Core Tools: The Alphabet Soup of

APQP, PPAP, FMEA, SPC and MSA

Jd Marhevko – Accuride Corporation, SVP QLMS

Shainin Medalist, ASQ Fellow, CSSBB, CMQ/OE, CQE, STEP Awardee
ASQ World Conference – Session T12 – May 1, 2018
The FIVE Core Tools
1. APQP: Advance Product Quality Planning:
Guidelines for a product quality plan to develop a product or
service that satisfies the customer
2. FMEA: Failure Modes and Effect Analysis: Methodology used to
ensure potential problems have been considered and addressed
throughout the product and process development process (Ex.
APQP). Traditionally includes the Control Plan (CP)
3. SPC: Statistical Process Control: Basic graphing statistical tools
that enable process control and capability for continual
improvement
4. MSA: Measurement Systems Analysis: Guidelines for assessing
the quality of a measurement system where readings are
replicated
5. PPAP: Production Part Approval Process: Ensures product
consistently meets customer engineering specification
requirements during production run at the quoted production rate
Other Sample Manuals
Core Tool inferences in ISO/IATF 16949:2016
Core ISO 9001:2015 IATF 16949:2016
Tool (Core Tools NOT Specified) (Core Tool Inferred/Referenced)
APQP 8.1 Operational Planning 8.1.1 Operational Planning and Control
and Control 8.2 Requirements for Products and Services
8.2 Requirements for 8.3 Design and Development of Products and Services
Products and Services 8.4 Control of Externally Provided Processes, Products
8.3 Design and Development and Services
of Products and Services
8.4 Control of Externally
Provided Processes,
Products and Services
FMEA 6.1 Actions to Address Risks 4.4.1.2 Product Safety
and Opportunities 6.1 Actions to Address Risks and Opportunities
8.3.5 Design and 8.3 Design and Develop of Products and Services
Development Output [8.3.3.3, 8.3.5.1, 8.3.5.2]
9.1. Monitoring, 8.5 Production and Service Provision [8.5.1.1, 8.5.6.1.1]
Measurement, Analysis and 8.7 Control of Non-Conforming Outputs [8.7.1.4, 8.7.1.5]
Evaluation General 9.1 Monitoring, Measurement, Analysis and Evaluation
General
9.2.3 Manufacturing Process Audit
10.2 Non-Conformity and Corrective Action [10.2.3,
10.2.4]
10.3.1 Continual Improvement
 Core Tool inferences in ISO/IATF 16949:2016
Core ISO 9001:2015 IATF 16949:2016
Tool (Core Tools NOT Specified) (Core Tool Inferred/Referenced)
CP* 8.3.5 Design and Development 8.3.5.2 Manufacturing Process Design Output
Outputs 8.5 Production and Service Provision [8.5.1.1,
8.5.1 Control of Production and 8.5.1.3, 8.5.6.1.1]
Service Provision 8.6 Release of Products and Services
8.6 Release of Products and 8.7 Control of Non-Conforming Outputs
Services 9.1.1.2 Identification of Statistical Tools
8.7 Control of Non-Conforming 9.2.2.3 Manufacturing Process Audit
Outputs 10.2.3 Problem Solving
Annex A. Control Plan
PPAP 8.3.4 Design and Development 8.3.4.3 Prototype Program
Control 8.3.4.4 Product Approval Process
*The Control Plan is not considered a “stand alone” Core Tool. Usually paired with the P-FMEA
Core Tool inferences in ISO/IATF 16949:2016
Core ISO 9001:2015 IATF 16949:2016
Tool (Core Tools NOT Specified) (Core Tool Inferred/Referenced)
SPC 9.1 Monitoring, Measurement, 8.3.5.2 Manufacturing Process Design Output
Analysis and Evaluation 8.6.4 Verification & Acceptance of Conformity…
9.1 Monitoring, Measurement, Analysis and
Evaluation
MSA 7.1.5 Monitoring and Measurement 7.1.5 Monitoring and Measuring Resources
Resources 7.1.5.1.1 MSA
7.1.5.2.1 Calibration/Verification Records
7.1.5.3 Laboratory Requirements
8.6.3 Appearance Items (inference)
 APQP
Advanced Product Quality Planning

DFSS DMAIC
APQP
What is it: The management of Product Development
Why do we need it: To understand what our customer
wants and to fulfill those wants
How is it done: Across a prescriptive “Five-Stage”, “Gated”
or “Phased” approach. Other iterations exist and are also
used so long as the foundational five are in place. The
process is required to be cross-functional in its development
and execution
The Typical APQP Stages/Phases
CONCEPT
PROGRAM
INITIATION/
APPROVAL
APPROVAL PROTOTYPE PILOT LAUNCH

PLANNING
PLANNING
PRODUCT DESIGN AND DEV.

PROCESS DESIGN AND DEVELOPMENT

PRODUCT & PROCESS VALIDATION

PRODUCTION

FEEDBACK ASSESSMENT AND CORRECTIVE ACTION

Planning Product Design & Process Design & Product & Process Feedback,
INPUTS Development Development Validation Assessment &
INPUTS INPUTS INPUTS CAPA INPUTS

Planning Product Design & Process Design & Product & Process Feedback,
OUTPUTS Development Development Validation Assessment &
OUTPUTS OUTPUTS OUTPUTS CAPA OUTPUTS
APQP Plan & Define Phase
Typical Inputs Typical Outputs
VOC Data Design goals
Marketing Strategy Reliability/Quality Goals
Product/Process Preliminary Critical
Assumptions Characteristics
Customer Inputs Preliminary Process Flow
Compliance Criteria Preliminary BOM
Etc. Etc.
APQP Product Design & Development Phase
Program Approval
Design Outputs APQP Outputs
DFMEA New Equipment/Tooling
Design for Mfg/Asm New Facility Needs
Design Verification Gage/Test Requirements
Prototype Built Final Critical Characteristics
Eng Drawings/Specs Etc.
Etc.
APQP Product Design & Development Phase

Prototype Outputs
Pkg Standards/Specs MSA/AAA
Product/Process Review Management Support
Process Flow Chart Cp/Cpk Plan
Floor Plan Work Instructions
PFMEA/DCP Etc.
APQP Product & Process Validation
Phase
Pilot. Sample Outputs
Significant Production Run Packaging/Preservation
MSA/AAA Production Control
Cp/Cpk Studies Quality Sign-Offs
PPAP Completion Management Support
Product Validation Testing Etc.
APQP Feedback, Assessment & CAPA
Phase
Launch Outputs
Reduced Variation
Improved Customer Satisfaction
Improved Delivery/Service
Lessons Learned
Standard Work Updates
Etc.
 Design FMEA
Design Failure Mode Effects Analysis
ALL Products & Processes Fail
Failure is ALWAYS a Design Requirement/Criteria

Determining HOW the design will fail, WHEN it will fail, and
WHY it will fail will allow a designer to incorporate failure as
an acceptable design constraint

Failure as an ACCEPTABLE design constraint =

Customer Satisfaction =
Design Quality

Planning for Failure

FMEA: Design (D) & Process (P)
What is it: A risk analysis of a part or process
Why do we need it: To identify the functions of a process and
the associated potential failure modes, effects and potential
causes. The vision is to prevent problems from occurring so that
defects are not incurred and no one gets hurt. It is used to
evaluate if the current planned actions are sufficient and effective
How is it done: Via the utilization of a cross-functional team
approach. Multiple iterations exist across industry. Within IATF,
the process is required to be cross-functional in its development
and execution. It is considered a “Risk-Based Thinking” (RBT)
tool. It often incorporates results from other methods such as
SPC, MSA, Fault Tree Analysis, etc.
FMEAS for Products & Processes
There are three (3) basic cases in which an FMEA is applied:
1. New designs, new technology or new process
2. New application of existing design or process
3. Changes to an existing design or process
• Design FMEA: A technique which analyzes system
functions within a defined boundary to address possible
design weakness and potential risks of failure. DFMEA
data is used in the creation of the PFMEA
• Process FMEA: A technique which analyzes processes
that can impact quality. These processes may be:
Receiving, Handling, Manufacturing, Assembly, Storage,
Transportation, Maintenance, Repair and Communication
Six (6) Steps of an FMEA (D or P)
Define System Function Failure Risk Optimiza
Scope Analysis Analysis Analysis Analysis -tion

1. Define Scope. Identify what is to be included in the

evaluation. (System, Sub-system, Component). Include
relevant Lessons Learned (LL) and reference materials.
Manage the five (5) T’s:
1. Team: Who will constitute the core team
2. Timing: When is it due. Gantt, lay-out timing plan
3. inTent: Why is the team there; Ensure skills/training
4. Tool: What reporting methodology will be used? Excel,
Software, etc
5. Task: What work needs to be done across the six steps.
Consider inclusion of effective documentation for
auditing and customer review
 Define System Function Failure Risk Optimiza
Scope Analysis Analysis Analysis Analysis -tion

2. Conduct System Analysis: Define the customer(s) wrt

End Users, Assembly, Manufacturing, etc.
1. Identify and break down the design into system,
sub-system, component and parts for functional risk
analysis. Note: A component FMEA is a subset of a
system FMEA. Ex. A brake pad is a component of a
brake assembly which is a sub-system of the chassis
2. Visualize the system via block (boundary) and/or
structure tree diagrams
Chassis

Structure
Block
 Define System Function Failure Risk Optimiza
Scope Analysis Analysis Analysis Analysis -tion

3. Conduct Function Analysis: Insures that the specified

and required functions are appropriately allocated to the
system elements. A function describes WHAT the item/
system element is intended to do.
1. Associates functions with the pertinent system elements
2. Overviews the functionality of the product
3. May describe functions in detail. May need to consider
interfaces and clearances wrt physical connections,
material exchange, energy transfer and data exchange
4. Allocates requirements/characteristics to individual
functions
5. Cascades internal/external customer functions with
associated requirements for intended use
 Define System Function Failure Risk Optimiza
Scope Analysis Analysis Analysis Analysis -tion

4. Conduct Failure Analysis: Identify failure causes,

modes, and effects, and show their relationships to
enable risk assessment.
Failure effects are the consequence of a failure mode
1. Identification of potential failures assigned to functions
in structural elements
2. Visualize failure relationships (FMEA spreadsheet)
3. Collaborate between the customer and supplier on
effects
Consider “Failure Chain” WHAT HOW WHY
approach. AKA the Failure Failure Failure
Golden Circle Effect (FE) Mode (FM) Cause (FC)

Marker dried out Cap Fell Off Barrel ID too Small

 Define System Function Failure Risk Optimiza
Scope Analysis Analysis Analysis Analysis -tion

5. Conduct Risk Analysis. Prioritize the risks by evaluating

Severity (how bad), Occurrence (how often) and
Detection (how well can we find it). Aka SOD. Each is on
a scale of 1-10. The multiplication of S x O x D is the RPN
1. A Risk Priority Number (RPN) is determined
2. Based on the RPN, assign preventive controls which
provide information/guidance as an input to the design
3. Assign detective controls to verify and validate
procedures previously demonstrated to detect the
failure
4. Completed SOD assessment
5. Collaboration between customer and supplier on
Severity
RPN, Criticality or Prioritization
Each method of evaluation has pros and cons. There is a
change in process towards an “Action Prioritization” (AP)
matrix which may incorporate Criticality (S*O). RPN will be
eliminated as a method of risk evaluation (AIAG, 2018)
AIAG currently references the SOD tables found in the
FMEA “Blue Book”. Many organizations have evolved to
their own form of prioritization tables
based on their own logic
 4th Ed SOD Summary for Design FMEA
NOTE: OEs & Other businesses often use their own SOD tables. This is a MODEL

# Severity Criteria Occurrence Criteria Opportunity for Detection

10 Failure to meet safety and/or regulatory requirements. Very high. New technology/new design with no No detection opportunity: No current design control. Cannot detect
Potential failure mode affects safe vehicle operation history. >= 1 per 10 or is not analyzed. Detection is almost impossible
and/or involves non-compliance with government
regulation without warning

9 Failure to meet safety and/or regulatory requirements. High. Failure is inevitable with new design, new Not likely to detect at any stage. Design analysis/detection controls
Potential failure mode affects safe vehicle operation application or change in duty cycle/operating have a weak detection capability. Virtual analysis is not correlated
and/or involves non-compliance with government conditions. 1 in 20 to expected actual operating conditions. Detection is very remote
regulation with warning

8 Loss or degradation of primary function. Loss of High. Failure is likely with new design, new Post design freeze and prior to launch. Product
primary function application or change in duty cycle/operating verification/validation after design freeze and prior to launch with
conditions. 1 in 50 pass/fail testing. Detection is remote
7 Loss or degradation of primary function. Degradation High. Failure is uncertain with new design, new Post design freeze and prior to launch. Product
of primary function application or change in duty cycle/operating verification/validation after design freeze and prior to launch with
conditions. 1 in 100 test to failure testing. Detection is very low
6 Loss or degradation of secondary function. Loss of Moderate. Frequent failures associated with Post design freeze and prior to launch. Product
secondary function similar designs or in design simulation and verification/validation after design freeze and prior to launch with
testing. 1 in 500 degradation testing. Detection is low
5 Loss or degradation of secondary function. Moderate. Occasional failures associated with Prior to design freeze. Product verification/validation after design
Degradation of secondary function similar designs or in design simulation and freeze and prior to launch with pass/fail testing. Detection is
testing. 1 in 2,000 moderate
4 Annoyance. Appearance or audible noise, vehicle Moderate. Isolated failures associated with Prior to design freeze. Product verification/validation after design
operable, item does not conform and noticed by most similar designs or in design simulation and freeze and prior to launch with test to failure testing. Detection is
customers (>75%) testing. 1 in 10,000 moderately high

3 Annoyance. Appearance or audible noise, vehicle Low. Only isolated failures associated with almost Prior to design freeze. Product verification/validation after design
operable, item does not conform and noticed by many identical design or in design simulation testing. 1 freeze and prior to launch with degradation testing. Detection is
customers (>50%) in 100,000 high

2 Annoyance. Appearance or audible noise, vehicle Low. No observed failures associated with almost Virtual analysis correlated. Design analysis/detection controls have
operable, item does not conform and noticed by identical design or in design simulation testing. 1 a strong detection capability. Virtual analysis is highly correlated
discriminating customers (<25%) in 100,000,000 with actual or expected operating conditions prior to design freeze.
Detection is very high
1 No discernable affect Very low. Failure is eliminated through preventive Detection not applicable; failure prevention. Failure cause or failure
control mode can not occur because it is fully prevented through design
solutions. Detection is almost certain

PFMEA 4th Edition. 2008. Chrysler LLC, Ford Motor Company, General Motors Corporation
 Define System Function Failure Risk Optimiza
Scope Analysis Analysis Analysis Analysis -tion

6. Evaluate for Optimization. The planning and execution of

actions to mitigate risk and assess the effectiveness of
those actions
1. Identify necessary actions
2. Assign responsibilities and timing
3. Confirmation of effectiveness of the actions taken
4. Continuous improvement of the design

Multiple other types of FMEA applications: System,

Concept, Environmental/Safety, Machinery, Software, etc.
DFMEA Sample Format
DFMEA formats vary widely based on OE criteria and
independent company expectations…Even though the AIAG
will add ~8-10 more columns to the current standard, the
general approach and intent will be the same; mitigate risk
through failure analysis
2. SYSTEM ANALYSIS 3. FUNCTION ANALYSIS 4. FAILURE ANALYSIS 5. RISK ANALYSIS 6. OPTIMIZATION

Ooccurrence (O)
Occurrence (O)
Severity (S)

Detection (D)

Detection (D)
Severity (S)
Potential Potential

Class

RPN

RPN
Controls Controls Recommended Responsibility & Actions Taken
Item Function Requirement Potential Failure Mode Effect(s) of Causes of
(Prevention) (Detection) Action Target Date Completion Date
Failure Failure

Barrel ID too Spec for Instron pull test

N/A 4 2 32 None at this time
small interference fit ABC
1,000 ft of
Marker Write continuous Cap Falls Off Marker dries out 4 Spec for 4 Instron pull test 2 None at this time
drawing N/A Cap ID too large 0
interference fit ABC
Use felt material 2 CTE lab test 3
Felt insert too
N/A with low CTE XYZ 0 None at this time
long
Other DFMEA Sources…
• http://quality-one.com/fmea/design-fmea/
• http://www.isixsigma.com/dictionary/dfmea/
• http://www.qmii.com/LT-
133%20ISO%209001_2015%20Risk%20Based%20Thinking.pdf
• http://www.iso.org/iso/home/standards/iso31000.htm (ISO Risk
Management)
• 86 Minute Video…very detailed
http://www.isixsigma.com/tools-templates/design-of-experiments-
doe/mark-kiemele-interview/
• AIAG APQP for DFMEA Checklist (2nd ed)
 Process FMEA & CP
PFMEA + Control Plan = Dynamic Control
Plan

Dynamic Control Plan

What is a DCP
A DCP is a blended format of a PFMEA
and CP. It leverages the common columns
in both tools and enables “linear” thinking
across the analysis of an individual process step
It saves time and increases the security of the system
• A PFMEA defines, identifies, prioritizes, and eliminates known
and/or potential process failures from reaching the customer. The
goal is to eliminate Failure Modes and reduce their risks
• A CP follows the PFMEA steps and provides details on how the
"potential issues" are checked for in the process
• A DCP is a living document which helps to prevent problems
• It saves time and increases process security
A DCP
A DCP lists a sequence of tasks used to produce a product or
provide a service by combining the PFMEA and CP. It:
1. Identifies process related Failure Modes before they occur
2. Determines the Effect & Severity of these failure modes
3. Identifies the Causes and probability of Occurrence of the
failure modes
4. Identifies the Controls and their Effectiveness
5. Quantifies the Risks associated with the failure modes
6. Develops and documents Action Plans to reduce the risks
7. Identifies the Type & Effectiveness of the Gaging system
8. Determines the necessary Inspection Frequency
 FMEA & CP in One Format
Plant Part/Produc t Name Customer PN Customer PN/Revision/Date
Site Address XXX XXX XXX 01/01/00
Site Address Proc ess PN: Flow Chart # PN/Revision/Date
Site Address XXX XXX XXX XXX 01/01/00
Prototype (X) Pre-Launch (X) Production (X)
Product/Process Characteristics Potential Failures and Effects Causes of Failure Current Controls
Char. or Process SC Failure Effects of Control - Detect Control Method to Responsible
No. Desc Characteristic Class Mode Failure SEV Cause OCC Failure Mode Prevent Cause DET RPN Recommendations Person/Timing
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
A B C D E F G H I J K L M N O

The format is completed CP “Side” P - AA

linearly from A – AA. This
Dynamic Control Plan (DCP) Revision/Date Core Team:
"C" 01/01/00 XXX Design Eng XXX Other
"B" 01/01/00 XXX Mfg Eng XXX Other
ensures inclusion of a gaging "A" 01/01/00
(Rolling top 3 levels)
XXX
DCP File Number:
Prod Mgr
XXX
XXX Other

system review and eliminates Orig New New New Ctrl Tool Gage Desc/ GRR & Insp Cpk & Reaction
the need to manage 2 forms SEV OCC DET RPN Fctr OWI# Fxt # Gage No. Date Freq Date Plans

** ** ** 0 0
** **
**Many sites modify 0
0
0
0

the format to fit their 0

0
0
0
0 0
own needs 0 0
0 0
P Q R S T U V W X Y Z AA
A Practice DCP
The fit of a marker cap…
1. Look at the cap and barrel of a writing marker
2. Review the step of assembling the cap onto the barrel
3. Complete relevant lines of the DCP wrt assembly
4. There can be two general failure modes:
a. The cap fits with an audible “click” and stays firmly in
place. It does NOT easily pull off
b. The cap does not stay secure and falls off
5. Each failure mode will have its own “DCP Stream” of
information
6. Follow across the format and complete the information
7. Work in teams across the format
 4th Ed SOD Summary for Process FMEA
NOTE: OEs & Other businesses often use their own SOD tables. This is a MODEL

# Severity Criteria (Customer Effect) Occurrence Opportunity for Detection

10 Failure to meet safety and/or regulatory requirements. Potential Very high. No detection opportunity: No current process control. Cannot detect or is not analyzed.
failure mode affects safe vehicle operation and/or involves non- >= 1 per 10 Detection is almost impossible
compliance with government regulation without warning

9 Failure to meet safety and/or regulatory requirements. Potential High. Not likely to detect at any stage. Failure mode and/or Cause is not easily detected.
failure mode affects safe vehicle operation and/or involves non- 1 in 20 Detection is very remote
compliance with government regulation with warning

8 Loss or degradation of primary function. Does not affect safe High. Problem detection post processing. Failure mode detection post processing by operator
vehicle operation 1 in 50 through visual, tactile, or audible means. Detection is remote

7 Loss or degradation of primary function. Degradation of primary High. Problem detection at source. Failure mode detection in-station by operator through
function. Vehicle operable at reduced level of performance 1 in 100 visual, tactile, or audible means or post-processing through attribute gaging. Detection
is very low
6 Loss or degradation of secondary function. Vehicle operable but Moderate. Problem detection post processing. Failure mode detection post-processing by
convenience/comfort functions inoperable 1 in 500 operator through use of variable gaging or in-station by operator through use of
attribute gaging. Detection is low
5 Loss or degradation of secondary function. Vehicle operable but Moderate. Problem detection at source. Failure mode or error detection in-station by operator
convenience/comfort functions at reduced levels of performance 1 in 2,000 through use of variable gaging or by automated controls in–station that will detect
st
issue and notify operator. Gaging performed on setup and 1 pc check. Detection is
moderate
4 Annoyance. Appearance or audible noise, vehicle operable, item Moderate. Problem detected post processing. Failure mode detection post-processing by
does not conform and noticed by most customers (>75%) 1 in 10,000 automated controls that will detect discrepant part and lock part to prevent further
processing. Detection is moderately high

3 Annoyance. Appearance or audible noise, vehicle operable, item Low. Problem detection at source. Failure mode detection in-station by automated controls
does not conform and noticed by many customers (>50%) 1 in 100,000 that will detect discrepant part and automatically lock part in station to prevent further
processing. Detection is high
2 Annoyance. Appearance or audible noise, vehicle operable, item Low. 1 in Error detection and/or problem prevention. Error cause detection in station by
does not conform and noticed by discriminating customers (<25%) 100,000,000 automated controls that will detect error and prevent discrepant part from being made.
Detection is very high
1 No discernable affect Very low. Failure is Detection not applicable; error prevention. Error cause prevention as a result of
eliminated through fixture/machine/part design. Discrepant parts cannot be made due to error proofing.
preventive control Detection is almost certain
PFMEA 4th Edition. 2008. Chrysler LLC, Ford Motor Company, General Motors Corporation
For Want of A Horse

SKF USA Inc, Reprinted with permission 02/2018

DCP or Fire-Fight?
Planning vs Fire-Fighting
When Planning is Secondary to Fire-Fighting When Fire-Fighting is Secondary to Planning

Continuous
Many Surprises Fire-Fighting

Significantly Fewer
Surprises
Resources

Resources
Planning
through DCP Smoother
Production
Minimal
Planning

Planning Launch Production Planning Launch Production

Project Timing Project Timing

Total time is area under the curve…Estimated monies are

7:1 with OT, Freight, Material/Equipment changes, T&E,
etc. Leverage the DCP to minimize fire-fighting after
release. Partner with functional teams
Case Study: Before/After DCP
Initial release and after DCP implementation of 3 products.
Was planning secondary to firefighting? What kinds of losses
were likely incurred? Was it worth it?

> June: Before DCP

> Sept: After DCP
> December:
Current Performance
 PPAP
Production Part Approval Process
PPAP
What is it: Requirements for approval of production parts
Why do we need it: To make sure that we understand all of
the customer requirements, and that we can meet them
under actual production conditions
How is it done: Based on customer direction, there are 5
levels of PPAP to secure product approval. An application
“cover sheet” is called a Product Sample Warrant (PSW)
which lists 18-20 different types of evidence that may be
required for submission. These can be customer and/or
product/process dependent. It is typical for a customer to
witness a launch and review PPAP records when on-site
PPAP Levels per AIAG 4th ed.
1. Warrant only for appearance items
2. Warrant with product samples and limited supporting data
3. Warrant with product samples and complete supporting
data
4. Warrant with other requirements specified by the customer
5. Warrant with product samples and complete supporting
data reviewing at the supplier’s manufacturing location
PPAP level details are typically arranged in advance between
the supplier and customer and will often depend on whether
the product is a new design or another revision of a tried and
true process
PPAP Components
1. Design records 11. Initial process study
2. Authorized Engineering 12. Qualified lab documentation
Change documents 13. Appearance approval report
3. Customer engineering 14. Sample production parts
approval 15. Master samples
4. Design FMEA 16. Checking aids
5. Process flow diagrams 17. Customer specific
6. Process FMEA requirements (CSR) records
7. Control Plan 18. PSW
8. MSA Studies 19. Bulk material requirements
9. Dimensional results checklist
10. Material/performance test 20. Special process audit
results results
PPAP Prep…All Hands on Deck
1. TAKES TIME and attention to DETAIL
2. Requires a cross-functional team
3. Insure a good understanding of the Customer Specific
Requirements (CSRs) in advance
4. Do WELL on the Appearance Approval Reports (AARs). While
the easiest “up front”, these are often the most expensive
later on. Take the time to develop boundary samples and
conduct Attribute Agreement Analysis (AAAs) to ensure skill
5. Attend to the full Measurement System Analysis (MSA) on
variables metrics. Include calibration, resolution and GRR
6. Enable sufficient lead time for the DFMEA, FMEA and CP
7. Insure statistical control of significant characteristics
8. Etc.
How to Organize
1. Many customers will dictate submission formats
2. Some companies establish binders/books
3. Some use formal organizing software

It is critical that:

1. More than 1 person has access/passwords

2. Proper security is enabled across those individuals
3. Proper revisions are sustained/maintained
 Cp/Cpk/Pp/Ppk
Process Capability Primer
 Process Capability 101
• Cp/Cpk: Also called “short term” capability
which is used to reliably determine if a
process is yielding good initial results by
taking a representative sample size.
➢Cp is based on the whole breadth of the process
➢Cpk is based on “half” of the process
• Pp/Ppk: Also known as “long term” process
capability. The key difference is that there is
much more data on hand for Pp/Ppk. AIAG
notes “90 shifts, 90 days”
 Dissecting the Bell
Lower Spec Limit Upper Spec Limit

6s (+/- 3 on each side of the average)

When there 34.2% 34.2%

is ROOM 6 5 4 3 2 1 1 2 3 4 5 6
for 6 13.6% 13.6%
sigma’s on
THEN you
EACH side 2.1% 2.1%
of the have “6s
average 0.15% 0.15% quality!”
+/- 1s 68.4%
before the
+/- 2s 95.6%
closest
target is +/- 3s 99.70%

hit… +/- 4s 99.994%

+/- 5s 99.9994%

+/- 6s 99.9997%
Calculating Capability
Cp (Pp). Measures the ability of Cpk (Ppk). Measures the ability of
the WHOLE bell to fit within the HALF of a bell (3 sigmas) to fit within
target limits the average and the closest target
limit
If the whole bell (6 sigmas) fit
within the target limits a total of 1 CpkU = (USL – Average) / (3 x s)
time, then the Cp = 1. CpkL = (Average – LSL) / (3 x s)
Ideally, 2 is preferred.
USL = 6, LSL = 0, s = 1
Cp = (USL – LSL) / (6 x s)
Cp = (6 – 0) / (6 x s) = 1
USL = 6, LSL = 0, s = 1 CpkU = (6 – 5) / (3 x s) = 1/3 (0.33)
CpkL = (5 – 0) / (3 x s) = 1 2/3 (1.67)

0 3 6 0 1 2 3 4 5 6 7 8
 Cpk Worksheet
Determine the Cp and Cpk for each situation…Remember, if the process is NOT
shaped like a bell, then sigma cannot be used (without special consideration)
and the Cp/Cpk cannot be properly determined
In each case either the # Avg s Cp CpkU CpkL %Non-Conf
average or sigma may
or may not change… 1 5.0 2.50
only the specifications
remain the same 2 5.0 1.67
3 7.5 0.83
4 5.0 0.83

1 2 3 4

0 5 10 0 5 10 0 5 7.5 10 0 5 10
 Shift Happens
Before
Cpk of 2 is desired for initial capability
Cpk
Long term capability is Ppk. This is the capability
after the process experiences “life” via multiple
material lot changes, set up and operator
variation, seasonality, etc. Ppk is usually
calculated after “90 days” (or with a significant Cpk = 2.0
quantity) of process data. It is the type of product
results that the long term process will represent +/- 1.5s
After
It is estimated that a process will “shift” by +/- Ppk
1.5s in response to those changes. As such, if a
process started ideally with a Cpk of 2.00, then it
is estimated that the resultant Ppk would be 1.33
to accommodate these types of affects
Ppk ~ 1.33
 MSA (GRR & AAA)
Measurement Systems Analysis
Measurement System Analysis
When we measure or make an assessment
of the goodness of an item, we need to be
sure that our result is correct. If it is not
correct, we take two risks:
➢ Alpha a Risk: We may inadvertently discard or
rework a good item (Aw, darn)
➢ Beta b Risk: We may inadvertently pass on a
bad item (Boy, that was Bad)
Why Do We Need to Know?
We need to know how much error there is in our
measurement processes for several reasons:
• Prevent a and b errors
• Reduce scrap/rework
• Understand what process Cp/Cpk we
need our processes to have
• It is our JOB to ensure that our people
are enabled to make the right pass/fail
decision EVERY time
• And of course…it is an inherent part of PPAP
• NOTE: EVERY item called out for measure or inspection on
a control plan is REQUIRED to have an MSA analysis
conducted.
 MSA Types: Variable & Attribute
Humans usually believe what they see and do not question a
value shown on an instrument. There are two typical types of
variables MSA used to determine the percentage of results error:
• Crossed Gage R&R (Repeatability & Reproducibility): One
instrument, multiple operators and multiple part samples
• Nested GR&R. Used for gage error in destructive testing
There is generally one type of Attribute MSA to
determine HOW right or wrong we are in our results:
• Attributes Agreement Analysis (AAA) is used for
items we assess visually or by go/no go or needs
to be categorized Is this window
broken? It still
opens. The
wooden frame is
in place
How Data Varies

Accurate & Inaccurate Accurate but Inaccurate &

Precise but Precise Imprecise Imprecise

Accuracy: Generally managed by calibration includes bias

(how far off), linearity (across the breadth of the measured
range) and stability (holding a measure over time)
Precision: Generally managed by Repeatability (gage) and
Reproducibility (human) aka GR&R
General MSA Notes
For a variables Measurement System to work,
three features are equally needed:
➢ Resolution: Ability to read the gage. (Discrimination).
Resolution needs to be at least 10% of the tolerance
(If not at 10% or better, additional actions are needed)
➢ Calibration: A check of bias, linearity and stability
(performed on a regular basis)
➢ GR&R: Amount of error in human and gage performance.
Typical GR&R <= 10% error on safety features. Included in
PPAP, it insures that the gage system will work as intended
BEFORE the process is launched. After that, it is conducted
on an as needed basis (verification of process, gage
system change, qualification of personnel)
Resolution and Cpk
What does Resolution do for you?
With a “10% resolution gage”, we
would accept a unit that reads 10.
But…it could be a 9 or an 11. We
are at risk 1/3 of the time for a b
error…IF the Cp/Cpk is 1
-10 0 +10 We would also reject an 11, (it
XXX could be a 10 or 12). We could
-10 0 +10 have an a error 1/3 of the
time...Again, IF the Cp/Cpk is 1
XXX
-10 0 +10

This is one of several reasons why a

Cp/Cpk of 1 isn’t good enough for
safety features
 Resolution With Cpk >1.33
Resolution with better process capability
With a more capable process, if
we still have a “10% gage”, the
process is not likely to generate
any units measuring a “10”. As
such, if we read an 8, it could still
be a 7 or 9. However, there is
-10 0 +10 now minimal risk for either an a or
XXX b error. In this case, the Cp/Cpk
-10 0 +10 is 1.33

This is one of several reasons of why

a minimum Cp/Cpk of 1.33 is
required for safety features
 Attribute Agreement Analysis
AAA Checks for the chances of 100%
agreement on three features:
➢ Within “myself”; Did I repeatedly call it good or bad in
a consistent manner (even if I was wrong)
➢ Between both me and “my peer”; Did both my peer
and I repeatedly call it good or bad in a consistent
manner (even if we were both wrong)
➢ Compared to “Standard”; Did I/we get it right
AAA Quick Notes
An AAA needs many Pass/Fail “Samples”;
Preferably 50 or more (pass/fail/borderline).
NOTE: One unit might have several samples on it
An AAA is a check for accuracy in human perform-
ance. The target for “Statistical Agreement” is >= 85%.
Another form of Agreement is called Kappa (K). AIAG
calls out for K >= 75%. AAA is done as a part of PPAP
to ensure that the review process will work as
intended; before the process is launched. It should be
treated as a “maintenance” action with regular review
to keep human assessors “calibrated”. Usually
quarterly
 AAA: What It Looks Like
AAA Gives a series of graphs to show how the
operators perform in general. While 100%
agreement is not feasible, (like 0% GRR Error),
industry norm is 85% for Statistical Agreement

Not an effective Statistical Agreement at < 85%

This team will be in statistical agreement about
68% of the time.
However, 95% of the time, they will
likely range from 47% in agreement to
85%
 SPC
Statistical Process Control
X-bar Chart Sample
12
10
8
6
4
2
0
-2 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

R-Chart Sample
5

4
 What’s Normal?
There are 6 main causes of Normal Variation for almost any
type of process…
This is NORMAL. Hence the “normal” or Gaussian
distribution.
Ma npower
Ma chine
Ma terial
Me thod
Me asurement
E nvironment
SPC; High Level Guidelines
1. SPC applies to both variables and attributes. It is a graph-
based statistical method to analyze and control a process
2. First step is to insure MSA effectiveness; whether for
variables (GRR) or attributes (AAA)
3. For variables, must insure that the process is capable
FIRST, prior to establishing a control chart (Cpk >= 1.33)
4. Determine any key patterns (common sense control) that
are meaningful to your process and train to those
conditions. These typically include: Shifts, Trends, Points
outside of the limits
5. After that, it’s a go/no go chart. The graphs help you to
know when the processes change (whether desired or not)
 After GRR & Cpk; Now We Can Chart
Moving X and Range chart plots data X-bar Chart Sample
across time along with its corresponding 12

ranges. Patterns are reviewed for 10

prevention purposes. 8
6

Most Common Signals: 4

2
• 5 or more points above or below the 0
average line is considered a shift (bell -2 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

has moved)
• 5 or more points continuously 5
R-Chart Sample

increasing or decreasing is considered 4

a trend
• Any point outside of the control limits.
3

2
These are considered non-normal 1
patterns and the process spread has
likely increased 0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

NOTE: Different references call out

varying control criteria X-bar and R charts are PREVENTIVE
and PREDICTIVE forms of process
management. They give an advanced
warning enabling proactive actions
 Pre-Control: No “Limits”
SPC is powerful and effective. Pre-Control is a step before that. It “forces” a
1.33 Cpk by requiring the process to “pre-act” when data signals are in-spec
but outside of the +/- 3 sigma range. While no control limits need to be
calculated, careful communication of WHY a person needs to react and
adjust the process for an in-spec part
USL
Red zone. Stop/Adjust process. Take next set of
readings. Recover process back to Green/Yellow zones
Yellow zone. Monitor for trends, shifts. Run as long
as results are within +/- 3 sigmas from the nominal

Green zone. Run as

long as results are
within +/- 2 sigma of
the nominal

Yellow zone. Monitor for trends, shifts. Run as long

as results are within +/- 3 sigmas from the nominal
Red zone. Stop/Adjust process. Take next set of
readings. Recover process back to Green/Yellow zones
LSL
 Attribute Charts; With a Good AAA
p-chart. A trend-based percentage chart. Must be paired with a Pareto or checksheet
to execute fixes. A p-chart typically follows a Weibull distribution because either 0 or
100 is optimal and a “half bell” is developed with bias towards one end or the other.
c-chart. This “counts” defects per unit. Ex. A application may have 3 typos, 2
smudges and 2 areas not filled out for 8 defects on 1 item. The next one may be
perfect. The c would equal 4 defects per unit. This is a highly effective method that
captures detailed data. It is powerful when paired with a Pareto. Again, checksheets
are often used. There is usually a high cost to capture this data. c-charts are usually
“turned on/off” to capture a timeframe of data and then rechecked later to verify the
effectiveness of the fixes
Trends:
• 5 or more points above or below the average line is considered a shift
• 5 or more points continuously increasing or decreasing is considered a trend
• Any point outside of the control limits. Spread has likely increased
p-Chart Sample c-Chart Sample p and c Charts describe what
60% happens AFTER the process has
50%
6 occurred. (identifying either
40%
4 scrap/rework). Losses are incurred.
30%

2
The intent of these charts is to see
20%

10%
if the corrective actions are
0
0% working
1 2 3 4 5 6 7 8 9 10
1 2 3 4 5 6 7 8 9 10
66
Common Types of SPC Charts
Chart Type Primary Usage What is Charted Typical
Sample Size
X-Bar & R Routine monitoring of high Plots the average of ~3 to 6
volume manufacturing the data set and its
processes range
Individual & Used when only sample is Plots the value and One
Moving possible. Common for the moving range of
Range transactional (monthly) the current and
(IMR) processes preceding values
p-Chart Routine monitoring of high Plots the percent Variable
volume processes where non-conforming
scrap/rework trends are
critical
c-Chart Used for deeply analyzing Plots the average Variable
non-conformities in a number of non-
product conformities in a
single unit
Where The Alphabets Fit…
APQP 5 Stages (or more)

DFMEA

MSA
PFMEA /CP

MSA

PPAP

SPC
The FIVE Core Tools
1. APQP: Advance Product Quality Planning:
Guidelines for a product quality plan to develop a product or
service that satisfies the customer
2. FMEA: Failure Modes and Effect Analysis: Methodology used to
ensure potential problems have been considered and addressed
throughout the product and process development process (Ex.
APQP). Traditionally includes the Control Plan (CP)
3. SPC: Statistical Process Control: Basic graphing statistical tools
that enable process control and capability for continual
improvement
4. MSA: Measurement Systems Analysis: Guidelines for assessing
the quality of a measurement system where readings are
replicated
5. PPAP: Production Part Approval Process: Ensures product
consistently meets customer engineering specification
requirements during production run at the quoted production rate
 Questions?
Jd Marhevko
Phone: (419) 704-5603
Email: [email protected]
 Concepts by ASQ Automotive Division
APQP: Advanced Product Quality Planning 04/2018; Jd Marhevko

FMEA: Failure Mode

PPAP: Production Effects Analysis.
Part Approval Design (DFMEA) &
Process Process (PFMEA)
Five Levels of PPAP Approval: The Golden Circle of Effect
1. Warrant for appearance items (symptom), Mode and Cause:
2. Warrant with samples and
limited supporting data The Alphabet Soup
3. Warrant with samples & all supporting data
4. Warrant with customer specified requirements
of Core Tools
5. Warrant with product samples & all supporting ASQ-Auto.org
data at the supplier location A blended PFMEA & Control Plan is a Dynamic
A Product Sample Warrant (PSW) acts as a cover Control Plan (DCP). The six FMEA steps are:
sheet to compile up to ~20 different types of the
process evidence (many covered by the core tools)

Variables & Attribute Charting. Prior to starting:

1. Confirm effective MSA (GRR/AAA)
2. Ensure Cpk > 1.67 or Ppk > 1.33 (for safety items)
MSA: Evaluate for shifts, trends or points outside the SPC:
control lines. If this happens, process control is
Measure- difficult to sustain. Sources vary on Statistical
Calibration, resolution and error management to
ment avoid a (rejecting good) and b errors (passing
USL
Shift
0 Process
bad). GRR & AAA manage error within an MSA.
System Variables Gage Repeatability and Reproducibility Trend Control
Analysis (GRR): Reduce error; the higher the error, the
worse it is. Attribute Agreement Analysis (AAA):
“how many” points create a shift or trend;
Secures agreement with oneself, with a peer, and
LSL (5 or more)…This is defined by the business
to a standard. The higher the value, the better it is
 Where The Alphabets Fit…
APQP…5 (or more) Stage Gates
CONCEPT
INITIATION/ PROGRAM
APPROVAL APPROVAL PROTOTYPE PILOT LAUNCH
DFMEA
1. PLANNING
…PLANNING
MSA (GRR/AAA)
2. PRODUCT DESIGN & DEV
PFMEA/CP (DCP) Living documents
3. PROCESS DESIGN & DEV
MSA (GRR/AAA)
4. PRODUCT & PROCESS VALIDATION
PPAP
PRODUCTION
SPC
5. FEEDBACK ASSESSMENT & CORRECTIVE ACTION

DFSS DMAIC
Concept by Jd Marhevko 02/2018