Pattern Inspection with Variable

Geometric Tolerance Limits

From Disregarding to Fully Applying

Variable Feature Bonus and Datum Shift
Tolerances in a Pattern Inspection
By Paul F. Jackson
 Variable Tolerance Limit Specifications and Variable Datum Shift

Maximum Variable
4X n 9.15`0.25 Tolerance Zone

j n 0.36m
0.36m A Bm
Bm
Ø0.86
For a Shaft Feature Size
@ LMC (Ø 8.9)

Minimum Variable Tolerance Zone

Ø0.36
A For a Shaft Feature Size
@ MMC (Ø 9.4)

n 25.6`0.2
b n 0m A
Maximum Datum Shift Minimum Datum Shift
Allowance ( Ø 0.4 ) Allowance ( Ø 0 )
When Datum Feature B
B
When Datum Feature B
Hole Size is LMC Ø25.8 Hole Size is MMC Ø25.4

Ⓜ & Ⓛ Symbols used in feature tolerance control frames make the feature
feature location or
feature’’s size.
orientation tolerances variable with respect to the feature
When they are used with datum features they provide freedom for the origin of measurement
to shift or rotate with respect to the size of the datum features.
features.
 Using a Single Pattern Feature Selected to Stop Rotation

Rudimentary pattern
inspections are often
performed by establishing a
coordinate reference
coincident with the Axis Feature # 3 Axis Feature # 4
intersection of the datum (∆X, ∆Y) Metric (∆X, ∆Y) Metric
feature’s planes, axes, or Y
(0.020, 0.170) (0.40, 0.16)
center-planes.

In this example where Axis B

X
rotation about the axis B is
unconstrained by a tertiary
datum feature one of the Tolerance specified @MMC 0.36
pattern features is often Position
aligned to the coordinate X Dev Ydev Feature Deviation Axis Feature # 1
system stopping the pattern 0.43 0.00 1 0.86
(∆X, ∆Y) Metric
rotation. Then the feature’s -0.16 0.24 2 0.58
0.02 0.17 3 0.34 (0.43,0.00)
coordinate deviations are
0.40 0.16 4 0.86 Set to Zero Degrees
use to figure the diameter of
the position deviation. Those
deviations are often Axis Feature # 2
erroneously compared to the (∆X, ∆Y) Metric
minimum variable tolerance (-0.16, 0.24)
to determine conformance.
Using a Single Feature for Rotation & Applying Individual Feature Bonus

Coordinate Measuring
Machines are often used to
collect both size and Feature # 3 Feature # 4
location of individual Size 9.24 Metric Size 9.14 Metric
0.52 ⊕ Tolerance
Y
0.62 ⊕ Tolerance
features. Most CMM
programs will figure the
variable portion of tolerance
( often referred to as bonus X
tolerance ) and compare
each feature’s actual
variable tolerance to the
position deviation.
Feature # 2 Feature # 1
See the next slide for the Size 9.14 Metric Size 9.04 Metric
analysis results. 0.62 ⊕ Tolerance 0.72 ⊕ Tolerance

Allowable Individual Feature Position Tolerance

= Individual Bonus + Tolerance @ MMC
= (MMC Size (9.4) – Actual Size) + (0.36)
The value “Residual Tolerance” is
simply the difference between the
position deviation and the feature’s
actual variable tolerance. CMM reports often
highlight the negative
Position Variable Residual residual tolerance
X Dev Ydev Feature Deviation Tolerance Tolerance
values as “OTOL” or
0.43 0.00 1 0.86 0.72 -0.14
-0.16 0.24 2 0.58 0.62 0.04 Out-of-Tolerance
0.02 0.17 3 0.34 0.52 0.18
0.40 0.16 4 0.86 0.62 -0.24 Rotation
0°

0.000, 0.000
Datum Shift

Basic Pattern Feature Axis Shift Allow ance Pattern Center

Feature 1 Tol Feature 2 Tol Feature 3 Tol Feature 4 Tol
Using Multiple Features for Rotation & applying Individual Feature Bonus

Most Coordinate Measuring Machine software

packages include algorithms that can simulate Cumulative Spacing
a pattern rotation and therefore minimize the Y Error Minimized
computed position error of the pattern features. Errors CW - CCW = 0

Some routines solve for a rotation correction

that makes the sum of the positive (CCW) and
negative (CW) rotation errors among all Axis B
features zero. Others may solve for a rotation
that equalizes the extreme CCW CW errors
minimizing the maximum individual error. Still
others may include each feature’s individual
variable bonus tolerances in the calculations.

See the next slide for the analysis results.

 Position Variable Residual
X Dev Ydev Feature Deviation Tolerance Tolerance
0.43 0.14 1 0.91 0.72 -0.19
-0.02 0.24 2 0.47 0.62 0.15
0.02 0.03 3 0.07 0.52 0.45
0.26 0.16 4 0.62 0.62 0.00 Rotation
-3.2°
Datum Shift
0.000, 0.000

In this example position deviations are minimized by iterating

candidate rotations to find the greatest number of positive
residual values while simultaneously minimizing the sum of
the two smallest residuals.

Basic Pattern Feature Axis Shift Allow ance Pattern Center

Feature 1 Tol Feature 2 Tol Feature 3 Tol Feature 4 Tol
Using Multiple Features for Rotation, Applying Individual Feature Bonus,
and Applying Pattern Datum Shift

Attribute gages (Position Go- Pattern Spacing

Gages) apply variable bonus Error and
tolerances physically by Allowable Datum Y Location Error
verifying fit of all feature’s Shift from Axis B Minimized
simultaneously to their virtual
conditions. Wiggling and X
rotating the part on the gage
iterates candidate shifts and
rotations that utilize all
available variable tolerance for Pattern
datum and pattern features Axis
relative to their actual sizes. Allowable Individual
Feature Bonus
Very few Coordinate Tolerance
Measuring Machine software Approaches
packages are capable of Pass / Fail
performing this analysis. Conclusions From
Attribute Gaging
See the next slide for the
analysis results.
Translation X Translation Y Rotation Ø Datum Ø Shift Tol Ø Shift
0.115 0.030 -2.400 25.64 0.24 0.24
Position Variable Residual
X Dev Ydev Feature Deviation Tolerance Tolerance
0.32 0.07 1 0.65 0.72 0.07
-0.17 0.21 2 0.54 0.62 0.08 Rotation
-0.10 0.04 3 0.21 0.52 0.31 -2.4°
0.18 0.13 4 0.45 0.62 0.17
Datum Shift
0.115, 0.030

In this example position deviations are minimized by iterating candidate

datum-shifts and rotations to find the greatest number of positive residual
values while simultaneously minimizing the sum of the two smallest
residuals. The amount of datum-shift is limited by the difference between
the datum feature’s actual and MMC size.

Basic Pattern Feature Axis Shift Allow ance Pattern Center

Feature 1 Tol Feature 2 Tol Feature 3 Tol Feature 4 Tol
 Notes / Cautions
• Simultaneous Requirements Datum-shifts must be applied in a single
common magnitude and direction to all drawing features whose feature
control frames declare the same modified datum features in the same order.
The CMM programming option to include datum shift tolerances for
individual feature conformance should be disabled unless the shift is applied
as described above.
• Continuous vs. Discreet At best rigorous continuous data analysis
techniques of variable tolerances can only approach what is physically
possible with attribute gaging because assumptions are made in data
acquisition about the individual feature’s form, orientation, and location.
 Dilemma
• Process capability estimations differ greatly between inspections done with
discreet data (attribute gaging) and those done with continuous data
(variables gaging). Continuous data capability estimations are commonly
preferred or demanded over discreet data capability estimations because
the sample sizes required to make the estimations are dramatically
different.

This is the dilemma, even though continuous data capability estimations are
preferred, they greatly underestimate conformance of process features that
have variable tolerance limits because the variable “bonus” tolerance is
excluded from the capability estimation formulas. The capability formulas
require constant tolerance limits and are not designed to handle a variable
tolerance limit. Attribute gages, on the other hand, verify fit to the variable
limit boundaries and therefore do include the variable bonus in the capability
estimation but their use is often prohibited because of the number of
samples needed to requirements.