SCMH Iaqg Aoql Values

SCMH Section 3.7.
2 Statistical Product
Acceptance
Revision Letter: D
Revision Date: 1 NOV 2018
www.iaqg.org/scmh Section 3.7
Statistical Product Acceptance

Guidance for Ensuring Quality Protection
Levels
(in support of IAQG 9138 implementation)
November 1, 2018
The intention of this document is to assist organizations with understanding the IAQG
9138 Statistical Product Acceptance Requirements standard.
This guidance was prepared by the IAQG 9138 writing team.
.
SCMH Section 3.7.2
Revision Letter: D
Revision Date: 1-NOV-2018
Page 2 of 13

Guidance for Ensuring Quality Protection Levels
TABLE OF CONTENTS
A. INTRODUCTION ................................................................................................ 3
B. TRAINED PERSONNEL .................................................................................... 4
C. SAFETY/CRITICAL CHARACTERISTICS ........................................................ 5
D. QUALITY PARAMETERS AND VALUES ......................................................... 6
D.1. DETERMINE MINIMUM ACCEPTABLE PROBABILITY OF CONFORMANCE . 6
Table 1 - Default Probability of Conformance Values for Parts .................................. 7
Table 2 - Default Probability of Conformance Values for Processes .......................... 7
Table 3 - Default Probability of Conformance Values for Mechanical Characteristics 8
D.2. CHOOSING A STATISTICAL PRODUCT ACCEPTANCE METHOD ................. 9
D.3. IDENTIFY THE QUALITY PARAMETER FOR A CHOSEN METHOD ............... 9
Figure 1 - Flowchart for Converting IRR to Quality Parameter ................................. 10
E. EVALUATION SYSTEM .................................................................................. 12
F. NONCONFORMANCES .................................................................................. 13
2
SCMH Section 3.7.2
Revision Letter: D
Page 3 of 13

A. INTRODUCTION
Statistical product acceptance is the practice of assessing product with respect to its
conformance to one or more specifications using statistical methods. Statistical theory
and methods are what allow the possible acceptance of product when only a subset is
assessed. Statistical methods enable a subset to represent the population.
There are potential savings in using statistical methods for product acceptance.
A variety of statistical methods have been developed and applied. Many applications,
despite correct statistical theory, have failed to realize savings or give enough assurance
of product conformity. The reasons are many. Some common categories of failure are
misapplication of a method, non-adherence to a method, and risks being unidentified,
unquantified, or underestimated.
The IAQG provides 9138 to make statistical product acceptance methods easier to
implement correctly and appropriately.
This SCMH Section 3.7 was written by the authors of IAQG 9138 to provide guidance
and recommendations for successful implementation of 9138 requirements and to
provide an easily obtainable source for frequently referenced sampling tables and
special statistical methods.
3
SCMH Section 3.7.2
Revision Letter: D
Page 4 of 13

B. TRAINED PERSONNEL
The guidance in 9138 clause D.1 provides guidelines for training for
• Lot sampling (D.1.2)
• Process controls (D.1.3 and D.1.4)
• Continuous sampling (D.1.4). This section also includes training for lot sampling
with switching rules and skip-lot sampling.
• Auditors for statistical methods (D.1.1.3, D.10)
The training plan should also identify when the training should be performed (generally
before sampling begins), when or under what circumstances the training will be
refreshed (often after the annual audit), how an organization will know if the training is
effective (occasionally by an in-class test), and where the records of the training will be
retained (which can be covered with the phrase “together with the production records”).
If the sample is to be representative without being random, then especially detailed

instructions must cover the fault tree analysis that makes non-random sampling be
representative, and what signs might indicate the fault tree to be indicating a need for
new action.
When non-published statistical methods are used, it is recommended that the authors or
approvers have mathematical and other relevant training similar to what is required for
the American Society for Quality Certified Quality Engineer certification or the United
Kingdom Chartered Quality Institute accreditation.
4
SCMH Section 3.7.2
Revision Letter: D
Page 5 of 13

C. SAFETY/CRITICAL CHARACTERISTICS
Critical and safety characteristics demand the highest levels of assurance methods.
The highest level of assurance would have the characteristic on every unit tested for
functionality. Ideally, such tests would be robust enough to not have to consider
statistical methods (such as when measurement uncertainties are small in relation to
safety margins).
Unfortunately, some characteristics are not amenable to such tests for example because
the tests destroy the usability of the product. When forced to predict the performance of
an untested characteristic based upon other testing, decreasing levels of statistical
assurance are described in clause D.2.
5
SCMH Section 3.7.2
Revision Letter: D
Page 6 of 13

D. QUALITY PARAMETERS AND VALUES
The recommended order of determining a quality parameter and its value is to;
(1) determine the minimum acceptable probability of conformance (D.1);
(2) determine or plan the methods to ensure the minimum is realized (D.2); and
(3) identify the quality parameter(s) that go with the method (D.3).
D.1. Determine Minimum Acceptable Probability of Conformance
New paragraph here about Engineering involvement.
The minimum acceptable probability of conformance is expressed by IAQG 9138 as the

"Initial Reliability Requirement" (IRR) also known as the "Inspection Reliability
Requirement." Clause 4.4.2 and Appendix A offer instruction on determining this IRR
including common industry values for many applications. The values from Appendix A
are replicated in this 3.7.2 as Tables 1, 2, and 3. The examples in Tables 1, 2, and 3
are not complete and may not be used to bypass any Engineering requirement. The
minimum acceptable probability of conformance for any product or product feature
should be determined by a cross-functional team consisting of manufacturing
engineering, quality engineering, and design engineering (where possible). If an
organization cannot perform this function, they should request assistance from the
design authority
Engineers, quality departments, or customers may specify values that are different than
these values. Any values given in Type Design drawings or specifications take
precedence over these default recommendations.
If a unit definition is not readily identifiable, engineering should define a unit of

inspection. (For example, engineering could specify the size of area to inspect for a
surface finish or paint adhesion application.)
6
SCMH Section 3.7.2
Revision Letter: D
Page 7 of 13

Table 1 - Default Probability of Conformance Values for Parts

PART, PROCESS, CHARACTERISTIC, OR MINIMUM RECOMMENDED
MECHANICAL REQUIREMENTS PROBABILITY OF CONFORMANCE
Machined parts (general features, except as noted under
97%
Mechanical Characteristics below)
Machined parts (special applications such as Key
99%
Characteristics, location holes, )
Conductivity 99%
Hardness 99%
Composite part - structural 98.7%
Composite part - nonstructural 95%
Cable form 98%
Connector assemblies 99%
Printed circuit boards 97%
Conformance of printed circuit board configuration 99%
Table 2 - Default Probability of Conformance Values for Processes

Sheet metal fabrication 95%
Receiving inspection of standards 97%
NDT (where permitted) 98.7%
Paint, anodize, and plating 92%
Electrical wire wrap 98%
Cleanliness test for printed circuit boards 97%
7
SCMH Section 3.7.2
Revision Letter: D
Page 8 of 13

Table 3 - Default Probability of Conformance Values for Mechanical

Characteristics
Fastener hole diameters tolerances 0.001 inch / .025mm or less 99%
Fastener hole diameters tolerances ≥ 0.001 inch / .025mm,
location on non-locating holes of > ± 0.01 inch / 0.25mm linear 95%
or equivalent diametral tolerance
Thickness with tolerance of ± 0.005 inch or less 95%
Thickness with tolerance of greater than ± 0.005 inch 92%
Linear dim. tolerance ≤ ± 0.01 inch 95%
Linear dim. tolerance greater than ± 0.01 inch 92%
Radii Total Tolerance Range less than 0.02 inch 97%
Radii Total Tolerance Range greater than 0.020 inch / 0.5mm 92%
Angle Total Tolerance Range less than or equal to ± 0.5 degree 97%
Angle Total Tolerance Range greater than ± 0.5 degree 92%
Concentricity, non-controlled contours, diameter or fillet radius
tolerances greater than ± 0.01 inch / 0.25mm, flatness / 92%
straightness, parallelism, perpendicularity, and profile
Controlled contour [Master Dimension Definition (MDD),
Master Dimension Surface (MDS)], diameter or fillet radius 95%
tolerances less than or equal to ± 0.01 inch / 0.25mm
Surface Finish Roughness (Ra) less than 32 µin or 0.8 micron 99%
Surface Finish Roughness (Ra) greater than or equal to 32 µin
95%
or 0.8 micron
Flatness Carbon Face Seals using optical flats 99%
Threaded parts:
Internal Profile Pre–load Locking 99%
Straight screw threads – total tolerance on pitch diameter 99%
Standard electrical connectors (such as EN, AN & MS parts
97%
or MH)
All other thread applications/characteristics 99%
Gears and Splines (all characteristics) 99%
Dimensions with only maximum or minimum values which
directly control material thickness of weldments, castings, or 99%
forgings; and at least one surface is machined
Absence of magnetism (ferrous parts) 99%
Spring rate (when specified by Engineering) 99%
8
SCMH Section 3.7.2
Revision Letter: D
Page 9 of 13

D.2. Choosing a Statistical Product Acceptance Method
Figure 2 in 9138 illustrates the decision process to determine what statistical method is
most appropriate for an application.
Clause F.1 in 9138 shows a list of guidelines for choosing the method.
The acceptance technique to use depends on the production plans. If production is for
one lot only, an Individual Lot Sampling technique should be used. For planned ongoing
production of lots, any of the tools in future IAQG 9138 Standard can be used. Each lot
in isolation is protected by Individual Lot Sampling. When a history of production
provides confidence in conformity, the methods of Sampling with Switching Rules can be
used. Controlled and capable processes can accept product on that basis whether or
not the product is made in lots. Other methods are available as Special or Alternate
Statistical Techniques.
D.3. Identify the Quality Parameter for a Chosen Method
Clause F.2 gives guidance for what quality parameters are appropriate for each method.
Figure 1 in SCMH 3.7.2 combines the flowcharts from Figure 2 and Figure G1 in 9138
with the quality parameters from 9138 Table F.2.
The organization may choose to apply the same quality parameter to the entirety of
inspected units or to apply different quality parameters or values to different features.
When the entire unit is inspected by one quality parameter and its value, the statistical
acceptance plan should satisfy the highest conformance need of all unit features.
Several quality parameters may be appropriate in order to cover the various products,
features, and processes at the organization’s facility.
The organization should document what quality parameters and values are assigned to
each statistical acceptance application. Showing how the quality parameters and values
are determined is also recommended. Characteristics and parts can be grouped in the
documentation if they have identical performance measures and criteria.
9
SCMH Section 3.7.2
Revision Letter: D
Page 10 of 13

Lot
Custom No or Batch Yes
IRR Design? Inspection
?
Yes
Item
-by-
Item
Attribute Attribute Yes Switching

or Switching Yes AOQL ≤ 1-IRR
Variable Rules? Rules?
?
Variable No No
Probability of
Process No Acceptance Outgoing
Control RQL ≤ 1-IRR No Quality
? ERP ≤ 1-IRR
Yes Measure?
LQ ≤ 1-IRR
EOQ ≤ 1-IRR Yes
PPM ≤ (1-IRR)*1000000
Yes Screen
AOQL ≤ 1-IRR AOQL ≤ 1-IRR Failed
EOQ ≤ 1-IRR Lots?
OQCL ≤ 1-IRR or No
PPM≤(1-IRR)*1000000
Salvage
Yes Conformances
AOQ[min[ pAOQL , pOQCL]] ≤ 1-IRR from Failed Lot
Samples?
No
OQCL ≤ 1-IRR
Figure 1 - Flowchart for Converting IRR to Quality Parameter
10
SCMH Section 3.7.2
Revision Letter: D
Page 11 of 13

If an organization elects to use a statistical technique, other than those stated in IAQG
9138, it must have verifiable mathematics that shows that the probability of conformance
meets or exceeds the IRR value. The statistical technique can take on any of the IAQG
9138 quality parameters.
If switching rules are to be implemented that change the sample sizes when some
defined criteria are met, then the appropriate quality parameter is the AOQL.
The OQCL is applicable when a failed lot is scrapped in its entirely. When sampling
does not screen entire failed lots, but attempts to salvage confirmed conformities from
the sample only, a mathematical combination of AOQL and OQCL parameters results in
the lowest sample sizes. If an organization wishes to make lot screening decisions for
failed lots at the time of failure, the OQCL parameter provides robust consumer
protection.
If an application utilizes variables data, then process control statistics become possible
such as found in Statistical Process Control (SPC) and associated control charts. If the
variables data can be characterized by a single distribution, thus showing the process is
predictable or "in control", then the quality parameters EOQ and PPM appropriately
relate the distribution to specifications.
Some quality parameters are known by different names. See the Definitions section in
9138 for some alternative names and definitions. By whatever name is used, the
statistical principles, practices, and supporting mathematics define the technique and
definitions.
Figure 1 in SCMH 3.7.2 does not include an AQL per 9138 policy. If an AQL was
specified, see 9138 4.4.2 and SCMH 3.7.8 Converting from Legacy Sampling Plans.
11
SCMH Section 3.7.2
Revision Letter: D
Page 12 of 13

E. EVALUATION SYSTEM
The following are key points from the guidelines for Evaluation Systems given in
9138 D.6.
• 4:1 minimum ratio for tolerance to measurement accuracy ranges
• Recommend study of sources of measurement variation
• Can consider Nondestructive test (NDT) costs versus product conformity costs
12
SCMH Section 3.7.2
Revision Letter: D
Page 13 of 13

F. NONCONFORMANCES
Normal Quality Management Systems for handling nonconformances should exist
independent of sampling systems.
Nonconformities from design tolerance discovered in the sample may indicate that even
larger deviations exist in the unchecked production.
Upon finding a nonconformance, material review procedures should cover the risks that:
• Other unchecked units may have nonconformances
• Other characteristics formed by the same portion of the manufacturing process

may also be nonconforming.
When a unit is conditionally passed because of a reduced, continuous, or skip-lot

sampling plan, the unit must be retrievable. This is an application of the IAQG 9100D
8.4.2 and 8.5.1q. stipulation "Where product is released for production use pending
completion of all required measurement and monitoring activities, it shall be identified
and recorded to allow recall and replacement if it is subsequently found that the product
does not meet requirements." The reason is that monitoring activities have not
concluded until the next production unit is inspected and passed. If the next unit or lot is
inspected and fails, then the conditionally passed units or lot(s) since the last inspected
and passed unit or lot are at least suspect of containing nonconformances. The suspect
units or lot(s) should be retrieved and inspected.
13

SCMH Iaqg Aoql Values

Uploaded by

Copyright:

Available Formats

SCMH Iaqg Aoql Values

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

SCMH Iaqg Aoql Values

Uploaded by

Copyright:

Available Formats

SCMH Section 3.7.

Statistical Product Acceptance

(in support of IAQG 9138 implementation)

This guidance was prepared by the IAQG 9138 writing team.

Statistical Product Acceptance

B. TRAINED PERSONNEL .................................................................................... 4

C. SAFETY/CRITICAL CHARACTERISTICS ........................................................ 5

D. QUALITY PARAMETERS AND VALUES ......................................................... 6

D.1. DETERMINE MINIMUM ACCEPTABLE PROBABILITY OF CONFORMANCE . 6

Table 1 - Default Probability of Conformance Values for Parts .................................. 7

Table 2 - Default Probability of Conformance Values for Processes .......................... 7

Table 3 - Default Probability of Conformance Values for Mechanical Characteristics 8

D.2. CHOOSING A STATISTICAL PRODUCT ACCEPTANCE METHOD ................. 9

D.3. IDENTIFY THE QUALITY PARAMETER FOR A CHOSEN METHOD ............... 9

Figure 1 - Flowchart for Converting IRR to Quality Parameter ................................. 10

E. EVALUATION SYSTEM .................................................................................. 12

Statistical Product Acceptance

Statistical Product Acceptance

• Lot sampling (D.1.2)

• Process controls (D.1.3 and D.1.4)

• Auditors for statistical methods (D.1.1.3, D.10)

If the sample is to be representative without being random, then especially detailed

Statistical Product Acceptance

Statistical Product Acceptance

D. QUALITY PARAMETERS AND VALUES

(1) determine the minimum acceptable probability of conformance (D.1);

D.1. Determine Minimum Acceptable Probability of Conformance

New paragraph here about Engineering involvement.

The minimum acceptable probability of conformance is expressed by IAQG 9138 as the

If a unit definition is not readily identifiable, engineering should define a unit of

Statistical Product Acceptance

Table 1 - Default Probability of Conformance Values for Parts

Table 2 - Default Probability of Conformance Values for Processes

Statistical Product Acceptance

Table 3 - Default Probability of Conformance Values for Mechanical

Statistical Product Acceptance

D.2. Choosing a Statistical Product Acceptance Method

D.3. Identify the Quality Parameter for a Chosen Method

Statistical Product Acceptance

Attribute Attribute Yes Switching

Figure 1 - Flowchart for Converting IRR to Quality Parameter

Statistical Product Acceptance

Statistical Product Acceptance

• 4:1 minimum ratio for tolerance to measurement accuracy ranges

• Recommend study of sources of measurement variation

Statistical Product Acceptance

• Other unchecked units may have nonconformances

• Other characteristics formed by the same portion of the manufacturing process

When a unit is conditionally passed because of a reduced, continuous, or skip-lot

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.