Six Sigma on Malaysian Technical Cooperation Program 2009.

1. MTCP Program
Six Sigma Introduction
15th June 2009
Topic: Six Sigma Introduction
Malaysia Productivity Corporation
(Statutory Body under MITI)
P.O Box 64, Jalan Sultan,
46904 Petaling Jaya
Website: www mpc gov my
www.mpc.gov.my
Consultant: Mohd Azlan Abas
Tel: +6 012 308 7421
Email: Mohdazlan.abas@gmail.com
1 of 60
Why Does One Need a Quality Initiative?
 Meet customer expectations for higher quality
 Provide a competitive differentiator in the service market
 Build greater pride and satisfaction in the team
 Drive other key goals: productivity and growth
Tangible Costs Intangible Costs
- Inspection - Expediting
- Scrap - Lost Customers
- Rework - Longer Cycles
- Warranty - Lower Morale
Enormous opportunity
2 of 60
1

2. Six Sigma Provides focus on
Critical to Quality (CTQ) Metrics in businesses
 Vital Few CTQs that apply to all Customers:
– Responsiveness
– Marketplace Competitiveness
p p
– On-time, Accurate and Complete Deliverables
– Product/Service Technical Performance
 Key CTQs for a company:
– Post Sales Issue resolution
– Service Delivery Span
– Contract Fulfillment
– Parts Fulfillment
– Pricing
– Customer Escalation Cycle Time
3 of 60
Six Sigma Provides focus on
Critical to Quality (CTQ) Metrics in businesses
Practical
Traditional
Problem
Approach
Define Measure
D fi & M
Statistical
Problem
Analyze
Statistical
Solution
6Quality Methodology
Systematic Approach Focusing p
Improve
on Statistically Significant Root
Practical
Causes & Solutions Solution
Driving Customer & Shareholder Benefits Control
4 of 60
2

3. What is Six Sigma?
Process to reduce defects per million opportunities
B DPMO
• From current levels to “Six Sigma”
• “Sigma” is standard deviation from the ideal 2 308,537
• Can be applied to all business functions
Can be applied to all business functions 3 66,807
66 807
» Manufacturing, Products, Transactions
» Service, Sales Support
4 6,210
Quantitative methodology 5 233
• Uses measurements and scientific process 6 3.4
3 to 6
... 20 000 times improvement ...
20,000
... A true quantum leap
5 of 60
Process = Hose
Four Important Properties:
(1) Centering
(2) Spread
(3) Shape
(4) Stability Over Time
10.5 10
9.5
Y axis = Weight (lbs)
6 of 60
3

4. Six Sigma Concept
Every Human Activity Has Variability...
Mean
Lower Upper
Customer
C t Customer
C t
Specification Specification
1 p(defect)
Target
Reducing variability is the essence of six sigma
7 of 60
Mean Specification
What is Sigma? Limit
Some
Chance of
Failure
1
The hi h th
Th higher the  3
number (Z) in front
of the sigma symbol
the lower the chance
of producing a Much Less
Chance of
defect Failure
1
6
Reducing variation is the key to reducing defects
8 of 60
4

5. Reducing Variation is the Key
What GE sees
Order by Order Delivery Times
Starting Point After Project
28 29
(Days) 18 6
6 10 13 17
23 12
5 4
8 10 What Customers feel
16 13 Mean
19 10
Big Change
h
33 20
11 13 30% improvement
Variance
Average 17 13 No Significant Change
9 of 60
Reference: Six Sigma Performance
3.8 Sigma Six Sigma
99% Good 99.99966% Good
• 20,000 lost articles of mail per • Seven articles lost per hour
hour
• Unsafe drinking water for • One unsafe minute every
almost 15 minutes each day seven months
• 5,000 incorrect surgical • 1.7 incorrect operations per
operations per week week
• Two short or long landings at • One short or long landing
most major airports each day every five years
• 200,000 wrong drug • 68 wrong prescriptions per
prescriptions each year year
10 of 60
5

6. Key Terms: Process or Activity
Process X’s
or Factors Outputs
X1
Y1
X2
X3
PROCESS Y2
Y3
X4
For any given product, procedure or transaction, there are inputs, a
process, and outputs. You will need to measure the outputs to quantify how
well you satisfy a CTQ requirement; the output measures are Ys To
Ys.
change the process performance however, you must find and change the
critical Xs.
Find and control the critical X’s
11 of 60
The People
GM Global
Quality
Master
Quality Leaders
Q
Black Belts
Greenbelts
Apply Six Sigma tools
and methodology in Develop tools and
Help choose projects,
everyday work. teaching materials.
interview Black Belt
candidates, tie projects Conduct training and
to business needs. communication
Remove barriers and Six Sigma
g sessions. Mentor
drive Six Sigma into the Black Belts Black Belts and their
culture of their functions. projects.
Project leaders, change agents,
expert application of tools, mentor
Green Belts and their projects.
12 of 60
6

7. DMAIC
Objectives Step A
• DEFINE Phase Purpose
– Step A: Customer & Project CTQs
– Step B: Team Charter (4‐Blocker)
– St C Th Hi h L l P
Step C: The High Level Process Map
M
• Change Acceleration Process (CAP)
– E = Q x A
– The Model
– Key CAP tools:
• ARMI
• GRPI
• Threat/Opportunity Matrix
• In/Out of the Frame
13 of 60
DMAIC
Customer & Project CTQs Step A
CTQ Definition and CTQ Elements
Product/
Cycle Time to Deliver
Process Drawings
Characteristic
From Notice To Proceed To
Customer Measure Delivery Time of Drawings
(Weeks)
Need
CTQ
On Time Target 13 Weeks
Delivery
Specification/
Tolerance 15 Weeks
Limit(s)
A performance standard translates customer needs into
quantified requirements for our product or process
14 of 60
7

8. A Project Team Charter DMAIC
Step B
Business Case (Problem statement) Potential Issues/ Speed Inhibitors
•Multilin has a small market share in Indonesia • Data collection
compared to the more established European relays
manufacturer. Need to improve our service to Project Team
customer and increase sales in Indonesia.
customer and increase sales in Indonesia
Name Org. Role % Ded.
Goal Statement
Vince Tullo Manager Champion
Stuward Thompson Manager Champion
•To increase sales figures in Indonesia in 2004 to
W.N. Yew Sales Leader 70%
$1.8M. Daniel Sutando Sales leader Member 10%
W.Y. Tan Sr App Engr Member 10%
• To ensure relevant projects are pursued and to Steven Tao BB Member 5%
improve the market coverage for distributors, EPCs,
and end users. Milestones
esto es Owner Date
Scope Included:
Project Charter Approved YWN /Tao Apr 30
• Multilin relays, Indonesia Market Measure /Baseline YWN/Tan July 30
Analyze WN/Tao Aug 30
Deliverables Improve YWN/Tao Sept 30
Control YWN/Tao Oct 30
Order growth for 2004 Close YWN/Tao Nov 30
15 of 60
DMAIC
WIRE MILL PROJECT Step C
2 3 4 5 6 7 8 9 10 11
12
correct wire on cold change wire correct wire on wire
size pulley weld reel speed size pulley build
stand # feet per on
out of damaged spool out of damaged reel
round pulley round pulley
1
measureme
nt accuracy
measureme
nt accuracy Xs
proper die
Input Legend
lubrication
l b i ti
direction of
critical
lubricant
flow
pH level controllabl
e
13 14 15 16 17 18 19 20 21 22 23 noise
correct accurat
e fat level
size
tension
setting pressure
out of temperature
round
Legend
1) Start of shift 13) Wire size change
2) Change final capstan felt 14) Stop machine
3) Check for excessive 15) Replace dies (as required)
slivers/fines
4) Check size
5) Check surface quality
16)
17)
18)
)
Restring new setup
Check wire size
Attach wire to spool
p
CTQ’s
A) Diameter
B) Out-of-round
)
Ys
6) Check stand supply
19) Set dancer air pressure C) Loops
7) Is reel spooled
D) Tangles
8) Connect wire to new reel 20) Crack valve
21) Check lubricant flow E) Pinchouts
9) Check size
22) Start spooler F) Surface
10) Check surface quality quality
11) Check for acceptable spool23) Start capstan
build
12) Store in enamel room
16 of 60
8

9. Objectives DMAIC
• Introduction to Measure
- Measure Phase Deliverables
- Using Statistics to Solve Problems
• Measure Step 1: Select CTQs
- Quality Function Deployment (QFD) Process
Quality Function Deployment (QFD) Process
- Process Mapping
- Failure Modes & Effect Analysis (FMEA)
- Pareto Chart
- Cause & Effect Diagram
• Measure Step 2: Define Performance Standards
- Performance Standard / Defect
/
- Basic Nature of Data
• Measure Step 3: Measurement System
• Introduction to Measurement Systems Analysis (MSA)
17 of 60
DMAIC
The Statistical Problem
Goal: Find the Relationship
Process - A Y = f(X1, …, Xn) Process - B


 
Shape of the Curves CHARACTERIZES the Process
p
Process B is Better than Process A*
* Assumes same scale
18 of 60
9

10. DMAIC
Select CTQ Characteristics Step 1
Quality Function Deployment (QFD) Process Mapping (Flow Chart)
Functional Requirements
(HOW’s)
Requirements
Customer
(WHAT’s)
Map Customer Needs to
Potential Hows
Identify Critical Few - Resource Planning Map “Information” Flow
Identify All Touch Points
19 of 60
DMAIC
Select CTQ Characteristics Step 1
Failure Modes & Effects Analysis (FMEA)
Process Step Potential Potential SE Potential O Current D R Action Owner
OR Failure Failure Effects V Causes C Controls E P Recommended
Part Number Mode C T N
Anticipate Potential Failures in Process / Products & Develop Proactive Mitigation Plans
D e fe c ts b y O p e ra tio n
20% of causes account for
1 00 0 0 00
100 80% of the problem
80
Percent
Count
60
50 0 0 00
40
20
0 0
W W
PE LO P LO
HA EF AT 1 INS 2 P H E3 RF
D e fe c t CH
.S
ST
AG CO HO
LE
X -R
AY HO
LE INS BE
NC
HO
L
LW
A TE
MA AL A
FIN FIN
C o unt 2 7 6 1 4 41 3 0 8 4 41 2 7 2 0 41 0 2 5 9 91 0 1 4 9 1 9 3 3 5 3 8 2 8 6 1 5 4 1 1 0 4 9 6 4 3 3 6 7 0 7
P erce nt 2 6 .2 1 2 .4 1 2 .1 9 .7 9 .6 8 .8 7 .9 5 .1 4 .7 3 .5
C um % 2 6 .2 3 8 .6 5 0 .6 6 0 .4 7 0 .0 7 8 .8 8 6 .7 9 1 .8 9 6 .5 1 0 0 .0
20 of 60
10

11. Performance Standards Objectives DMAIC
Step 2
• Determine the Performance Standard
• Define a Defect
- What are the customer’s acceptance criteria for the part/product or process?
• Established How to Measure the Quality of the Part /
Product or Process
- Where are the data coming from?
- How do you measure the process?
- What are the units of measure?
- Is it a discrete or continuous measure?
• Gained Consensus On the Performance Standard
21 of 60
DMAIC
What’s Performance Standards Step 2
• Are Requirement(s) or Specification(s) Imposed by the
Customer on a Specific CTQ
• Translate Customer Needs into Measurable
Characteristic
– Have Clear Operational Definition, i.e. Specifies What to Measure, How to
Measure & Collect the Data
– Specifies Target or Mean
– Impose Specification Limits
– Have Clear Defect Definition
CTQs Quantified – Everybody on Same Page
22 of 60
11

12. DMAIC
More Performance Standards . . . Step 2
Target
Good Product
Defective Product Tolerance Defective Product
= USL - LSL
LSL USL
Lower Spec Limit Upper Spec Limit
Specification limits are set in order to divide customer satisfaction from customer
disappointment. While the exact limits may not be explicitly stated by the customer
(and captured in Step 1), their specific values come from what the customer defines as
a defect.
23 of 60
DMAIC
The Basic Nature of Data Step 2
• Continuous Data
- Characterizes a product or process feature in terms of its size, weight,
volts, time, or currency
- The measurement scale can be meaningfully divided into finer and finer
increments of precision
- Distributions: To apply the normal distribution, one must necessarily use
continuous data
• Discrete Data
- Counts the frequency of occurrence: e.g., the number of times something
happens or fails to happen
- Is not capable of being meaningfully subdivided into more precise
increments
- The validity of inferences made from discrete data are highly dependent
upon the number of observations. The sample size required to characterize
a discrete product or process feature is much larger than that required
when continuous data is used.
- Distributions: The Poisson and binomial models are used in connection
with this type of data
24 of 60
12

13. DMAIC
Making Data Driven Decisions Step 3
• Six Sigma is all about reducing defects for the customer. Fewer defect result in
a more satisfied customer.
• The project team must decide what needs to be done to improve quality for
the customer based on actual data – measurements of the product or process.
• The measurement system (gauge) used to collect the project data has to be
The measurement system (gauge) used to collect the project data has to be
sufficiently good to allow the project team to make the correct decisions.
• A measurement system must deliver data that accurately represents the
project or product. It is defective if it does not.
• The Six Sigma project team becomes the customer for the measurement
process. There are many CTQs a project team must consider when evaluating
the quality of a gauge…
If the gauge isn’t good enough for the needs of the project,
th i ’t d h f th d f th j t
you have to fix it (using DMAIC) before you can move on!
25 of 60
The Measurement Process DMAIC
Step 3
In order to improve a product or process, we must measure it. We have
submit the output from that process to a second measurement process.
Parts
(Example) As a result, all of our
observations of the original
Inputs
Process
ocess
Outputs p
process are distorted byy
errors in our measurement
Inputs
system. We need to make
sure these error don’t
dominate our view of the
process!
• Observations
Measurement Outputs • Measurements
Process • Data
26 of 60
13

14. DMAIC
Accuracy (Bias) ‐ Shift in the Average Step 3
Inputs Process Outputs Inputs Measurement Outputs
Process
Actual(Part) + Meas. System = Observed(Total)
True Avg Bias Obs. Avg
Measurement System Bias –
Determined through “Calibration Study”
27 of 60
DMAIC
Measurement System Precision Step 3
Inputs Process Outputs Inputs Measurement Outputs
Process
Product Variability Measurement Total Variability
(part) Variability (Observed total)
actual(part) + meas. system = observed(total)
Measurement System Variability –
Investigated through “R&R Study”
28 of 60
14

15. DMAIC
Accuracy vs. Precision Step 3
x xx
x xx
x xx Accurate?
Accurate? x Yes No
Yes No
x xx
xx
x xx
Precise?
Precise? x Yes No
Yes No
Measured Value
The True Value
x
x x x
Accurate?
Accurate? x x x Yes
Yes No
x x x
No
x x x x x Precise?
Precise? x x Yes No
Yes No x
x
x
x
x
Accuracy: the difference between the observed average and the truth.
Precision: the amount of inconsistency between measurements.
29 of 60
Characteristics of a Measurement System DMAIC
Step 3
• Accuracy : Differences between observed average measurement
and a standard.
• Resolution : The smallest scale of the measurement.
• Stability : Do measurements change with time?
• Linearityy : Is the measurement proportional to the magnitude (size,
p p g ( ,
weight, etc) of the sample.
• Precision : Noise of the measurements.
– Repeatability: variation when one person repeatedly measures the
same unit with the same measuring equipment. Also called Equipment
Variation (EV).
– Reproducibility: variation when two or more people measure the same
unit with the same measuring equipment. Also called Appraiser
Variation (AV).
Variation (AV)
30 of 60
15

16. WHAT IS MINITAB?
• Statistical software package
• Has many of the tools needed to successfully analyze data with
rigor
- Graphs
- Statistical tools for data analysis
- Six Sigma Reports
• Descriptive Statistics
• Gage R & R
• Capability Analysis (ZST/ZLT)
• Graphing - Many Types! (Try them all!)
• Pareto, Fish bone
• Hypothesis Testing
• Product & Process Six Sigma Reports
• Generation of Test Plans for Designed Experiments
• Analysis of DOE results
• Statistical Process Control
A Great Toolbox for Six Sigma Projects!
31 of 60
What does Minitab Look Like?
Menu Bars: For
quick access to
common
commands.
Session Window:
Shows Minitab text
output (One only is
present)
Worksheet
Window: Kind of
like an Excel
worksheet (at
least one is
always present). A
tool is available to
manage multiple
worksheets.
32 of 60
16

17. Objectives DMAIC
• Calculate baseline capability of the process using either
continuous or discrete data
• Statistically define the improvement goals
y p g
• Generated a list of Statistically Significant Xs based on
analysis of historical data
• Identified which Xs to further investigate in the Improve
phase
• Gained consensus with the project team on the list of Xs for
Gained consensus with the project team on the list of Xs for
investigation
33 of 60
DMAIC
Step 4: Establish Process Capability Step 4
USL
Y = f (X 1 . . . X n)
The variation inherent to any dependent variable (Y) is determined by
the variations inherent to each of the independent variables. (X)
Poor Excellent
Process Capability Process Capability
Very High Very High Very Low Very Low
Probability of Probability of Probability of Probability of
Defects Defects Defects Defects
LSL USL LSL USL
Low Z High Z
Z is a Measure of Process Capability
34 of 60
17

18. Step 5: Define Performance Objective
p(x) Benchmark
Entitlement
Baseline
Defects
Benchmark: World-Class performance
Entitlement: The level of performance a business should be able to
achieve given the investments already made
Baseline: The current level of performance
Benchmarking Sets the Ultimate Goal, while Baselining
Takes Current Measurements to Monitor a Process
35 of 60
DMAIC
Step 6: Identify Sources of Variation Step 6
Y= f (X)
To get results, should we focus our behavior on the Y or X ?
n Y n X1 . . . Xn
n Dependent n Independent
n Output n Input-Process
n Effect n Cause
n Symptom n Problem
n Monitor n Control
Historically the Y, … with Six Sigma the Xs
36 of 60
18

19. Process Capability DMAIC
Step 4
Determining Process Capability (Steps 1 – 4)
1. Identify Requirements on the Customer or Internal Y
These are often expressed quantitatively as target values plus
specification limits. - Anything “outside of the specification limits” is a defect.
Target Value
g
(lower specification limit) LSL USL (upper specification limit)
90 100 110
Defects Defects
Tolerance
2. Determine Process Distribution
We expect variation in our process.
Mean = 102
The Y’s measured on a large number
Ys  = 5.0
of parts will form a distribution. For
simplicity, we’ll assume that they form
a normal distribution with a known
mean and standard deviation.
85 90 95 100 105 110 115
37 of 60
Process Capability DMAIC
Step 4
3. Superimpose your process distribution on to the target
and specification limits.
Target Mean = 102
 = 5.0
LSL USL
Probability of Probability of
a Defect a Defect
85 90 95 100 105 110 115
4. Apply some Basic Statistics.
1. We can relate our distribution to the standardized normal distribution.
• The total area under the standard curve = 1 (or 100%)
2. If we can determine the Z l corresponding to a specification li i
2 d i h Z-value di ifi i limit,
then we can calculate the area beyond that limit (out of specification)
• The area beyond a given specification limit is the fraction of the
population that is defective wrt that limit.
38 of 60
19

20. Process Capability DMAIC
Step 4
Putting it all together (Steps 1 – 4)
Total Area Outside of Specification Limits = (0.055 + 0.008) = 0.063
6.3% of what we produce is defective
90  102 Notice also that our 110  102
Z(LSL)   2.4 Z(USL)   1.6
5 process is not centered 5
Probability  0.008 on the target Probability  0.055
Target Mean = 102
 = 5.0
LSL USL
85 90 95 100 105 110 115
39 of 60
Single-
Single-Tail Z Table (values from 0.00 to 3.99)
Z 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
0 0.5000 0.4960 0.4920 0.4880 0.4840 0.4801 0.4761 0.4721 0.4681 0.4641
0.1 0.4602 0.4562 0.4522 0.4483 0.4443 0.4404 0.4364 0.4325 0.4286 0.4247
0.2 0.4207 0.4168 0.4129 0.4090 0.4052 0.4013 0.3974 0.3936 0.3897 0.3859
0.3 0.3821 0.3783 0.3745 0.3707 0.3669 0.3632 0.3594 0.3557 0.3520 0.3483
0.4 0.3446 0.3409 0.3372 0.3336 0.3300 0.3264 0.3228 0.3192 0.3156 0.3121
0.5 0.3085 0.3050 0.3015 0.2981 0.2946 0.2912 0.2877 0.2843 0.2810 0.2776
0.6 0.2743 0.2709 0.2676 0.2643 0.2611 0.2578 0.2546 0.2514 0.2483 0.2451
z 0.7 0.2420 0.2389 0.2358 0.2327 0.2296 0.2266 0.2236 0.2206 0.2177 0.2148
0.8 0.2119 0.2090 0.2061 0.2033 0.2005 0.1977 0.1949 0.1922 0.1894 0.1867
0.9 0.1841 0.1814 0.1788 0.1762 0.1736 0.1711 0.1685 0.1660 0.1635 0.1611
1 0.1587 0.1562 0.1539 0.1515 0.1492 0.1469 0.1446 0.1423 0.1401 0.1379
1.1 0.1357 0.1335 0.1314 0.1292 0.1271 0.1251 0.1230 0.1210 0.1190 0.1170
1.2 0.1151 0.1131 0.1112 0.1093 0.1075 0.1056 0.1038 0.1020 0.1003 0.0985
1.3 0.0968 0.0951 0.0934 0.0918 0.0901 0.0885 0.0869 0.0853 0.0838 0.0823
Calculates the 1.4 0.0808 0.0793 0.0778 0.0764 0.0749 0.0735 0.0721 0.0708 0.0694 0.0681
1.5 0.0668 0.0655 0.0643 0.0630 0.0618 0.0606 0.0594 0.0582 0.0571 0.0559
probability of a USL 1.6 0.0548 0.0537 0.0526 0.0516 0.0505 0.0495 0.0485 0.0475 0.0465 0.0455
1.7 0.0446 0.0436 0.0427 0.0418 0.0409 0.0401 0.0392 0.0384 0.0375 0.0367
defect above the 1.8 0.0359 0.0351 0.0344 0.0336 0.0329 0.0322 0.0314 0.0307 0.0301 0.0294
1.9 0.0287 0.0281 0.0274 0.0268 0.0262 0.0256 0.0250 0.0244 0.0239 0.0233
given value of z 2 0.0228 0.0222 0.0217 0.0212 0.0207 0.0202 0.0197 0.0192 0.0188 0.0183
2.1 0.0179 0.0174 0.0170 0.0166 0.0162 0.0158 0.0154 0.0150 0.0146 0.0143
2.2 0.0139 0.0136 0.0132 0.0129 0.0125 0.0122 0.0119 0.0116 0.0113 0.0110
2.3 0.0107 0.0104 0.0102 0.0099 0.0096 0.0094 0.0091 0.0089 0.0087 0.0084
2.4 0.0082 0.0080 0.0078 0.0075 0.0073 0.0071 0.0069 0.0068 0.0066 0.0064
2.5 0.0062 0.0060 0.0059 0.0057 0.0055 0.0054 0.0052 0.0051 0.0049 0.0048
2.6 0.0047 0.0045 0.0044 0.0043 0.0041 0.0040 0.0039 0.0038 0.0037 0.0036
2.7 0.0035 0.0034 0.0033 0.0032 0.0031 0.0030 0.0029 0.0028 0.0027 0.0026
2.8 0.0026 0.0025 0.0024 0.0023 0.0023 0.0022 0.0021 0.0021 0.0020 0.0019
2.9
29 0.0019
0 0019 0.0018
0 0018 0.0018
0 0018 0.0017
0 0017 0.0016
0 0016 0.0016
0 0016 0.0015
0 0015 0.0015
0 0015 0.0014
0 0014 0.0014
0 0014
3 0.0013 0.0013 0.0013 0.0012 0.0012 0.0011 0.0011 0.0011 0.0010 0.0010
3.1 9.68E-04 9.36E-04 9.04E-04 8.74E-04 8.45E-04 8.16E-04 7.89E-04 7.62E-04 7.36E-04 7.11E-04
3.2 6.87E-04 6.64E-04 6.41E-04 6.19E-04 5.98E-04 5.77E-04 5.57E-04 5.38E-04 5.19E-04 5.01E-04
3.3 4.83E-04 4.67E-04 4.50E-04 4.34E-04 4.19E-04 4.04E-04 3.90E-04 3.76E-04 3.62E-04 3.50E-04
3.4 3.37E-04 3.25E-04 3.13E-04 3.02E-04 2.91E-04 2.80E-04 2.70E-04 2.60E-04 2.51E-04 2.42E-04
3.5 2.33E-04 2.24E-04 2.16E-04 2.08E-04 2.00E-04 1.93E-04 1.85E-04 1.79E-04 1.72E-04 1.65E-04
3.6 1.59E-04 1.53E-04 1.47E-04 1.42E-04 1.36E-04 1.31E-04 1.26E-04 1.21E-04 1.17E-04 1.12E-04
3.7 1.08E-04 1.04E-04 9.96E-05 9.58E-05 9.20E-05 8.84E-05 8.50E-05 8.16E-05 7.84E-05 7.53E-05
3.8 7.24E-05 6.95E-05 6.67E-05 6.41E-05 6.15E-05 5.91E-05 5.67E-05 5.44E-05 5.22E-05 5.01E-05
3.9 4.81E-05 4.62E-05 4.43E-05 4.25E-05 4.08E-05 3.91E-05 3.75E-05 3.60E-05 3.45E-05 3.31E-05
40 of 60
20

21. Identify Variation Sources DMAIC
Step 6
Goal: Process on Target with Minimum Spread
Desired Problem with Spread
Accurate but Current
not Precise Situation
LSL T USL
Problem with Centering – Not on Target
Current
Desired
Situation
Precise but not
Accurate
LSL T USL

Is the Problem Centering, Spread, or Both?
41 of 60
DMAIC
Continuous Data Analysis Road Map Step 6
One Sample Two Samples Multiple Samples
Study Stability
(Run Chart)
Study Shape
(Histogram, Dot Plot, Normality)
Study Spread Data paired?
(Chi Square-Test) Study Spread
(Homogeneity of Variance)
Study Spread Study Centering
(F test)
(F-test) (Paired t test)
t-test)
Study Centering Study Centering
(1 sample t-test) (ANOVA)
Study Centering
(2 sample t-test)
42 of 60
21

22. Descriptive Statistics DMAIC
Step 6
Stat > Basic Statistics >
Display Descriptive
Statistics
Provides summary report
of basic statistical
information
p-value > .05 Cannot
Histogram Reject H0
Minitab Output No evidence that
data is non-normal
Dot Plot
Basic Statistical
Information
Confidence Intervals
43 of 60
DMAIC
X and Y Data Correlation Step 6
25
Strong Positive Correlation
20
15
Y
10
5
0
0 5 10 15 20 25
X
25
No Correlation
20
15
Y
10
5
0
0 5 10 15 20 25
X
44 of 60
22

23. Improve Phase Objectives
• The benefits of Design of Experiments (DOE’s)
• Key concepts and terms associated with DOE’s
• Performing a simple full factorial and fractional DOE’s and
interpreting the results
• Awareness of screening designs and higher level response
surface designs
45 of 60
DMAIC
What’s Improve Phase About. . .
• Develop an Improvement Strategy
• Determine which candidate x’s identified in the Analyze Phase
are truly “critical X’s”.
• If possible, determine a quantitative transfer function that
relates your Y to these critical X’s
• Identify Improvement Actions
• Determine optimal settings for the X’s
• Show the impact of the changes on meeting
project or business objectives. Y = f(x)
• Validate the Improvement
• Demonstrate the validity of your identified improvement
actions via additional experiments or a pilot study
• Develop a Plan to Implement the Change
It’s More than Just Designed Experiments
46 of 60
23

24. DMAIC
Common “Improve” Tools
Basic Intermediate Advanced
 Process Map  DOE  DOE
 Fishbone  Full Factorial  Response Surface
 Box Plot  Fractional  Taguchi (Inner /
Factorial Outer Array)
 Time Order Plots
 Intro to  Simulation Models
 Hypothesis Tests Response
 Linear Regression Surface  Already Covered
 Covered in Improve
 Mistake Proofing  Multivariate  Covered in DFSS
Regression  Adv. Level III e.g. ProModel
LOW Problem Sophistication
• Complexity • Risk HIGH
Match the Tool to the
• Business Impact Problem • Data Availability
47 of 60
DMAIC
DOE – Design of Experiments Steps 7-8
Y = f (x1, x2, x3,……xn)
Response (Y) Factors (x’s)
(x s)
• The measured outcome of an • The critical X’s which determine the
experiment response,Y
• The value observed for the CTQ • They can be categorical or
being explored numerical
Ranges Levels
• The extreme values for each factor • In DOE’s we investigate the effect
determines the range for that factor of each factor at more than one
- the region of interest/investigation setting or value
48 of 60
24

25. DMAIC
Benefits of DOEs Steps 7-8
Classical Approach
OFAT - One Factor at a Time
• Change one variable, X2, 100 90
while holding all others
constant.
80
• Find a maximum 70
• Hold X2 at the 60
“maximum effect” level and
Factor X1
repeat the process for the other variables.
OFAT
• Requires more experiments than a DOE
• Becomes unmanageable as the number of factors increases
• Can be very expensive and time consuming – and may not work very well
49 of 60
Benefits of Design of Experiments DMAIC
Steps 7-8
DOE Approach
• Select factors and levels
• Select mathematical model
designed to obtain maximum
information for the number 100
00 90
of factors/levels selected.
• In your experiments change 80
the factor levels in a systematic 70
manner so that all coefficients in 60
the model can be uniquely computed.
(Orthogonality) Factor X1
• Solve the resulting set of simultaneous equations to obtain the coefficients.
coefficients
• Use statistical tests to determine if the coefficients are statistically significant,
and if the resulting model (transfer function) is adequate.
• Use the results of your DOE to plan the next DOE (if needed).
50 of 60
25

26. DMAIC
Screening Designs Steps 7-8
The Team’s understanding at
the beginning of the project What Mother Nature Knows
The following 7 Factors may be critical X's Yield = 64.25
+
A. Temperature (160C - 180C) 11.50*A
B. Monomer Concentration (20% - 40%) -
C. Catalyst Vendor (Sally - Ed) 2.50*B
D. Stirring Speed ( 50 RPM - 100 RPM) +
E. Monomer Purity ( 90% - 98%) 0.75*C
F. Pressure ( 100 PSI - 500 PSI) +
G. Acetone/Methanol Ratio - ( 0.25 - 0.50) 5.00*A*C
We need an efficient method for screening these
"candidate critical X's"
so that we can identify the 'Vital Few"
51 of 60
DMAIC
Objectives
• In the physical world, the law of Entropy explains the gradual loss of
order in a system. The same law applies to business processes.
• Unless we add “energy” (in the form of documentation and ongoing
process controls), processes will tend to degrade over time, losing the
gains achieved by design and improvement activities.
• The quality plan is the structure through which we add this “energy”
to business processes. This is Control and the main objectives:
– To make sure that our process stays in control after the solution has been
implemented.
– To quickly detect the out of control state and determine the associated
special causes so that actions can be taken to correct the problem before
special causes so that actions can be taken to correct the problem before
nonconformance are produced.
52 of 60
26

27. DMAIC
Control –
Control – Keep It On Target
• Focus on the Right CTQ
• Quantify the Problem
Define • Determine the Drivers
Measure Y = f(X)
Analyze • Identify Needed Change
Improve • Implement the Change
• Validate Measurement System (Xs)
• Determine Process Capability
• Develop/Modify Quality Plan
Control > Process Documentation
> Process Controls
• Implement Process Controls
• Audit Plan Established
• Transition to Operating Owners
53 of 60
DMAIC
On the Lookout for Special Cause
• Common cause variation
– Natural variability
– Random
– Inherent in the process
• Special cause variation
– May be caused by operator errors, adjusted machines, or
defective raw materials
– Generally large when compared to the common cause variation
– Considered an unacceptable level of process performance
• Special causes tend to cause a process to shift out of
control where. The output does not meet the desired
specifications.
54 of 60
27

28. DMAIC
What is a Process Control System?
• A Process Control System (PCS)
– strategy for maintaining the improved process performance over time
– identifies the specific actions and tools required for sustaining the
process improvements or gains
• A control system may incorporate
y y p
– Risk Management
– Mistake‐proofing devices
– Statistical process control (SPC)
– Data collection plans
– Ongoing measurements
– Audit plans
– Response plans*
Response plans
– Product drawings
– Process documentation
– Process ownership
55 of 60
Statistical Process Control DMAIC
Step 12
Statistical -- Probability based decision rules.
Process -- Any repetitive task or steps.
Control -- Monitoring of process performance.
SPC will signal when the
process is “out-of- control”.
Your Mission is to find out why and take
corrective action!
56 of 60
28

29. Control Chart Components DMAIC
Step 12
Average Chart
Statis (mean/defects)
0.065
Upper Control Limit
measured
Grand Average
0.060
Central Line
stic
Lower Control Limit
0.055
0 5 Sample / Subgroup (time ordered)
10 15 20 25
Monitors Shift
Variation Chart
e/Sigma)
0.010
Upper Control Limit
red
Statistic (Range
measur
0.005
0 005
Average Range/Sigma
0.000
Central Line
Lower Control Limit
Sample / Subgroup (time ordered)
Monitors Drift
57 of 60
Significance of 3s limit DMAIC
Step 12
A Control Chart is a graphic display of a
continuing two tailed test with HO and HA
defined as:
Ho: iHa: i
/2
/2
UCLx
X
LCLx
/2
• For 3 limits,  = 0.00135. approximate confidence level is 99.7%.
• 3  limits provide good sensitivity to change with low potential for over-
reacting when the process is stable.
58 of 60
29

30. Types of Control Chart DMAIC
Step 12
Variable Chart (Continuous) Attribute Chart (Discrete)
• Uses Measured Values • Defects: Number of non
conformance in a part
– Cycle Time, Lengths,
• Defective: Pass/Fail,
Diameters, Drops, etc.
Good/Bad, Go/No-Go
• Generally One Characteristic Per
y Information
Chart • Can Be Many Characteristics
• More Expensive, But More Per Chart
Information • Less Expensive, But Less
Information
High or Low
Volume? Constant Variable
Lot / Unit Size Lot / Unit Size
Low
L High
Defects Poisson
c u
Individuals & X-Bar &
Moving Range
Defective np p Binomial
Range
59 of 60
THANK YOU
60 of 60
30