that size. For example:
At 2.960 the feature is allowed to be out of perpendicularity to A .034. The sum of
these numbers still equals 2.994.
So, we should now come to the conclusion that it is the job of the inspector at this stage
of the definition to assure all concerned that datum feature B will indeed, not only comply
with the tolerance, but that its tolerance, if complied with, will allow the feature to fit into
a cavity as described (with a minimum area available for insertion of 2.994) while
seated on datum feature A with three points of high point contact minimum. This is a
virtual condition boundary that represents the worst case assembly condition for datum
feature B.

 Page 26

FIGURE 3-3

FIGURE 3-4

 Page 27
Sometimes, in this day and age, we forget it is the ultimate goal of the inspector to give
assurances that a part or feature on the part being inspected not only meets tolerances, but
that in doing so proves that the part or features will perform desired functions. By that
same logic, we must look for multiple messages about functional requirements from the
final geometric control on this part, the control of position.
The essence of any position control, when used on patterns of features, is to hold the
desired angle and distance between the features within the pattern. The use of the position
symbol relates the features within the pattern to each other. When datums are used, the
relationship of the features to these datums must also be held. Sometimes, when the
relationship between the features is more important than their relationship to the datums,
composite position tolerancing is used (two levels of position controls sharing one
position symbol). Or if position of the features to each other and to one or two of the
datums is more important that the positional relationship to all of the datums, multiple
single segment position controls are used (each level using its own position symbol). But,
as is the case here, where not only must the features mate, but the datum features of size
must mate (and with the same interfacing part in the assembly), one level of position is
used to relate the features to each other and to the datums to within the same tolerance.
All relationships stated within the control are of equal importance. The reason for that
here is related by reading the control.
This positional control can be read as that we must hold the position of diameters that
mate while the part is seated on A and the holes are centered to B while B mates. Again,
the maximum material condition symbols have been read as an implication that these
features and datum feature B all mate with the same interfacing part in the assembly. If
this was not true, then the MMC symbology should not have been used in the way it has
been.
We have already gone over what datum feature B mates with; a cavity with a minimum
diameter of 2.994. But now we know this cavity probably contains five pins (or five holes
that will contain five pins eventually) that mate with the five holes being positional on the
part we are discussing. Four of these pins are dimensioned on a bolt circle that matches
the bolt circle on our part of 1.600, and 90 degrees to each other, while the other pin is
centered in the cavity.
If the holes on our part have a minimum diameter (MMC) of .570 and may move out of
position a .020 at the MMC size, then the area available for a pin to insert into right at
the perfect (true) position is the difference between those two numbers, or .550. If the
holes were produced at a .590, they would be allowed .040 of movement, again
creating a worst case boundary for mating conditions that is the difference between those
two numbers of .550.
That being so, we must assume that the pins in the cavity have been di-

 Page 28
mensioned and toleranced to have a worst mating condition boundary of their own that is
.550 or smaller (if a clearance is desired). They will mate with the holes on our part
while the plate that the pins are mounted on (the bottom of the cavity) seats on datum
feature A on our part. They will do so while datum feature B mates with the cavity.
The positional control could also have been read that we are positioning five holes to
each other for angle and distance, and to A for perpendicularity, and to B for location, to
within a minimum diameter tolerance zone per hole of .020, if the holes are produced at
MMC (a diameter of .570) or a maximum of .040 per hole if the holes are produced at
their largest size of .590 (LMC).
To make this explanation a little more clear, let's change the illustration to one with four
holes instead of five. Although the meaning will be the same in most respects to the last
illustration, to keep them separate, let's switch the primary datum to the letter L and the
secondary to datum D.
Since the numbers are all the same (just four holes instead of five) as in Figure 3-4, the
boundary explanation is the same for both.

FIGURE 3-5

 Page 29

FIGURE 3-6
Virtual condition boundaries (MMC concept) are generated by the collective effect of the
maximum material condition of a feature of size and the geometric tolerance applicable at
that size. The surface of the feature, such as the holes shown on the parts discussed, must
not violate the virtual condition boundary (.550) to comply with the individual
positional tolerance given to it. The pattern of virtual condition boundaries theoretically
begins perfectly centered to the secondary datum feature axis D (.800 radially from
center), but because of the maximum material condition symbol used after datum feature
D in the positional control for the hole pattern, this pattern may shift as datum feature D
departs from its virtual condition boundary of 2.994.

 Page 30
This pattern shift is only allowed while maintaining the relationship between the four
virtual condition boundaries of 90 degree angles to one another (and the primary datum)
and their 1.600 bolt circle distance.
Since datum feature D may be made as small as 2.960 and could, in theory, be perfectly
perpendicular to datum plane L at that size, the maximum pattern shift that may be
experienced by the four virtual condition boundaries as a group is a .034 or .017
radially. If the full amount of pattern shift is allowed, the pattern of four holes may be
shifted off the center of secondary datum feature axis D radially an amount equal to one-
half of the pattern shift zone (which is a .034) or .017. This has the effect of allowing
the four virtual condition boundaries as a group to deviate from their .800 basic
dimension distance from the secondary datum feature axis of D by the amount of radial
shift.
If this part mates with a part consisting of four pins in a cylindrical cavity, this pattern
shift allowed by the maximum material condition symbol used after datum feature D
(inside of the four hole's positional feature control frame) allows some parts to pass
inspection that can only be assembled off center of the mating cavity. Possible negative
effects of this are imbalance for spinning parts and nonuniformity of fit (unequal airspace
between datum feature D and its counterpart cylindrical cavity). This pattern shift of the
four holes as a group is in addition to the individual positional tolerance of movement
around the .550 virtual condition boundaries afforded to each of the holes. Size limits
of each hole's MMC and LMC are verified separately from the conditions discussed here.
These geometric tolerances and the virtual condition boundaries they generate tell the
inspector what he or she needs to know about the functional requirements of the parts
described here. And whether the part has four or five holes and whether the datums are
labeled as A and B or as L and D, we can generate a vision of worst case boundaries that
can be represented in gages. These gages may be generated in software and a hard gage
may be produced. Although the mating part may not be sized at exactly .550 for the pins
and 2.994 for the cavity, if our part as produced was capable of mating with such a
gage, it may also be capable of fitting the worst mating part (if it was also dimensioned
properly) without interference. This is so because what we have described here is a worst
case scenario type receiver gage or known as a functional gage. This functional gage
would be capable of checking both the perpendicularity control shown and the positional
control simultaneously. It would not check feature size limits.
This gage may or may not be needed for purposes of inspection. Other inspection tools
and equipment may be used to gather variables data. However, the process we just went
through (visualizing the worst conditions allowed) is important and allows us to
understand exactly what the drawing asks for and why.
The untoleranced gage for one of the parts discussed (the one with five holes) may
appear as shown in Figure 3-7.
With only some of the dimensions and tolerances, the gage may appear as shown in
Figure 3-8.

 Page 31

FIGURE 3-7

FIGURE 3-8

 Page 33

4
Functional Product Definition Creates Functional Inspection
4.1 Inspecting Parts as They are Used
I visited a company a few years ago that made large, soft rubber parts. As soon as I
entered, one of the quality engineers was waiting for me with product in hand. He shook
it at me. It looked like a large piece of gelatin with holes. Our inspectors are supposed to
inspect the location of these holes, he said as the piece quivered and rippled in the air.
What do you think? he asked, smiling. Believe we can get it to conform to specs? His
point was well taken and graphically displayed.
I asked the designer how the part was used, and while he explained, I took notes. When
he finished, I had the beginning of a good drawing note. It began, These features are to
be inspected in a restrained state, and went on to explain brieflybut specificallyhow the
part was restrained during use. The note was put on the drawings and a success story was
born.
Recently, a client company complained that while their very long, thin beams inspected
well within specified limits for parallelism, when in use they sagged so far in the middle
that anything placed on them slid to the lowest point. I asked to look at some, and sure
enough they looked pretty bad. The engineer scratched and shook his head as we stared
up at the structure. I don't get it. The thing inspected great according to the inspectors, he
said.
I notice it is suspended on each end with no support in the middle, I responded. He shook
his head.
How is it inspected? I asked.
He shrugged his shoulders in a you got me fashion. I suggested we find out, so we visited
the inspection department. As soon as we entered the room, we

 Page 34
saw one of the beams sitting flat on a long surface plate, entirely supported along its
length. An inspector was running an indicator over the surface of the beam's opposing
side. He looked up at us and grinned. Checks out great, he said. I got a look at the
drawing and realized the designer had used an entire surface of the beam as a datum
feature instead of using each end of that surface as datum target areas.
Thin-walled materials such as sheet metal, as well as some plastics and rubber parts, are
often prone to wide variations in form when released from the restraints used to machine
or form them. Holding these features to specific tolerances of geometric control requires
close examination of when and how the part dictates the need for these controls. For
example, if a sheet metal part is to be used in a restrained state (bolted, riveted, or
otherwise clamped against several surfaces of the part), then perhaps the final shape or
position (form, orientation, runout, profile, or location) of features should be judged to
be in or out of control while in the restrained state.
A note added to the drawing, such as This feature shall be inspected while restrained
(bolted, clamped, etc.) to the specified datum(s) with the following torque, will allow the
controlled features to be judged based on a simulation of the actual conditions the part
will experience while in use. As such, any geometric tolerances assigned the features take
on a realistic value, and the datum features are chosen and utilized based on proper
criteria:
1. Function
2. Representation of the mating situation
3. Accessibility
4. Repeatability
4.2 Functionality Can Be Used To Suggest Manufacturing Procedures
To Begin
Form control tolerances (like flatness and cylindricity), whether given as a direct
geometric tolerance or deferred to the feature size limits under Rule #1, should be easily
discernible.
Likewise, the tolerance on interrelationships (like angles and distance) between all
features that comprise a part should be either stated directly or calculable.
Be Complete
One should look at the individual features (surfaces) first. Look at each one individually
and determine whether or not its form is sufficiently toleranced.

 Page 35
Next, look at interrelationships. Is the tolerance between features easily distinguishable?
On many drawings I encounter, it is not possible to determine the desired relationships
(dimensions) between features (surfaces, holes, shafts, tabs, slots, etc.), much less a
tolerance on that relationship (angle and distance tolerances).
A Well-Documented Story
A drawing, like a part being produced or inspected, should be able to be read like a
bookwith a beginning, a middle, and an end. Each dimension and each feature control
frame describes how the hero goes from a nothing chunk of material to the dazzlingly
brilliant and functionally correct individual we grow to love (and/or put into our
assembly).
Simultaneous Requirements
Sometimes he/she wins the lottery and is instantly transformed, like a part with many
simultaneous requirements, which has everything punched, machined, or molded at once.
Step-by-Step
But at other times, a drawing can lead you through, chapter by chapter, saying things like
This feature/part works while it lays on this surface (primary datum feature) and is
centered to or measured from this hole (secondary datum feature) along a line that is
parallel or perpendicular to this surface [angular orientation (pattern rotation) datum
feature].
In English
In other words, lay it on this surface (primary datum feature) and push it up against this
surface (pattern rotation datum feature listed as third in the control) while you measure
from the center/axis of this hole (secondary datum feature of size). Or, as in the case of
three-plane datum features, lay the part down on this surface (primary), then push it up
against these two rails when you punch or drill these holes (three point minimum high
point contact for primary datum, two point minimum high point contact for secondary,
and one point minimum high point contact for the tertiary datum).
A Manufacturing Map
The controls on a drawing can map the way to proceed in manufacturing, and although
the manufacturing people can follow alternate procedures, it lets them know what is
functionally important and to within how much. They can then use their experience and
expertise to hold those relationships and shapes to within tolerance.

 Page 36
Choosing Datums
Seating surfaces are often made primary datum features and, to prevent them from
rocking in assembly, are given flatness controls. These flatness controls are usually great
places to begin manufacturing compliance. Making them flat gives us stable surfaces to
lay the part on to perform subsequent manufacturing operations.
Working from the Datums
In manufacturing, it is often beneficial to examine the datum features and to proceed with
providing for their needs first. If holes are to be drilled/punched into a plate, and the
holes are positioned to three planar datum features, the map reads:
1. Provide form control for the primary datum feature (flatness).
2. Provide orientation (perpendicularity) control for secondary datum feature to primary
datum feature.
3. Provide orientation (perpendicularity) control for tertiary datum feature to primary and
secondary datum features.
4. Position all holes in pattern while oriented to and/or located from the chosen primary,
secondary, and tertiary datum features.
Datums should be chosen based on functional requirements, but don't let that stop you
from reading their usage as a detailed outline on suggested order of manufacturing events.
Even if the designer didn't do his/her research on manufacturing capabilities, the given
geometric controls and the datums listed within those controls that use them can lend clue
after helpful clue as to the manufacturing methods (and their order) needed to produce a
functionally acceptable part at the lowest possible cost.
4.3 Multiple Possibilities of Dimensioning and Tolerancing for Functionality and
Inspectability
I've written a lot in this text about measuring parts as they function to be assured, as much
as possible, that if a feature measures within its functional boundaries, it will, indeed,
function. But there are times when features can be toleranced in a variety of ways, and
some of those ways are easier to inspect than others. In fact, sometimes features on a part
can be defined and toleranced in a less than functionally optimal manner and through
simple mathematical calculations can be shown as acceptable as a proof of functionality.
These deviations from an absolutely functional approach are almost always more
restrictive in terms of the amount of tolerance given and/or in the relationships to be held.
This methodology also involves a bit of lying. Oh, it's

 Page 37

FIGURE 4-1

 Page 38
lying in geometric symbology, but it's lying nonetheless. Let's face it, geometric
symbology is a language we use to describe how a part works and what relationships
must be held, and to within what tolerances in order for it to work. If we create false
origins of measurement, origins that serve no purpose in the way a part seats or mates or
aligns, we are saying a relationship must be held that serves no direct functional purpose.
But these origins of measurement or datum features can often be used to drive functional
requirements in an indirect way. Still, we have to accept that holding these false
relationships while buying us something such as ease of manufacture or ease of
inspection will cost us something as well, such as making tolerances smaller or rejecting
parts that may work quite well but that don't meet the requirements we have specified.
Still, these sacrifices are, at times, deemed acceptable to obtain the attributes they also
possess.
Ideally, the part shown in Figure 4-1 would have two primary datum features; one for the
clearance holes and another for the threaded holes. If this was the approach used and, as
indicated at the top of the figure, a gage was used to inspect the position of the M8X1.25
threaded holes, a different gage or checking procedure would be necessary to inspect the
clearance holes. However, if we choose to use one gage to inspect all clearance and
threaded holes simultaneously, we must use the same datums for all 10 holes. This will
work if we use a parallelism control on the surface that should have been used as one of
the datum features for the threaded holes but wasn't. Since the parallelism can't be perfect,
this will also make it necessary to reduce the positional tolerance available for the
threaded holes.
4.4 I Can't Measure That!
I received a call a while back from an irate inspector.
Those designers and tolerancing guys are always putting specs on parts that we can't
measure. I want you to call them and tell them what's what.
What do you mean? I asked.
You know, tell them to put on only the specifications that we can measure in the real
world.
I knew I was in trouble. I always cringe at the phrase in the real world because I know
the person who uses it is referring to the few square feet he or she occupies. It's the world
they have created for themselves in their own mind. To them it is all of the world that
matters. They can't seem to think themselves out of their narrow little box, so whatever
anyone else is doing or suggests is not, in their opinion, feasible in the real world. How I
hate that phrase.
The specification you want me to advocate, I began, is that how the part functions?
What do you mean? he said defensively.

 Page 39
I mean if I was successful in getting them to change the specification to the one you say
you can measure, would your measurement procedure assure that the part would or
would not function?
What do you mean? he said defiantly.
I mean, does this new specification have anything remotely to do with how the part
works?
It's what can be measured, he said, as though he was now dealing with a small child.
You know, I began, the original specification is one I've measured many times on many
different part configurations. And as I recall, the part you're discussing is of no special
geometry. It seems like it would be pretty easy to measure on a coordinate measurement
machine.
We don't have a CMM, he said, as if I was from Mars.
Well, how about an optical comparator, or a video comparator? I persisted. They should
be able to do that job.
We don't have that stuff either. His exasperation was building.
I thought I'd give it another try. Okay, then let's get back to the basics. This part has
curvatures and some hole locations that are important. If you were willing to create a
template as a master and a fixture to stabilize the part He stopped me in midsentence.
You're not going to help me, are you? He was irate, fuming. I had become his enemy.
You're just as bad as those knotheads that put this stuff on the drawings in the first place.
I'm not sure what you want from me, I answered.
I just want this stuff to be measurable.
To make your job easier, I finished the sentence for him.
Why not? he said.
A couple of reasons come to me right off the top of my knothead. Has it ever occurred to
you that if you're successful in getting these changes made, the way you will be measuring
the part has nothing to do with the way it works? Oh, it'll be easy to measure all right, but
your measurement procedure will be of no use. You might as well not perform it. Just
take a coffee break instead. When I was an apprentice die maker, one of the most often
repeated pieces of advice was if you don't want to work, go home.
I work hard, he yelled.
Maybe. But right now you're working harder at making your job one that adds no value to
the product. You'll be out there working up a storm, flinging elbows in all directions, but
it'll be a lot of sound and fury signifying nothing. I have a white mouse that runs around
on a little ferris wheel all day. He's working hard too, but he isn't getting anywhere. Did it
ever occur to you to ask for better equipment, so that you could measure these functional
specifications?

 Page 40
He was clearly offended now. I knew I had gone too far, but I couldn't seem to stop
myself.
He said, I won't need better equipment, if I get this junk off the drawings. I'm just trying
to save the company some money. These pieces of equipment cost a lot.
I said, So does shipping parts that don't work, or fall apart because they haven't been
properly tested or inspected. My advice is to put in for the equipment. Spend a little
money. Everything can't be inspected by eyeballing it from 30 yards away and saying
Yeah, it looks good to me.
At that point I heard a click in my ear. Oh well. It wasn't the first time I'd been hung up
on. It reminded me of the time I was teaching a concurrent engineering course at a
company. A bunch of managers sat in on the first day of the course. The company
problems were legion, with departments at war with one another. There were no
interdepartmental communications that didn't start with four-letter words and escalate to
the questioning of one another's parentage. It took a while, but when the first day ended, I
thought I had made significant progress. I did think it a little strange that the managers
had sat in the back of the room all day not saying a word to me or the employees being
trained, but instead mumbling among themselves and occasionally glaring or sneering at
me.
The next morning, before the class started, I was summoned into another meeting room.
When I arrived, I found managers and V.P.'s in abundance. Twelve of them sat around a
huge boardroom type table. They asked me to sit across from them. Them on one side
and me on the other. This was clearly not going to be pleasant. They sternly informed me
that what I was teaching may be what other companies call concurrent engineering. (This
group must have determined this from all the textbook quotes and film clips I had shown
from major corporations around the world about what was considered a successful
concurrent engineering structure.) But, they assured me, it was not what they wanted
taught to their employees. They ordered me to go back and tell their employees that I had
just been kidding the day before, and that what their company was already doing was
really concurrent engineering.
The exchange that followed wasn't much fun for either them or me. After a long and
unsuccessful attempt on their part to compromise my values, I offered to leave, but
pointed out my contract stipulated I would still have to be paid.
This so enraged them that instead of paying me, a person they now clearly hated, for
doing nothing, they decided to let me finish the training program and said they would
merely practice damage control after I left. They followed with a long letter to my
company about how unresponsive I was. One thing they didn't know, I guess, was that
my company was simply that: my company.
The funny thing was, a few months later I received several letters from both management
and the employees saying how much better the company was running, and that although
they hadn't employed all of the suggestions I had put

 Page 41
forth, they had changed many of the ways they did things to be more in line with the
concurrent engineering model we discussed. They had stopped looking at people who
worked in different departments as the enemy and their communication had grown
congenial and cooperative. One of the employees said he realized in the classroom that
the other people there weren't monsters and idiots as he had been told. They were just
people like him, with a job to do, a job that ultimately had an effect on his job and his
security in life. He realized that if they didn't work together and use the skills and
knowledge that each possessed, the product would suffer and eventually the company
would fail, and most important to him, he would be out of a job.
All said change was hard, but that to do things as they had done them in the past was
killing them, as individuals, and as a company.
This isn't a contest to see how much personal power we can accumulate within a
company, although many play it that way. It isn't a war between departments and
employees who work in the same company, although many believe it to be. It's a
business. And the only way to hope to succeed is to make stuff that works great and costs
as little as possible. And the only way I know to do that is to cooperate with others whose
skills, even though we might hate to admit it, are as vital to the product as our own. The
goal is not to make our own jobs easier if that can only be accomplished by hurting the
product and the company that keeps us employed. This isn't a me thing. It's an us thing.
And if you don't believe me ask the formerly employed who now stand on corners with
signs that say, Economically challenged. Will work for whatever you've got. They know
that what you have, they once had and thought it would last forever.
4.5 Functionality vs. Producibility
One day, many years ago, I went from manufacturing parts as a tool and die
maker/machinist to designing and tolerancing parts. Until that day, I had felt like I was
dictated to by the designers, but Now, I said to myself, I have the reigns! I was
determined to dimension every part to be as producible as possiblesomething I felt had
not been done by most designers before me.
Almost every day I made a trip to the shop floor to be with my buddiesthe machinists and
the inspectorsand to solicit their advice. I would explain my most current project and ask
them what they wanted to be the origins of measurement (datum features). Of course,
they were only too happy to explain to me in great detail how machining and set-up
procedures of the day dictated the most accessible and repeatable origins of measurement
and orientation. I was (coming recently from their midst) easily convinced. So, I readily
took their advice and every part drawing I turned out was practically the most producible
they had ever seen. I can tell you for a fact that those parts zipped through manufacturing
and inspection without a hitch in what some considered record time.

 Page 42
Everyone was happy until a somber-faced assembly foreman came to visit me one day
and said the words I'd never considered hearing. Hey, he droned in a deep voice, none of
your parts work! I was devastated. It seems my parts wouldn't function. As a matter of
fact, if they went together at all, they did so at odd angles. The loads were not being borne
by the big broad surfaces capable of bearing them but often by little spindly pins, bosses,
and screws.
I felt betrayed, so I went to visit those I thought had let me down. I dragged what was left
of my ego to the shop floor and confronted my aforementioned buddiesthe machinists
and inspectors. I said Buddies pals, you guys have let me down. None of the parts work!
It was at that time that I got the coldest stares I can ever remember seeing. And from the
faces of those I had, to that point, called friend, I heard the words All those parts fit the
drawings, pal. We just manufacture and inspect them to that drawing. It's your job to
make certain they work. Remember, you're the designer now.
I felt defeated for a while, but I learned a great lesson. If the parts were to function, I was
the one who must make it happen, and I had to do it in the design and dimensioning and
tolerancing phase of the project. It was smart to solicit the advice of the machinists and
inspectors as to what was most producible and inspectable to insure datum feature
accessibility and set-up repeatability. But that advice must be taken in context and never
should I, as the designer, sacrifice functionality to get producibility. I must examine the
surfaces along which the parts mate and make the datum features representative of the
mating situation. In other words, we must dimension and tolerance the parts to be as
producible as possible without sacrificing design requirements for form, fit, and
function. In the final analysis, we are judged on more than just costthe parts have to
work!
4.6 The Things I've Seen (Where Tolerances Come From)
Going from company to company, product to product, and part to part, it's amazing how
many similarities in both good and bad dimensioning and tolerancing practices I come
across.
It is almost the same thing as when archeologists find different cultures in separate
regions of the world that share similar legends or ceremoniesthey begin to speculate. Were
these cultures growing at the same rates and reacting to the same environmental and
astrological forces, or was there some vagabond traveler who spread the word like an
over-the-fence neighborhood gossip?
I picture that same type of thing happening with dimensioning and tolerancing practices. I
don't buy the separate cultures growing along the same lines theory, though. I think the
vagabond employee theory holds up here. I

 Page 43
keep running into the same guys throughout my travels. They stay with one company just
long enough to muck up the works, then move on. One guy carries around the kielbasa-
shaped tolerance zone (plus and minus location in one direction and true position
cylindrical in the other).
Then there is the person (a short, balding guy with a fake-looking bushy mustache and
fan-like eyebrows) who goes around telling people that a composite positional control
must repeat only the primary datum feature in the lower tier (feature-to-feature relating)
tolerance.
The more I think it over, it could be the same guy with a good disguise kit. You'd be
surprised what one can do with just a pair of Groucho glasses and elevator shoes.
Of course, I don't blame everything on this person (or persons). I know a lot of people
are just out there trying to fake it. It goes back to what my long-departed (fired)
acquaintance Charlie told me over 30 years ago when I asked him about tolerances.
Charlie, I asked with the wide-eyed innocence of youth, where do tolerances come from?
Charlie put his big, hairy arm around my shoulders and said, Kid, tolerances are a
mystical kind of thing. After 20 or so years in the business, it just sort of comes to you.
But until that day, you just come to your old pal Charlie, and I'll tell you what the
tolerance du jour is.
And, of course, I believed him. So, every day I went to him and inquired about the daily
tolerance. And I noticed the answer varied only with his mood. If he was in a good one, I
got plus or minus .010. But if he was grumpy, it reduced to plus or minus .005. Later, I
found out Charlie didn't know a tolerance calculation from third base, but I think he has
disciples out there still spreading the gospel according to Charlie.
And how many times have you seen a tolerance inside a feature control frame that was a
diameter of .028 or .014? This is the evolved Charlie Principle of taking a tolerance that
was calculated incorrectly to begin with (read that plucked from the air) and multiplying it
by 1.4.
I've had so many manufacturing and quality people look on in incredulous disbelief as I
removed those mystical tolerances and replaced them with ones that were calculated using
better information.
How can you just take that out of there? It was put there for a reason, wasn't it? they ask.
Yes. It was. Unfortunately, the reason was quite often to get it off his or her desk, rather
than preservation of form, fit, or function.
I have spoken to more than just a few people who thought the secret of good tolerancing
technique was to copy it from old drawings of parts that sort of look like the one I'm
doing now.
Datum selection is another telling detail about the level of a person's ex-

 Page 44
pertise. Many dimension and tolerancers don't think of a feature control frame as a
physical description of how a part functions and mates. That's why for positional controls
you'll see only locational datums, nothing to orient the part and features: no
perpendicularity datum because they don't see position as capable of anything but
locatingand never an angular orientation datum to control pattern rotation, for the same
reason.
One of the most common true position feature control frames includes datums that
control perpendicularity, location, and angular orientation when positioning hole patterns
from a datum hole. It's easy to tell that, when all this is needed, if the only datum feature
they decide to include in the feature control frame is the hole for location, and nothing for
perpendicularity or angular orientation, that they are more than just a little wet behind the
ears.
I don't exactly believe there is a conspiracy at work to throw a monkey wrench into our
dedicated efforts to dimension and tolerance parts correctly. But at the same time, if you
meet a guy with a bushy red mustache, or even someone named Charlie with fan
eyebrows and Groucho glasses, don't take any unnecessary chances.
4.7 Hog-Tied and Cow-Chipped (Importance of Geometric Definition)
Sometimes, going from company to company, I feel like someone in a mall doesgoing
from store to store with their face pressed up against the windows. I peer in briefly,
glance around, comment on what I like and what I don't. It all seems so clear when you
don't have to contend with the day-to-day static that tends to camouflage the real issues.
I don't have to worry about the personal problems of the staff or that what is needed to
make a functional product sends someone from manufacturing yelling to the plant
manager because he misunderstands the complexities. I see the needs so clearly. I often
listen patiently for long periods of time while someone from design explains just how
extremely complicated his products are. I think to myself at those times that part of what
he is saying is true, part frustration, and part that he just feels the need to impress upon
me that no simple solution will solve what he perceives as an unsolvable problem with a
set of circumstances entirely unique to the world.
My job is pretty simple, compared to his. There are only so many geometries that occur
and they seem to all fall into the special little categories that I have created for myself over
the years. This geometry calls for a variation on that categoryand that geometry a small
tailoring of another. There are, quite simply, some absolutes in industry.
Once you realize that, the rest is clearing away the debris to allow those absolutes to take
place. The result may be, initially, some change in the way a fa-

 Page 45
cility does its daily business, but improvement in the product and the skills and creativity
of the company personnel is usually the wonderful final result.
So much of the creative energies of a person are drained by dealing with the politics of an
organization, that often the best part of him isn't left to influence the actual product. I just
finished a workshop with a great group of individuals who had allowed themselves to be
talked into putting on their drawings not what the products needed to actually work, but
rather exactly word for word in a drawing note what their mold maker (a vendor) told
them was the nature of parts as they react in the molding process.
I picture the mold maker examining the initial sketches that outlined their functional
requirements in the finished parts. He probably smiled condescendingly, put his arm
around them, and led them aside to impart the breadth of his great wisdom. He told them
that parts in a mold change size at a rate no one can predict.
So, he probably continued, don't worry your pretty little heads about putting any specific
tolerances on those dimensions. Simply write a drawing note that says the tolerances may
increase as each individual hole or shaft gets farther and farther away from the origin of
measurement.
So, they did just that, and as each hole got farther and farther away from the datums, they
were required to be less and less accurate in their location. They did the same for the
mating part. And when they brought the two together, each progressive hole aligned with
its mating hole with less and less accuracy, until very few of them were well enough
aligned to put a screw through. This was the same group that had laughed when I told a
story in the workshop about the drawing note I had encountered at another company that
read This part must function, regardless of the dimensions and tolerances given on this
drawing.
I asked them if they were willing to match-drill the holes. They said no. They were
committed to having the parts come out of the mold and work as is. I asked if they were
designing the molds or the finished parts. They said the vendor supplied the molds (and
designed them) and all they did was pay for the parts hundreds of thousands of them.
And these parts currently don't function? I asked. No, they replied.
Hundreds of thousands of them? I asked.
It's been a real problem, they replied. We have them stored.
I went with them to visit the mold maker. I told him the relationship between the first and
last hole had to be held to within the same tolerances as the first and second hole. They
work together, all at once, with one mating part. I explained. He gave me the same smile
and put his arm around me. I smiled and put my arm around him. New chums.
That will cost them more money, he counseled, his smile getting bigger all the time.
You can make the parts cheaper your way, I said.

 Page 46
He nodded. Now you are beginning to understand, he said. Like a big brother, I thought.
Did you know that none of those parts will actually mate? That they haven't been able to
use even one of the parts that came from your molds? They appreciate the fact that you
didn't charge them very much for each part but would really like the parts to actually
work.
He became defensive and then he repeated a phrase I have heard hundreds of times
before. Well, they fit the drawing requirements they gave me.
Which are what you dictated they put on the drawings, I said, my new chum status
quickly dwindling.
And legally, I don't have to buy even one of them back, he shouted, after removing his
arm from around my shouldersno longer pals.
Understood, I said, still smiling. Nothing personal here. These are the new part drawing
requirements, I continued, spreading out the same pictures he had seen beforeonly with
specified datums, basic dimensions and feature control frames that held some very
specific tolerances.
His eyes immediately ran to the drawing note column, and scanned for his note. The
white portion of his eyes seemed unusually largebut it could have been my imagination.
Still, he reminded me of a reptile when he put his face very close to my own and said,
Where's my tolerance note?
I was still smiling buddies, when I put my arm around him. The one about the tolerances
sort of being whatever you can hold at that moment in space/time? I asked. He
backpedaled to escape me, but I stuck with his moonwalk movement as though he was a
calf about to be thrown and have his hind legs tied. Nevermore, I croaked like Poe's
raven. Nevermore.
4.8 Correct and Complete (Tips on Functional Product Definition)
I have been having a lot of fun lately. I have been fortunate enough to work with many
good people and on some really stimulating projects. I've learned some interesting things
and also had some disturbing insights. I want to share them with you.
I have found over the years that the most effective method of training is not just the
generic course that leads the trainees through the dimensioning and tolerancing standard
and explains the general uses of each symbol. The type of training that has the most
impact on a system and the culture within a company is the kind that uses their drawings.
If you take their drawings, the ones in progress or those ready to undergo change, and
ask the course participants the right questions about part functionality, manufacturability,
inspection procedures, and, in general, the completeness of geometric definition and then
change the drawing right before their eyes using the information that they feed you, the
impact on the

 Page 47
product and the participants is amazing. The course material is no longer considered just
another training exercise. It becomes real and exciting and cannot be ignored. It is also at
times a little threatening. It threatens the little bubble that some tend to build up around
themselves that says, This stuff is okay in theory, but can't be applied in my special,
unique situation. Any instructor who is there to do anything but blow smoke and show
others how bright he is has to burst that bubble, and working on their parts right in front
of them is the most effective way to do it. It works even better if, after the course, the
instructor leads a work group of company representatives key to the drawing and product
creation and plows through more company drawings making suggestions for change that
make them complete and correct. Having a large number of really good product drawings
around will not only improve the current projects, but will influence all work that is done
in the future. People will follow a good example just as many will follow and perpetuate a
bad oneif that is all that is available from which to learn. I have been doing a lot of this
type of work/training lately, and given a manageable-size working group, we have been
averaging about 30 minutes per drawing to correct and complete. Some of these drawings
have caused troubles for years. Who knows how many man-hours have been spent in
heated discussions necessitated only by the fact that this type of process of integrated
design procedure wasn't conducted in the first place.
Many of us have spent a lot of time wondering why some individuals and even some
companies have had such a tough time turning out drawings that well define their product
needs. Why is it that when a drawing package is put in front of me, I expect it to be
absolutely atrocious in the way the parts have been defined? Why are the people turning
out parts that have measurements taken from opposing ends? Why do they use geometric
symbology to control surface form, when their size limits already control it better? Why
are many features shown coaxial, but only a few given a tolerance of allowed deviation
from perfect coaxiality? Why are the All angles are equal to general tolerance notes on the
drawings so often left blank? Why can't the parts be given datum references that really
reflect how the parts function, so that when the parts are inspected and found to comply
with the drawing, we can rest assured they will actually work? I could go on for the rest
of this section, but I think you get the point. I always say, and truly believe, that most
people just don't ask themselves the correct questions when they dimension and tolerance
the part. Questions like: Does my feature control frame truthfully describe how my part
works?
I suggest to people that they ask themselves to mentally strip the drawing of all geometric
controls and even all numbers, to initially see what physical configuration they have
drawn. Many times, this enables them to see that many features on the part have been
dimensioned without the use of numbers. For example, if features are drawn at 90
degrees, they are considered dimensioned 90 degrees. They must then be toleranced.
That's what the general tolerance block

 Page 48
or feature control frames are capable of doing. If features are shown coaxial or
symmetrical to each other, they are dimensioned. We often term this type of dimensioning
a basic dimension of zero. If a dimension other than zero was desired, it would have to be
inserted on the drawing. Still, once it is realized that these features are dimensioned for
desired location (a basic dimension of zero), they must then be toleranced. Tolerancing
for location of holes and shafts, slots, and tabs is accomplished with feature control
frames. Plus-and-minus toleranced dimensions are quite useful for showing size
tolerance, but are not used to show location of features of size. Location of these size
features is shown with basic dimensions (some of which may be implied to be zero) and
tolerances are put in feature control frames that include geometric characteristic symbols.
People often say they don't know which geometric control to use. For example, they say it
is hard to choose between circular runout, total runout, concentricity, and position. I say,
at least you are choosing some type of tolerancing that will get the job done, rather than
relying on the psychic abilities of the drawing interpreter to guess at the tolerance. Some
people say they will rely on good workmanship, but believe me, even if we want to do
the right thing, a hint at what an acceptable deviation from perfection is would be greatly
appreciated. Many designers or engineers or drafters don't want to choose wrong, so they
don't choose at all. Most of the time there are several right answers. Some are better than
others, but for certain, no choice at all is not the correct way to go.
There are many rules of thumb that can be helpful. A few I use to begin my decision-
making process are: Does the controlled feature mate with anything? If it does, use a
geometric characteristic symbol that allows you to put a maximum material condition
(MMC) symbol next to the geometric tolerance. For example, in the choice of runout
controls vs. concentricity vs. position, only position allows the use of the MMC concept,
so if it mates with something and does not spin while in use, use position. Once chosen,
position also allows the use of an MMC symbol next to any datum features of size. But
does the datum feature of size mate with anything at the same time that the controlled
feature does? If so, then use the circled M, but if not, use the RFS concept instead. In
other words, if an alteration in size does not increase your chances for mating by shifting
things around, then don't say that it does.
As I have said so often, the feature control frame is nothing more than a physical
description of how a part and its controlled features work. It can be read like a sentence.
The sentence must tell the truth. If it does not tell the truth, then one of two things can
happen: (1) either the parts can fit the drawing requirements and not function or (2) all
parts that do fit the drawing will function, but the parts cost too much.
Another good question to ask is Do the parts spin? If so, a symbol that requires the part to
be inspected while rotating may be a good choice. It allows us to inspect the part as it is
used. Both circular and total runout are in-

 Page 49
spected with a full indicator movement in a 360-degree rotation about a datum axis.
I see many feature control frames that are surface controls that include a diameter symbol
after the geometric tolerance. That is not correct. I always ask participants in my courses
to read the diameter sign that appears before the geometric tolerance as of the axis. We
only have three types of controls currently. They are: (a) surface, (b) axis, and (c)
centerplane controls. If your desire is to control the surface of a feature, then reading of
the axis inside of a feature control frame should tip you off that a grammatical error has
been made. Likewise, when a surface control uses a MMC symbol, it implies that a
change in size somehow alters surface form, which is not true. Again, a tip that an error
has been made. Many times a MMC symbol is placed next to a controlled feature
tolerance or next to a datum feature that has no size, but is instead just a planar surface.
Ask What is the MMC of that surface all by itself? and if the answer is that it has none,
then that is a sure sign that the symbol has been misused.
I am sometimes frustrated that when a group graduates from a training course they are
reluctant to show me their progress on the job later. I want to see what they are doing, to
let me suggest adjustments in their work, to keep them on the right track. I care.
And I know how the pressures of the job can sometimes coerce them into doing things
they learned not to do, and that sometimes they may forget to ask themselves the same
questions that I would ask themthe questions that would have led them to correct
answers. I understand how much they have bled for that design, how much they have
invested in it, and to have me look at it while they watch for my reaction, hoping my eyes
will light up and proclaim them brilliant and tell them they have surpassed anything I
could ever have done, is enough to make my heart sink with theirs when I can't. I tell
them that many of the things they have done are wonderful, insightful, and better than
previous efforts, but it is not enough praise. They want me to say it is perfect. I would
settle for being able to say it is sensical, logical, and legalthat it is complete. I have learned
to be very flexible over the years. Just because the procedure is not the one I would have
followed is not usually even worthy of comment. Will it work, and not cost a fortune? I
ask myself. If the answer is yes, the birds sing!
Often, letting me examine the drawings of those who consider themselves in-house
experts is not something to which they wish to subject themselves. I only want to help. I
want to ask them all those questions I used to asked them during training, to remind them
that they apply in the real world. I want them to continue to progress, so that someday
they really won't need my help, not just say that they don't. I know they are smart. I know
they can do other things that I never will be able to even comprehend. But this is what I
do. I have seen the mistakes common to part definition so often that they jump off of the
page at me. I was once called into a company by a well-meaning person and had
drawings

 Page 50
spread before me for comment. As I began, the drawing's creator entered the room out of
breath, looking wild-eyed, like a creature cornered. He yelled that he didn't want any
outsider telling them what to do on their parts. He insisted I not comment, while I stared
at thousands of dollars in potential problems on those drawings. I saw interference
between mating parts, unproducible controls, nonsensical geometric statements and
missing tolerances by the dozens, but by then the person who had called me in had lost
his resolve and wouldn't allow me to tell them what I saw. It was so sad. I wanted to say,
I am on your side. I am your pal. Don't be intimidated. No one is perfect. We learn as we
go.
I wanted to say, You and those like you have taught me so much that I have revised my
books time after time and added over a thousand pages to my original offerings. I don't
want to tell you what to do, I just want to ask you the questions I have learned to ask
myself over the yearsthe questions that will lead you (as they have led me) to do the right
things on your drawings, in your environment, using your knowledge of your parts. But
by then, they were yelling at one another. Saying things like How dare you question my
judgment! This is my responsibility, not yours! If I want your help, I will ask for it! and
other such team-building type phrases.
I was reminded of what Lee Iacocca said when asked what the single most important
thing was that he had learned from the Japanese. He replied that a design had to be a
process that thoroughly integrated the knowledge of all levels and departments of the
workforcethat it was the only way for us to stop problems before we build them into the
products. We need each other.
I have to turn this chapter over to my trusted associate now for comments. I always hope
she will say this latest section is fantastic, the most wonderful ever. I always wait and
watch while she reads them, anticipating nothing short of unbridled, enthusiastic
compliments. It never happens. She always finds something that needs changing, things
that must be improved. She assures me that it will be all rightthat she is on my side. It is
always a little disappointing at first.
We are all pretty much the same slightly insecure, at least a little sensitive, and to a person,
we are all imperfect. But we keep trying.
4.9 Miles to go Before we Sleep (Importance of Geometric Definition)
I've worked in several companies that vend work out to be manufactured. Several times
I've received phone calls or visits from representatives of these vendors with similar
messages. They've said things like, Jim, we really appreciate the contracts we've received
from your company, and we thoroughly understand everything important to the
production of your parts. There is just one thing we've been curious about, though. It's a
small thing. What are all those funny little symbols all over the field of the drawings?

 Page 51
Certainly, a scary question. One that leads to another question I am forever being asked:
How many people out there really understand this stuff?
I'm never able to answer that one. The only people I come into contact with are those
currently using or interested in learning to use ASME/ANSI Y14.5 symbology. There
must be pockets of individuals out there, potential users who have never been exposed to
or never been interested enough to pursue the topic. I have no concept of what the ratio
of users to nonusers is.
The second facet of that is How many of the current users understand it well? I believe
the numbers are growing. Still, it takes diligence and study and teamwork to use it
properly. Just as we were not born with an inherent knowledge of plus-or-minus
tolerancing and had to learn it, we are likewise not genetically infused with a knowledge
of how to use or interpret geometric controls. It takes a lot of hard work. But, unlike
many changes we resist, this one has too many benefits to ignore. Rarely have I seen such
a vast difference in the capabilities between two choices of product. This language of
symbols allows us to precisely geometrically define parts as never before possible. It
allows us to increase tolerances based on produced part size without endangering
functionality or to reduce tolerance on only one facet of geometric characteristic without a
multitude of lines of drawing notes.
We can specify the loosest tolerances of form, orientation, profile, runout, or location
possible to reduce costs and still preserve necessary fits and balance. We are now capable
of leaving no doubt as to which origins of measurement reflect the conditions under
which the parts actually operate and mate. Repeatability from machine to machine and
department to department is easily accomplished with those features desired for set-up
clearly marked and accessible. These are but a few of the benefits available through the
use of geometric controls. And all of those funny symbols used allow us to reduce the
number of potentially ambiguous drawing notes.
I often ask groups what the word buck means to them, and over the years I have had at
least 50 responses covering those concerned with finances, natures fauna, rodeo riders,
Harry S. Truman (The buck stops here), and many more. If that one word means so many
different things to so many different people, how many ways can the average drawing
note column with dozens or even hundreds of words be interpreted and perhaps
interpreted differently than the designer intended.
It frustrated me in my years in manufacturing that I would receive so many drawings that
were so ambiguous and/or incomplete that I would have to visit the designer. I would say
something similar to Hey, Joe, I've thoroughly considered all aspects of this drawing and
studied what it says. Now I want to know what you really meant.
I realize now that it was not the fault of the designer but rather the language he had
decided to or had been forced to use. Good old plus-and-minus tol-

 Page 52
erancing was not capable of adequately defining parts or locating features on those parts.
Joe was a good designer, but it was as though many of his ingenious ideas were doomed
to remain in his mind or at least never to find their way onto the field of the drawings. He
was like a potentially great novelist confined to the use of a very limited number of words
or a talented machinist not given the tools necessary for the job.
We have come so far. Yet, there is a way to go. We must work together to spread the
knowledge, to share experiences and obstacles faced and to be overcome.
4.10 Plastic Parts and Stretch Tolerances
A few years ago, I was asked to consult and teach at a company highly regarded in the
area of home electronics. They said they were having some problems with assembling the
plastic housings that held their main product.
When I first arrived, I was taken on a tour of the assembly facility by the manager of that
area. During the tour of the assembly line, a worker approached and told the manager that
there was no more room in the scrap area where they held their rejected parts. The
manager clasped his hands behind his back and rocked back and forth on the balls of his
feet for a while as he stared at the ceiling. Then he said to the worker, Build the walls
higher.
Excuse me? the somewhat perplexed worked asked.
Build the walls around the scrap area higher so it can hold more, he responded.
I wondered at how this guy ever got to be a manager.
After the worker left, the manager turned to me and said, It's taking longer than we
thought.
What is? I asked.
For the molds that make our parts to wear their way in.
Wear their way in?
Yeah, he said. Right now, the mold maker, who designs our molds and makes our
housings from them, is still shipping us the two halves that comprise our housing, but the
darn things still won't assemble. But he assures us that any day now the molds he
produces the housings in will wear their way in and we will have housings that assemble.
Wear their way in? I repeated.
Yeah. You know plastics are a very difficult product to work with. No one can really
predict what a plastic part will look like when it comes out of a mold or is extruded. It
bends, twists, and shrinks.
They do, I said. But don't you put tolerances on your piece part drawings, to show what
the finished parts, the two housing halves, have to conform to in order to actually go
together?

 Page 53
He was indignant. Of course we put tolerances on our parts. We're not boneheads! It's just
that with plastics, he now acted as though he was speaking to a 6-year-old, the tolerances
must be proportionally larger as you get farther away from the origins of measurement.
You know, he said, stretch tolerances.
Stretch tolerances? I asked with a blank look on my face.
Of course. Our mold maker says specific arithmetically calculated tolerances can't be
assigned to features on a plastic part like they can on other parts, so they have to stretch
just like the part does in the mold.
Allowing the features on the two halves of the housing that are made in two different
molds to misalign more and more as they get further away from the datum measurement
origins, I summarized.
Finally, he smiled and nodded vigorously. That's it!
But doesn't that mean you won't be able to get the two parts to mate?
That's our problem right now. Until the molds wear their way in we're getting nothing but
scrap parts out of the molds.
How many of these parts have you bought from the mold maker? I asked.
All of them, he said. It's in his contract. We've got to eat the bad parts until the good ones
start to roll in.
What numbers are we talking about? I asked.
He thought for a while and then said, Well, so far we've bought roughly two hundred and
fifty thousand parts from him.
And they're where?
He pointed. In the scrap area over there.
Are you going to rework them?
He shook off that suggestion. No time for rework. Any day now, we'll start getting good
parts.
Some people are just gullible. Some are so desperate for an answer for what goes wrong
or so needy to believe their situation is vastly unique and harder than anything anyone
else has ever faced that they will take any answer provided them. Later, they'll say, Hey it's
not my fault. It's the situation, the product, the material, the industryanything but them.
What we have to accept is that our individual situations are no more difficult than the
situation of others. Life is tough. Wear a helmet! But take the responsibility and don't
accept ludicrous answers that can potentially drive your company out of business.
What this man failed to understand was that his company wasn't supposed to be creating
in-process part specification drawings, but rather finished product information, with
which the parts, when they were received, would comply. Specific tolerances, not stretch
tolerances, were required, and it doesn't matter if his parts are made out of plastic or
cream cheese.
Eventually the stretch tolerances on their drawings were replaced by spe-

 Page 54
cific calculated tolerances. And, yes, the mold maker was angry and refused to comply.
But the way I figure it, he was a scam artist anyway who probably took his profits from
the hundreds of thousands of nonfunctional housings he sold this company and retired to
Acapulco. The next mold maker didn't flinch at the tolerances, chuckled when told the
history of problems, tried harder, and easily produced better molds and consequently
better housings from them. And surprise, they all assembled, met the tolerances, and
worked great.
You can take off your helmet now. Life is getting easier.
4.11 Design, Inspection, Production, and Prototype Needs and Capabilities Regarding
Dimensioning and Tolerancing
Although geometric characteristic symbols and datum features are to be selected to
preserve the functionality of features, there are times when more than one choice of
symbols or one set of datums will, if selected, serve that purpose. In these instances,
manufacturing and inspection techniques and limitations should be considered. This
makes one's choices better in that it may allow the part to be more producible and
inspectable.
One such example is the choice between concentricity and circular runout. Both of these
geometric characteristic symbols are capable of controlling coaxiality. Concentricity is
potentially a more inspection-intensive control in that it asks one not to reject a part
feature that is centered but not round. Since an indicator, if kept in constant contact with
the feature's surface, picks up not only a type of coaxiality but also roundness, a full
indicator movement in a 360 revolution may exceed the concentricity tolerance. This
forces the inspector to differentiate between concentricity characteristics and roundness in
order to accept or reject the part. This can be a time-consuming procedure. Still, it will
not force parts to be rejected that are on center, balanced, and able to spin true to the
datum axis, but are simply not round. Circular runout, on the other hand, is potentially
more manufacturing-intensive. It requires a part surface not only to be coaxial to the
datum axis but also to be round. A part that fails to be round enough is rejected no matter
how true it spins. This means manufacturing must take greater care to make the part
surface round as well as coaxial to the datum axis if the part is to be accepted by the
inspector. This circular runout, on the other hand, does not force the inspector to
differentiate between roundness and coaxiality. If the full indicator movement in a 360
revolution exceeds the runout tolerance, the part is rejected. The choice between these
symbols may boil down to where the tolerancing engineer decides to spend the most
moneyin manufacturing or inspection. Many geometric characteristics are chosen on a
similar basis. A prototype part or a small production run of products may be able to
withstand a larger amount of inspection time; therefore concentricity may be chosen. If a
large production run may not be able to with-

 Page 55
stand a lot of inspection time, circular runout may be chosen. However, it may be the
decision of the tolerancing engineer that a large batch of very coaxial parts that fail
inspection because they have been given a circular runout control and happen to have
been produced out of round is not a situation to be tolerated.
Datum feature selection is primarily based on functionality, but again, it may be true that
there is more than one set of datum features that would preserve part feature
functionality. When this is true, producibility can be used to decide between them. If, for
example, one set of datum features includes very accessible planar surfaces and the other
set of datum features includes datum features of size, the amount of parts to be produced
can be a deciding factor. In a situation where only a few parts are to be made, centering a
pattern of holes to the axis of a datum hole may be an acceptable expenditure of time. If,
however, hundreds or even hundreds of thousands of parts are to be produced, unless
special fixtures are made to do the centering, this type of datum feature (a hole or slot)
may not be acceptable. Given a choice in a long production run, three planar datum
features may be more attractive. Requiring only a table on which to lay the part (on the
primary datum feature) and two planar rails against which to push the secondary and
tertiary datum features, the set-up time is greatly reduced.
In summary, designing in functionality is critical. It does no good to design, dimension,
and tolerance to produce a low-cost product that simply does not work. Still, many
considerations of producibility, inspectability, and cost cutting can be infused into the
dimensioning and tolerancing techniques without adversely affecting functionality. A
great amount of thought and collaboration between departments should go into the
process before the drawing design requirements are finalized. Those individuals or
departments who believe that they are knowledgeable enough in all fields of design,
manufacturing, and inspection (among others) to complete the task on their own are
mistaken.
A concerted concurrent engineering effort is needed to do the best job possible and to
avoid high costs in production and/or inspection. One must remember the parts not only
have to function but must be affordable to the customer. Regarding prototype vs. long
production runs, one must keep in mind that a prototype machinist can perform feats that
would require magic in a production environment. If a part is first made in prototype and
subsequently put into a long production run, it is generally a good idea to allow the same
vendor/shop to do both. In that way, they can foresee difficulties that will occur during
the production run while they are producing the prototype.
This information can be used to improve the design, dimensioning, and tolerancing on
the finalized design drawing to allow the part to be more easily produced and inspected.
This collaboration can be even more efficient if data bases are connected between design,
production, and inspection facilities so that each can monitor and input suggestions
during the design, dimensioning, and tolerancing phase. This should not only improve the
part and lower costs, but

 Page 56
also improve greatly the relationship between these different disciplines. It may even
allow them to more easily realize that they have goals in common and are not required to
battle one another. All want the company to do well, sell product, and build a good
reputation that generates more business. Once this realization occurs, a team can really
begin to form and work well together. The like a well-oiled machine description of how a
great team operates can only be achieved when the collaboration begins and the bickering
ends.
The tolerancing method shown in Figure 4-2 uses two holes within a pattern as the
origins of measurement and part orientation for all other part features, including other
holes of the same size. This is an alternative method (and generally considered easier to
inspect) to using the entire pattern of holes or shafts from which to construct a datum axis
as illustrated in Figure 4-3. The simultane-

FIGURE 4-2

 Page 57

FIGURE 4-3
ous gaging requirement in Figure 4-2 is in effect for all features that use exactly the same
datum structure. This is reinforced by the SIM REQT note.
4.12 Implying a Manufacturing Sequence on Complex part Configurations
Using Compound Pattern Datums and Profile
The part in Figure 4-4 mates with a similarly amoeba-shaped cavity that contains 7
pins/shafts. These shafts insert into the holes on this part while the cavity entirely
encompasses the outer periphery of the external amoeba.

 Page 58

FIGURE 4-4
The manufacturing facility has requested we create a series of feature control frames that
suggest a manufacturing sequence. The part will be punched out of a blank rectangular
piece of sheet metal in three separate presses. Each press will either punch out one of the
hole patterns or the outside periphery's profile. The manufacturing sequence will be
suggested by the datums, or lack thereof, in each feature control frame. If a control uses
datums, then that implies the datums

 Page 59
have been created/punched out in a press prior to the feature or features controlled to that
datum or datums.
The four-hole pattern is labeled datum D. The holes are positioned and use datum A only.
It is read, The axes of these .250.260 holes are positioned to one another and must
maintain their basic hole-to-hole location dimensions and their perpendicularity to datum
plane A perfectly if produced at maximum material condition (a .250). It also implies
that one should not produce the holes at their .250 size. If produced at that size, the
required perfection of hole-to-hole location and perpendicularity to datum plane A would
be impossible with which to comply.
If, instead, the holes are produced at a larger size (within their limits of size), an
additional tolerance of position is gained. A maximum of .010 of positional tolerance is
available to each hole if the hole is produced at .260 (least material condition). Since
both the three-hole pattern's positional control and the profile control use datum D, the
four holes (datum pattern D) must precede them. The four-hole pattern is punched out
(created) in the first press.
Four real holes are manufactured. These holes are imperfect but, hopefully, good enough
to meet their size requirements and not violate the perfect imaginary boundaries of virtual
condition. These virtual-condition boundaries consist of four cylinders (a diameter of
.250 of an inch) with their axes located exactly as dictated by the basic hole-to-hole
dimensions displayed on the drawing and perfectly perpendicular to datum plane A.
These virtual condition boundaries have a central axis, which is datum D, but for all
practical purposes, it is the axes of the boundaries that serve as the origin of measurement
for features controlled to datum pattern D. The concept of a central axis for datum D is
more difficult to deal with in reality than the task of establishing datum pattern D for
purposes of fixturing and measurement.
For example, if we merely visualize the second pressing operation, in the bed of the press
exist four pins .250 in diameter, perpendicular to the bed and separated by the basic
dimensions shown on the product drawing. If I place the still-rectangular part into the
press, slip the four holes punched in press one down over the four pins in press two, and
bottom out in the bottom of the press's bed, I have accomplished two things: (a) I have
fixtured the part by establishing datum plane A and datum pattern D, and (b) I have
inspected/gaged the position of the four holes to one another and datum plane A.
The second pressing operation is implied to be based on datums that exist at that
timedatums A and D. Since the three-hole pattern is positioned to datum plane A and
datum pattern D, the second press punches out the 3 holes .510.520. They form datum
pattern E. Again, datum pattern E generates a central axis dictated by the axes of the
individual holes' virtual condition boundaries located at their true position from datums A
and D. Just as before, the datum axis is difficult to pinpoint and measure from, but that
task is not necessary. We must establish the pattern datum, if it is subsequently listed in a
feature control frame, and gage the features controlled from the pattern datum. To do this,

 Page 60
we don't have to find the centroid axis E. As a matter of practicality, once the pattern
datum is established, we can measure/gage the features controlled to it by using the
physical embodiment of datum pattern Ethe fixture/gage pins. These pins are represented
at their true position to datums A and D and made at .510 (their virtual condition size).
There remains one press and one feature control frame. The outside of the part is profiled
to datum plane A for line element perpendicularity and located from compound pattern
datum D-E invoked at MMC. Any deviations allowed in the basic size, form, and location
of the profile are stated in the final feature control frame. This part may only shrink from
these basic dimensions and relationships because the profile tolerance is unilateral inside.
The reason for this is that the mating cavity uses all the same basic dimensions and
relationships to shafts (instead of holes) with the same virtual condition boundaries. It,
however, is profiled unilateral outside (the cavity may only grow). This will give a mating
minimum gap (worst case) of zero and a maximum gap (not including pattern shift) of the
.010 on this part plus the allowed profile tolerance on the mating cavity. If one were to try
to accomplish a similar mating design condition for irregular shapes such as these,
without basic dimensions from specified datums and feature control frames of position
and profile, it would be infinitely more difficult, if possible at all. Plus and minus
tolerancing for all this geometric definition is anemic by comparison because of
ambiguity and tolerance accumulation.
We still must complete the final pressing operation. To do this, in the bed of the last press
we must establish the compound pattern datum D-E at maximum material/virtual
condition. To do this, we will have seven pinsfour at .250 and three at .510. These
pins are perpendicular to datum plane A represented by the press bed and separated by
the basic dimensions taken from the part drawing representing the true position of the
holes being fixtured and simultaneously gaged. If the holes pass over the virtual condition
pins and seat on the datum feature simulator for datum A, the part is fixtured for the final
pressing operation to profile the part, and the holes have been gaged in their relationship
of perpendicularity to datum A and their position to one another.
Once the part is fixtured, the final press is activated and creates the profile of the
periphery. The sequence of events is complete and the part finished. Three geometric
definitions are toleranced in three feature control frames, sequentially built on one
another and accomplished in three punch presses. The manufacturing operations are
implied by this type of sequential feature definition. Of course, the manufacturing facility
does not have to follow the implied sequence and can choose an entirely different
procedure; but, certainly, one has been implied. The final part, no matter how it has been
manufactured, must comply with the tolerances shown on the drawing.
In reality, another entirely different approach could have been used and given us a part
that would conform to the same geometric definition. For example, see Figure 4-5.

 Page 61

FIGURE 4-5
Figure 4-5 implies a part produced in one pressnot three. There are no controls that imply
they are complied with first. No leaders, so to speak. All holes are positioned and the part
outside profiled as a simultaneous requirement. The interrelationships are created by this
requirement for implied simultaneous creation and required simultaneous inspection of
all features defined by the three controls. They are all part of one pattern, interrelated to
one another and to datum plane A.

 Page 62
In both of the drawings shown, the profile tolerance will not be allowed a growth. It is a
surface control. No surface controls are allowed a bonus tolerance (a growth in the
geometric tolerance zone) as the controlled feature changes size (grows or shrinks). The
profile of the surface tolerance is defined as .010 and remains at that regardless of the
feature profile produced size. However, the circled M that appears after the datum features
D and E in Figure 4-4 shows the inspectors that a pattern shift is allowed for the profile as
datum features D and E depart from their MMC. Also, the simple fixturing of both parts
in the two illustrations with all holes controlled at MMC shows that each hole may drift
from its true position as it grows.
For the second illustration (Fig. 4-5), the allowed shift of profile is less blatantly called
out but still logically deduced. The holes are positioned to one another at MMC. The axes
of the virtual condition boundaries are the origin of measurement, but they are imaginary.
They are located at true positionbut to one another. If the holes grow, they may move
away from true position by the amount of their growth.
Still, who is to say where the virtual condition axes reside. If the holes all grow and all
move in the same direction, the virtual condition boundaries may be viewed as at that
new location. If this is true, the result is that the profile, which is simultaneously
measured from the axes of the virtual condition boundaries, is viewed as having shifted
with the movement of the virtual condition boundaries. Therefore, the profile experiences
the same advantage as the pattern shift allowed the profile in the first illustration (Fig. 4-
4) when the datum features were controlled at MMC in that profile control.
The controls shown in both illustrations create the same tolerance zones with the same
compliance requirements, while each implies a different manufacturing procedure.
Inspection of either part may be dealt with to within various levels of satisfaction using an
optical comparator with a comparator chart (overlay/template) made specifically for this
part, or templates with fixtures and dial indicators, or fixtures and CMM probing with
appropriate software. And to a certain degree, the mating functional requirements are
capable of being inspected adequately with a receiver-type functional gage that consists of
seven virtual-condition-sized pins located at their true position in a cavity that is the size
and configuration shown by the basic dimensions on the external profile. If our part
drops into the cavity, clears the pins (sized at 4 .250 and 3 .510) and datum feature
A contacts the bottom of the cavity (with three high points of contact on datum feature
simulator A), all controls have been inspected except for the internal boundary of the
profile tolerance zone.

 Page 63

5
Principles of Measurement for Geometric Characteristics
5.1 Geometric Characteristic Symbols

 Page 64

5.2 StraightnessForm Category

Straightness can be two distinctly different controls: (a) a surface control or (b) a control
of derived median planes or derived median lines.
In both types of control, the feature control frame will appear in a view where the
controlled surfaces appear as a straight line.
When used as a surface control, for example, for a planar surface (see Fig. 5-1), the
tolerance zone is considered the distance between two parallel lines. The lines are
accepted as perfectly straight and a distance apart equal to the tolerance in the feature
control frame. Each line element of the considered surface must lie between two parallel
lines, which are the allowed tolerance apart shown in the feature control frame. These
controlled line elements are inspected in the same plane represented in the view showing
the control.
In the case of surface straightness of a cylindrical feature, each longitudinal element of the
controlled surface must lie between two parallel lines that are the allowed geometric
tolerance apart and in a plane common to the nominal axis of the feature.
As seen in Figure 5-1, unless otherwise specified, for any part controlled with surface
straightness, the envelope of perfect form at MMC must be preserved. So, where the part
is produced at its maximum material condition (MMC) limit, the straightness tolerance
must not extend above that point (it is not additive to the MMC size). However, it may
have its full straightness tolerance within the confines of that MMC envelope (below the
MMC size). Least material condition must not be violated at any cross section.
Since this control is a refinement of the form control imposed by the size limits, it must
have a straightness tolerance less than the size tolerance in order

 Page 65

FIGURE 5-1
An example of straightness of line elements of a planar surface.
to have meaning. No material condition symbols (modifiers) are allowable since the
geometric tolerance of straightness is not related to, nor additive to, the size. This means
no bonus tolerance can be drawn from the size as the part deviates from MMC. Since the
envelope of perfect form at MMC cannot be violated, no virtual condition or resultant
condition is created by this geometric surface control.
For straightness of the surface of a cone and its verification techniques, see Figure 5-2. It
shows a straightness control applied to line elements of a surface. It applies independently
to each line element in the view shown. The tolerance zones consist of two parallel lines
0.1 apart within which each point on the line under test must reside. Each line element of
the surface has its own tolerance zone and is independently verifiable.

 Page 66

FIGURE 5-2
Straightness of line elements of a conical surface.
The inspection set-up shown consists of one center, one fixed support, and one adjustable
support. The supports are set to allow the line element being inspected to be as parallel as
possible to the surface plate/test table. A dial indicator held on a stand traverses the line
under test while the stand slides along on top of the surface plate. The full indicator
movement may not exceed 0.1 for each line inspected. This straightness control limits the
pitting, bumping, and curving of each line element of the surface in the view shown.

 Page 67
The tolerance zones for straightness of any one surface are infinite. Each line element of
the surface must, independent of all other line elements, lie within its own tolerance zone,
which consists of two parallel lines. The surface may be readjusted for the inspection of
each line element to optimize the procedure and minimize the full indicator movement for
that individual line element. Once a line element has been inspected and accepted (with
the part stationary and the indicator moving longitudinally over it), the part is rotated and
another line element is inspected in the same way. This continues until all line elements
are inspected and accepted, or a bad one is found, or more realistically (if none are found
to be out-of-tolerance), until the inspector is assured that had all line elements been
inspected, all would have been accepted.
In straightness of a derived median line, the tolerance zone begins as cylindrical in shape
and runs through the feature's center. The controlled feature's derived median line must
lie within this tolerance zone. This derived median line, which is generated by the
controlled feature's surface, may bend or distort from a perfectly straight counterpart, but
must never violate this cylindrical tolerance zone (when the MMC material condition
symbol is stated, this tolerance zone may grow). When applied on an MMC basis, the
maximum straightness tolerance applicable to the produced feature is the specified
tolerance plus the amount the actual local size departs from MMC. As each produced local
size departs from its MMC, an increase in the local diameter of the straightness tolerance
zone occurs that equals the amount of the departure from MMC.
The perfect form at MMC envelope generated by the MMC size limit may be violated. It is
no longer sacred as it was in the case of straightness-of-surface element controls.
However, in straightness of a derived median line or derived median plane, a new
envelope is formed, and this new boundary must never be violated. In the MMC concept,
the new boundary is generated by the MMC size of the controlled feature and its
straightness tolerance (an external feature's boundary equals MMC plus straightness
tolerance, and with internal features the boundary equals MMC minus straightness
tolerance). These boundaries are often used to determine worst mating condition sizes for
mating features or functional gages (as in the MMC concept usage).
In these cases, size must still be verified but MMC limits are checked the same way LMC
limits are checked. They are both checked at every cross section for two point-opposing-
point violations of size. Once size is verified as being within specified limits, the form
(straightness) must be checked. If the feature of size is controlled with the MMC material
condition symbol, a gage can be made to check the form. This gage is sometimes referred
to as a functional gage or a receiver gage. This gage will be made to the virtual condition
of the feature of size it is to check. If the controlled feature is a male feature (shaft, tab,
boss, etc.), the receiver gage will resemble its female counterpart in shape (hole, key-

 Page 68

FIGURE 5-3
Straightness of a derived median line. Virtual condition,
functional gaging, and bonus tolerance.

 Page 69

FIGURE 5-4
Straightness of a derived median line (regardless of feature size and
boundary exceeding maximum material condition envelope allowed).

 Page 70

FIGURE 5-5
Straightness of the derived median plane verification.

 Page 71
way, ring, slot, etc.). And, of course, a controlled female feature will have a functional
gage that resembles its male counterpart.
Figure 5-5 shows the specification of a derived median plane control. the feature control
frame is associated with the size requirement. The geometric control includes a
straightness symbol, a geometric tolerance, and either an implied (RFS) or specified
material condition symbol (in this case, an implied regardless-of-feature-size modifier).
The straightness tolerance zone generated is the distance between two perfectly flat
parallel planes within which the derived median plane must reside. The derived median
plane may be discerned from the set-up shown. All median points may be found by
averaging every set of 180 opposed surface points. The inspection set-up shown is one
that employs adjustable supports, fixed supports, a surface plate, and dial indicators. The
median points are found by comparing the readings on the indicators at M1 and M2. This
may be viewed as each indicator recording shows a height from the surface plate to the
surfaces. Once the center point is found, it should be plotted on a graph. If enough
median points are plotted, a graphic display of the derived median plane emerges and is
assessed for compliance with the .025 geometric straightness tolerance.
All median points must reside within the two parallel planes of the tolerance zone to be
accepted. Since no maximum or least material condition symbols have been used in the
feature control frame, no bonus tolerance is allowed. Had such modifiers been employed,
growth in the tolerance zone would have been allowed locally as each set of opposing
points departed from MMC or LMC (as appropriate to the modifying symbol used).
Figure 5-6 shows the specification of a derived median line control. The feature control
frame is associated with the size requirement. The geometric control includes a
straightness symbol, a diameter sign, a geometric tolerance, and either an implied (RFS)
or specified material condition symbol (in this case, MMC modifier).
The straightness tolerance zone generated begins as a perfectly cylindrical zone, may grow
locally, and must contain the feature's derived median line. In this case, the tolerance zone
is a diameter of .001 at MMC. The derived median line may be assessed by averaging
every set of diametrically opposed points. The inspection set-up shown is one that
employs centers and indicators. The median point is found by comparing the readings on
the indicators at Ma and Mb. One way to view this is that each of these readings uses a
vernier scale or electronic device (such as may be part of a height gage) to give the height
of the surface at Ma and Mb. These height readings are averaged to find the center and the
point plotted on a graph to show the centerpoint. If this is done many times, for many
different sets of diametrically opposed points on the surface, and all are plotted on the
same graph, a relatively accurate assessment of compliance with the geometric
straightness control can be made.
We will describe one method for taking differential measurements for a cylin-

 Page 72

FIGURE 5-6
drical feature with a straightness control of its derived median line. First, the axis on each
end of the feature must be established. Let us say, for this particular part, centerdrilling of
the ends is permissible. (Keep in mind this process is only as accurate as the centerdrilling
process.) On a lathe, each end must be closely chucked or colleted with as little of the part
as possible protruding. This is to allow the feature to turn true on the end and have a
better chance of hitting true center. Likewise, the centerdrill should not protrude from the
chuck in the tailstock farther than neces-

 Page 73
sary. Even though the centerdrill is hardened, it is prone to a minor amount of bend or
wobble if not properly chucked (which makes centerdrilling less accurate).
After being centerdrilled, these tapered holes can be used to place the object between
lathe centers for machining and then used to establish a perfectly straight axis with bench
centers for inspection purposes. Or, if the machinist is doing his own inspection, the part
can be rough checked while still between centers on the lathe.
Differential Measurements for Waisted or Barreled Features Only
Once the set-up is established, the cylindrical object should have the indicator run across
its surface from one end to the other while the cylinder is stationary. The inspector
records the variations at every cross section. For example: .000, +.001, +.002, +.001,
+.000, etc. (See Fig. 5-7.)
Then the object should be rotated 180 and the same routine followed, recording the
longitudinal variations of surface straightness. When this is complete, the inspector
should rotate the part periodically and record surface point variations for many opposing
line elements (at 180 separations). Then, with many sets of figures, a comparison of
these opposing figures should be made, subtracting one cross-sectional figure from its
opposing cross-sectional figure and dividing by two to get radial deviation from perfectly
straight.

For example:
These figures reflect the location of the points on the derived median line at these cross
sections and the line's deviation (if any) from perfectly straight. If any of these differences
when graphed or examined in terms of a diameter is greater than the tolerance in the
feature control frame, the part must be rejected or reworkedif possible.

FIGURE 5-7

 Page 74

FIGURE 5-8
In this control, the derived median line is allowed .009 out-of-straightness, and we can
see the maximum out-of-straightness (for these sets of opposing figures) is .005 radially
in both the positive and negative direction, or a diameter of .010. Therefore, this part is
rejected or reworked. It is important to note it isn't always necessary to take differential
measurements to establish derived median line or derived median plane straightness. This
method is used when the feature surface is: waisted (Fig. 5-9a), barreled (Fig. 5-9b), or
complicated by other odd surface variations. If, however, the feature of size is simply
bowed (Fig. 5-10a) the feature derived median line should be closely reflected in the
surface and perhaps, for your needs, an indicator check of the surface from end to end in
several places will suffice (Fig. 5-10b).

 Page 75

FIGURE 5-9

FIGURE 5-10

 Page 76

5.3 Flatness Form Category

The regular flatness control looks like this: , which means the entire surface of
the controlled feature must lie within a total wide tolerance zone that is the .001 distance
between two parallel planes. Size requirements are separately verifiable and must be met.
The flatness tolerance is most often merely a refinement of the form control given by the
size requirements of an object (under Rule #1). Therefore, on rigid objects, the tolerance
given in the feature control

FIGURE 5-11

 Page 77
frame for a single planar surface must be less than that tolerance given for feature sizes
that involve that single surface in combination with an opposing surface.
Flatness is a geometric control in which a part surface (feature) is compared to a perfectly
flat geometric counterpart of itself. A part surface is real; therefore, it has flawsridges,
grooves, pits, bumps, etc. Since a feature cannot be flat to something else, as in
parallelism or perpendicularity of one feature to another, no datums are proper or allowed
in the feature control frame. So, we check the surface to see how much irregularity an
indicator registers when it is run across that surface. Since we are not checking for
parallelism, we may not get an accurate reading if we simply set the part down on its
opposing surface when running the indicator over the controlled surface and reject the
part on that basis (although it could be accepted on that basis). Therefore, at times, we
must look into other ways of setting up the part.
We could, for instance, use three jackscrews set on top of a surface plate and at the same
time underneath the part. The jacks could be adjusted until the top surface (controlled
surface) is parallel to the surface plate top (see Fig. 5-12 and Fig. 5-13). Once the
controlled surface is as level (parallel) as possible, an indicator on a height gage or
surface gage (which sits on and runs across the top of the surface plate) is put in contact
with the controlled surface and pulled

FIGURE 5-12
Flatnessverification techniques.

 Page 78

FIGURE 5-13
Flatnessverification techniques.
alongregistering surface deviations. The deviations registered must be smaller than or
equal to the tolerance in the feature control frame.
Another alternative for registering flatness deviations is to set the controlled surface in
contact with a surface plate that is equipped with a plungertype indicator protruding from
its surface and move the part over the indicator point, noting the full indicator movement
(see Fig. 5-12). Since this type of setup is rare, an option is to put parallels on the surface
plate and put the controlled surface on the parallels. Then, with the height gage and the
indicator, indicate the controlled surface for deviations in flatness.
The tolerance zone is to be considered the distance between two parallel planes.
Consequently, all elements of the controlled surface must be between two parallel planes
that are the distance apart reflected by the tolerance in the feature control frame.

 Page 79

5.4 Circularity Form Category

Circularity is a two-dimensional surface form control. The surface is assessed to
determine that all circles that make up the surface being controlled are within the given
circularity tolerance. Each circle is separately verified and without regard to datum
features. The circularity tolerance zone consists of two concentric circles the distance
apart radially the amount that appears in the feature control frame. The centers of the
measured polar profile may be used to determine the out-of-roundness. The centers that
are used, unless otherwise specified, are determined by the minimum radial separation
center. Minimum radial separation (MRS) is the center for which the radial difference
between two concentric circles that just contain the measured polar profile is a minimum
(also known as the center for minimum full indicator movement, FIM).
Other methods of finding the centers used for surface evaluation are least squares circle
(LSC)the center of a circle from which the sum of the squares of the radial ordinates of
the measured polar profile has a minimum value; maximum inscribed circle (MIC)the
center of the largest circle that can be inscribed within the measured polar profile; and
minimum circumscribed circle (MCC)the center of the smallest circle that will just contain
the measured profile. Most feature control frames contain only the minimum information
such as: but the ANSI B89.3.1 Standard for Measurement of Out-of-Roundness
allows for a much more complete control, such as: , which means that
this surface shall be round within .0002 in. As assessed by the LSC method with 150
cycles per revolution response using a .003 radius stylus tip.
The minimum number of out-of-roundness measurements that are required to define the
surface of a three-dimensional body is not specified in the Y14.5 Standard. Sufficient
measurements should be taken to ensure that the measured profiles are typical.
Circularity is an interesting control, being at a glance relatively easy to understand, but on
closer scrutiny it can become so complex to accurately check that it warrants its own
ANSI Standard (ANSI B89.3.1).
Circularity is an unrelated control. Each feature is compared to a perfectly round
geometric counterpart of itself. No datums are required or proper. It is a surface control
wherein each circular line element of the surface is considered in a cross section
perpendicular (normal) to the feature center at that cross section.
In this two-dimensional radial control, each element of the surface must lie within the
boundaries formed by two concentric circles that lie the distance apart reflected by the
tolerance in the feature control frame.
The diameter surface in Figure 5-14 can be viewed as constructed by an infinite number
of circular line elements. Each line element has been given its own tolerance zone. These
tolerance zones each consist of two coaxial circles the distance apart radially that appears
in the feature control frame. The size tolerance

 Page 80

FIGURE 5-14
Circularity used on a cylindrical part.

 Page 81
given the diameter 1.2251.235 dictates the extent to which the surface may grow or
shrink, taper, or be out-of-straight.
Since each circular line element of the surface is independently verifiable, it is
conceptually possible that a surface may be acceptable that looks bowed, is shaped like an
hourglass, a barrel, or even a stack of coins with each coin displaced from what would be
a common axis for the cylinder (since each circular line element may be assessed for out-
of-roundness from its own center). The envelope of perfect form at MMC of 1.235 may
not be violated and the LMC may not violate a micrometer-type measurement (every two
diametrically opposing points may not measure less than a 1.225).
The conical feature shown in Figure 5-16 is placed on a set of supports to create the
optimum orientation for each measured cross section. The supports may be readjusted for
each measured section to allow the least deviation of circularity.
1. Either the spindle rotates 360 for a circular measurement of each cross section or the
table on which the part rests rotates (see Figure 5-15).

FIGURE 5-15
Verification techniques.

 Page 82

FIGURE 5-15
Continued
2. The probe or indicator is reset (zeroed) for each measured cross section of the part
surface. Each section is separately verifiable and must reside within its own tolerance
zone. Each tolerance zone consists of two coaxial circles 0.2 apart radially. Each measured
cross section of the surface has a tolerance zone appropriate to the size of the part at

 Page 83
that cross section. A point at the middle of each tolerance zone exists that is representative
of a point on the spine of the surface, from which the circularity of the cross section is
evaluated.
Circularity (roundness) constructs a series of tolerance zones that must not be be violated
by all points on all circular line elements under test that comprise the surface of the
controlled feature (see Figures 5-14 and 5-15). Circular elements of the surface are
established by cross sections perpendicular to a spine. A spine is a point or a simple non-
self-intersecting curve. In terms of inspecting a circularity control, measurements can be
taken perpendicular to this spine as

FIGURE 5-16
Circularity of a cone.

 Page 84
constructed by a simple surface. The spine can be thought of as similar to a derived
median line, and circular elements of the surface can be assessed for roundness from
points that comprise the spine. For a sphere, the spine is a point, but for a cylinder or
cone, the spine, as already described, is a simple, non-self-intersecting tangent-continuous
curve.
Each set of concentric circles of the circularity tolerance zones is centered on and in a
plane perpendicular to the spine at the circular element of the surface being measured.
The center from which each circular line element of the surface is measured may be reset.
This would allow configurations such as bowed surfaces, snaking surfaces, and surfaces
created by a stack of coins, wherein each coin is offset, to measure round, whereas
surfaces such as these would generally not meet a more three-dimensional form
requirement such as cylindricity. The actual value of circularity for any controlled circular
line element is the minimum radial separation of the two concentric circles closest
together still able to contain all elements of the circular line element being measured.

A simple circularity control may appear as: This means the two concentric circles
that form the tolerance zone must be .003 apart radially.
Each cross section of the circular feature must be such that all its surface elements fit
between its appropriate set of these two circles, regardless of the feature size. The center
from which the tolerance zone is established is important and can be established in many
different ways. The deferred practice is the center of the minimum radial separation
tolerance zone. This is the center for which the radial difference between two concentric
circles that just contain the measured polar profile is a minimum. Another method, which
creates the center from the sum of the squares of the radial ordinates of the measured
polar profile as a minimum value, is also popular. This is known as a least squares circle
center.
The method known as maximum inscribed circle (for holes) or minimum

FIGURE 5-17

 Page 85
circumscribed circle (for shafts) creates a center from which to measure. If the feature is
an external one, a shaft or cylinder, or a cone or sphere, one way the outer circle of the
tolerance zone can be established is by the minimum circumscribed circle or the smallest
perfect circle that can be drawn around the feature cross section touching high points.
The inner circle of the zone is then established concentric to the outer one. If, however,
the feature is an internal one (a hole or inside tapered cone or sphere), the inner circle can
be established first by the maximum inscribed circle or the largest perfect circle that can
be drawn inside the cross section touching high points. The outer circle is established
concentric to the inner circle.
The tolerance in the feature control frame is a refinement of the size control. Therefore, to
make sense, it must be smaller than the size tolerance, keeping in mind that circularity is a
radial control and size tolerance controls form on a diameter basis. The envelope
boundary of perfect form at MMC must not be violated. Therefore, the circularity
tolerance is not additive to the size tolerance (or size limits) and no virtual condition is
derived.
The MMC size limit is checked for any violation of this envelope and the LMC size limit
is checked at cross sections (every two opposing points). If the size limits are violated,
there is usually little reason to check the roundness of the feature because the feature
hasn't met its first requirement and must be rejected or reworked.
No material condition symbols are proper because circularity is always to be considered
regardless of feature size. Because of this (no allowed use of the MMC concept), no
bonus tolerance is ever possible.

FIGURE 5-18
Evaluation of roundness. (These procedures are not recommended
for final verification.)

 Page 86
Circularity (Roundness) Measurement Instruments
The Measurement of Out-of-Roundness Standard (ANSI B89.3.1) in discussing
instruments for measurement makes statements to the effect that out-of-roundness as
defined is usually measured by methods involving a stylus in contact with the part
surface. Analog or digital techniques are used to construct for graphical recording
(usually on a polar chart) the magnified radial movements of the stylus, as either the
stylus or the part is rotated around an accurately defined axis. (See illustrations in this
unit.)
While other types of roundness inspection are not forbidden, measurement of roundness
with two-point measurement instruments, such as a micrometer and vee-block
measurements (as shown in Figure 5-18), are flawed in that two-point measurement
methods can determine the out-of-roundness value only where the part is known to have
an even number of uniformly spaced and uniformly sized lobes or undulations around its
periphery. For parts having an odd number of lobes, the difference in diametral
measurements generally will be smaller than the true radial out-of-roundness value and
will diminish to zero for uniform symmetrically shaped lobing. Parts having an even-
lobed surface will produce diametral out-of-roundness values larger than the true value.
Vee-block measurements can be somewhat more useful than diametral measurements;
however, shapes with a known odd number of uniform symmetrically shaped lobes of
equal size and uniform distribution can be related by a conversion factor to the out-of-
roundness value when the vee-block of proper included angle is used. This is based on
the number of lobes and the vee-block included angle. However, there is no single vee-
block angle that will cover all numbers of odd-lobed parts. The vee-block measurement
system is not as useful for shapes with even numbers of lobes as the diametral method.
The vee-block tends to diminish the total indicator readings of even-lobed shapes,
sometimes to nearly zero. The two-lobed shape and the four-lobed shape in the 60 vee
will show only a slight variation in each total indicator reading. The major disadvantage
of the vee-block method is that it is not sensitive to all types of lobing.
It should be mentioned also that vee-block measurements are not two dimensional, but
instead the part rides on its highest peaks along the contact length with the vee-block
surfaces. One common failing of both the two-point and vee-block measurement
methods is the lack of a fixed center. This makes it incapable of stabilizing an axis or
centerpoint from which to assess the surface roundness. It is measuring from a point in
space that keeps moving.
Circularity is a condition of a surface of revolution where, for a cylinder or cone, all
points of the surface intersected by any plane perpendicular to a point on a spine are
equidistant from that point. This would indicate that a garden hose, for example, may be
round. A garden hose-type feature would be considered as round enough where all points
on every circular line element are equidis-

 Page 87
tant from its median point and lie within a tolerance zone of two concentric circles that
has a minimum radial separation equal to or less than the circularity tolerance specified in
the feature control frame.
When giving feature size limits, the abbreviation AVG. may be placed beside or beneath
the limits. For example, see Figure 5-19. This control involves maximum and minimum
dimensions that are averaged to determine an average dimension that must not lie outside
the stated average limitsfor example, the 1.000'' to 1.010'' stated in the figure.

FIGURE 5-19
The roundness tolerance is used to (among other things) set the maximum upper limit size
in the free state for any part made and also the minimum lower limit size in the free state
(unrestrained) for any part made. Besides lending other control, such as form, the
geometric (circularity, in this case) tolerance may also be used to recheck the average
dimension. For example, if you take two size checks (at least four at each cross section is
the usual minimum to establish an average for that cross section) and get, say, 1,003 on
the first and 1.009 on the other, and then average them:
1.003
+ 1.009
2.012 divided by 2 = 1.006 AVG. (see note below)
Because 1.006 is the average, one must not exceed a maximum feature dimension of 1.006
plus the .010 roundness tolerance or 1.016" or a minimum feature dimension of 1.006"
minus the .010 roundness tolerance or .996".
Note: No dimension on any of the parts checked may exceed:
1.010 + .010 or 1.020 high limit, or 1.000 -.010 or .990 low limit.

 Page 88
ApplicationsChapter Example
Size and Circularity Tolerance
Using the drawings in Figure 5-20, answer the questions that follow.

FIGURE 5-20

 Page 89

5.5 Cylindricity Form Category

As shown in Figure 5-21, cylindricity is a surface control of three-dimensional form. It
controls circularity, straightness, and taper of the surface.
The feature control frame contains a cylindricity geometric characteristic symbol and a
geometric tolerance. This cylindricity control is (for rigid parts) a refinement of the form
control given in the size tolerance of the cylinder. Unlike the size tolerance, which is given
as a diameter and would allow, for example, a flat on one side of the part to be as deep as
the entire size tolerance, the cylindricity tolerance limits any deviation from perfect form
radially to within the amount given in the feature control frame.
The tolerance zone consists of two coaxial cylinders different in size radially by the
amount shown in the cylindricity control (in the case of the part shown at the top of
Figure 5-21, a .001 radial difference). The entire surface of the cylinder regardless of its
size (which is separately verified) may not have surface variations that exceed the
cylindricity tolerance zone. Pits in the surface may not be deeper than .001. Bumps may
not be taller than .001. Flats may not exceed .001. Line elements may not curve more than
.001. The part may not taper more than .001 on each side.
The tolerance zone cylinders, while maintaining their coaxiality to one another and their
radial difference in size of .001, may grow or shrink to contain the surface of the part.
Even though the size limits say the parts may measure from .990 to 1.010, each
individual part may not differ in size measurements on the diameter by more than .002 (as
dictated by the cylindricity control shown here). A taper of more than .001 per side of any
one part would create a situation where the entire surface being controlled would violate
the cylindricity tolerance zone.
The part may be inspected in many ways. The inspection set-up shown for the part at the
bottom of Figure 5-21 consists of fixed and adjustable supports being used to orient the
part so that its axis is coaxial to the axis of a rotating spindle. Attached to the spindle is an
extension arm holding a dial indicator. Once oriented on the machine table, the part is
inspected by rotating the spindle as it moves up or down the surface of the part with its
attached indicator in contact with the cylinder under test. The full indicator movement is
then compared directly to the geometric cylindricity tolerance allowed by the feature
control frame. The tolerance may not be exceeded if the part is to pass inspection.
Unless otherwise specified, neither part shown can violate an envelope formed by the
boundary of perfect form at the MMC size. When checking the size limits on both
controls (before you've checked for form compliance), the feature is checked for any
violations of that MMC envelope and for any violations of LMC at any cross section at all
180 opposing points. (A vernier caliper or micrometer is usually sufficient for the LMC
check.) Features controlled for

 Page 90

FIGURE 5-21
Cylindricity.

 Page 91
circularity and cylindricity are set up in a similar fashion to check for these form
requirements.
Many checking procedures are used to determine if, indeed, all elements of the controlled
surface are inside the tolerance zone. Placing the feature inside a vee-block and rotating
the feature while deriving a full indicator movement (FIM) is a rough check used to
check circularity and cylindricity. In checking circularity, the indicator is often stationary
as the feature is rotated 360. But in checking the cylindricity, the indicator is moved
down the feature longitudinally while the feature is rotated 360. This rough check is not
always an accurate indication of a feature's compliance to the stated tolerance. However,
because of the possibility of football shaped or other oddly configured features, a center-
to-center check, using the indicator FIM as described in the circularity unit, is a betterbut
still roughcheck for cylindricity (keeping in mind the longitudinal movement necessary
for the cylindricity check but not for the circularity check). In this center-to-center
inspection method, the FIM should reflect the radial out-of-cylindrical error.
The polar graph method is a more accurate inspection technique used quite successfully
by many manufacturing facilities. With this method, a precision spindle with work-
holding capabilities is used to grip the feature being checked. The feature can be rotated
360 and the uppermost portion of the feature aligned to the lowermost, creating an axis
to which perpendicular readings can be taken.
For example, see Figure 5-22.
For circularity checks, a probe will contact the workpiece at different cross sections
during separate 360 rotations. For cylindricity, the probe will move up and down the
controlled feature during 360 rotations. While this process takes

FIGURE 5-22

 Page 92

FIGURE 5-23
Measurement of the deviation from cylindricity with an
electric sensor and the specimen fastened between two tips.
place, a drawing (record) of the feature surface readings is being made. Afterward, an
overlay with concentric circles at specified distances apart (for example, .001 in. apart)
can be placed over the recorded surface profile to determine if all the feature surface
elements lie within the specified tolerance zone (at every cross section for circularity or
over the entire surface for cylindricity). The cylindricity feature control frame is directed
to the controlled feature by a leader line and may be shown in either view.
In Figure 5-23, position the specimen between the tips. Make sure that the slide bar is
parallel to the line connecting the two fastening tips. Scan the cylindrical surface to be
measured continuously with a constant scanning rate according to the generatrix method.
Draw up the measured profiles as a linear diagram.
In Figure 5-24, position the specimen between the tips. A length measurement device is
assigned to each test point on the generatrix to be measured, so that all test points of a
generatrix are determined simultaneously. Set all length measurement devices to the same
indicated value with the help of the reference cylinder. Scan the cylindrical surface to be
measured.

 Page 93

FIGURE 5-24
Measurement of the deviation from cylindricity with several
length measuring instruments and the specimen fastened
between two tips.

5.6 Parallelism Orientation Category

Parallelism is a member of the orientation family. It can be used to control the orientation
of line elements, surfaces, axes, and centerplanes. If used on features of size (such as
holes, shafts, slots, or tabs), it is normally employed as a refinement of, or to augment, a
positional control.
In Figure 5-25, the tolerance zone is the distance between two parallel planes separated by
.005 and both parallel to the datum plane (A). Datum plane A is formed by the three
highest points (minimum) of the datum feature. It is common for the primary datum
feature to be controlled for flatness so as to limit the amount of rock possible when
forming the datum plane. Lacking such a direct control of flatness, the form of the datum
feature is relegated to the size limits of 1.5001.520. This, albeit somewhat deficiently in
this case, makes the datum feature flat, depending on the size at which it is actually
produced to within .020 (if produced at the LMC of 1.500). The larger the part's produced
size, the flatter the datum feature must be. If the part is uniformly produced at 1.520
(MMC), not only would the datum feature have to be perfectly flat, but the controlled
surface would have to be perfectly straight, flat, and parallel to the datum (per Rule #1).

 Page 94

FIGURE 5-25
Parallelism. Surface to a datum plane.

 Page 95
Unless otherwise specified, a rigid part must have perfect form at MMC. Rule #1 does not
control the perpendicularity of the sides of the part. That would have to be toleranced
either on the field of the drawing or relegated to a general tolerance note for angles
(usually located in the title block of the drawing).
The illustration shown in Figure 5-26, although not a complete part definition, is more
realistic and complete than is necessary to discuss the concept of

FIGURE 5-26

 Page 96

FIGURE 5-27
Parallelism verification.

 Page 97
parallelism of an axis to a datum plane. The hole must first be positioned on the part (or
the part must be located to the hole), and then any refinements of parallelism applied. The
positional control creates a stationary cylindrical tolerance zone a diameter of .030 for a
hole produced at its maximum material condition (.369).
This tolerance zone is allowed to grow, but not to move, as the hole grows. If the hole is
produced at a diameter of .381 (LMC, the positional tolerance zone grows to a diameter of
.042. The positional zone is .750 from and parallel to primary datum plane A, intersects
secondary datum axis B, and is parallel to datum plane C (a fourth datum plane
established by a minimum of two high points of contact on the datum surface C).
The actual axis of the hole must reside within both the positional tolerance zone and the
smaller refining (RFS in this case) tolerance zone, which is a diameter of .005 (regardless
of the feature sizethe smaller .005 parallelism tolerance zone may not grow as the hole
grows). This smaller tolerance zone, which refines the parallelism of the hole, is parallel
to datum planes A and C but is not a stationary zone (as was the positional zone).

5.7 Perpendicularity Orientation Category

Perpendicularity is a member of the orientation family. It can be used to control the
orientation of surfaces, axes, and centerplanes. If used on features of size, it is often used
as a refinement of, or to augment, a positional control. It is also often used to orient
secondary datum features of size to primary plane datums. As shown in Figure 5-28,
perpendicularity often is a control of flatness at a 90 angle to the datum plane.
Another common use of perpendicularity is in the control of a hole or shaft to a datum
plane (see Fig. 5-29, illustration of perpendicularity of a .370.375 shaft to within a
diameter of .003 to datum plane A). In this type of control, the shaft (a feature of size)
may be modified with a MMC symbol. At .375 (MMC), a .003 perpendicularity applies.
The shaft axis must lie within a cylindrical tolerance zone that is .003 in diameter and
perfectly perpendicular to the datum plane formed by at least three high points of datum
feature A. Datum feature A is not a feature of size. Therefore, no material condition
symbol may be placed with it in the feature control frame.
As the .370.375 shaft departs in size from MMC (.375) to LMC (.370), the .003 tolerance
zone grows as follows:
.374 has a .004 tolerance zone
.373 has a .005 tolerance zone
.372 has a .006 tolerance zone
 .371 has a .007 tolerance zone
LMC .370 has a .008 tolerance zone

 Page 98

FIGURE 5-28
Perpendicularity. Planar surface to a datum plane. Tolerance zonethe
distance between two parallel planes.

 Page 99

FIGURE 5-29
Perpendicularity of an axis to a datum plane.
The actual shaft axis may lean from perfectly perpendicular to datum plane A only
enough that it never violates this cylindrical zone. In its interrelationship with datum plane
A, a virtual condition boundary is generated. The virtual condition boundary is equal to
.375 (MMC) plus the perpendicularity tolerance at MMC (.003). This equals .378.
The shaft surface may not violate the .378 virtual condition boundary. This cylindrical
boundary is constructed perfectly perpendicular to datum plane A.

 Page 100

FIGURE 5-30
Perpendicularity. Centerplane to a datum plane. Verification
technique. The gage block is the largest that will fit inside the slot.
The indicator runs perpendicular to datum plane C. The indicator
should only travel longitudinally on the gage block as far as the
slot is deep. If the FIM exceeds the reading in the feature control
frame (.003), then the feature is unacceptable. Conversely, if the FIM is
.003 or less, the feature perpendicularity is acceptable.

 Page 101
ApplicationsChapter Examples

FIGURE 5-31
1. Are Drawings 1 and 2 (Fig. 5-31) significantly different in what they control? If so,
how? If not, how are they similar?
Answer:
They are very similar in that they both create an envelope of perfect orientation at
MMC.
2. If parts were produced and inspected per Drawing 2, would it be similar to
constructing a functional gage for Drawing 1? If so, what are the similarities? If not, what
are the differences?
Answer:
Yes, both would verify an envelope of perfect orientation at MMC.

 Page 102

5.8 Angularity Orientation Category

Angularity is often called the father of the orientation controls because it can be used for
tolerancing all angles other than those that are perpendicular or parallel. It can tolerance
the angles of surfaces, centerplanes, or axes to datum planes, centerplanes, or axes. If
used on features of size such as holes or shafts, it is often a refinement of a specified
positional control. Although the tolerance zone is commonly the distance between two
parallel planes, a cylindrical tolerance zone may be used for cylindrical features and often
is when more than one datum feature is specified. It is also possible to specify a tolerance
zone that consists of two parallel lines by using a local note such as Each line element
beneath the feature control frame. In any of the above cases, the angle to be toleranced
must be shown as basic.

FIGURE 5-32
Angularity. Surface to datum plane.

 Page 103

FIGURE 5-33
Angularity. Verification of surface to datum plane.
The produced, controlled surface in Figure 5-32 must reside within the specified .010
tolerance zone to comply with the angularity control. This controls the surface's
straightness and flatness, as well as its angular deviations from 30, all to within the
specified geometric tolerance.
In Figure 5-33, a secondary datum has been added to the control for more restrictive
movement of the part's rotation so as to not have to rely on the view as an implied datum.
The sine plate represents datum A, and the rectangular set-up block stops rotation of the
part on datum A and represents datum B.
An example of a hole angular to a primary datum axis and secondary datum plane is
shown in Figure 5-34.
The positional control shown places the hole on the part, eliminating all 6 degrees of
spatial freedom to within a diameter of 0.5 mm to datum axis A (to hold the 45 angle and
10mm distance), to B (to hold the 30 mm distance), and to C (to hold the orientation).
The cylindrical tolerance zone is a logical component of the control to contain the axis;
thus the diameter symbol is included in the

 Page 104

FIGURE 5-34
A hole angular to a primary datum axis and secondary datum plane.
control. Likewise, in the angularity control (although not required), it is logical the
cylindrical tolerance zone is advantageous in refining the angular relationships to datum
axis A and datum plane C to within a diameter of 0.25 mm. The application of datum C as
a secondary datum stops rotation of the tolerance zone (and thus the hole's axis, which
must reside within the tolerance zone) about the primary axis A, fully restricting
orientation.
Orientation of Line Elements
In Figure 5-35, the angle of each line element must be within 0.01 of the basic 60 angle
to datum axis A. A simple way of inspecting this to estimate compliance using common
surface plate tools would be as follows:
1. Set the datum diameter in a vee-block. Adjust and clamp the part.
2. Put the vee-block onto a sine plate. Build the sine plate up with gage blocks to the
appropriate angle.
3. Use an indicator on its stand to traverse a single line element. If further adjustment of
the part in the vee-block is necessary, adjust until optimized (since no other datum feature
prevents rotation).

 Page 105

FIGURE 5-35
Inspecting angularity of each line element.
4. If the single line element is within the tolerance zone (which is two parallel lines 60
from datum axis A), repeat Step 3 for each line element that comprises the surface until
all are inspected and separately verified, or until one is found to be out-of-tolerance.
Angularity, perpendicularity, and parallelism can be applied to a surface as one entity.
When used as a line element control, it releases the lines that comprise the surface in the
view shown from all having to reside within one tolerance zone. It allows each line
element to reside within its own tolerance zone.

 Page 106
Each tolerance zone consists of two parallel lines that are oriented to the datum or datums
shown in the feature control frame. Each set of parallel lines is separated by the amount
shown as the geometric tolerance of line element orientation.
5.9 Profile of a Surface Profile Category
The Power and Versatility of Profile to Control Irregular and Unusual Features
Profile is the Y14.5 Standard's most versatile geometric control. It is capable of handling
many of the most difficult situations. When a person is having trouble geometrically
defining tolerances of size, form, orientation, and/or position of irregular features for
those configurations that the other geometric characteristic symbols are unable to handle,
profile may be unleashed upon them. Quite often, profile turns an otherwise difficult
problem of geometric definition into one easily solved. Some of the advantages of profile
are:
It may be used as a three-dimensional surface control as in or a two-dimensional line
element control with .
The tolerance zone may be either equal bilateral, unequal bilateral, unilateral inside, or
unilateral outside.
Datums may be used to orient or locate tolerance zones, or not used at all.

FIGURE 5-36
Step 1.

 Page 107
Step 1:
This controls the size and shape of the cavity in Figure 5-36 by setting the size and form
of the tolerance zone. This tolerance zone is all around and unilateral outside. The
controlled profile must be equal to or larger than the basic dimensions that define the
inner boundary of the profile tolerance zone.
Step 2 (Fig. 5-37):
This controls the size, shape, and perpendicularity of the surface elements of the cavity.
The tolerance zone consists of two uniform

FIGURE 5-37
Step 2 (top) and Step 3 (bottom).

 Page 108
Mating Parts

FIGURE 5-38
Profile: mating part dimensioning and geometric tolerancing of complex
feature surfaces. Step 3 (top) mates with Step 4 (bottom).

 Page 109
boundaries. The inner portion of the zone is sized by the basic dimensions. The outer
zone is uniformly .005 larger all around. The zone boundaries are perpendicular to datum
plane A. To conform, the produced surface must reside within (between) the boundaries
of the profile zone.
Step 3:
This controls the size, shape, perpendicularity, and location of the surface of the cavity.
Basic dimensions from the added datums B and C locate a profile tolerance zone that
already had size, form, and (line element) perpendicularity to datum plane A. If the
controlled surface, as produced, resides within the profile tolerance zone, it is within its
limits of size, form, orientation, and (now) location. If the mating feature in the next
illustration is dimensioned and toleranced in a similar manner, with its profile zone given
as unilateral inside, the controlled features stand an extremely good chance of mating at
assembly.
Step 4:
Figure 5-38 has the external counterpart for the cavity to mate with the design in Step 3.
The external feature fits into the internal feature and the datum features are representative
from part to part. How different would this be if the dimensions on both parts were all
given as plus and minus dimensions?
5.10 Composite Profile vs. Two Single-Segment Profile Controls
Like their counterpart controls in the position category, composite profile and two
(separate) single-segment profile controls can take on a difference in meaning in some
situations, while having the same meaning in others. For example, a composite profile
control, such as:

where datum A is used to control perpendicularity and datums B and C are for location,
the meaning is exactly the same as two single segment profile controls, such as:

 Page 110

FIGURE 5-39
Composite profile tolerance vs. two single-segment profile tolerances.
In both cases, if no datums appear in the lower level controls, the upper-level control acts
to confine the location to datums B and C and the orientation (perpendicularity) to datum
A of the oddly configured feature, while the lower-level control acts to define the limits
of the feature's size and shape/form/profile. Even if the two types of controls allow the
primary datum to appear in the lower-level controls, such as:

 Page 111

Composite Profile
Tolerance

Two Single-Segment
Profile Tolerances
the control remains identical in meaning. The addition of datum A in each of the lower-
level controls simply means that the job of refining the perpendicularity tolerance of the
line elements of the oddly configured profile now falls to the lower-level control.
Since the tolerance in the lower level of each control happens to be smaller, it is the one
that more closely controls the angle to A and, of course, still controls the size limitations
and the shape/form/profile tolerance. Since datums B and C have not been brought down
into the lower-level control (with the tighter tolerance), the relationship to B and C is
controlled by the upper-level control that includes B and C (with the looser tolerance).
In this example, the difference in meaning between composite profile tolerancing and two
single-segment profile tolerancing arises only when the datums used for location (B and
C) in the upper level control are brought into play in the lower-level control, as shown in
Figure 5-39 and below.

Composite Profile
Tolerance

Two Single-Segment
Profile Tolerances
These two controls are not identical in meaning. The composite profile control uses
datums in the lower control only to refine the angular relationship of the oddly
configured feature to those datums and may only use datums in the lower control and in
the same order as have been used in the upper-level control. Two single-segment profile
tolerances may use datums in the lower-level control that have been used in the upper-
level control in the same order or may use them in a different order. For example:

 Page 112

In addition to the versatility of being able to use the same datums in a different order, the
lower-level control in two single-segment profile tolerances may use entirely different
datum features than the upper-level control (which composite profile tolerancing is not
allowed to do). For example:

At any rate, the datums used in the lower-level control for two single-segment profile
retain their full implication and are not limited to an angular tolerance refinement of the
controlled feature. A datum that has a basic dimension originating from it to locate the
feature retains that locational relationship in a refining level control. Datums used in a
lower-level control with a tighter tolerance of profile refine the relationship between that
datum and the controlled profile for angle or location, not just for angleas is the case with
composite profile tolerancing.
In the example shown below for the composite profile control,

the upper-level control creates a stationary, total-wide, all-around tolerance zone that
confines the movement of the controlled feature surface. The surface of the controlled
feature must reside within the 0.5 tolerance zone located as shown by the basic
dimensions from datums B and C and perpendicular to primary datum plane A. The
smaller lower-level, total-wide, all-around tolerance zone of 0.25 may float in its location
to datums B and C but must maintain its perpendicularity to datum A and its angular
orientation (rotational relationship) to datum B. The actual surface of the controlled
feature must reside within both the 0.5 and the refining 0.25 tolerance zone to be in
compliance with this composite profile control.

 Page 113
If, instead, the example was of two single segment profile controls, such as

the upper-level control would create a stationary, total-wide, all-around tolerance zone,
which (like composite profile) confines the movement of the controlled feature surface.
The surface of the controlled feature would have to reside within the 0.5 tolerance zone,
which is located as shown by the basic dimensions from datums B and C and
perpendicular to primary datum plane A.
The smaller lower-level, total-wide, all-around tolerance zone of 0.25 may float in its
location only to datum C, but must maintain its locational relationship as shown by the
basic location dimensions to datum B (since B is listed in the lower-level refining control,
which also refines size/shape/profile) and its perpendicularity to datum A. This also has
the effect of (by default) refining the rotational relationship to tertiary datum plane C.
(This is a more restrictive control than the composite profile control described.) The
actual surface of the controlled feature must reside within both the 0.5 and the refining
0.25 tolerance zone to be in compliance with this two single-segment profile control. See
Figure 5-40 for more information on composite profile tolerancing.
5.11 Coplanarity (Profile for Coplanar Surfaces) Profile Category
When a surface is entirely interrupted by, for example, a groove (as shown in Figure 5-
41), it becomes more than one surface. Once it is classified as more than one
surface/feature, simple form or orientation controls often become ineffective. These
controls, such as flatness and parallelism, are unable to create one tolerance zone within
which multiple surfaces must reside. Enter profile. Profile of a surface, when used in this
type of situation, is capable of creating one tolerance zone within which more than one
surface must reside.
As shown in Figure 5-41, the tolerance zone of profile consists of two parallel planes .003
apart within which both surfaces must reside to conform. If in conformance, both
surfaces will not only be straight, flat, and parallel to one another to within .003, but they
will also be coplanar to within .003. If the tolerance on the 1.000 dimension was as tight
as the profile tolerance, the profile tolerance could be eliminated. The size tolerance
would, if that tight (a total of .003), control all of the same factors that the profile
tolerance currently controls.

 Page 114

FIGURE 5-40
Composite profile tolerancing.
That is not the case in Figure 5-41. The size tolerance is a total of .060 (plus or minus
.030). If not for the profile control, this would allow one of the surfaces to be lower than
the other by .060. The profile of the surface control allows the part to uniformly grow or
shrink by .060 (while not exceeding its size limits), but keeps the two surfaces under
control (flat and straight while also parallel and coplanar to each other) to within .003.
More control is possible, for example, if a datum feature were added as shown in Figure
5-42.

 Page 115

FIGURE 5-41
Coplanarity.

FIGURE 5-42

 Page 116
The controlled surfaces are now flat, straight, coplanar, and parallel, not only to one
another, but also parallel to datum plane A. Even more control is possible, for example, if
a basic dimension was to replace the toleranced height (see Fig. 5-43).
The controlled surfaces are now flat, straight, coplanar, and parallel to one another and
also parallel to datum plane A and located to datum plane A. The tolerance zone is located
equally distributed about the basic dimension of 1.000 located from datum plane A. Both
surfaces may be no farther than 1.0015 from datum plane A and no closer than .9985.
An equivalent control could use a dimension origin symbol instead of a profile control
and a datum feature. For example, see Figure 5-44.
This is a simple-looking control, yet it does what the basic dimension of 1.000 with a
profile control and a datum feature would do. The tolerance zone for both is as shown in
Figure 5-45.

FIGURE 5-43

FIGURE 5-44

 Page 117

FIGURE 5-45

5.12 Conicity (Profile for Cones) Profile Category

In Figure 5-46, a method is introduced for the creation of a three-dimensional-type
control of the surface elements of a cone. Just as cylindricity controls generate a tolerance
zone consisting of two concentric cylinders with a radial separation of the tolerance in the
feature control frame, the use of profile of a surface on cones can create a tolerance zone
consisting of two concentric cones, both with a basic angle of the cone given in the
drawing with a radial separation of the tolerance zone cones of the amount given in the
feature control frame.
As Figure 5-46 shows, the feature surface at the large diameter must stay within size limits
1.9902.010 and must lie between two coaxial boundaries .005 apart RFS. The tolerance
zone boundaries have an included angle of 20 basic. If the entire surface resides within
the confines of the tolerance zone, the surface is considered to be conical enough to pass
inspection.
A rough check of this control could be to insert the cone into the vee of a vee-block, then
place the vee-block on a sine plate. Build up the basic angle through the use of gage
blocks. This set-up could be accomplished by placing the sine plate, gage blocks, and a
dial indicator (with stand) on a surface plate. Once the set-up is complete, the part could
be rotated in the vee-block while the indicator traversed the highest line on the rotating
cone. The conicity may be assessed from the full indicator movement. Care must be taken
when using a vee-block set-up of this kind in that the vee-block is not capable of
stabilizing the axis of the cone. Surface imperfections may cause a false reading in that the
axis of the cone would move, causing the surface to moveadding to the reading of the
conicity error. Or, perhaps the axis would move and prevent the surface from recording
certain variations in conicity. When possible, stabilization of the axis is recommended
during inspection.

 Page 118

FIGURE 5-46
Conicity. Profile tolerance for a conical feature. Individual feature.
The vee-block is tilted up to attain the included 20 angle and the indicator run across the
surface as the part rotates 360.
To calculate gage block build-up height for a cone in a vee-block on a sine plate:
Formula #1: The angle to rotate the sine plate equals one-half the angle of the cone plus
the angle calculated by Formula #2.
Formula #2: The angle is calculated by the sine of one-half the cone times the cosecant
of one-half the angle of the vee-block.
Formula #3: The build-up height for gage blocks on the sine plate (or sine

 Page 119

FIGURE 5-47
Vee-block set-up for Figure 5-46.
bar) equals the length of the sine plate times the sine of the angle described in Formula
#1.
So, for example, a 20 cone and a 90 vee-block using a 5-in. sine plate would use the
following: Sine of 10 = .1736 Cosecant of 45 = 1.4142
Steps:
(Per Formula #2) .1736 1.4142 = .2456
(Per Formula #2) .2456 is the sine of 14.215
(Per Formula #1) 14.215 plus 10 equals 24.215
(Per Formula #3) The build-up equals 5X SIN 24.215
B.U. = 5 .4101618
B.U. = 2.0508
If, instead of using a Vee-block, the conical feature had been placed directly on the sine
plate, the calculation for gage block build-up would have been 5 SIN 20.
5.13 Profile of a Line Profile Category
Figure 5-48 deals with the profile of a line geometric characteristic. For profile of a line
geometric controls, in each cross section parallel to the plane of projection, oriented to the
appropriate datums, if any, the considered line profile must

 Page 120

FIGURE 5-48
Profile of a line: unilateral inside.
be contained between two uniform lines a distance apart of the given profile geometric
tolerance. The 2-D tolerance zone is defined with basic dimensions. This is different than
what we have discussed for profile of a surface. For profile of a surface geometric
controls, the considered surface must lie between two uniform boundaries of the given
profile geometric tolerance. The tolerance zone is defined by basic dimensions which
control 3-D form and is capable of controlling size, orientation, and location to datums.

 Page 121
This profile of a line control in Figure 5-48 applies to line elements between T and V. The
datums used do not control the height/location of the line elements from datums A and B
but, rather, the orientation of the line elements. The profile control has the effect of
slicing the surface up into an infinite number of lines, each of which may vary in their
height from datum A by the tolerance shown in the limit dimension of .865 to .885.
However, the shape of each line element must fall within its own profile tolerance zone,
consisting of two uniform line elements .007 apart and of the configuration indicated by
the basic radii and angles shown in Figure 5-48. This could allow abrupt surface
variations from line element to line element (cross section to cross section) since each
tolerance zone may begin at a different location (within the .865.885 range). It is also
permissible for the surface to taper from front to back within the size limits of the height
dimension (.865.885). The datums used in this instance will stop the tolerance zones for
all line elements from angling to the datums and, consequently, to one another.
As the tolerance zones maintaining their orientation to the datums approach the maximum
(.885) and minimum (.865) size of the part, the usable portion of the tolerance zone
diminishes, until at each of those heights the zone becomes zero. In other words, the
tolerance of part height is inviolate and may not be exceeded. It is, therefore, possible that
the two parallel lines that form each tolerance zone, and begin with a fully usable .007
zone, may see portions of that zone close up and become unusable while other portions
of the same zone remain with the .007 usable zone. These zones are used to control the
form and orientation of each line element.
ApplicationsChapter Examples
1. In a profile control of a curved surface, the profile must be dimensioned with:
a. plus and minus dimensions.
b. limit dimensions.
c. radial dimensions.
d. basic dimensions.
Answer: d.
2. What is the difference between profile of a line and profile of a surface?
Answer: Profile of a line is a 2-D element control. Profile of a surface is a 3-D surface
control.
3. For profile controls, datums are:
 a. allowable but not required.
b. not allowed.
c. required.

 Page 122

FIGURE 5-49
Answer: a.
4. If a datum (or datums) was used for a profile control, what purpose would be served?
Answer: Orientation or location.
5. Are material condition symbols allowed for the feature being controlled? Why or why
not?
Answer: No. Never for surface controls. However, profile does allow datum features
that use MMC or LMC concepts. (See Figure 5-49.)
6. If both profile of a surface and profile of a line are used on the same feature, which
must have the smaller profile tolerance?
Answer: Profile of a line must have the smaller tolerance.
7. How is profile checked?
Answer: There are many methods. Some of the most effective are optical comparators,
video comparators, coordinate measuring machines, templates and indicators, masters
with fixtures and indicators, and even receiver gages.
8. What is the tolerance zone shape of profile of a line?
Answer: Two uniform lines separated by the profile tolerance.

 Page 123
9. What is the tolerance zone shape of profile of a surface?
Answer: Two uniform boundaries separated by the profile tolerance.
10. How is profile tolerance designated as bilateral or unilateral?
Answer: If phantom lines are shown, they designate a direction of tolerance.
11. How is the extent of where the profile control applies on a surface designated?
Answer: With notes or between symbols or all around symbols.
12. Suppose the entire outer periphery of the part profile must be controlled. How may it
be done?
Answer: With either all-around symbols or an ALL OVER note. All-around is view
dependant, while the note ALL OVER applies in all views simultaneously.
13. Sketch below an optical comparator overlay that could verify the given simultaneous
requirements for Figure 5-49. Dimension tolerance zone size and virtual condition
boundaries being inspected by this overlay.
Answer: (Figure 5-50.)

FIGURE 5-50

 Page 124
14. What tolerances can this overlay not inspect?
Answer: MMC or LMC on datum feature B, LMC on the other two holes, part
thickness (.035.045), the flatness tolerance on datum feature A, and the
perpendicularity of all features to datum plane A.
15. What is needed to inspect those?
Answer: A GO gage for MMC of B, small hole gages for LMC of all holes, a height
gage for part thickness MMC of .045, a micrometer for thickness LMC of .035, and a
dial indicator for flatness and perpendicularity are some of the possibilitiesbut the
possibilities are many.
16. Are any special fixtures needed for any of these inspection procedures?
Answer: No. Generic tools found in most inspection departments should suffice.
Using Figure 5-51, answer the following questions.

FIGURE 5-51

 Page 125
17. How flat must datum feature A be when produced?
Answer: .002.
18. For the two holes INDICATED X, what are they being positioned to besides datum
feature A?
Answer: Each other.
19. How many holes does the profile control tie the profile to?
Answer: All fourtwo datum holes and two Z holes (through SIM REQT).
20. What acts as the origin of measurement for distance for the two holes that are
INDICATED Z?
Answer: B.
21. What characteristic of geometry does datum A tolerance when used in each feature
control frame?
Answer: Perpendicularity.
22. Is the profile control a total wide tolerance zone of .010 or .010?
Answer: .010.
23. What is the virtual condition of the X holes? the Z holes?
Answer: X holes = .281; Z holes = .281.
24. What acts as the angular orientation datum?
Answer: C.
25. If the holes are produced at .300, how much positional tolerance are they entitled to?
at .295? at .305?
Answer: .019; .014; .024.
26. Would it be more difficult for a coordinate measurement machine to establish the
datums shown or a pattern datum consisting of all four holes? Why?
Answer: Pattern of four holes; harder to find the central axis.
27. What measurement equipment would you prefer to inspect the profile control, and
why?
Answer: Optical comparator with dedicated comparator chart; ease and accuracy.
28. If the part thickness is measured with a micrometer at .374 and the median plane bows
.003, is the part within its size tolerance?
Answer: No, it exceeds the envelope of MMC.
29. If the two holes INDICATED X are produced at a size of .300 and with their axes a
distance apart of 1.982, are they acceptable?

Answer: 1.963 .019 1.963 1.963

+ .019 - .019
1.982 1.944 Yes.

30. If the profile size shown as a basic 3.000 is produced at 2.980, does it exceed its
tolerance?
Answer: No.

 Page 126
31. If the two Z holes are produced at .305 and their axes at a distance of 1.099 from
each other, are they within tolerance?

Answer: 1.123 .024 1.123 1.123

+ .024 - .024
1.147 1.099Yes.

32. If the holes are produced at .300 and out of perpendicularity to datum A .015, are
they out of tolerance?
Answer: No, at .300 they are allowed .019.
33. If the surface profile has a bump that exceeds the basic radius shown by .005, is the
surface to be rejected as out of tolerance?
Answer: Yes. It is out of tolerance (profile is all inside) unless datum shift could bring
bump into tolerance zone.
34. If the holes measure with a micrometer a .299 and can receive a gage pin (without
regard to orientation or location no larger than a .292, do they exceed their size
tolerance?
Answer: Yes. They violate Rule #1 (must receive .295).
35. If the largest gage pin that can be inserted into datum feature B measures a .305, the
largest gage pin that can be inserted into datum feature C measures a .295, and a vernier
caliper measurement from the outermost points on the gage pin diameters is a distance of
2.282, would that indicate that they exceed their positional tolerance?

Answer: .305
+
.295
.600
(.600 divided by 2 =
.300
.300)
+
1.963
 2.263.019 (half of .038)
(2.282 OK)

5.14 Circular Runout Runout Category

Circular runout is a two-dimensional control capable of maintaining to within a specified
tolerance the circularity and coaxiality of a feature to a datum axis. If used on a surface
that is 90 to the datum axis, it is capable of controlling wobble of that surface (of
stopping more material from residing on one side of the axis than the other). It is
inspected in 360 rotations at circular line elements. Each line element is inspected
independently.
A datum, compound datums, or primary and secondary datums are options available to
establish and orient the required datum axis. The controlled

 Page 127

FIGURE
5-52
Circular runout.
feature is then rotated 360 about the datum axis. A dial indicator is used on the feature to
establish a full indicator movement (FIM) at every cross section. If, at any of the cross-
sectional checks, the FIM exceeds the tolerance allowed in the feature control frame, the
feature is not acceptable. If, however, at every cross-sectional check, the FIM does not
exceed the allowed tolerance, the feature is accepted, provided it conforms to its size
requirements.
If used as a type of coaxiality control, this surface-to-datum axis characteristic can control
roundness (circularity) and concentricity.
For a control such as , where the surface being controlled is a diameter and datum
A is an axis, the following applies:
1. The surface must be round/circular to within .002 radially (per side). This would allow
no pits deeper than .002, no bumps taller than .002, and no flats flatter than .002.
However, a micrometer-type (two point, opposing point) measurement may record a
difference in the diameter of each circular line element of .004. (It must be remembered
that size tolerances must be met, as separately verifiable requirements, and may not be
violated per Rule #1.)
2. The effect of this control on points of the derived median line is that each point must
not be radially off the datum axis (eccentric) by more than .001 (eccentricity) or a
diameter of .002 (concentricity).
3. The requirements of number one (for the roundness) and number two (for the
concentricity) must be simultaneously met in order for the surface to be considered within
tolerance.

 Page 128

FIGURE 5-53
Circular runout verification.

5.15 Total Runout Runout Category

Total runout is a three-dimensional control capable of maintaining to within a specified
tolerance the form of a surface (most often the cylindricity) and coaxiality of the feature
to a datum axis. If used on a surface that is 90 to the datum axis, it is capable of
controlling the flatness and the perpendicularity of that surface to the datum axis. It is
rotated about the datum axis. The indicator is then moved down the controlled surface.
From end to end and in all the revolutions, the FIM may not exceed the tolerance that
appears in the feature control frame.
Total runout is a composite control of surface elements. It is a three-dimensional control
of roundness (circularity), straightness, angularity, taper, coaxiality, and profile of a
surface. For features constructed at 90 angles to the datum axis, it controls
perpendicularity and flatness.
The same procedure for establishing the datum axis for circular runout applies to total
runout. But in total runout, when the controlled feature is rotated 360 about the datum
axis, the dial indicator is used to establish a FIM for the entire surface. While the indicator
is in contact with the surface and this controlled feature is rotated 360, the indicator is
moved longitudinally down the length of the feature. Total runout can be used to control
the entire length or any portion thereof. If only a certain segment of that surface is to be
controlled, a thick chain line can be drawn outside and next to the surface profile. This
chain line will use basic dimensions for length and location to define the segment of the
feature being controlled.
For a control such as as shown in Figure 5-54, where the surface being controlled
is a cylindrical diameter and datum A is an axis, the following applies:
1. The surface must be cylindrical to within .002 radially (per side). This would allow no
pits deeper than .002, no bumps taller than

 Page 129

FIGURE 5-54
Total runout.
.002, no flats flatter than .002, a total tapering of the surface of no more than .002 per
side, and allow the surface to be out-of-straight no more than .002. However, a
micrometer-type (two point diametrically opposing point) measurement may allow a
difference in measurements over the entire length of the diameter of .004. If the size limits
are less restrictive than this control, some geometric imperfections, such as surface taper,
which may be allowed by the size tolerance, would not be allowed (and would be
overridden) by the total runout control.
2. The effect of this control on points of the derived median line of the controlled surface
must not be off the datum axis (eccentric) by more than .001 or a diameter of .002
(concentric).
3. The requirements of number one (for cylindricity) and number two (for concentricity)
must be simultaneously met in order for the surface to be considered within tolerance.
Additionally, size tolerances must not be violated.

FIGURE 5-55
Total runout verification.

 Page 130

FIGURE 5-56
ApplicationsChapter Examples
Part 12135Bushing
1. Using the following variable data, assess the part (Part 12135Figure 5-56) and make
recommendations pertaining to:
a. the status of the part itself (what to do with it).
b. the manufacturing procedure being used to produce subsequent parts, if
appropriate.
Available Information:
Outside diameter measures .525 at its largest and .519 at its smallest.
Part outside diameter is round to within .0002.
FIM of I.D. while centered on and rotating about datum axis D is .002 at each cross
section.
Inside diameter is .2550 at its largest and .2549 at its smallest.
Part thickness varies between .196 and .195. It is rigid.
Part thickness has a slight uniform bow. It has a centerplane straightness of .003.

 Page 131
(Note: Some pieces of collected data may conflict with others. A major portion of this
exercise is to find out these conflicting pieces of data, state how they conflict, and
speculate as to how and why they may have occurred.)
Answer:

Outside
diameter
measures:
.525 at its
largest. OK
.519 at its
.525
smallest. OK
.519
.006 actual taper allows
taper of .005 per side (a .010)

Part outside diameter is round to within .0002. OK

FIM of I.D. while centered on and rotating about datum axis D is .002 at each cross
section. OK
Inside diameter is:
.2550 at its largest. OK
.2549 at its smallest. OK .002 must be off center (eccentric) because inside diameter
indicates shape is good.
Part thickness varies between .196 and .195. It is rigid. OK
Part thickness has a slight uniform bow. It has a centerplane straightness of .003. No.
Exceeds MMC envelope.
Three-Dimensional vs. Two-Dimensional Controls (Figure 5-57)
2. The inspection procedure for this part yielded the following data.
Evaluate with comments on measurement uncertainties.
MMC envelope verified with GO gages on outside and inside diameters
 Size checks taken at cross sections
I.D. = Between .769 and .771 down full length
O.D.= Between 1.260 and 1.261 down full length
FIM

.0003 at each circular cross section

I.D. =
checked
.0005 from end to end while
=
continually rotating 360
.0015 at each circular cross section
O.D.=
checked
.003 from end to end while
=
continually rotating 360

(Note: Some pieces of collected data may conflict with others. A major portion of this
exercise is to find these conflicting pieces of data, state how they conflict, and speculate
as to how and why they may have occurred.)

 Page 132

FIGURE 5-57

 Page 133
Answer:
MMC envelope verified with GO gages on outside and inside diameters.
Size checks taken at cross sections

Between .769 and .771 down

I.D. =full length
OK. Allowed .760-.780
between 1.260 and 1.261 down
O.D.=full length
OK. Allowed 1.252-1.272

FIM

.0003 at each circular cross section checked

I.D. =
OK for roundness
.0005 from end to end while continually
rotating 360
=OK for cylindricity but conflicts with I.D.
measurements of .769.771. The FIM should be
.001.
.0015 at each circular cross section checked
OK for circular runout. This indicates that the
O.D.=
diameter is off-center. .0005 or a .001 when
compared to its size measurements.
.003 from end to end while continually rotating
360
OK for total runout but this piece of
=information is in conflict with the other pieces
of information of size measurements and
circular runout measurements. The most the
FIM should be is .002 (.0005 + .0015).

Part 767-23Stud (Figure 5-58)

3. Using the following variable data, assess the part (Part 767-23Stud) and make
recommendations pertaining to (a) the status of the part itself (what to do with it) and (b)
the manufacturing procedure being used to produce subsequent parts, if appropriate.
Available Information:
Outside diameter measures between .5002 and .5003.
Pilot hole sizes vary locally between .126 and .128.
The pilot hole has a maximum inscribed cylinder of .120.
The C bore measures no more than .224 at all opposing points.
The C bore will receive a cylindrical pin to its full depth that measures no more than .220
in diameter.
The origin of measurement surface has some pits in it that are .007 deep.
All runout controlled surfaces have a FIM that does not exceed .0006.
The conical feature has a pit in it that is .004 deep and is eccentric to datum A .003.

 Page 134

FIGURE 5-58
(Note: some pieces of collected data may conflict with others. A major portion of this
exercise is to find out these conflicting pieces of data, state how they conflict, and
speculate as to how and why they may have occurred.)
Answer:

Outside Size
diameter: limits
.5010
.4990 OK

Pilot hole OK*

The pilot hole Mating size violates MMC envelope.*
The C bore measures OK
The C bore will OK
The origin of measurement OK
*If MIC is .120 and local measurements are no greater than .128, then the axis snakes (is out-of-
straightness) by about .008. to within .0006 says that isn't the case; therefore, the information in
incorrect on one of these three items.

 Page 135
All runout controlled OK*
The conical feature:
.004 (pit within limits)
.003 [eccentricity is out of limits (allowed only .0025 eccentricity)]
5.16 Concentricity Location Category
Concentricity is a geometric control of the median points of all diametrically opposed
elements of a figure of revolution. It is also a control of all correspondingly located
elements of two or more radially disposed features. If perfectly concentric, these median
points/elements coincide exactly in all their parts with the datum axis (or centerpoint).
This tolerance zone generated is cylindrical or spherical and coaxial with the datum axis
or centerpoint. The cloud of median points created when all diametrically opposed points
of the diameter are measured and the mean points found must reside within the tolerance
zone for the feature to conform to the concentricity tolerance. It is always applied on a
regardless of feature size basis. It is sometimes used where dynamic balance is needed for
rotating parts. Since parts may be accepted as within the concentricity tolerance with a
simple FIM inspection method, but not rejected without taking differential measurements
(because FIM also reads any out-of-roundness, which is not part of the concentricity
control), it is often recommended that runout be used instead. Runout controls are
capable of controlling concentricity characteristics while also controlling surface form
characteristics, such as roundness/circularity (circular runout), and cylindricity (total
runout).
To reject parts controlled with concentricity, it is necessary to determine where the actual
diametrically opposed elements and their cloud of median points for the controlled
feature are in relation to the datum (axis or centerpoint). This is done by taking
differential measurements on the controlled feature. This method filters out deviations in
the surface form. Differential measurements take readings of surface elements of the
controlled feature at 180 opposing points. The median points located through this
method must reside within the cylindrical (or spherical) geometric tolerance zone.
By touching the surface with probes or indicators on the datum feature as in Figure 5-59
(1), we are probably not able to comply with the same accuracy in establishing an
appropriate datum axis as when we use a collet-type chucking device as shown in Figure
5-60. However, when the use of such a device is not available, probes or indicators may
be used. When probes or indicators are used, as many points as possible should be taken
and the highest points probed used to establish the datum axis from the datum feature.
In the example shown in Figure 5-59, one should align the object with the measurement
equipment making the simulation of the datum feature axis coincident with the
equipment's rotating axis. Then (2) record the difference in radius at a predetermined
number of diametrically opposing points on the controlled feature

 Page 136

FIGURE 5-59

 Page 137

FIGURE 5-60
Differential measurements.
to determine the location of the median points. The departure from perfect concentricity
is calculated by using these median point locations and comparing them to the location of
the datum axis (as simulated by the axis of the measurement equipment). This method can
be applied to internal as well as external diameters. Its reliability is enhanced by
computer-assisted measurement equipment. It is also helpful if a polar diagram is
generated for a visual display of collected data.

 Page 138
ApplicationsChapter Example
Coaxiality-Type Options
1. Let us assume parts have been produced for each of the options listed. Although not
shown in Figure 5-61, the designer has allowed the machinist to center drill each end of
the parts. The inspector has chosen to use those centers, having first verified they are
(within acceptable limits) representative of the axis that would be generated by surfaces C
and

FIGURE 5-61

 Page 139
D if appropriately chucked simultaneously. Based on the inspection data given, assess the
parts using the chart at the bottom of Figure 5-61.
5.17 Symmetry Location Category

FIGURE 5-62

 Page 140
All median points of all opposed elements of the controlled feature of size (the slot)
shown in Figure 5-62 may form an entity that waffles, pits, bumps, or is uniformly
curved. It reflects the middle of every two opposing points of the slot and, so, the entity
goes where the surface takes it. In this example, all median points must reside within the
tolerance zone created perpendicular to primary datum plane C and centered about
secondary centerplane A. This tolerance zone is between two parallel planes that are .005
apart and equally distributed about datum centerplane A (.0025 on each side). If all
median points of the slot reside within the tolerance zone, the slot is within its symmetry
tolerance. The zone receives neither growth (no MMC symbol is allowed after the
geometric tolerance of .005) nor shift (no MMC symbol is allowed after datum feature of
size A). The regardless of feature size concept is implied throughout the control.
If a centerplane control of the slot's mating size is desired, a position control may be used
instead of a symmetry control. The position control is allowed to use the MMC, LMC, or
RFS concepts. As with all geometric controls, if the RFS concept is desired, no circled S
is needed, since it is implied (unless the circled M or L is stated).
ApplicationsChapter Examples
Symmetry
1. The following data has been collected with the set-up shown in Figure 5-63. The slot
size has been measured locally. Of all the two point opposing point measurements taken,
the slot has uniformly measured 19 mm wide. In position 1 shown, the height of the
surface is 120.47 from the surface plate. In position 2 (a 180 rotation from position 1),
the height of the opposing surface is 120.53 from the surface plate. Does this indicate that
the slot is outside of its symmetry tolerance?

Answer: No.
120.53
- 120.47
0.06 Deviation
vs. 0.08 allowed

Positional Symmetry
2. The following data has been collected with the set-up shown in Figure 5-64. The width
of the gage block (which is the largest that can be inserted into the slot) is 19 mm. The top
of the gage block measured in position 1 shown is 139.54. Does this indicate that the slot
is outside of its positional tolerance?

 Page 141

FIGURE 5-63
Position to Control Symmetry
3. Answer the following questions about the part in Figure 5-65 as produced:
a. Where is the centerplane of the part?
Answer: Centerplane of part is at .999.
b. What is the size of the part, not including the slot size?

Answer: .624
+
.621
1.245

 Page 142

FIGURE 5-64

 Page 143

FIGURE 5-65
c. What is the slot size?

Answer:1.998
- 1.245
.753 Size of
slot

d. What is half the slot size?

Answer: .753 2 = .3765 Half of Slot
e. How far is the slot centerplane from the top of the part?

Answer:.6240
+ .3765
1.0005 Slot centerplane from
top of part

f. How far is the slot centerplane from the bottom of the part?
Answer:.6210
+ .3765
.9975 Slot centerplane from
bottom of part

 Page 144
g. Starting with the answer to a) as a mean dimension, give the limits of the tolerance
zone.
_____ _____ as plus and minus.
_____ - _____ as a limit dimension.
Answer:

.999 .999.999 .001

+ .001 - .001
.998 - 1.000 Extent of centered
1.000 .998
tolerance zone

h. Is the slot centerplane within tolerance?

Answer: No.
i. If not, what is the minimum that can be removed from the top of the slot to be made
acceptable per the symmetry control? Use your answers to the questions in (a) through
(g) to prove your answer.
Answer: Slot must be reworked. Slot is .0005 too close to bottom of part. Remove
minimum of .001 from top of slot:

.623
+ .621
1.244
1.998
- 1.244
754 Size of slot
.754 divided by 2 = .377 Half of
slot
.623
+ .377
1.000 Slot centerplane from top
of part
.621
+ .377
998 Slot centerplane from
bottom of part
.999 .001 or .998 1.000

Centerplane of slot in tolerance zone

5.18 Inspecting Threaded Holes
The location of a threaded hole's axis is often incorrectly checked by putting an
unthreaded gage pin into the hole's minor diameter. The gage pin is then probed for
location only at the surface of the part. This method gives two-dimensional information
about a three-dimensional feature, and it gives that information about the location of a
point on the axis of the minor diameter rather than locating the entire axis of the hole's
pitch diameter (which is what is required).
Also, the tolerance zone, if not removed entirely from the threaded hole and

 Page 145

FIGURE 5-66
projected only outside of the part through the height of the axis of the screw or stud
above the interfacing/seating surface, will give information that cannot be relied on to
determine whether the part will actually assemble. The axis of the screw takes on the
location of the hole it is threaded into, but the angle is exaggerated. (see Figure 5-66).
This may allow a threaded hole to pass an inspection procedure, but if a gage screw was
threaded into the hole and its axis inspected for violations of a projected tolerance zone, it
would fail. The part must be inspected as it functions, or as close to that as is possible, or
functionality cannot be verified. Projected tolerance zones are recommended for threaded
holes or tight-fitting/press-fitting holes where a screw or pin will be inserted in a later
operation.
If a part using conventional positional tolerancing was given position using a projected
tolerance zone, it would appear as shown in Figure 5-67.
This assumes the threaded fastener goes through the mating part and the mating part has a
thickness of 1.500 or the threaded fastener (once screwed into the

 Page 146

FIGURE 5-67

FIGURE 5-68
controlled hole) will protrude 1.500 up from the top surface of the part. The overriding
factor in deciding the height of the projected tolerance zone is the height that the screw or
stud protrudes beyond the surface of the threaded hole. Some mistakenly believe the
tolerance zone, once projected, resides not only outside of, but also within, the part
surface (in the threaded hole). This is not true. Once projected, the positional tolerance
zone resides only beyond the confines of the threaded hole. The projected tolerance zone
could also be shown with the use of a chain line as in Figure 5-68. In Figures 5-67 and 5-
68, the tolerance zone applies above the surface of primary datum A. A functional gage
that simulates the mating part situation may be used to determine whether the feature axis,
when projected 1.500 above the part's primary datum plane, lies within the .005 tolerance
zone depicted in Figure 5-69.

 Page 147

FIGURE 5-69
Threaded Holes Treated as Screws to Calculate Virtual Condition
The virtual condition of a threaded hole controlled with perpendicularity or position with
a projected tolerance zone at MMC should be calculated exactly as though it were a screw.
Use the MMC size of the screw that will fit into the hole, then add that MMC to the
geometric tolerance ( or ) to get the virtual condition (MMC concept). (Note: This
also works for RFS conditions to calculate the worst mating boundary size.)
This virtual condition may be used to build a functional gage to check the tapped hole.
The functional gage can consist of a plate to represent the A datum. It will have a hole in
it the calculated virtual condition size. If the control is true position instead of
perpendicularity, the axis of the gage hole will be at true position as simulated from the
datum planes that make up the complete datum reference frame. With the actual part in
the datum reference frame, if the screw made for the tapped hole can pass through the
hole in the gage and screw into the actual part, the part is good.
The question is often asked as to whether the RFS or MMC concept should be employed
in threaded holes. When a screw is inserted into a threaded hole, there is some play (slop,
if you will) during assembly. This slop between the screw and threaded hole is the result
of the pitch diameter of the screw being smaller than the pitch diameter of the threaded
hole, and this slop or play helps us to more easily assemble the parts. Hence, there is more
slop if the class of fit is a 1 B instead of a 2B, and even less for a class 3B threaded hole.
The larger the pitch diameter of the hole, the more play (slop) we derive.

 Page 148
If the play is utilized to help us assemble the products, it should be allowed as additional
tolerance during the inspection procedure. This is bonus tolerance, and using the MMC
concept in the control recognizes that. If you can't measure the bonus tolerance, the
dimensioning and tolerancing engineer can suggest gaging procedures that will
automatically accommodate the allowed minimum bonus tolerance available for that hole
(for example, functional receiver-type gaging).
If you do not wish to utilize the bonus tolerance because you can't quantify its exact
number of thousandths of an inch or portions of a millimeter, that is your decision. Use
of the RFS concept in the control will negate it in the inspection procedure. Still, for a
tolerancing engineer (as merely a describer of physical phenomena) to say it doesn't exist
by putting in an RFS modifier instead of an MMC is not recommended. The RFS concept
implies that the hole contracts to meet the screw sizeand that is rarely true.
In visualizing the act of a screw and threaded hole assembling, don't view the additional
play (as the hole's pitch diameter grows) as a tolerance zone growth, rather see it as
allowing you to use the available play to push the screw over into the projected tolerance
zone of the hole. In inspection, see the gage screw as entitled to the same advantage,
because, however it is viewed, additional tolerance exists even if it is small and we can't
ascribe a number to it. As tolerancing engineers, we are tellers of physical truths. A
feature control frame is nothing but a sentence in the form of symbols strung together
like words and phrases. If we lie in a geometric control, the result is usually either: (a)
parts that work are rejected, or (b) parts that don't work are accepted.
Projected Tolerance Zones for Tight-Fitting Holes

FIGURE 5-70
Projected tolerance zones (verification without
functional gaging).

 Page 149
The reason we use projected tolerance zones on tight-fitting holes is to assure that once
the shaft is pushed into the hole, it won't lean too far to increase its effective mating size
and not allow it to mate with (fit inside of) the hole(s) on the mating part(s). The fit of the
shaft, although tight fitting into one part, is not necessarily a tight fit into the other
(mating) part.
Besides the use of functional gaging to check the location of this threaded hole,
verification can also be accomplished through the use of a pin threaded on one end (to
the depth of the hole) and with a smoothly ground (1.5 in. long) periphery on the other.
The threaded end could be screwed into the full depth of the

FIGURE 5-71
True position locators for threaded holes.

 Page 150
controlled hole, the part oriented to and positioned from datums A, C, and B and the axis
of the 1.5 in. length checked with conventional methods for location to determine
whether the axis of the hole, as dictated by the pitch cylinder, when projected over the
height specified (1.5 in.), exceeds the confines of the .005 diameter tolerance zone.
Threaded plug gage pins are often used to inspect the position of threaded holes that have
a projected tolerance zone. The pin is screwed into the hole. Then the CMM probe or the
dial indicator determines the hole location over the height of the tolerance projection.
5.19 The Boundary Concept
For meaning of Figure 5-72, see Figure 5-73.
The boundary concept used with conventional position controls has been expanded to
include geometric configurations not considered features of size

FIGURE 5-72
Elongated holes.

 Page 151

FIGURE 5-73
Virtual condition boundaries.
by definition. In the past the sole exception to the precedent of positional tolerancing on
features of size only was elongated holes. In past issues of ANSI Y14.5, positional
controls have been confined to cylindrical features, spherical features, features consisting
of two parallel plane opposing surfaces (all considered features of size), and the one
aberration allowed a position control: an elongated hole.
In all instances, except that of the elongated hole, it was considered that a tolerance zone
confined the axis or centerplane of a feature of size. In every case, the tolerance generated
a boundary that controlled the surface movement. In the case of the MMC concept, this
boundary is considered the virtual condition. For an internal feature of size (for example,
a hole) this boundary is smaller than the MMC by the geometric (position) tolerance that
applies at that size. For an external feature of size (for example; a shaft) this boundary is
larger than the MMC by the geometric (position) tolerance that applies at that size.
Only in the case of an elongated hole was the tolerance zone never considered at the axis
or centerplane, but rather only as a boundary that the feature surface could not violate.
With the latest revision of the ANSI Y14.5 Dimensioning and Tolerancing Standard, this
boundary concept has been ex-

 Page 152
tended to include any configuration, in the same way as it applies to the elongated hole.
For example, the 1994 ASME Y14.5 Standard shows an example of an oddly configured
hole defined by basic dimensions for size, form/shape and location (similar to Figure 5-
74). This feature is given a profile tolerance to create an inner and outer zone that controls
growth and shrinkage (size and form deviations). A primary datum feature is used in this
profile control to stabilize the part for purposes of manufacture and inspection of the
profile control. It is considered that this profile tolerance generates a type of maximum
material condition contour suited to this oddly configured hole. A positional tolerance is
then assigned to the hole with the word Boundary placed beneath the feature control
frame.
The Boundary note clarifies the fact that this is a specific type of positional requirement
that is not to be assessed for compliance on any basis other than whether the controlled
feature surface violates the boundary generated by the MMC of the contour minus the
geometric positional tolerance. The surface of this internal contour (hole) must lie outside
of the boundary generated. In this surface control, the boundary is generated by
subtracting only half of the positional tolerance from the perfectly positioned MMC of the
contour at the surface. Since this example is one continuous surface, the entire boundary
contour is created/calculated by reducing the MMC of the contour by the full amount of
the positional tolerance overall, but by only half of it in any radial direction.
A composite profile control would not have been able to accomplish the same effect. In a
composite profile control the tolerance to locate the feature from the datums would have
been the larger of the tolerances in the two levels of control, thereby allowing the second
level of the profile tolerance to tighten the form/contour tolerance (just the opposite of
what the example of the combined profile and position controls shown in Figure 5-74
accomplished).
Also, when a profile control is used with datum features to locate a contour, each level of
profile (no matter how many are used) generates two boundaries, an inner and an outer
boundary, which follow the basic profile and confines the surface of the contour. The
MMC positional control creates only one boundary (an inner boundary if it controls an
internal feature, or an outer boundary if it controls an external feature). This makes the
use of a positional control to limit the contour's movement/location less restrictive than
profile and perhaps easier to inspect, even opening the possibility of using a functional
gage, if desired.
A functional gage could easily be created to inspect the one constant boundary created by
the positional control. However, if a physical functional gage were to be employed to
inspect a profile control, it would be difficult for it to check both the inner and outer
boundaries created by profile. Still, it should be noted that functional gaging can be
accomplished by computer-generated mod-

 Page 153
els of boundaries and compared against data collected on parts checked by various means
such as computerized coordinate measuring machines. This methodology is often referred
to as soft gaging and is in many instances quite capable of inspecting a controlled contour
for one or two boundaries, thus making it able to verify either a position or a profile
control. For a visual example of this concept see Figure 5-74.

FIGURE 5-74
The boundary concept.

 Page 154

5.20 Composite Positional Tolerancing vs. Two Single-Segment Positional Tolerancing

Controls
One of the most sweeping additions made to the new ASME Y14.5 Standard is found in
the section distinguishing between composite positional tolerancing and two single-
segment positional tolerance controls. Examples follow.
If datum feature B is a planar surface from which a basic dimension originates to locate a
pattern of holes, these two controls are not identical in meaning in the tolerancing of such
a pattern.
Composite Positional Tolerance:

Two Single-Segment Positional Tolerance:

The essential element of holding the positional relationship between features within a
pattern is kept by both. The smaller tolerance between the two levels of control is the one
that (at least) refines the positional tolerance of the features to each other (the hole-to-hole
tolerance, for example). Also, if no datums are included in the lower-level control, or if
the only datum included happens to be the primary datum used in the upper-level control
and that is a datum used to control perpendicularity, then the two types of control [(1)
composite and (2) two single-segment position] are identical in meaning. However, if a
datum is used in the lower-level control that is either not used in the upper-level control
or is used in a different order than it is used in the upper-level control, or if the datum
used in the lower-level control is one used for location in the upper-level control, the
differences become evident.
In a composite positional tolerance, only datums used in the upper-level (pattern locating
tolerance zone frameworkPLTZF) control may be used in the

 Page 155

FIGURE 5-75
Composite positional tolerancing vs. Two single-segment positional
controls.
lower-level (feature-relating tolerance zone frameworkFRTZF) control. And, if used in
the FRTZF, they must be used in the same order as they were used in the PLTZF control
above. For example, a primary datum in the upper-level (pattern locating tolerance zones)
control may not become secondary in the lower-level (feature-relating tolerance zones)
control.

 Page 156

FIGURE 5-76
Composite positional tolerancing vs. Two single-segment
positional controls.

 Page 157
With two single-segment positional tolerance controls, the datum features used in the
lower-level control do not have to be in the same order or even be the same datum
features used in the upper-level control. In fact, since the two controls are entirely
separate positional tolerance controls for the same pattern of features, the use of datums,
or the lack thereof, is completely up to the discretion of the tolerance assignor (the
tolerancing engineer).
In a composite positional tolerance control, any datums used in the lower-level (feature-
relating tolerance zones) control are used to refine orientation only to those datums.
Angle tolerance refinement is the only use for datums in the lower level of a composite
positional tolerance control. Datums used for orientation/angle in the upper-level (PLTZF)
control refine orientation/angle in the lower-level (FRTZF) control. Datums used for
location in the upper-level (PLTZF) control refine only orientation/angle to that datum if
used again in the lower-level (FRTZF) control. This, of course, is in addition to what
might be considered the main purpose of the lower-level (FRTZF) control and that is to
refine the relationship between all features within the controlled pattern (for example,
hole-to-hole or pin-to-pin tolerance).
With two single-segment positional tolerance feature control frames, a datum used in the
upper-level positional tolerance control may have that tolerance refined in the lower-level
control, or an entirely different set of datums may be used. If datums used in the upper-
level control are also used in the lower-level control (and in the same order), they refine
the tolerance on whatever the datum was used for in the upper-level control.
Orientation/angular datums refine tolerance on the angle and location datums refine
tolerance on the location to those datums.

 Page 158

FIGURE 5-77
Composite positional tolerancing feature control frames. Only one possible
displacement of the feature relating tolerance zone framework (FRTZF) relative
to the pattern locating tolerance zone framework (PLTZF) is shown.

 Page 159

FIGURE 5-78
Two single-segment feature control frames.

 Page 160

5.21 Rules Sheet

FIGURE 5-79

 Page 161

6
Comparison of Controls
6.1 Food Chain of Symbology

FIGURE 6-1
Food chain of symbology. Geometric characteristics that control
form (excluding position and concentricity which do not control surface from).

 Page 162

6.2 2-D vs. 3-D Controls

Many two-dimensional controls have three-dimensional counterparts: straightness of a
surface (2-D) vs. flatness (3-D), circularity (2-D) vs. cylindricity (3-D), circular runout (2-
D at the surface) vs. total runout (3-D at the surface), as well as Profile of a Line (2-D) vs.
Profile of a Surface (3-D). These two-dimensional controls are not meant to tie together
the individual crosssections of a surface and put them into the same tolerance zones (as
their 3-D counterparts do). Think of the two-dimensional controls as a can of Pringle's
potato chips.
These two-dimensional line element controls give a separate tolerance zone to each potato
chip (cross-sectional line element) and subsequent zones are independently verified.
Consequently, the transitions between the line elements can vary within other constraints
placed on the part, like within the overall size tolerances, which are often plus or minus
toleranced.
A line element that holds its appropriate form in the view controlled is a good one, while
in other views the part surface can look like the Pringle's have been removed from the
can and subsequent chips displaced from others in the rowcausing the dreaded banana or
snake effect. Or, if some chips are smaller or larger than others, the transition from one
chip to the next is abrupt, staggered or waisted, barreled or conical.
In short, the line elements not being required to all be in the same tolerance zone (as
would be required of all elements of the surface in the three-dimensional counterpart type
control) but each being in its own separate zone, let the chips fall where they may.
Side note: They say a little knowledge is dangerous. T. H. Huxley (182595) once said If a
little knowledge is dangerous, where is the man who has so much as to be out of danger?
And as someone else once said The longest journey begins with a single step (probably
the same guy who said Go West, young man). Of one thing I am certain, we know more
now than we have ever known in the history of mankind, and yet we are as babes on the
verge of taking our first steps. My advice is to learn what you can now, use what you feel
comfortable with, and then learn more. One piece of knowledge should be viewed as the
beginning of a lifelong process. Learning is infinite. P.S. Huxley was a busy guy. He also
said, The great tragedy of Sciencethe slaying of a beautiful hypothesis by an ugly fact.
6.3 Comparison of Coaxiality Controls (Position, Concentricity, Circular Runout, Total
Runout)
Position
Position is an axis to datum axis control. It is potentially the least restrictive of the
coaxiality controls. It is the only coaxiality control that allows the use of the

 Page 163
maximum material condition or least material condition symbols after the geometric
tolerance and hence a bonus of geometric tolerance drawn from the pool of feature size
tolerance. It should be noted that the use of the maximum material condition symbol
affects the cost of the product in a positive way, lowering the potential product cost. It
can, however, allow some parts to be accepted that mate in a nonuniform manner.
More unequal airspace is possible between shafts and holes when bonus tolerance is
allowed, and when the maximum material condition symbol is used after datum features
of size referenced in the feature control frame, parts may be acceptable (within tolerance)
that can only be assembled off-center. This could endanger balance on spinning parts.
Therefore, the maximum material condition symbol used after datum features of size
referenced is usually reserved for nonspinning mating features. Position tolerances
impose no surface form controls. They are strictly controlled feature mating size axis to
datum axis controls (when used for coaxiality).
Concentricity
Concentricity is a median points to datum axis control. No maximum material condition
symbols or least material condition symbols are allowed inside the feature control frame.
Consequently, no bonus of geometric tolerance may be drawn from the feature's size
tolerance. This control is excellent to help in the balancing of spinning parts. It is an
easier geometric control for manufacturing to comply with than is a runout control,
because unlike runout, concentricity imposes no surface form controls. A full indicator
movement may be used to accept features as within geometric tolerance of concentricity,
but would not be valid for the rejection of a feature.
Since no roundness control is imposed by concentricity and a full indicator movement
registers both concentricity (centering) and roundness (form), the roundness aspect of
FIM would first have to be filtered out of the reading before a feature diameter is deemed
in violation of the concentricity tolerance. Because of this, it is considered by many to be
difficult to inspect. It is, however, only difficult to inspect if it is not round enough to be
accepted with a FIM. Then a procedure termed taking differential measurements is
commonly used, which compares 180 diametrically opposed points on the controlled
diameter to determine if the median point of every two diametrically opposed set of
points on the surface lies within the concentricity tolerance zone. This tolerance zone (as
with a positional zone for coaxiality) is cylindrical and centered on the datum axis. It is
interesting to note, though, that many advocate the use of runout instead of concentricity.
Actually, runout is a more restrictive control than is concentricity, since it combines the
geometric controls inherent in circularity and concentricity.
Generally speaking, since runout is more geometrically restrictive, one may assume that in
many instances it will be more costly to manufacture a part

 Page 164
that complies with a runout control than one that complies with a concentricity control.
However, since circular runout is usually considered easier to inspect, concentricity is
often viewed as the more difficult control overall. It would seem a choice must be made
as to whether or not the circularity aspect of runout is really needed functionally, and if
the answer is no, a decision must be made as to whether the parts produced should be
manufacturing intensive (in which case runout is chosen) or inspection intensive (in
which case concentricity is chosen). At any rate, these questions should be given their due
amount of consideration before concentricity is dismissed as the geometric control to be
used on any part.
Circular Runout
Runout is a surface being controlled to a datum axis. It allows no use of the maximum
material condition symbols or least material condition symbols inside of the circular
runout feature control frame. No bonus of geometric tolerance may be drawn from the
controlled feature's limits of size tolerance. Runout is considered a good control to assist
in the balance of spinning parts and, indeed, is inspected in a series of separately
verifiable 360 revolutions to read full indicator movements of each circular line element
of the controlled surface.
As mentioned previously, runout is often viewed as an easier-to-inspect control than
concentricity because runout combines the geometric aspects of circularity and
concentricity, which is what a full indicator movement (FIM) in a single 360 revolution
measures. Every circular line element of the controlled feature is expected to measure
within the tolerance stated within the feature control frame. This is done by directly
comparing the FIM to the tolerance of circular runout given in the feature control frame.
Even as the circularity aspect of this control makes a part easier to inspect than a
concentricity control, it often makes it more difficult and consequently more costly to
manufacture. One must decide whether he needs the circularity aspect of form control, as
well as the centering aspect of concentricity, before choosing this control, or is simply
willing to spend extra in the manufacturing process to make certain the part is round to
reap the benefits of this easier-to-inspect (runout) control versus plain concentricity.
If chosen, circular runout uses a FIM sometimes known as a total indicator readout (TIR)
in 360 rotations to validate or reject each circular line element of the controlled surface.
One FIM is required for each surface cross section (circular line element). The indicator
is reset at zero for every cross-sectional check so that each can be independently verified
in its relationship to the datum axis.
Total Runout
Total runout is an entire surface being controlled as an entity to a datum axis. It allows no
use of the maximum or least material condition symbols inside of the

 Page 165
total runout feature control frame. No bonus of geometric tolerance may be drawn from
the controlled feature's limits of size tolerance. Total runout is considered a competent
control to assist in the balance of spinning parts, and indeed is inspected with a
continuous revolution of the part about the datum axis while an indicator contacts the
controlled surface and moves down that surface in a longitudinal manner. All elements of
the controlled surface must simultaneously reside within the tolerance of FIM given
inside of the feature control frame.
Total runout controls cylindricity (if used on a nominally cylindrical surface) and
concentricity. It is a control of both form (roundness, straightness, and taper) and location
(centering). A good but geometrically restrictive control often used to assist in the form
and balance of spinning parts, it is usually considered a more difficult control for
manufacturing compliance than circular runout. It is also considered more difficult to
measure than circular runout, especially when the tolerances given are sufficiently tight as
to require the assistance of computer aided equipment and its accompanying software
(since more memory is required to assess the collected data).
Summary
Axis to datum axis control. Least restrictive of coaxial controls because it allows the
and bonus tolerance. could endanger balance on spinning parts. Least expensive
because it imposes no surface form controls and allows bonus tolerance. Used for
nonspinning mating features.
Median line to datum axis control. Allows no 's or bonus tolerance. Good balance
control for spinning parts. Easier to manufacture than runout controls. Imposes no
surface form controls. FIM (TIR) valid only for acceptance of features. Differential
measurements needed for feature rejection because of roundness factor. Considered
difficult to inspect.
Surface to datum axis control. Allows no 's or bonus tolerance as with all surface
controls. Controls roundness and concentricity (form and centering/coaxiality). Good
balance control for spinning parts. Considered easier to inspect than , but because of
roundness requirement, more difficult/expensive to manufacture. FIM (TIR) in 360
rotations valid to accept or reject parts. One FIM required for each surface cross section.
Reset indicator at each cross section.
Surface to datum axis control. Allows no 's or bonus tolerance. Controls cylindricity
(roundness, straightness, and taper) and concentricity (centering/coaxiality). Most
restrictive/expensive control to manufacture of the four coaxiality controls listed here.
Good for balance and form of spinning parts. Considered harder to manufacture and
inpect than but easier (if tolerances are loose enough) to inspect than . FIM (TIR) in
continuous revolutions while traversing entire longitudinal surface valid for part
acceptance or rejection.

 Page 166

6.4 Concentricity vs. Runout

Over the years you have probably either been involved in, or at least overheard, a
discussion or heated debate about concentricity versus runout.

Every year or so I write down an analysis of the controls to reaffirm my own knowledge.
Every year I come to the same conclusions, which is reassuring. But after a few months
and countless opinions from others, my fortitude begins to diminish, and eventually I
have to write it down again. Here it is for this year (see Fig. 6-2):

When the FIM or TIR is .005, it is a radial deviation from perfect runout of .005, which is
a .005 tolerance zone per opposing surface (per side).
This means the feature surface is simultaneously round to within .005 radially and
concentric to within a diameter of .005. The allowed radial deviation from perfect
coaxiality is .0025.
Example 1 (a good part): If size limits are met, a .005 bump or pit on each side of the part
at opposing points would be acceptable because that would be within the .005 radial out
of circularity component of the circular runout control, and at the same time keep the axis
to within the tolerance allowed (centered).
Example 2 (a bad part): A bump on one side of .005 and a pit opposing it (on the other
side) of .005 would not be acceptable because, although it would fall within the .005
radial component of circularity (out of roundness) allowed, it would not fall within the
coaxiality (concentricity) component of the control. It is only allowed to be out of
coaxiality .0025 radially or a diameter of .005. On this part, the indicator would read a
FIM (which we used to call TIR) of .010 (plus .005 on one side and minus .005 on the
other side) and the part would be rejected.
Example 3 (a good part): A bump on one side of .005 that opposed a perfectly smooth
point on the other side would be accepted. It would be out of round .005 radially, but
only off-center (out of concentricity) by .0025 radially (a diameter of .005) at that point,
which is acceptable. The FIM of .005 would register a part surface that is within the
allowed tolerance of circular runout to within .005.

 Page 167

FIGURE 6-2
People sometimes ask about the effect of using runout on the ability of a feature to mate.
Even though positional controls are normally associated with mating parts, even for
coaxial features, at times position is not the right control. For example, let's say we have a
multidiameter part composed of coaxial shafts that spin about a datum axis formed at the
center of one of the diameters. If I choose to give one of the other diameters a runout
control, how do I calculate what size diameter hole it will spin inside of?
For example: This part will mate with and spin inside of a counterbored hole that allows
at least a diameter of .510 for datum feature A and a hole that allows at least a diameter of
.265 for the controlled feature.

A total indicator readout (full indicator movement) of .005 of circular runout only allows
a radial deviation from perfect coaxiality/concentricity of .0025 (commonly referred to as
eccentricity). This is equivalent to a concentricity control that is a diameter of .005.
So, for analysis of coaxiality only (without the circularity/roundness as part of the
consideration) a circular runout of .005 is equal to a concentricity of .005.

 Page 169

7
Datums
7.1 When Sweet Seeds Grow Rancid Fruit (How to Select Datum Features)
Often the functional surfaces are completely ignored for use as datum features, and it isn't
until way down the line that someone realizes this can cause assembly and other
functionality problems. Many times, although the manufacturing procedure has been
adequately described on the drawing, it has nothing to do with whether or not the parts
will workand if inspected per those manufacturing datums shown on the drawing, the
only information or assurances that can be given by the inspectors are that the parts fit the
drawings. But they still will have no idea whether the parts will mate or function. And it
gets worse.
As one person told me, We have a problem at our company that could put us out of
business! We have been using datums that have nothing whatsoever to do with the way
our parts work.
Sometimes, when this is brought to light, the solution is suggested by someone: Add a
multitude of geometric interrelationships between those no-function datums and the
functional surfaces that have not been used. Extra parallelism, perpendicularity,
angularity, and profile controls are added, until the drawing is so full of geometric
symbology that the part configuration is hard to make out.
No problem, they say, add more views and more pages, a few more drawing notes here
and a home phone number there in case of confusion. Of course, these extra controls
mean extra manufacturing operations. Machining, forming, or molding costs grow
geometrically. And the end result is that looking at our drawings gives a visual impact so
full of geometric confusion it sends everyone for the Tums and aspirin. What can I do?

 Page 170
The purpose of the designing for manufacturability wave was never to increase product
cost and design out functionality. It was to get the design and manufacturing personnel
communicating as part of the same team, realizing their common goals, sharing their
knowledge and experience to turn out low-cost designs that preserve part function. As the
ads say, sometimes a great idea gets lost.
As I am fond of explaining, If manufacturing wants to use a catapult to fling my parts
high into the air and cannon fire to pepper them with holes, and when they fall to the
ground they wish to release rats to scurry up and gnaw in the final details, as long as I can
hook up a team of horses to drag the remains over to the inspection department and they
tell me the parts fit the drawing requirements (which reflect the functional needs), then I
will expire with a smile on my face.
A drawing should be a physical description (a factual statement) of how a part works, and
the datum selection is key to making it such. Concessions can and should be made to
manufacturing when they do not impair the functionality. Some additional controls will
occasionally be added to allow surfaces that are not as functional to be used as datums,
but for the most part, this is viewed as overdesign. Sometimes, the judgment of the team
will justify these concessions, but they must be made with the overall goal in minda low-
cost, functional product.
I would like to relate just a few of the procedures I have used as rules of thumb over the
years for the selection of symbology, datum features, and material condition symbols.
Please use these to spur your decision making, not to replace it. Like any tools, these are
only as good as the craftsman using them.
1. Select those surfaces that seat on one another as primary datum features if you wish
them to do so with good (for example three-point high-point) contact. Control these plane
surfaces with flatness (unless size limits do so sufficiently). Remember, we orient to and
measure from datum planes, and generally speaking, the better formed the datum
features, the closer they will be to these primary planes.
2. Watch not only the surfaces that seat on one another, but those that fit inside one
another, like functional diameters. They too make good datum features. If these diameters
are used as primary datum features, consider giving them a cylindricity control, to make
them more repeatable, to taper less, to be able to strike datum axes from which to
measure and orient. If a diameter is used as a secondary datum feature, consider either a
perpendicularity control back to the plane primary surface (if all other features on the part
are to be measured from these), or a positional control (if this datum feature is to be
measured from others for location).
3. Remember, not all functional features can be datum features. Sometimes it is easier to
work with a small number of datum features and derive your

 Page 171
functional requirements by controlling other important features back to that same set of
datums.
4. When several choices seem equally functional, choose the one that has the most surface
area and/or is the most accessible on which to set up.
5. A datum reference frame must not only relate to the drawing reader the important
interrelationships, it must stabilize the part and create an orientation that is a physical
description of how the part is oriented in its assembly.
6. Switching from one datum reference frame to another on one part or specifying
composite positional tolerancing often indicates that it is unimportant for all datum
features on mating parts to align. An error in alignment between original and subsequent
datum reference frames must be expected, as tolerance accumulation is created when
switching from one set of datums to another. This datum alignment between mating parts
is also not a priority when composite positional tolerancing is used.
7. Switching datums in midstream is just as likely an indication that the part has more
than one mating component. In these situations, one mating component may need to
relate to one set of datums, while another relates to a different set.
8. Just as often, if one mating part with many mating features mates with a single mating
part and its corresponding mating features, only one set of datums (one datum reference
frame) is used.
9. Try not to allow your datum features to conflict with one another. There are, after all,
only six degrees of freedom. Datum features that try to eliminate the same degrees of
freedom as others already in that control often conflict and/or confuse. Besides,
whichever datum gets there first gets the job.
10. Use the maximum material condition (MMC) symbol only when an alteration in the
size of the feature(s) or datum feature(s) really will have an effect on the allowed
geometric imperfection of the controlled feature or pattern of features. I always ask, Does
it mate? If so, and all I care about is that they go together, not with pins in the center of
holes (anywhere in the holes will do), and not with the datum features of size centered
between the mating parts, then circled M's all around usually works fine. When I see
people use the MMC symbols for parts that don't mate, or next to datum features of size
that don't mate while the controlled features mate, I automatically question it. Sometimes,
this is done only for the convenience of functional gage use, and although I can
understand that need in some instances, how functional can the gage be if it is checking
the part on a false basis? (Saying, in essence, that an alteration in the size of the datum
feature has an effect on whether or not the part will matewhen it really has none.)
11. Use of the least material condition symbol is usually an indication that mating is not
the most important requirement, but that it is the preservation of material on the part. One
example of this is when wall thicknesses are the

 Page 172
main consideration. Another might be on casting drawings where it is vital to preserve
enough material to allow clean-up to occur in accordance with a machining requirement
found on a subsequent drawing.
12. The regardless of feature size (RFS) symbol (or implied symbol) is of use when
describing that the worsening of a geometric shape or geometric relationship based on
feature or datum feature size is undesirable. For example, when balance is required, the
use of the RFS concept usually makes it easier to balance a part than would use of the
MMC concept (although RFS rarely assures balance by itself). It is especially useful after
cylindrical datum features of size when they are used for location on spinning parts. The
RFS concept is also useful for uniformity of fit, equal airspace between mating features
and uniformity of pressure for hydraulics.
7.2 Part Stabilization (Datums for Flexible Parts)
One goal of datum targeting is to achieve part stabilization for parts that are inherently
unstable, such as flexible sheet metal and some plastic parts. If these datum points, lines,
or areas of contact can also reflect the functional surfaces where a part seats in an
assembly, we have also achieved one of the other major goals of datum targets, or datum
referencing in general. We seek to construct a datum reference frame from which to
measure the subsequent part features. This is often a sequential process wherein the
datum targets or features are first defined, then formed and interrelated. Once this
structure is complete, we then reference other features on the part from the datum
reference frame. For flexible parts it is often necessary to define more than three points of
contact for the primary datum feature, more than two for the secondary, and more than
one for the tertiary datum feature. Some flexible parts are so flexible that they sag and
twist so freely that they may require more support for stabilization. Areas of contact are
frequently used to support the part and to restrain the part during inspection. Since all
parts are to be inspected in the free state unless otherwise specified, when restraint is
necessary it is important to write a note and include it as a local note for the feature or
features to be inspected, or to include this note as a general note applying to all features of
a part. The note may be as specific as you think is necessary to create part stability and
preserve functionality. Sometimes the simplest notes are used, just saying to restrain the
part during inspection, while other times the notes are detailed even to the point of
including the number of a special fixture to be used to create the datum reference frame.
Also helpful is the inclusion of a specific amount of force to be used in restraining the
part to the datum features or datum targets.
The use of restraint during inspection is often mistakenly thought of as restraint of any
kind, and since parts are not supposed to be restrained unless restraint is specified, many
think this rule is not practical. However, the use of

 Page 173
restraint is common on rigid parts during inspection because we know the restraint will
not distort the part and change the measurement data collected. So, to extend this
principle of logic, we can deduce that the type of restraint that is forbidden to flexible
parts that have no note that states the part is to be restrained during inspection is the type
of restraint that would distort the part and make it measure differently because of the
restraint being used. Any restraint that does not alter the inspection data is as valid for a
flexible part as it is for a rigid part.
These rules for restrained versus free-state inspection are so misunderstood in the
measurement community that a symbol has been created to reinforce the rules. A circled F
may be included in a feature control frame as a clarifying redundancy, to alert readers that
the part feature is to be inspected in the free state. Again, the rule is that the part is not
supposed to be restrained to such a degree that it would be distorted unless a note is
included to indicate restraint during inspection is permitted. But the use of the free state
symbol can save the involved parties a lengthy discussion after the fact about what the
intent of the designer and/or tolerancing engineer was.
The following is an example of a workpiece that uses notes to alert people to the fact that
the rules of drawing interpretation are to be in compliance with the ASME Y 14.5M-1994
standard on dimensioning and tolerancing, but that the part is to be restrained during the
inspection procedure. Also included is a general tolerance of profile of a surface that
applies simultaneously to all features not toleranced on the field of the drawing. In the use
of this note it is important that math CAD data be considered basic. If this is not stated in
the CAD file for this part, another note may be included on this drawing that states,
Unless otherwise specified all math CAD dimensions are to be considered basic or Unless
otherwise specified all dimensions in the CAD file for this part are basic.
One unusual item included in the section view of Figure 7-1 is the representation of a
fixture or gage block seated on datum target area A1 as reference. This type of depiction
is meant to help those who set the part up for inspection to understand that a fixturing
block is most appropriate in each datum target area to help establish the primary datum
correctly. Some measurement professionals mistakenly believe they are supposed to
probe the datum target area directly only in the center of the area, instead of the fixture or
gage block that contacts the high points of the surface. Sometimes using a datum feature
simulator such as a fixturing block is not practical and direct probing is necessary. But
when this is the case, the inspector must remember the goal is to simulate the plane that
would be given from the highest points of all primary datum target areas simultaneously.
Figure 7-1 uses three datum target areas for the primary datum, a hole for the secondary
datum feature (which will generate the second and third datum planes as the axis of the
hole), and two datum target areas for the third datum feature (generating a fourth datum
plane to orient the two crossing at the axis of

 Page 174

FIGURE 7-1

 Page 175
the hole) to give an angle of measurement. No measurements are actually to be taken
from the plane formed by the two datum targets labeled C1 and C2, but they will create
the angular orientation of the measurements taken from the datum axis of the secondary
datum feature B. Some might ask why two targets are used for C instead of one. Again,
we must remember we are beyond the tertiary datum plane when we get to C, and even if
we weren't, we are dealing with datum target areas here, not points. So, there is no reason
to believe we should use only one point or one area here. Each area generates an infinite
number of points. Our goal is to stabilize the part and consequently the measurement
plane, and if that takes one area or two or a hundred and two, we must do what is
necessary to create part stabilization.
Figure 7-1 uses the increasingly popular method of a minimum number of views (usually
one of which is an isometric) to show the configuration of the part, the datum features,
and the geometric controls. The other dimensions can be located in the CAD file for this
part. This has the advantage of creating a less cluttered drawing, with much of the critical
information. It is clear and simple and uses less paper. However, some have voiced
concerns over this trend in that it created the need for everyone who must make and
measure the part to have access to a terminal and the CAD file for the part.
Another tolerancing method that has been employed on this drawing is the implied
simultaneous requirement. All features controlled to the same datums in the same order,
using the same material conditions for those datum features referenced, are to be
considered as one pattern of features and inspected in one set-up or with one gage. This
simultaneous gaging requirement is created by the fact that all features on this part either
are the datum features referenced or are referenced from those datum features in the
order of A, then B at maximum material condition (virtual condition really, by virtue of
being secondary), then C. This is a more restrictive requirement than if they had been
referenced as separate requirements. As the datum features rock or as datum feature B
departs from its virtual condition, every feature that is part of the simultaneous
requirement must meet its tolerance in the exact orientation or shifted direction as every
other feature that is part of this requirement. This assures us that all features will function
as a pattern, for example assembling with one mating part in the assembly. Had features
been allowed to meet their tolerances while rocked in different orientations, or shifted in
different directions, we could not have been certain the part would interface properly
with one mating part. Each feature would essentially have been measured as though it was
an independent agent, perhaps each mating with a different part. The simultaneous gaging
requirement is a very powerful rule that is not always well understood by the creators and
readers of drawings. For that reason I sometimes choose to include the SIM REQT
abbreviation or the entire note Simultaneous requirement below feature control frames as
a clarification even when it is redundant. It helps those who are not as steeped in the
language and rules of geometric dimensioning and tolerancing as others. It is just as
important

 Page 176
that one include a SEP REQT or Separate requirement note when the same datums have
been used by multiple patterns of features but that may be verified by using the less
restrictive method of the separate gaging requirement. Separate requirements may be
verified at different orientations or different shifted locations as their controls allow. Also,
each may be verified by a separate gage.
7.3 Sequential Part Definition vs. Simultaneous Requirement (The Game of Leaders and
Followers)
One of the most effective, easy-to-follow methods of part definition is to sequentially
define datum features and then controlled features methodically from the primary datum
feature until the last part feature is tied either directly or indirectly into all previously
defined features. This linear progression of defined features assures that every part
feature has been defined and tied into every other part feature. Many product drawings do
not follow any particular method of this kind and therefore the definitions are chaotic and
incomplete. This linear progression may also be used to measure part features, measuring
the part as it has been geometrically dimensioned and toleranced and in the same order.
This is often called the game of leaders and followers. The datum feature defined as
primary leads into the datum feature that is secondary and so on, until the part is
completely defined and/or measured. Parts that are not sequentially defined run the risk
of either being poorly defined or being incompletely defined.
The part shown in Figure 7-2 uses a linear progression to begin with, defining the datum
features and their interrelationship to one another, but it also has a bit of chaos.
The progressive part begins with defining the datum features. Datum feature A is clearly
defined as three datum target areas. But there is also a bit of chaos here. Datum target
areas A1, A2, and A3 are not profiled merely to each other, but are also tied to datum
pattern B (the two-hole pattern). In a linear progression it could be said that this is
impossible since datum pattern B does not yet exist when datum targets A1A3 are being
defined, therefore they can only be related to each other. A more linear approach would
be to profile these three datum target areas only to one another using no datum features
inside their profile control. This would serve the purpose of forming each surface and
orienting and locating them to each other. Three tolerance zones would be formed at a
basic angle and distance from each other by using one profile control with three leader
lines, one to each target area. Once this was finished, datum pattern B would be defined
as shown, positioning the two holes to each other (for distance) and to datum A for
orientation. No further datum features are necessary since all other features on the part are
peers (equal) and will form a pattern of features all simultaneously defined and measured
from the datums formed by datum targets A1, A2, and A3 and datum pattern B. Since
every other feature will be measured

 Page 177

FIGURE 7-2
from these datums as a group, they all end up related to each other through their
tolerances to exactly the same datums.
Although it is common to tolerance parts in the somewhat out-of-sequence way this part
has been done (datum targets A1, A2, and A3 referenced to datum B before datum B
exists), it should be understood that when the primary datum target areas reference the
secondary datum feature, nothing additional is gained in terms of geometric
interrelationships. If datum pattern B is related to datum plane A, then the targets that
create datum plane A are automatically related to datum pattern B in the same way that if
you are 6 ft from me, then I am automatically 6 ft from you. On this part the use of datum
B in the profile control used on datum targets A1A3 doesn't hurt anything geometrically,
but it doesn't gain us anything extra either. There is something to be said for not
referencing B for the datum targets in that we aren't defining in circles if we use a more
linear progression. Defining in circles becomes a dizzying habit to get into and is hard to
break, even to the point of convincing ourselves that we gain more by doubly relating
everything back and forth to each other over and over again. If taken to extremes, this
methodology becomes confusing and clumsy. The example given above isn't too bad, but
anytime the sequencing is broken in such an obvious fashion, clarity of definition suffers
a little and those reading the drawing for purposes of manufacturing or measurement are
being asked to do something at least unnecessary and at most impossible since you are
essentially relating something to another thing that is not yet in existence.

 Page 178

7.4 Datum Features

Datum features should be:
functional (serving a purpose in preserving how a part operates),
representative of mating or seating features and/or alignment edges (to assure that if
inspected and accepted while oriented to, or located from, datums constructed from those
features, the controlled features will, indeed, mate and/or seat,
accessible to those in manufacturing and inspection,
repeatable from manufacturing operation to manufacturing operation and department to
department. Datum features are often made more repeatable in the planes, points, or axes
they generate by the form and orientation controls they are given. For example, the better
formed a primary datum feature, the more likely a subsequent set-up will be to recreate
the same datum. One should not give up part functionality to gain producibility or
inspectability, but producibility and inspectability should occupy places high on the
priority list when datum features are selected.
Datum features are chosen from real physical features on the part. They are chosen from
which to construct datums (imaginary points, axes, and planes). Datums are origins of
measurement. These imaginary datums do not exist in the real world, so we use
manufacturing and inspection equipment to simulate their existence. We recognize that
this equipment is not perfect, as are the imaginary datums, but is the best we can do to
recreate the perfect geometric counterpart of the datum surface(s). This simulation
equipment creates a reasonable facsimile of a point, axis, or plane from contact with the
datum feature surface. In making this contact, the part is stabilized. We then orient or
locate other feature surfaces, axes, centerplanes, or tangent planes from these simulated
datum points, axes, and planes.
Datum features are either permanent on the part or temporary (to be removed or altered
during the manufacturing process). Temporary datums are often used in-process from
which to measure other features. These other features are sometimes used as datum
features and establish the permanent datums to be used in the final inspection process.
Sometimes, because of rough surfacessuch as forgings or castingswhere machining is out
of the question because of cost factors, datum target points, lines, or areas are used
instead of entire surfaces from which to establish datums. This is also sometimes done
with weldments and similar parts prone to warpage or other distortions. Datum targets are
even occasionally more functional than would be the use of entire surfaces for the
establishment of datums for orientation or location. These portions of surfaces will often
act to create a datum reference frame to eliminate the six degrees of freedom (three
rotational

 Page 179
and three linear) a part experiences in free space. A drawing should be a physical
descriptiona factual statementof how a part works, and the datum selection is key to
making it such. Concessions can and should be made to manufacturing for ease and
consistency of set-up when these concessions do not impair the functionality. Some
additional controls will occasionally be added to allow surfaces that are not as functional
to be used as datum features, but, if overdone, this can make parts harder to produce.
Sometimes, the judgment of the team will justify these less functional datum features, but
this must be done with the overall goal in minda low-cost, functional product. Generally,
when this indirect datuming method is used, geometric tolerances must be tightened to
hold the more functional relationships needed, than if features are controlled directly to
the functional datums.
Establishing Datums with the Coordinate Measurement Machine
The establishment of datum reference frames has come under close scrutiny recently by
members of the ANSI and measurement communities. The Dimensioning and
Tolerancing Standard (ASME Y 14.5) makes statements to the effect that:
When magnified, flat surfaces of manufactured parts are seen to have irregularities;
contact is made with a datum plane at a number of surface extremities or high points.
Therefore, it is necessary to establish a method for simulating the theoretical reference
frame from the actual features of the part.
Planes are simulated in a mutually perpendicular relationship to provide direction as well
as the origin for related dimensions and measurements.
A datum feature is selected on the basis of its geometric relationship to the toleranced
feature and the requirements of the design. To ensure proper part interface and assembly,
corresponding features of mating parts are also selected as datum features whenever
possible.
If not sufficiently accurate, datum features from which datum planes and datum axes are
formed may need to be controlled by specifying appropriate geometric tolerances. When
control of the entire feature becomes impractical, use of datum targets may be considered.
But, if datum targets are not used, datum planes and axes are established by datum feature
surface extremities or high points. For example: a planar primary datum feature relates the
part to the datum reference frame by bringing a minimum of three high points on the
surface into contact with the first datum plane. The part is further related to the frame by
bringing at least two high points of the secondary planar datum feature into contact with
the second datum plane. The relationship is completed

 Page 180
by bringing at least one high point of the tertiary planar datum feature into contact with
the third datum plane.
The problems faced in complying with datum establishment procedures are complicated
by imperfect datum surfaces that rock or wobble when contacted at the high points. This
creates a candidate set of planes or axes from which to choose and from which to
measure. Where a nominally flat surface is specified as a datum feature, the
corresponding datum is simulated by a plane contacting the high points of that surface. If
irregularities on the surface of a primary or secondary datum feature are such that the part
is unstable (that is, it wobbles) when brought into contact with the corresponding surface
of a fixture, the part may be adjusted to an optimum position, if necessary, to simulate the
datum.
Unfortunately, the statement may be adjusted to an optimum position is so ambiguous it is
difficult to comply. Should one merely rock the surface one way or another and apply
pressure until it stabilizes, or should the surface be shimmed up until it no longer wobbles
or rocks? If it is to be shimmed, is one to use shims of equal heights on each end, shim
up only one end, or shim to create an equal angle (if the rock point is not in the middle of
the part)?
The ANSI committees have considered this problem and issued clarifying statements.
These statements guide us to the effect that a rocking datum feature (primary) or set of
datum features (primary and secondary) will generate more than one possible datum
plane or axis each, and/or multiple datum reference frames that may be established for a
single feature control frame. Another way of looking at it is that the part may move in a
datum reference frame. The set of datum reference frames may by viewed as a candidate
datum reference frame set from which one must choose to measure. If the controlled
feature complies with its geometric tolerance to at least one datum reference frame in the
set, the feature is viewed as in compliance with its geometric tolerance.
Because of these insights and clarifying statements (mostly originating in ASME Y14.5.1),
compliance with may be adjusted to an optimum position is much better defined and, as
such, much less ambiguous, and constructing a valid datum plane or axis is easier. One
such recommendation is that (in simplified terms) the primary datum feature be divided
into thirds, and any candidate datum plane may not have all of the high-point-contact
points residing in either of the outer thirds of the surface. This type of clarification on
constructing a valid datum plane makes for a more complete explanation.
Another problem in establishing repeatable datum planes simulated by a plane contacting
the high points of that surface has come into evidence with the widespread use of the
computerized coordinate measuring machine (CMM). In

 Page 181
the past, many operators of these machines have chosen not to fixture their parts, to
establish datums from high point contactfor example up against mutually perpendicular
flat rails or angle platesfrom which probed points can be taken. Since it is faster to probe
arbitrary points on the surface instead of taking the time to fixture the part and probe the
fixture, many CMM sales people push surface probing to establish datum planes or axes.
Even though good fixtures touch the datum surface high points and, if probed, the fixture
can simulate quite accurately high point mutually perpendicular datum planes and/or axes,
a person trying to make a sale does not always care about truth, justice, and the good of
mankind. If one salesperson recommends fixturing, it might cost a sale to another
salesperson who says fixturing on his or her particular machine is not necessary.
It is a self-perpetuating falsehood that if one expects to sell a CMM, one must tell, because
others willand there is a lot of competition. So many say, with a straight face, With a
CMM it is never necessary to fixture a part. They are wrong, of course, because without
fixturing you don't stand much of a chance of forming datum planes or axes from high
point contact. You will, instead, get, for example, a plane that is the average plane formed
between the arbitrary points you have probed. But it is faster to surface probe instead of
doing it correctly and CMM users want to believe it works because it is easier andspeed is
a great selling point. Datum planes (if formed in this way) are often not repeatable from
set-up to set-up; and, therefore, results in measurement data collected from a feature may
vary greatly.
Nonrepeatability is not the only problem when no fixturing is used and arbitrary points on
the feature surface are probed. Constructing datum planes in this way, when high point
contact is called for, is simply not in compliance with the Dimensioning and Tolerancing
Standard. And, if the primary datum surface is not sufficiently well formed and
secondary and tertiary surfaces are not well enough formed and oriented, parts may be
accepted that do not meet the geometric tolerances specified (had the feature been
inspected properly), or parts that may be technically acceptable may be rejected as being
out of tolerance. Parts that will not function or mate are sometimes accepted; then when
assembly is attempted, since measurements have not been taken from the datum feature
high point planes, the high points of the datum surfaces interfere with other datum
surface high points of mating parts.
On the other hand, some of the parts that have been rejected may function and assemble
just fine. Clearly, this is a problem wherein the speed of initial measurement must be
weighed against the cost of scrapping good parts, reviewing and rechecking parts that
have failed the initial improper inspection procedure, and/or accepting low quality,
nonfunctioning parts. Proper fixturing is recommended so that the fixtures, which contact
the datum surface high points,

 Page 182

FIGURE 7-3
Uncertainties in datum establishment: datum planes incorrectly
established.

 Page 183

FIGURE 7-4
(A) CMM probing angle plates to establish datum planes from surface
high points. (B) CMM using chucking device to establish datum axis by
probing inside of chuck.

 Page 184
can be probed instead of having to probe the actual datum surface at arbitrary points.
In the previous examples, illustrations were shown to demonstrate the possible errors
experienced when proper fixturing of a part is not established for measurement from
datums. These illustrations show two possibilities for the fixturing of datum features
(both planar and cylindrical surfaces). In these ways, the fixtures may be probed instead
of bringing the probe into direct surface contact with the datum features. The fixturing
devices shown in Figure 7-4 (angle plates, CMM table, and chuck) touch the datum
features at their highest points of contact, as the datum features referenced ask. Probing
of the fixtures rather than direct probing of the datum features stands a much better
chance of rendering reliable, repeatable measurement data that are in compliance with the
appropriate establishment of datums per the ASME Y14.5 Dimensioning and Tolerancing
Standard's specifications.
The main purpose of a design drawing is to convey functional requirements of the part(s)
to the manufacturing and quality personnel, so they can produce a high-quality, functional
product at the lowest possible cost. Even though the drawing isn't usually dictatorial in
what machining practice or tool is to be used to obtain a desired shape or finish, the
tolerancing engineer must realize that when he assigns datums he is loading the machine
for the manufacturereven telling the inspector how the part is to be inspected. Some
latitude on the manufacturer's and inspector's part is, of course, built into the system. For
example, it is up to the inspector whether to use receiver gages, surface plate set-ups,
CMMs, optical inspection, paper gaging, or a combination of these tools and other
methods. Still, even these decisions are influenced by the symbols and their placement
used by the tolerancing engineer. The designer must consider his/her own requirements
and then those of the manufacturing and inspection facility.
Processing, manufacturing, and inspection capability need to be taken into consideration
during the design stage. Otherwise, products may not be producible or inspectable or
simply be prohibitively expensive to produce or inspect. So, as stated earlier in this unit,
to combine those requirements, the tolerancing engineer has an obligation to make certain
the datum features assigned are:
functional (serve a purpose in the proper restraint of the part),
accessible (so the manufacturer or inspector can simulate datum points, lines, planes, or
axes from the datum features identified).
repeatable (for purposes of consistency and economy), and
representative of mating and seating features and alignment edges when appropriate (to
assure that when the part is manufactured and inspected to the given specifications, it will
assemble and seat correctly with mating parts in the rest of the assembly).

 Page 185
The designer must realize the manufacturing facility has to simulate with real tools those
datum features that generate perfect geometric points, axes, and planes that exist only in
space. To do this, one uses datum feature simulators (surface plates, chucks, angle plates,
parallels, mandrels, etc.). With this in mind, the question should be asked, Can it be
done? and if so, Could it be done easier and faster if I datumed it differently, while still
satisfying my design requirements? Concerns are often voiced about striking repeatable
planes in a production situation because of rough surfaces or datum features that rock.
Even drill-through or mill-through requirements force machinists to deviate from setting a
part on a datum feature as drawings often intend.
For these and many more reasons, datum target points, lines, and areas are sometimes
used. These tools of the designer enable one to untie the hands of those trying to simulate
the required datums. For example, a target area can be drawn to realistically represent
equipment used in a manufacturing environment.
Chain lines can show an area the same size and shape of the riser blocks or parallels so a
part can be set on them when being machined and inspected. This makes the datums
something the manufacturer doesn't have to strain to simulate. It allows the datum to be
repeatable and, at the same time, it often satisfies design requirements. Unless
functionality absolutely dictates it, the machinist should not have to make allowances for
unattainable or unrealistic datum features. The machinist's changes or allowances might
not be the same as those of the next machinist, or the next.
7.5 Simultaneous and Separate Requirements
When more than one pattern of features is located by basic dimensions relative to the
same datum features in the same order of precedence and those datum features are not
subject to size considerations, the patterns of features are to be considered one pattern of
features and checked with the same set-up or functional gage (if a functional gage is
used). This means that one set-up is required to check all features within this unified
pattern. If, for example, the primary datum feature rocks on the primary datum plane,
only one orientation of the surface may be used to verify all features (holes, shafts, etc.).
This tying together of the patterns during inspection is based on a simulation of the way
they function. For example, one mating part with many patterns of mounted pins may all
mate with holes on one given part. Therefore, the multiple patterns actually function as
one. This treatment of multiple patterns as one in an inspection procedure is more
restrictive than if the patterns were treated as separate requirements.
If treated as separate requirements rather than as simultaneous requirements, one pattern
would be allowed to rock on the datums in a differ-

 Page 186
ent direction than another pattern of features to be accepted. Conceivably, more parts
would pass inspection if this were allowed, but some accepted parts may not assemble
with unified mating pin patterns. The separate requirement treatment (discussed later in
this chapter) would be more appropriate if separate component parts were expected to
mate with each separate hole pattern. This rule also applies if the datum features are
subject to size tolerances if the same material condition symbols for those datum features
of size are used in the considered pattern's feature control frame. See Figures 7-5 and 7-6.
In multiple patterns located by basic dimensions relative to the same datums in the same
order of precedence in Figure 7-5, the datum features are not subject to size
considerations. They are planar surfaces. The datums will be planes constructed from the
high points of the surfaces (as dictated by the datum reference frame).
In multiple patterns located by basic dimensions and referenced to the same datums in the
same order of precedence, the datum features in Figure 7-6,

FIGURE 7-5
To be considered one pattern of features.

 Page 187

FIGURE 7-6
To be considered one pattern of features.
two of which are subject to size tolerances, are given the same size considerations
(material condition symbols) in all patterns.
If datum features of size are used with MMC symbols (rather than at regardless of feature
size), a controlled pattern shift sometimes occurs. When datum features of size depart
from MMC (if primary) or virtual condition (if secondary or tertiary), patterns of features
controlled to them may shift as a group away from the datum axis or centerplane.
If a simultaneous requirement is invoked because of the same use of datums in what
otherwise might be considered separate patterns of features, any shifting of the patterns
must occur simultaneously and in the same direction. This allows unified mating pin
patterns to simply be moved by shifting the mating datum feature (off the datum axis or
centerplane), aligning all pins and holes, and assembling the parts. If the patterns of
features were not unified but treated as separate requirements, the separate patterns would
be allowed to shift in opposite (different) directions. This would make it difficult to
assemble all holes with one unified pin pattern.

 Page 188
If, however, the patterns of holes mate with separate component pin patterns, a unifying
of the hole patterns during the inspection process might be too restrictive, causing
rejection of features that would function fine as separate patterns. If the more restrictive
simultaneous inspection requirement of inspecting the patterns of features as a unified
pattern is unnecessary, then using the datums in a different order or using a different
modifier (material condition symbol) after the datums for the separate patterns would
disconnect them. Also, a local note used beneath each control that says Separate
requirement would disconnect the patterns, allowing more parts to pass inspection. As
long as the patterns do not function as one unified hole pattern (for example: they do not
mate with one unified pin pattern), this option may be desirable.

FIGURE 7-7
Unifying multiple patterns.

 Page 189
Although a simultaneous requirement is implied for the situations described in this unit
for multiple patterns located by basic dimensions and referenced to the same datums in
the same order that use the same modifier after datum features of size, this requirement
may be clarified. It is permissible to include a local note beneath related patterns such as
SIM REQT, standing for simultaneous requirement. Unlike the SEP REQT note, SIM
REQT when used in situations already implied as a simultaneous requirement is
considered a clarifying redundancy.
In both Figure 7-5 and Figure 7-6, if you do not wish the patterns of features to be
considered (and checked) as one pattern, simply add a local note to designate a separate
requirement under each feature control frame. For example:

For a composite positional control, the statements in this unit apply only to the upper-
segment datum-to-pattern control. They do not apply to the feature- to-feature control. In
other words, if two patterns of features are controlled with composite positional
tolerancing, they can be considered one pattern of features only when checking the
datum-to-pattern requirement. When checking feature-to-feature requirements, each
pattern must be considered separately.
This rule may be changed through the use of a local note that states SIM REQT
(simultaneous requirement) listed after (to the right of) the lower segment of each
applicable composite control. These lower segments may also be changed to a
simultaneous requirement through the use of one composite control with multiple leader
lines each leading to patterns to be treated as one. As shown in Figure 7-7, it is sometimes
considered useful to clarify the total number of features within the final simultaneous
pattern through the use of a local note. In this method the hole sizes are listed on the
drawing separated from the composite positional control.
7.6 Datum Target Assignment
Datum target areas of any configuration may be used to establish a datum. Circular areas
of contact, as shown, are common. When flat areas of contact such as these are used, they
require an attempt at full flat area contact by the datum target area simulator.

 Page 190

FIGURE 7-8
Datum target assignment. (A) Point contact, (B) line contact. (C) The
datum target point is commonly contacted with a spherically tipped
pin or the pointed end of a cone. (D) The datum target line is established
by contacting the surface with the side of a cylindrical pin as shown.
(E) Area designations.

 Page 191

FIGURE 7-9
Datum reference frame for a part using datum target areas.

FIGURE 7-10
Datum targetsarea contact: flat contact required to simulate datum.
(This frame has protruding pins that correspond to the size and location
of the datum target areas defined on the drawing.)

 Page 192

FIGURE 7-11
Datum planes established by point contact. (For fixture simulation,
see Fig. 7-12.)

FIGURE 7-12
Datum reference frame target point
contact.

 Page 193

FIGURE 7-13
Line and point contact: primary datum plane a step datumpoint
contact, secondary datum plane established by line contact, tertiary
datum plane established by point contact.
The part is placed on the fixture in an unrestrained condition, unless restrained contact is
specifically spelled out in a drawing note. For target areas full area contact is attempted,
but irregularities in the part surface will relegate the fixture to contacting high points
within the target areas simultaneously with priority given to the order of the datum
structure. In other words, the primary datum is created/simulated first, the secondary
second and perpendicular to the primary, and the tertiary third, perpendicular to the
primary and secondary planes of contact.
Datum targeting is an excellent way to stabilize a part that is rough or warped without
having to machine the datum features. It is also a way to stabilize a part and construct a
datum reference frame for parts that have only free-form curved-type surfaces.

 Page 194

FIGURE 7-14
Area, line and point contact: primary datum planetarget area contact,
secondary datum plane established by line contact, tertiary datum plane
established by point contact.

 Page 195

FIGURE 7-15
Datum target areas.

 Page 196

7.7 Oddly Configured and Curved Surfaces as Datum Features

It is now legal to identify a compound curve or contoured surface as a datum feature.
Whereas before these surfaces were not mentioned as possible datum features in the
Y14.5 Standard, except with the use of datum targets, one may now identify the entire
surface as a datum feature. As an extension of the principles of datum target points, one
may speculate as to the limitless possibilities this concept opens to us. If, for example, a
primary datum feature can be used to construct a primary datum plane from a minimum
of three points of contact (high points if an entire planar surface is labeled as the datum
feature, or specific points if datum targets are used), one might ask: What is the maximum
number of points that can be used?
The answer is, of course, there is no maximum. If no datum targets were used and
(theoretically) a surface was produced perfectly flat, all points, being of equal height and
lying in the same plane, would be high points. If datum targets are used, especially on
very flexible parts, such as sheet metal panels, it is not unusual to use more than three
specific points to establish the datum plane. It is also quite common to use step datums,
with some of these datum target points given height dimensions from the other target
points to allow the panel to nest well. If the surface is a curve that has been
mathematically defined, and if using datum targets at varying stepped heights, we would
pick the number of points we felt necessary to stabilize the part and then specify each as a
datum target point. We would identify the location and height of each, perhaps with basic
dimensions (toleranced dimensions are also allowed).
Suppose, instead, we decided we needed more than the three-point minimum to properly
nest the part, more than five, and more than 100 or even 1000 points. At some moment
we may reach the decision that it would be functionally appropriate, as well as less
cluttered on the drawing, to specify an infinite number of points on the surface to stabilize
the part, establish the primary datum plane, and to allow the part to nest in the same
manner in which it functions, even as it seats in assembly.
Upon reaching this decision, it seems logical to simply call the entire surface the datum
feature, using all the basic mathematically defined points on the curved or otherwise
complex surface to define the primary datum feature. We would create a simulation of a
perfect representation of this basically defined surface on which to seat the actual (less
than perfect) produced surface. This simulation of a perfect counterpart for the surface
could be thought of as a fixture on which to nest the part. In other words, a fixture could
be made to the basic dimensions given and the contoured surface seated on the datum
feature simulator (fixture, in this case) to establish the datum (or datums).
In summary, in the Y14.5 Standard, it is now technically legal to identify a

 Page 197
compound curve or a contoured feature as a datum feature. This mathematically defined
surface must be able to be related to a three-plane datum reference frame and, as stated, is
represented by the true geometric counterpart of the surface's shape to establish the
datum.
ApplicationsChapter Example
1. The ASME Y14.5M-1994 standard allows mathematically defined curved surfaces to be
used as datum features. Datum feature A on Figure 7-16 demonstrates this concept. It is a
thin plastic part about to be seated on the fixture for stabilization.
a. How many spacial degrees of freedom will datum feature A eliminate?
Answer: Five degrees of freedom: 3 rotational and 2 linear
b. How many datum planes does A alone create?
Answer: Two datum planes
c. What is the virtual condition of each hole that is positioned to A and B?
Answer: 9.87 and 11.87

FIGURE 7-16

 Page 198

7.8 Secondary and Tertiary Datum Features of Size

FIGURE 7-17

 Page 199
ApplicationsChapter Examples
Using Figure 718, answer the questions that follow.
1. What are the largest acceptable produced dimensions for the outside profile of this
part?
Answer: .255 2.050 3.175
2. What is the maximum axial separation allowed between datum features B and C?
Answer: 2.010
3. What is the minimum axial separation allowed between datum features B and C?
Answer: 1.990
4. What is the maximum allowed distance between the axis of datum feature B and the
edge of the part dimensioned .500 basic from it?
Answer: .500 + .025 (half profile tol.) + .005 (half .010 tol. shift from datum B made
at LMC) = .530
5. What is the minimum allowed distance between the axis of datum feature B and the
edge of the part dimensioned .500 basic from it?

FIGURE 7-18

 Page 200
Answer: .500 -.025 -.005 = .470
6. What is the maximum allowed distance between the left-hand edge of the part and the
surface of the .120 hole closest to it if all holes are produced at MMC?
Answer: .500 -.060 (half .120 MMC) + .025 (half profile tol.) =.465
7. What is the minimum distance allowed between the right-hand edge of the part and the
axis of the .130 hole closest to it if all holes are produced at LMC?
Answer: 3.125 -.500 -2.250 =.375
.375 -.025 (half profile tol.) -.005 (edge shift) -.005 (hole axis shift) -.005 (half hole
pos. tol.) =.335
8. What is the minimum wall thickness allowed on this part for the most endangered
wall?
Answer: .130 (LMC) + .010 (geo. tol. at LMC) =.140 sweep +.010 (shift from B if sept.
req.) =.150 (shift & sweep)
.075 = half sweep & shift
1.625 -1.375 -.025 (half profile tol.) -.005 (half shift from B of part edge) -.075 (half
hole sweep and shift) =.145 min. wall thickness

FIGURE 7-19

 Page 201
9. What is the maximum distance allowed between the right-hand edge of the part and the
axis of the .130 hole closest to it if all holes are produced at LMC?
Answer: 3.125 -.500 -2.250 = .375
.375 + .025 (half profile tol.) +.005 (half shift from B of part edge) + .005 (half pos.
tol. of .130 hole) + .005 (half shift of .130 hole axis from B) =.415
10. What is the maximum distance allowed between the left-hand edge of the part and the
axis of the .130 hole closest to it if all holes are produced at LMC?
Answer: .500 (basic dim.) + .025 (half profile tol.) + .005 (half shift from B of part
edge) + .005 (half Pos. Tol. of .130 hole) + .005 (half shift of .130 hole axis from
B) = .540
Using Figure 7-20, answer the questions that follow.

FIGURE 7-20

 Page 202
11. What are the largest acceptable produced dimensions for the outside profile of this
part?
Answer: .255 2.050 3.175
12. What is the maximum axial separation allowed between datum features B and C?
Answer: 2.010
13. What is the minimum axial separation allowed between datum features B and C?
Answer: 1.990
14. What is the maximum allowed distance between the axis of datum feature B and the
edge of the part dimensioned .500 basic from it?
Answer: .500 + .025 (half pos. tol.) + .005 (half .010 bonus tol. from datum B made
at LMC) = .530
15. What is the minimum allowed distance between the axis of datum feature B and the
edge of the part dimensioned .500 basic from it?
Answer: .500 -.025 -.005 = .470
16. What is the maximum allowed distance between the left-hand edge of the part and the
surface of the .120 hole closest to it if all holes are produced at MMC?
Answer: .500 -.060 (half .120 MMC) + .025 (half pos. tol.) =.465
17. What is the minimum distance allowed between the right-hand edge of the part and
the axis of the .130 hole closest to it if all holes are produced at LMC?
Answer: 3.075 -.500 -2.250 =.325
.325 -.025 (half pos. tol.) -.005 (half bonus tol. for B) -.005 (half hole axis shift) -.005
(half hole pos. tol.) =.285
18. What is the minimum wall thickness allowed on this part for the most endangered
wall?
Answer: .130 (LMC) + .010 (Geo. tol. at LMC) =.140 sweep +.010 (shift from B) =.150
(shift and sweep)
.075 = half sweep and shift
1.950 -.375 -1.375 -.025 (half pos. tol.) -.005 (half bonus tol. for B) -.075 (half hole
sweep and shift) =.095 min. wall thickness
19. What is the maximum distance allowed between the right-hand edge of the part and
the axis of the .130 hole closest to it if all holes are produced at LMC?
Answer: 3.175 -.500 -2.250 = .425
.425 + .025 (half pos. tol.) + .005 (half bonus tol. of B) + .005 (half pos. tol. of .130
hole) + .005 (half shift of .130 hole axis from B) =.465
20. What is the maximum distance allowed between the left-hand edge of the part and the
axis of the .130 hole closest to it if all holes are produced at LMC?

 Page 203
Answer: .500 (basic dim.) + .025 (half pos. tol.) + .005 (half bonus tol. of B) + .005
(half pos. tol. of .130 hole) + .005 (half shift of .130 hole axis from B) = .540
7.9 Free-State Variation in Sheet Metal Parts
Some concern has been expressed by designers of sheet metal parts about free-state
variation. Thin-walled materials such as sheet metal, as well as some plastics and rubber
parts, are often prone to wide variations in form when released from the restraints used to
machine or form them. Holding these features to specified tolerances of geometric control
requires close examination of when and how the part function dictates the need for these
controls. For example, if a sheet metal part is to be used in a restrained state (bolted,
riveted, or otherwise clamped against several surfaces of the part), then perhaps the final
shape or position (form, orientation, runout, profile, or location) of features should be
judged to be in or out of tolerance while in this restrained state.
A note added to the drawing, such as This feature shall be checked while restrained
(bolted, clamped, etc.) to the specified datum(s) with the following torque , will allow the
controlled features to be judged based on a simulation of the actual conditions the part
will experience while in use. As such, any geometric tolerances assigned the features take
on a realistic value and the datum features are chosen and utilized based on proper
criteria:
function
representation of the mating situation
repeatability
accessibility
The top of the surface plate in Figure 7-21 simulates datum plane A from the high points
of both surfaces simultaneously while the part is in the free state. Distortion due to free-
state variation is not controlled. The part is not clamped onto the datum features to try to
compensate for the warpage experienced by the part during the inspection procedure.
The dial indicator traverses both surfaces controlled for coplanarity with the profile
characteristic. This is accomplished while the part is in the free state to determine whether
the surfaces comply with the 0.1 profile tolerance. The surfaces must be coplanar
(straight, flat, parallel, and in the same plane) within a tolerance zone that consists of two
parallel planes 0.1 apart. The tolerance zone must contain all elements of both surfaces
simultaneously.
The tolerance zone is parallel to datum plane A simulated by the surface plate. Note that
since no coplanarity control has been applied to the datum features that form datum plane
A from high-point contact, the datum surfaces

 Page 204

FIGURE 7-21
Free-state inspection of flexible parts.
may be angled to and depart from the surface plate that simulates the primary datum plane
A. If the circled F was not added to the control, the part would still be inspected in the
same manner, since all geometric controls are considered to apply in the free state, unless
a restrained state condition is specified on the drawing. If this free-state inspection
condition is deemed nonfunctional and/or inappropriate, then a restrained state inspection
condition may be used instead. This is accomplished by a local or general note. See
Figure 7-22.
The top of the surface plate simulates datum plane A from the high points of both
surfaces simultaneously while the part is in the restrained state. Distortion due to free-
state variation is controlled. The part is clamped onto the datum features to try to
compensate for the warpage that may otherwise be experienced by the part during the
inspection procedure.
The dial indicator traverses both surfaces controlled for coplanarity with the profile
characteristic. This is accomplished while the part is in the restrained state to determine
whether or not the surfaces comply with the 0.1 profile tolerance. The surfaces must be
coplanar (straight, flat, parallel, and in the same plane) within a tolerance zone that
consists of two parallel planes 0.1 apart. The tolerance zone must contain all elements of
both surfaces simultaneously while the part is restrained as prescribed in general note 1.
Any free-state variation of the datum features is negated by the restraint created by the
clamping devices. This restraint should be, whenever possible, reflective of how the part
is re-

 Page 205

FIGURE 7-22
Restrained state control of flexible parts.
strained during actual use (for example, assembly). The tolerance zone is parallel to
datum plane A simulated by the surface plate.
The part shown in Figure 7-23 may be fixtured with a device similar to the one shown in
Figure 7-24. This fixture may be used during inspection of the part to establish the
appropriate datum reference frame. The part will seat on the surface that represents datum
A, while the shafts that represent the virtual conditions of datum features B and C are
inserted into the holes that are datum features B and C. Only the width of datum feature C
will be used to establish the orientation of the measurements originating from the axis of
datum feature B. The axis of datum feature B is simulated by the center of the cylindrical
pin/shaft. The centerplane of the width of the rectangular shaft will serve as the angular
orientation datum only.
If the geometric controls used the datum features at regardless of feature size (RFS), the
fixturing pins would be required to be expanding pins. This would locate the part and the
holes on the part on the fixture without the shift allowed by the datum features controlled
at MMC. This shift is equal to the difference in the actual size of datum features B and C
and the fixture pin size (the virtual condition of the features).
The part shown in the top portion of Figure 7-25 may be fixtured with a

 Page 206

FIGURE 7-23
Fixturing sheet metal parts.

FIGURE 7-24

 Page 207

FIGURE 7-25
Fixturing sheet metal parts with step datum targets.

 Page 208
device similar to the one shown below it. This fixture may be used during inspection of
the part to establish the appropriate datum reference frame. The part will seat on the
surfaces that represent datum A. This is termed a step datum. One of the seating surfaces
(A3) for the establishment of datum A will be 16 mm taller than the other two (A1 and
A2). The shafts that represent datums B and C are inserted into the holes that are datum
features B and C. Only the width of datum feature C will be used to establish the
orientation of the measurements originating from the axis of datum B. The axis of datum
B is simulated by the center of the cylindrical pin/shaft.
If the geometric controls used the datum features at RFS, the fixturing pins would be
required to be expanding pins. This would locate the part and the holes on the part on the
fixture without the shift allowed by the datum features controlled at MMC. The radial
shift away from the datum axis allowed because of the datum features being invoked at
MMC is equal to one-half the difference in the actual size of datum features B and C and
the fixture pin size (the virtual condition of the datum features).
ApplicationsChapter Example
1. Using Figure 7-26, fill in the chart on the following page.

FIGURE 7-26

 Page 209
Allowable Allowable
Max. Min. Avg. AcceptReject
Max. Min.
[Examples]
1.010 .998 1.004 1.014 .994 X
1.009 .999 1.004 1.014 .994 X
1.0101.0051.0075 1.0175 .9975 X
1.002 .998 1.000 1.010 .990 X
1.0251.0061.0155 X
[Problems]
1.0211.0001.0105 1.020 1.0005 X
1.019 .991 1.005 1.015 .995 X
1.0081.0071.0075 1.0175 .9975 X
1.001 .982 .9915 1.0015 .990 X
1.011 .995 1.003 1.013 .993 X
1.030 .970 1.000 1.010 .990 X
1.010 .990 1.000 1.010 .990 X
1.020 .989 1.0045 1.0145 .9945 X
1.004 .997 1.0005 1.0105 .9905 X
1.0181.008 1.013 1.020 1.003 X


 Page 211

8
Material Conditions (Modifiers): Meaning and Uses
8.1 Maximum Material Condition
Maximum material condition (MMC) is the condition of a feature wherein the feature
contains the most material. It is often thought of as the heaviest feature. It is the smallest
hole or the largest shaft. In the ASME Y14.5, MMC is described as the condition in which
a feature of size contains the maximum amount of material within the stated limits of
sizefor example, minimum hole diameter, maximum shaft diameter.
A shaft or other external feature such as a pin, a tab, or a feature made up of opposing
external sides of a part are at MMC when they are at their largest allowable size. Some like
to refer to it as when the feature is the heaviest. For example, a shaft with a size of .125
.010 would be at MMC when it measures .135 or its largest size.
A hole or other internal feature such as a keyway, slot, or the feature made up of
opposing internal sides of a part is at MMC when it is at its smallest allowable size. If a
hole was drilled bigger than its smallest allowable size, more material would have to be
removed from the hole; therefore, it wouldn't have its maximum amount of material. If a
hole with size limits .480 .500 was produced at .480, it would be at its smallest
allowable sizeits MMC.
In feature control frames, the MMC concept (circled M) is usually used for mating
features. If a shaft is to be inserted into a hole, the shaft's geometry of form, orientation,
or location need not be as perfect if the shaft is made at a smaller sizejust as the hole's
geometry can be less perfect if the hole is produced at a larger size. This additional
geometric tolerance based on size departure from MMC is often termed bonus tolerance.
It allows more parts to pass

 Page 212
the inspection procedure based on functional requirements. It lets the inspectors buy off
parts that will mate.
Many of these parts may not have been allowed to pass the inspection process if they had
been controlled on a regardless of feature size basis, even though they may easily have
been able to mate with mating features. The MMC concept allows the acceptance of parts
that have mating features that may mate with pins more off to one side of the holes rather
than directly in the center of the holes. It allows the acceptance of pins that may lean more
inside a mating hole as the pin shrinks in size. If the need is simply that all parts that mate
are to be accepted, then the MMC principle may be employed without endangering
functionality.
If used after a geometric tolerance in the feature control frame, the effect on part balance
is not usually a major factor because it does not require a shift of the part off-center to be
assembled. However, the use of the MMC control, after a datum feature of size, allows the
acceptance of parts that may require the shifting of the part's weight off the center of the
datum axis. This must be used with caution, as it may endanger the balance of a rotating
part.
The MMC symbol is usually used to preserve a mating boundary, a constant boundary of
virtual conditionthe worst mating condition. For example:

Formula: The maximum inscribed cylinder size minus the allowed geometric tolerance at
that size equals the worst mating boundaryvirtual condition.

.250 - .014 = .236 .255 - .019 = .236

.251 - .015 = .236 .256 - .020 = .236
.252 - .016 = .236 .257 - .021 = .236
.253 - .017 = .236 .258 - .022 = .236
.254 - .018 = .236 .259 - .023 = .236
.260 - .024 = .236

As the hole moves or leans out of perfect position, it has the effect of closing the hole's
mating boundary at true position by the amount that it moves or leans. It may not measure
outside of its size limits of .250.260, but the effect of having less than perfect position or
orientation closes the space available at that location and angle for a feature expected to
mate with the hole at that location and angle. It is closed by the amount the hole moves or
angles away from perfect. The result is that the mating shaft must reside within the
boundary of .236 and the hole may not violate that boundary if the features are to mate
without interference.
If it is also necessary to calculate a minimum wall thickness between this hole and another
surface, it would be helpful to know how much larger an area

 Page 213
this hole is allowed to sweep (in a worst case situation) than the biggest hole. The
addition of the allowed geometric tolerance to the largest hole size allowed (LMC) would
calculate this sweep boundary. For example: .260 + .024 = .284. If one wished to
calculate a minimum wall thickness of this hole's surface to, say, an edge on the part from
which it was given a basic dimension, the .284 could be divided by two and the answer
subtracted from the basic dimension given from datum B or C to the center of the hole (to
begin the process).
General Rule of Thumb
Use the MMC symbol when mating is most important. Use it when you wish the inspector
to be able to accept all parts that can be assembled. It allows the acceptance of all parts
that assemble, whether they be assembled with the pin in the middle or near the side of
the mating hole. For example:

If used after the datum feature of size, the MMC symbol allows some parts to be accepted
that can only be assembled off center, imbalanced, with unequal gaps between mating
features. For example:

If a pattern of bolt holes is positioned to a datum feature of size specified at MMC, as the
datum feature of size departs from MMC (or virtual condition if the datum feature of size
is secondary instead of primary), the pattern of bolt holes may shift location (as a group)
away from the datum feature axis. If shifting occurs and assembly is attempted with the
mating datum features of size centered to one another, the bolt holes on one part will
probably not align with the bolt holes on the mating part. However, since there will be
clearance between the mating datum features (because of the datum feature's departure
from MMC), you may now shift the entire part off the datum center to align the bolt holes
on the mating parts. This will allow you to assemble the part, but the entire weight of the
part will be off to one side. If the part is a spinning part, it will be more difficult to
balance in this assembled condition. If the parts are panels to be assembled, such as
vehicle panels, gaps may occur on opposite sides that are not equal, making the panels
appear ill fitting (either a cosmetic or functional defect). It's

 Page 214
not that the fits will be looser than if they had been controlled at RFS, just that the fits
may not be as uniform.
Use the MMC symbol after the positional tolerance to protect a constant mating boundary.
This boundary is the collective effect of the actual feature size and its allowed geometric
tolerance. For example:

Hole Allowed positional Equals mating

Minus
size tolerance boundary
.250 - .010 = .240
.251 - .011 = .240
.255 - .015 = .240
.260 - .020 = .240

This .240 boundary is a constant boundary referred to as a virtual condition boundary.

Any mating pin to be assembled at this location, at this angle, must be compatibleIn other
words, with a worst mating condition of its own that is of a .240 or smaller (unless a
transition or an interference fit is desired).
8.2 The Maximum Material Condition Symbol and its Ramifications
Features that mate can often take advantage of the use of the MMC concept by using the
MMC symbol in the feature control frame next to the geometric tolerance and next to any
datum features of size referenced that also mate. This concept allows a growth and shift in
the geometric tolerance originally assigned without adverse effect on part functionality.
This is allowed because of the boundary of the collective effect of the worst size of a
feature for purposes of mating (MMC) and any applicable geometric tolerance at that size.
This boundary of worst mating condition, often referred to as virtual condition (MMC
concept), must not be violated if features are expected to consistently mate.
The allowed growth of tolerance zones by the use of the MMC symbology after the
geometric tolerance recognizes that this worst mating boundary, which exists as inviolate
between the mating features, can be approached by the actual

 Page 215
holes and shafts, and that as the holes grow from their smallest size and the shafts shrink
from their largest size, they may each move or lean or just in general be more out of
geometric perfection before reaching that boundary.
For example, in the case of a position control, this stationary boundary exists at true
position and does not change in size; therefore, as a hole is produced around that
boundary, the larger the hole is produced, the more it may move away from being
perfectly centered around that boundary before it actually comes into contact with the
boundary. This is important because should it violate the boundary, it would then occupy
space that is supposedly reserved for the mating shafta situation that could cause material
interference. If allowed (by the use of the MMC symbol), this growth in the tolerance
zone drawn from a departure from the features' MMC limit of size is usually one for one.
In other words, as the feature departs from MMC (as produced), the geometric tolerance
zone will grow by an amount equal to the departure.
This is not always true. As in the case of threaded features (and other similar self-
centering features) modified by the MMC symbol, since the point of engagement in the
mating situation is the pitch cylinder, the enlargement of a threaded hole's minor diameter
would not mean an equal growth (allowed movement) of the hole. We know that drilling
the minor diameter size a little bigger will not allow us to move the mating screw around
in the hole a like amount. There is a bonus of tolerance that is gained as the hole's pitch
cylinder grows, but calculating that growth is not normally done. However, it is often
taken advantage of in functional gaging procedures and then in assembly situations to be
able to assemble parts that might not otherwise assemble.
Just as the MMC symbol allows geometric tolerance zone growth when used next to the
geometric tolerance on what we refer to as the controlled feature, the MMC symbol when
used next to referenced datum features of size, can also allow zones to shift either
individually (if there is only one controlled feature) or as a pattern (where more than one
controlled feature is involved). This pattern shift is also a physical description of what is
allowable during assembly to get the parts to mate.
For example, if a pattern of holes is positioned to the centerplanes of a rectangular male
part (as in Fig. 8-1), the outside edges of that part generate those centerplanes. Now, if the
outside of the rectangular part should be made smaller than the cavity into which it fits,
you may be able to shift the mating male feature around inside of the larger female cavity.
This shift of one part inside of the other might be useful in assembling the parts.
For example, if the pattern of holes in the rectangular male part mates with a pattern of
pins in the female cavity, there is a possibility that by moving the parts from side to side
and misaligning the datum centerplanes of the two parts, they may assemble. It might be
impossible to assemble them in any other way. This occurs when the pattern of holes has
shifted off-center as a group, maybe due to movement

 Page 216

FIGURE 8-1
of the set-up during manufacturebut while maintaining proper location from hole to hole.
A multiple spindle drill press or a multiple punch press will consistently hold appropriate
hole-to-hole locations, but if the set-up shifts (such as in rail movement or loose clamps)
allowing parts to drift, the location to datums may change.
This concept of using the MMC symbol after referenced datum features of size takes
advantage of the departure of the datum features from MMC. It allows parts to be
accepted based on the fact that this size departure from MMC would actually allow the
parts to mate during assembly. However, if the parts must be assembled while centered,
such as in spinning assemblies rotating around the aligned central datum axes, use of the
MMC symbol after datum features of size could endanger balance. It would allow
inspectors to buy off (accept) all parts that will assemble, whether they assemble centered
or off-center. In these situations, the regardless of feature size (RFS) concept should be
considered for use after the datum feature of size. This will assure that no matter what
size the datum feature is produced at, no pattern shifting away from the datum axis will be
tolerated. Only each hole or shaft's individual tolerance of movement is allowed; no addi-

 Page 217
tional pattern shifting. The RFS concept used after the datum features of size will assure
that all parts passed by inspection can be assembled with their datum axes or centerplanes
aligned (on-center)providing better balance for spinning parts.
The MMC symbol is a magnificent tool for any manufacturer. The designer automatically
reduces the cost of any part when he assigns the symbol to a feature or features of size. In
terms of tolerance gained, this material condition symbol can be used in different ways.
When assigned to the feature of size being controlled (for example, a shaft given position
with this feature control frame),

we automatically know several things:

1. We know when the feature is made at its MMC size, its axis is allowed to be out of
position anywhere within a 0.25 cylindrical tolerance zone constructed around true
position.
2. We know this tolerance zone sets up a boundary that cannot be violated by any portion
of the actual feature. This boundary (which is discussed at length in the next chapter) is
the collective effect of size and geometric positional tolerance. In this case (an external
feature of size), that boundary is the MMC size of the feature plus the original 0.25
tolerance. This boundary is also constructed around the feature's true position (as defined
by basic dimensions from datum planes in this case, but in other cases by coaxiality to a
datum axis, or perhaps by position's brand of symmetry to a datum centerplane).
3. We know as the feature departs from its MMC size, extra (bonus) tolerance may be
drawn from the feature's size limits.
4. We know this extra tolerance is equal to the shaft's actual minimum circumscribed
cylinder's size deviation from the MMC size. Therefore, we can calculate the total
geometric tolerance for either holes or shafts from the following formulae:
Hole
(1) (2)
Actual size [max. inscribed cylinder] Bonus tolerance
+ Original geometric
- MMC
tolerance
Bonus Geometric Tolerance Total Geometric Tolerance
Shaft
(1) (2)
MMC Bonus tolerance
- Actual size (min. circumscribed + Original geometric
cylinder) tolerance
Bonus Geometric Tolerance Total Geometric Tolerance

 Page 218
The MMC symbol is allowed where it has been determined that function and fit between
mating features of size will not be endangered by the addition of a proportional extra
deviation from perfect form, orientation, or position as the features depart from MMC
(staying within their size limits). The examples given in this chapter dealt with a positional
control, but, as stated, the formulae for calculating total geometric tolerance are valid for
any appropriate application of the MMC symbol to the feature of size being controlled
(although with straightness controls at MMC, actual size is actual local size).
A different situation arises when the MMC symbol is applied to datum features of size in
feature control frames of other features being controlled to that datum. For example:

or

You can see in these two controls we are not worried about the extra tolerance to be
gained as the features being controlled depart from MMC because they are controlled
closely by the RFS concept. The features get no extra expansion in the size of their
geometric tolerance zones as the features are made larger or smaller (within size limits).
However, each has a datum feature of size called out at MMC that it is being controlled to.
This means that as the datum feature of size departs from its own MMC, an additional
shift in the tolerance zone of the feature or pattern of features being controlled to this
datum is allowed.
Sometimes, the amount of this additional shift is easy to figure. When the part and feature
geometry is very simple (perhaps, for example, one feature controlled to one datum
feature of sizelike a plain surface to a shaft, or a clearance hole to a clearance hole, or
even a coaxial situation of position with one shaft diameter controlled to another shaft
diameter), the additional shift of the feature's tolerance zone is easy to calculate because it
is usually a shift directly proportional to the datum feature's departure from its MMC size.
But, in some situations, this allowable shift is more difficult to visualize. The more
complicated part geometries (the more features and datum features of size involved) make
it somewhat more difficult for a floor inspector to determine the effect this allowable shift
has on part acceptance without the use of receiver (functional) gaging.
The use of receiver gages makes the calculation of bonus tolerance and allowed shift of
tolerances unnecessary for part acceptance or rejection. The gage automatically does these
calculations by either accepting or rejecting the part.

 Page 219
However, it must be noted that unless the functional receiver type gages are soft gages
(computer generated in software), the physical gage gives only attribute data (good vs.
bad) and does not give variable data (how good or bad and why). Also, coordinate
measuring machines are getting better at determining pattern shift because of software
improvements.
(Note: Paper gaging has been used by some inspection departments as a useful,
inexpensive tool to augment their inspection procedures. When used correctly, it can, in
many instances, be used in place of receiver gages. Paper gaging is used in conjunction
with open set-ups (variable data collectors, such as probes and indicators).
ApplicationsChapter Examples
1. Figure 8-2 shows examples of how to calculate bonus tolerances.

FIGURE 8-2
Examples of calculating bonus tolerances.

 Page 220

FIGURE 8-3
2. For a 2526-mm shaft, the completed table (Figure 8-3) shows the allowable total
tolerance using the following formula: Bonus + original tolerance = Total tolerance
8.3 Allowed vs. Actual Deviation from True Position
The allowed deviation from true position is equal to the size of the controlled feature's
tolerance zone. In the case of those features with cylindrical tolerance zones, the allowed
deviation is given on the basis of a diametrical tolerance zone constructed about the true
position axis. Allowed deviation takes into consideration the original tolerance given in
the feature control frame plus any additional (bonus) tolerance drawn from the feature
size.
This bonus tolerance is allowed for those holes or shafts using the MMC symbol. As they
depart from MMC toward LMC, extra tolerance for the allowed deviation from true
position is gained. In other words, the tolerance grows. The actual axis of the hole or
shaft as produced must lie within this cylindrical tolerance zone, which is the allowed
deviation from true position (diameter basis) that is acceptable.
The actual deviation from true position may be determined by an open set-up. Coordinate
measuring machines, optical and video comparators, height gages, gage pins, indicators,
and other inspection equipment may be used to locate the actual hole or shaft axis. Once
this is found, other calculations are re-

 Page 221

FIGURE 8-4
quired. First, measuring in a straight line from the location datums, it must be determined
how far the feature axis has been produced from its true position. This deviation must be
found first measuring along the X-axis, then along the Y-axis from the datum planes (see
Figure 8-4).
For example, if the feature's true position is drawn as in Figure 8-5, we know exactly
where true position is.
After the part is produced, we must determine how far the actual hole axis is from datum
C. Let's say we've checked and found the maximum deviation point of the axis is 50.12
from datum C. This is a deviation of 0.12 from true position along the X-axis. Checking
along the Y-axis from datum B, we find a maximum deviation point of 63.4, or 0.1 from
true position. If we use those figures (0.12 and 0.1) in the hole's actual deviation from
true position radially,

FIGURE 8-5

 Page 222
But since we are really interested in how large a diameter would have to be drawn around
true position to encompass the actual hole axis, we must multiply this answer by two. The
following formula is more appropriate for our needs.

Once computed, this actual diametrical deviation from true position must be compared
against the allowed deviation. If the allowed deviation (or actual tolerance zone size) is
larger than the actual diametrical deviation, the feature position is acceptable. If not, the
feature must be reworked or rejected. Reworking is sometimes possible. For example, if a
hole is not already made at LMC, it may be opened up (drilled larger). If the hole is
modified with a MMC symbol, this procedure will enlarge the allowed deviation from
true position (tolerance zone size).
The control given in Figure 8-6 can, if examined closely, explain not only how to
calculate allowed additional positional tolerance but also why it is allowed. In the
inspection of a part, we are often trying to simply discern whether

FIGURE 8-6

 Page 223
or not the produced part or parts will function. If so, we accept them. If not, rejection or
rework of the part is recommended.
To make a good judgment as to the capability of a part to function, one would like to
know how the part is to be used. Sometimes, the inspector has only the design drawing
from which to work. It is unfortunate when one is not given an intimate knowledge of
features to be inspected. But, the reality is that this is often the case. The inspector must
read the drawing as though it tells a story about the needed characteristics for
conformance in order to function. If the drawing is well done, the datums and feature
control frames can, indeed, give the inspector the information necessary to make good
judgments about the part and also to make valid recommendations to manufacturing
regarding adjustments to part processing to improve these desired characteristics.
If we read the aforementioned drawing and focus on the symbols/components of the
positional control, it can tell the story of what is expected of the controlled hole. It says in
part, Position of a diameter that mates while the part is seated on surface A and the hole is
located from B and C. Of course, it could be read in the more traditional manner: The
axis may be out of position a diameter of 0.4 if the hole is produced at maximum material
condition holding a relationship of perpendicularity to primary datum plane A, distance
from secondary datum plane B, and distance from tertiary datum plane C.
But to decode the function of the hole, the first reading and its subsequent logical
implications are probably more helpful. If we derive from the positional control that we
are positioning a hole that mates while A seats with a three-point high point contact and
we locate/measure the hole from planes B and C, we must then ask a series of questions.
If the round hole mates, what does it mate with? Without a quantum leap in logic, we may
speculate that a round hole mates with a round shaft. We might then ask, What is the
largest shaft that, in theory, could enter this hole at the desired angle and location from the
datums if the hole was made within its limits of size and also out of perpendicularity and
location to the listed datums the maximum amount allowed at that size?
The answer is to be found by simply subtracting the allowed geometric tolerance from the
produced hole size in each instance. If this is done, one finds that a constant boundary of
virtual condition (MMC concept) has been protected on each hole produced. This
boundary is perfectly cylindrical, perfectly oriented to datum plane A, and perfectly
located from datum planes B and C. The job of the hole is to stay outside of this
boundary. The mating pin/shaft is designed, dimensioned, and toleranced (for size and
position) to reside within this boundary. If this is discovered to be the case in parts that
have been produced, the inspector accepts the parts with a fair amount of confidence that
he has proven that the inspected features will work/assemble. To determine the boundary
size on the parts

 Page 224

 Page 225

 Page 226
under discussion, one may refer to the columns labeled actual hole size (maximum
inscribed cylinder) and column A, which is the allowed diameter of the deviation from
true position, and subtract these two numbers. For example:
Actual hole
A
size
- 0.40
(1) 11.75 11.35
=
- 0.45
(2) 11.80 11.35
=
- 0.55
(3) 11.90 11.35
=
- 0.65
(4) 12.00 11.35
=

The mating boundary remained constant even though the size of the produced hole and
geometric positional tolerance allowed changes. The inspector is saying, If the hole does
not violate this boundary, it will mate with the worst case mating shaft if it also does not
violate the boundary. If the hole resides outside the boundary and the shaft resides inside
the boundary (which has its center/axis at true position), no interference of material will
occur. Therefore, the parts will assemble and should be passed on by the inspector to
assembly.
ApplicationsChapter Examples
1. Is the part in Figure 8-7 within tolerance? (See figure for answer).
The part in Figure 8-8 has been produced. The center hole has been measured with a
vernier caliper. The recorded diameter is from .751 to .752. When viewed from above,
its shape is three-lobed.
Measurement Example 1
2. Is the method given in Figure 8-9 valid to accept or reject the hole for size violation? If
yes, why? If no, why not? Can you suggest a better method?
Answer: No. The envelope of perfect form at MMC has not been inspected for
violations. A .750 gage pin would be better to represent the MMC envelope.
Measurement Example 2
3. Is Measurement Example 2 (Figure 8-10) the proper way to establish a datum axis? If
yes, why? If not, why not?
 Answer: No. The axis of the maximum inscribed cylinder that is perfectly
perpendicular to datum A is not well simulated in the fashion shown.

 Page 227

FIGURE 8-7
4. A gage pin was inserted into the center hole of a different part produced for Figure 8-8.
The gage pin diameter was .753 (maximum inscribed cylinder). The following data was
collected: measurements from datum B at the axis of the gage pin closest to the top of the
hole (hole circle farthest from datum A) were 1.008 and similarly at the bottom of the
hole (hole circle nearest to datum A) from datum B were 1.019.

 Page 228

FIGURE 8-8

FIGURE 8-9

 Page 229

FIGURE 8-10
a. Do these measurements (given in question 4 on page 227) indicate that any of the
size requirements of the hole have been violated?
b. Does this information indicate the perpendicularity tolerance has been violated?
c. Has the positional tolerance been violated if measurements from datum C reveal that
the axis at the top of the hole is at 1.002 and at the bottom of the hole at 1.000?
Answer:
a. No
b. No

.003
.753 = Actual Bonus
=
+.010
-.750 = MMC Original Geo. Tol.
=
.003 = Bonus .013Total Geo. Tol. allowed
Geo. Tol. =for perpendicularity

1.019
-1.008
Actual amount of
.011=out of
perpendicularity

c. Yes. Reject.

 Page 230

8.4 Least Material Condition (Meaning and Uses)

Least material condition (LMC) is the condition of a feature wherein the feature contains
the least material. It is often thought of as the lightest feature. It is the largest hole or the
smallest shaft. LMC is in many ways the opposite of MMC. ASME Y14.5 lists it as the
condition in which a feature of size contains the least amount of material within the stated
limits of sizefor example, maximum hole diameter, minimum shaft diameter.
So, instead of a shaft or tab or other external feature being at its largest size as with MMC,
LMC is when that external feature is at its smallest size (having the least material) within
its size limits. For example, a shaft with size limits of .500.510 is at its LMC when it
measures .500. An internal feature, however, like a hole or keyway, is at its least
material condition when it is at its largest size. For example, a hole with size limits of
.750.760 will be at its LMC when produced at .760.
In feature control frames, the LMC concept (circled L) is usually used for features when
the preservation of material is of great importance and cost is a significant factor. It is
used when wall thickness is thought to be endangered and the holes stand a chance of
approaching a breakout condition. It is also used on casting drawings to assure that in
subsequent machining operations enough material is available to allow a machine cut to
clean up the part surface. It is capable of alleviating the machinist's frustration of starting
out machining material and ending up machining nothing but air. Since wall thickness,
instead of mating, is of major concern, if a hole is made smaller and a shaft made bigger,
a greater positional tolerance can be allowed. This additional geometric tolerance, based
on size departure from LMC, is often termed bonus tolerance. It allows more parts to pass
the inspection procedure based on functional requirements than if the feature had been
controlled on a regardless of feature size basis.
It could, for example, for a positional control, let the inspector buy off parts that have
enough preserved material at that location to assure breakout will not occur. Many of
these parts might not have been allowed to pass the inspection process if they had been
controlled RFS, even though they might easily have had thick enough walls and enough
preserved material either to not endanger wall thickness or to clean up in a machining
process. The LMC concept allows the acceptance of holes and shafts that are more out of
position than the original geometric tolerance would have allowed, and it does this on the
basis that the shafts are bigger and the holes are smaller than the LMC. They have more
material and therefore have to be less well located. If the need is simply that all features
have no walls that are too thin or have enough material to clean up during a machining
operation, the LMC principle may be employed without adversely affecting functionality.
The LMC symbol is usually used to preserve a sweep boundarya con-

 Page 231
stant boundary of virtual conditionbigger than the biggest hole and smaller than the
smallest shaft.
For example:

Formula: The actual produced size plus the allowed geometric tolerance at that size
equals the sweep boundary/virtual condition.

.260 + .014 = .274 .255 + .019 = .274

.259 + .015 = .274 .254 + .020 = .274
.258 + .016 = .274 .253 + .021 = .274
.257 + .017 = .274 .252 + .022 = .274
.256 + .018 = .274 .251 + .023 = .274
.250 + .024 = .274

As the hole moves or leans out of perfect position, it has the effect of widening the
cylindrical boundary the hole sweeps over at true position by the amount that it moves or
leans. It may not measure outside of its size limits of .250.260, but the effect of having
less than perfect position or orientation widens the area the hole covers centered about
that true position location and, therefore, thins wall thickness. It is widened by the
amount the hole moves or angles away from perfect. The result is that the hole must not
violate that boundary of .274 if material integrity and part strength are to be maintained.
If a mating boundary must also be calculated, as well as a sweep boundary when the LMC
symbol is used, it is easily done. The geometric tolerance allowed to the smallest hole size
is subtracted from that MMC size to determine the worst mating boundary. For example:
.250 - .024 = .226 worst mating boundary for the geometric control given the hole in
the example used in this section.
A sweep boundary could be used to calculate a minimum wall thickness. If, for example,
a minimum wall thickness needed to be calculated to a plane datum such as datum B or C,
half the sweep boundary of .274 is a radius of .137. This .137 can be subtracted from
the basic dimension that emanates from the datums and locates the center of the hole's
true position (to begin the process).
8.5 Regardless of Feature Size
The regardless of feature size (RFS) concept is used and its symbol implied when the
geometric tolerance is to remain the same no matter what size hole or

 Page 232
shaft is produced within its limits of size. The geometric tolerance never changes with the
size of the feature it controls. The size of the feature must be met and the geometric
tolerance must be met as separately verifiable requirements.
The RFS concept is often used where balance is importantfor example, on parts that
rotate. Spinning parts could be functionally endangered if tolerance of location, such as
centering, were allowed to vary with the size of the feature. As with the MMC symbol (to
say, for example, as a hole grows or a shaft shrinks that it can be further off-center on a
spinning part), this could adversely affect a part's balance. Therefore, if true, the RFS
symbol should be implied (as is the case under current national and international rules)
instead of specifying the MMC symbol.
If function must be preserved in such a manner as to have mating features assemble in a
specific way, such as a pin more in the center of a mating hole than the MMC symbol
would maintain (since the MMC symbol allows the hole to move out of its perfect
position more as the hole grows and the shaft to move out of its perfect position more as
the shaft shrinks), then the RFS concept should be used. Where the MMC symbol (if used
after datum features of size) allows a shift of feature patterns making it necessary to shift
a part's weight off-center of the datum axis to assemble the mating parts, the RFS concept
(if used after datum features of size) allows no pattern shift as the datum feature departs
from its MMC. This means that the only parts the inspector can accept are parts that can,
at least, be assembled on center. Although it does not assure that the parts will be
assembled on-center in a balanced condition, it does assure that the parts accepted for
assembly can be assembled on-center or pushed back to center. This makes any balancing
procedures that are subsequently attempted easier. It is also worth noting that the use of
the RFS concept protects the same worst mating boundary as does the MMC concept and
the same worst sweep boundary for wall thicknesses and material preservation as the
MMC concept.
The RFS concept would protect the minimum wall thicknesses as well as the LMC
concept if two other rules did not interfere. One rule is that with the RFS concept no
envelope of perfect form is required at LMC, but one is required at MMC. The second
rule is that positional controls applied at RFS are inspected by verifying the location of
the axes or centerplanes of their mating size. For example, a gage pin of the maximum
inscribed cylinder generates an axis that must reside within the tolerance zone for a hole
that is positioned. As is shown in Figure 8-11, the result is a thinner wall if the hole is
produced large but curved.
The following calculation is based on the dimensions shown in Figure 8-11. and the
geometric tolerance below.

 Page 233

FIGURE 8-11
MMC or RFS worst case analysis.

 Page 234
However, for the calculation in Figure 8-11, if we assume the feature uses the RFS
method of tolerancing as follows:

.250 + .010 + .010 + .014 = .284

worst outer boundary
or

.125=half the MMC envelope

size tol. all on one size on
+.010=
bottom as depicted
.135
+.007=half the positional tolerance
.142=Half worst outer boundary

For example: .250.260 hole

.236 worst mating boundary

.250 -
=(same as if the MMC symbol
.014
was used)
.284 worst sweep boundary
.260 +
=(same as if the MMC symbol
.024
was used)

So, while it is important to note that the RFS concept preserves balance better than either
the MMC or LMC symbol concepts, protects mating as well as the MMC symbol concept,
and protects wall thicknesses and material preservation as well as the MMC symbol
concept (but not as well as the LMC concept), it does create tighter, more restrictive
tolerances on the average produced part. This runs the risk of increasing the cost of the
overall product and should be used only when it is determined the use of the MMC or
LMC concepts may endanger part function.

 Page 235

9
Virtual and Resultant Condition Boundaries
9.1 Virtual Conditiona Functional Boundary
Virtual condition has always been seen as a way of describing a boundary that is
important to the functionality of a feature of size. For example, with the use of a
maximum material condition (MMC) symbol in a feature control frame, it was considered
that the boundary generated by the MMC of the feature and the allowed geometric
imperfection at that size was important to be able to calculate for the purpose of mating. A
hole that is controlled with the MMC concept (or symbol) is usually considered a feature
wherein any change in size is important to its function because it affects what size shaft
will mate with it at a certain location or orientation (or form when the MMC envelope
does not confine it). With these features, if the tolerance zone is not inspected, the
functional boundary generated by size and geometric control (virtual conditionMMC
concept) is usually inspected, often by what we call functional gaging. The functional
gage size is derived by calculating the virtual condition boundary size (MMC concept) and
then applying it as the gage size (to which gage tolerances are given).
With the advent of the least material condition (LMC) concept, the circled was
allowed to be included in the feature controls. This happened in the ANSI Y 14.5
Standard in the 1982 revision. This LMC concept was most often used for the
preservation of material for wall thicknesses. For example, if someone was creating a
casting drawing and was concerned that the casting have enough material to be cleaned
up during a machining operation to be conducted according to requirements called out on
a subsequent drawing (used by the machine shop), the LMC symbol was often utilized. It
had the advantage over the MMC symbol of stating that the more material a feature was
produced with the

 Page 236
more imperfect it could be, for example, in its location or orientation. The use of this
LMC symbol created a whole new functionality. The most important boundary under
those circumstances was material preservation, as might be the case with a hole in a
drinking strawthe smaller the hole in the straw, the more it may move out of perfect
coaxiality with the outside diameter without endangering the minimum wall thickness
required for part integrity.
The electronics industry terms these boundaries that are created by holes or shafts moving
as inner and outer locus. Working over a period of years, the Dimensioning and
Tolerancing Committee (Y14.5) determined their own names for these boundaries of
inner and outer locus. The most functional boundary generated is termed the virtual
condition, and if that happens to be, for example, the inner locus, then the outer locus is
termed the resultant condition. The measurement committees of B89 and the
Mathematization Committee of Y14.5.1 discussed the fact that if a boundary was to be
inspected, it would most appropriately be the one termed virtual condition (to protect part
functionality).
The formulas that could be applied to these concepts are many. Some are given in this
book and summarized below. Virtual condition boundaries are constant size. Resultant
condition boundaries are not constant in size, but for our worst case calculations we will
derive the worst case value.
MMC and LMC Formulas for Virtual Condition and Resultant Condition
MMC Concept (circled M in the control)
For holes, the virtual condition boundary is the inner locus.
Virtual condition = MMC - the specified geometric tolerance at MMC
For shafts, the virtual condition boundary is the outer locus.
Virtual condition = MMC + the specified geometric tolerance at MMC
For holes, the resultant condition boundary is the outer locus.
Resultant Condition = LMC + the specified geometric tolerance at MMC + the difference
between the MMC and LMC sizes
For shafts, the resultant condition boundary is the inner locus.
Resultant condition = LMC - the specified geometric tolerance at MMC - the difference
between the MMC and LMC sizes
LMC Concept (circled L in the control)
For holes, the virtual condition boundary is the outer locus.
Virtual condition = LMC + the specified geometric tolerance at LMC

 Page 237

FIGURE 9-1

FIGURE 9-2

 Page 238

FIGURE 9-3

FIGURE 9-4

 Page 239

FIGURE 9-5
For shafts, the virtual condition boundary is the inner locus.
Virtual condition = LMC - the specified geometric tolerance at LMC
For holes, the resultant condition boundary is the inner locus.
Resultant condition = MMC - the specified geometric tolerance at LMC - the difference
between the MMC and LMC sizes
For shafts, the resultant condition boundary is the outer locus.
Resultant condition = MMC + the specified geometric tolerance at LMC + the difference
between the LMC and the MMC sizes
9.2 Surface Vs. Axis Verification
It had long been believed that the verification of a virtual condition boundary generated
by a feature controlled for position tolerance at MMC was equivalent to verifying the
feature's axis location. It was believed that if the axis moved out of its perfect location,
since it took the surface with it, there was no differ-

 Page 240
ence in whether an inspector chose to inspect the surface location (say through the use of
a functional gage) for virtual condition boundary violations or the axis location through
verification of the cylindrical geometric tolerance zone using the location of the axis of
the hole's maximum inscribed cylinder (actual mating size). Figure 9-6, however, shows
that in some instances a feature could check within tolerance using one method and out of
tolerance using the other method.
As can be seen from Figure 9-6, the functional gage that checks the virtual condition
boundary would accept this hole as within its geometric toler-

FIGURE 9-6

 Page 241
ance, but the axis of the actual mating envelope would not reside within the allowed
positional tolerance zone of 0.1 at MMC. Depending on how the part is inspected, it
may be viewed as either a good or a bad part. In the case above, it is recommended that
the hole be accepted as within its geometric tolerance.
9.3 Zero Positional Tolerancing at MMC for a Greater Yield
Considerations Involved in Tolerancing Location (To Zero Tolerance or Not to Zero
Tolerance)
Zero positional tolerancing at MMC is often not appropriate when a designer wants a
press (tight) fit between mating features, but one of the features has no locational
datumshence, no positional tolerance of its own (for example, a free-floating pin, shaft,
or tab)and the location of its mating feature (hole or slot) is relatively unimportant in
relation to the location datums.
Situations where zero tolerancing at MMC or LMC is not appropriate are usually easy to
identify as they arise. For example, coaxial features that rotate in use should commonly
not use tolerancing at MMC (or MMC symbology in general) because of a loss of balance
as the part departs from MMC. Also, in the tolerancing of tapped holes and countersinks,
zero tolerancing is not appropriate because of centering effects. In those situations, it is
known that tolerance drawn from size limits as the feature departs from MMC cannot be
counted on.
However, if all the designer cares about is respective mating features going together at
assembly, then zero positional tolerancing (if used with proper dimensioning and datum
assignment techniques) may often be the best way to accept all parts that will mate at
assembly. It means not having to reject a feature based solely on nonconformance to a
size specification that is not the worst case mating situation.
For example, let us consider the following specification:

If, on this part, the main concern is the mating of this shaft (in its proper orientation and
location to the datums) with its mating hole (which has a location to datums
representative of those used to locate the shaft), we could, and quite often will, reject
parts that will mate at assembly. We know the worst mating condition

 Page 242
(virtual condition) for this shaft is MMC plus the geometric tolerance at MMC (.290 +
.040, or .330). We also know we have designed the hole that mates with this shaft to have
a worst case mating condition of .330. Remember, the hole will theoretically accept a shaft
with a worst case mating condition of .330 (or smaller).
Let us say we have produced three shafts listed, A, B, and C:
Actual diameter(feature Shaft axis out Worst mating
mating size without regard of position a condition as
to orientation or location) diameter of produced
Part A .320 .005 .325 diameter
Part B .310 .010 .320 diameter
Part C .291 .001 .292 diameter

After inspecting these parts, we know all three parts will mate with their properly oriented
hole, given the required datum orientation to the specified datum reference frame. Each
hole's worst case mating condition is .330 diameter.
The worst of these shafts has an as produced worst case for mating of .325 diameter. The
best is a .292 diameter shaft to fit into a .330 diameter hole. We know the features will fit
together (mate at assembly). The parts work!
But we cannot accept any of them because all exceed the specified MMC of .290 diameter.
They must all be rejected. The designer didn't mean for the MMC to stop the acceptance
of parts that would mate at assembly, but that is exactly what has happened. Consider the
alternative: if the design was altered (redesigned or originally designed as follows); let's
look at the ramifications.
Suppose, instead, the requirement was stated:

How many of the parts we produced (A, B, and C) would be acceptable? The answer, of
course, is that all three parts would be acceptable. The size limits have been met. The
virtual condition is still .330 diameter. The LMC is the same. The

 Page 243
material strength has not been endangered. We have allowed more material, not less,
when the position (orientation/location) of the feature improves. This can only strengthen
the feature/part. The MMC has been changed by the amount of tolerance in the feature
control frame (additive to external features such as shafts or subtracted from internal
features such as holes).
All tolerance of position comes from any as produced departure from the features' MMC.
For example, if the shaft, whose design was changed to zero positional tolerancing at
MMC, was produced at its previous MMC size of .290 diameter, its positional tolerance
would be .040 diameterjust as it was before the MMC was altered to .330 diameter. Holes,
when changed to zero tolerancing at MMC, experience the added advantage of a wider
range of drills we may use to drill the hole (because of the broadening of the size limits).
Zero tolerancing at MMC can result in the acceptance of more usable parts that will mate
at assembly.
As previously stated, it is not always an appropriate control for the function of the part.
However, if used properly and with good judgement that comes from experience and
simple logic, it can be a useful and very profitable design tool much of the time.

 Page 245

10
GagingPrinciples, Terms, and Examples
10.1 Gage General Principles
Attribute gages that determine whether size envelopes of perfect form at maximum
material condition (MMC) or geometric boundaries of virtual condition (MMC concept)
have or have not been violated are all designed on similar principles.
GO gages determine compliance with the MMC envelope. This envelope is required by
the American National Standards (ANSI) on measurement (B89 and B4) and
documentation (Y14). The American Society of Mechanical Engineers (ASME) ANSI-
approved Standard on Dimensioning and Tolerancing Y14.5 section 2.7 discusses this
principle.
Functional gages inspect for violations of the constant virtual condition boundary (MMC
concept), which is commonly thought of as a worst mating boundary. This boundary is
created by the collective effect of the worst feature size allowed for mating purposes
(MMC) and the allowed geometric imperfection at that size (geometric tolerance).
Functional gages are made at the virtual condition size of the features they inspect. These
functional gages, as with the GO gages that verify MMC, require tolerance that accepts no
bad parts. Therefore, the tolerances that are applied to gage pins are all on the side that
would add material to the pin. Tolerances given to gage holes are also on the side that
would add material to the gage hole (allowing it to get only smaller than the envelope or
boundary it attempts to verify has not been violated by the surface of the hole). In other
words, with gage pins the pin may be no smaller than the boundary it is verifying and
with gage holes the hole may be no larger than the boundary it is verifying. This means
that al-

 Page 246
though in theory gages do not reject good parts and do not accept bad ones, in reality
gages reject parts that are on the verge of violating their MMC envelopes or their MMC
concept virtual condition boundaries.
However, the goal of gaging remains that gages accept all good parts and reject all bad
parts. The tolerances assigned to gages divert us from the pure goal, but are necessary for
one to actually produce a gage. Therefore, the realistic goal of gages is to accept as many
good parts (parts that are produced within design specifications) as possible, while
(following the ANSI recommended practice of absolute gage tolerancing) never accepting
a bad one. This goal requires a gage design with size tolerances and geometric tolerances
as small as economically feasible. As is generally true with any part, the smaller the
allowed tolerances for the gage, the more expensive it is to manufacture. The small gage
tolerance may also cause the gage to wear out faster than one with a larger tolerance. As
soon as it wears beyond acceptable limits, it begins to accept technically bad parts. Gages
must be inspected periodically and replaced before this happens.
The cost of a less expensive gage with larger tolerances must also consider that a gage
given larger tolerances, although initially less expensive to produce, will tend to be made
at a size that will reject more technically acceptable border-line parts than the more tightly
toleranced alternative. It is therefore extremely important that the gage designer consider
the break-even point, deciding on the best balance between the gage with prohibitive up-
front costs to produce that tends to wear out faster and the prohibitive long-range costs
due to good (within specifications) parts being rejected.
Although in this chapter it will be recommended that the gage designer consider 5% of the
part tolerance as gage tolerance with an additional 5% considered as wear allowance, it
should be remembered that these are only guide-lines from which to begin the decision-
making process on what gage tolerance will be assigned. The function of the parts being
gaged must be taken into consideration, as well as the safety and economic ramifications
that might arise from assigning various tolerances to the gages and in a variety of
distributions (for example, between size and geometric requirements for orientation and
location). These are the same considerations that must be used in the tolerancing of
mating features in assemblies of parts.
Ideally, fixed-limit gages will accept all workpieces dimensionally conforming to
specification and reject all workpieces that do not conform. The GO gage and the
functional gage should fully receive the workpiece to be inspected. The NO-GO gage
should not receive the workpiece in any position.
GO plug gages should enter the hole over its full length when applied by hand without
using excessive force. If a full-form plug gage is not possible to use, or the rule
concerning the requirement of perfect form at MMC is not in effect, GO segmental gages,
if used, should be applied to the hole in axial planes

 Page 247
uniformly distributed around the circumference. Unless otherwise specified, perfect form
is required at MMC for rigid features, necessitating the usage or simulation of full-form
MMC sized cylindrical plug gages for holes and full-form MMC-sized cylindrical ring
gages for shafts. When nonrigid workpieces such as thin-walled parts are gaged,
considerable care is required as the application of too great a force may distort the hole
and give a false result. It should be remembered, however, that with nonrigid features,
perfect form at MMC is not required.
NO-GO gages are designed to inspect violations of the least material condition (LMC)
limit of size. The least material limit of the workpiece features is checked with this gage or
a simulation of this gage that performs cross-sectional verifications. This NO-GO gage
should not be able to pass into or over the work-piece at any position. If it is determined
that this type of measurement is to be performed, several gage designs are considered for
use. Some configurations can easily be simulated with common inspection tools such as
micrometers, telescoping gages, small hole gages, calipers, etc. Also, a spherical plug gage
may be used to best represent the requirement. This spherical plug gage should not enter
the hole when applied by hand without using excessive force. The hole should be
checked at both ends, if possible. A NO-GO gage with segmental spherical gaging
surfaces should be introduced into the hole by tilting it and it should not be possible to
erect the gage in the hole without using excessive force. The inspector is responsible for
all sets of opposing points inside the hole.
A GO cylindrical ring gage should pass over the complete length of a shaft when applied
by hand without using excessive force. If a cylindrical ring gage cannot be used because
the perfect form at MMC rule has been eliminated for a specific workpiece or feature on
the workpiece, and a GO snap gage or its simulation is to be used to check cross sections
for compliance, the GO snap gage should pass over the shaft, the axis of which is
horizontal, under its own weight or the force marked on the gage, or the gage should pass
over the shaft, the axis of which is vertical, when applied by hand without using
excessive force. Excessive force may be viewed as force that would distort or scar the
surface of the shaft. The goal is to inspect the shaft for compliance, not to force it or
distort it into compliance.
A NO-GO snap gage that inspects for cross-sectional violations of LMC should not pass
over a shaft whose axis is horizontal under its own weight or the force marked on the
gage and should not pass over the shaft whose axis is vertical when applied by hand
without using excessive force.
A functional gage pin should be able to enter the hole being gaged over the entire depth
of the hole without excessive force being applied. A functional gage hole (full-form ring
gage) should be able to receive the shaft being gaged over the entire length of the shaft
without excessive force being applied. If planar datum features are simulated on the gage,
the datum features on the workpiece

 Page 248
must contact the datum feature simulators on the gage, as appropriate (for example a
minimum of three points of high point contact on a primary planar datum feature, a
minimum of two points of high point contact on a secondary planar datum feature, and a
minimum of one point of high point contact on a tertiary planar datum feature). If
restraint is to be applied to the datum features, the need for restraint must be specified in
the workpiece drawing specifications or be restrained so as not to alter the measurement
readings of the same part measured in the free state using no restraint.
Gage Tolerancing
Just as with any manufactured feature or product, gages must be toleranced. There are
three major gaging policies practiced throughout the world. They are called many things,
but will be referred to in this text as: (a) absolute tolerancing (also known as pessimistic
tolerancing), (b) optimistic tolerancing, and (c) tolerant tolerancing.
Absolute tolerancing is the practice of tolerancing gages to ensure complete random
interchangeability of mating parts inspected by these gages. This is done by applying the
gagemaker's tolerances, the wear allowances, the measurement uncertainties, and the form
tolerances of the gage all within the workpiece limits of size and geometric control. These
gage tolerances add material to the gage, beginning at the limit (MMC or virtual condition
MMC concept as applicable) of the feature that is being gaged. This is the policy that is
recommended by ANSI. This method is practiced on the premise that gages will accept
most part features that are within tolerance, reject all part features that are not within their
tolerances and reject a small percentage of borderline part features that are technically
within their tolerances.
Optimistic tolerancing is the practice of tolerancing gages that ensures all part features that
are within their tolerances being gaged are accepted by the gage. This is accomplished by
applying the gagemaker's tolerances, the wear allowances, the measurement uncertainties,
and the gage's form tolerances all outside of the workpiece limits of size and geometric
control. Gage tolerances subtract material from the gage, beginning at the limit (MMC or
virtual condition MMC concept as applicable) of the feature that is being gaged. This
policy is not recommended by ANSI. This method is practiced on the premise that gages
will accept all part features that are within their tolerances, reject most part features that
are not within their tolerances, and accept a small percentage of borderline part features
that are technically in violation of their tolerances.
Tolerant tolerancing is the practice of tolerancing gages that ensures most part features
that are within their tolerances being gaged are accepted by the gage and most part
features not within their tolerances being gaged are rejected by the gage. This is
accomplished by applying the gagemaker's tolerances, the

 Page 249
wear allowances, the measurement uncertainties, and the gage form tolerances in such a
manner some of the tolerance on the gage is within the workpiece limits of size and
geometric control and that some of the tolerance on the gage is outside the workpiece
limits of size and geometric control. Gage tolerances both add and subtract material from
the gage, beginning at the limit (MMC or virtual condition MMC concept as applicable) of
the feature being gaged. This policy is not recommended by ANSI. This method is
practiced on the premise that gages will accept most part features that are within tolerance,
reject most part features that are not within their tolerances, accept a small percentage of
borderline technically out of tolerance features, or reject a small percentage of borderline
technically within tolerance features.
Gaging Policies (as Practiced throughout the World)
1. Accept all good parts, reject most bad parts (accept some borderline bad parts). Not
ANSI preferred.
2. Accept most good parts, reject most bad parts (reject some borderline good parts and
accept some borderline bad parts). Not ANSI preferred.
3. Accept most good parts, reject all bad parts (reject some borderline good parts). This
policy is preferred by ANSI.
GO and Functional Gaging Notes
Receiver gages = GO gages and functional gages.
GO gages check MMC only (perfect form at MMC envelopes).
Functional gages inspect geometric tolerances and virtual condition boundaries (MMC
concept).
Functional gages are made to the virtual condition (MMC) concept of the features they
inspect.
Recommended gage-tolerancing policy is that no bad parts will be accepted; therefore, a
small percentage of borderline technically good (in tolerance parts) must be rejected.
If so, gage tolerances add material to the gaging feature.
ANSI B4.4 (Inspection of Workpieces Standard) suggests 5% of the part tolerance as gage
tolerance with an optional 5% for wear allowance for a total of 510% of the part tolerance
to be used as gage tolerance.
Part tolerance for GO gages is equal to the difference between the inspected feature's
LMC and MMC.
Part tolerance for functional gages is equal to the difference between the virtual condition
(MMC concept) and the LMC of the inspected feature.
Virtual condition MMC concept for all holes (all internal features of size): MMC of the
hole minus the geometric tolerance applicable at MMC equals the virtual condition

 Page 250
Virtual condition MMC concept for all shafts (all external features of size): MMC of the
shaft plus the geometric tolerance applicable at MMC equals the virtual condition
10.2 Gaging Size Limits
As explained in the previous section, three methods are used throughout the world to
tolerance gages. One method subtracts material from the gage by applying the
gagemaker's tolerances, the wear allowances, the measurement uncertainties, and the form
controls all outside the workpiece limits of size. The second method distributes some of
these tolerances on the gage outside the workpiece limits of size and some within the
workpiece limits of size. The third, and most often used method within the United States,
employs a practice known as absolute tolerance. Absolute tolerance is a practice of
tolerancing workpieces that ensures complete random interchangeability of parts by
applying the gagemaker tolerances, the wear allowances, the measurement uncertainties,
and the form controls all within the workpiece limits of size. This adds material (through
gage tolerancing) to the fixed limit receiver type gage.
A fixed-limit gage is a device of specific geometric form and size used to assess the
conformance of a feature or features of a workpiece to a dimensional specification, such
as MMC. If gaging MMC, this type of gage is termed a GO gage. A gage of this fixed
variety that gages LMC is often referred to as a NO-GO gage or a NOT-GO gage.
The principle of GO and NO-GO gaging is often referred to as the Taylor principle. The
MMC (maximum material condition) of the feature undergoing inspection is checked with
a plug gage or ring gage of a length equal to the length of the feature or longer (or the
length of engagement of the workpiece to its mating part). In a diameter inspection, the
diameter of the gage is theoretically exactly equal to the MMC of the workpiece. This GO
gage should be able to fully pass into (for a hole) or (for a shaft) over the feature being
inspected.
The LMC limit of the feature under inspection is checked with a gage designed to contact
the workpiece at two diametrically opposite points separated by a distance exactly equal to
the LMC limit of the feature. This NO-GO gage should not be able to pass into or over
the workpiece at any location.
In the case of holes, the diameter of the largest perfect cylinder that can be inscribed
within the hole so that it just contacts the high points of the surface should be no smaller
than the MMC limit of size. At the same time, the maximum diameter at any location in
the hole (internal feature of size) must not be larger than the LMC limit of size at any two
diametrically opposite points.
In the case of shafts, the diameter of the smallest perfect cylinder that can be
circumscribed about the shaft so that it just contacts the high points of the surface should
be no larger than the MMC limit of size. At the same time, the mini-

 Page 251
mum diameter at any location on the shaft (external feature of size) must not be less than
the LMC limit of size at any two diametrically opposite points. This imposes a rule of
perfect form at MMC for rigid features, unless specified otherwise.
For a cylindrical feature produced uniformly at MMC, it must be perfectly round, straight,
and untaperedin other words, a perfect cylinder. Departures from MMC within limits of
size allow a less than perfectly formed feature by the amount of that departure to a
maximum of the full value of the diametral tolerance value specified.
If these errors of allowed geometric imperfection would prove too large to allow
satisfactory functioning of the assembled parts, for example, separate geometric
tolerances should be specified to refine the controls. Such refinements may be achieved
through the use of circularity, straightness, or cylindricity, or a combination of these
controls as appropriate to functional requirements. As always, judgement and practicality
must be employed in the gaging of any feature.
It may be determined, for example, that a full-form gage is not always necessary when
gaging MMC when it is known that the error of form is so small that it will not affect the
character of fit of the assembled workpieces. It is also possible that in gaging LMC a two-
point opposing point checking device is not always necessary when it is impractical to
design and manufacture such a suitable two-point measuring device. If the user is willing
to accept the possibility of a gage that is less restrictive than the control calls for, a full-
form cylindrical NO-GO gage may be chosen.
Workpieces may be inspected by these fixed-limit gages or limit-indicating gages (also
known as comparators). They are used to assess the conformance of a feature of a
workpiece to a specification, and incorporate a sensing element provided with limit
markers or signalsfor example, bore gages that have a fixed anvil and a sensor or other
standard measuring instruments. Factors to be considered when choosing a method of
inspection should include the following:
1. The purpose of inspection should be stated.
2. The number of workpieces to be checked should be great enough to justify the cost of
manufacture of the gage.
3. Plain limit gages may be designed to match the shape and size of the workpiece.
4. Some measurement instruments may give the workpiece size in the measuring position
but may not be capable of checking the geometry of the workpiece.
5. Direct measurement or limit-indicating gages may be preferable to plain limit gages
because:
a. They generally permit sampling inspection by variable data to warn when the sizes are
approaching one of the workpiece limits during a production run.
b. There is a need for one or more workpiece samples.

 Page 252
The principle of alignment should be followed as strictly as possible in all instruments for
measuring linear dimensions. The principle is that the axis or line of the dimension being
measured should be aligned with the scale or other dimensional reference of the
measuring instrument (when appropriate to the instrument being used).
It should also be remembered that all measuring operations given are understood as being
referred to a zero measuring force, and that if a measurement is carried out with a force
other than zero, its result should be corrected as appropriate. A correction is not required,
however, when, for comparative measurements carried out with the same comparison
methods, the same comparison force between similar items of identical material and
roughness of the surface exists.
ApplicationsChapter Examples
1. Unless otherwise specified, individual features of size must:
a. Not violate size limits of MMC at any cross section
b. not violate an envelope of perfect form at LMC
c. not violate an envelope of perfect form at MMC
Answer: c
2. The concept discussed in Question 1 deals with the:
a. Conrad principle
b. Taylor principle
c. Johnson principle
Answer: b
3. Least material condition on most parts is checked
a. At cross sections
b. For an envelope of perfect form at LMC
c. Never checked because they serve no purpose
Answer: a
4. When checking for MMC violations in a hole with size limits of .500.510, what size
and type of gage would you use?
a. A full-form plug gage with a diameter of .500
b. A full-form plug gage with a diameter of .510
c. A telescoping gage capable of checking size at any two opposing points.
Answer: a
5. What effect on size checks would adding the note Perfect form at MMC not required
have?
Answer: It would change the envelope/GO gage requirement at MMC to a cross-sectional
check/requirement.
Choosing the Right Tool
Using the list below, choose the correct tool to inspect the items required. Additionally,
answer the other questions.

 Page 253
a. Micrometer
b. GO gage
c. Peripheral measuring tape
d. Proper tool not listed (fill in the proper tool)
6. .250 .010 AVG. This is a hole in metal. It is the I.D. of a thinwalled tube.

What is What tool may inspect it? awith a

the MMC? telescoping gage (AVG eliminates
.240 envelope of perfect form.)
What is
What tool may inspect it? awith a
the LMC?
telescoping gage.
.260

7. .500 .005 This is a flexible rubber shaft.

What is
What tool may inspect it? a (The envelope
the
of perfect form at MMC does not apply to
MMC?
flexible or nonrigid features of size.)
.505
What is
the
What tool may inspect it? a
LMC?
.495

8. .625.635 This is a slot milled into a thick rigid plate.

What is the
What tool may inspect it? b
MMC? .625
What is the What tool may inspect it? awith a
LMC? .635 telescoping gage.

9. .010 .002 This is a steel tab that is nonrigid.

What is the MMC? What tool may inspect

.012 it? a
What is the LMC? What tool may inspect
.008 it? a

10. 1.000 .001

This control refers to a glass shaft outside diameter. A hole runs through its center making
it look like a drinking straw. It is mounted on a plate.

What
is the
What tool may inspect it? a
MMC?
1.001
What tool may inspect it? doptical
What comparator; part would have to be fixtured
is the to rotate; hard to do. (Use of the LMC
LMC? concept eliminates the envelope of perfect
.999 form at MMC and creates an envelope of
perfect form at LMC.)

11. 3.750 .020

 Page 254
This is a metal rod (shaft) that is 25 ft long.

What is
What tool may inspect it? a (Straightness
the
of a derived median line eliminates the
MMC?
envelope of perfect form at MMC.)
3.770
What is
the
What tool may inspect it? a
LMC?
3.730

12. 10.00010.100 This forms two sides of a rectangular part. It is considered an external
feature of size. It is rigid.

What is the MMC? What tool may inspect

10.100 it? b
What is the LMC? What tool may inspect
10.000 it? a

13. Answer the questions in Figure 10-1.

10.3 Terms
Absolute Tolerance (Pessimistic)A practice of tolerancing gages that ensures complete
random interchangeability of parts by applying the gagemaker's tolerances, the wear
allowances, the measurement uncertainties, and the form tolerances all within the
workpiece limits of size and geometric control. Gage tolerances add material to the gage.
Attribute DateThis is information that is obtained from an inspection process that
determines whether a part is acceptable or unacceptable. The format the data is related in
is good vs. bad. A feature is either acceptable for use or not acceptable for use. It either
meets its requirements or it does not meet its requirements.
CalibrationThe process of inspecting and then adjusting a gage or other measurement
tool to meet a specific parameter.
CertificationThe process of inspecting a gage to verify that it meets a specific parameter.
Fixed-Limit GageA device of defined geometric form and size used to assess the
compliance of a feature or pattern of features of a workpiece with a dimensional
specification and its accompanying tolerance. Also referred to as a limit gage.
FixturesDevices used to hold parts securely at the desired orientation and location for
purposes of manufacture, inspection or assembly.
Functional FixturesEssentially a gage with only the datum features simulated. These
simulators make physical contact with workpiece datum features to hold parts as they are
to be assembled. These simulated datum features are identified in product definitions of
the workpiece and represented on the fixture. Fixture datum features commonly represent
mating part datum features and therefore interfacing features in the product assembly.
Functional GageA fixed-limit gage used to verify a feature or a pattern of feature's
compliance with virtual condition boundaries (MMC concept) that are

 Page 255

FIGURE 10-1

 Page 256
generated by the collective effect of the feature's MMC and the applicable geometric
tolerance at that size.
Gagemaker's ToleranceA manufacturing tolerance allowed a gagemaker. It is applied to
fixed gages and comparator setting masters. Gagemaker's tolerances are generally
recommended to be about 5% of the part tolerance on the feature being inspected.
GO GageA fixed-limit gage that checks a workpiece at the MMC.
NO-GO GageA fixed-limit gage that checks a workpiece at cross sections for violations
of the least material condition. It is also referred to as a NOT-GO gage.
Part/WorkpieceA general term denoting a specific item, subassembly, or final assembly.
Separate RequirementA condition where all of the controlled features or patterns of
features that are inspected are not required to meet their tolerances in one set-up or on
one gage. They may meet their requirements in separate gaging orientations and be shifted
in different locations as their separate gages allow. A simultaneous gaging of all features
in one orientation and location would be more restrictive and potentially accept a smaller
yield of parts than the separate gaging requirement. Since a simultaneous requirement is
created by multiple patterns of features that use the same datums in the same order of
precedence and the same material condition symbols after datum features of size, if a
separate gaging requirement is desired, the abbreviation SEP REQT is placed under the
feature control frame.
Simultaneous RequirementA condition where all of the controlled features or patterns of
features that are inspected are required to meet their tolerances in one set-up or on one
gage to assure that they will function as one pattern. They must meet their requirements in
one gaging orientation and/or, if shifted as allowed by datum features of size specified at
MMC, must meet their requirements simultaneously in one shifted location.
Variable DataInformation obtained from an inspection procedure that indicates the level
of a feature's deviation from perfection or nominal. This level of deviation is given as a
numerical value.
Wear AllowanceAn additional amount of tolerance that may be considered to account for
the wear of the gage over time. The wear allowances and the gagemaker's tolerances are
normally equal in size (5% each of the part tolerance on the feature being inspected) and
combined they make up the total gage allowances applied to the MMC (GO or functional
gage) limits of size. The calculated gagemaker's and wear allowances are usually
combined for a total of 10% of the part tolerance and then assigned to the GO or
functional gage. The wear allowance will extend the useful life of the gage, and should be
used in applications where wear is critical.

 Page 257
Workpiece/Part ToleranceFor tolerancing a functional gage, it is the difference between
the virtual condition (MMC concept) and the LMC of the feature being gaged. For
tolerancing GO and NO-GO gages, it is the difference between the LMC and the MMC of
the feature being gaged.
10.4 A Few Major Measurement Concepts
Virtual condition (MMC concept) for all internal features of size is calculated by
subtracting the geometric tolerance applicable at MMC from the MMC of the feature.
Virtual condition (MMC concept) for all external features of size is calculated by adding
the geometric tolerance applicable at MMC to the MMC of the feature.
Measurement Uncertainty
Measurement uncertainty may be considered the difference between the corrected
measured size and the correct size. In cases where there is adequate information based on
a statistical distribution, the estimate may be associated with a specified probability. In
other cases, an alternative form of numerical expression of the degree of confidence to be
attached to the estimate may be given.
In the case of receiver gages (the family of functional gages and GO gages), one major
factor in measurement uncertainty is the tolerance on the gage, which is recommended to
be 510% of the tolerance on the feature being gaged.
Distribution
Gage tolerances may be divided between size and geometric controls. The distribution of
gage tolerances should be to optimize the manufacture of the gage and the acceptance of
all gages within the extremes of the range of total gage tolerance. This may, at times, call
for geometric tolerances assigned to the gage to be zero tolerance at MMC. However, if
some of the tolerance appears in the feature control frame as well as some in the size
limits, a RFS concept may be employed.
Gaging Application
There are many different applications for gages. Most of the discussion in this book deals
with gaging finished product requirements. There are also requirements for in-process
gaging procedures. It is often considered good gaging practice to have available two sets
of gages. One set will be used for in-process gaging. Normally, this set of in-process
gages, which are used by manufacturing personnel, will be provided with a larger wear
allowance than the final-inspection gages. This

 Page 258
is done because the in-process gages will receive more use and be used in an environment
more hostile to optimum gage handling and preservation. These gages tend to wear out
faster than gages used in an inspection-controlled environment.
If workpieces are rejected by the in-process gages, they can be set aside for a final
inspection procedure using the more tightly toleranced final inspection gage. Since this
gage is stored in a controlled environment more conducive to gage preservation and
appropriate usage, it is generally the more reliable of the gages and used as the final word
in the status of the workpiece.
In-Process Gaging vs. Final Acceptance Gaging
In-process gaging has several uses. One use is to spot-check the product of a controlled
process. An attribute gage will not show the actual quantitative value of the part; however,
it will show whether a part is outside of the acceptable limits. Another benefit is that the
process tooling can be used in place of building nearly duplicate tooling. If the process is
not reliable, gaging 100% of the product as final acceptance may be required. Gages will
not satisfy the quantitative data collection required to support, statistical process capability
studies; therefore, alternate inspection methods must be used.
Usage
Gages should be used in a manner that closely duplicates how the feature being gaged will
function. If it is to be assembled, the gage used should duplicate assembly conditions.
Environmental Conditions
Temperature
All part dimensions and tolerances apply at a temperature of 20C (68F). If both the gage
and the workpiece are at 20C, there is no measurement uncertainty caused by
temperature. For other conditions the effects of thermal expansion on the gage and the
workpiece must be considered.
Should the gage and the workpiece be at the same temperature that is other than 20C,
both will expand by an amount that can be calculated by the formula: KL (T - 20), where
K = coefficient of expansion, L = length, and T = temperature in degrees Centigrade.
With the same coefficients of expansion, no measurement uncertainty caused by
temperature is introduced. With different coefficients of expansion, temperature-related
measurement uncertainty is introduced. If the gage or the workpiece is constructed of
more than one component and these components have different coefficients of
expansion, the structure should be examined to see whether an additional uncertainty
could occur because of twist or bend.

 Page 259
Functional Gaging
Gages that check envelopes or boundaries are all designed on similar principles whether
inspecting MMC size violations or virtual condition. GO gages determine violations of the
MMC envelope that are required by Rule #1the Taylor principle. Functional gages inspect
for violations of the virtual condition boundary created by use of the MMC concept.
Functional gages are made to the virtual condition of the features they check. These
functional gages, as with the GO gages that verify MMC, require tolerance that accepts no
bad parts. Therefore, the tolerances that are applied to gage pins are all on the side that
would add material to the pin. Tolerances given to gage holes are also on the side that
would add material to the gage hole. In other words, with gage pins the pin may be no
smaller than the boundary it is verifying and with gage holes the hole may be no larger
than the boundary it is verifying. This means that although in theory gages reject no good
parts and accept no bad ones, in reality gages reject parts that are on the verge of violating
either their MMC envelopes or their MMC concept virtual condition boundaries.
It is important to remember the unattainable goalgages that accept all good parts and reject
all bad ones. The three methods of gage-tolerancing philosophy available and used
throughout the world are:
Gages that accept most good parts, reject all bad parts, and reject a small percentage of
borderline good parts.
Gages that accept all good parts, reject most bad parts, and accept a small percentage of
borderline bad parts.
Gages that accept most good parts and reject most bad parts but sometimes accept a small
percentage of borderline bad parts and sometimes reject a small percentage of borderline
good parts.
Functional Gage Design
Our gaging policy in the United States (often called a pessimistic gaging policy by the rest
of the world) states that, when designing gages, all of the tolerance goes against the part
being gaged. The amount of tolerance has often been suggested as 510% (5% of the
tolerance on the part plus an additional optional 5% for wear allowance per ANSI B4.4,
the Inspection of Workpieces Standard) of the part tolerance. In reality, we know that
percentage of the part tolerance is only an arbitrary figure. The actual amount of tolerance
chosen depends largely on the amount of parts that will work (which many times means
will mate at assembly with their mating parts) which we are willing to reject. The larger
the gage tolerance, the more usable parts the gages made could conceivably reject.
We are, in this chapter, going to direct our attention to two types of gages.

 Page 260
First, the GO-type gages used to check MMC envelopes and then the functional (receiver-
type) gages used to check geometric controls (for example, orientation or position).
The ultimate goal of any U.S. gage is, most often, to accept as many good parts (parts
produced within print/drawing specifications) as possible, while never accepting a bad
one. This goal requires a gage design with size tolerances and/or geometric tolerances as
small as economically feasible. As with any part, the smaller the allowed tolerances for
the gage, the more expensive it is to produce/manufacture. The small gage tolerance also
will cause the gage to tend to wear out faster than one with a larger tolerance. As soon as
it wears beyond acceptable limits, it begins to accept technically bad parts. Gages must be
checked periodically and replaced before this happens.
However, we must also consider the cost of a less expensive, larger-toleranced gage. The
gage given larger tolerances, although less expensive to produce, will (as previously
mentioned) tend to be made at a size that will reject more per-drawing parts than its
tightly toleranced alternative. So, the designer must give great consideration to the break-
even point and decide on the correct balance between the gage with prohibitive up-front
costs and prohibitive long-range costs due to good (per drawing) part rejection. A gage to
determine violations of the MMC envelope will, for gages measuring internal features
(holes, slots, etc.), be made to have a minimum size that is the MMC (smallest size) of the
feature it is checking. For example, the gage pin used to check a hole's MMC may be
slightly larger than the hole's MMC, but never smaller. If it were allowed to be smaller, it
would accept some bad parts (holes that have been produced smaller than the hole's
MMC). So, for those gage pins, the size and tolerance given the pin will be equal to the
hole's MMC plus something and minus nothing. For example, see Figure 10-2.
To engage the entire hole's length at once, the gage pin must be at least as long as the
maximum hole length.
As you can see in Figure 10-2, if the gage pin were produced (as it most probably will be)
slightly larger than .250, it would reject (not enter) a perfectly cylindrical .250 hole,
even though that hole theoretically should be accepted because it complies with the
engineering drawing. Therefore, careful consideration must be given to gage cost, part
(being gaged) cost, and gage wear (because the closer the gage pin is produced to the
hole's MMC, the more often it will have to be replaced due to being worn down in size
beyond the acceptable limit). By this same token, a gage being designed to check the
MMC envelope of an external feature (shaft, tab, etc.) may be slightly smaller than the
external feature's MMC but never larger. So, for the full-form ring GO gage (hole) used to
check the MMC envelope of a pin, the size and tolerance given the gage hole will be equal
to the pin's MMC plus nothing and minus something. For example, see Figure 10-3.
As is shown in Figure 10-3, caution is used to make certain the gage is no

 Page 261

FIGURE 10-2
(A) Hole to be checked; (B) gage to
check MMC envelope.

FIGURE 10-3
(A) Pin to be checked, (B) gage to check
MMC envelope.

 Page 262
larger than the pin's MMC. If it were, it would accept parts that would not comply with
the pin's design requirements (pins that exceed the MMC envelope). As should be evident
by now, these gages will reject a very small portion of the technically acceptable parts so
as to never accept a bad one.
Functional gages used to check geometric controls (MMC concept virtual condition
boundaries) follow essentially the same rules as those designed to check MMC
boundaries. However, there are extra design considerations that must be analyzed and
dealt with while designing functional gages. For example, see Figures 10-4 and 10-5 for
parts for which we will design gages.
Part 1 (Figure 10-4) has some interesting qualities simply because is has a straightness of
a derived median line control. Unlike surface straightness controls, derived median line or
derived median plane type straightness controls are additive to the size limits in that,
although the controlled feature may not exceed the stated MMC and LMC limits, MMC is
not checked for an envelope but at cross-sectional diametrically opposing points (perhaps
by a vernier caliper or a micrometer). So, even though cross-sectional opposing point size
is controlled by the size limits, form is not held in check by the MMC. Straightness of a
derived median line and straightness of a derived median plane, not being related to
datums, allow the form of the feature to bow outside of the normally established MMC
envelope to a larger MMC concept virtual condition envelope (boundary).
Part 1 is, therefore, first checked for size violations differently than similar features given
only size limits or similar features given size and surface straightness controls. Both MMC
and LMC are checked for cross-sectional opposing point size violations instead of the
usual LMC cross-sectional opposing point and MMC envelope size checks. Then, of
course, if the size limits are not violated, geometric control (derived median line
straightness within a diameter of .020 at MMC) must be checked. Obviously, this
geometric control stops the shaft from looking like a doughnut.

FIGURE 10-4
Part 1.

 Page 263

FIGURE 10-5
Part 2.
Because the MMC symbol is used in the feature control frame, a functional gage may be
designed (if cost-effective) to check the straightness control. Functional gages work well
for MMC controls and not so well for RFS controls. The reason that is true is because a
fixed-size functional gage will figure out bonus (additional) geometric tolerance for the
person using it (tolerance drawn from size by virtue of any departure from MMC of the
controlled fea-

 Page 264
ture). No, it doesn't add it up and write it down. But if the controlled feature fits the gage
(after size has been verified as within limits), you know that the feature has not exceeded
total allowed geometric tolerance (straightness for Part 1), which is the original geometric
tolerance (Part 1 has a .020 diameter original geometric tolerance) plus any bonus
geometric tolerance, because the functional gage has been made to the virtual condition of
the feature it is checking. This MMC concept virtual condition boundary that the
functional gage simulates is a constant boundary that represents the worst mating
condition of the controlled feature. The MMC symbol allows the feature to geometrically
change in form (as in the case of Part 1), orientation, or location to meet that virtual
condition boundary.
For example, Part 1 may bow more the smaller (within size limits) it is madeuntil it
bumps up against that virtual condition boundary of 1.040 diameter (1.020 MMC plus.
.020 geometric tolerance applicable at MMC equals the 1.040 virtual condition). However,
in a feature-controlled RFS, even though it has the same worst case mating boundary of
1.040, its as-produced size (within size limits) has no effect on increasing the allowed out-
of-straightness to more than the original .020 diameter. Therefore, the size of the
produced part has an effect on the boundary it may bow to. That boundary will not
always be 1.040. It will be the actual local size plus .020 diameter. Therefore, a functional
gage made at 1.040 would not be appropriate to check the straightness unless the shaft
was made at a consistent size of 1.020 (MMC). At any other produced-part size the
functional gage, if used, would be allowing bonus tolerance, whereas RFS controls
provide none.
An example of this will be given after we design the gage. Let's do that now. The
functional gage to check straightness for Part 1 would be a hole at least 3.020 deep with a
diameter of 1.040. It would, of course, have to have tolerance on the hole's diameter in
order to be made. For reasons mentioned earlier, all of that tolerance would have to go
against the part. The smaller the tolerance, the fewer usable parts (parts produced per
drawing specifications) the average gage would reject. Straightness of the gage hole is
protected under ASME Y14.5 Rule #1 (the size controls the form).
With this gage, if Part 1 were produced at 1.000 diameter, the MMC symbol would allow
an out-of-straightness of .040 diameter (.020 original plus .020 bonus). The gage hole is
in theory 1.040, so if the 1.000 part bows (is out-of-straightness) a diameter of .040 or
less, it theoretically fits into the gage and the gage tells us it is a good part (has not
exceeded allowed geometric tolerance).
If the part were controlled RFS instead of MMC and produced at 1.000, it would only be
allowed to bow the original .020 (no bonus .020) and would, in theory, only be able to
bow to an envelope (boundary) a diameter of 1.020. If we used the gage designed to
1.040 diameter, the shaft would fit inside, but the gage wouldn't tell us what we need to
know. The gage would tell us the shaft was

 Page 265

FIGURE 10-6
The gage.
straight to within a diameter of .040, but it would only be allowed to be out-of-
straightness .020so that information would be useless. Unless we were prepared to make a
functional gage hole that could adjust to produced size (whatever that was) plus .020, we
would not use functional gaging to check the geometric control. Granted, on a simple part
such as Part 1, a rough simulation of this RFS gage is conceivable (perhaps as simple as a
height gage and surface plate), but as the geometric controls become more complicated,
the feasibility of functional gage design for features controlled MMC becomes apparent
and the difficulties in producing even rough simulations of functional receiver-type gages
for attribute-only information (good parts vs. bad) for RFS controls reveal themselves as
most often economically prohibitive.
For example, let's examine Part 2 (Figure 10-5). This part actually requires two functional
gages to check the two separate geometric controls. (Size must also be considered; LMC
of .510 at cross-sectional opposing points and MMC of .520 for an envelope.) The
functional gage to check the perpendicularity requirement, because it is zero toleranced,
also will check the MMC envelope.
If functional gages are made to the virtual condition of the features they check and a zero
perpendicularity tolerance at MMC is given the shaft, we know the shaft's worst
perpendicularity mating case (virtual condition) is the same as its MMC (both are a .520
diameter). This means the part that fits the perpendicularity gage also complies with the
MMC envelope requirement. In this case, the virtual condition perpendicularity envelope
is an MMC envelope with a specific orientation (90 to datum plane A). It must also have
enough flat area to contact the high points of datum feature A to establish datum plane A
and be thick enough to cover the maximum height of the shaft.
The gage in Figure 10-7 has been simplified so as not to include the hole's locational
requirement (a necessary portion of the gage design). The reason for this is to focus
attention on some of the other things specific to checking perpen-

 Page 266
dicularity. The simulation of datum feature A must be large enough, and the hole in the
gage located closely enough, so that once the pin (shaft) goes into the hole, the gage
surface A (datum feature simulator) contacts the high points of the perpendicularity
datum feature of the part being checked.
The rest of this gage is pretty straightforward. The hole is made to the virtual condition of
the shaft it is gaging (.520) with all tolerance going against the part (all minus) so as to
never accept a bad part. The larger the gage's size tolerance, the greater the possibility the
gage will be produced to reject more actually acceptable (per print) features. So the object
of tolerancing the gage is to keep that size tolerance as small as is economically wise.

FIGURE 10-7
A simple and incomplete version of a perpendicularity
functional gage.

 Page 267
The gage hole also needs to be controlled for orientation (perpendicularity).
Consequently, we assign a feature control frame for that control (since none exists in size
controls). The more tolerance of out-of-perpendicularity we allow, the more chance we
have of producing a gage that will reject parts that are theoretically acceptable per the
engineering drawing. Consequently, we must hold that perpendicularity tolerance to a
minimum.
The material condition symbol chosen (either MMC or RFS) is another choice left to the
gage designer and is based on one of two criteria. If RFS is chosen, the designer has
essentially come to grips with the fact that he or she is going to play the percentages. The
designer knows the chances are better of getting a gage produced from the design that is
closer to perfect and, if so, will reject fewer parts that are within design limits (parts
produced per the engineering drawing). RFS says no matter what size the hole is
produced to, it can only be out-of-perpendicularity (lean) the original prescribed
tolerance. Still, even though the risk of producing a worst case gage (the hole has a worst
case condition of its own that reduces its size to even less than the virtual condition of the
shaft it is checking) has been reduced, a gage worst case exists.
If, on the other hand, the designer decides to use MMC, more of a chance exists that a
gage will be made and accepted that approaches that same gage worst case situation
(virtual condition MMC concept of the gage hole). Still, logic dictates that the designers
realize they aren't hurting themselves any more in a worst case situation than they would
have if they had used RFS. Let's face it; we are willing to accept a gage produced at the
gage's worst case condition, and the worst case mating condition, whether we use RFS or
MMC, is the same.
The only real difference with RFS is that as the hole departs from MMC (gets larger), the
gage hole [not being allowed to be any more out-of-perpendicularity because of that
MMC departure (growth)] accepts more produced parts that comply with the engineering
drawing. As the hole gets bigger, the gage gets better at accepting the borderline good
shafts it is gaging. It has toto be accepted per its own design requirements (RFS).
However, if MMC were used to define the gage hole perpendicularity requirement, as the
hole departed from MMC (got larger), it would be allowed more out-of-perpendicularity
to compensate for that growth and to be able to reapproach the gage worst case situation
(virtual condition MMC concept of the gage hole). This simply means that the chance of
accepting a gage that rejects more good parts (shafts that are within perpendicularity
requirements) increases if the MMC concept is used on the gage.
But let us not be too pessimistic. The percentage of per print (good) parts being rejected
by this gage using either the RFS or MMC principle is smallif small tolerances have been
used for gage hole size and orientation. The other gage needed for Part 2 (Figure 10-5) is,
of course, to check the positional requirement of the shaft.

 Page 268

FIGURE 10-8
Functional gage for the position control on Figure 10-5.

 Page 269
In addition to the geometric controls given to the gage, others are needed to properly
complete the gage design. (a) Datum feature A could use a flatness control. (b) Datum
features B and C need to be given perpendicularity controls to datum A and to one
another. The note is included to make certain enough surface area exists to establish a
suitable datum feature A simulator to contact the high points of datum feature A on the
part being checked by the gage. The MIN. dimensions will be measured from
perpendicular planes creating a rectangular area a minimum of 10.060 by 12.060. The
tolerances given to the gage hole for size and position work together to effect the number
of per print parts accepted, the same as previously explained for the perpendicularity gage
design. The larger the tolerances, the more the possibility of (good) per print parts being
rejected. Also, the material condition symbol used (RFS or MMC) relies on the designers
using the same criteria to make that decision as they did for the perpendicularity gage
design.
Another useful and, in fact, often wise option in completing a functional gage design for
checking orientation or location is zero tolerancing at MMC. Keeping in mind that the
functional gage has a worst case condition that is a combination of the MMC concept
virtual condition of the feature it is checking and any size tolerance and geometric
tolerance given the gage, we come to realize the worst case condition of the gage is
different than that of the feature it checks. In the case of a gage hole checking an external
(shaft) feature, the worst mating condition of the gage hole is smaller than the worst case
of the shaft it gages.
A gage shaft used to check an internal feature has a larger virtual condition (MMC
concept) than, for example, the hole it checks. If one decided to use a MMC modifier to
control the orientation or location of the gage feature(s) of size, then it stands to reason
that if we give all tolerance to the size limits of the gage feature and zero orientation or
location tolerance at MMC, we could maintain the same gage feature constant acceptance
boundary as we would have had we divided the tolerance between the gage size and the
geometric control.
Therefore, this gage will reject no more parts made per the engineering drawing
requirements than if we had taken some of the tolerance from the gage feature size limits
and transferred it to the gage feature control frame. As a matter of fact, the hole to check
an external feature could be made smaller and if, by chance, its location was made closer
to perfect than a larger acceptable gage hole, this smaller, better-located gage hole could
possibly accept more of the technically good per-engineering drawing male (external)
features, because, under those circumstances, the as-produced worst mating condition of
the gage hole could be larger and more closely match the virtual condition of the external
feature being checked.
The same goes for external gage features made to check internal fea-

 Page 270
tures of size. Zero tolerancing at MMC of the male gage feature would present the
opportunity to create and accept a functional gage that might otherwise be rejected
because it doesn't conform to size limits. But that rejected gage feature might more closely
approach the virtual condition of the internal feature being checked. Therefore, the gage
feature with expanded size limits and made larger (due to zero tolerancing at MMC) is
accepted, but can, at times, be produced better oriented and/or positioned to reflect an
overall mating condition that is smaller and, therefore, more closely resembles the virtual
condition of the features it checks. And, because of that, the gage that might otherwise
have been rejected can accept more parts produced per the engineering drawing.
There is one last thing to consider, for those who never design or use functional gages. It
is important to be able to look at a part design (whether yours or someone else's) and
visualize the functional gage it would take to check the geometric controls. This simple
exercise in spatial visualization will quickly help to determine the design intent and
coherency (whether or not the geometric controls make sense) and, in general, helps one
to thoroughly understand the drawing. It can speed the process of comprehending the
application of ANSI Y14.5 and allow the designer of parts to make certain the engineering
drawing says exactly what was intended.
Tolerance on the Work (Ten Percent of What?)
Ten percent of the tolerance on the work is often suggested as tolerance for gages
[approximately 5% gagemaker's tolerance and 5% wear allowance (optional) for a total of
10%; see ANSI B4.4M for more detailed guidelines and specifics]. Functional gages and
other inspection gage tolerancing should never be applied on such an arbitrary basis
without taking into consideration part function and the economics of a gage that might
reject more parts made per print than one is inclined to allow. Still, that mystical figure of
10% is often an acceptable one. But the question repeatedly asked is How does that 10%
come into play when geometric tolerance is also a factor? To delve into the answer to that
question, let us take a standard positional control and examine the components. This
should help us to realize what 10% of the tolerance on the work means.
To create a simple functional gage to check the location of the feature (hole) in Figure 10-
9, we must start with a datum reference frame simulation and then a pin (shaft) that is
representative of the worst mating condition of the hole being checked. It might appear as
shown in Figure 10-10 (some dimensions have been omitted to focus attention on the
topic at hand). We note that in the size and the feature control frame the tolerances are not
yet filled in. Let's examine some possible gage tolerances for both.

 Page 271

FIGURE 10-9

FIGURE 10-10

 Page 272
A straightforward approach might be to leave each tolerance in its respective position. For
example, the size limits of the hole are .250.255, or a total of .005. If we take 10% of that,
we arrive at a figure of .0005. We could fill in the gage pin size limits with that and end up
with:

As for the positional tolerance, if we take 10% of the .015, we would derive .0015; and
the feature control frame for the gage pin position could read:

So we have:

for the gage pin tolerance of size and location.

On the other hand (as in converting features to zero tolerancing at MMC), we could
realize that the tolerance available is .0005 plus .0015, or a total of .002. If we choose M as
the material condition symbol for the gage pin feature control frame, and also since we
know that size and geometric (positional) tolerance work together to create a worst case
situation (virtual condition) for the gage, then we might have the choice of dividing that
.002 in many ways or not at all. For example, suppose instead we gave the entire tolerance
to the size and used zero positional tolerancing at MMC. The gage pin size and feature
control frame might appear as:

If this seems logical to you, then the term tolerance on the work as it deals with this type
of gage refers to a combination of the hole's size tolerance and its positional tolerance.
Therefore, 10% of .020 equals a tolerance on the work of .002. The virtual condition of
both gage designs is .252 (.002, or 10% larger than the virtual condition of the hole being
gaged). An example of when

 Page 273

FIGURE 10-11
this type of thinking might become most useful would be when features to be gaged have
been themselves given zero positional tolerance at MMC.
If we are assigned the task of designing a functional gage to check parts produced per
figure 10-11, it might start out as in Figure 10-12. Of course, at 10% we realize that a total
of .001 gage tolerance is available. If we wanted to give a portion to the size and a portion
to the positional control, that would be an acceptable gage design as well.
There are other pertinent questions often asked. One question is Should we begin the size
of the gage pin at a slightly larger size than the hole's virtual condition and go up from
there?

For example, instead of shaft, we could use

This is acceptable if in the gage designer's opinion the chance of the gage pin being
produced at .375 and perhaps wearing quickly to something below the acceptable .375 is a
concern. You could even begin at .3752 or more (whatever you deem appropriate). Also
to be considered is whether one may wish to allow the finished gage pin to be no larger
than .376 (.001 larger than .375).
Another, more complicated question is whether the size tolerance should

 Page 274

FIGURE 10-12
even be considered when calculating the tolerance on the work when designing a
functional gage to check feature of size position. For example, in Figure 10-9, we could
have not included the .005 drawn from .250.255 limits of size in calculating the tolerance
on the work. In that case, had we not included the .005, we would only have had the .015
positional tolerance from which to calculate the 10% gage tolerance, or .0015.
Figure 10-13 includes both of those concepts.
Reason: 10% of .015 = .0015
.2352
-.2350
.0002fudge factor
+.0003size tolerance
+.0010positional tolerance
.0015Total

A little logic applied to the technique of not including the size tolerance in the tolerance
on the work puts the technique in question. When calculating tolerance on the work for a
GO gage to check MMC envelopes, we call the size tolerance the tolerance on the work.
For example, see Figure 10-14.

 Page 275

FIGURE 10-13

FIGURE 10-14

 Page 276
10% of .005 =
1.005
.0005
-1.000
.005

This .0005 is to be subtracted from the MMC (1.005.0005 = 1.0045) and 1.0045 is to be
used as a gage hole size limit (MMC of the gage hole, with 1.005 or smaller as the LMC).
We also know that we can often take positional tolerance from a feature control frame
and transfer it to the size limits. For example, in Figure 10-9 size and feature control
frame tolerances can be manipulated back and forth by the designer in many ways.
It now reads:

(virtual condition .235)

It could be changed to read:

(virtual condition .235)

What is the tolerance on the work now? Or

(virtual condition .235)

The combinations are endless, while keeping the worst case mating situation the same. As
you can see, the size tolerance and positional tolerance work closely together to create a
logical part tolerance. One can hardly be considered without its relationship to the other
(especially in mating situations). No matter what tolerancing method is utilized, caution
must be exercised to frequently check the gages making certain they have not worn down
beyond acceptable limits. Functional gages should not accept parts that exceed the virtual
condition boundaries of the features they check. Otherwise, the gages cease to be
functional.
Pushpin GagesAdvantages, Tolerance Distribution
There are some distinct advantages when checking hole (and other internal feature of
size) locations in choosing pushpin functional gages over fixed-pin gages.

 Page 277
With pushpin gages, the part is first oriented in the datum reference frame, making certain
proper contact is made between each simulated datum feature and datum feature. After
this location and orientation of the part is established and, if appropriate, it is clamped
into place, pins (external gage features of size) are either (a) pushed through gage holes
and then into controlled holes (or other internal features)) on the part, or (b) pushed
through controlled holes on the part and then into respective gage holes.
In either gage design, the result is, if the pushpins fit within both gage hole and the
internal feature being gaged, that the controlled feature is within the required positional
tolerance. Besides providing far better part orientation than fixed-pin gaging, pushpin
gages make it easier to determine which feature(s) in a pattern is out of tolerance and
sometimes even show in which direction it has strayed.
Caution must be used in the design of pushpin gages to make certain tolerances given to
the gage holes and the pins that are used in them provide for (a) a pin that easily enters
and can be easily extracted from its gage hole, yet with a minimum of airspace, and (b)
for holes controlled at MMC, the size of the gage pin is at least as large as the controlled
hole's virtual condition. Tolerance on the gage pin size is to be all plus and no minus.
Obviously, the gage hole receiving the pin must have tolerance as well. Just as obviously,
its size must be at least as large as the gage pin's MMC if we wish the gage pin to always
enter its gage hole. So, the gage hole's size tolerance is equal to the gage pin MMC plus
some tolerance and minus none. In addition, some positional tolerance is given to the
gage hole.
It is recommended that projected tolerance be used, since the gage hole gives orientation
to the gage pin. The amount of tolerance used has the effect of possibly virtually
increasing the size of the gage pin (MMC concept virtual condition), consequently
infringing further beyond the controlled hole's virtual condition boundary (MMC
concept). This has the effect of creating a gage pin virtual condition larger than the virtual
condition of the hole it checks. The more tolerance given the projected tolerance zone of
the gage hole, the more the chances increase to reject controlled part holes made
theoretically acceptable per the engineering drawing. Size tolerances given the gage pin
and the gage hole also increase this possibility. Consequently, these tolerances should be
kept to a minimum. Consult other chapters in this book for recommended criteria for
selecting functional gage tolerances.
Figure 10-15 shows one example of how to design, dimension, and tolerance a functional
gage that will inspect a positional control with a datum feature of size that has been called
out at MMC. The guidelines that have been discussed in this unit have been respected.
The results are as shown in Figure 10-15.

 Page 278

FIGURE 10-15
Datum feature at MMC.

 Page 279
Considerations of Functional Gage Design, Dimensioning, and Tolerancing
The decision-making process one goes through to design, dimension, and tolerance a
functional gage involves many factors. Among them is whether or not to conform to the
guidelines regarding the suggestion that 510% of the part tolerance be used as gage
tolerance. Again, part tolerance (of the part being inspected) is being defined for purposes
of a functional gage as the difference between a feature's LMC and its virtual condition
(MMC concept). Questions also arise about whether the percentage of part tolerance
should be 510% of each feature for its corresponding control.
For example, in Figure 10-15, the gage shown used 10% of the part tolerance used on the
flatness control for the gage flatness tolerance, 10% of the part tolerance used on the
size/perpendicularity control of the part for the counterpart gage feature's size and
perpendicularity tolerance, and 10% of the part tolerance of the part's hole size and
positional tolerance for the gage pins' size and positional tolerance; or, should the 10% be
a total of just 10% of the holes' (which are being positioned and, subsequently, inspected)
part tolerance, with this 10% divided up between all tolerances for the gage. For this
example, it was decided that the first, more liberal option (as shown in the gage) be given
as preferred. ANSI B4.4 suggests that the total tolerance on the page not exceed 50% of
the tolerance on the feature(s) being gaged.
Other options for these types of gages are as follows:
1. Tolerances may be divided between the size and the feature control frame as
appropriate, instead of the feature control frame being given a zero at MMC geometric
tolerance and having to draw all its geometric tolerance as bonus tolerance from the size
limits.
2. If some of the tolerance appears in the feature control frame as well as some in the size
limits, a RFS concept may be employed.
3. A pushpin gage design may often be more desirable than the fixed-pin concept shown
in Figure 10-15. The pushpin concept allows the part to first seat appropriately in its
datum reference frame before the gage pins are attempted to be inserted into the gage and
the part being gaged. With the fixed-pin design, it may, at times, be difficult to discern
whether the required datum high point contact is, in fact, being made. If this pushpin gage
design is employed, the part tolerance must be divided between the gage pin size limits
and its counterpart gage hole's positional tolerance. Thought must also be given to the fit
between the gage pin and its counterpart gage hole.
4. The application of the wear allowance, which is described as the additional 5% of the
part tolerance allowed that takes the gage tolerance from 5% to the optional 10% of the
part tolerance as gage tolerance, is always

 Page 280
worthy of consideration. As used in the example gage, the additional 5% for wear of the
gage was used the same way as the first 5% gage tolerance.
5. However, it is recognized that the additional 5% for gage wear may be used to change
the size of the gage pins to make them initially larger by 5%, and, in the case of gage
holes, the initial size of the gage holes may be made smaller by 5%. Then, one may
proceed with the tolerancing of the gage feature's size and geometric tolerance with only
the 5% left. This would potentially produce gages that would reject more of the
borderline technically good parts, but would assure that a gage element would not be
produced at exactly the virtual condition (MMC concept) of the feature being gaged and
run the risk of immediately wearing out beyond acceptable limits to allow the gage to
accept technically bad parts.
ApplicationsChapter Examples
Functional Gage Design
1. Describe the process required to verify the hole's size limits in Figure 10-16.
Answer:
Cross-sectional verification only is required for both MMC and LMC.
2. A functional gage pin capable of checking the hole's axial straightness requirement
(Figure 10-17).
3. What controls the gage pin's roundness, straightness, and taper?
Answer:
Its size tolerance per Rule #1.
4. Would this gage be sufficient to check the axial straightness requirement if the hole was
controlled with RFS instead of MMC? Why or why not?

FIGURE 10-16

 Page 281
Answer:
No. In an RFS condition, the boundary is not constant and would have to reflect the
actual local size minus .030.
5. Perpendicularity of an axis is often a refinement of a positional control. For the
moment, let us concentrate on size and perpendicularity. Using Figure 10-18, explain the
procedures needed to verify size conformance, then design a functional gage to check the
perpendicularity requirement. For this exercise, assume the surface area of your gage will
be sufficient to cover the datum feature. Include size and perpendicularity controls for
your gage hole. Do not give the gage a hole position. See Figure 10-19 for answer.
6. Dimension and tolerance the functional gage at the bottom of Figure 10-20. See Figure
10-21 for answer.

FIGURE 10-17

FIGURE 10-18

 Page 282

FIGURE 10-19
Answer to Question 5.

 Page 283

FIGURE 10-20

 Page 284

FIGURE 10-21
Answer to Question 6.

 Page 285
Receiver Gages
7. Using Figure 10-22, design, dimension, and tolerance:
a. GO gages for all size features to inspect MMC envelopes.
b. Functional gages to inspect each calculable virtual condition boundary.

FIGURE 10-22

 Page 286

FIGURE 10-23
Answer to Question 7a.

 Page 287

FIGURE 10-24
Answer to Question 7b.
Functional Gage Design
8. Design, dimension, and tolerance a functional gage (Figure 10-25) to verify position of
the four-hole pattern. Draw a sketch of the gage.
Functional Gage Design
9. Design, dimension, and tolerance a functional gage (Figure 10-26) to verify each
positional requirement for the four-hole pattern. Draw a sketch of the gage.

 Page 288

FIGURE 10-25

 Page 289

FIGURE 10-26

 Page 290

FIGURE 10-27
Using Figure 10-27 for the part to be gaged, answer Questions 10, 11, and 12 about the
gages.
10.
a. Should the gage hole position tolerance use a projected tolerance zone (Figure 10-28)?
Answer:
Yes
b. What is the gage pushpin diameter for D?
Answer:
.870
11.
a. Should the gage hole position tolerance use projected tolerance zone (Figure 10-29)?
Answer:
Yes
b. What is the gage pushpin diameter for the perpendicularity control of D?
Answer:
.900

 Page 291

FIGURE 10-28

FIGURE 10-29

 Page 292

FIGURE 10-30
12.
a. What is the gage fixed-pin diameter to represent datum feature D (Figure 10-30)?
Answer:
.900
b. What is the gage pushpin diameter for the position control of the four-hole pattern?
Answer:
.500
Functional Gage Design
13. Design, dimension, and tolerance a functional gage (Figure 10-31) to inspect each of
the two positional requirements. Draw a sketch of the gage.
Gaging to Represent Compound Datums vs. Primary, Secondary, and Tertiary Datums
14. Design, dimension, and tolerance one functional gage (Figure 10-32) capable of
gaging Option 1. Describe how the gage to check Option 2 might differ.

 Page 293

FIGURE 10-31

 Page 294

FIGURE 10-32
Part to be gaged.

 Page 295

FIGURE 10-33
Answer to Question 14 (Option1). In Option 2, the gage to inspect
this option might differ from Option 1 in that it could use a
diamond pin to represent datum feature D. The diamond pin
would simply be smaller in the direction that is not required to
control angular orientation.
Calculating X and Y Coordinates When a Bolt Circle and Angles between Holes Are
Known
The formula in Figure 10-34 is to be used to convert all hole location dimensions to X
and Y coordinates.
10.5 Gaging a Complex Sheet Metal Part
In this book we have spent a great deal of time going over the major principles of the
language of geometric definition and how to measure geometric tolerances. But just as it
is important to take a progressive approach to assigning geometric

 Page 296

FIGURE 10-34
controls to a product, it is equally important to have a well-laid plan to measure it. To do
this we must not only know the tools we plan to use but also how all the principles of
geometric definition will work together and consequently impact or alter our technique to
measure a workpiece. To put this in perspective, let us examine the workpiece in Figure
10-35, discuss its major principles, what they mean together, and how we will inspect the
part.
Size limits would be one item to consider first. One of the limits of size that may easily be
overlooked is material thickness. The sheet metal is purchased from a company or
division specializing in that sort of thing. This company produces the sheet metal, perhaps
in rolls, and has its own standards for all charac-

 Page 297
teristics of the metal including thickness. But when we receive the metal we work it,
punch it, form it, bend it, draw itall the operations necessary to make it into what the
designer envisioned. When we do this, the metal changes, stretching sometimes to the
point of breaking or tearing. A visual check is always performed to determine whether
fractures have occurred. But often no other inspection of material thickness or strength is
performed. This can be a mistake. If the material is thin enough, although not broken, it
can fail to withstand the stresses and strains of performance it receives once on the
vehicle.
The profile controls often are so large in their tolerances, say, for example, 1 mm per
surface, they were never meant to control material thickness and consequently strength.
Profile on a rigid thicker part regularly controls these material characteristics, but on sheet
metal parts it may take an additional thickness control that will be inspected locally, at
critical areas where bends may cause weakness. This can be performed with as simple a
tool as an appropriately anviled micrometer.
Next we can consider hole sizes. Since this is a thin part. it should be easier to inspect for
hole size violations than in a very thick part. Rule #1 violations, of perfect form at MMC,
even if in effect because the holes themselves happen to be rigid, could be verified at one
level. We don't have to worry about a banana-shaped hole or other 3-D violations of
form. Simply probing the hole with a coordinate measuring machine probe taking enough
points should be sufficient to determine size violations. Even simple hand-held tools such
as a dial bore gage or a telescoping gage set at the MMC and then at the LMC and used as
a GO and NO-GO gage may be sufficient. Or if a more theoretically correct tool is
needed, a pin could be used for the MMC condition and if it (GO) goes into the hole, we
could rest assured the envelope of perfect form required under Rule #1 for rigid features
has not been violated.
However, if the holes were toleranced with a positional control of zero at MMC, a fixture
if used to establish the datum reference frame could substitute for a GO gage and a
functional gage (to inspect the positional requirement between the two holes that form
datum pattern B and their angular relationship to A). This is true for zero tolerancing
situations because the MMC and the virtual condition boundary have been merged. They
are of the same size. Essentially, if zero toleranced, the two MMC envelopes would have
been oriented to datum plane A and separated from one another by the basic distance
(located in the CAD model/math CAD file) making them into virtual condition boundaries
that can be inspected with a checking fixture that represents datum targets A1A3 and
datum hole pattern B. However, as can be seen, in this situation the holes are not zero
toleranced at MMC. Since they have a tolerance of 0.5 at MMC to each other and datum
A, the size is to be inspected as a separately verifiable requirement from the position.
Still, the attributes of using a checking fixture with datum target areas A1A3, represented
by blocks of the target area sizes and two pins at the virtual

 Page 298

FIGURE 10-35
Gaging a Complex Sheet Metal Part.

 Page 299
condition size of 29 (separated by the basic distance given in the CAD file/model) to
represent datum pattern B, are many.
Consider the alternative: datum pattern B generates datum axis B, which is the central axis
between the two 29 mm virtual condition boundaries. Finding this axis without a fixture
would be difficult. Since the virtual condition boundaries are theoretical and only
physically embodied by such a fixture or gage, if a coordinate measuring machine is used
and the fixture is available, the CMM operator can simply probe the two fixture pins and
find the central axis halfway between the two. But without the fixture, the two-holes
would actually be probed and a halfway point between the two assigned as the datum
axisnot exactly the same. Still, with only a two-hole datum pattern, it might be close
enough to the same for most of us. However, if the datum pattern consists of more than
two holes, the gap widens between what the fixture would reap as a central axis and the
axis probing the holes would give.
Since every other control on this part used datum features A as primary and B at MMC as
secondary, two things result. The first is that some pattern shift is allowed. As the two
holes that form datum pattern B are produced well positioned using less and less of their
positional tolerance, and as these holes grow from their MMC size toward LMC, every
other feature on the part controlled to datums A and B at MMC receives an allowed
pattern shift. This means that the difference between the virtual condition of the holes and
their actual mating size determines how far the features controlled to them may have their
tolerance zones move not grow, but move.
Also, since all other controlled features use exactly the same datum features, in the same
order of precedence, with the same material condition symbols used after those datum
feature references (MMC used after B), they form a simultaneous gaging requirement as
redundantly noted by the SIM REQT local notes. This means all these features must be
accepted in the same orientation from A and in one optimum selected shifted location of
the part from datum feature axis B. In other words, any airspace existing between the two
29-mm fixture pins and the two holes of datum pattern B may be used to shift the part
for a more favorable measurement of the produced features, but all features must be
shifted in the same direction. Features that can only meet their individual tolerances
shifted in a different direction than the others are not in compliance with the simultaneous
gaging requirement rule and therefore are unacceptable. All features that are part of the
SIM REQT must maintain their relationship to one another almost as though they were
one continuous tolerance zone being simply shifted from side to side until all were able to
contain their applicable surfaces, centerplanes, or meet positional boundary requirements
as well as their relationship to the datums as a group. These measurements could be
accomplished through the use of a computer-assisted coordinate measuring machine, but
it would have to use a software package ca-

 Page 300
pable of simulating a tolerance zone pattern shift unless the CMM operator was willing to
shift the actual part on the fixture by hand until all features were in compliance with their
geometric tolerances simultaneously, or until it was determined they could not comply.
It is critical that the notes be read thoroughly, since they require the part to be restrained
on the datum targets during inspection. Such restraint, if it distorted inspection data (and
part configuration) would not be allowed without such a note. Ideally, the maximum
force of the allowed restraint would be specified in the note.
Also, since the general tolerance block includes a profile control of 1.4 mm to datum
features A and B with B referenced at MMC, this invokes (as is redundantly noted) a
simultaneous gaging requirement for all other part features not controlled on the field of
the drawing. These features must comply with their tolerances in one optimum orientation
and shifted location. Furthermore, they must comply with their tolerances in the same
gaging SIM REQT set-up/orientation/location as all features defined on the field of the
drawing controlled to the same datums in the same order and using the same modifiers (B
referenced at MMC).
10.6 General Principles and Terms of Tolerance Compliance
Statistical Tolerancing
Statistical tolerancing is a concept new to the dimensioning and tolerancing ASME Y14.5
Standard. It allows the use of tolerances on an assembly based on statistics.
A commonly used assembly tolerance formula is: the square root of the sum of the
squares of the individual tolerances. This method has the advantage of increasing
individual feature tolerances for the reduction of part costs. This method may be
employed when a manufacturing system is under statistical control through the use of
such methods as statistical process control. It should not be used when a manufacturing
facility or process is out of statistical control or in statistical chaos, producing parts
without appropriate repeatability and/or accuracy. The more restrictive tolerances derived
by arithmetic stacking limits should be employed in these situations.
A note may be added to a drawing such as Features identified as statistically toleranced
using the ST symbol shall be produced with statistical process controls, or to the more
restrictive arithmetic limits, in which case a feature will be listed with two tolerancesone
tighter tolerance for use in facilities out of statistical control and one a looser sta-

 Page 301

FIGURE 10-36
tistical tolerance (ST) to be used by manufacturers under statistical control and using
statistical methodology in their production procedures. If the tolerance is a statistical
geometric tolerance, the symbol is used in the feature control frame after the geometric
tolerance and any modifier it uses.
When Tolerances Apply
Any dimensions and tolerances given a feature to be plated or coated shall specify if the
dimensions apply before or after plating/coating. A general or local note to this effect is
recommended.
All dimensions and tolerances apply in the free state and at 68F (20C) unless specified
otherwise. Parts to be restrained while inspected must state so with a drawing note. As a
clarifying redundancy, it is permissible to use the free state symbol (a circled F) inside the
feature control frame to indicate the part is measured in the free state. However, even
without the use of the free-state symbol, the part is measured in the free state unless a
restrained condition note is specified.
The same is true of the temperature at which a part is measured. Unless a note is used to
specify that a part's toleranced dimensions apply at a temperature other than 68F (20C),
the part must meet its dimensional requirements at that temperature.
All dimensions and tolerances apply only at the drawing level at which they are shown.
For example, dimensional requirements given at the detail drawing level do not have to
be met at the assembly drawing level. Therefore, it is recommended that each level of
drawing retain or specify dimensions, tolerances, and relationships important at that level
that may otherwise not be met.

 Page 302
Absolutes
All limits of a dimension are absolute, regardless of the number of decimal places. It is as
though an infinite number of zeros existed after the last number given. For
example: 15.3 means 15.300
15.0 15.000
When the part is inspected, the measured value is compared directly with these limits.
Any measured dimension outside the limits is in violation of the tolerance, regardless of
the number of decimal places shown in the dimensional limits.

 Page 303

11
Merging Theory with Reality
11.1 Merging Theory with Reality (Having the Right Measurement Tool)
There is a great section in ANSI B4.4, the Inspection of Workpieces Standard, that I
summarize at least a dozen times in each Dimensioning and Tolerancing workshop I
conduct. The gist of it has to do with the way we merge the guidelines put forth by the
standards of our nation and our workplace realities of limited time and resources.
In essence, its message is that the theories of part and tolerance zone verification set down
in the standards must be complied with as closely as the confines of our environment will
allow. Even the most sophisticated pieces of inspection equipment do not often check
every element of a surface to determine that they do indeed reside within the tolerance
zone prescribed. The inspector invariably checks as many elements as he or she feels
necessary to achieve a confidence level sufficient to create the belief that had all elements
been inspected, they, too, would have checked good.
To carry that theme a step further, should we not possess in our environment the
inspection equipment necessary to completely comply with the most correct method of
verification, we improvise. Cooks often do the same thing and quite regularly turn out
acceptable results. This is the real world, as the saying goes. I'm all for it. Still, there is
that boundary of common sense that says don't substitute cyanide for sugar.
I consistently hear about the unique situations ASME/ANSI Y14.5 users believe they are
in. They are all correct. No matter what the nature of your parts whether they are plastic
or sheet metal, very small or extremely large, have tight

 Page 304
tolerances or irregular surfacesyour environment is special. However, just as it is the
nature of your business to work on such parts, it is also the nature of your business to
equip yourselves with the necessary tools to properly manufacture and inspect those parts
according to the design requirements. If your verification procedures are not closely
reflective of those needed, you are only succeeding in giving yourself information no one
has asked for, and supplying yourself with a short-lived, false sense of security.
For example, checking size features for maximum material condition (MMC) envelope
violations with a micrometer just isn't close enough. A size has controlled for form check
is needed and a cross-sectional check doesn't give that information.
Vee-blocks to hold datum features for runout, concentricity, and positional coaxiality
checks when the datum features are so far out of perfect form that they jump around like
live fish on a hot skillet don't stabilize the axis and are, therefore, not close enough.
I've visited some companies where as soon as I began to discuss functional or receiver
gages, I have been stopped with We don't have or use them. So, I moved on to coordinate
measuring machines and have again been stopped with We don't have or use them. Okay,
I say, let's talk about dial indicators and height gages. And, you guessed it, I've gotten We
don't have or use them again. Spend a few dollars. You can't inspect geometric controls
consistently with a keen eye and sleight of hand tricks.
Reality imposes some harsh realizations on purists who expect every inspection technique
to precisely recreate the set-up theory conveyed by geometric or other controls. A little
substitution must be expected and, indeed, innovation and creativity are necessary to
merge the standard's guidelines for verification and your own particular environment.
Still, don't fool yourself into believing your parts are so unique that you shouldn't try.
Comparing those Y14.5 users over the years who use that excuse, I have found them no
more uniqueand usually less creativethan the rest. Reality is a tough thing, but those most
successful in dealing with it are usually those who try the hardest.
11.2 Theory vs. Reality (Uncertainties)
All measurement procedures are flawed in some way:
The instruments we use are not perfect.
The procedures the inspector performs are not exact or entirely appropriate.
The fixturing devices don't meet the requirements to establish the datum reference frame
(not a problem for circularity since no datums are allowed)or worse, no fixturing is used
at all.

 Page 305
The computer-assisted measuring instruments are programmed with faulty, inadequate
algorithms.
The probes do not touch all points on the surface and those touched are often improperly
distributed for the geometry commonly produced by a particular manufacturing
procedure.
The inspectors are often trained to perform tasks in an inappropriate manner or perhaps
not trained at all. Often they only know what the machine can do, not what the geometric
control on the part to be inspected really means and what operations must be performed
to verify compliance. And the worst thing is, inspectors often believe that what they are
doing is absolutely correct. In this state of blissful ignorance, there is no reason to study,
learn, explore, or elevate capabilities. Others often see these misguided efforts at part
verification and marvel at their inadequacies. Occasionally, an undiplomatic urge compels
them to say, Don't bother performing that measurement procedure. Why don't you take a
coffee break instead? Often this angers the inspector. What are you talking about? he
yells, I'm working here! Can't you see I'm working? They see it, but what they see is an
operation that collects information so vastly lacking in the area of part functionality that it
may as well not be performed at all.
Vee-blocks are an example of well-intended instruments that may produce inadequate
measurements regarding functionality. Vee-blocks, incapable of part axial stabilization,
allow the axis to move as one attempts to measure from it. This movement can give a
false impression of a feature's circularity or cylindricity, since, by definition, we attempt
to assess those geometric characteristics of the surface from a stabilized center. If the vee-
block is used to establish a datum axis for the purpose of measuring a controlled feature's
circular or total runout, concentricity, position, or conicity, it can add the datum feature's
surface deviations from perfection and the subsequent datum axis movement during
rotation to the FIM of the controlled feature. These factors make the resultant
measurement data suspect. A football-shaped object, for example, may read round, or a
surface well within its runout tolerance may read out of tolerance. For these reasons, (a) a
bad part reading good and (b) a good part reading bad, the vee-block can only be relied
on to give a rough estimate of part compliance to given geometric tolerances.
If two vee-blocks with different Vee angles are used on the same part, they are more
likely to discover that a bad part is reading good (for example, round). Likewise, two-
point opposing point measurements are unreliable for checking form deviations like
roundness and straightness or perfect form at MMC requirements. A part that is produced
as three-lobed is liable to measure similarly at different locations on the surface if
measured with a micrometer. It may appear to be round if only opposing points are
measured. This method is, therefore, only recommended as a very rough check of
geometric compliance. Likewise, similar

 Page 306
methods attempting to verify the MMC envelope under the Taylor principle (Rule #1) may
easily fail to dectect axial out-of-straightness, allowing a feature to pass when it actually
violates the perfect form at MMC requirement.
11.3 Reactions to a Nonperfect Datum Establishment
A somewhat irate caller phoned to say he was upset about the explanation given in the
Y14.5 Standard for the establishment of a compound datum axis (or, as it's sometimes
referred to, a simultaneous datum). An example is shown in Figure 11-1.
The inspector complained, This type of control calls on me to establish an axis in a way
that I cannot comply with.
The theory is explained as a common axis formed by both features simultaneously.
Minimum circumscribed cylinders contacting features C and D would give this axis.
Chucking both simultaneously is one way to think about it. The axis achieved is then
rotated about and deviations from perfect (in this case) total runout (cylindricity and
concentricity) are recorded with full indicator movement. Chucking both datum features
simultaneously isn't easy.
This business of simultaneous datums makes a lot of people angry. Some people indicate
them in. This inspector said he could do that, but it wouldn't be perfect. Likewise,
centerdrilling each separately and then putting them between centers offended his
sensibilities. It is understandable. It's less correct than indicating them in. Both are done.
Still, as the man said, It isn't perfect! No kidding! We live in an imperfect world.
He was asked, What if only one datum feature was used in the control?
One diameter?
Yes, only C, for example. What would you do to establish that axis? he was asked.

FIGURE 11-1

 Page 307
Grab it, he said.
With what?
A chuck, he said.
What kind?
Three jaw, he said
Not good enough, was replied. It would not give the axis of the minimum circumscribed
cylinder.
Ten jaws, he said.
Not good enough. Even ten jaws wouldn't encompass all elements of the surface. Even a
collect would miss a few that peeked out between its seams.
At that point, the message was becoming clear. Just as when ANSI B4.4, the Inspection of
Workpieces Standard, relates the theory of GO gages, then subsequently says even though
the theory calls for it, a full form gage is not always necessary or used (See Section
1.4.2.2) if it is known that, with theprocesses used, the erroris so small that it does not
affect the character of the fit of the assembled workpieces.
Let's face it, when measuring force is explained, it says, all measuring operations given in
or in any way associated with this standard are understood as being referred to a zero
measuring force. All readers understand some force is applied, but the error is ignored or
compensated for. Compensation is a tricky, nasty business, and most people pretend the
error isn't there or ignore it because they believe that it really isn't harming or negating the
data being collected. And, in many cases, they are absolutely correct.
When an inspector checks any surface geometric control with a dial indicator or probe, he
or she is expected to touch every element of the surface. And, although theoretically
possible, no one has ever done so. Inspectors check as many elements as they need to,
and assure themselves that, had they checked them all, all would have been within the
tolerance zone. It's a confidence builder. They want to be able to tell themselves, No one
out there is going to find an element of this surface that falls in the forbidden zone. They
know they are responsible for them all, but, as inspectors, they cannot afford to spend the
rest of their lives checking one part. Sothey settle. There are recently developed
guidelines for certain part configurations that help them decide when to settle. But, let's
face it, they still settle.
Once they realize what they're doing is not perfect, not exactly complying with the theory
of the control for set-up or inspectionever, the whole picture comes into a better
perspective. Just as a designer realizes he/she must tolerance a part because one can never
make perfect parts, the inspectors must realize the inspection set-up is full of errors. No
piece of equipment they use is perfectjust some are better than others. There are no
perfect set-ups. But they try. The inspectors examine their guidelines, standards,
environment, equipment, and budget, and then hopefully infuse some knowledge of part
function and come up

 Page 308
with the best they can do that day under the given circumstances. It's judgment and
professionalism and fear.
They know if they settle for too much imperfection, eventually it will come home to
roost, laying goose eggs in the dollar sign column of their company's account book.
Business goes away in one form or another if they don't try hard enough to comply with
known requirements and theoretically perfect setups or if they don't have a firm grasp of
them to begin with. Things can afford to be perfect only if they are imaginary. In the
harsh light of reality, these theories of perfection sometimes merge poorly with the
capabilities of our environment. Still, inspectors continue to keep in mind those
unattainable goals and strive to achieve them as best they can.
The inspection procedure will never be perfect, but in order to even aspire to perfection,
one must first know what perfect is. It's something to keep in mindthe bull's eye to shoot
for. It reassures them that they have tried to do the right thingtheir very best.

 Page 309

12
How to Improve Quality
12.1 The Zen of Quality
Before aspiring to perfection, one must first know what perfection is. This is the
frustration and challenge of the manufacturing environment. Manufacturers often wish to
produce what was in the mind of the designer, but are saddled with the very human frailty
of not being able to climb inside that mind and walk around the object viewing all its
functional requirements. Instead, a design drawing is produced, and whether on paper or
in the data base, it becomes what represents perfection for that particular product.
Unfortunately, this representation is not exactly as it appeared to the designer. He has used
the language of illustrations and words and symbols to simulate this perfect idea, but in
the translation, much is lost. The representation lacks the clarity of the idea, and even if it
was translated perfectly, it may not be produced perfectly. So, it receives tolerances.
Boundaries are set, within which it is assumed that if the part resides, it will function. Yet,
the person setting these boundaries is often unsure of how they should be arrived at. How
far from perfect may my idea be produced and still function as I perceived it? And so,
sometimes the boundaries are guessed at.
These guesses are transmitted to the manufacturer and the product produced. The
physical embodiment of the guesses, the produced part, is then measured per the
documented geometric definition and its boundaries, the dimensions and tolerances. If the
part is perceived as within these tolerances, it is passed along for assembly or consumer
use. Still, if the tests performed to determine compliance are not reflective of how the part
functions, it may pass these inspections and still not function at all, or perhaps function
for a while and then break down. Clearly, this is a highly inefficient way of doing
business.

 Page 310
There is a saying that business is war. Unfortunately, sometimes the most decimating
battles are fought within one's own ranks. These internal squabbles drain the best part of
the employees and leave little to create and solve. Innovation and pride in one's work take
on secondary importance to survival. If business is war, it would seem that many do not
recognize the enemy is not within one's own company.
Lasting success can only be achieved with harmonyin mind, spirit, and environment. This
harmony begins with communication and cooperation between members of one's own
department and between departments and with extended members of the company family,
the vendors. Until this harmony occurs, politics and pettiness will reign. Confusion and
frustration will be ever-present. And inferiority of product will be the inevitable result,
and failure the destiny of the company. This suicide is committed by companies every
day. Sometimes it is slow in coming. At other times swift. But however it manifests itself,
it is therewaiting. The symptoms are easily visible to the outsider looking in. To a visiting
consultant, it is exactly like watching the soldiers within an army turn on one another and,
to assert their own dominance, kill off one another before the real enemy, the competitor,
is ever engaged in battle. When the competing army of corporate warriors finally comes
over the horizon, it sees only the ragged remnants of a once potentially formidable foe.
The battle is over before it begins.
The company that has forged a synergistic strength within its ranks thinks and acts as one.
Its goals are ultimately the same goals. Mistakes are made, but they are few, and they are
quickly repaired. Blame is not placed, but causes found and eliminated, replaced with
well thought out, interdepartmentally agreed upon strategies. The products are made
stronger, better, and quicker. The company not only endures, but flourishes. And being
an integrated part of the company, the employees flourish with it.
Teams are forged over time. They are forged with great effort and sometimes with
individual goals of recognition and achievement taking their place far behind the
company's.
Concurrent efforts must be instituted during the design process to achieve the synergy
that comes of having all members of the product team involved and contributing their
experience and knowledge of their particular disciplines. In this way, manufacturing
knows exactly what must be produced in order to function optimally, and inspection of
that product is carried out in a manner that is most reflective of how the product will be
asked to function by the consumer. Boundaries within which a part must reside in order
to perform optimally are not guessed at, but rather worked out by the team members.
Producibility and inspectability are achieved while not adversely affecting functionality.
The process flows like a gentle stream. Employees look forward to their work, even look
upon it as a gift. Individuals, while maintaining their individuality, become part

 Page 311
of something greater than themselves, and realizing they are achieving at a level that could
only be attained with the other members of the company's family, they respect one
another. With this respect, which manifests itself in the way they are treated and treat one
another every day of their working lives, comes the harmony that is needed to work and
succeed optimally.
12.2 Tying Measurements to the Machine
Measuring machines can integrate the measuring function into the machining process by
being physically attached.
In-process gaging offers several benefits and advantages to manufacturers.
It gives immediate feedback on and control of product features during the manufacturing
process.
An in-process system, once installed, requires relatively little set-up or adjustment.
It provides increased control, which reduces the probability of producing out-of-control
parts.
It provides an historical data base of how machines, parts, pallets, and fixtures behave
during the actual machining process.
The gage computer communicates the information back to the machine computer and any
necessary adjustments are made automatically and immediately, virtually eliminating
scrap.
12.3 How Many Points Should I Probe?
An Electronic Measuring Devices, Inc. study conducted in 1988 showed the following:
Six one-inch-square blocks were tested.
They represented many different deformities.
Each was scanned in a grid that ended up with about 2500 evenly spaced data points.
It took between 80 and 357 points (average 191) per square inch to achieve a 90%
certainty of form.
It took between 48 and 227 points (average 102) per square inch to achieve a 75%
certainty of form.
12.4 The Search for Common Sense (Statistics Out of Control)
An employee of a large organization called me recently and frantically proclaimed, My
company has lost its collective mind! Entire departments have

 Page 312
been created that never existed before. And most of them exist to feed themselves, like
some alien thing that sucks the life out of the company and all the employees within.
They send out representatives that act as their tentacles and they attach themselves to you
and your environment like giant drinking straws draining off all common sense and
possibilities of doing a good job. Common sense is no longer common! Everything done
goes into feeding these Gods of Statistics. I know statistics, if used appropriately, serves a
purpose in life and in industry, but the way I see it being used is killing me and ultimately
will destroy our product. And anyone, and I mean anyone, who speaks out against the
absolutely insane way it has taken over may as well fall heart first on a rat tail file. It is
political suicide to challenge the Gods of Stats. They calculate the measurement
uncertainties of our inspection techniques, and then instead of correcting them or even
treating them individually, they just add them in as one lump number to the final
deviation from perfection of our product. This has the effect of causing perfectly good
parts to be rejected. Since less of the yield of our product is accepted, and little effort is
made to correct the problems as individually treatable symptoms of the overall
uncertainty, the situation continues on and adds to the overall cost of our product. We're
losing market share to our competitors, who have yet to institute such a system. My only
hope is that they catch up soon, so that they too can reject a bunch of perfectly functional
parts. Help!
Shortly after that, while at another company, I was taken aside by an engineer nearing
hysteria, who whispered, The statisticians have invaded my company and eliminated
every maximum material condition [MMC] symbol from every drawing we create. They
say that there is no gage repeatability with the MMC concept and that there is no way they
can chart bonus tolerance on their statistical process control charts [SPC]. They don't care
that functionality allows additional tolerance based on larger hole and smaller shaft sizes.
They don't care that these things allow reduced product costs that we can pass on to the
customer. They don't care that the elimination of the MMC concept makes all those
perfectly functional fixed-size gages and fixtures no longer allowable to use and instead
ask us to use expanding fixtures for everything. They say this increases repeatability and
reproducibility, but I've used these things, and I know that even if we could afford them
(they're very expensive), these expanding fixturing and gaging elements aren't repeatable
and the data they eventually give us is not reproducible. Sometimes when you expand a
fixturing or gaging element it moves the part and at other times it just locks up on the side
of a feature not moving the part at all. And I'm not talking about different parts either.
You can use the same checking fixture on the same part six times and get six different
measurement results. And since the mating parts in the assembly don't expand, we aren't
recreating the functional situation in any of these expanded locations and orientations, so
what are we accomplishing?
Also, since we can't use fixed-size gages for anything, we are forced to

 Page 313
use a coordinate measuring machine [CMM] for everything, which is what they want so
that they can again chart all the data, no matter how poorly and nonfunctionally it has
been collected. Which brings me to my next problem. All the CMM operators have been
told by the CMM sellers that fixtures are entirely unnecessary when using a CMM, to just
probe a few random points on the surfaces. Forget all the studies that have been done that
say you need over a hundred points within each square inch of surface to even achieve a
75% certainty of form. They are told to believe they can probe the datum features any
place they choose and get the same datum planes and axes that the fixtures would give if
probed while resting on the highest points of contact. What in the world is going on? I
feel like one of the few people around here with anything above the ears besides mushy
oatmeal. There are others, but they've been intimidated so badly, they've decided to just
go along to get along. But this is wrong, wrong, wrong!
I'll tell you what I think. I think we are creating a monster that doesn't use any common
sense whatsoever. It doesn't care about how the parts function. It doesn't care about the
rising cost of our product. It just cares about the act of performing statistical calculations.
It only cares about the satisfaction it receives from rows and rows of ultimately
functionally meaningless numbers it can churn out and put on a chart to show the
admiring masses. But what can we do? Merely the act of voicing a concern can end your
career. You're labeled as not a team player. But what we don't realize is we aren't the team.
The statisticians have become the team. All we are is the equipment, no more important
than the copy machine and far less important that those damnable SPC charts that they
carry around like the Holy Grail.
The thing that really upsets me is that these guys have skills, just no common sense. They
think the means is the end, not a tool to improve the end result, which should be the
improvement of the overall company and its products. Higher costs don't improve the
company standing, especially if those higher costs are based on nonfunctional processes
created only to feed the Stat Monster. This thing has gotten way out of hand. Is there any
hope for us?
My response to both was the same. There is nothing wrong with statistics as a discipline
of mathematics, but there is often something wrong with the way it is applied. And, of
course, any suggestions for improvement made from the data collected are only as good,
and as reliable, as the data itself. You may find some comfort in the fact that I have been
reading a number of authors recently who say that they think fool proofing will someday
make the need for many statistical procedures unnecessary. One example they give is
using fixtures that both inspect and hold a workpiece for the next manufacturing
operation simultaneously. I love telling this story because it makes all those who have
made statistics their God nervous and consequently angry (no matter how badly or ineptly
the statistics are practiced). They turn red in the face and hiss through clenched teeth that
this type of thing will never happen, and anyone who would spread

 Page 314
such a tale must not be a team player. Still, it is worth considering. If anything can reduce
costs, improve quality, save time, and improve the mental health of a company's
employees, it has to be worth a look.
What if all manufacturing and design considerations had to consider functionality first?
Products would be designed to work well and be easy to manufacture. We wouldn't deny
the manufacturing people any of the functionally acceptable tolerances just so an SPC
chart could turn out an acceptable statistical number. Yes, we would still monitor our
processes, but not throw out parts that don't fall within a functionally meaningless
statistical range. There would be jobs for the statisticians to perform, but they would not
be able to change fully functional, low-cost, common-sense manufacturing procedures.
All measurements could be directly related to whether or not a product would work well.
Manufacturing and measurement procedures would be improved by targeting specifics
and changing them to be better, not just identifying a vague nonrepeatability or
uncertainty and adding that lump number to the measurement deviations of the product
features. Product yields would be greater, and yet all increases in yield would be of
products that functioned just as well as those rejected by previous nonfunctional
requirements. Fixturing would be used whenever possible to recreate the way a product
fits into its overall assembly. We would no longer take the word of nonsensical statistical
data that claims to be able to interpolate points on a parts surface that have not been
probed or inspected. Gages would play a greater role in part inspection, and in many
instances, the fixtures would become the gages. Variables data collectors like the CMM
would be used optimally, not alone, but with help, the help of fixturing devices and
augmenting gages that would allow the datums to be formed in a more efficient manner
than if the CMM probes were brought in contact with a few surface points and from those
randomly selected points datum planes and axes formed, pretending to be the same as, or
as good as, the high point planes and the axes created by probing the high points.
Isn't it interesting how we will try to repair a manufacturing procedure based on data
collected in a grossly nonfunctional manner. And in doing so, data collected after the
correction are often worse than the data collected before, or the data seem to be better, but
the parts don't work as well as before the correction. A resurgence in the use of and
interest shown for gages and fixtures around the world in the last couple of years is
probably not just a coincidence. I think some, many perhaps, have just plain lost faith in
these nonfunctional procedures that don't seem to have a basis in good, sound common
sense. And some folks are running to procedures and tools that they know, from their
many years of solid experience and judgment, are trustworthy. Most often they don't want
to get rid of the new procedures altogether. They know they have some merit. But what
they are rebelling against is a mindless adherence to this new cult-like subservience to
procedures and data that seems to have no basis in good, sound logic. They want to test
it, to examine it, with some of the tried and tangible tools and techniques

 Page 315
they grew up with. And when they do that and the information comes out different from
the information they are being fed by these techniques that have been forced down their
throats, they want to discuss it, improve it, and not just blindly go where they are being
forced to go no matter what the consequences just because it happens to be the program-
of-the-day. These programs come and go, but years of experience and logic and good
common sense are things we have come to rely on and need to use. Search for the sense
in an act, don't just perform it. Use all your judgment and reason to determine if
something has merit before accepting it as your mission in life. And remember that
throughout history, I was just following orders has never been an acceptable defense.
You are a thinking, knowledgeable human being, the culmination of a life-time's
experience. Don't just throw that away.
12.5 Quality Issues from the Afterlife (Slogans vs. Programs and Planning)
Several years ago a company was making printers that, when you turned them on, would
shock you to death. Kill you! Shouts could be heard from the great beyond, Hey, I just
wanted to run a hard copy! and You guys can forget about repeat business from me!
When I brought up the issue at that company, the employees got mad. Who the heck are
you to tell us how to do our jobs? they sneered.
I'm the consultant you hired to help you to stop killing people, I said. Besides, I bought
one of those printers. Would you like to turn it on for me?
I don't get it. When people do a rotten job, they get angry at those who point it out. I'm
nave enough to believe this is information they need. Even that, by giving them this
information, I am helping in some small way.
It's interesting in that people hire me to look at what they have done and critique it. But
what they really want is for me to look it over and breathlessly proclaim it the most
perfect, wondrous piece of work I have ever beheld, a blend of Einsteinian brilliance and
Rembrandt's aesthetics. Tell me how great I am, their dancing eyes plead. I say, if you ask
a question, make sure you want the answer.
There should be an awareness by us all that we are human and therefore flawed. We
should all have among our goals to improve, to learn, to growto listen, to be accountable.
Employees of large companies often have this false sense of security that if they do
something illegal or unscrupulous that results in injury or death, the company will protect
them from the victims and the law. The reality, as I have observed it, is that when they are
caught, the company can't get out of the way fast enough. There he is, they say. We had
no idea what he was doing. Take him away. We can't stand the sight of the little crook!
I did some work for a firm that made heart pumps. When you walked into the worksite,
you could hear the thumpthumpthump of the product. It

 Page 316
was as if the shop floor itself was alive. There was a wall covered with photographs. Each
was a picture of a person who was currently hooked up to one of their heart pumps. Next
to each picture was a history of who the person was, what had happened to him, and how
he was doing. I was deeply moved by the firm's commitment to quality and
accountability. What employee could read that wall and do less than his best each and
every minute he was on the job?
Great companies have an intimate connection between their products, their employees,
and their customers.
Some comments I hear at other companies are scary. One employee of a company that
made knee implants shocked me. I mentioned that they must feel a tremendous sense of
fulfillment, producing something that made it possible for invalids to walk again. His
response was We love it when the winters are bad. The ice and snow cause so many
accidents that our product demand soars. I'm pretty sure that's not the slogan portrayed in
the company's television ads. I must point out that he was not representative of the
company's overall employee attitude, but he was a virus in their system. Others laughed
and nodded at his comment. It's cool to be insensitive.
Some companies have unhealthy cultures. If TV ads were truthful, their spokesmen
would proclaim, Buy our junk. We're prone to bouts of hysteria or We don't correct
anyone's mistakes here. We're too polite or We think we're great. And no one better tell us
different or we'll kill them (politically at the very least) or We're the government. We can
do anything we want! or If we don't manufacture what we designed, we just change the
specs! or Who cares about the customer, we have great marketing. People will buy
whatever garbage we make.
Don't get me wrong. There really are some great companies out there. The problem is that
the public often can't distinguish which products are actually great from those whose
marketing simply says they are.
Take software for example. When you buy a software package, do you actually take the
time to research whether or not it can actually perform the tasks (correctly) it claims to be
able to perform? Or do you simply take their word for it like most folks? See, this makes
it hard for a truly great product to be successful. The manufacturer might spend all their
money on research and development, or manufacturing and quality. Then they come into
a marketplace and realize success isn't necessarily based on what you can do, but rather
what you claim you can do, in a very loud and glitzy marketing campaign. It can be
discouraging to those with actual scruples. Many companies, thank goodness, aren't
burdened with those. If you consider that they teach M.B.A.'s the purpose of business is
to make a profit, and that salesmen are paid on commission, the fact that it all has a
seamier side shouldn't be surprising.
The truth is that Quality is Job I and Zero Defects are merely slogans. It's much easier to
think up slogans than to improve a product or change the

 Page 317
deeply ingrained beliefs of a workforce. It would be comical, if managers who think this
stuff up, didn't take it so seriously. Believe me, most workers don't. But they are usually
smart enough not to voice how simplistic they know the slogan approach is. They know
entire departments are born out of these things. Their future promotion can easily lie in
how fervently they repeat the mantra of the day.
They might even head up a task force on that some day! It sure beats working.
Hey boss, I just had to let you know how excited I am about our team-work approach to
total quality management [TQM] and how it relates to our QS 9000 certification efforts.
Promote that guy! He can talk the talk.
I once asked a quality manager who had mentioned TQM to me, What are the
components of your TQM program? She just looked at me with a wide-eyed stare. She
tried to tough it out, but the stench of fear was in the air. She was talking the talk, but had
no idea what the talk was.
If you think I'm being cynical, tell me whatever happened to quality circles? Before you
panic, let me tell you the politically correct answer. Quality circles evolved over time to
become team meetings. I know, that sounds plausible. You see, I can talk the talk too.
12.6 Focus and Excel (Maximizing your Abilities)
Somewhere, perhaps a long time ago, someone such as a college professor told you that
to be a good engineer, designer, measurement expert, or professional in general, you had
to be able to think about a hundred different things at once and do it well. That person, or
persons, whoever they were, lied to you. I've trained more than twenty-five thousand
people in my career and only about five of them could think globally and do it well. And
they weren't happy people! There is a high suicide rate in this IQ category. Not something
I covet, nor could ever achieve. Most people in the professions we (the people likely
reading this) occupy are above average. That's what we are; above average. Even if you
are a genius, chances are you are only a genius in a couple of things, not in everything.
Think of the man so many of us associate with genius: Albert Einstein. Even he wasn't a
genius in everything. Would you, for example, want him as your hair stylist?
Most of us would be wise to maximize our intelligence by focusing our attention on one
thing at a time, giving that thing all of our attention, accomplishing the task it presents,
then moving on and finding something else to do, until all facets of a project are
complete. It's like putting together a jigsaw puzzle. We'll keep in mind what the picture on
the front of the puzzle box looks like, but we can't forget that we have only one piece of it
in our hand at a time, and focus all of our attention on that piece until we either fit it into
the proper

 Page 318
place or decide to put it down and focus on another piece. Eventually, we will complete
the entire puzzle. We won't miss even the smallest detail. By following this tactic we will
do a better job. It will be more accurate, more error-free, more complete, and believe it or
not, it will be done faster.
People who look at things globally most often diffuse their intelligence so much that they
become less competent, often incompetent. They panic and freeze up or do things slowly
or in a chaotic frenzy. An example is when someone complains about his workload. He
looks at a desktop buried in projects and proclaims, My God! I'll never be able to do all
this! You sometimes read about someone bringing a gun to work to punish those who
have caused his troubles. Even then, that approach probably reflects the employee's lack
of focus, shooting anyone who happens to be there. More than likely, the source of one's
problems looks back at him from the mirror each morning. If he had just asked himself
the right questions in the first place, instead of trying to exceed his intellectual abilities, he
probably would have been okay. Yes, there is a stack of projects on your desk, but: (a)
Isn't one of them due first? and (b) In that file that is due first, isn't there one task that
needs to be done before any others? In other words, begindo something, but not just
anything. Take the time to put things in order, then focus on accomplishing the ordered
items. Recognize your weaknesses and make them your strengths. No, unless you have
infinite experience and a stratospheric IQ, you can't accomplish or even think about
everything at once. Quit crying. It's a fact. Now that you're over that hump, focus on what
you are good at, doing one thing at a time and doing it well, maybe even better than
anyone else could do it (at least anyone available today). Begin, focus, persevere, be
tenacious, finish. Begin again. It doesn't even matter what your job is, this works. Just do
it. Quit staring at it. The pile isn't getting smaller. Unless your boss starts giving you less
work, and if so, watch your back. The old manager's adage is to give the work to those
able to accomplish it. If Joe has a lot of projects and you don't, it isn't time to celebrate.
It's time to update your rsum. The boss has discovered your depth. He's standing in it
and his socks aren't even wet.
You can be less talented, less brilliant than those around you and still run rings around
them. Just focus all your abilities, all your attention, and peck away at it. Soon you'll
realize you can do almost anything, especially if you're willing to admit you don't know
everything in the whole wide world and ask others for advice from their areas of
expertise. Don't worry about showing weakness of knowledge. In the long run you won't
look weak. You'll look curious and confident and successful. And if they don't want to
help you, and berate you for not knowing something already, ask someone else. To
paraphrase Henry Ford, I might not know the answer to that, but give me five minutes
and I'll hire someone who does. Take the criticism of those crevice dwellers who would
tear down your ego and wave at them as you rise above and watch them sinking slowly in
their own self-assigned importance. And if you have a bad day here

 Page 319
and there, it goes with the job of pushing air in and out. Take it seriously, but don't
agonize over it. Take it from a guy who watched his life's blood flow out of an ulcer the
size of death's door. With the right perspective, life and the job can be fun and even the
tribulations become amusing. William Shakespeare portrayed that perspective in Macbeth.
He wrote: Tomorrow, and tomorrow, and tomorrow, creeps in this petty pace from day to
day, to the last syllable of recorded time; and all our yesterdays have lighted fools the way
to dusty death. Out, out, brief candle! Life's but a walking shadow; a poor player, that
struts and frets his hour upon the stage, and then is heard no more: it is a tale told by an
idiot, full of sound and fury, signifying nothing.
12.7 Competing in a Global Economy (Great Products at an Affordable Cost)
In 1982 I was introduced to a Japanese businessman who gave me some sage advice. He
said, It seems like you Americans are worried about competing in a global economy.
There is no need to worry if you are willing to do what is necessary to succeed. To
prosper you only have to do a few things. You have to make a product that the public
needs. It must work extremely well. And it must cost as little as possible. He was right
then and he's still right today. The problem was we had many companies that weren't used
to the concept or the competition.
In 1980 I was called in to help a defense contractor that had to fill a large government
contract for a weapons system. Each system cost millions of dollars and they were to
make thousands of them. Each system was to be deployed right on the battlefield to
engage the enemy. When I arrived, they told me they had already built one hundred of
these systems and they were in the field being used.
They said there was an initial problem with the system, in that it housed soldiers and they
had made it out of magnesium. The first time it was hit by unfriendly fire, it burst into an
inferno-like flame, lighting up the night sky and baking anyone unlucky enough to be
inside. I asked them if they hadn't owned a chemistry set when they were kids, and
thought back on when you lighted the magnesium strip in the set, it burned down the
house. They said magnesium was chosen because it was light and easy to transport, but
they had switched to aluminium after the first tests were conducted.
Now they faced another problem in that every one of the first one hundred systems they
had built were prototypes. I asked what that meant, and they responded that the systems
had no interchangeable parts. A breakdown occurring in the field meant they had to
remove the nonfunctioning part, return it to the factory, measure it, build another that
size, and then return it for installation. I asked if that didn't slow the battle down a little.
I was hired to train several thousand employees in some critical technical areas, including
dimensioning and tolerancing, and to help institute some

 Page 320
changes in the design specifications. We stopped counting at 28,000 changes. And since
each change cost two thousand dollars to push through, the drawing changes alone cost
the American taxpayer fifty-six million dollars.
This, of course, did not include the cost of changing the manufacturing procedures, the
fixtures, and replacing the gages. I guess they could have taken the time to get all this
right the first time, but it seems that in some organizations (many, in fact) there is more
time and money to correct problems than to prevent them.
Many of these military contractors were left without a customer or salable product when
the government cut defense spending about the time the Soviet Union dissolved. So they
began scrambling to find a product they could produce and sell to the private sector. A lot
of these organizations were so used to working on a cost overrun basis and getting jobs
based on political clout that they had trouble competing. The companies they had to
compete with did business differently, even efficiently in some cases. So, it was
recognized that a change had to take place. I was asked by one such company to come in
and conduct a training course on concurrent engineering. Their design department was
running the show, with manufacturing and quality having no input as to how the product
was designed. Consequently, manufacturing personnel were constantly at odds with the
designers as were the quality people. The new products were overly expensive to make
and inspect, so to make a profit they had to be sold at a price few seemed willing to pay.
But when the designers were told this, the finger pointing began, and the departments
went to war with one another. I guess no one realized that successfully destroying each
other would mean the death of the company and loss of their own jobs and income.
Everyone was represented on the first day of the concurrent engineering training
program, including the managers and vice presidents. On that day examples of operating
principles of many successful companies were discussed. Filmed guidance and
testimonials were shown as to what was necessary to practice teamwork and to engineer
products with concurrent input from all concerned parties, especially the customer. After
the first day of training, I was summoned to a high-level managerial meeting where I was
ordered to stop showing what others had defined as concurrent engineering and to
henceforth simply tell everyone in the training sessions that their company was already
practicing a perfect model of concurrent engineering. The premise was that since they
weren't willing to change to the correct model of conduct, the model had to be changed. I
declined, and eventually dragged them kicking and screaming all the way over to the
model of concurrent engineering. They tried it. It was successful, and so were they, but
they hated me for it (at least in the beginning). But that's okay. I think they just needed
someone to vent their anger on, and better me than to vent it on one another.

 Page 321

13
Inspection Planning
13.1 Dimensional Inspection Planning
The concept of a dimensional inspection process designer creating a dimensional
inspection plan is recommended for the verification of most, if not all, product designs.
Times when design needs for part functionality are abandoned because of inferior, ill-
conceived inspection plans and a lack of knowledge of equipment potential belong in the
past. This unit leads one through a process for the creation of a dimensional inspection
plan. It shows some of the items worthy of consideration when writing the step-by-step
process a part must go through to assess tolerances and cosmetic compliance and assure
functionality. It explores the uncertainties common to the inspection processes and what
uncertainties are and are not permissible. It raises the point of what to do with
information gained from the inspection process to improve the manufacturing procedure
and quality of parts produced from that time on. A step-by-step example is given for
several possible dimensional inspection plans for parts to be produced.
For each workpiece design, the measurement process designer may prepare a dimensional
inspection plan. The plan should include a list of measurements to be made, what gage to
use for each measurement, the procedure for each measurement, and the gaging limits for
each measurement. It is recommended that the measurement process designer document
logic for the plan. The dimensional inspection plan may be developed in the following
steps:
1. For every part to be inspected:
a. Learn how the part functions.
b. Decide which dimensions will be inspected and which dimensions will not be
inspected.

 Page 322
c. Determine the ramifications of approving a dimension that is not within its tolerance
and of rejecting a dimension that is within its tolerance.
2. For every dimension that is to be put under test:
a. Determine whether an inspection plan already exists that will suffice if minor
modifications are introduced. If one exists, tailor it to the new part as needed, and go to
Step 2(e). Otherwise,
b. Find out how the part is produced and what errors of geometric perfection are
common to that manufacturing procedure.
c. Determine the best inspection approach/scheme to follow.
Decide the most appropriate tools, gages, and major equipment needed
d. Discover which measurement uncertainties will be introduced.
Does your measurement approach/scheme contain inherent uncertainties?
Do the gages and/or inspection equipment and/or inspector knowledge have
uncertainties that need to be considered?
Does the environment within which the part will be inspected have the capability of
introducing uncertainty?
e. Decide on the acceptable probabilities of accepting a bad feature or rejecting a good
feature.
f. Analyze the most likely distribution of measurement data.
g. Calculate/analyze the gaging limits.
3. Create a dimensional inspection plan for every part design.
The Dimensional Inspection Plan Format
A dimensional inspection plan should be created for each part design. Among the items to
be considered are:
1. Which characteristics of which part features will be measured.
2. The tools needed for the measurement of each feature characteristic.
3. When it is in the best interest of the feature, the position and number of points on the
surfaces where sample data will be taken may be specified.
4. The steps to be followed in the measurement procedure.
5. How collected data will be analyzed.
6. What to do with the collected data to improve the manufacturing process.
The inspection planner may consider it wise to document the reasoning behind each
decision in the plan. This may explain to those scrutinizing the plan the logic behind the
decisions made by the planner.

 Page 323
Plan Development
To determine the best possible plan for the inspection of any part, the designer of the
measurement process should know:
1. How the part functions.
2. Which characteristics of which features must be inspected to insure the part's
functionality.
He or she will also need to determine how lots will be measured and what type of
sampling is necessary to insure optimization of inspection time and collection of the most
valuable data. Whether the information is used merely to insure the functionality of that
one part or is to be used as variable data for the continued production of parts within
geometric tolerances, the inspection plan is a vital piece of the process. It can help insure
functionality, interchangeability, and a product produced at the lowest possible cost.
Each Measurement
The portion of the dimensional inspection plan that deals with how each measurement
should be taken should consider:
1. Has a previous plan been developed either for this part or for this type of part? If so, it
may be used as a guide to rewrite the plan, or to create the plan.
2. Should the feature be measured? Sometimes measurement of a feature is not necessary.
For example:
a. If the dimension is controlled by a proven die or mold.
b. If attribute can be verified by a means other than measurement. For example: strength
vs. diameter.
c. If dimension is unimportant to workpiece function and is known to be not a factor
worth consideration in the batch or part under test.
d. If a feature can be either accepted or rejected at a lower cost by simply seeing if it will
fit into the assembly, it is sometimes better to do so (for example, if interchangeability is
not a factor).
What is Being Verified?
In the case of geometric controls applied per the ASME Y14.5 Standard, if the feature is
to be inspected, verification of either boundaries or tolerances zones is required. When
the designer specifies a positional maximum material condition (MMC) control, for
example, the surface of the feature may not lie outside a boundary of virtual condition or
else it often will infringe on space supposed to be occupied by the surface of the mating
feature. Position of the virtual condition

 Page 324
boundary may be fixed or movable (as in the case of features controlled to datum features
of size at MMC).
In some controls, line elements are to reside within tolerance zones. In others,
centerplanes, surfaces, axes, or points must reside within the specified tolerance zone. In
most instances where a centerplane or axis is being controlled, verification of either the
tolerance zone or the virtual condition boundary it generates is acceptable. Verification of
both is rarely necessary. Although it is generally understood that the verification of
tolerance zones or boundaries is appropriate, and considered to be roughly equivalent,
these two concepts are not always mathematically equivalent. Where there is considered
to be a conflict between the two concepts for verification, the virtual condition boundary
concept is given greater weight. It is considered that the MMC concept virtual condition
boundary is usually more descriptive of the space needed for assembly in mating
situations.
The intent of the designer should always be clear and interpretable per the ASME Y14.5
Standard. Where intent is not clear, the designer should be contacted for clarification
whenever possible.
Hard vs. Soft Gages
To verify feature compliance with size or geometric tolerance, we may use either hard or
soft gages. Hard gages are mechanical, such as, for example, coordinate measuring
machines (CMM), micrometers, vernier calipers, ring gages, and snap gages. Depending
on the situations, these gages are capable of taking either direct measurements or
comparative (to a standard) measurements.
Hard gages are often used to determine coordinates of a set of points on a feature surface
to estimate the true shape of the feature surface. This information is sometimes then
compared to mathematical soft gages in computer software. The comparison can be either
direct (measured values to the ideal values) or comparative (measured quantity
determined to be either larger than, equal to, or less than the standard). If the information
fed to the computers involved in this method is correct, soft gages can perform
verification techniques for geometric controls quite well. Remember, however, that the
verification is only as good as the procedures followed during the collection of data. Hard
gage measurements can be compared directly with gaging limits to determine acceptance
or rejection.
Soft gages can use a knowledge of the manufacturing process for features to be able to
augment data that is represented by the part configuration. For example, with regard to
deviations that generate smooth profiles, these deviations from geometric perfection may
be caused by thermal bending of the machine or workpiece during machining, clamping
distortion, or bad fixturing. Deviations that tend to form a trend like a periodic pattern
over the workpiece may be

 Page 325
caused by machine tool feed, tool form errorsas in rolling componentstool stiffness,
cutter form errors, or cutter alignment problems.
Dividing these errors from one another and analyzing their causes can allow one to take
action that will prevent them from recurring.
Choosing Gages
The choice of appropriate gages should be based on: (a) capability (b) availability, and (c)
cost-effectiveness.
We would all like to check our parts in the most accurate manner available without
spending more inspector time and machine time than the part can warrant. For this
purpose, the dimensional inspection plan designer should know:
1. How many parts are being made from this drawing?
a. In this run.
b. In future runs.
2. What gages do I have available from other inspection procedures that could be used?
3. How long will set-ups take using the various gages available?
4. How valuable is the machine time needed to inspect this part in relation to its
importance?
5. What is the cost of investment of acquiring the use of gages not currently available?
6. Is training of personnel a factor in the implementation of this inspection plan?
7. Is the environment required available for use during inspection?
Determining Uncertainty
1. Uncertainty of the measurement plan
Has every piece of important information concerning this part and its measurement been
used to create the plan?
2. Uncertainty of the gages
What is the quality of the gages that have been chosen?
What is the repeatability and the accuracy of the gages to be used?
Are there software errors?
With the advent of computer analysis of collected information about part features, we
were faced with telling the computerized mechanisms how to collect data, how much data
to collect, and what to do with the data that was collected. Sometimes this was done by
those with skills in one area of expertise, such as computer programming and/or
mathematics, but who were not knowledgeable in the guidelines that existed in the
standards for what was being inspected for (or what uncertainties were introduced by
real-world mechanisms

 Page 326
and environments), not to mention operator error. Standards are now being written
worldwide to combine the knowledge of those involved in all aspects of the work
environment.
3. Environmental uncertainty
a. Temperature. All part dimensions and tolerances apply at a temperature of 20C (68F).
If both the gage and the workpiece are at 20C there is no measurement uncertainty
caused by temperature. For other conditions the effects of thermal expansion on the gage
and the workpiece must be considered.
Should the gage and the workpiece be at the same temperature that is other than 20C,
both will expand by an amount that can be calculated by the formula: KL (T-20),
where:
K = coefficient of expansion
L = length
T = temperature in degrees Centigrade
With the same coefficients of expansion, no measurement uncertainty caused by
temperature is introduced. With different coefficients of expansion, temperature-related
measurement uncertainty is introduced. If the gage or the workpiece is constructed of
more than one component and these components have different coefficients of
expansion, the structure should be examined to see whether an additional uncertainty
could occur because of twist or bend. Among the many other factors to consider are
the following:
b. Slowly changing temperature. Should the air temperature slowly change over time, and
thermal conductivities that are in the structures of the gage and the workpiece happen to
be high, uncertainty caused by temperature can be scrutinized based on the premise that
the temperatures of the gage and the workpiece are uniform, but not equal.
c. Quickly changing temperature. If minor, fast air temperature changes take place and if
the gage and the workpiece are of large mass, the effect of the temperature changes could
be small. In these situations, not enough heat flows in and out of the gage and the
workpiece to change the temperature significantly. Big, fast air temperature fluctuations
may impose temperature changes on the gage and the workpiece that can cause them to
twist and bend. The uncertainty of this must be taken into consideration and, if possible,
avoided.
d. Other effects on temperature can be caused by:

 Page 327
Radiant energy such as sunlight and artificial lighting. Sunlight should be avoided.
Artificial lighting and radiant energy outside the visible spectrum should be brought to
a minimum. Indirect lighting is often effective. Lighting should be as uniform as
possible to prevent uneven heating of gage and workpiece. One of the most prevalent
problems caused by radiant energy is in the flatness of large surface plates.
Workpieces not stabilized to the inspection environment.
Air from heating or cooling ducts.
Handling.
4. Other environmental uncertainty effects
a. Vibration. The characteristic of vibration that causes uncertainty is the relative motion
between the gage sensing or contacting element and the measured point on the workpiece.
b. Magnetic fields
c. Gravity
d. Acoustic noise
e. Operator skill
f. Electronic drift
A Dimensional Inspection Plan Example
The set of measured values made on the workpiece constitutes the data upon which
calculations are carried out to determine position, size, and departure from nominal
(desired geometric) form. Data collection may be under manual or automatic control. The
use of too few data points or data points inappropriately distributed may provide an
unreliable, nonrepresentative description of the actual feature being considered.
Also to be given consideration is the knowledge and experience of the inspector. Whether
operating a CMM or collecting data on a surface plate (or any other) situation, the
inspector may know the likely deformations of a particular workpiece. The measurement
procedure should take into account that existing user knowledge. (See Fig. 13-1.)
An Example of a Dimensional Inspection Plan
Run of 20,000 Parts
One hundred parts per shift, 2 shifts per day, 5 days per week over 20 weeks.
Ten parts per shift will be inspected (10% of production per shift) for variable data
collection.

 Page 328

FIGURE 13-1
One part per shift will be paper gaged.
All other parts will be inspected for functionality by use of a positional functional gage
for four-hole pattern's feature control frame.
Brief Manufacturing Procedure Summary for Inspector Information
1. Part is Blanchard ground top and bottom for thickness.
2. Four holes are drilled on a four-spindle drill press.
3. Outside diameter is turned on a CNC lathe, while fixtured to hole pattern.
Brief Functional Description of This Part
This is Part 2 of a two-part assembly. The mating part consists of a cylindrical cavity with
four pins inside. The parts assemble and do not spin. This subassembly is internal and
will not be viewed on the finished product (a helicopter blade assembly).
Features to be Measured
1. For size:
Step 1 2.9602.990
Step 2 4 .570.590
2. For geometric tolerance:
Step 3 Perpendicularity to within a diameter of .004 at MMC.
Step 4 Position to within a diameter of .020 at MMC to L and D at MMC.

 Page 329
Gages Needed
1. For size:
One 23-in. external micrometer for LMC verification.
One 2.990 diameter GO gage for external envelope check.
One 01-in. external micrometer for LMC verification with one set of telescoping gages for
.570.590 holes (small-hole gages may be substituted).
One .570 diameter GO gage pin for internal envelope check.
2. For geometric tolerance:
One surface plate.
One height gage.
One indexing head with inspection surface plate chuck, 90 indexing capability, and
magnetic face plate attachment.
One gage pin set.0005 increments between .5700 and .5900.
One dial indicator (nonplunger type), 1-in.-long spindle, ball not to exceed .100.
One functional gage for positional control verification of the four-hole pattern and the
perpendicularity of D to L.
Graph paper and vellum for paper gaging.
Measurement Procedural Steps
1. Use micrometer to verify 2.960. Record sizes and recommendations for pass, fail,
rework.
2. Use GO gage to verify 2.990. Record pass/fail.
3. Use micrometer and telescoping gages to verify .590. Record sizes and
recommendations for pass, fail, rework.
4. Use GO gage to verify .570. Record pass/fail.
5. Use indexing head with magnetic face plate, height gage, and dial indicator to verify
perpendicularity tolerance. Take differential measurements.
a. Chart results and recommendations.
b. Replace magnetic face plate with chucking device. Chuck up on datum feature D while
perpendicular to datum L. Record inside diameter of chucking device for use as pattern
shift tolerance if needed.
6. Use chucking device, pin gages, height gage, and dial indicator to verify position of
four-hole pattern.
a. Mount part in chucking device:
Up against datum feature L, with a minimum of 3 high point contact.
Grabbing datum feature D, minimum circumscribed cylinder perpendicular to L.

 Page 330
b. Put gage pins in holes, maximum inscribed cylinders.
c. Use one hole (gage pin) for angular orientation (if hole pattern does not check within
tolerance in this orientation, other orientations may be tried).
d. Mount dial indicator to vernier height gage and set both (along with chuck) onto
surface plate.
e. Zero indicator out at height of chucking device's axis and record vernier scale reading.
f. Go to top of gage pin from which orientation has been taken and find same indicator
zero while placing indicator as close to the surface of the part as possible. Record vernier
scale reading.
g. Subtract vernier scale readings, and from the difference also subtract one-half the
diameter of the gage pin.
h. Record the difference between the .800 desired location and the actual location from
center of chucking device to center of hole (this hole will only be out of position on the
Y-axis since it has been used for orientation).
i. While in this orientation, all other hole deviations should be inspected for deviations
from the desired height of each from the chucking device's axis and deviations from
perfect recorded.
j. The chucking device should then be rotated 90 and deviations of each hole's axis from
the desired location recorded by following the same procedures as were used in the first
orientation (again, the first hole will not be out at all in this second orientation unless the
chucking device has created error while rotating).
k. Each hole may be validated as within tolerance by using the formula of 2 times the
square root of the sum of the squares of the deviations on the X-and Y-axes.
l. If the answer to step k is less than or equal to the .020 positional tolerance in each case,
the holes have acceptable axial position at the top where inspected. Position must be
verified in a similar manner along the axis of each hole. Access to the hole may be made
through direct contact with the dial indicator if it is known that the hole is cylindrical to
such an extent as to not affect the validity of data collected. If not, differential
measurements may be taken to give more valid data.
If holes check to be outside of the original .020 diameter positional tolerance, then each
hole should be considered for individual bonus tolerance drawn one-for-one as much
as that hole has grown from .570 without exceeding its size limits. This bonus
 tolerance will be determined by the difference in size of the largest gage pin the hole
will accept and the .570 MMC.

 Page 331
If all holes are still not within their allowed positional tolerances, pattern shift may be
considered. It is the deviation of the minimum circumscribed cylinder of datum feature
D that is perpendicular to datum plane L (the difference between the inside diameter of
the chucking device recorded in step 5b and the 2.994 virtual condition of datum
feature D. To use this pattern shift for part acceptance, it is recommended that paper
gaging be used.
13.2 Methodical Measurement
Measurements should be performed in a methodical, well-planned manner to determine
whether or not a part will function. Just as the best dimensioning and tolerancing
approach is one that builds from size and form controls to interrelating datum features to
positioning and, perhaps, profiling critical part features, measurement procedures should
mimic this natural progression. It should verify size requirements have been met, then
that the first primary datum feature is well formed, then that the interrelationship
tolerances between datum features have been complied with, then go on to the
verification of tolerances for features or patterns of features held to the datum reference
frame generated by those features.
Then, if other datum reference frames exist, the interrelationship between the datum
reference frames should be inspected and the process begun again. Planning is the key.
Chaos has no place in the measurement of parts, just as it has no place in the
dimensioning and tolerancing of parts.
ApplicationsChapter Examples
GROUP EXERCISE: Recognizing that (as with most real-world design drawings) the
design drawing in Figure 13-2 is not perfect in its use of dimensions and tolerances:
1. Interpret the drawing and discuss its meaning.
2. Discuss how each dimension and tolerance will be verified or otherwise accepted.
3. Discuss the measurement strategy in terms of gaging and procedural uncertainty:

b) Gage
a) Strategy uncertainty
uncertainty
d) Equipment
c) Environmental uncertainty
uncertainty
e) Operator (inspector) f) Permissible
error/knowledge probabilities
g) Recommeded design change
proposals (if any)

 Page 332

FIGURE 13-2
4. Summarize discussion in notes for presentation to other groups and for a comparison
in findings.
Answers to Questions 14:
Discussion question to allow completion of inspection plan.
5. List the steps required for the inspection procedure needed to verify or reject the part in
Fig. 13-2.
Answer to Question 5:
STEPS
a) Verify all size requirements for parallel planar surfaces with a micrometer for LMC and
a height gage and surface plate for MMC envelopes.

 Page 333
b) Verify all hole LMC requirements with appropriate small hole gages and/or telescoping
gages and a micrometer. Verify all MMC hole requirements with gage pins of either a
.375, a .406, or a .640 as appropriate.
c) Use gage pin set ranging from .375 to .690 to determine hole bonus tolerances. Gage
pins should run in increments of .001.
d) Use dial indicator to determine datum feature C compliance with perpendicularity
control. Use protractor to determine all other angular tolerance requirements from general
tolerance note.
e) Use dedicated fixture or equivalent angle blocks to establish datum reference frame A,
B and C for .640.690 hole.
f) Use CMM to measure hole location with touch probe. Touch probe must probe fixture
to establish datum reference frame. Do not bring touch probe into direct contact with
datum feature A, B, or C.
g) Position refinement to A is just a perpendicularity control, since only one hole exists in
this .640.690 pattern. Verify compliance.
h) Use dedicated fixture to establish datum reference frame A, D , C for the .406.475
holes and the .375.418 hole as a simultaneous gaging requirement. Once fixtured, use
the CMM to establish hole positional tolerance compliance.
i) All other measurements may be verified using a vernier caliper.
TOOLS NEEDED:
a) 6-in. vernier caliper
b) Set of micrometers ranging from 0 to 4 in.
c) 6-in. or taller height gage
d) Surface plate
e) Small hole gages from .375 to .690 and/or telescoping gages
f) Gage pin set in increments of .001, ranging from .375.475 to .640.690
g) Two dedicated fixtures to establish the two datum reference frames listed
h) A CMM with suitable touch probes
i) One small-angle plate to use when inspecting the perpendicularity of datum feature C
j) One dial indicator and stand
k) One protractor
6. List the steps required for the inspection procedure needed to verify or reject the part in
Fig. 13-3.

 Page 334

FIGURE 13-3
Answer to Question 6:
The inspection plan (Note: Many features on this part either will not be inspected or will
be rough checked only, as follows.)
STEPS
a) Use a vernier caliper to measure the .375, .490, and .750 lengths and MMC and LMC of
the .090, the .182 LMC, and the .056 and .312 depths.
b) Measure the .13832 UNC-2A threaded shaft for thread specifications with a threaded
hole gage for attribute information only. This gage should be equipped with a coaxial
feature that slides over datum feature D with an inside diameter of .194 and a feature
that seats on datum feature C with primary contact. It should also have a feature that
verifies the .11240 UNC-2B hole simultaneously for size and position to datums and
interrelationship to the .13832 threaded shaft.
c) The flatness of datum feature C need not be inspected.
d) The position of the hexagonal feature need not be inspected.

 Page 335
e) The runout control may be inspected with a V-Block and dial indicator.
f) Any other unmentioned feature tolerance need not be inspected.
Inspecting Parts with Profile, Position
Inspection planningcreating the procedure
7. This part may be destroyed during inspection (Fig. 13-4). How would you inspect it?
This involves a production run of 100,000 parts. List the inspection plan steps.
Answer to Question 7:
STEPSSizes
a) Use a micrometer to verify .9 LMC of part thickness.
b) Lay part on surface plate. Zero dial indicator on stand to a stack of gage blocks 1.1
high. Run indicator over part to verify 1.1 MMC envelope has not been violated.

FIGURE 13-4

 Page 336
c) Use small hole gage and micrometer to verify .388 LMC of all holes.
d) Use .380 gage pin to verify compliance of each hole with MMC envelope.
STEPSLocation
e) Find largest gage pin that can be fit into full depth of each hole. This will be used for
positional bonus tolerance if part is inspected with dial indicator and height gage.
f) Set datum feature A against an angle plate while pushing the two B datum targets
against a surface plate and target C against a mutually perpendicular angle plate. Clamp
part.
g) Use dial indicator and height gage and, when possible, gage pins to inspect all hole
locations from datum feature simulator B. Record readings.
h) Rotate angle plate 90 to bring datum target C into contact with surface plate. Inspect
each hole location from C. Record readings.
i) Use formula and or Table 8-1 from page 224 to determine actual deviation from true
position.
j) Add bonus tolerance to .010 original position tolerance for each hole to get total
allowed deviation from true position.
k) Compare allowed to actual deviation from true position for compliance with geometric
control.
l) Break down positional set-up.
m) Put part on optical comparator staging table onto datum feature A.
n) Use dedicated optical comparator chart with profile tolerance zone for this part to
determine compliance with profile zone. Part may be moved or rotated as required to fit
zone as long as appropriate three-point high point contact is maintained on datum feature
A.
Note: No need to destroy part.
Major uncertainty: Because part is thick, perpendicularity of profile surface elements to
datum A is in question and may require further evaluation with dial indicator.
TOOLS NEEDED:
a) Micrometer 01 in.
b) Surface plate
c) Dial indicator and stand
d) Gage block stack of 1.1 in.
e) Small hole gage for .380.388 holes
f) Gage pins ranging from .380 to .388 (a minimum of 1.1 in. long) at .001 increments

 Page 337
g) Height gage
h) Two angle plates or equivalent with enough surface area to cover datum features
i) Two C-clamps
j) Optical comparator
k) Dedicated optical comparator chart with profile tolerance zone Represented
Functional Gaging vs. Variable Data Collection
8. If the concept of verification of a geometric control that utilizes the MMC symbol is
conceptualized by a functional receiver type gage made at virtual condition and only
thought of in theoretical terms (without gage tolerances or cost factors), it serves a useful
purpose in helping us understand the design requirements. Why, then, does a real
functional gage sometimes not fulfill our inspection needs?
Answer:
A receiver gage will not usually collect enough information to help control a
manufacturing process. It commonly only gives good vs. bad (attribute type) data.
9. With the introduction of the CMM, a computer was used to analyze points taken on the
part and to recreate simulated reconstructions of the actual surface. Its ability to reorient
this part simulation without actually moving the real part and to compare the geometry of
the simulation to the desired part geometry would seem like the answer to all our
inspection needs. Does it ever let us down? If so, how and why?
Answer:
It can let us down through an insufficient collection of points on a feature or datum
feature to accurately represent a surface. It can also let us down through improper
programming. Collected data can be improperly processed to give a false impression of a
feature's tolerance compliance.
10. Optical comparators are capable of taking the actual surface profile, and through the
use of staging tables, light reflections, mirrors, lenses, and screens merging its image with
an etching overlay of the desired part geometry. Does it ever give a less than perfect
analysis of the compliance of a part's geometry with stated tolerances? If so, how and
why?
Answer:
Yes. The image of a feature is dependent on the way the light or camera looks at the part.
If shining in, or looking in at only one angle (view), it can miss surface pits or tapers that
depart from the viewing angle.
11. Surface plate type set-ups with angle plates, height gages, dial indica-

 Page 338
tors, chucking devices, dividing heads, bench centers, parallels, veeblocks, sine plates,
and the hundreds of other augmenting tools available have served us for many decades.
Why would we ever need to use anything else?
Answer:
These techniques are slow.
Statistical Process Control and the Inspection Process
12. What types of inspection procedures are most compatible with the SPC process and
why?
Answer:
Inspection procedures that collect as-measured dimensions (variables data) that can be
charted are more compatible than attribute gages that give no numbers to chart.
13. Which types of inspection procedures are least compatible with the SPC process and
why? Also, can these conflicts be minimized somehow? If so, how?
Answer:
Functional and GO type receiver gages are least compatible, but the conflicts can be
minimized if these gages are computerized as soft functional and GO gages. This would
involve a collection of measured distances and sizes that could be charted and also
merged with a computer's soft functional gage.
Inspection of Parts with Plane Datum Features, Datum Features of Size at MMC, and
Datum Features of Size RFS
14. If inspected with a touch probe on a CMM, would Design 1 (Fig. 13-5) benefit from a
fixture to establish the datums or would probing the datum features be as reliable?
Answer:
A fixture would be beneficial in establishing the datums. By probing the highly accurate
planar fixture surfaces that seat against the datum features, one would have a better
chance of establishing high point planes. By probing the datum features directly, the
chances of probing the high points are remote. See answer to question 16.
15. Could a CMM establish centerplane datums B and C in Design 2 (Fig. 13-6) by direct
probing of the surface?
Answer:
Yes, but a lower degree of uncertainty will occur if many points are probed and the
software is capable of using the highest sets of oppos-

 Page 339
ing points to create the centerplane datums. However, probing the properly sequenced
sets of planar parallel and perpendicular fixturing rails that are in contact with datum
features B and C and are also perpendicular to plane A, as probed from the CMM
table, would probably be more reliable and repeatable.
16. Describe the simplest method (inspection plan) you can think of to inspect each of
these three positional controls from Design 1, Design 2, and Design 3 (Figs. 13-5 to 13-7).

FIGURE 13-5

FIGURE 13-6

 Page 340

FIGURE 13-7

FIGURE 13-8
Answer:
a) Design 1: A functional gage consisting of three sides to represent the datum features
and a .240 gage pin located the specific basic distance from B and C and perpendicular
to A. If an open set-up is used, consider the one below.
Alternate open set-up for inspection of Design 1 to be used with a CMM. See Figure 13-8.

 Page 341
b) Design 2: A functional gage consisting of five sides and a .240 gage pin. The gage pin
would be centered to B and C and perpendicular to A. Datum feature B is simulated at
2.990 and C at 1.880.
c) Design 3: A CMM that probes enough points on the datum features to find centerplanes
B and C which are perpendicular to A and each other. The CMM then probes the hole at
many points at many levels to determine if the hole's axis is within tolerance. As
described in question 14, probing fixturing planar rails that contact the datum features
would, if highly accurate within fixture and gage maker's tolerances, be more reliable and
repeatable.

 Page 342

 Page 343

14
Sources of Measurement Error
14.1 What We Know is Often Untrue and Other Sources of Measurement Error
Consider the inspector who wants to do a good job, but also seeks a bit of comfort in his
life. He oversees the construction of an environmentally controlled room in which to
store his inspection equipment and to conduct his measurement procedures.
He makes certain the room temperature is set at a constant 68F/20C. The first time the
thermostat is mounted, he stares at it proudly, then walks the 30 or 40 ft to his CMM and
places the first of a series of workpieces to be measured onto the machine table. He
pauses for a moment to turn his face toward the window and bathe in the warm sunlight,
then opens his eyes to admire the view of the nearby electrical power plant. Soon he
notices the air conditioning vent overhead, which emits a soft blast of cold air that
eventually reaches the thermostat and settles the room temperature at the appropriate
number of degrees.
He feels the floor vibrate with the passing of a forklift just outside the confines of the
room and smiles at his isolation from all that. The sound of presses pounding out parts is
muffled by the walls of his room and even the grinding of burrs off of sheet metal is faint
although the sparks are bouncing against the window of his door. The grinding dust is so
thick outside this entrance that it engulfs his friend as he enters to admire the new set-up.
The inspector beams with pride as he laughs and nods in agreement, and places his hand
on the part to be measured, leaning his weight and warmth into it.
Eventually the friend leaves and the inspector begins the measurement. He probes a
minimum number of points on the part's surfaces to establish the datum

 Page 344
reference frame, the three perpendicular planes the computer will create and that will act
as his origins of measurement. He inadvertently probes some low as well as high points.
The computer gives equal weight to each in establishing the datums. If, on the second
part he measures, he decides to probe more points to establish his datum planes, the
computer algorithms in the program he's using will establish an average best fit plane
from all the points probes. These are not the high points of the surface most likely to seat
in the assembly and orient the part. If he had taken the time to fixture the part and probe
the fixture or parallels and angle plates that would naturally rest against the highest points
on the surface, his results would be more representative of the actual feature locations as
described in the ASME, ANSI-approved Y14.5 Standard on dimensioning and
tolerancing, but the person who sold him the machine told him the computer negated the
need for fixturing.
He finishes his measurements of this part in record time and decides to use the time he
has saved to take a coffee break.
Maybe he should have taken the coffee break instead of measuring the part. Each act has
the same amount of value added to the product. Or perhaps the coffee break, having less
negative value, is the one that, in the future, should take the higher priority.
We all want to believe that we are hard workers and competent in what we do. Sometimes
the reality of what we think we know just doesn't happen to be true or even remotely
correct. But don't tell this inspector. He might take it as a personal affront. After all, to
him, it seems as though you're trying to negate his knowledge and skills, a major part of
his existence. His ignorance is blissful.
The sad truth is that messengers of what could be liberating knowledge are often killed.
Study your history!
Still, the problem with doing nothing to improve these errors is that it eventually shows
up in product quality. Let's face it, the job of the inspector is to validate whether or not a
feature meets its requirements and contributes to the overall part functionality. And if the
measurement errors are, as I have described here, legion, how can we place even the
remotest confidence in the findings?
What are the sources of his measurement uncertainty?
14.2 CMM Error Compensation
If a coordinate measurement machine (CMM) is very repeatable, but not accurate, it is
possible to compensate for the inaccuracy. One method is to identify the sources of the
problem and physically eliminate them. This method is known as reduction. Another
approach is to mathematically eliminate the errors. This method is known as error
compensation. The following is a summary of a study and its subsequent report
conducted by Electronic Measuring Devices, Inc.
Computers can help compensate for the errors, but first we must know their

 Page 345
sources. CMM technicians are often tasked to analyze individual sources of error. This
treats each source as though it is a separate entity that is found and eliminated, which then
allows the technician to move on to the next source of error. Again, it is found and
eliminated and the next source soughtand on, and on.
Other methods of error compensation make no attempt to separate these error sources,
but rather make the assumption that these system errors are part of an overall problem
that can be dealt with as though it were one problem. This method deals with the error
sources in combination and bases its algorithms on this assumption. It uses the smallest
number of artifacts to find the errors. It refuses to measure angles, except for 90-degree
angles for a preliminary evaluation of perpendicularity. This system is meant for the
CMM user and may be utilized at any time. The user must have the appropriate software
to perform the procedure. If the equipment consists of granite and air bearings, the
smoothness of mobility may be used as an advantage, because periodic errors in the
movement of components to each other will not be a factor. This advantage can also
assume that measurements taken in, for example, 1-in. increments, are proportionally
reliable to each other. If one finds all the displacement errors existing on the perimeters of
the measuring volume, then all errors inside the measuring volume are captured in
combination. Since the recommended measurement error is usually no more than 10%,
the error compensation should reduce the error to that amount or less. To detect part
errors at a specific three-dimensional location inside the measuring volume, calculate the
proportions of the machine error that are at that spot in all three directions from the
machine origin. Assume at this origin that all errors are zero. All calculations are
controlled by the software. All measured points taken are fully corrected.
For example, to begin the process a National Institute of Standards and Technology-
certified Moore lug bar is used to measure each CMM axis's full travel, in both transverse
directions. Analog scanning is recommended. At least three runs are taken to assure a
good set of averages. The runs are made in the center of the measuring volume for each
axis. Any error in the bar is taken as acceptable, allowing the CMM's straightness to be
compared to a perfect geometric counterpart of the bar.
When an average one has confidence in is secured, the software automatically applies
mathematical correction values, to size and location to correct straightness errors. All
corrections may be checked through more scanning of the Moore lug bar. The errors are
corrected within the repeatability of the CMM. The physical error still exists, but is not a
consideration in measurement accuracy.
Since straightness has been corrected, perpendicularity/squareness must now be
corrected. The Moore bar is used, now measuring distances. The bar is angled at 45
degrees to an axis in the center of the measuring volume. The distance between the
furthest lugs that can be reached with the probe head is mea-

 Page 346
sured. This is repeated a minimum of three times to establish an average that is reliable.
The bar is then repositioned to the other 45-degree position and the same two lugs
measured again. When the readings are the same, the two axes are square to an acceptable
accuracy.
This process is performed between all of the axes and the deviation entered into the
software. Perpendicularity corrections are calculated and applied in all subsequent
product measurements. This occurs in real time as if all of the axes were perpendicular to
each other.
The last error compensation step is to measure the displacement errors that are found on
each of the 12 edges of the parallelepiped that is the measuring volume. There are four
positions at right angles to each other at the level of the worktable, four similar positions
a distance above the table (for example, 1 ft above), and four vertical positions
connecting their corners. The software determines the distances between all lugs in each
set-up. A minimum of three runs is recommended to establish a reliable average.
The software is fed the correction values that were discovered when the Moore bar was
certified. These are taken away from the observed errors discovered while the
compensation procedure was being conducted.
The data collected in 1-in. increments all around the measuring volume allow a grid of
data points to be created that no longer has the bar errors.
Three displacement data sets that intersect the corner that has been defined as the machine
zero are chosen. They are referenced by all of the others. The data collected in the X-axis
has corrected at this location, so no errors exist within the 12-in. measured length. On the
opposite side of the cube the results of the measurement in the same X direction will be
different. These differences are the errors in X that are caused by the geometric
aberrations within the system. The source of these errors is ignored. Their total amount is
all that matters.
The premise is that all errors existing in X are proportional to the distance between the
two sets of readings. Still, the Y-axis readings are necessary to create the coordinate
position of the workpiece. This is important to be able to discern what part of the total X
error to give to it. Consequently, the same procedures are performed in the Y as were
performed in the X direction.
After these procedures are completed, the grid of data is created on 1-in. centers at the
height of the CMM table. And the amount of errors that exist from one side to the other
can be established in both the X and Y directions. If the process is performed up the Z
direction, a 3-D grid of data points is created. The assigning of errors in 3-D can then be
accomplished.
The intersection of the measurement planes (the corner of the X, Y, and Z) is used as the
reference for all corrected readings. These corrected readings are proportional to their
location inside of the measurement volume.
This approach should compensate for errors of measurement and reduce the error in that
way approximately 9095%.

 Page 347

14.3 Certification vs. Calibration

Calibration is the process of checking the dimensions and tolerances of a gage, or the
accuracy of a measurement instrument by comparing it to a like instrument that has been
certified as a standard of known accuracy. The person or team performing the calibration
is then to detect and adjust any discrepancies in the gage or instrument's accuracy to bring
it within acceptable limits.
Accuracy is how close measurements taken are to the true feature measurements. It is
commonly defined by the factors of systemic error, also known as bias, and precision,
also known as reproducibility. (a) Systemic error can be defined as the difference
between the average of a series of measurements and the true measurement. (b) Precision
is the degree of repeatability of the measurement process. It is how close derived values
are to identical when measuring the same characteristic of a feature time after time.
Reliability of the measurement of dimensions is often seen as having random errors.
These errors can be due to human factors such as not reading scales appropriately, poor
set-up of parts, incomplete or inaccurate knowledge of a technical requirement, mistakes
in recording readings, and many others.
Calibration is done over time, according to the usage of the gage or instrument and the
materials of the part and the gage. The dimensions and tolerances of the instrument or
gage are inspected to determine whether it has departed from the previously accepted
certified condition. If it has deteriorated to a point that it does not meet the requirements,
it can be downgraded and used as a rough check, or it can be reworked, or it can be
scrapped. If it is reworked, it must be recertified.
If a gage is in frequent use, it will require more maintenance and more frequent
calibration.
Certification is performed prior to the gage's first use and later to reverify that if it has
been reworked, it again meets its requirements. Most often certification is given by a
comparison to a reference standard whose calibration is traceable to an accepted
organization such as the National Institute of Standards and Technology. It is
recommended that such reference standards have been certified and calibrated as masters
not more than 6 months prior to use.
ApplicationsChapter Examples
Coordinate Measuring Machines
1. If a CMM uses a ball probe to establish the axis of a cylindrical feature from which to
measure, what, if any, kinds of errors may occur?
Answer:
The size of the probe's radius and the number of points and the distribution of probed
points could introduce error.

 Page 348
2. If the software uses the best fit algorithm, do you see that as possibly causing problems
with verification? If so, what?
Answer:
It ignores the high point plane or axis and could accept parts that are nonfunctional with
high points that were ignored, causing assembly interference. It could also reject parts that
are technically good and functional
3. Can you think of a better algorithm?
Answer:
A better algorithm would be one that forms planes and axes from the high points probed.
4. If a probe is used on a surface to establish a plane from which to measure:
a. what are the possibilities of error?
b. Do you see these errors as more or less serious than measurements of features taken
from an axis, established by the best-fit algorithm from a cylindrical feature?
Answer:
a. It could miss the high points.
b. Both have the problem of missing and/or not using high point established datums.
5. If a tapered probe is used in a hole to establish a datum axis from which to measure
other features:
a. What are the possibilities for error for thin parts? for thicker parts?
b. Can these deficiencies be minimized? If so, how?
c. Can knowing the manufacturing procedures and capabilities help minimize the
deficiencies in measurement? How?
d. If the deficiencies in measurement are not minimized and are potentially significant,
what use can be made of the data complied?
Answer:
a. The thicker the part, the worse the possibility of a problem. But, in either case, the axis
is being formed from a 2-D circle, not a 3-D, cylinder.
b. Yes, by not using a tapered hole.
c. Yes; if the processes are known to create certain geometric shapes or deviations from
perfect, one can search for those particular deviations.
d. The data can be used to suggest changes in the procedures of part inspection by
comparing it to inspection data that has been complied more correctly (to point out the
deficiencies).

 Page 349
Verification of Tapped Holes and Datum Targeting
6. The following information has been collected concerning the part produced to comply
with Fig. 14-1. Please state your recommendations concerning compliance and inspection
procedures.
Hole's distance from C= 1.377 Distance from B = .751
This information has been obtained by mounting the part down on datum feature A,
pushing datum feature B against a rail and datum feature C against a perpendicular rail.
An unthreaded gage pin has been put into the hole's minor diameter and the distance from
the datums measured at the top of the hole, as close to the curved surface as possible (that
surface farthest from datum plane A). Please comment on the way this data has been
collected.
Answer:
In place of the procedures listed above, the gage pin should have, instead, been threaded
and screwed in so that it contacts the pitch diameter. Also, the gage pin should have been
checked as it protrudes from the side of the hole as indicated by the chain line of 3.000
MIN. It should have been inspected over the length of the projection, not just at one end
of the projection.

FIGURE 14-1

 Page 351

15
Graphical Inspection Techniques and Examples
15.1 Graphical Techniques and Inspection Reporting
Graphical inspection techniques are an easy way to verify part compliance with
complicated geometric tolerances. They are capable of analyzing data collected in a visual
manner that shows pattern drift and allows correction of manufacturing procedures
before parts in a run violate tolerance zones. They are a cost-effective substitute for
expensive functional gaging and can be used to augment other types of gaging. In this
unit, step-by-step procedures are given to create paper gages and gaging report formats
are suggested for variable data analysis.
Regarding Figure 15-1:
1. Use open set-up to determine the actual location of the hole axes in relation to the
datums.
2. On a piece of graph paper, draw lines or points (as appropriate) to represent datums.
3. Pick true position points to represent the of the controlled features on the graph
paper.
4. Each feature will have four quadrants. Plot the actual position of each feature axis in
the quadrant that represents its position in relation to the datums. (Each line on the graph
paper may represent .001 or .0001 or whatever graduation you deem appropriate.)
5. Measure the size of each controlled feature. If the maximum material condition (MMC)
symbol is used for the feature, calculate the feature's departure from MMC (within size
limits). Add this departure to

 Page 352

FIGURE 15-1
An example of paper gaging.
any tolerance already in the feature control frame. This will give the allowed deviation
from true position.
6. Adjust your compass to reflect one-half that amount using the graph paper graduations
as your scale.
7. Place the compass point at the feature's true position point. Draw a diameter around the
true position point.
8. If the actual plotted axis location of the feature is within that circle, the feature axis is
within tolerance. If not, the feature must be reworked to gain more positional tolerance
(perhaps by enlarging the hole for any internal features) or scrapped. See Tables 15-1, 15-
2, and 15-3 for a complete analysis of Figure 15-1 as produced.
Composite Positional Controls (Checking Feature-to-Feature Requirements)
Regarding Figure 15-2:
First, follow Steps 18 above to check the datum-to-pattern requirement. To check the
feature-to-feature requirement:
1. Place a piece of see-through paper over the graph paper. Transfer the true position of
each feature to the see-through top piece of paper.
2. In Step 5 of the datum-to-pattern check, the size of each feature was measured and the
departure from MMC noted (for features controlled at MMC). Now add this departure to
the tolerance in the lower tier of the feature control frame. Do this for each feature.

 Page 353
TABLE 15-1 Feature Location
FEATURE LOCATION
Feature
XAxis YAxis
Number
SpecifiedActualDeviationSpecifiedActualDeviation
1 1.000 1.001 .001 2.600 2.598 .002
2 2.000 2.001 .001 2.600 2.601 .001
3 3.000 3.002 .002 2.600 2.599 .001
4 1.000 1.002 .002 4.250 4.252 .002
5 2.000 2.001 .001 4.250 4.248 .002
6 3.000 3.003 .003 4.250 4.251 .001

TABLE 15-2 Positional Tolerance and Feature Size

 Page 354
TABLE 15-3 Inspection Remarks

 Page 355

FIGURE 15-2
3. Using the graph paper graduations as a scale, adjust the compass to one-half of the
tolerance derived for one of the features in the preceding step.
4. Take the see-through piece of paper and put the compass point at that feature's true
position. Draw a circle around the true position point.
5. Repeat for each feature.
6. With the see-through paper on top of the graph paper, move the tolerance zones drawn
on the top piece of paper around while the bottom (graph) paper is stationary. If you can
move the top paper so that the plotted actual feature axis points on the graph paper fall
within their respective feature-to-feature (top paper) tolerance zones, the features are
within tolerance. If not, they must be reworked or rejected. (See Figures 15-8 through 15-
11 for examples of these steps.)

 Page 356
Pattern of Features Controlled to a Center Datum Feature of Size
Regarding Figure 15-3:
Extra shift of positional tolerance zones due to a center datum feature of size's departure
from MMC (within size limits) is shown in Figure 15-3 as it applies to the datum-to-
pattern relationship, not to the feature-to-feature relationship. Once you've checked for
any violation of datum-to-pattern requirements, if one or more of the controlled features
lies outside their individual tolerance zones, you can take advantage of an additional
allowed shift of the tolerance zones as a group if the following conditions exist:
In the controlled feature's feature control frame, the datum feature of size is referenced at
MMC (or LMC).
The datum feature of size has departed from its MMC (or LMC) (within size limits).
If these conditions exist, the procedure is as follows:
1. Place a transparent sheet of paper over the graph paper used to plot the actual feature
axis positions.
2. Transfer the actual axis positions of the controlled features and then the datum feature
of size's true position point onto the transparency.
3. Measure the datum feature's departure from MMC or virtual condition, as applicable (if
referenced at MMC).
4. With the compass point on the (transparency) datum's true position point, draw a circle
the diameter of that departure.
5. With the graph paper stationary, move the transparency around while keeping the
datum's true position point on the graph paper within the transparency's datum shift
tolerance zone.
6. If this can be done and all actual axis positions from the transparency fit inside each
feature's respective tolerance zone on the graph paper (at the same time), the features are
within tolerance.
Figure 15-3 has been inspected. Hole sizes are:
Hole #1 = .387
Hole #2 = .384
Hole #3 = .386
 Hole #4 =.385
Datum hole #5 =.500
Hole positions are:

 Page 357
Hole From datum D on From datum D on
# X-axis Y-axis
#1 1.005 .376
#2 .996 .374
#3 .998 .377
#4 1.002 .369

FIGURE 15-3

 Page 358
If you were to use this variable data to construct a graphical display, you would plot these
deviations from true position on graph paper. Include datums B and D. Datum B is for
angular orientation, datum D for position. Sketch the datum-to-pattern tolerance zones.
Ask, Is the datum-to-pattern position of each hole in tolerance? If not, plot the position of
each feature on a piece of see-through paper. Include a line for datum B and a point for
datum D. Draw a circle around datum point D. The circle should be the diameter of the
datum feature D departure from MMC. Put the transparent paper over the first graph.
Keeping the datum B lines on both sheets parallel, shift the transparency around, trying to
shift the hole axes into their tolerance zones on the first graph. During this shift, the axis
point of datum D on the first graph must stay within its tolerance zone on the
transparency. (See Figures 15-12 through 15-16 for completed examples of this
technique.)
When determining additional pattern shift for paper gaging of a hole pattern controlled to
a center datum feature of size that has its own tolerance of form or location and is
controlled for orientation to the same datum as the pattern of features, we must determine
the orientation of the datum feature of size to the orientation datum. If the datum feature
of size's (in this case, a hole) feature control frame read (see Fig. 15-3):

and A was the orientation( )datum, we would want to know how much out of to A
the datum feature of size was. If it was not out of to A at all (perfect to A), we
would automatically derive an allowed shift of the hole pattern to the center datum feature
of size of a diameter of .010.
Then, in addition to that, if the datum of size departs from its MMC of .500 to .520
(LMC), then .020 more would be added to the .010 for a total of a possible .030 diameter
shift of the pattern of features to the datum feature's axis. Essentially, this process is the
same as determining the difference between the orientation virtual condition of the feature
as produced and the real virtual condition as explained below. The usual way to calculate
virtual condition boundaries if features are controlled at MMC is:
For shafts: Add the tolerance in the feature control frame to the shaft MMC.
For holes: Subtract the tolerance in the feature control frame from the hole MMC.
In calculating the orientation virtual condition of the part as produced:
For shafts: Add the actual amount of out-of-perpendicularity to the actual shaft size.

 Page 359
For holes: Subtract the actual amount of out-of-perpendicularity from the actual hole size.
The difference between the orientation virtual condition of the datum feature as produced
and the virtual condition of the datum feature as defined in ASME Y14.5 would be the
diameter of the allowed shift of the pattern of features as a unit to the center datum. This
shift exists only for the hole pattern as a unit. It is separate from any individual tolerance
allowed for each feature (by the hole pattern's feature control frame and, if controlled at
MMC, any bonus tolerance derived from that feature's departure from its MMC within
size limits).
On the other hand, if the datum feature of size is a primary datum, any shift of the hole
pattern as a unit is derived only from the center datum feature of size's departure from
MMC (within size limits). Obviously, these statements apply only if the datum feature of
size is referenced at MMC in the pattern of the features' own feature control frame.
Paper Gaging with Datum Features of Size
When paper gaging parts that use datum features of size at MMC, the representation of
those shift tolerance zones that are created when the datum features of size depart from
MMC is, at times, not well understood. It does create a unique problem, in that the shift
allowed, whether it be for movement or for both movement and pattern rotation, must be
represented by a zone of specific extent.
The zone created by a cylindrical datum feature of size's departure from MMC (or virtual
condition), for example, can be shown as a circle on the paper gage. Each square on the
graph paper can represent whatever denomination seems appropriate, but once that
denomination is decided on, the shifting tolerance zones and the tolerance zones applied
to the individual features must use that scale.
A different scale is used when assigning the location of the tolerance zones to each other
and to the datum features of size. Also, the length of each datum center plane that must be
represented, if datum centerplanes are to be shown, must rely on those same proportions
being applied. In other words, the length of each datum plane shown (for these
noncylindrical-size datums) must be in the same proportion as other datum planes shown
(for other noncylindrical-size datums), for both length and location. This proportion is
used for controlled feature locations as well, for representing true positions.
Another factor that applies to all parts being paper gaged is that although the gage is a
two-dimensional object, the axes being gaged may be located three-dimensionally on the
part and that information used on the paper gage. This is done, for example, by
determining the hole location's deviation from true position at the top and plotting it on
the graph, then determining the hole's deviation from true position at the bottom of the
hole and plotting it also on the graph. These two (or more) points are then connected by a
line. The length of this line

 Page 360

FIGURE 15-4
represents the amount the hole's axis is actually out of perpendicularity. This has the
effect of projecting each hole's three-dimensional axis onto a two-dimensional plane (the
graph paper). The individual tolerance zones (usually shown as circles) may not be
violated by their respective actual hole axes.
To summarize, for the situations discussed above:
The datum feature of size location and the true position of each feature should be in
proportion to one other and to the part drawn.
Each square on the graph paper may be equal to any numerical value you desire.
If the hole or shaft being graphed is out different amounts along its axis, multiple points
may be graphed (usually twoone at the top and one at the bottom), connected, and
represent a projected line (axis) on the graph paper.

 Page 361
The length of each datum plane shown to represent datum centerplanes must be in
proportion to other datum centerplanes for length and also use the same proportion as
hole-to-hole locations and datum-to-hole locations.
Shifting of hole or other controlled feature patterns must be done without allowing datum
axes or centerplanes to violate their shift tolerance zones whether a diameter or a
rectangle of specific extent.
ApplicationsChapter Examples
Practice Part 1
1. Using the information given (regarding a part produced per Figure 15-4), we have:
a. Completed the following inspection charts.
b. Graphed the results. (Note: You may choose your own scale to plot each hole's true
position. Because all datums are planar, the scale to plot true position is not a
significant factor. Each square on the graph equals .002 for tolerance zone size and
actual hole axis location.)
Answer to Question 1:

FIGURE 15-5
Practice Part 1Inspection charts.

 Page 362

FIGURE 15-5
Continued.

FIGURE 15-6
Practice Part 1Inspection chart.

 Page 363

FIGURE 15-7
Practice Part 1graph.

 Page 364

FIGURE 15-8

 Page 365
Practice Part 2
2. Using the information given (regarding a part produced per Figure 15-8), we have:
a. Completed the following inspection charts.
b. Graphed the results. (Note: Although not reflected here, the scale to plot each hole's
true position may be three squares equals .100. Each square on the graph equals .0015
for tolerance zone size and actual hole axis location.)
Answer to Question 2:

FIGURE 15-9
Practice Part 2Inspection charts.

 Page 366

FIGURE 15-9
Continued.

FIGURE 15-10
Practice Part 2Inspection chart.

 Page 367

FIGURE 15-11
Practice Part 2graph. This part checks good with the smaller
zones centered. If this had not been true, the pattern of four smaller
zones could have been shifted off center.
Practice Part 3
3. Using the information given (regarding a part produced per Figure 15-12), we have:
a. Completed the following inspection charts
b. Graphed the results. (Note: To plot each hole's true position, each square equals
.050. Each square on the graph equals .001 for tolerance zone size and actual hole axis
location.)
Part 3 (Fig. 15-12) has been inspected. The hole sizes are:

Hole #1= .387

#2= .384
#3= .386
#4= .385
Datum Hole #5= .500

Hole positions are:


 Page 368

FIGURE 15-12

FIGURE 15-13
From datum D on X- From datum D on Y-
Hole#
axis axis
#1 1.005 .376
#2 .996 .374
#3 .998 .377
#4 1.002 .369

 Page 369
Graphing Instructions:
a. Plot these deviations from true position on graph paper. Include datums B and D.
Datum B is for angular orientation, datum D for position. Sketch the datum-to-pattern
tolerance zones.
b. Is the datum-to-pattern position of the holes in tolerance? If not, plot the position of
each feature on another piece of paper. Include a line for datum B and a point for
datum D.
c. Now draw a circle around datum point D. The circle should be the diameter of
datum D's departure from MMC. Run a transparency of this second piece of paper. Put
the transparency over the first graph. Keeping datum B lines on both sheets parallel,
shift the transparency around, trying to fit the hole axes into their tolerance zones on
the first graph. During this shift, the axis point of datum D on the first graph must stay
within its tolerance zone on the transparency. If the holes are within their tolerance
zones without the shifting procedure described above, just graph all necessary
information on the one piece of graph paper.
Answer to Question 3:

FIGURE 15-14
Practice Part 3Inspection charts.

 Page 370

FIGURE 15-14
Continued.

FIGURE 15-15
Practice Part 3Inspection chart.

 Page 371

FIGURE 15-16
Practice Part 3graph.
Practice Part 4
4. Using the information given (regarding a part produced per Figure 15-17), we have:
a. Completed the following inspection charts.
b. Graphed the results. (Note: To plot each hole's true position, each square equals
.050. Each square on the graph equals .001 for tolerance zone size and actual hole axis
location.)

 Page 372

FIGURE 15-17
Establishing datum center planes from datum
featuresdatum reference frame (three-plane concept). Note:
Drawing purposely shown incomplete.
Answer to Question 4:

FIGURE 15-18
Practice Part 4Inspection charts.

 Page 373

FIGURE 15-18
Continued.

 Page 374

FIGURE 15-19
Practice Part 4Graph.

 Page 375

15.2 Comparison of Measurement Techniques and Equipment (Advantages and

Disadvantages)
There are many ways of gaging features of size controlled by geometric positional
definitions of location. All have certain attributes, drawbacks, and measurement
uncertainties.
Nonvisual Techniques
Open set-up, e.g., the use of such instruments as precision height gages, dial indicators,
angle plates, and surface plates, does not rely on an extreme initial out-lay of capital to
obtain, but is time consuming. This becomes costly, especially on a long-run production
basis. The more parts that must be inspected in a particular lot or run, the more the
inspection costs soar.
Functional or receiver-type gages are initially expensive to manufacture or buy, but in
the long run save a great deal of time in telling a good part from one that will not
function. Still, many times, drawing changes can make expensive functional gages
obsolete before they have paid for themselves. They are also usually reserved for features
controlled by MMC symbols and, moreover, are not capable of giving inspectors variable
information that might be useful in determining patterns of deviations that could be
corrected before parts in a production run exceed allowed tolerances.
Coordinate measuring machines (CMM) are very helpful if used with the same common-
sense set-ups that are employed with surface plate type set-ups. But, often datums are
established by arbitrary contact with datum features instead of high point contact, or the
algorithms the computer uses are inconsistent with definitions in the Y14.5 Standard and,
therefore, prone to a wide variation in answers. So, depending on what points and how
many points are probed and what the computer does with that information to determine
conformance, CMM's can be a blessing or just a way of allowing the inspector to get the
wrong answer faster. Another drawbackat least initiallyof the CMM is that they can be
very expensive.
Visual Techniques
Paper gaging is an excellent method of verifying some otherwise complex positional
controls. It is also relatively inexpensive in and of itself. However, paper gaging is used
after the collection of points on the surface of the part by the tools mentioned above
(open set-up equipment and/or CMM). Once feature center-planes or axes are located in
relation to one another and to their datums, they are plotted on graph paper and
sometimes on overlays to show not only tolerance zone location and size, but also
allowed pattern shifts.
Optical comparators are among many additional alternatives. Optical ver-

 Page 376

FIGURE 15-20
Partial part drawing.

FIGURE 15-21
Overlay of virtual condition boundaries. (Note: The
location of datum feature D to datums A, B, and C
has already been verified prior to this.)

 Page 377
ification of features of size that are controlled by positional tolerances has some distinct
advantages. For holes, whether controlled at MMC or RFS, optical comparators can be
particularly useful, especially on very small features. Boundaries of virtual condition are
generated by these controlled features of size. These boundaries can be easily recreated on
transparent overlays located from their theoretical datums (whether the datums are planes,
lines, axes, or points). These overlays can be used in the same manner as other optical
comparator visual inspection techniques that are more commonsuch as checking sizes or
the form of surfaces. The part is fixtured on the comparator table to establish orientation
and location to the datums. Magnification requirements are assigned in the standard way.
Once the part is in place and the datums drawn on the overlay aligned with the datums
simulated from the datum features, the actual controlled feature's worst mating surface
boundary (as produced) can be compared to the allowed virtual condition boundary or
tolerance zones drawn on the overlay. Should any portion of the actual feature's surface
infringe on the allowed boundary or go outside of the tolerance zone, the feature can be
considered in violation of the allowed geometric tolerance (see Fig. 15-22).
The optical comparator is also capable of taking into consideration allowed pattern shifts
due to a primary datum feature of size's departure from MMC (or virtual condition if it is
secondary or tertiary) (see Fig. 15-20 for partial part drawing). For example:

Because the datum feature of size D has a circled M after it, a pattern shift may bring
feature axes in the pattern into their tolerance zones. In this case, the datum feature and
the features controlled to it are represented at virtual condition on the overlay used to
check the controlled features (see Fig. 15-21). Should the datum feature of size D not use
all of its virtual condition boundary due to any deviation from MMC within its size limits
or not use all of its allowed orientation tolerance to datum feature A, an extra shift of the
controlled feature's boundaries of virtual condition is allowed. In other words, the pattern
of holes may be produced farther to one side or another as a group than they would
other-wise be allowed.
This can easily be simulated by shifting the overlay, or part, to allow the controlled
features to fall within their respective boundaries of virtual condition. This is allowed as
long as datum feature D's surface does not infringe on its virtual condition boundary (see
Fig. 15-23). In any case, should the controlled features not infringe on their respective
virtual condition boundaries, they are considered within their assigned positional
tolerance.
This overlay method can be used for single features, patterns, or even multiple patterns of
features with simultaneous positional and/or profile require-

 Page 378

FIGURE 15-22

FIGURE 15-23
ments. It can be used for simple positional controls or even composite positional
tolerances.
In composite positional tolerancing, the datum-to-pattern relationship would require one
overlay and would be checked with the previous described procedure. Once that
relationship was verified as being in boundary, a second overlay would be required to
check the feature-to-feature refinement. The only datum simulations required on the part
for this check would be for refinement of orientation requirements. The overlay has
boundaries of virtual condition as directed by the feature's MMC and the second-tier
feature-to-feature requirement

 Page 379
of the feature control frame. The boundaries would be centered around the true position
points dictated only by the basic dimensions between the controlled features. These
boundaries would be matched against the controlled features' actual worst condition
surface boundaries (as produced).
As with the other positional verification methods initially described, this method also has
limitations. It is recommended for through internal features, such as holes and slots. Blind
holes or shafts and other external features may be more difficult to verify using this
method, depending on the sophistication of the optics used.
The overlay of virtual condition boundaries (Fig. 15-22) of datum feature D and the
pattern of holes located to it are positioned on an optical comparator screen over the
image of the actual datum feature and hole pattern as produced. Dark areas represent
actual surfaces of holes in violation of virtual condition boundaries.
In Figure 15-22, the actual surface of datum feature D is now shown to demonstrate that
the surface of the actual feature did not infringe on its virtual condition boundary. This is
an indication that a shift of the pattern of holes away from that datum axis is possible.
This additional shift of the four holes as a group is allowed, as long as datum feature D
does not begin to violate its boundary and datum feature B is kept parallel to its
simulation on the overlay. If, as is indicated in Figure 15-23, the part, once shifted, shows
no feature violations of the virtual condition boundaries, the controlled features conform
to the required geometric control.

 Page 381

16
Teamwork Issues
16.1 Unity of Purpose Affects the Bottom Line
So often we are asked to justify the cost of properly training staff members in the various
aspects of applying, interpreting, and inspecting geometric controls. In recent times I have
grown to dislike the term geometric dimensioning and tolerancing, because it implies that
there is some other nationally or internationally accepted form of dimensioning and
tolerancing products. There isn't.
In the United States we have the ASME/ANSI Y14.5 Standard that covers all of the basic
guidelines on the dimensioning and tolerancing of parts. There are standards such as
ANSI B4.4 (Inspection of Workpieces) that explain size limit inspection and the type of
gages available and suitable for use. There are also a series of B89 standards
(Dimensional Metrology), some of which are still in process, that explain the procedures
for verification. One such standard is B89.3.1 (Measurement of Out-of-Roundness),
which gives us a more in-depth understanding of that symbol and its ramifications.
These standards, however, are not ways of circumventing the techniques given in Y14.5,
but rather are additional information to augment it, explain it, and put it into the proper
terminology geared to the person's job category that may be involved in the
implementation of a particular facet of the dimensioning and tolerancing language at the
particular facility.
Internationally, dimensioning and tolerancing are handled by such standards as ISO
1101(Technical Drawings-Geometrical Tolerancing), ISO 5459 (Datums and Datum-
Systems for Geometrical Tolerances), ISO 2692 (Maximum Material Principle), ISO 8015
(Fundamental Tolerancing Principle for Dimensional, Angular, Form, Position
Tolerances, and Size Control), ISO 5458 (Tech-

 Page 382
nical DrawingsGeometrical TolerancingPositional Tolerancing), ISO 10578 (Technical
DrawingsTolerancing of Orientation and LocationProjected Tolerance Zones), and ISO
10579 (Technical DrawingsDimensioning and Tolerancing Non-Rigid Parts). These ISO
and our ANSI standards have a few differences, but not many, and every time a new
revision of either is issued, the differences diminish.
Although each country considers itself somewhat unique, as do all companies (and to a
certain extent they are), whenever possible, we all essentially try to follow similar
guidelines. And the reason is simple. It's good business. These guidelines are well
researched, efficient, and universally interpretable. They provide a book of accessible
rules by which to describe products that meet functional requirements. Manufacturers
may then produce parts that fall within these tolerances and be assured, if the designer has
described the product efficiently, that parts will then function.
Often, people in manufacturing and quality get so caught up in trying to fulfill the
drawing requirements that they forget to question whether the drawing requirements are
also the design requirements. They just take that as gospel. If the parts don't work, but fit
the drawing tolerances, they feel they have done their part. They get so wrapped up in
their own jobs that they forget what the overall goals of product design, manufacture, and
inspection area cost-effective, fully functional product that will provide the company with
a profit, loyal and satisfied customers, and a good reputation.
The product's design drawings should be a basis for the sustained health of any
organization. They are graphic descriptions of the configurations and parameters within
which the company's products must be produced in order to function. They should
therefore be sophisticated enough descriptions of the finished products to accurately
describe all parts that can be accepted and that actually work. Hopefully, manufacturing is
involved in the design process, lending vital advice on producibility and being advised on
part functionality. They can then take that design drawing and actually make a functional
part. The drawing requirements should therefore fully reflect the part's functional
requirements. Unfortunately, this is not always the case. Education and communication
are still the answers.
Understanding FunctionalityA Key to Part Inspection
Unless the inspectors understand the functional needs of the product, they can only
inspect to the drawing. This often takes away factors vital in the verification of any
product or feature of that productfactors that can help determine whether it will really
work. The bottom line here is that if all the inspectors see their job as only a
determination of whether or not the part fits the drawing requirements, they have lost
sight of one of the prime effects on the bottom line:

 Page 383
that is, whether or not the part will actually fulfill the most basic desire of the customers.
They want it to work. An inspector with a firm understanding of what the part does,
whether it must fit, perform, sustain, balance, be cosmetically pleasing, or do cartwheels,
is going to do a much better job.
When, for example, a coordinate measuring machine (CMM) operator uses that tapered
pin to determine whether or not a hole is in the proper locationwithin its positional
tolerancejust as the CMM manufacturer has instructed him to do, he is often using some
misguided logic to determine only whether the hole fits the drawing requirement. This
method will not tell the operator that, no matter what the CMM manufacturer has said. He
would be able to see that if he would only ask himself the question How much of this
hole has to function? For example, if the hole is part of a pattern that mates with a pin
pattern, the logical question might be How much of the hole must mate? If the answer is
The entire hole, then why on earth would he be inspecting only the very top lip of the
hole? The answers are, I believe, because it is what the operator was taught, it is easier,
and he is, in his own mind, verifying compliance only with the drawing requirements. If
he reminded himself each time that the drawing requirements are only there to describe to
within what parameters the part must conform to actually function, and the inspector
knew what the function of that part was, the operator would understand that no matter
how many computers, bells, and/or whistles are attached to a CMM, if you don't feed it
enough information, or you feed it the wrong information, it can't tell you if the part will
function. Garbage in is not gospel out.
If an inspector does not fully understand part functionality, is not sufficiently schooled in
the inspection procedures necessary to verify each dimension and tolerance on that
partwhether it be a simple size check or the most complex positional controlhe/she will at
times reject fully functional parts and at other times accept parts that can never work.
Inspectors Led Astray
The rate of error in the inspection process due to incorrect instruction by peers and
manufacturers of gages, measuring machines, inappropriate fixturing devices, and
instrumentation is staggering.
I have recently researched tens of thousands of dollars worth of textbooks, training films,
sales literature, and training packages and have personally observed the inspection of
thousands of parts at hundreds of facilities. It has been one of the most unsettling
experiences of my life. The amount of misinformation out there is amazing, and the lack
of awareness as to the availability of our most useful and accurate documentssuch as the
international guidelines and research already in existence, some old (ANSI Y14.5
Dimensioning and Tolerancing), some new (ISO/TR 5460 Guidelines for Verification
Principles of

 Page 384
Geometric Tolerancing), (ASME Y14.5.1 Dimensioning and Tolerancing
Mathematization), some currently evolving (ASME B89.3.2 Dimensional Measurement
Methods) and (ASME B89.3.6 Functional Gaging and Fixturing for Geometric
Tolerances), and some being translated from other languagesis truly disturbing. This
information must be disseminated as soon as possible.
As for those texts, prepackaged training programs, and films being distributed (often by
the people who make the inspection equipment) that show quick, easy, and vastly
incorrect ways of inspecting dimensions and tolerances, I am going to buy a load of the
stuff. Then, I am going to show it to as many people as I can, so that I can say, This is
what they are teaching you! Don't do this! It's wrong! And I know very few will argue the
point. They will use logic, their acquired knowledge, and their years of experience. I will
present them with descriptions from the ANSI standards of the tolerance zones they are
trying to verify and simply ask if these methods are capable of even a fairly close
approximation of part conformance. There are smart people out there. They will see the
point.
At first they will feel betrayed by those who have led them astray. They will feel angry
and vulnerable for what they have done in the past. But, hopefully they will never forget
those feelings. They will cling to them like a bad suit, and the next time they are
confronted with a decision on technique or methodology, they will be contributing on all
cylinders.
The bottom line is a term I can easily relate to when I see hundreds of hours being
absolutely wasted by inspection procedures being performed that cannot possibly tell
whether or not a part can function. This type of activity is movement without positive
effect on the company goals. It is a mouse running a circular treadmill. He's going like
crazy but getting nowhere.
The successful companies are not stupid. Their goal is not just a paycheck at the end of
the week for putting in time. Their goal is not just departmental. Their goal is
companywide and, in a more global sense, nationwide. The best product for the least
money is accomplished when every employee knows what the end product is and what it
does and is not only allowed, but encouraged, to input his or her knowledge and
experience, contributing to the best design for the most functional and producible
product and foreseeing difficulties before they occur.
The best product is produced by the most-skilled, accurately informed, and best-educated
employees who can take all that knowledge they have accumulated over the years and use
it to think, deduce, foresee, suggest, and initiate change to improve the product drawing,
the manufacturing procedures, and the inspection processes. They don't ask, How did we
do this yesterday? to determine how to do it tomorrow. They ask themselves and their
team members (all reliable sources), How can we do this the best, most correct, most
efficient way possible? Then, if the answer is that no one within ear range knows, they
take the time to find out.

 Page 385
This is something we have not consistently done well in the past. The reasons are many:
(a) We are afraid to admit to ourselves we don't know everything. (b) We are afraid of
being ridiculed. (c) We don't want to rock the boat. (d) We never have enough time to do
it right the first time. (e) We have so many bad examples and so much bad advice to
follow, that it is easier to follow those than to research the correct way. (f) We don't know
where to go to find the answer, because we are not aware of what guidelines, standards,
and technical reports exist. (g) We are not used to thinking of the job in a global sense,
and therefore deducing from all our experience and all our education and our good
common sense, the best, most functional, and most cost-effective way of approaching a
job or task.
We have some designers designing parts who have never seenmuch less studiedtheir
manufacturing capability, who consider the people who encounter their drawings (the
manufacturers and the inspectors) as something a cut above subhuman and certainly
without the intelligence to comprehend functionality. That hardly creates or promotes a
team atmosphere.
I have taught thousands of people to understand the symbology of Y14.5, and in that time
have seen a marked majority in these courses of design type personnel. It is not unusual
to have an entire course composed of designers and drafters. Where are the inspectors?
They should be represented in mass. They should be exposed to the latest correct material
on the appropriate inspection of geometrically controlled parts. I have found people with
30 years and more of inspection experience who did not know the correct way to inspect
a length or size dimension, and in those 30 years they have never been made aware of the
standards that tell exactly how to do it. How many others have they passed along their ill-
advised techniques to? The answer, I am finding out, is plenty.
How is the bottom line affected by inspectors performing incorrect, uncalled-for, and
misleading procedures to determine part functionality? Part costs soar. Profits plummet.
And business ultimately goes elsewhere.
16.2 Troubled Times and Technical Leaps
The famous author Henry Miller wrote in 1941, Over and over again I have said that there
is no way out of the present impasse. If we were wide awake we would be instantly
struck by the horrors which surround us, we would drop our tools, quit our jobs, deny
our obligations, pay no taxes, observe no laws, and so on. Could the man or woman who
is thoroughly awakened possibly do the crazy things which are now expected of him or
her every moment of the day?
Well, Henry, I'd like you to know that here it is 1998 and things aren't any better yet. The
airline seats are too small and we have to use them too often and they serve plastic-
covered and plastic-tasting food! We institute the best programs of design, manufacturing,
and quality into our systems by teaching our

 Page 386
employees the buzz words but not making certain they are clear on the concepts. They
walk around speaking in what must seem to them completely noncoherent tongues,
saying things like TQM, Baldridge, SPC, just-in-time, GD&T, and team building. They
say these things because they are part of the industry jargon they know is needed to
survive within their department/company/environment, but they lament over the poor
souls (not themselves, of course) who won't be able to deal with these complicated
concepts (translationDo we really have to do this stuff? I'm too busy doing the real work
to try out your stupid newest program-of-the-day.). But they wear a really concerned look
when they say it, which they spent a great deal of time trying out in the mirror that
morning, and consequently seem so sincere that another insecure soul who was thinking
exactly the same thing thinks it's all right to agree with him and, therefore, not do the
follow-through necessary to make the program really work.
Well, I understand their real concern. There is a whole lot of new stuff we have to learn
to survive in today's industry. It is difficult and frightening and wonderful and exciting.
And contrary to Henry's suggestion, throwing up one's hands and running for the safety
of hiding under the covers is only one of the options.
It's time to get on with the business of business, which, I'm afraid, in this day of chips
(that are not of bovine origin) means the business of education. Cast aside your memories
of mean old Mrs. Crookshank making you stand in front of the class holding two heavy
books at arms length until you suddenly, and some-how miraculously, were struck with
the answer to the question, or soiled yourself from the fear and injustice of it all. These
are new times and as the old song goes, You're not a kid anymore. Learn. Pack up your
troubles in an old kit bag and study, study, study.
I've often said Talk the talk. we'll work on the walk later. Well, it's later and maybe later
than you think. It's time to put into action all the concepts behind the buzz words you
have spent so much time using on people. Become a champion for the cause of better
drawings, better manufacturing techniques, better communication between departments
and a higher-quality product at a lower price to the consumereven if you are the
consumer.
Apply what you have learned, learn more, extend the research, personalize the techniques
to your product line, and spread not only the word, but the zeal and the commitment.
Make it work and teach others how it does. Become a mentor. Don't hoard the
knowledge, thinking it gives you power over others. Share the knowledge and confer.
Confide that you don't know everything there is and allow others to add a missing piece
to your puzzle.
Open up and admit to yourself and to others that you are human and, therefore,
flawedthat you haven't really believed you were perfect since you were about 16 years old
and that you need the help of others for you to continue to learn and to extend what is
currently known (hopefully to share with the rest of us).

 Page 387
We really do need one another, and until you truly believe that, teamwork will just be
another of those words that you mouth to get you through the next day of politics while
you continue to do what you think is the real work.
I inherited it brick and left it marble [of the city of Rome]Emperor Augustus 63 B.C.A.D.
14.
16.3 Concurrent EngineeringKey to our Future
The real challenge of concurrent engineering (CE) and the teamwork approach it entails is
getting people to cooperate who understand nothing about one another and don't even
speak the same language. I am always fascinated by the number of acronyms used within
each job category. We've got SPC, CAE, CAD, GD&T, CNC, CMM, DARPA, DICE, and
on and on and on. Each job category cloaks itself in an endless array of letter
abbreviations and uses them to speak to one another in secret codes. These codes and
terms specific to their jobs define the people to others of their kind. If you speak the
language, you are part of us; if you don't, you are an outsider.
CE efforts that only attempt to link computers with compatible software, I believe, are
missing the point. Besides computer interlinks, we need to be able to sit down and talk to
one another, learn something about not only the terms and codes but the needs and
obstacles with which those in other job categories have to deal. We used to use cross-
training to bridge these communication gaps. But that supposedly became too expensive.
Then specialization took on an air of job category arrogance, and cliques formed that
spawned competing departmental goals and the natural progression of
thatinterpersonal/interdepartmental arguments and prejudice.
CE programs need to take on the job of teaching people how to work together again, but
not in a touchy-feely way. It should be done with a technical training approach. Each job
discipline should be educated enough in the day-to-day technical operations of other
departments to understand what they are saying, why they are saying it, and how their job
disciplines can interact to help one another achieve goals that are good for the overall
company goals and product needs.
People need to be taught to ask questions about terms and techniques used by others they
do not understand. Those they ask need to be taught that because people ask these
questions does not mean they are stupidbut just the opposite. They are taking the first
steps necessary to allow intelligent interaction between departments and job disciplines
that must be able to cohesively interact in order to concurrent engineer anything. I think
we might be missing the point. To work together, we must respect one another. To respect
one another, we must understand what the role of others is in the system, how they
impact the product, and how they go about it. Only then can we put our full and best
efforts (and our hearts) into helping them.

 Page 388
The Defense Advanced Research Projects Agency (DARPA) has spent close to a hundred
million dollars in the last 5 years on research to computerize and Americanize tactics used
successfully by Japan's Tiger Teams in concurrent engineering. This effort was launched
in 1988 under a program called DARPA Initiative in Concurrent Engineering (DICE), an
acronym based on an acronym. The goal of the DICE project is to develop a computer
architecture for CE that would provide people in a large, dispersed organization the same
freedom of interaction and information transfer enjoyed by a small team that is in the
same location.
Although hub centers are a great idea, and help a lot, the idea that people all in the same
location have an easy freedom of interaction and information transfer, in reality, is rarely
the case. I have seen more companies commit suicide with the sicknesses of bickering
and disrespect between departments and their employees within, than any other problem.
CE needs a people approach to solve this. If you want to learn to ski, you must skito play
an instrument, you must practiceand if you want to learn to work together to develop and
engineer a product with others concurrently, you need to practice that, too, and believe in
itnot just pay it lip service.
Often there are prizes deep within the cereal box that we may not see right off but that
have a way of naturally surfacing as we consume the product. People who are raised in
different environments and not exposed with enough frequency to others (even those
who directly may affect their own environment) tend to look upon those strange other
beings who inhabit those environments as being not as smart, competent, or rational as
those within their own everyday world.
This is true in general society, but it is also true within smaller worlds, such as companies.
We have trouble communicating with one another across job categories. We don't even
seem to speak the same language. The terminology is different, and at times we think
those outsidersthose people from awayare speaking gibberish, mere children without a
knowledge of the real world (which is, of course, the one in which we live). This tragedy
of thinking them inferior, or even viewing as adversaries those who are different from us,
cannot only divide and destroy a society but can have as devastating an effect on a
corporation.
Many skills that are essential to communicating between job categories are not taught in
colleges and universities across the nation. Worse, they are often not taughtand even
discouragedon the job. It seems hard to believe, in this day, that the concept of teamwork
is often paid only lip service. But, still it is sadly true. For this reason, people recall with
great fondness the days when white collar and engineering types were required to spend
time on the shop floor. I have found over the years that sometimes training courses can
help to bridge that gap. Courses that bring people of different technical disciplines
together to learn a technical skill can have a powerful side effect. The people involved
gain a

 Page 389
knowledge and a respect of others from different job disciplines. Barriers come down.
And knowledge flows.
16.4 Magical, Mystical Stuff (Doing Something vs. Doing Nothing)
During a recent workshop I was faced with an individual I felt I had encountered at many
other companies in my travels. He wears a different face each time. Sometimes he is a
she, at times older and at others younger, but the mentality remains the same.
I had just spent a significant span of time explaining the workings, interrelationships, and
other pertinent facts concerning a product, with the goal in mind to involve the group in a
discussion about what dimensioning and tolerancing was needed. So, with all necessary
groundwork laid, I said, with a great deal of anticipation, Okay, what should we do first?
The immediate response I received came from the front row. That depends on the
function of the part, he screeched, as though his very existence was at stake. We can't
touch that part. The designer is the only one who can decide these things!
I smiled. I am the designer, I said. I have just spent a great deal of time telling you exactly
what the function of the part is and how it relates to the assembly. You have an assembly
drawing in front of you, showing exactly how the part interrelates. You have a detail
drawing of this part. I will give you any additional information you think you need to
begin suggesting tolerancing controls to use. Let's just discuss it. If you'll suggest
something, the group will go over the advantages and disadvantages of the control and
from that decide if it is the best in this situation and for that feature. If it is, we will use it
and go on to control other features until the part is complete. I smiled again reassuringly
and waited for him to make a suggestion.
He did. His suggestion was This is not possible! If we change anything, we could be
inflicting unknown difficulties into the design. It's best to just leave it alone and let the
people in manufacturing make it. They will know what we need.
Through psychic guidance? I asked. We are the ones who know what it needs. You know
the function. The part needs to be dimensioned and toleranced. For goodness sake, just
do it! Normally, I am not the excitable type, but this Anything but action, Excuses,
excuses and abuses, We can't do anything about that because stuff is about to grind me
down.
I know this fellow has made a career explaining to his peers and bosses why things are so
bad and why nothing can be done about it, while his competitors have seen his nonaction
as an opportunity to design and build betterand at a lower cost.

 Page 390
I often put up overhead transparencies of a company's products (after a few days of
training them on generic products) and generate discussions about what is good, what is
not so good, and what is blatantly incoherent and in conflict with the functional
requirements. Some people amaze me with their knowledge and flexibility. It doesn't
matter to them what is on the current drawing, whether they have drawn it, or it was
drawn by someone sitting to their right or by someone who retired 20 years ago. I can
look into their new eyesso filled with new knowledgeand tell that all the lights are on and
that everyone is home. They are writing furiously on the drawing copy they have in front
of them and verbally making suggestions to lead the conversation. They are blending the
recently studied examples, folding them into their products right before my eyes, applying
what they have learned at breakneck speed to rectify that which they can now easily see
was not in the best interest of their products and their company and themselves.
And I also see others. The looks on their faces horrified! Years of sidestepping and loud
talk, of pretense and ignorance being washed away in the cleansing light of newfound
knowledge. These are the ones who would rather retire than learn, destroy colleagues
rather than have them shine, and build the walls in the scrap area higher rather than
change the process. They are insecure little people with more power and bluster than they
are entitled to, standing in the road of progress, mean and afraid. They are in the way.
I don't want to hear why it can't be done, only ideas and suggestions as to how it can. I
want healthy, cooperative discussionsnot arguments. Give me creativity and inspiration. I
don't want to hear that the problem is too big and the people and resources too small to
solve the difficulties. I needwe all needtenacity and dedication, open-mindedness, and
deductive reasoning. We need to work together, not be threatened by one another.
I have seen the panic in their eyes when the solutions to age-old (and what in their minds
have seemed to be unsolvable) problems just naturally float to the top in discussions in
my classroom, inspiration striking the person next to themthe ones with the lights in their
eyesand yet eluding them completely. They want to protect what was theirs, the bliss of
their ignorance, their position as top intimidator, the loudest yeller, the one with the boss's
ear. And yet, they are helpless as their power melts before them, their protests drowned
out by the ideas to improve, the zeal to correct, the willingness to adapt, and the insight to
apply displayed by their peerswhom they once considered less than their equals and who
are quickly, and for all to see, becoming the future of the company. Their voices, once so
strong and confident in ignorance and negativity, becoming tinny and whinny sounding
until they are nothing but a whisper, a Neanderthal grunt ignored or finally recognized for
what it is, what it has always been.
These are wonderful and sad situations, like watching vile, loathsome creatures sinking in
quicksand, throwing them a rope only to see them shrink

 Page 391
from it as though you had thrown in a venomous snake. They sink lower and lower, until
they are hard to see at all. But they are not gone. And yet, if we are lucky, we will never
hear from them again.
16.5 Avoiding Tiger PitsWhile Helping a Leopard Change Its Spots
I have been traveling a lot lately and have seen and heard some interesting views and
ways of doing business. One company was especially disturbing.
It was populated by extremely intelligent individualsdesigners, drafts people, inspectors,
and manufacturing typeswho absorbed the information I put forth regarding
dimensioning and tolerancing techniques with very little effort. I would explain,
demonstrate, and ask for questions, and they would proceed to apply the principles to a
wide variety of products.
They understood everything I presented. They used deductive reasoning to extend
principles to pertain directly to their product line. They took their existing, vastly
inadequate, and incomplete drawingsand changed them to be correct and complete right
before my eyes. I beamed. It had been a long time since a group was able to do so much
in the span of a short workshop.
The manufacturing people loved it and voiced their opinions on how much easier it
would make their jobs to have the engineering drawings convey so much more in terms
of functional requirements. Quality loved it and explained to everyone in the room just
how it would improve product quality to inspect a part as it is used.
They glowed with anticipation at how, if instituted as suggested, a step-by-step plan of
action could be initiated immediately to use the information gathered during
manufacturing and inspection to improve the drawings and to make the parts more
producible, inspectable, and functional. They wanted to have more input into the design
process to be able to lend their experience about real-world manufacturing and inspection
difficulties faced. They hated the fact that each department viewed itself as autonomous
and at odds with the goals of the other departments within their own company and saw
this as a way of beginning a healing process. One company, one goal, one team.
At the end of the workshop, I felt great, until one of the quality people sidled up to me,
head down, staring at her feet and said, The designers who were here said that they don't
think we need this. They said it would add up-front time to the drawing procedure.
That's it? I asked. They thoroughly understand it. They know how you and the rest of
your colleagues feel about the advantages to the company not only in quality,
communication, and morale, but also that time spent up-front to make the functional
requirements evident and achievable will be returned a hundred-fold down line and for
the life of the product, that it will produce positive

 Page 392
examples for all who come after to follow, that it might even save your company's life
considering you have a competitor who is following these procedures and producing a
lower-cost, better-functioning product than yours. And they say they don't need it?
That's what they said. She looked heartsick, as did others from the group that gathered
around me, their ray of sunshine eclipsed by the same old tired caste system that has
plagued our industries for long enough to see many of them become sick and dying.
Years ago, there were those in design and management who considered every person in
manufacturing and every person in inspection (in fact, all blue collar workers) knuckle
draggers and mouth breathers who needed guidance that could only be relayed verbally
by a lot of shouting and arm waving supplied by a slightly superior foreman. Educating
them was considered a waste of time.
I remember when a peer on the shop floor taking night courses at a local college was
looked on in puzzlement by his bosses as to who gave him the okay to do such a thing. I
was looked on as an oddity myself because I was a tool and die maker with a college
degree. I had a foreman yell at me when his frustration was no longer containable, What
the hell are you still doing here? Why aren't you using that degree? My answer was a
simple one. I am using the education to help me enjoy life more and do a better job here.
This is a noble profession. I trained for 8,000 hours to learn my trademore time than it
took to earn the degree. This is what I do. And aside from the tirades thrust upon me by
him and a few others, I loved that job. I continued to study my trade and those trades
related to it.
Eventually, I found it easier to converse with the designers and drafters, many of whom I
had sat next to in courses on design and drafting and mathematics and other topics in
which we shared a common interest.
Then came the clichs from management. If you don't want to work, go home. But I was
working. I was opening lines of communication between the shop floor and design. You
do your job and let them do theirs. Can't they see, I thought, that we are part of the same
job, all integral elements of a team that cannot succeed without one another?
But the answer was No, they don't see that at all. They saw themselves at odds with one
another, obstacles in one another's way. Someone they had to grind under their heels to
attain their own goalswhatever minor, short-sighted goals those were. Power, prestige,
personal glory. It worked both ways. The shop resented the white collar workers for their
ignorance of real-world manufacturing and quality techniques, and the engineers,
designers, etc., resented the shop workers for their ignorance of design and lack of formal
education.
But forgive me, I have been speaking of the past again to lend perspective on just how far
we have come since those days of the prevalent caste systems and division of power,
fiefdoms within each department, and loyalty of the

 Page 393
workers encouraged and coerced to be only given to one's own. Defend your territory
was the watchword of the time.
And, of course, we have learned from our mistakes. Things are so much different now.
On the surface anyway. We know the buzz words. We know the programs. We know the
lessons. We just don't believe in these cures enough to dedicate the time to instituting
them long range. We tend to feel we have to play catchup and that we are under such a
time crunch to do so that we really don't have the time to think long range. The feeling is
often that since the competition is so much ahead, if we don't do something quickly, we
won't even be around long range. And therein lies a great deal of our problem.
Short-range thinking, short-range planning, short-range solutions, Band-Aids to stop
what has become a full-fledged hemorrhage. In our hearts, we still don't believe we can
change, and because of the deep-down belief, it becomes the reality. And no Band-Aid
approach can repair that damage. We must come to understand that our methods of
conducting business, of design and documentation, of manufacturing and quality are not
always superior, and are sometimes not even competitive.
And until we can truly look ourselves in the mirror, believe and know that we must
change our behavior, our modes of interdepartmental interaction, our thoroughness, and
our genuine feelings for those we must rely on, we will not change. Our managers will
continue to institute training that begins the cure and then not do the follow-up necessary
to carry it through to fruition. We will continue to learn the words, but not understand the
ways of success. Success is ours to save or sell off.
I continue to hope we will rise to the needs of the crisis. Let's see some action out there.

 Page 395

17
Training Issues
17.1 I Wonder As I Wander
Years ago, I gave a workshop at a New England company that shall go nameless. The
engineering manager did not meet me at the entrance on the first day as my itinerary
indicated he would. After repeated calls and paging from the guard shack, the guard
finally went to the manager's office and brought him to me. He did not seem happy, but I
attributed it to personal problems. I had to run to keep up with him as he guided me to
the room where the workshop was to be held. I peered into the cramped, closet-like room
and saw 20 chairs and no tables.
We will have about 30 people running in and out of here. Ten will have to stand, he said.
I nodded. For 4 days?
He nodded. The old engineering manager arranged this workshop before he was
promoted to the company's headquarters. I'm the engineering manager now, and I've
decided we will hold the workshop because of our contract to do so, but we aren't going
to use this stuff. He was smiling now, having gotten to air his point of view.
But it's good stuff, I said. He sneered at me as though he smelled something bad. Does
that mean we can't have more chairs? I asked.
His face got red. I have real work to do. He spun on his heels. I can't be bothered with
this, he said as he disappeared down the hall.
I looked around. There was no audiovisual equipment, no board or chart pad to write on,
and people were beginning to arrive. I pictured violent acts taking place, the weaker
attendees being forced to relinquish their seats by burly guys with one eyebrow. I asked
the early arrivals for help and things began to

 Page 396
work out. We got the needed equipment and chairs, even the luxury of tables for them to
sit at.
However, I noticed the engineering manager lurking in the halls, passing by the open
door, peeking around corners while we held the sessions, incredulous, I'm sure, that we
were overcoming the barriers he had so carefully constructed.
He heard the course participants discussing the topics at break times, and feeling their
enthusiasm and knowledge levels growing, he decided on another tactic. He would
ambush the group during each subsequent break, berating their enthusiasm and belittling
their knowledge. I heard him ranting, We're not going to use that junk around here!
Still, the people were very bright and took to the topic and unbridled encouragement
without a hint of reluctance or shame.
The creature in the hall was losing momentum to the crowd. His hair fell into disarray and
his shirttail wasn't properly tucked in. I noticed a frenzied look overtake his face and his
movements about the hallway took on a scurrying mode, like a scared creature looking
for a gathering wherein he might be welcomed. But he was having more and more trouble
getting people's attention. Even waving his arms didn't seem to help much. He went home
sick before lunch on the third day.
When I called for him 2 weeks later, I was told he was no longer an employee of the
company. The new engineering manager, one of the course graduates, assured me
implementation of the course principles was proceeding smoothlyand that the old
engineering manager was assumed lost during a storm.
A storm? I asked.
Yes, he said with a hint of amusement. It took place in the Vice President's office a few
days after you left.
17.2 Training: An Ongoing Process
An On-Going Process, Not a One-Shot Deal
To successfully implement and carry out a training program, a company must realize it is
an ongoing process, not a one-time training course. Employees must first be trained and
then reinforcement training must occur within 6 months to a year. Reinforcement training
may be in the form of:
1. Taking the same program again.
2. Taking more-advanced courses.
3. Taking courses more specialized to job categories.
4. Taking courses more tailored to a specific product line.

 Page 397
Subsequent reinforcement programs should follow at the rate of at least one per year for
the first several years, if users are expected to be really proficient. All job categories need
to follow this procedure of reinforcement.
An open line of communication is important between course instructors and students, so
that when questions arise between courses, there is a vehicle by which to have these
questions answered.
Individuals within a company should form either a formal or informal network of users
to answer one another's questions and to be able to bounce ideas around.
Manufacturing, design, and quality departments should, in the best of scenarios, be able
to communicate concerns to one another in a discussional, rather that a confrontational,
manner.
On Dimensioning and Tolerancing
If these procedures are followed, maximum success is probable if:
1. Everyone who must interpret dimensioning and tolerancing for manufacturing parts
and inspection of parts is sufficiently trained.
2. Everyone who creates design drawings, gage drawings, or any drawings that have
dimensions and tolerances is also sufficiently trained.
3. The person doing the training is a professional trainer in geometric dimensioning and
tolerancing (GD&T)an expert, not a person who sort of knows it. The instructor is very
important in the process. If the best is not chosen, the information obtained will be either
incorrect or poorly and improperly conveyed.
4. Gains made are viewed in the long term. Training is an investment in time and money
that pays off in huge dividends over the lifetime of a company. It reduces product costs
and increases quality. It builds teamwork and shortens production time. But for this
process to begin, it takes a visionarysomeone who can see beyond the next few days,
someone who is willing to take the necessary steps to reap the longterm rewards.
To be successful in creating a complete, correct, and legally defensible drawing package
using GD&T, a company must make a commitment to training. Many companies sending
a few people to a public workshop or holding one onsite training program feel that their
journey to perfection is complete. Really, it has just stalled on the side of the road. GD&T
cannot be treated like just another program of the day, being in political vogue one day
and out the next. It has to be a way of everyday business, a shared vehicle of
communication between manufacturing, design, quality, and all other individuals who
have to read the draw-

 Page 398
ings. The drawings tell the story of the productits functional requirementsand suggest,
through these needs, how to manufacture the product in order to comply.
Even though manufacturing procedures aren't usually stated on the drawings, they are
implied by functional requirements, as are ways the product must be inspected or gaged
to determine compliance with those requirements. Without this common vehicle of
communicationthe language of GD&Tcompanies are often thrust into arguments of
misunderstanding and suspicion. Chaos occurs. This chaotic way of communicating with
one another without the common language of GD&T is a burden that impedes their
chances of success. But individuals are often so close to the situation, so used to the
arguments and lack of cooperation, they can't recognize it as an unnecessary burden.
Competitors can see it, but they, themselves, cannot.
It's like the story of two donkeys that meet at a hitching rail in a western town. One of
them is a town donkey with nothing on his back but a saddle. The other is a prospector's
donkey, loaded down with packs, camping and cooking gear, and four 50-lb sacks of ore.
His back is bent into a concertina shape from the weight. The town donkey says, That's
quite a load you got there! The prospector's donkey says, What load?

 Page 399

18
Statistical Process Control
18.1 Faulty ImplementationWonderful Concepts
I have often complained about the faulty implementation of wonderful concepts.
Statistical process control has been the latest of these programs not optimally
implemented. So the obvious question is: What is the solution? One problem that could
be rectified is the charting of a size to represent a critical dimension. In most
circumstances, size alone is not worth charting. The reason I say that is that the actual
local size (two-point-opposing-point size) is not usually critical. This is especially true
when the mean is thought of as the difference between the maximum material condition
(MMC) and the least material condition (LMC) and/or the spread of dimensions is charted
as the spread of the actual locally measured sizes. This is usually easy-to-collect
information, but is such a simplistic dimension devoid of functional value as to make it
unworthy of charting, and even misleading as to how functional a feature may be.
Perhaps as a tiny factor in a sea of more complex pieces of data it can help, but to endow
it as the only critical requirement is usually a mistake.
The collective effect of the geometric tolerance and the size together, considered as one
entity, is much more telling for functionality. If we consider both, we end up considering
the difference between the result of the aforementioned boundary generated by size and
geometric control as an item worthy of consideration. This boundary, whether generated
by the MMC concept or the LMC concept, is known as the virtual condition. The
boundary opposite the virtual condition is the resultant condition. Virtual condition is a
constant worst case boundary created by the collective effect of the size and geometric
control, while the resultant condition is the worst case nonconstant boundary that is
simultane-

 Page 400
ously generated by these controls of size and geometry. Perhaps a truer functional design
mean should be considered the difference between these two boundaries rather than
merely the difference between the MMC and the LMC. However, since we should be
considering the mean difference between the worst and best cases, we should instead be
considering the mean difference between one of these worst case boundaries and the size
considered most friendly to the functional needs. For example, in an MMC concept
situation, we specify a MMC symbol next to a geometric tolerance such as a positional
tolerance. This generates a worst case mating boundary. The best case mating condition
would be LMC size (using none of its geometric tolerance). So, the true design mean in
such a circumstance would be the halfway size between the virtual condition (MMC
concept) and the LMC size. Deviations from this mean would seem to be more
functionally vital than just the deviations from the mean between MMC and LMC alone.
Likewise, charting or graphing a deviation such as the actual produced worst case
boundary and the design Virtual Condition boundary would be more telling and
productive than what is currently done. I would then define an as-produced worst case
mating boundary as the minimum cylinder that can be circumscribed about a shaft with
the cylinder's center at perfect orientation and location.
With the as-produced worst case mating boundary determined, one could then compare it
to the virtual condition boundary (MMC concept) to deduce the difference between the
allowed worst case boundary (virtual condition) and the actual worst case boundary. This
would be a comparison of two three-dimen-

FIGURE 18-1

 Page 401

FIGURE 18-2
sional pieces of data and therefore more functionally meaningful than just a two-
dimensional charting of 2-dimensional size deviations. For accuracy we would commonly
consider the difference between the mean size (halfway between MMC and LMC) and the
actual size. Instead, I propose charting the difference between the functional mean (the
halfway mark between the virtual condition and the LMC) and the actual mean (the
halfway mark between the as-produced worst case boundary and the LMC).
For repeatability, we commonly chart the local size deviations. A more meaningful three-
dimensional size would be to chart deviations of the mating sizes (the minimum cylinder
that can be circumscribed about a shaft or the maximum cylinder that can be inscribed
into a hole). We should also keep track of the deviations in the axial locations of the as-
produced actual mating envelopes without their centers at true position.
For example, in Figure 18-3, the axis of the 27.75 will always be at true position, since it
is the actual mating size with regard to (true) position. This is the axis of the as-produced
worst mating condition. But the axis of the 28.25 is the actual mating size without
perfect orientation or (true) position. This axis could be charted from part to part for
deviations from one another. Values could be given in X and Y locations of the points on
the actual mating envelope axis nearest the primary datum and also farthest from the
primary datum (but still in the hole).
As already mentioned, rather than simplistic size differences, meaningful comparisons
might better be thought of as factors like the difference between the virtual condition
(MMC concept) and the LMC compared to the actual deviation

 Page 402

FIGURE 18-3
of the feature axis from true position (related in terms of a diameter). Since the difference
between the virtual condition and the LMC is equal to the total allowed feature tolerance
(geometric tolerance plus the size tolerance), it is easily comparable to the amount of
tolerance that has been used. The diameter of the actual deviation of the controlled feature
axis from true position is the amount of tolerance that has been used.

 Page 403
Potentially, features that are holes and shafts may move (have a diameter of airspace) in
or on a functional gage by the difference between their LMC and their virtual condition
since the functional gage is designed at the virtual condition (MMC concept) of the
feature it gages.
Actually, controlled features such as holes and shafts may move (have a diameter of
airspace) in or on a functional gage by the difference between their produced mating size
(perfectly oriented to the primary datum) and their virtual condition (MMC concept) size.
Therefore, the amount of tolerance they have used up is the difference between how
much potential movement (diameter of airspace between virtual condition and LMC) they
have on the gage and their actual remaining movement (diameter of airspace) on the gage.
So, used tolerance = P - A, where P is the potential movement (diameter of airspace
between virtual condition and LMC) on the gage and A is the actual remaining movement
(diameter of airspace) of their produced mating size (perfectly oriented to the primary
datum) on the gage.
If LMC was the desired size, then the difference between virtual condition (MMC
concept) and LMC would denote a perfectly sized, located, and oriented feature. But since
the mean size may be the desired size, then the mean feature may be designated as having
a movement (a diameter of airspace) in or on the gage equal to one-half the difference
between the feature's virtual condition (MMC concept) size and the LMC size of the
feature.
I tend to use gages to illustrate these concepts only because they are generally more three-
dimensional and easy to understand than visualizing these concepts with other methods of
measurement. But gages aren't necessary to employ these concepts.
Certainly, these concepts of modeling situations of the perfect versus the actual can be
broken down into components as has been traditionally done. For example, size can be
considered separate from geometric tolerance, but when this is done more emphasis is
placed on the MMC size limit than it generally deserves. One must ask why MMC is even
considered as a functional factor. We know LMC is a major component of material
strength and wall thickness and that virtual condition (MMC concept) determines whether
features will mate/assemble. MMC, unless it is also the virtual condition (which occurs
when features are referenced in a feature control frame at zero tolerance at MMC), is most
often a meaningless drafting requirement that causes fully functional features to be
rejected as in violation of a size (MMC) requirement. Therefore, one must question it
being used to set a functional mean, as when the halfway mark between MMC and LMC
is used as a target for which to shoot. A better target and a more functional mean would
be the halfway mark between the LMC and the virtual condition (MMC concept).
Having gone over these concerns, I will now offer a more traditional version of statistical
process control. Let's begin with the concept of Cp. Cp is a mea-

 Page 404
sure of the specification range in comparison to the spread of the population. Cp can be
defined as the difference between the upper specification limit and the lower specification
limit divided by 6 times the standard deviation. Or, in formula symbology:

where U is the upper specification limit. For size, distance, radius, and angle tolerances U
equals the upper limit specified. And for geometric tolerances such as form, distance,
angle, and runout (controlled regardless of feature size) U is the specified value of the
geometric tolerance. And L is the lower specification limit. For size, distance, angle, and
runout (controlled at regardless of feature size) L is considered zero. s is the standard
deviation.
Another index of process capability known as Cc may be considered as a measure of the
location of the average of the population as it deviates from the desired midpoint of the
specified tolerance range of size, distance, angle, or radius. But for geometric tolerances,
it is the deviation from the perfect geometry of form, orientation, location, runout, etc.
This process capability index (Cc) can be expressed by the following formula:

where t = the target value = (U + L)/2 and Cc = m/U if the target value t = L = O. In these
formulas m = mean, U = upper specification limit, and L = lower specification limit.
Another process capability index is known as Cpk. Cpk is a measure of both the location
and the spread of the population. Cpk can be expressed as:
Cpk = min (Cpl, Cpu)
where Cpl equals (m - L)/3s and where Cpu equals (U - m)/3s. Again, in these formulas,
m = mean, L = lower specification limit, U = upper specification limit, m = standard
deviation, and min (Cpl, Cpu) = the minimum of the enclosed items. Cp and Cpk use to
calculate s. The importance of this is that to calculate , you need a control chart so you
have a measure of process stability.
Another deviation index is expressed as a root mean square (RMS) deviation index. This
index, known as Cpm, can be expressed as:

 Page 405
The Cpm is equal to the Cp when the mean (m) equals the target value (t). Cpm is a
measure of the root mean square deviation of the population variable from the target
value.
In these definitions, the term population is considered the totality of items under
consideration.
Other useful definitions given in the International Organization for Standardization (ISO)
documents follow.
1. Assignable cause: A factor (usually systematic) that can be detected and identified as
contributing to a change in a quality characteristic or process level (from ISO 3534-
2:1993). And from the AIAG Manual on SPC, special causes (often called assignable
causes) refer to any factors causing variation that are not always acting on the process.
That is, when they occur, they make the (overall) process distribution change. Unless all
the special causes of variation are identified and acted upon, they will continue to affect
the process output in unpredictable ways. If special causes of variation are present, the
process output is not stable over time.
2. Chance cause: Factors, generally many in number, but each of relatively small
importance, contributing to variation that have not necessarily been identified (from ISO
3534-2:1993).
3. Distribution function of a random variable X: A function giving, for every value x, the
probability that the random variable X be less than or equal to x: F(x) = Pr[X<x] (from
ISO 3534-1:1993).
4. Parameter of a population: A quantity used in describing the probability distribution of
a random variable (from ISO 3534-2:1993).
5. State of statistical control: A state in which the variations among the observed sampling
results can be attributed to a system of chance causes that does not appear to change with
time (From ISO 3534-2:1993).
The following section covers more of the concepts of statistical process control. It is
included here to augment one's basic understanding of SPC, but does not attempt to cover
all of its aspects.
18.2 Concepts of Statistical Process Control
Normal Curve
Figure 18-4 represents the total population of the expected population. Since not all parts
in a production run will always be measured, the normal curve distribution makes
estimates based on the measured sample about the total population of the run.
In the histogram shown in Figure 18-5, the process is well centered and the parts are well
within specifications. But, in the long run, a few parts will be at or near the specification
limits. Rarely will any of the parts be beyond the specification limits.

 Page 406

FIGURE 18-4

FIGURE 18-5
Capable Centered Process
Histograms
The histogram bar graph plots frequency distributions. It reflects a process in terms of
one specific characteristic of the process, for example feature size measurements. It shows
the proportion of times that measurements occur and how far these measurements are
from the specification minimum and maximum limits and the specification mean. It
shows process centering, spread, and the pattern of variation.
Calculating Average (Arithmetic Mean)
The following formula is useful in calculating an average, also
referred to as an arithmetic mean.

 Page 407

The same formula in mathematical form is:

Calculating a Median
The median of a group of numbers represented by is calculated by putting the numbers
in ascending order and counting to the middle value. However, if the group has an odd
number of values, the location of the median can be derived as follows:

If the sample size is even, the median is the average of the two middle-most values.
Standard Deviation from Sample Data
Standard deviation is a way of quantifying how the values in a distribution de-part from
the average value of the distribution. The formula that follows may be used to calculate a
sample standard deviation. A sample standard deviation is the standard deviation
calculated from sample data.

To use this formula to calculate the standard deviation using the sample values of 2, 4, 6,
8, 10, and 12, first calculate .

 Page 408

Then subtract from each value, square the results, and sum the squares:
:
2 - 7 = -52 =
-5 25
4 - 7 =
-32 = 9
-3
6 - 7 =
-12 = 1
-1
8 - 7 = 1 12 = 1
10 - 7 =
32 = 9
3
12 - 7 =
52 = 25
5
Sum of
= 70
squares

Next divide the sum of the squares by the number of values minus one, and then calculate
the square root.

Attribute Data
In situations where product is evaluated as either conforming or nonconforming, the
sample results are expressed as either the fraction nonconforming (p) or the fraction
conforming (q).

After many subgroups have been collected, preferably representing a minimum of 1000
parts, the average fraction nonconforming, or average fraction conforming, is
calculated.

These values represent the average process performance over time, and if the process
shows stability, or would represent process capability.
In addition to the sample statistics of p or q, and their average values, and , a measure
of expected variation around these average values is required. This statistic, as before, is
the standard deviation. For the binomial distribution, the standard deviation is calculated
as follows:

 Page 409

FIGURE 18-6
A normal curve divided into eight standard
deviation units. The area under the normal
curve represents probability or expected
variation. For example, if a process is stable
and normally distributed, 99.73% of the process
output will fall within plus or minus three standard
deviations from the average.
The Average Chart for Testing the Stability of a Process
The average chart has a control limit calculated from sample data derived from the
process. The center line is called the grand average (the average of the averages). The
control limits are at from the center line. (Note: This is the standard deviation of the
averages, not the individual measurements.) If the sample averages vary according to the
expected random normal pattern, the process is assumed to be unchanging with no special
causes present, thus the process is predictable into the future. If the sample averages are
found to be outside the control limits or fail to follow a normal random pattern between
the control limits, this is a powerful indication that the process is unstable and special
causes of variation exist. Calculations of capability from unstable processes should not be
done.

 Page 410

FIGURE 18-7
Two distributions with same average but
different spreads.
Only plotting and testing sample averages isn't enough for a process stability check. A
process average can be stable for a short period while its spread or variation changes.
This is represented in Figure 18-7 where the stippled areas show a widening of the
spread.
A range chart is needed to augment the sample averages chart.
Range Chart
The range chart is based on the premise that ranges calculated for small samples from a
single distribution tend to be normally distributed. On a range chart, the center line is
and represents the average of the sample ranges. The control limits are placed from the
center line. But, for subgroups that have sizes 6 or less, no lower control limit exists for
the range values.
When sample ranges follow the expected distribution, the process spread is deemed
stable. If the sample ranges fall beyond the control limits or vary non-randomly between
the control limits, it is considered powerful evidence that the process spread is unstable.
Therefore, the and R chart is a sample-average chart and a sample-range chart
combined into one data sheet. Both charts are plotted from common data. Output from a
process is plotted on an and R chart. If operating in a stable manner, it follows a
random pattern within control limits. This process is said to be in statistical control.
Stability vs. Accuracy
A process stability testing procedure is performed. If the process is determined to be
stable, data from the and R chart needs to be used to characterize the

 Page 411

FIGURE 18-8
Expected distribution of sample range.
process capability. Since stability and accuracy are not one and the same, if the process is
producing features that are beyond the product's specifications, the process must be
improved.
The and R chart can show a graphical display of a relationship between process average
and process standard deviation.
Normal Curve Chart
The process depicted in Figure 18-9 will yield acceptable parts. Were the process to shift,
out-of-spec parts could be produced.
Process Shifted from Mean
The process depicted in Figure 18-10 has shifted so as to produce features that fall outside
of the lower specification limit.
Capability Analysis Using Variable Data
Capability analysis predicts the variation one should expect from the process or portion of
the process under analysis. The variation is then compared to a specified tolerance to
determine the process's ability to meet the specification. The three types of variation
studies conducted in SPC are:

 Page 412

FIGURE 18-9
Normal curve describing a capable process.

FIGURE 18-10
Normal curves showing shift of process
from specification mean.
1. Process capability
2. Process potential/machine capability
3. Measurement system capability.
Let's examine three common process capability indices. All these calculations assume a
stable, in-control process. These will quantify the degree of process capability.
1. The process capability ratio (PCR)
2. Cp index
3. Cpk index

 Page 413
The PCR is calculated as follows:

If the PCR is multiplied by 100, it can be interpreted as the percent of the tolerance the
process is capable of holding. However, although capable, the PCR doesn't use the
process average's location. The process could be stable and have a small PCR value
(which is good) but still produce a large percentage of out-of-tolerance parts. PCR is a
measure of the potential of the process if it was centered on nominal.
If the PCR value is 1, it means that the process is capable of producing 99.73% of the
product within the tolerances, if the process were perfectly centered. Still, since in reality
the process average will most likely be off the nominal on occasion, we need to examine
the remaining process capability indices.
The capability index system will provide a more complete measure of process
performance and capability, especially for measuring location of the process in relation to
the engineering specifications. To accomplish this, we must understand the Cp index and
the Cpk index.

A Cp value of 1 shows a process capable of producing 99.73% of the product within

tolerance. The Cp used with the Cpk index gives a more complete analysis of the
performance of a process.
Cpk quantifies the process spread and the centering of the process to the upper and lower
specification requirements. It measures actual performance rather than potential.
For bilateral tolerances, Cpk can be calculated with the following formula:

Cpk compares the process average to the nearest specification limit. For bilateral
tolerances:

 Page 414

A Cpk value of 1 indicates a process that is producing at least 99.73% of its yield within
tolerance. A positive Cpk value of less than 1 indicates the process average is within
tolerance but at least one of the three sigma points is beyond the tolerance limits. A Cpk
value of zero indicates the process is centered on one of the tolerance specification limits.
A negative Cpk value indicates a process average beyond one of the specification limits.
If processes are known to have unilateral tolerances and distributions that are not normal,
care must be taken in trying to use the Cpk. However, if the tolerance is unilateral and a
normally distributed process, the following equations may be used.
If requirement is a maximum spec:

If requirement is a minimum spec:

18.3 Calculating Percent of Nonconforming

If a process Cpk value is determined to be less than 1, the percent of the product not
conforming to tolerance may be of interest. The two stages of determining this are as
follows:
1. Convert the distance between the specification limits and the process average into
standard deviation units known as Z values.
2. A table of Z values is created to predict the percent of parts beyond the tolerance.
One standard deviation above the process average has a Z value of 1. Two standard
deviations above the process average has a Z value of 2, and so on. Z values may be
either positive or negative depending on which side of the mean is being considered. The
following formulas are recommended:

 Page 415
The Z values derived from these formulas are expressed in standard deviations. A normal
curve is a probability distribution with the total area under the curve equaling 1 (100%).
The Z values when graphed on the curve represents the percent of parts not conforming
to tolerance. Their fractional values are listed in Tables 18-1 and 18-2.
To use Tables 18-1 and 18-2, the values are entered into the tables using the method of: a
negative Z value of 3 is viewed as -3.0 and .00 (total of -3.00).

 Page 416

So it is looked up as -3.0 on the left side and .00 across the top for a value of .00135, or
.135% below the specification. The percent above the specification can be determined by
entering the other Z value on the plus table. For example, a Z value of 2.58 is entered as
2.50 on the left side of the table and .08 at the top, for a value of .9951. This is the total
area to the left of 2.58 standard deviations. To find the area to the right of 2.58, subtract
.9951 from 1.0000. 1.0000 - .9951 = .0049, or .49% above specification. The total percent
outside the specification is then:

 Page 417

Machine Capability Studies

The difference between a process capability and a machine capability study is that a
machine capability study concentrates on consecutive pieces in a run of one batch,
without tooling changes, with one operator, one inspector, and one gage, whereas a
process capability study is conducted over an extended period, with multiple batches,
allowing normal change cycles, with all operators, all inspection personnel, and all gages
appropriate for the process. It is recommended that a machine capability have a Cp greater
than or equal to 1.33 (based on a sample of 100 pieces) and that a process capability have
a Cpk greater than or equal to 1.33 over an extended period of time.
For machine capability studies, it may, at times, be more practical to use a sample
selection of less than 100 pieces. If so, a higher requirement of Cp should be given.
Prior to and R chart construction, a histogram can be formed using the data for an
impression of capability and distribution. Any distribution that is not normal is
investigated and causes eliminated. Data from normal distributions are used to calculate
control limits and plot charts.
For stable processes, (the average of the averagesgrand average) can be used to
estimate machine average/process potential and estimated process potential (machine
spread) can be determined with:

PCR and Cp are calculated using estimated standard deviation and average values.
To summarize:
1. Choose pieces consecutively.
2. Measure and record results in the same production sequence.
3. Create a histogram for normality.
4. Determine stability.
5. Calculate average and standard deviation.
6. Calculate PCR and Cp.
Variables Gage Capability Studies
The total variation due to variables gages repeatability and reproducibility can be
determined through studies. Other characteristics of variables gages to be studied include
accuracy and stability.

 Page 418
The first method is shorter and provides an estimate of measuring system variation. It
shows a combined error of repeatability and reproducibility, but doesn't differentiate
between the two for a breakdown of where error is originating. To carry out the shorter
method, it is recommended that more than one gage appraiser be used, 510 parts be
measured, and the measurement system be calibrated using appropriate masters or
National Institute for Standards and Technology traceable standards.
If two gage appraisers are used, each part is measured once and the range recorded. The
difference in readings taken on each part by each gage appraiser is the range (high minus
low readings). This is done for all parts in the study. The ranges are then added together
to get the sum of range values.
If the sum of the ranges was .005 and five parts were measured, then:

Gage system repeatability and reproducibility error (GRR) can be calculated as follows:

GRR=4.33

4.33
=
.001
=.00433

The number 4.33 used in the formula is a constant for a sample of five parts and two
appraisers. See Table 18-4 for other constants.
If we divide the gage system error (.00433) by the feature tolerance and multiply by 100,
the gage system error is expressed as a percentage of the tolerance. If the tolerance was
.03, the calculation would be:

To many companies, a %GRR under 10% is acceptable, 1030% is marginal, and over 30%
is unacceptable and the gage should not be used. If the shorter method of gage analysis
produces results in the marginal or unacceptable percentages, it is recommended the
appraiser continue the study with the longer method of gage analysis. The longer method
of variable gage study differentiates between the components of total error. It breaks the
error down into equipment (repeatability) error and appraiser (reproducibility) error. This
makes corrective action more specific to the problems.
For this method, it is recommended that at least two variables gage appraisers be used, 10
parts be measured, and the measurement system appropriately calibrated as described for
the shorter method. As with the shorter method, all parts are measured once by each
appraiser and the measurements recorded. The range for each set of readings is calculated
and recorded.

 Page 419
R = XHIGH - XLOW
The average range for each gage appraiser is calculated and recorded.

The grand average is calculated by adding the average ranges for all appraisers together
and dividing by the number of appraisers. For example, if three appraisers had average
ranges of: .0004 for appraiser #1, .0006 for appraiser #2, and .0004 for appraiser #3, then
the average range for all samples could be calculated by:

The upper control limit (UCL) range for the individual ranges can then be calculated and
then recorded using:

D4 is a multiplier determined by the sample size. See Table 18-3 for the appropriate
number.
Stability is then tested. If any individual range exceeds the upper control limit range
(UCLR), the causes must be found and the trials repeated or the affected data tossed out.
The affected average ranges are then recalculated for special cause. This procedure is
continued until all test requirements are complied with. If more than 20% of data points
are discarded to meet the requirements, measurement equipment may need
improvements.
Then each appraiser's average is calculated and recorded by the formula:

or in words, the average of measurements equals the sum of the individual measurements
divided by the number of measurements.
The range of averages is computed by:

The next steps involve calculating:

1. Repeatability (equipment) variation using:
and repeatability variation as a percent of tolerancing using:

 Page 420

 Page 421
TABLE 18-4 Range Multiplication Constants
Number of appraiser
Number of parts 2 3 4 5
1 3.65 2.70 2.30 2.08
2 4.02 2.85 2.40 2.15
3 4.19 2.91 2.43 2.16
4 4.26 2.94 2.44 2.17
5 4.33 2.96 2.45 2.18
6 4.36 2.98 2.46 2.19
7 4.40 2.98 2.46 2.19
8 4.40 2.99 2.48 2.19
9 4.44 2.99 2.48 2.20
10 4.44 2.99 2.48 2.20
Under 10% = acceptable; 1030% = marginal;
over 30% = unacceptable.

constants are obtained from Table 18-3 and are dependent upon the number of the sample
size. K2 numbers, which are range multiplication constants, are to be taken from Table 18-
4. As shown in Table 18-4, K2 constants are dependent upon the number of parts and the
number of appraisers.)
Gage Repeateability and Reproducibility
(R&R) equals the square root of the quantity of the equipment variation (EV) squared
plus the appraiser variation (AV) squared. In formula form:

A %GRR over 30% means the measurement system should undergo mandatory corrective
action. For a %GRR between 10% and 30%, the largest contributor to error should be
determined and possible improvement instituted.
18.4 Statistical Tolerancing
Arithmetic Mean
All manufactured part features are subject to deviations from perfection. If we were to
take all the actual dimensions for a part feature that have been produced

 Page 422
in a production run, we could derive a number that represents the mean size of each
feature. This mean size is known as the arithmetic mean. It is often represented by the
symbol . If the individual measurements are represented by X1, X2, X3, and so on, then
the formula to derive this arithmetic mean is:

Normal Distribution of Tolerance and the Standard Deviation

Standard deviation is a measure of dispersion. It is often represented by the Greek letter
for sigma (s).
The area beneath a bell-curved normal distribution of parts (or part features) represents
100% of the production. The height of this bell-shaped curve represents the number of
occurrences of values of the variable individual components. The standard deviation and
the area beneath the bell curve share a relationship that may be expressed as a percentage.
The arithmetic mean plus or minus one standard deviation (s) is 68.26% of the total
production. The arithmetic mean plus or minus two standard deviations (2s) is 95.46% of
the total production. The arithmetic mean plus or minus three standard deviations (3s) is
99.73% of the total production. Natural tolerance is defined as the limit of plus or minus
3s.
Statistical Probability for Tolerance Stack-Up Analysis for Positional Geometric
Tolerances
Establishing tolerances using the statistical probability method can be expressed as the
tolerance of the assembly (TA) and is equal to the square root of the sum of the squares of
the individual component tolerances (T1, T2, T3, and so on). In the form of a formula, it
is:

Statistical probability can predict how a group of independent part tolerances combine to
create an assembly tolerance. Statistical probability methods of applying tolerances work
best if subsequent manufacturing procedures are closely monitored using statistical
process control (SPC). If processes ever become out-of-statistical-control (in statistical
chaos), one cannot be confident of meeting specifications sufficiently to allow parts to fit
into the assembly.
Normally operating processes produce parts with dimensions that
follow a natural variation reflecting pure chance. A large percentage of parts that are
produced will have dimensions that are close to the average dimension. When

 Page 423

FIGURE 18-11
Standard bell-shaped curve.
spread and magnitude of all dimensional variations are graphed or plotted, a standard
bell-shaped (Gaussian) curve of a distribution is formed.
For a Two-Component Assembly
Statistical tolerance of the assembly can be computed using:

18.5 Glossary of Terms

Attribute Data:
The type of good versus bad information receiver gages give. Attribute data lacks
variables data. It does not relate how good or how bad a feature is, just the presence or
absence of a characteristic. The information is related in the number of conforming
versus nonconforming pieces or number of defects per piece.

 Page 424
Average or Arithmetic Mean:
The value derived when a group of dimensions is taken, the measured values are added
together, and then divided by the size of the group.
Bell-Shaped Curve:
A distribution showing a central peak and a smooth, symmetrical tapering off on either
side, such as a Gaussian curve.
Bias in Measurement:
Systemic error leading to a difference between the true value of the population of features
being measured and the average result of measurements.
Calibration:
Adjusting an instrument using a more accurate reference standard.
Capability:
The ability of a gage, machine, or procedure to hold a certain percentage of products
within specification limits. Many companies consider a process capable when 99.73% or
more of the features or parts being measured fall within the specification limits.
Capability can be expressed in Cp, Cpk, CR (also known as PCR) and other methods.
Cause-and-Effect Diagram:
A diagram showing the relationships between all process inputs and their resulting
problem(s) that affect the process.
C Chart:
A control chart showing the number of defects found in a subgroup of fixed size.
Characteristic:
A geometric trait or product specification that is measured or examined to determine
conformance. A dimension or parameter of a part that can be measured and then
monitored for capability and control.
Control Chart:
A representation, usually graphical, used to keep track of outputs.
Control Limits:
Statistically calculated boundaries that are used to determine whether a process is in or
out of statistical control. Control limits and tolerance limits are not the same.
Control of a Process:
A process is termed in statistical control when the process exhibits only random
variations. When control charts are used, a state of statistical control is assumed to exist
when all monitored points remain between stated control limits. This is a stable and
predictable process.
Cp:
A capability index derived by the following formula: CP equals the tolerance range
divided by 6s.

 Page 425
Cpk:
A capability index that combines Cp and k is a measure of difference between the process
mean and the specification mean (nominal). Cpk is used to determine whether the process
will produce units within the tolerance limits. Cpk equals the lesser of the following two
formulas: Formula 1: (The upper specification limit minus the mean) divided by 3 times
the standard deviation of a sample. Formula 2: (The mean minus the lower specification
limit) divided by 3 times the standard deviation.
CR:
The inverse of Cp. CR equals 6s divided by the tolerance. The smaller the value, the more
capable the process (also known as process capability ratio).
Distribution:
A display of values to show frequency of occurrence.
Mean:
The numerical value in a distribution of values calculated by adding all values, then
dividing by the number of values that have been added. Same as average or arithmetic
mean.
Median:
The middle value in a group of numbers.
Range:
The difference between the lowest and highest values in a set of values.
R Chart:
A control chart that shows the range of variation of the individual subgroups.
Reliability:
The probability that a product will properly function for a defined period of time under
certain conditions.
Repeatability:
The variation in measurements that is obtained when one inspector using the same tool(s)
measures the same feature characteristic of the same part.
Reproducibility:
The variation in the average of measurements made by a variety of operators who use the
same tools measuring the same characteristics of the same parts.
Sigma:
The standard deviation of a statistical population, often characterized by the lower-case
Greek letter s (sigma). The upper-case Greek letter for sigma (S) stands for summation
(n1 + n2 + n3 + n4, and so on).
Specifications:
The required properties of a workpiece. Specifications may include the upper and lower
limits of a dimension, a texture of surface finish or anything required of a feature.
Standard Deviation:
A measure of the variation of the members of a statistical sample.

 Page 426
Statistical Process Control:
Methods of statistics used for the analysis and control of variation in a process. The use
of control charts to determine significant changes in a process. It is used to assess how
stable a process is.
Variables Data:
The measured values of a feature. Quantitative data capable of measuring (having a
value) anywhere within a given range of values.
18.6 Symbology for SPC Formulas

 Page 427

 Page 428

known as a Z distribution formula

s (sigma) is the population standard deviation. Population means all parts produced, not
only a sample of the parts. However, when more than 30 samples are taken, a sample
standard deviation (s) and a population standard deviation will give the same result.
The following formulas can be used to calculate s and s:

 Page 429

19
Letters: Common Problems and Their Solutions
The Good, The Bad and the Really Weird
Dear Jim:
I have been sitting at my work station all morning applying geometric symbology to my
parts. I never thought I would say this but I am actually having a lot of fun using the
symbols. It's as if they are wordsonly clearerto convey the parameters within which my
parts will function. Since taking your course I look at the drawings, as you like to say,
with new eyes. I see all that is there, but now I also see all that is not there. I can't believe
all the assumptions I made before, assuming those reading my drawings would somehow,
psychically I guess, be able to discern my needs. Now I simply take a little more time up
front to tolerance. This symbology, once you have worked with it for a while, makes it
easy to convey precisely on the detail drawings how the product seats and orients itself in
the assembly, using the functional datum features. I can't get over how easily it can isolate
different geometric needs of the feature and handle each individually. Early returns from
our vendors have been excellent. I was genuinely, but pleasantly, surprised at how many
do use and understand the geometric symbology. I have gotten quotes for product
production based on my new drawings that are lower. That was a real shock, but the way
the vendors explained itthe use of the specified datums told them exactly how to set up
the parts to produce them, the use of basic dimensions got rid of tolerance accumulation,
and the use of realistic tolerances that are calculated gave them confidence that the
number of in-process changes would be reduced. They also said they were now able to
quote us prices based not only on the original tolerances stated in the controls, but also on
additional positional tolerances that could potentially be drawn from

 Page 430
the size limits based on the bonus tolerancing principle. As I said, I'm having a lot of fun
and thought I would take the time to write and let you know that I agree with youit is
good stuff!
Dear Jim:
My boss keeps watching me. He even has other people watching me. Sometimes I just
can't take it and scream out, Leave me alone. Stop watching me. But this just makes them
watch me more. I'm pretty sure they are listening in on my phone conversations, too. We
are in glass cubicles, and when I get on the phone, at least one of the others is always on
the phone. Even though they move their lips, I know it's just an act to make me think they
are talking to someone, when they are really just listening in on my phone conversations.
It must be that they are all out to get me. I have a few weapons and I know how to use
them. What should I do?
Dear Writer:
Take a vacation.
Dear Jim:
Ever since I started studying the ANSI Y14.5 Dimensioning and Tolerancing Standard, I
can't sleep at night. I've got all these geometric shapes and symbols going around inside
my head. Tolerance zones grow and shift, taking hole patterns with them, getting close to
walls, almost breaking out, but then I take out a circled M and insert a circled L or S and
the walls thicken up, but then draft angle occurs and leans in on my holes, and by that
time I'm sitting up in bed with all the lights on trying to figure out what the heck I'm
going to do. My appearance is beginning to sufferdark circles under my eyes, weight loss.
I'm starting to look as though I'm wearing my bigger brother's clothesthe pants bunching
up around my waist, my belt hanging down to my knees with about a million new holes
punched in it. The reason I'm writing is I can't figure out how to sketch up the hole
alignment on my belt. I don't want to just launch into punching more new holes without a
drawing, because the results I have gotten with this no geometric interrelationships/no
sketch approach have been unsatisfactory. I just can't concentrate in meetings because I'm
always thinking about what symbology to use in this situation. What do you think I
should use?
Dear Writer:
Sominex.
I See the Light at the End of the Tunnel, But It Could Be a Train
Dear Jim:
Life is a slasher movie and I'm being chased by a guy in a hockey mask! He calls himself
my boss. Yesterday, he came into my office, slapped down a

 Page 431
copy of the Y14.5 Dimensioning and Tolerancing Standard, and told me to immediately
begin using it. I don't mean to be ungrateful to those who wrote this thing, but it might as
well be written in Swahili. Help!
Dear Writer:
I was a tool and die maker many years ago, and at the time taking design courses in
college hoping to broaden my knowledge and marketability. I remember asking my
instructors about dimensioning and tolerancing techniques and was told not to bother
them about the trivial while they were trying to make an inventor out of me. I know what
you are going through. In college they don't teach dimensioning and tolerancing, or they
teach it poorly, and on the job the boss expects you to have already acquired that
knowledgesomewhere, somehow. But, of course, you didn't, so you copy what has been
done by your departmental peers, which may or may not be correct because they are in
the same boat as you.
People are not inherently endowed with a knowledge of how to dimension and tolerance
parts, and the standard was not written as a self-teaching learning guide. Be secure
enough in yourself and the skills you possess in other areas of technical disciplines to
admit that you are not the world's most knowledgeable person when it comes to
dimensioning and tolerancing parts. If you ask around, you might even find enough other
honest and interested people in your company to warrant a request for training.
Dear Jim:
My colleagues say that a 3- or 4-day course in dimensioning and tolerancing is not
enough to make an expert of anyone, so that is a good reason not to take it. They say a
little knowledge is dangerous.
Dear Writer:
T. H. Huxley (192595) once said, If a little knowledge is dangerous, where is the man who
has so much as to be out of danger? And as someone else once said, The longest journey
begins with a single step (probably the same fellow who said, Go West, young man). Of
one thing I am certain, we know more now than we have ever known in the history of
mankind, and yet we are as babes on the verge of taking our first steps. My advice is to
learn what you can now, use what you feel comfortable with, and then learn more. One
course should be viewed as the beginning of a lifelong process. Learning is infinite.
P. S. That Huxley was a busy guy. He also said, The great tragedy of Sciencethe slaying of
a beautiful hypothesis by an ugly fact.
Explosives and a Cell Mate Named Bubbaor Training
Dear Jim:
Recently my wife left me, my dog bit me, and I smashed my car into a big sap-leaking
tree. It all happened after my company began right sizing, which

 Page 432
used to be called laying people off. The good news was I didn't get sized. The bad news
was I had to take on the responsibilities of those who did. The bosses walked by a
coordinate measuring machine one day and realized it wasn't running. One big boss
turned to a smaller boss and asked why. The miniboss said the operator had been one of
those who were sized. The large boss nodded slowly, then looked over in my direction
and cocked his head like a dog does when he's trying to figure out some new furry critter
he's just spotted. Hey, you, he bellowed. Come stand over here where the other guy used
to stand. Me? I said, glancing around. Yeah, right over here, he pointed with his big,
empty head. From now on, you run this machine. Then the herd of little and big bosses
lumbered away, leaving me to my new job. I didn't know the first thing about running
that machine. I'm a chemist, for goodness' sake! I looked at the buttons, screens, bells,
whistles, parts to be inspected, and drawings (which I couldn't read) that they were to be
inspected to and, after a time, I LOST MY MIND! I got real surly. I yelled at my wife,
kicked the dog, contemplated homicide on bosses of every size (when I should have been
keeping my eyes on the road), and generally displayed a mean and angry disposition. I
blame this whole thing on the great cosmic karma of my life. It's dumping on me and I
want to get even. Is there anything I can do (that won't involve explosives) to set things
right again?
Dear Future Potential Inmate:
Life isn't always fair. Bosses aren't always smart. The sun doesn't always shine. And the
future isn't always bright. Still, unless you seek occupancy within a state or federally run
institution, apologize to your wife, buy the dog a steak, and tell the insurance man the tree
didn't signal its left-hand turn. When companies down size or right size or kick able
bodied workers out into the street (my new term), there is often an expectation that the
workers left standing will take up the slack and do more work. Some of the specialized
positions we once occupied go awayor become just a part of what we must do. The
strong survive and often even benefit by the new skills they are forced to learn. Just make
certain you are one of them. You were not born with the skills you now possess. They
had to be learned. Unless you have an alternative job to go to, or someone willing to
reconsider your current assignment, learn to do the job. Study the books. Take courses.
Learn to run the machine, read the drawings, and learn the concepts of part measurement.
Consider this a positive alternative to bunking with a big guy named Bubba, who thinks
you're sort cute.
Shakin' Like a Dog Passin' Peach Pits
Dear Jim:
My boss is always telling me what to do. I hate being told what to do. I like to come in
every morning and stare out my office window, drink six or seven

 Page 433
cups of coffee, and smoke about half a pack of cigarettes. So what if it takes up the first
hour of the work day, I do my job! I not only schedule meetings back to back all day
long, I overlap them and double book. I think that shows dedication. All he does is
complain. He says I'm late to the meetings I attend and then leave early. He says I miss
half the meetings I schedule. Then he has the nerve to say my job is to design stuff, and I
never dothat all I do is go to meetings, most of which are unnecessary to begin with. He
says he never knows where I am and other bosses complain that I bother their workers.
He doesn't even like the way I smell. I think I should quit. What do you think?
Dear Writer:
I think your boss would be eternally grateful if you did.
Dear Jim:
Why do we need geometric dimensioning and tolerancing anyway? No one understands
it. Everyone uses it wrong. It causes endless arguments about what it means. Vendors
charge us more when we put it on our drawings. It takes years to be any good at it.
Drawings without it look simpler. I think it's big pain.
Dear Writer:
Why don't people use GD&T? No one really understands plus and minus tolerancing with
its endless ambiguities and reams of drawing notes. There is no way a part can be defined
well without the use of a universal standard that has geometric symbols and datums.
Everyone interprets the tolerance zones and set-ups differently for plus and minus
toleranced hole locations. Tolerance accumulations abound with plus and minus
tolerances and are eliminated or easily calculated with GD&T. Vendors often charge less
for a part that properly uses GD&T because of bonus tolerances and nonambiguous
origins of measurement. Drawings may not look simpler with symbology, but are actually
much clearer in intent than without it. Drawing notes are minimized with GD&T. The
parts are much better defined. Tolerances are based on calculations, not fear, and
therefore, they are usually as liberal as functionally possible. I think drawings that don't
use GD&T symbology are painfully devoid of information vital to the manufacturing of
high-quality products.
Dear Jim:
I'm deep in high cotton and happier than a pig in slop. I just got a contract to teach a
thousand people geometric dimensioning and tolerancing. And it's so close to my
homestead I can mosey on home at noon for biscuits and red eye gravy. Can you lend me
a video on this subject? I been cogitatin' on this, and since I never done geometric
anything before, I figure I ought to study up on it before I commence to doin' it.
Dear Writer:
You're in something deeper than high cotton.

 Page 434
YO, Jim:
Lookee here. I got no time to be learnin' new junk. Aren't there computer programs that
just spray this GD&T stuff on my designs? I'm way too creative to spend my precious
time studying symbols. I spent my time in school, paid my debt to society, and am ready
to rock, but my boss is constantly screaming at me. I think he should chill and leave me
to create.
Dear Writer:
I don't know of a computer program that just sprays on the most appropriate geometric
symbology. But I'll keep my head to the ground and will contact you should anything
crawl in my ear.
Dudes:
Like what is GD&T? It sounds really cosmic. Does it have to do with channeling? I know
this chick does a transformation into a 3000-year-old warrior from Atlantis who has lived
on in the winds of time and space. He's called Gaga Din Tin or GD&T for short. Billy Jim,
Mississippi
Dude:
They let you talk like that in Mississippi?
My Boss Just Died and I Feel Fine
Dear Jim:
I stand at the threshold of a new world order. My horizons are broadened and my future
is bright. I have been chosen as the new manager of my department. I've had the job only
a month and already half of my 35 staff members have threatened to quit because of the
pressure I have put them under. Little do they realize my magnificence, my sheer genius!
Sure, they have to come in early and leave late without compensation. And yes, they have
double the workload compared to what the old simpleton boss used to give them. But, I
have a greater good in mind my own. I think it's humorous that employees really believe
their boss cares about their well-being, values, and goals. I tell them whatever they want
to hear to get them to do the job. I am planning on stepping on every neck available to
climb my way to the top. And, I'm going to get there if it kills them. What a wonderful
world I live in, that these peasants are there for me to use. Besides, I think watching me in
action will finally teach them what the real world is all about.
Dear Writer:
I think I used to work for you.
Dear Jim:
Life stinks! My boss says I have to be trained to dimension and tolerance parts and how
to communicate better with other departmental personnel. I'm a

 Page 435
college graduate, for goodness' sake! Why in the world would he think I need more
training? I've been in school my whole life. I say if they couldn't teach me all I need to
know in almost 20 years of school, then the system has failed me and I should sue. But I
don't believe that is true. I do know all I need to know. It's my boss who needs to be
trained to show the proper respect for my degrees. He doesn't have nearly the formal
education that I have. Oh sure, he's been on the job for as long as I've been alive, but all
that means to me is that his formal education has become obsolete. He should be learning
from me! As a matter of fact, he asked me to write you this letter to inquire about the
availability of training courses, but just having written it has convinced me more than
ever that I need to march right back into his office and tell him just what I think of him. If
he wants someone to dimension and tolerance parts, let him hire someone else to do it
and let me do the things I was educated to do. As far as communicating and getting along
with people from other departments, forget it. They're all a bunch of stupid fools anyway.
Dear Writer:
I think you used to work for me!
Bad Meatand No Mayo
Dear Jim:
Why does everyone hate me? My heart's in the right place. I tell the truth about
everything. I want to improve my company, its products, and the people with whom I
work. So, when I see something is wrong, I tell someone about it. For example, when my
co-worker started wearing too much makeup, I asked her if she was running away with
the circus. When my boss started wearing too much cologne, I asked him if I could keep
the door to his office open during my performance evaluation because I was starting to
lose consciousness, and when he refused I projectile vomited all over his desk. More to
the point of this letter, though, when the people in my department said they didn't want to
learn or need to apply geometric tolerancing to our drawings, I told them they were so
ignorant of the topic and its importance that I thought their parents must be brother and
sister. I'm just trying to do the right thing. Why am I suddenly getting death threats on my
E-mail?
Dear Writer:
Lighten up. We might all be dead tomorrow. At least, you might. Sometimes tactful is
better than truthful.
Dear Jim:
The company I work for has begun to create design drawings that do not reflect the
functional requirements of our products, but rather simply document the manufacturing
procedures that are used to produce them. The problem is that

 Page 436
all the parts pass the drawing requirements once produced and inspected, but none of our
products actually work. The people in charge of the design-for-producibility effort have
taken over. Any requests for drawing changes must go through these people and they
quickly reject any efforts to design in the functional requirements. We are shipping
replacement parts that couldn't be wedged into the product with a crowbar and
sledgehammer. And, even if we could get them in there, they would just squeal like a
wounded animal when we started the mechanism up. Help! Our stock prices are plunging
into negative numbers. People are being laid off, the furniture is being sold, and the
attrition rate now includes leaps off the company water tower! Shouldn't we not only
worry about how the parts are produced, but also that they have at least a remote chance
of working?
Dear Writer:
Before you take the leap, try writing memos (signed or not) to the powers that be that
outline the problems and your suggested solutions. They may listen. They may not. There
is no predicting how a group or an individual will react when his or her authority or
judgment is questioned. The secure, intelligent ones will be open to suggestions for
improvement. The insecure political beast may try to fire bomb your officeor at least
destroy you as a threat to them and their quest to take over as captain of the sinking ship.
Good luck, but remember that there is no proof to the saying that they can kill you, but
they can't eat you. Political bullies have a voracious appetite.
Dear Jim:
We have begun to put a note on all our design drawings that says we are in compliance
with the ASME, ANSI-approved, Y14.5 Standard on Dimensioning and Tolerancing.
That's all. Just the note. We don't actually use the thing or the symbols or concepts
contained in it. But my boss says all we need is the note to show everyone that we know
what we are doing. He thinks it will make it easier to get our ISO 9000 series certification.
Don't we actually have to use the stuff, not just say we do? Won't people know when they
look at the field of the drawing that we don't have a clue as to what a well-documented
design drawing is?
Dear Writer:
One would hope your company would document well, not just say they do. But nothing
surprises me any more.
Dear Jim:
I just want you people to know the living hell I am forced to endure every working day
of my life because of geometric dimensioning and tolerancing. I don't have the time for
this crud. I have real work to do. Just because we've had some difficulties recently
regarding the quality of our products, my manager wants me to head up an effort to
redraw all our components. She wants me to re-

 Page 437
define them using the symbols and principles in the Y14.5 and Y14.5.1 standards. I have
read over these documents and find the material therein offensive and difficult to
understand. I believe applying this garbage to our drawings would be time-consuming
and tedious. The people who have written this material are trying to change our whole
way of defining our parts. They must be arrogant and loathsome individuals indeed, to
think they can force us into doing such a thing. If I ever meet one of these people, they
will rue the day they became involved in the writing of these standards.
Dear Writer:
You mailed this letter in an envelope that had your company name on it. I did a little
research and found out you folks haven't made a profit in the last 5 years. Since you
wrote me, I'll offer a little advice: If your hind end is on fire, don't shoot the man carrying
the water bucket. He just may be trying to help.
Dear Jim:
My life has not been easy since people found out I am an ANSI savant. It was discovered,
in a workshop held at my company over a year ago, that I have real aptitude for GD&T. It
happened during the seminar. Whenever the instructor asked a question, I knew the
answer. The whole process of geometric definition was logical to me. I knew what he was
going to ask to best define the products before he even asked it. Ever since then, my co-
workers have looked at me funny and treated me like some kind of freak. They come
around all day long and ask me questions to help with their drawings and part
tolerancing. They try not to be seen talking to me, so my E-mail is always full of messages
seeking advice on this stuff. Everyone tries to trivialize the importance of what I know,
but everyone seems to need my help. All of the time! I feel isolated by this curse. No one
will sit with me at lunchtime. They talk behind by back, and call me Geometric George.
It's not my fault I have this aptitude. Should I pretend to be ignorant? I've practiced
cocking my head from side to side, like my dog does when I ask him a question. In the
mirror at home, I try to take on the look of shock and fear one sees on a deer's face when
it gets caught in someone's headlights. Still, I don't feel comfortable with this empty-
headed charade. Help me. I've taken up hiding out in the men's room to avoid ridicule,
and I can only do that for so long. It smells bad in there!
Dear Writer:
I'm afraid I'm not going to be much help on this one. I'm known as Geometric Jim.
Maybe we should start a club.
Low Foreheads, Bad Instincts, and Hangdog Looks
Dear Jim:
I've got a real problem at work, and I'm afraid it's going to get me canned. It's my boss.
Don't get me wrong. He's one of the sweetest, brightest guys on

 Page 438
God's green earth, but I can't be in the same room with him since his midlife crisis started.
There's so much I want to tell him but don't know how he'll take it. I want to tell him he
ought to push his new hairpiece back farther on his head instead of wearing it like an
overseas cap. That it should match his swarthy skin and eyebrow coloring better than
bright yellow. That piling hair up that high makes it look like a vanilla Dairy Swirl. That
when his mustache grew in slow and thin, coloring in a one-inch-by-six-inch area below
his nose with black magic marker wasn't the answer. That his little chin beard looks like
armpit hair. That purple nylon socks don't go with Roman sandals. That his pants always
ride up too high in back because his suspenders are way too tight. Every day it gets
weirder. I need help, bad!
Dear Writer:
Granted, the behavioral quirks you describe may seem extreme, until you factor in the
trauma of a new job search. Unless you're ready to start writing a rsum, you may wish
to consider that what constitutes appropriate business attire is often best left to personal
taste (or the lack thereof). My advice is to hold thy tongue, shuck ideas about rambling on
down the road, and burrow your head into the job like a tick into a hound dog's hide.
Dear Jim:
The people in my company are so ignorant about geometric tolerancing that they think
runout refers to a wildcat labor strike. Product definition around here is so entirely
comprised of caveman techniques they might as well scrawl the drawings on our walls
with a flame-charred stick.
Dear Writer:
Give a copy of this book to all who can read. Eventually, those of protruding brow and
dragging knuckle may come around and seek evolution through training.
Dear Jim:
Recently, I have noticed that some of the basic skills we have taken for granted in the past
are missing in our employees. People cannot add or subtract without a calculator. They
often cannot tell you which of two numbers is larger if each has a different number of
decimal places. The process of using simple logic to discern whether a calculated answer
is even in the ballpark is foreign to them. What is going on?
Dear Writer:
This is truly the computer/calculator generation. I was behind a gentlemen close to my
own age who was returning a rented car when I overheard the young man behind the
counter attempt to charge him $238 for gas. The car he had

 Page 439
rented had been in his possession for only one day. We older guys spent the better part of
the next half hour convincing the youngster he had his decimal point in the wrong place
and only $23.80 was owed. The guy could have driven to the Arctic Circle on $238 worth
of gas, but the kid had no concept of any of that. He just kept pointing to the calculator
and smirking at the obvious ignorance of us old guys. To many people, the computer or
calculator is the final word and logic has no place in a discussion. Whether the computer
or calculator is in a store or part of CAD, CAM, CNC, or CMM operations, they don't
understand that garbage in isn't gospel out.
Dear Jim:
In my company, mechanical engineering types are strictly coach passengers. Our
company's focus is on chip technology and software engineering. I am one of those
shadowy figures strapped to a milk crate and paid in bags of peanuts, while the chip techs
and soft injuneers gargle champagne and scarf fish eggs. Those in first class fail to see the
importance of a well-documented engineering drawing with standardized geometric
controls. Help!
Dear Writer:
Most companies that begin with an electrical, chemical, or software base have similar
problems. An elitist faction forms that fails to recognize that these gadgets not only often
need documentation of their own, but have to be housed or packaged in or on something
that requires standardized engineering drawings. These drawings must be correct, legally
defensible, and uniformly interpretable for purposes of manufacture and inspection/test.
I have seen many companies assign the task of creating these drawings to people not even
familiar with orthographic projection, much less acceptable dimensioning and tolerancing
techniques. The results are commonly comic tragedy. Before beginning to correct
dimensioning and tolerancing, I have often suggested the drawings first must all be
redrawn using standardized, interpretable views. This is expensive and would easily have
been avoided had the powers that be recognized that all facets of the product design are
important enough to be done correctly the first timeby trained and qualified professionals
operating in their own areas of expertise.
Here is one of the newer attention-getting arguments for standardized drawings: the
companies not using standardized product drawings and manufacturing and quality
procedures wishing to do business in Europe in the future (I was told by European
delegates at the ISO meetings) will not be welcome!
Dear Jim:
I recently moved to South Carolina. I thought the drawings at some of the companies in
the Northeast were bad, but by comparison with what I have to deal with here, they were
technically advanced. So far, having reviewed about 50

 Page 440
machining and assembly drawings, I've seen nothing but chaos and confusion. Any
attempts at geometric control have been in note form and entirely wrongin fact,
incomprehensible. What has become of our educational system?
Dear Writer:
I was recently asked by a mechanical engineering professor to speak at a major university
about geometric dimensioning and tolerancing. He suggested that about an hour should be
enough to properly educate the students about dimensioning and tolerancing techniques
and geometric symbology use. That certainly points out one of the things wrong with our
educational system: topics vital and complex are often treated as unimportant and
simplistic.
I declined his request, but I don't believe he would have liked my message. It is one I
have repeated many times. Without a thorough understanding of how to express oneself
using the concepts and symbology contained in the Y14.5 Standard on Dimensioning and
Tolerancing to convey functional product requirements and that suggest manufacturing
and inspection procedures, you are illiterate in the language of the profession you have
chosen.
Dear Jim:
The example included in your last newsletter depicted an MMC modifier in the feature
control frame for a threaded hole. As the locating feature of the hole is the pitch diameter
of the threadsand not readily measurable, it has been my practice to specify RFS for
positional tolerance on threaded holes. I believe that if a feature of size is to have an
MMC callout, that size has to be measurable and the correct bonus tolerance determined.
Please advise.
Dear Writer:
This is a portion of a broader argument. The argument is whether a design drawing
should include geometric controls that describe how the part/feature actually works, or
rather what I can measure given my environment. The answer is, hopefully, both.
However, the main job of a person giving geometric controls to a part is to describe how
it functions. I always say designing is the creative aspect of the job, while tolerancing is
the act of describing what the product physically does when you try to assemble it or
operate it.
When a screw is inserted into a tapped hole, there is some playslop, if you willduring
assembly. This slop between the screw and tapped hole is the result of the pitch diameter
of the screw being smaller than the pitch diameter of the tapped hole, and this slop or
play helps us to more easily assemble the parts. Hence, there is more slop if the class of fit
is a 1B instead of a 2B, and even less for a class 3B threaded hole. The larger the pitch
diameter of the hole, the more play/slop we derive.
If the play is utilized to help us assemble the products, it should be allowed as additional
tolerance during the inspection procedure. This is bonus tolerance.

 Page 441
If you can't measure the bonus tolerance, I, as the dimensioning and tolerancing engineer,
will suggest gaging procedures that will automatically accommodate the allowed
minimum bonus tolerance available for that hole.
If you wish not to utilize the bonus tolerance because you can't calculate an exact number
of thousandths of an inch or portions of a millimeter, that is your decision. But please
don't ask me, as merely a describer of physical phenomena, to say it doesn't exist by
putting in an RFS modifier instead of an MMC. To me, the RFS symbol says the hole
contracts to meet the screw sizeand that is rarely true.
In my way of visualizing the act of screw and tapped hole assembly, I don't really view
the additional play (as the hole's pitch diameter grows) as a tolerance zone growth so
much as I see it as allowing me to use the available play to push the screw over into the
projected tolerance zone of the hole. And in inspection, I see the gage screw as entitled to
the same advantage. Because, however it is viewed, additional tolerance exists even if it is
small and we can't ascribe a number to it. We are tellers of physical truths as tolerancing
engineers. A feature control frame is nothing but a sentence in the form of symbols
strung together like words and phrases. If we lie in a geometric control, the result is
usually either: (a) parts that work are rejected, or (b) parts that don't work are accepted.
Dear Jim:
In the ASME Y14.5 standard, when it discusses the possibility of a rocking datum feature,
it suggests the feature should be adjusted to an optimum position to strike the datum
plane. Just what does that mean, exactly?
Dear Writer:
This is the topic we have been studying in ASME Y14.5.1, the Committee on
Dimensioning and Tolerancing Mathematics. What happens when a surface rocks on its
high points while seated on a machine table, a surface plate, or in the assembly is that a
candidate set of datum planes is created. This candidate set of datum planes are all
different because they are struck at different positions as the part rocks. All are taken
from surface high points. It is just that the high points contacting the table change as the
part continues to rock.
It is up to us to choose a particular datum plane within the candidate datum plane set that
is best for a product. Hopefully, it is one that assures us the part, if used, will function. It
used to be thought by many that the line adjust to an optimum position meant just shim it
up until it doesn't rock any more. But that hardly gives a unique answer (depending on
how the shims are used), nor one that is always how the product functions (for example,
how it is bolted into the assembly). So now we show a mathematical equivalent of adjust
to an optimum position that incorporates better criteria based on what is optimum for the
product functionality.

 Page 442
Slip Sliding Away
Dear Jim:
I saw a guy with a bumper sticker that said, Ask me about GD&T. Thinking he might be
an expert or at least another interested practitioner of (G)eometric (D)imensioning and
(T)olerancing, I followed him. Every time I figured he was about to stop, he would take
another dusty side road or go down a dingy alleyway. Before I knew it, it was dark and I
was 50 miles outside of civilization. He finally pulled into the parking lot of a really
sleazy-looking joint called (G)irls, (D)rinks, and (T)rouble.
Come to think of it, I was right after allwe did have a lot in common. Still, finding this
out took us on a 3-day escapade, most of which I don't recall (although I'm certain some
explosives and sirens were involved somehow). Anyway, I am left with the faint
recollections and physical evidence of being (G)ross, (D)isorderly, and (T)attooed. Know
any way to have the words U.S. Amy removed from one's chest?
The tattoo artist apparently never won a spelling bee. My problem is that I've never
known anyone named Amy and, besides, even if he had gotten that word right, what I
seemed unable to remember at the time was that I was in the Navynot the Army.
Dear Writer:
(G)ive (D)iscipline and (T)herapy a try. Then, (a) date only women named Amy or (b)
consider a battleship as a cover-up tattoo.
Dear Jim:
This morning my boss stormed into my office, kicked out all my friends, started shaking
his fist at me, and got all red in the face. He yelled, From now on, it's my way or the
highway. I play the tune and you dance to it. If you step over the line one more time,
youll be wearin' a smaller shoe size. Any more of your lollygaggin' and I'll kick you in the
bee-hind so hard you'll have to unbuckle your pants to change your mind. My question is,
what's his problem?
Dear Writer:
He's suffering from OC or over clich. It is a common malady among bosses. Still, if I
were you, I'd watch my P's and Q's.
Dear Jim:
I recently attended one of your workshops on geometric dimensioning and tolerancing. I
realized that in college when I was studying to be a mechanical engineer I was not taught
how to properly dimension and tolerance parts. They made me a good enough inventor,
but I'm mad as hell that they didn't consider, as an essential part of the curriculum,
providing the students with a thorough understanding of current dimensioning and
tolerancing practices and geometric symbology.
Certainly, most of the managers I have worked for since graduation have

 Page 443
just expected me to know it. Now that I do, I find myself noticing just how few of my
colleagues are competent in its application. About all they seem to know is the name of
the symbols. We should all sue.
Dear Writer:
After they attend a thorough training session, I have had many participants tell me they
look at engineering drawings, documentation, and interdepartmental interaction methods
with new eyes. I'm not certain litigation is the answer to the ills of our education system,
but maybe those alumni with close ties to their former universities can exert enough
influence to get them to consider changing and improving courses of study to stress the
importance of this topic.
Dear Jim:
I am a design engineer. After attending your workshop, I attempted to get to know the
manufacturing procedures better by visiting the shop floor. It was noisy, smelled bad, and
I got oil on my new tie with the flower pattern. The people looked at me as though I was
spying on them, and one of them spit a load of tobacco at my leg. Are you sure this stuff
works?
Dear Writer:
It does. But as with anything worthwhile it takes time. My advice is to toss the tie, practice
your side shuffle, and keep up the vigil. Introduce yourself, ask a lot of questions. Tell
him/her you draw the parts and want to see if he/she has suggestions as to how they could
be made easier. Find out if they understand the drawings you have created and, if not,
take the time to explain. You are fortunate. Some don't have ready access to the shop
floor.
Dear Jim:
My Grandmother works on the milling machine next to me. She constantly swears a blue
streak, drinks potato mash whiskey (which she brews in her basement), and chews
Redman tobacco. She has no teeth and her nose runs constantly. Every time I turn my
back, she walks over and wipes her drippings on my shirt sleeve. The other day a nice
young design engineer was walking through the shop and she spit a stream of brown
juice on his pant leg. I can't take it any more. Should I turn her in to my foreman (a nice
little man but no match for Granny) or to my mom (who drives a truck and arm wrestles
semi-professionallyand could pound her if she got huffy)?
Dear Writer:
I make it a policy not to get involved in domestic squabbles.
GD&T Questions I Can't Take Any More
Q: I don't have to use this, do I?
A: At least let me show it to you.

 Page 444
Q: How many people really use this stuff?
A: A lot. (The actual answer to this question is much more precise.)
Q: How is that poor guy out in the shop supposed to understand this? (Translation: How
am I supposed to understand this?)
A: Learn it.
Q: This standard doesn't apply to (Fill insee list below), does it? List (choose one): plastic
parts, sheet metal parts, real big parts, simple parts, real small parts, my parts.
A: Yes.
Q: Where's my part in this book? I don't see my part!
A: Squint.
Dear Jim:
At my company our drawings are atrocious. I came here from a company that thought
tolerancing was important. Here, all we do is document the manufacturing procedures and
the tolerances they hold. If the tolerances on the drawing are violated, all we do is change
the drawing to what has been produced. We even call it reverse engineering. Well, maybe
it doesn't matter anyway. After all, the original tolerances for the product weren't
calculated, just pulled out of midair. When we tolerance parts, we never take into
consideration what tolerances and worst mating conditions are on the parts in the
assembly that our part interacts with. Yet, in our commercials on television and in
magazines, all they talk about is the space age tolerances we hold. Why is it that
tolerancing seems to be a great marketing tool, but a joke on our drawings and on the
shop floor?
Dear Writer:
People do and teach others to do what they know themselves. They often make fun of or
downplay the importance of the things they don't know how to do well. In college, I
remember taking courses from professors on how to design parts and assemblies. I took a
bunch of these courses and always came away feeling cheated. So finally, I began asking
the questions that had been bothering me. Professor, I would say, when I'm not in school,
I work as a journeyman diemaker. I make parts all day long, and I've never been able to
make even one perfect part. So, could you tell me please, what about tolerances? In your
class we never have calculated tolerances, or even discussed how to accurately tolerance a
part or an assembly. The professors always answered me in the same way. We don't have
time for that. This is only a 4-year program. We must make the most of the time we have
together. We have bigger fish to fry. We are making a designer, in fact an inventor, out of
you. You should appreciate the information we are giving you.
Appreciation was never my problem. I loved to learn. But I knew there was a major piece
of the puzzle missing. Finally, from what I learned later, I re-

 Page 445
alized that they were teaching what they knew. I learned that designing products and
tolerancing them took completely different skills, and that just because you knew how to
do one, didn't mean you knew how to do the other. They were good at teaching and even
actually doing the work of design. But they couldn't have toleranced themselves out of an
ankle-deep mud hole.
On the job, I found out the problem was the same. People didn't know how to tolerance
anything. Furthermore, they didn't care about it. They thought of it as grunt work. They
were too important to worry themselves over lesser skills than they provided daily. They
thought that what they did was enough. But it wasn't. Products failed, wouldn't assemble,
blew up, killed people. But they all blamed others. Its the fault of that other guy, always
the other guy. The designers blamed manufacturing, saying they all needed directions to
put their shoes on the correct feet in the morning, that they should be more intuitive about
the needs of the product. They said a design drawing was, at times, only a loose set of
guidelines. Besides, the tolerances didn't matter. If they had just made better parts, they
would have worked fine. The manufacturing guys blamed it on a bad design with
conflicting requirements, and they blamed the people from inspection for not inspecting
the parts under the same conditions in which they were to be used. The inspectors blamed
it on the designers, saying they inspected it according to the drawing requirements, but the
drawing requirements were so ambiguous, they could mean anything, and the tolerances
so crude they might as well have been scrawled on a cave wall with a charred stick.
And yet, when asked to improve their technical abilities and their team-work and
concurrent engineering skills, many of them grimaced. Do we have to? they asked. We
have far too much work to do. We have meetings, and deadlines, and E-mail to respond
to, and fax messages to send. You know, the real work, in the real world. Are you really
going to force me to sit through a training class? I have a job to do! I'm trying to keep it
real. And the customer responds, Yes, you are trying to keep it real. Real stupid. Real
dangerous. Real low quality. I just wanted to use your product, not be killed by it, not to
have it explode in my face or fall apart in my hands. I've seen your commercials. What
happened to those space age tolerances? And the workers respond, Oh that's just
marketing. We never talk to marketing. And the customer says, Apparently you never talk
to manufacturing either. What's wrong? Is it too dirty down there on the shop floor? Are
you afraid to ask for their advice, afraid they may find out you don't know everything
there is to know in the whole world, afraid you'll look too human, afraid you will be
thought of as fraternizing with the enemy? What is it? Are you afraid they may ask you
where you got your tolerances and you'll have to show them your dart board?
I guess the truth is not to be found in a television advertisement or on a billboard
someplace. The truth is found in the products, and in the words of the consumer, and
unfortunately, sometimes on the nightly news.

 Page 446
Dear Jim:
Is geometric dimensioning and tolerancing the answer to all my company's problems as
they pertain to design, manufacturing, and inspection?
Dear Writer:
No, of course not. It is the answer to many problems though. Not only does it give
unambiguous design criteria, allow tolerance analysis to be easy, parts to mate in
assembly with interchangeability, and make the inspection procedure meaningful by
allowing workpieces to be measured in the same way as they will function, but it also
gives a person an understanding of the important dimensions, tolerances, boundaries, and
features that affect quality and product endurance.
Among the things that interest me are the way we look at our everyday jobs. Some do
what they are told and don't think much about it. Some don't do much of anything at all.
Some do what they are told and grow to believe in the procedures, whether or not they
have functional merit, even to the point of becoming zealots and proponents of the
procedures.
Some are knowledgeable beyond the programs they are told to institute and genuinely
want to improve them. Sometimes these employees are seen as rebellious, but nothing
could be farther from the truth. There is no rebellion in their hearts, only a desire to do a
good job. They study what they are doing and look for ways to improve the system and,
consequently, the products and the company. These people are the curious visionaries a
company needs to create a better way and survive in a competitive world.
Geometric dimensioning and tolerancing is a tool that is capable of giving insight to the
curious. It shows how to calculate instead of guess, to deduce rather than mindlessly
perform. I have always believed that the people in the technical trades those likely to be
reading this occupy are capable of performing great feats of logic. But logic is only as
good as the knowledge and experience of the one who employs it. To reason things out,
we need a good basis of knowledge. For a product line, this basis has to include a piece
of knowledge about practically every aspect of a system. We must make it our business to
learn about the procedures our company and their vendors use in design, manufacturing,
and inspection. We must, in fact, learn about all the job categories that relate to or input
information that directly affects the creation of the product. Otherwise, our logic will be
flawed. A piece of the puzzle that we need will be missing when we attempt to reason out
the best course of action.
Geometric tolerancing, when applied correctly, cuts across all job categories. It makes the
tolerancing engineer create datums (origins of measurement and part orientation) that take
into consideration how the part fits into the assembly and how it will best function, which
lends information to the inspector as to the best way to determine whether or not parts
produced will, indeed, perform as

 Page 447
needed. GD&T allows the tolerancing engineer to calculate tolerances and do tolerance
analysis prior to making the product and it allows the inspector to determine whether or
not these tolerances and boundaries have been violated or will soon be violated if the
same course of manufacturing is continued. It gets the people in manufacturing involved
with the inspectors, seeking their input as to measurements taken and how to change the
manufacturing methods to improve the part and its measured features. It links the
designer, tolerancing engineer, manufacturing people, and inspectors in a never-ending
discussion as to how to improve the product, its tolerances, the manufacturing methods,
and the inspection procedures.
Part of this is to be found in the standards of the world on dimensioning and tolerancing,
such as the ASME Y14.5M-1994 standard. But that is only the beginning of the process.
The ANSI standards on documentation (Y14) and measurement (B89) and their
counterpart International (ISO) standards provide basic information on what tools are in
the box, but the individuals and teams of users of these documents must provide their
knowledge of their products and environments to best utilize these tools. The standards
are only a small part of the over-all formula that makes a company and a product line
successful. But, it is a part that is essential, and makes the other subsequent components
of success possible. Standards are not the answer to your problems, but they are the
beginning of the answer. The programs you employ, whether GD&T or SPC or any of a
long line of programs, will only be as successful as the way they are implemented allows
them to be. They need your judgment, your knowledge, your experience, and your
caution. You need to take a good idea and make it better. And then take that idea and
improve on it. And so on. Progress is often possible because of people like you and me
taking all of the knowledge and experience and ideas of those who came before us, and
adding our little pieces of reason and logic to make the program fit our situation.
No program is going to be perfect for your system without your help. We stand on the
shoulders of the giants who came before us. But all of us have the ability to, one day, be
considered one of those giants whose shoulders are stood upon by those yet to come.
Dear Jim:
My boss wears really big shoes. Don't misunderstand me, his feet aren't big. In fact, he is
a rather small man. It's the shoes that are big. They are so big that they curl up on the end.
It makes him look like he's wearing skis or elf shoes or something. In fact, his nickname
(not to his face) around here is Ski Boots. He not only doesn't know anything is wrong,
he is so proud of his shoes that he wears pants that are far too short for him in order to
emphasize the shoes. Also, he isn't a naturally graceful man. He would have trouble even
if he wasn't wearing big shoes. But with the shoes he's always stumbling around tripping
over

 Page 448
things, even to the point of injuring himself and others. Bandages have become an
expected part of his attire. When we enter a meeting with people who have never met him
before, they are shocked and amused to the point that we never get anything
accomplished. One guy recently just came out with it. He said, What's with the shoes?
You lose a bet or something? My boss was offended. He said, These shoes cost more than
your car is worth, pal.
What should I do? This is getting ridiculous.
Dear Writer:
This is a tolerancing newsletter. I don't know how to solve big shoe problems.
Dear Jim:
I sit in an area without natural lighting. There are only fluorescent lights that sit in rows
above my head and suck the energy from my body, and I think they are slowly draining
my life away. Every day that I look in the mirror I have less hair and fewer teeth. My skin
is gray and hangs loose on my bones. I am wasting away into nothingness. But, today
something happened! As I stared at my computer screen (which radiates a greenish glow
that seeps into my brain and does God knows what), I noticed a ray of sunlight break
across my hand (as it rested gently on my mouse). It seemed to defy all physical laws as it
crept around and through the hundreds of other cubicles to reach me. My skin came alive
at its warm and sensuous touch. My flesh grew pink, and then rosy with actual blood
pulsing beneath. I could see my heart (which I long ago thought turned to stone) beat in a
rhythm worthy of a symphony. I realized at that moment that this job was killing me. I
needed to get out into the world. What light, through yonder window breaks? I asked.
Could it be nature's message to go forth and begin anew? To start again with different
priorities and goals? To reassess the values of my existence and finally realize I have
worshipped the wrong gods? To know that moments spent with family and friends are
what really matters? To appreciate every flower and every sunset as a wonder of
existence? To seek serenity and joy? To explore the beauty that I know abounds in art and
music and nature, but that I have always been too busy rotting away in this tomb to even
think about? All these things passed through my mind as that single miracle shaft of light
ran across my skin like it was the finger of God. And thenand then I came to my senses
and finished my report! It was a stupid idea anyway. A guy's gotta eat. OOPS, I just lost
another tooth.
A Pragmatist in Pocatello
Dear Prag:
I didn't know pragmatism could be so depressing. Next time you get an insight, either act
on it or keep it to yourself. Excuse me, I have to go and find a field of flowers I can run
barefoot through.

 Page 449
Dear Jim:
I have a problem with the measurement of holes and shafts that are positioned to a pattern
datum (a pattern of features used to construct one datum axis). It seems my coordinate
measurement machine (CMM) people say that they can't do what the drawing requirement
means. They say that a receiver gage could be constructed to do the job, but if the CMM
is used it can't be done. They say that no matter what appears on the drawing, they are
going to pick one of the holes to measure from and one other hole to orient to. I don't get
any respect around here. I feel like the Rodney Dangerfield of dimensioning and
tolerancing. Take my wife Please! Oh, wait a minute, I think that was Henny Youngman.
Dear Writer:
What it is, is old. As far as the technical question is concerned though, using a pattern of
features to construct one datum axis (often known as the central axis of the MMC concept
virtual condition boundaries) can be a difficult measurement task for a CMM. Yes, they
are correct in that a receiver (functional) gage would do a nice job, but it would collect no
variable data. But to substitute, as they want to do, also has its problems. The
measurement of subsequently controlled features from only one feature within a pattern
datum is subject to the error that one feature may experience in its relationship to the
others within the pattern. In other words, there is an accumulated error with which to
deal. This error is not one that is easy to compensate for either. One can use a CMM
probe on a bunch of features within a pattern and ask the computer for an average axis. If
the actual pattern of features (the real, flawed surfaces) is probed, the error can be gross,
depending on the feature configuration, orientation, and location errors. However, if the
part is mounted on a fixture, and the fixture is probed, the error is reduced by the amount
that the fixture is better than the pattern it fixtures (hope-fully 1020 times better, according
to the gaging standard's recommendation of 510% of the part tolerance as gage or fixture
tolerance). This fixture idea is better because the fixturing pins represent the virtual
condition boundaries (MMC of the hole minus the geometric tolerance).
Sorry about your bleak circumstances, but if the inspection folks begin ignoring the
design requirements and make up their own requirements for part measurement, you will
experience a karma akin to being sucked down a black hole in space that spits you out
somewhere in the sewers of New York City. I believe you have only a few good choices.
You could change the drawing requirements to reflect what they are going to measure
anyway (make one feature a datum feature for location and use another feature as an
angular orientation datum feature), but write a drawing note that says you are going to
fixture to the entire pattern for all subsequently controlled features for measurement
purposes. This, in a way, employs the concept of simultaneous gaging requirements. It
says that since the holes in the pattern not chosen as the datum features are controlled

 Page 450
to the same exact datums as the other subsequently controlled features, they must be
treated as a single pattern of features and either gaged as one or fixtured in one set-up for
measurement purposes. Another option is that you could keep the drawing requirement as
is and employ the use of the fixture to lessen the problem of accumulated error (by either
measuring from one post in the fixture and orienting to another or by probing all fixture
posts and asking the computer for an average datum axis). I know these solutions aren't
perfect, but sometimes measurement is just doing the best you can on a particular day,
under a given set of circumstances.
On a different note, you are correct in your overall assessment of life. We go through it
being pushed around, trying to keep what little dignity our morbid childhoods left us
with. Then at lunch one day, we choke to death on the chicken salad and our co-workers
leap over our twisted, lifeless corpse, trying to be the first to apply for our jobs and office
space. It's hard being cool when rigor mortis forms your hand into a claw and your
friends begin to use it as a back scratcher or drink holder to amuse the rest of the
lunchroom crowd. Have a nice day!
Your cheerful Pal,
Jim Meadows
For more letters, common problems and shared wisdom, see the quarterly issue of the
American Society of Tolerancing Engineers (ASTE) Tolerancing Engineer Newsletter at
http://www.geotolmeadows.com.

 Page 451

Index
A
Absolutes, 302
Accuracy vs. stability, 410
All Around, 57, 107
Allowed vs. actual deviations from true position, 13, 220, 223, 224, 225, 227, 229
Ameoba shaped parts, 58, 61
Angularity, 102
axis, 104
line element, 104, 105
surface, 102
Angular orientation, 174, 294
Arithmetic mean, 406, 421
Attribute data, 408
Average dimensions, 87, 208, 406
B
Basic dimensions, 10
Bonus tolerance, 13, 26, 212, 217, 220, 226, 229, 231, 440
Boundaries, 2, 12, 150, 153, 227, 235, 402
Boundary vs. tolerance zone, 13, 227, 401
C
Calibration, 344, 347
Capability analysis, 411, 417
Centerplane controls, 139, 143, 187, 339
Centers, 92, 93
Certification, 343, 347
Chain lines, 146, 195
Circularity, 79, 132, 305
Coaxiality, comparison of controls, 54, 162, 166
Coefficient of expansion formula, 258, 326
Comparison of measurement equipment, 337, 375
Comparison of symmetry and position, 141, 142
Comparison of 2-D vs. 3-D controls, 162
Competing in a global economy, 319
Components of a feature control frame, 19, 21, 23
Composite tolerancing
position, 154, 158, 188, 352, 364
profile, 109, 114
Composite vs. two single segment controls, 109, 154
Compound datum features, 138, 294, 306
Concentricity, 54, 135, 163
Concurrent engineering, 55, 309, 320, 387
Conicity, 117, 134

 Page 452
Converting from plus and minus to true position tolerancing, 7
Coordinate Measurement Machine, 149, 179, 182, 337, 344, 375, 449
Coordinate Measurement Machine error compensation, 344
Coplanarity, 113, 115
Correct and complete (tips on product definitions), 46
Cost savings, 319
Cylindricity, 89, 130
D
Datum target symbols, 174, 177, 189, 207
Datums
angular orientation, 125, 174, 198
centerplanes, 139, 187, 339, 372
compound, 57, 138, 294, 306
feature of size representation, 29, 187, 198, 213, 216, 218, 339, 356, 368
holes, 57, 356
oddly configured and curved surfaces, 196
partial, 177, 195
pattern, 57, 177, 294
pattern shift, 29, 213, 216, 218
rockers, 180, 441
selection of, 9, 34, 55, 169, 172, 178
simulation, 9, 35, 55, 179, 182, 198, 228, 449
switching, 198
targets, 174, 177, 189, 207, 349
Definitions, 254
Differential measurements, 70, 73, 137, 138, 140
Difficulties in measurement, 38
Dimension origin symbol, 116, 134
Dimensional inspection plan, 321, 327
Dimensioning and tolerancing for functionality and inspectability, 36
Dimensions, 7, 8, 302
Distributions, 257, 273, 410, 422
Doing something vs. doing nothing, 389
E
Elongated holes, 150
Envelopes of perfect form at MMC, 1, 245
Envelopes of perfect orientation at MMC, 101
Environmental conditions, 258, 301, 343
Explosives and a cell mate, 431
F
Feature control frames
components of, 19, 21, 23
reading, 19, 21, 23
Fixturing, 3, 183, 190, 191, 192, 194, 198, 206, 207
Flatness, 76
Flexible parts, 34, 52, 57, 87, 172, 203, 295
Focus and excel, 317
Form, 64
Free state variation, 34, 87, 172, 203, 208, 298
Full indicator movement, 128, 129, 132, 136, 138, 305
Functionality vs. producibility, 33, 41, 48, 319, 382, 389
Functional gages, 2, 31, 68, 219, 245, 259, 337, 375
G
Gages
fixed rail vs. sliding rails, 292
functional, 2, 31, 68, 219, 245, 259, 337, 375
GO, 1, 245, 250
NOGO, 2, 246, 250
push pin vs. fixed pin, 276
repeatability, 419
reproducibility, 420
softgaging, 219, 324
tolerancing, 31, 245, 248, 259, 270, 279
Gage blocks, 100, 142
General principles, 160
Geometric characteristic symbols, 19, 22, 63, 161

 Page 453
GO gages, 1, 245, 250
Graphical inspection, 351
GRR, 419, 421
H
Histogram, 406
I
Implying a manufacturing sequence, 35, 57, 176
Importance of geometric definition, 44, 50
Indicating features with letter designations, 56, 124
In-process vs. final acceptance gaging, 258
Inspecting parts as they are used, 33
Inspectors led astray, 383
Introduction to measurement principles, 1
L
Least material condition principle, 14, 230, 235
Letters, 429
Light could be a train, 430
Linear approach to inspection, 176, 322, 331
Logical approach to tolerancing, 44
M
Machine capability studies, 417
Magical, mystical stuff, 389
Material condition symbols
least, 14, 230, 235
maximum, 14, 211, 232, 235
 regardless of feature size, 20, 172, 216, 232
Maximizing your abilities, 317
Maximum material condition, 14, 211, 232, 235
Mean, 411
Measuring parts as they function, 33, 321, 382, 400
Measurement error, 86, 182, 228, 257, 304, 325, 343, 383
Measurement, inspection, and gaging principles, 7
Median, 407
Merging reality with theory, 38, 303, 306, 343, 444, 448
Methodical measurement, 331
Minimum wall thickness calculations, 231, 233
N
Needs and capabilities, 54
Normal curve, 405, 409, 411
O
Optical comparators, 122, 125, 337
Orientation, 93
P
Paper gages, 351
Parallelism, 93, 96
Pattern datums, 57, 58, 61, 177, 294
Pattern shift, 29
Perpendicularity, 97
axis, 99
centerplane, 100
surface, 98
Plastic parts and stretch toleances, 52
Polar coordinate dimensions converted to X & Y, 295, 296
Polar graphing, 81, 82
Position
boundaries, 2, 12, 150, 153, 227
coaxiality, 162
composite positional tolerancing vs. two-single segment controls, 109, 154
composite tolerancing, 154, 158, 188, 352
mating features, 12, 14
non-cylindrical features of size, 143, 187
odd configurations, 150, 153
symmetry, 143
Probes, 3, 228, 338, 349
Process stability, 409
Producibility vs. functionality, 41, 48, 319, 382, 435

 Page 454
Profile, 57, 106
Projected tolerance zones, 37, 144, 148, 291, 348, 440
Push pin gages, 276, 291
Q
Quality improvement, 309
Quality issues, 315
R
Range, 410
Reactions to a nonperfect datum establishment, 306
Reading a feature control frame, 19, 21, 23, 225
Refining geometric controls, 242
Regardless of feature size, 20, 172, 216, 232, 440
Relationships between individual features, 101
Repeatability, 419
Reproducibility, 420
Restrained state, 34, 52, 57, 87, 172, 205
Resultant condition, 235
Roundness, 79, 132, 305
Rules, 34, 160, 170, 301
Runout
circular, 126, 134, 138, 164
total, 128, 138, 164, 306
S
Sample data, 407
Search for common sense, 311
Separate gaging requirements, 185, 189
Set-ups, 9, 183
Sequential part definition, 176
Sheet metal, 34, 52, 57, 87, 172, 174, 177, 295
Shifting tolerance zones, 29
Simultaneous gaging requirements, 35, 38, 56, 61, 122, 175, 185, 199
Sine plate set-ups, 103, 104, 118
Size, 1, 13, 160, 212, 245, 250, 334
Slip sliding away, 442
Slogans vs. programs and planning, 315
Software, 3, 4, 219, 348
Spindle set-ups, 90, 128, 129, 132, 136, 137
Stabilization of parts, 172
Stability vs. accuracy, 410
Stack-up analysis, 422
Standard deviation, 407, 422
Statistical Process Control, 311, 399
SPC terms, 423
Statistical tolerancing, 300, 338, 421
Straightness, 64
derived median line, 68
derived median plane, 70
surface, 65
Surface plate set-ups, 16, 96, 98, 103, 119
Surface vs. axis verification, 13, 239
Symbols, 19, 22, 63, 161
Symmetry, 139
T
Tables
actual deviations from true position, 223, 224, 225
areas under normal curve, 415
factors for control charts, 420
Taylor principle, 13, 160, 245, 250
Teamwork, 309, 381, 388, 434
Temperature, 258, 301, 326, 343
Technical leaps, 385
Terms, 254, 423
Theory vs. reality, 38, 303, 304, 306, 343, 444, 448
Threaded features, 144, 215, 348
Tolerance origins, 42
Tolerances, 8, 12, 43, 44, 52
Tolerance stack-up analysis, 422
Training, 381, 391, 395, 431, 440, 443
Tying measurements to the machine, 311
U
Uncertainties, 86, 182, 228, 257, 304, 325, 343, 383