Dfss jerry fix apr12

Design For Six Sigma
A Brief Overview
April, 2012
St. Louis ProductCamp
Presented By: Jerry Fix

Agenda
> Introduction
> What is DFSS (and “Lean” and 6Ϭ and…) ?
> Overview of the DFSS Process (DMADV/IDOV)
> Identify
> Design
> Optimize
> Validate
> DFSS Successes
> Some Best Practices
> Wrap Up

2

Introduction
> 10+ years in Product Management and
Marketing
> Practiced Lean and DFSS at Cooper
Bussmann, Energizer and Parker Hannifin
> DFSS is a philosophy – a way of doing
things
 It is NOT a certification, a specification or standard

> DFSS needs to be knowledge based
 Focus on knowledge generated to streamline the process –
not adherence to strict timelines
 Technical competence versus procedural compliance
 Become innovation driven
 Empower employees to make decisions based on knowledge
3
gained

What is DFSS (and lean, 6Ϭ etc…)
> Let’s start with 6Ϭ

4

> What is “good” or “bad” ?
 Ultimately the customer determines this
 Six Sigma is a way to compare the “Voice of the process”
with the “Voice of the customer”

5

> What does Six Sigma actually mean ?
 A measurement of process capability
 Reduce variations (defects)

6

> OK – So what’s Lean ?
 Lean is related to six sigma – but NOT the same
 Lean focuses on reducing the steps or parts of the process
 Simplify the “flow” of a process

7

> So Lean and Six Sigma Together
 Reduce steps in the process (simplify) and center the
process (reduce variations) so that the number “defects”
Lean (ie: unwanted results) gets very, very small

8

> Now that we’ve looked at 6Ϭ and Lean, what
about DFSS
 We now know that six sigma is a set of tools to reduce the
variability in a process
 Lean is a set of tools to reduce the complexity (steps) in a
process
> The end result of the “process” being
standardized and simplified is your product
or service
> The traditional focus of six sigma and lean
is the manufacturing or production or
delivery (service) part of the process
> DFSS is focused on the research, design
and development part of the process
> Intent is to “design in” reduced variability
and increased simplification
9

> There is a cost associated with optimizing
and simplifying a process
> Not all costs are created equal, though

10

> DFSS has 4 main goals
 Reduce cycle time in the design and development process
 Reduce “Time To Money” or TTM
 Reduce the cost of poor quality
 Improve predictability (in cost, quality and delivery)

> Does it work ? GE thinks so
> When they implemented DFSS, they
documented the following:
 A +1ϭ increase in quality at launch versus previous
designs
 A 25% or greater DECREASE in time to market versus
previous launches
 An estimated total cost savings in resources utilized of 20%
- 40%

11

> So finally, why DFSS ?
 Lean Six Sigma is intended to fix known problems
 DFSS is intended to PREVENT UNKNOWN problems

UNKNOWNS 12

The DFSS Process
> The DFSS process is a structured way to
step through the design process
 Applying a variety of tools along the way

The traditional DFSS process steps

D M A D V
Define Measure Analyze Design Verify

The GE modified process steps

I D O V
Identify Design Optimize Validate

13

The DFSS Process

14

The DFSS Process
> Identify Phase
 The DFSS Scorecard
 Voice of Customer (VOC)
> Design Phase
 Translate VOC (requirements flow-down)
 Concept generation and selection
 TRIZ, Axiomatic Design, Pugh Concept Selection
> Optimize Phase
 Multiple Response Optimization
 Expected Value Analysis (using Monte Carlo sim.)
 Parameter Design
 Tolerance Allocation
> Validate Phase
 High Throughput Testing
 Discrete Event Simulation

15

The DFSS Process
> MANY tools are available to determine
outputs for each of the phases
 We will just highlight a few along the way

16

The Identify Phase
> Key questions to ask in the IDENTIFY
phase

1. Why are we working on this project ? What need does it
address and how has it’s value been determined ?
2. Who are the external customer who will benefit from this
project… who may NOT benefit from it ? Who are the
internal customers (stakeholders) that may be affected ?
3. What are the prioritized functional requirements (Critical
To Customer) in measurable terms – how do they relate to
our VOC data ?
4. How will we determine success ? What are the measureable
attributes of the project – and what are the goals for those
measurements ?

17

The Identify Phase
> Key questions to ask in the IDENTIFY phase

5. Do measurement systems exist to support the project ? If so,
what is the capability of the system ?
6. Have the performance CTC’s been entered into the DFSS
scorecard (more on this later) ? If so, is the data good ?
7. Is there a gap between the performance and goals ? If so,
what is the risk of this gap ?
8. Do we have the right people/skills to undertake this project ?
Are there resource gaps – and if so where are they ?

18

The Identify Phase
> The DFSS Scorecard

19

The Identify Phase
 Each identified items is then measured for its Defects per Unit
(DPU), target values are identified and then samples taken to
determine the mean values, deviations and ultimately a defects
per million (yield) which give us the “sigma” of the process

Part Scorecard

Process Scorecard

Performance Scorecard

20

The Identify Phase
 The end result is a rolled-up scorecard that provides a sigma
for each component and for the overall product/project

> Software Scorecard
 This same process can be done for software as well. It’s too
involved for this presentation – but software AND services can
be fully comprehended by DFSS

21

The Identify Phase
> Voice of the Customer (VOC)
 Basis can be viewed via the Kano model of customer
satisfaction

22

The Identify Phase
> Identify your customers – external AND internal
> Segment them in ways that make sense for the
project
 Age, income, geography, price etc.
> Obtain the customer requirements
 Some will be given to you by the customer directly
 However, customers do not always know what they want or cannot
verbalize it – so consider observing or interviewing
 Put yourself in their shoes
 Consider using existing customer data 3rd party research, user tests etc.
> Using your customer (functional) requirements,
you identify the performance requirements
 The performance requirements are a measure of the customer’s
functional requirement

23

Afford- Appeal Repair- Reli- Stor-
able ing able able able Safe Fun
Quality

Customer Attributes

Toy is reliable

Toy is inexpensive
Toy is easy to repair
R

Toy is easy to store away
el
at

Toy is safe for pre-schoolers
iv
Toy is fun for parent and child
e
Performance Requirements

Toy appeals to parent and child
Im
po
rta (CTC)
nc
e

0.06
0.25
0.07
0.10
0.07
0.20
0.25

Maximum launch velocity

Maximum mass launch capability

Storage volume
The Identify Phase

Mean cycles to failure

Breaking force of critical components

Repair part cost

Tensile strength of home-applied adhesive bond

Time for disassembly and reassembly

Material cost

Time to manufacture
> You now have the start of a VOC based House of

24

The Identify Phase
> The CTC’s are ranked by using the importance
rating. These ranked CTC’s from the HOQ1 then
become the side of HOQ2. This then repeats.

25

The Identify Phase
> The end result is a collection of data that allows
you to identify in order of importance the critical to
customer requirements. These are then used to
flow down through the design and manufacturing
process to identify all the critical performance
items.

26

The Design Phase
> Key questions to ask in the Design phase

1. What are the conceptual designs and technologies
necessary to support the functional requirements ?
2. What are the risks associated with these designs ? Have we
done FMEA or Pugh on the designs ?
3. What are the potential trade-offs to eliminate any identified
failure modes ?
4. Have we completed a requirements flow down maintaining
links to the CTC’s ?
5. What are the key input variables that affect the standard
deviations ?
6. What are the transfer functions for each critical output or
CTC ?
7. Have the DFSS scorecards been updated ?

27

The Design Phase
> Design Concepts
 Create alternative designs that fulfill the CTC’s
 Compare designs with the functional
requirements
 Choose the best design
 Assess risks of chosen design
> Selecting design concepts
 Pugh Concept Selection
 Functional Analysis System Technique (FAST)
 Axiomatic Design
 TRIZ
 Transfer Functions

28

The Design Phase
> Pugh Concept Selection
 Applicable to a variety of situations
 Prevents team from focusing on small number
of “pet” projects
 Customer requirement driven
 Excellent tool to aid decision making
 Provides good documentation

29

The Design Phase
> FAST
 Allows users to quickly design the key
functionality of the system

Heat Generate Cause Flood Break Press
Beverage Heat Reaction Chemical Water Cell Button

HOW ?

WHY ?

30

The Design Phase
> Axiomatic Design
 Helps design teams evaluate the “goodness” of
designs
 Decomposes customer requirements into 4
“domains”: Customer, Functional, Physical,
Process
> TRIZ
 Formed by Altshuller – Russian patent examiner post
WWII
 Breaks down all design into 39 problem parameters and 40
“inventive principles”
 Results in a 39 by 39 matrix that is then populated with the
inventive principles

31

The Design Phase
> TRIZ

32

The Optimize Phase
Key questions to ask in the IDENTIFY phase

1. Have we determined the optimal parameters for our design
?
2. How sensitive is the CTC performance measure to changes
in parameters ? Have done a sensitivity analysis ?
3. What are the sources of variability in our system ? Which
can we control and which can we not control ?
4. Is the variability supplier dependent ? If so, how can we
minimize this variability ?
5. Have we confirmed that the design is actually producible
(process capability) ?
6. What gaps and risks still exist ?
7. Have we updated the DFSS scorecards ?

33

The Optimize Phase
> Determining optimal parameters
 Multiple Response Optimization simulation
 Done computationally – example is SimWare ProAge, income,
geography, price etc.

Design Parameter 1

Design Parameter 2 Requirement 1

Requirement 2
Design Parameter 3 Process

Design Parameter 4 Requirement 3

34

The Optimize Phase
> Determining optimal parameters
 Monte Carlo simulation can also be used
 Expected Value Analysis (EVA), Robust Parameter Design, Tolerance
Allocation
> EVA
 If we know the input variable distributions (mean, std dev., shape) and
the transfer function, we can estimate the output and it’s distribution
characteristics
> Robust Parameter Design
 A process of finding the optimal mean settings of input variables to
minimize the resulting dpm (defects per million) or maximize the
“sigma” of the process
> Tolerance Allocation
 Look at all the input standard deviations of the system.
 Determine which have the largest impact on the output variations
 Focus on controlling those with the largest impact

35

The Validate Phase
> Key questions to ask in the VALIDATE phase

1. What validation testing has been done, and what are the results ?
2. Is there a large gap between actual capability and predicted
capability from the DFSS scorecard ?
3. What is the gap between current performance and the CTC’s ?
4. Is the existing production system under a state of quality control
?
5. What unintended consequences or problems of the design exist
or could exist ? What is the contingency ?
6. How has ALL information been documented and communicated
to stakeholders
7. What is the total project benefit in terms of quality, cost and
delivery ?

36

The Validate Phase
> Validate performance
> Perform sensitivity analysis
> Compare predicted and actual capability
> Do a gap analysis
> Update scorecards

> Validate Performance
 Actual values for critical parameters are compared
against predicted values
 Methods to use: prototyping, lab scale production,
test fixturing of sub assemblies

If the validation provides poor results…. A gap analysis
needs to be conducted

37

The Validate Phase
> One method to validate is called High
Throughput Testing (HTT)

38

The Validate Phase
> One method to validate is called High
Throughput Testing (HTT)
 Used to test many factors at once
 Uses a minimum number of runs
 Software driven
 Not DOE – though similar terminology
 Started in the chemical industry where many
compounds are combined together at different
strengths to produce a specific reaction

39

DFSS Success Stories
> General Electric GEMS Lightspeed CT Scanner

40

> Xerox “Green” Paper

41

> Xerox Printer Belt Tensioning System

42

Characteristics Of Successful
DFSS Implementation
> Commitment and leadership from the top

> Measureable “stretch” goals for each project

> Accountability for project success

> Involvement and commitment from EVERYONE

> Training and support for all the tools and
knowledge necessary to do it right

IT IS VERY EASY TO FOCUS ON THE LAST
ITEM – BUT WITHOUT THE FIRST 4 THE DFSS
IMPLEMENTATION CANNOT SUCCEED
43

DFSS Best Practices
> Train leadership to build organizational
commitment and momentum
> Begin using DFSS tools and techniques on ALL
product development projects and provide
COACHING
> Weave DFSS tools into your NPD process right
away
> Review progress based on DFSS metrics and
re-align as necessary
> Build the bridge to innovation – build and use
the transfer function
> Build a strong capability in DOE / HTT
> Establish a good process capability database –
both in house and with suppliers
> Link DFSS to business success
> View DFSS as a cultural change – a mindset –
not just a set of tools 44

Thank You !

Jerry Fix
(314) 255-9938
jcfix@att.net

45

Dfss jerry fix apr12

More Related Content

What's hot (17)

Viewers also liked (9)

Similar to Dfss jerry fix apr12 (20)

Recently uploaded (20)

Dfss jerry fix apr12