SlideShare a Scribd company logo
6
1
Six Sigma
an
Overview
6
2
Emerging Paradigm
• Rapid Pace of technology
change
• Shrinking Life cycle
6
3
How to cope with the Emerging
Paradigm?
Lean Thinking!
6
4
Lean Thinking
‘Lean’ is for increasing the speed by
eliminating wastes, where as
‘SIX SIGMA’ is for improving Quality
Combined together they make our
Operations excellent.
6
5
WR RR
WW RW
BUSINESS EXCELLENCE
Strategic
Excellence
Operations
Excellence
6
6
Strategy Excellence
(Doing the right things)
1.Corporate Level Strategy
• What businesses to go into
• Merger and Acquisition
• Diversification
• Allocation of Resources
• Such plans concerning
matters of vital, long term
and continuing importance to
organization
(After studying the Economical, Technological,
Social and Political facets)
6
7
Strategy Excellence
(Doing the right things)
2. Business Level Strategy
(SBU/Unit)
• Competing in a given business or
Product area through
OPERATIONS EXCELLENCE
6
8
Operations Excellence
(Doing things right)
Includes :
• Marketing Excellence
• D & E Excellence
• Manufacturing & Service
Excellence
6
9
SIX SIGMA
- A means to realize the
philosophy and values
associated with TQM
6
10
- the ultimate aspiration
• SIX SIGMA
- the threshold of excellence
• ZERO DEFECTS
Dr. Mikel J. Harry
6
11
Missing Strategy for QI
Tools &
Techniques
Strategy
Discipline
Philosophy
& Vision
Quality Vision
existed for years
(ISO, “O defects”…)
What was
missing --
Six Sigma!
Many tools
existed for years
(SPC,DOE...)
6
12
What Is Six Sigma ?
 IT IS A STATISTICAL
MEASUREMENT
(HOW GOOD ?)
- Dr. Mikel J. Harry
 IT IS A BUSINESS
STRATEGY
(COMPETITIVE
EDGE).
 IT IS A PHILOSOPHY.
(WORKING SMARTER
AND NOT HARDER).
6
13
Example: OFFICE IN-PUNCH time
Historic data of IN-PUNCH timings
for the last 50 weeks.
50 Weekly Average values
Target time = 9.00hrs
USL = 9.18 hrs
LSL = 8.42 hrs
DATA:
Understanding Six Sigma
6
14
906
902 910 914 918
846
842 850 854 858

 = 06
USL
LSL
= 900
= 900

T
3
6
15
T

906
902
900
910
846
842 850 854 858
838
834
= 900
= 852
USL
LSL
6
16
61
T

1
USL
LSL
1 = 03
= 900
= 900
905
907
909
853
851 855
857
859 901
903
842 918
6
17

 = 06
USL
LSL
3
= 900
909
851
1
USL
842
1= 03
61
918
= 900
LSL
906
902 910 914 918
846
842 850 854 858
T

6
18
Quality Improvement
is
Problem with Spread
Accurate but not Precise
Desired
Current
Situation
LSL USL
T
Shrinking the variation
&
Problem with Centering
Shifting the mean to the target
d
Precise but not Accurate
USL
LSL
Current
Situation
Desired
6
19
Sigma Scale
(Distribution Shifted 1.5)
2
3
4
5
6
308,537
66,807
6,210
233
3.4
Z
Defects
PPM
7.5 0.001
6
20
Six Sigma Centered
Short Term
Zst = 6.0
DPPMst = .002
Zlt = 4.5
DPPMlt = 3.4
USL
LSL
3.4 ppm
LSL USL
4.5
T

.000 ppm
Six Sigma Shifted 1.5
Long Term
Short Term / Long Term
USL
LSL ± 6
.001 ppm .001 ppm
USL
T

7.5
6
21
SIGMA SPELLING
3 1.5 misspelled words per
page in a book
4 1 misspelled word per 30
pages in a book
5 1 misspelled word in a
set of encyclopedias
6 1 misspelled word in all
of the books contained in
a small library
Practical Meaning
6
22
Understanding The Difference
4 Capability:
Defect Dots = 1849
6 Capability:
Defect Dots = 1
6
23
d
1 2 3 4 5 6 7 8 9 10
a          
b          
c          
d          
Units = 10
Opp/unit = 4 DPO = 0.25
TOP = 40
Defects = 10 Z = 2.17
DPO  Probability of Defect
c
b
a
Z calculation for Discrete Data
6
24
1 2 3 4 5 6 7 8 9 10
a          
b          
c          
d          
Units = 10
Opp/unit = 4 DPO = 0.025
TOP = 40
Defect = 1 Z = 3.46
DPO  Probability of Defect
d c
b
a
Z calculation for Discrete Data
6
25
6
26
SIX SIGMA
WHAT IS DIFFERENT ?
• SIX SIGMA ORGANIZATION
• DISCIPLINED BREAKTHROUGH METHODOLOGY
• UNIQUE TRAINING METHODOLOGY
• RECOGNITION SYSTEM
6
27
The Six Sigma 0rgn.
Grand
Champion
Master
Black Belts
Full Time
Six Sigma Black Belts
Full Time
Six Sigma Green Belts
Part Time
Deployment Teams
Champions
6
28
Role of Champion
Create the vision of Six Sigma for the
company
Select high-impact projects
Review regularly to see that Black Belts
/ Green Belts complete the project.
Support development of “statistical
thinking”
Ask Black Belts / Green Belts many
questions to ensure that they are
properly focused.
Realize the gains by supporting Six
Sigma projects through allocation of
resources and removal of road blocks.
Hold the ground by implementing Black
Belt / Green Belt recommendation.







6
29
Role of MBB
Assist champion in the
identification of projects.
Develop and deliver training
to Black Belts / Green Belts.
Coach and support Black
Belts in project work.
Participate in project reviews
conducted by the champion
to offer technical expertise.
Facilitate sharing of best
practices across the
company.
*
*
*
*
*
6
30
Role of Black Belt
Be the change agent of the
company-(full time at least
for 2 years)
Be the Six Sigma Break though
strategy expert and enthusiast.
Support Green Belts in addition
to doing own’s projects.
Influence without direct authority.
Identify barriers and stimulate
champion thinking.





6
31
Role of Green Belt
Do projects on a part time
basis, while performing the
regular duties
Continue to learn and practice
the six sigma methods after
project completion.


6
32
SIX SIGMA
WHAT IS DIFFERENT ?
• SIX SIGMA ORGANIZATION
• DISCIPLINED BREAKTHROUGH METHODOLOGY
• UNIQUE TRAINING METHODOLOGY
• RECOGNITION SYSTEM
6
33
2 to 3 Sigma = 5X improvement
3 to 4 Sigma = 10X improvement
4 to 5 Sigma = 27X improvement
5 to 6 Sigma = 70X improvement
So... 3 to 6 Sigma = 19600X improvement!!!
Breakthrough Improvement
6
34
Sweet Fruit
Design for Six Sigma
Bulk of Fruit
Process Characterization
and Optimization
Low Hanging Fruit
Seven Basic Tools
Ground Fruit
Logic and Intuition
3 Level
Achieving Six Sigma is like reaching for the fruit at
the top of a tree...it gets progressively harder to do!
4 Level
5 Level
6
35
Product = ƒ (design, manufacturing process)
Focus on this is
DMADV
Focus on this is
DMAIC
The benefits can only be realized if the ‘design’
and ‘manufacturing Process’ can play together
Six Sigma Methodologies
6
36
Define
Improve
Control
Measure
Breakthrough Methodology
Analyze
Operations
Six Sigma
6
37
Define
Design
Verify
Measure
Breakthrough Methodology
Analyze
Design for
Six Sigma
6
38
Design Opportunity
Difficult to see/predict
Easy to fix
Easy to see
Costly to fix
Defects are:
$
Research Design Prototype Production Customer
Most DMAIC Six Sigma effort is here.
Cost
to
Correct
Quality
Moving upstream (DFSS) increases return
on investment
6
39
SIX SIGMA
WHAT IS DIFFERENT ?
• SIX SIGMA ORGANIZATION
• DISCIPLINED BREAKTHROUGH METHODOLOGY
• UNIQUE TRAINING METHODOLOGY
• RECOGNITION SYSTEM
6
40
Six Sigma Training-
What is different?
• Six sigma Training is project based
• The projects are linked to the bottom line
• The computer is used extensively in Six Sigma
training
• Six Sigma training methods involve roadmaps to
explain the application of improvement tools
• Six Sigma training focuses on how to apply the tools
to improve processes and not on the tools specifically
• Six Sigma instructors rely on a variety of media to
deliver material
6
41
Which factors have the greatest effect on results?
Are All These Factors Vital ?
4
Pin Position
Rubber Band
Stop Angle
1
3
3
Start Angle
1
“Wiffle”
“Solid”
Ball type Cup position
Hook position
6
42
SIX SIGMA
WHAT IS DIFFERENT ?
• SIX SIGMA ORGANIZATION
• DISCIPLINED BREAKTHROUGH METHODOLOGY
• UNIQUE TRAINING METHODOLOGY
• RECOGNITION SYSTEM
6
43
Recognition System
•No body gets promoted to an executive position at
GE without Six Sigma Training.
•“Get only your best people in the Six Sigma program
and give them the options”.
- Jack Welch
6
44
Latent Defects.
‘BILL SMITH’ an Engineer at ‘MOTOROLA’ was
studying the correlation between a product’s
life and how often it had been repaired during
the manufacturing process. In 1985 he
presented a paper that concluded that even if
an item had been found defective and
corrected during the manufacturing process,
other defects were bound to be missed. These
would then show up in the product’s early
days with the customer. When, however a
product was manufactured error- free, it rarely
failed during the aforementioned early days.
6
45
0%
5%
10%
15%
20%
25%
3 4 5 6 7
Cost of Poor Quality (% of Revenues)
Sigma level
Cost of Poor Quality Versus Sigma Level
- Thomas Pyzdek
6
Find and Control the Critical X’s
 Y
 Dependent
 Output
 Effect
 Symptom
 Monitor
 X1 . . . XN
 Independent
 Input-Process
 Cause
 Problem
 Control
Putting Science into Everything We Do!
f (X)
Y=
Establishing Y=f(X)
The Framework for Six Sigma
6 The Impact of Added Inspection
Escaping
PPM
Number of Consecutive Inspectors
The Y axis represents the undetected defects-per-million defects.
Each curve represents the inspection efficiency per independent inspection.
99% 90% 80% 70%
Example: If the likelihood of
detecting the defect is 70% and we
have 10 consecutive inspectors with
this level of capability, we would
expect about 6 escaping defects out of
every 1,000,000 defects produced.
1
10
100
1000
10000
100000
1000000
1 2 3 4 5 6 7 8 9 10 11
6 3.4 ppm
Inspection is an expensive and time
consuming way
to get to Six Sigma !
Focus must be on doing it right the
first time...
6 Unmasking the Hidden Factory
“Theoretical Cycle Time: The back-to-back process time
required for a single unit to complete all stages of a task
without waiting, stopping, or setups.”
T = test A = analysis F = fix
Product
A
F
A
F
Step
1
T Step
2
Floor Space Floor Space
Floor Space
T
Value Added
Non-Value Added
The Hidden Factory
6
49
Project Documentation
Why document?
To preserve the project details
To ensure recipe is not lost and problem does
not come back
To share the knowledge with others
To prevent re-inventing the wheel
Also an ISO 9000 requirement





6
50
Project
Monitor
Project Documentation….Contd.
Contents of the project binder
 How the project was selected (brief summary)
 Team members and contacts
 Discussions held with the various relevant depts. regarding
the projects
 All forms of data collected
 R-0 form
 Presentation details after characterization
 Presentation details after optimization
 Relevant office orders if any, issued for implementing the
recommendation
 Financial savings details
 Many more details beyond the presentation materials.
6
51
Chain of Causation
PROCESS VARIATION LEADS TO AN INCREASE IN DEFECTS,
COSTS AND CYCLE TIME.
- Dr. Mikel J. Harry
OUR SURVIVAL IS DEPENDENT UPON GROWING OUR BUSINESS.
OUR BUSINESS GROWTH IS LARGELY DETERMINED BY
CUSTOMER SATISFACTION.
PROCESS CAPABILITY IS GREATLY LIMITED BY VARIATION.
Q, P & D IS CONTROLLED BY PROCESS CAPABILITY.
CUSTOMER SATISFACTION IS GOVERNED BY Q,P & D.
6
52
 A Vision
 An Organization
 Champions
 Black Belts ,Green Belts and the Teams
 Master Black Belts
 Training and Application
 BB/GB must have a project
 Entire company must be involved and trained
 Business relevance is key to project selection
 Review Mechanisms
 Champions and operational managers responsibility
 A Rigorous Methodology
 DMAIC
 DMADV
 Recognition System
 Supplier Involvement
 Communication
 Six Sigma website
Operationalizing Six Sigma is the Key to Success
To summarize the elements of this strategy:
Summary
6
6
53
Methodology
Define
6
54
• Has a clearly defined problem
statement
* Is clearly linked to the customer
* Is relevant to the business and will
have positive impact
* Relates directly to the GB’s job
responsibilities, Gs&Os
* Has a clearly defined and
measurable defect (Y)
* Is manageable in scope
A Well-Defined Project:
Problem Statement
Definition looks easy - difficult for many
6
55
Good Example
In the last 3 months, 12% of our customers are late,
by over 45 days in paying their bills. This
represents 20% of our outstanding receivables &
negatively affects our operating cash flow
Poor Example
Our customers are angry with us and thus
delay paying their bills
What
When Consequence
Magnitude
Defining a Problem Statement
6
56
 Are you aware of any problem that a customer is having with the
products/services your organization offers?
 Is the Quality from competitive products/services better?
 Do you have a persistent problem that you have attempted to fix in
the past with limited success?
 Are your cycle times too long in any process?
 Are your costs too high in any process?
 Do you have regulatory/compliance problem?
Any of these questions can be addressed through a project
based on Six Sigma thinking
Project Selection Checklist
6
57
General problem: Electric
motor reliability is poor
Project scope: Reduce variation
in brush hardness
Electric motor issue
Brush wear issue
Brush hardness
variability
Project Focussing
6
58
Y1 – Electric motor reliability depends on:
X1 – Motor reliability
X2 – Controller reliability
X3 – Mechanical mounting integrity
X4 – Specific application or use
Y3 – Brush reliability depends on:
X1 – Assembly stack up issues
X2 – Brush brittleness issues
X3 – Contamination of the brush assembly
X4 – Spring rate or condition issues
X5 – Brush dimensional issues
X6 – Brush hardness issues
Y2 – The reliability of the motor itself depends on the reliability of the:
X1 – Stator X2 – Rotor X3 – Brush X4 - Housing
Y4 – Brush hardness issues depend on:
X1 – Mean brush hardness
X2 – Variation in brush hardness
Y1 = X1+X2+X3+X4
Y2 = X1+X2+X3+X4
Y3 = X1+X2+X3+X4 +X5 +X6
Y4 = X1+ X2
Project Focussing (contd.)
6
59
Airplane Factory
3.4
What is our factory’s
quality level?
Dents
0 40
Airplane
Wings Controls Fuselage
Sheet Metal Wheels Altimeter Throttle Doors Windows
Size Pressure Tread Cracks Calibration Grips Intermittent Air leaks Falls off Cracks Size
29 1080
40
3 40
1
16 280
5 480
40
1 40 2 160 0 160 0 40 3 40 0 40 5 40 1 80 2 80 0 40 1 40
0 40
2 320 8 120 7 120 3 160 1 80
Defects during
plane assembly
1 120
0 0 5 40 2 40 0 0 0 0
7 280
40
2
1
# of opportunities
Total # of defects
Total opportunities
Roll-up Example
3.4
6
60
Projects Selection
Workshop
6
61
Functions – (Individuals)
a) What for your function exists?(Role)
b) What is Quality in your function ?
 due date, cycle time, errors etc.
c) Who are your “customers”?
d) What are your customers’ major complaints/expectations
(Some times perennial complaints, other times sporadic).
(If you have records on these pl. go & collect the same)
e) Can we convert the complaints into measurable form?
(pl. try)
15 minutes.
6
62
Products
(Concerned Individuals)
a) List out the end Products your external customer
receives.
- 10 minutes
b) List out the complaints you have received on these
Products
(you may pl. go back to your work place and gather
information from your records)
- 30 minutes
6
63
Manufacturing Processes
(Groups)
a) List out the Manufacturing Processes
- 10 minutes
b) Make Process maps for all the above processes
- 15 minutes
c) Make it more detailed by walking through the
Processes - 30 minutes
d) Identify the steps which are yielding less
(lot of rework and also scrap)
(you may pl. go back and refer your records)
- 30 minutes
6
64
Transactional Processes
(Intra or Inter -Functional)
(individuals or groups)
a) Write the process map.
- (10 minutes)
b) Walk through the process and make the above
process more detailed and realistic.
(You may pl. go back to your places and sit with
your people.)
- (30minutes)
6
65
Typical Good Projects
1. Reduction of variation of Hfe of NPN transistors (BC547, BF
494) at low currents and operating currents.
2. To improve the wire bonding process in IC division. (focused on
BEL 1895 device)
3. Defect reduction in C1Bellow Assembly (Vac interrupters)
4. Reduction in rejection of sealing flange 7051 04850102 of V1
tubes
5. Minimisation of Insertion and Return loss in Multiplexers.
6. CT reduction from Sanction to Award of contracts in Services.
7. Reduction of cycle time in Internal Mailing system. (Same day
delivery)
8. Cycle time reduction from enquiry to quotation for spares in HF
division.
6
66
Methodology
Measure
6
67
 We don’t know what we don’t know.
 If we can’t express what we know in the form of
numbers, we really don’t know much about it.
 If we don’t know much about it, we can’t control it.
 If we can’t control it, we are at the mercy of chance.
Measurements Are the
Foundation for Six Sigma
0
20
40
60
80
100
1st
Qtr
2nd
Qtr
3rd
Qtr
4th
Qtr
East
West
North
6
68
Data: A collection of any number of related
observations on one or more variables.
Raw Data: Information before it is arranged or
analyzed by statistical methods.
Data Array: The arrangement of raw data by
observations in either ascending or descending order.
Data Point: A single observation from a data set.
What Is Data?
6
69
Population: A collection of all the elements we are studying and
about which we are trying to draw conclusions.
Sample: A collection of some, but not all, of the elements of the
population under study, used to describe the population.
Representative Sample: A sample that contains the relevant
characteristics of the population in the same proportions as they are
included in that population.
Advantages of sampling: a. costs less ,
b. takes less time ,
c. case of destructive testing,
The wolf did not have to eat the whole ox
to know the meat is tough.
Population And Sample
6
70
Statistics is a collection of techniques useful for making decisions
about a process or population based on an analysis of the
information contained in a sample from that process or
population.
When summarizing large data it is useful to distribute data into
small groups or classes.
Frequency distribution shows the number of observations in the
data set that fall into each distinct classes.
Histogram is a graph of the observed frequencies that fall into the
different classes.
Statistics means never having to say
you are certain !
What Is Statistics?
6
71
Discrete Data: (Attribute Data)
1. Data that take discrete values.
2. Always expressed in whole numbers.
3. Countable data.
4. Requires large sample size to characterize
product or feature.
e.g. No. of defects, good / bad , go-nogo etc.
Probability distributions applicable are
binomial , Poisson etc.
Continuous Data: (Variable Data)
1. Data that has continuous values
2. Can be expressed easily in
fractions or precise increments.
3. Measurable data.
e.g. Resistance of coil , thickness of
coating ,Voltage etc.
Probability distributions applicable
are normal, exponential etc.
Data : Discrete and Continuous
6
72
A measure of location indicating
the value of a typical middle
point of a distribution.
e.g. Mean , Median , Mode etc.
1. Mean: A central tendency measure representing the arithmetic
average of a set of observations.
2. Median: If the data set is arranged in order of magnitude ,the
middle value that divides the data set into halves is called median.
Quartiles : Values that divide data into four equal parts ( Q1, Q2
and Q3).
Percentiles : Values that divide data into 100 equal parts .
3. Mode: The value most often repeated in the data set. It is
represented by the highest point in the distribution curve of a data
set.
Measures of Central Tendency
6
73
Dispersion: The spread or
variability in a set of data.
Measure of Dispersion: A measure describing how the observations
in a data set or distribution are scattered or spread out. Measures the
precision of a process.
1. Range: The distance between the highest and lowest values in a data
set.
2. Deviation: The difference of a data point from the mean .
The sum of deviations of a data set from their mean is zero.
3. Mean Absolute Deviation: Average of the absolute deviations
from the mean .
Measures of Dispersion
6
74
4. Sum of squares ( SS) : Sum of the squared deviations
from the mean.
5. Variance: The average of the squared deviations from
the mean.
6. Standard Deviation: The root mean square deviation.
The positive square root of the variance .
Widely used measure of dispersion.
7. Inter-quartile range (Q3-Q1) / 2
Measures of Dispersion
6
75
Parameters And Statistics
File : wt-ss4.mtw.
1. Find mean, median and mode for weight data.
2. Find variance and Standard deviation also.
3. Check that sum of the deviations of all observations from its mean is zero.
Measure Statistics Parameters
(of sample) (of population)
Mean x 
Sum of squares ( x - x )2 ( x -  )2
Variance s2 = ( x- x )2 2 = ( x -  )2
n - 1 N
Std.Dev. s = (x - x )2  = ( x -  )2
n - 1 N
No. of observations n N
6
76
Symmetrical: A characteristic of a distribution
in which each half is the mirror image
of the other half.
Skewness: The extent to which a distribution of data points is
concentrated at one end or the other; the lack of symmetry.
Measure of the degree of asymmetry.
Positively skewed - curve that tails off toward the high end or right.
Negatively skewed - curve that tails off toward the low end or left.
Kurtosis: The degree of peaked ness of a distribution of points.
Other Characteristics
of a Distribution
6
77
1.
2.
3.
4.
5.
Leptokurtic
Mesokurtic
Platykurtic
Distribution shapes
6
78
 = s =
( X i - X )
2
S
i = 1
n
n - 1
?
If we have 4 numbers and you
know the average and 3 of the
numbers, you can derive the 4th
(You don’t have that 4th DF)
Degrees of freedom (n-1) is employed to derive an unbiased
estimator of the population standard deviation. (n-1) is the number of
independent contrasts out of n observations.
Obs. value contrast
1 6
2 4 2
3 6 -2
4 1 5
The fourth contrast (6-1) has no independent existence because its
value is known from adding the three contrasts (2-2+5) = 5.
Thus the degrees of freedom for four observations is three.
Degrees of Freedom
6
79
The Nature of Variation
5
4
3
2
1
5
4
3
2
1
Mean is Centered
( it is On Target), but
there is Large Variation
Precise but not
Accurate
Accurate
but not Precise
Mean is not Centered
( it is Off Target);
the Variation is Small
Accuracy refers to the closeness of the average of the
Measurements to the target value.
Precision refers to the closeness among
the individual measurements
6
80
Understanding Process
Variation
1.233 1.235 1.237 1.239 1.241 1.243 1.245 1.247
USL
T
LSL
Recognize that the process
width is independent of the
design width. In other words,
the inherent precision of a
process is not determined by
the design specifications.
Centered...
but
large variation!
6
81
Understanding Process
Centering
USL
T
LSL
Recognize that the process
center () is independent of the
design center (T). In other
words, the ability of a process
to repeat any given centering
condition is independent of the
design specifications.
1.233 1.235 1.239 1.241 1.243 1.245 1.247
1.237

Increase in
nonconformance due to
shift in process centering
5
4
3
2
1
Small variation...
but not centered!
Without deviating from the norm, progress is not possible!
6
82
What’s the probability of making it to the Limo before he
walks into the lamp post ?
- Parking lot is perfectly rectangular
- The end of the limo just touches the fence at the end of the parking spot
Probability
The Blindfold
Pub
Wind direction, straight out of
Northwest,10 MPH
Region of
better light
He walks at 50 ft per minute.
30 ft 60 ft
10
ft
50
ft
100 ft
Limo 15 ft x 6 ft, corner parking spot
Perimeter=1ft
6
83
The Blindfold...
Use this page if you wish to sketch out his path in order
to determine the answer
Pub
Probability…Contd.
Don’t use statistics as a drunken man uses lamp
post , for support rather than illumination.
6
84
Understanding the Histogram
A total of 1,000 parts were produced.
22 parts were greater than the USL.
31 parts were less than the LSL.
The probability of defect is 53/1000 = .053
The process yield is 1 - .053 = .947, or 94.7%
300
1.247
Height of the bar corresponds to the number of
times a measurement was observed within the
given interval width (on the X axis).
Total width of the
histogram is an estimate of
the process capability.
This is the range over
which the process operates
most of the time.
0
50
100
150
200
250
1.233 1.235 1.237 1.239 1.241 1.243 1.245
USL
LSL
Probability…Contd.
Process
Capability
Report
6
85
DPU and DPO
= defect
= 1 unit = 1 opportunity
1 unit = 4 opportunities
DPU = 1
dpo = 0.25
DPU = 2
dpo = 0.5
DPU = 3
dpo = 0.75
6
86
Unit of
Product
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
1 •
2 •
3 • • •
4 • • • • •
5 •
6 •
7 • • •
8 • • •
9 •
10 •
11 • •
12 • • •
13 • • •
14 • •
15 • • • •
16 • • •
17 • •
18 •
19 • •
20 • • • •
21 •
22 •
23 •
24 • •
25 • •
26 •
27 • •
28 • • •
29
30 •
Class Exercise: PCB Production
Each PCB
Has 10
Opportunities
for a Defect.
We Have
Produced a
Lot of 60
PCBs and
Inspected for
Defects:
DPU and DPO…Contd..
= Defect
= Opportunity
•
=
6
87
Do The Following
 How many total defects?
 How many total units?
 What’s the average Defects/Unit?
 How many Total Opportunities/unit?
 What’s the Defects/Total Opportunities?
 Fill in the observed ...
defects/unit observed
0
1
2
3
4
5
DPU and DPO…Contd..
6
88
Nature Of The Problem
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
1 •
2 •
3 • • •
4 • • • • •
5 •
6 •
7 • • •
8 • • •
9 •
10 •
11 • •
12 • • •
13 • • •
14 • •
15 • • • •
16 • • •
17 • •
18 •
19 • •
20 • • • •
21 •
22 •
23 •
24 • •
25 • •
26 •
27 • •
28 • • •
29
30 •
1) The defects are randomly distributed
2) 60 defects were observed out of 60 units produced
3) The defects-per-unit is 1.0
4) There are 10 opportunities for defect per unit of
product
Given the facts, what is the
likelihood of producing a
part with zero defects? In
turn, this guarantees no
rework or repair.
DPU and DPO…Contd..
6
89
Given the facts, what is
the likelihood of
producing a part with
zero defects? In turn,
this guarantees no
rework or repair.
In other words....
What is the probability of 1 unit having all 10 of its
opportunities as good ones.
Note: If an opportunity has a 10% chance of being bad,
then it has a 90% chance of being good.
DPU and DPO…Contd..
6
90
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
1 •
2 •
3 • • •
4 • • • • •
5 •
6 •
7 • • •
8 • • •
9 •
10 •
11 • •
12 • • •
13 • • •
14 • •
15 • • • •
16 • • •
17 • •
18 •
19 • •
20 • • • •
21 •
22 •
23 •
24 • •
25 • •
26 •
27 • •
28 • • •
29
30 •
= Defect = Opportunity = Unit of Product
1-.10 =.90
.10
Probability the Opportunity is Defective =
Defects per Opportunity(DPO)
.9010 = .34867844
Thus, the likelihood that any given unit of product will
contain zero defects is 34.87%
Defects: 60
Production: 60 Units
PCB Part
Probability the Opportunity is not Defective
DPU and DPO…Contd..
6
91
A given process has two operations. Each operation has a first-
time yield of 99 percent. The rolled-throughput yield equals:
There is a 98% probability that any given unit of product
could pass through both operations defect free.
Extending The Concept
Process Centered
Op 1
99%
Process Centered
Op 2
99% 98%
Output
x =
YRT = e-DPU
YFT =
Rolled-Throughput Yield
Classical First-Time Yield
s
U
Rolled-Throughput Yield
S = good parts of the final step, U = total parts of final step
6
92
The hidden operation
37 Units are guaranteed to pass because they contain no defects. A total of
63 units contain one or more defects. Of these, some will be repaired while
others will become scrap. In this case, 53 were repaired and 10 were scrap.
37 Units
d = 0
dpu = 1
100 Units
Submitted
d = defects
u = units
The
Hidden
Operation
Defect Number
Type Observed
A 30
B 10
C 0
D 20
E 40
Total 100
90 Units
Passed
(Final yield)
Verify
Not
OK
Scrap 10 Units
Operation
Rework
Scrap
Y .tp
Throughput yield
(Units with Zero defects)
63 Units
d  1
53 Units Repaired
6
93
Rolled-Throughput Yield
YIELD DECREASES WHEN COMPLEXITY INCREASES
DIE
BONDING
WIRE
BONDING
PROCESS
MAP
6  Process/Product
4  Process/Product
100
P(d) = 6.2 10-3
Yield = (1-p)400
Yield = 8,3 %
400
COMPLEXITY
Process Steps
or CTQ’s
QUALITY
P(d) = 3.4 10-6
Yield = (1-p)400
Yield = 99.8 %
P(d) = 6.2 10-3
Yield = (1-p)100
Yield = 53 %
P(d) = 3.4 10-6
Yield = (1-p)100
Yield = 99.96%
4  6 
DICING MOULDING TESTING
6
94
6
95
Why Six Sigma?
Yields thru Multiple Steps/Parts/Processes
Zst
(distribution shifted 1.5)
# of parts,
steps, or
processes
3 4 5 6
1 93.32% 99.38% 99.9767% 99.99966%
5 70.77% 96.93% 99.88% 99.9983%
10 50.09% 93.96% 99.77% 99.997%
20 25.09% 88.29% 99.54% 99.993%
50 3.15% 73.24% 98.84% 99.983%
100 53.64% 97.70% 99.966%
200 28.77% 95.45% 99.932%
500 4.44% 89.02% 99.830%
1000 0.20% 79.24% 99.660%
2000 62.79% 99.322%
10000 9.76% 96.656%
Yields thru Multiple Steps/Parts/Processes
Zst
(distribution shifted 1.5)
# of parts,
steps, or
processes
3 4 5 6
1 93.32% 99.38% 99.9767% 99.99966%
5 70.77% 96.93% 99.88% 99.9983%
10 50.09% 93.96% 99.77% 99.997%
20 25.09% 88.29% 99.54% 99.993%
50 3.15% 73.24% 98.84% 99.983%
100 53.64% 97.70% 99.966%
200 28.77% 95.45% 99.932%
500 4.44% 89.02% 99.830%
1000 0.20% 79.24% 99.660%
2000 62.79% 99.322%
10000 9.76% 96.656%
Reduce
Parts/Steps
Improve
Sigma per
Part/Step
Rolled-Throughput Yield
6
96
Probability Models
•Binomial Distribution
•Poisson Distribution
•Normal Distribution
Three Models That Have Frequent
Application for Products &
Processes:
Probability & Statistics:
If you have knowledge of a Population, Probability
is a tool you can use to determine how likely you
are to draw a certain Sample
If you have knowledge of a Sample, Statistics is a
tool you can use to draw conclusions about an
unknown Population
6
97
Normal Distribution
 Distribution of a continuous random variable.
 Gaussian distribution.
 Fits into actual observed natural phenomena
e.g. human characteristics,
process outputs etc.
 Unimodal and bell shaped.
 Mean lies at the center of the curve & is the highest point.
 Symmetrical about the mean.
 Median and mode coincides with the mean.
 The two tails extent indefinitely and thus never touch the
horizontal axis.
 Area under the curve = 1.
 Only mean and sigma required to make predictions.
 Shape is related to frequency distribution & histogram.
6
98
The Focus of Improvement
Leverage
variables
which
control
the Mean
Leverage
variables
which
control the
Standard
Deviation
Y = f ( X 1 , ... , X N )

 Scale of Y
Characterization
Some Xs might
affect the mean,
some might affect
the spread ()
some might affect
both.
Very Low
Probability
of Defects
Very Low
Probability
of Defects
LSL USL
Excellent
Process
Capability
Very High
Probability
of Defects
Very High
Probability
of Defects
LSL USL
Poor
Process
Capability
Capability
Low Z High Z
6
99
Standard Deviation.

Point of Inflection
1
T USL
p(defect)
3
The distance between the point of inflection and the mean
constitutes the size of a standard deviation. If three such
deviations can be fit between the target value and the specification
limit, we would say the process has “three sigma capability.”
Z =
USL - 

6
100
-3 -2 -1 0 +1 +2 +3 Z
1 and -1
PROPERTIES OF SD
1. 68.27 % of the observations are included between +
3. 99.73 % of the observations are included between +3
2. 95.45 % of the observations are included between +2 and -2
 and -3
Properties of Standard Deviation
6
101
Using Z as a Measure of Capability
As variation decreases, the
standard deviation (s) gets
smaller and capability
increases, which in turn
decreases the probability of a
defect.
s
z
USL
Xbar
3 Capability
USL
6 Capability
Xbar
Z =
SL - Xbar
s
Z = 6
Z = 3
6
102
What Is Probability Of a Defect when Z = 0.97?
Area to the right of
Z0 = 1.66x10-1
Probability = 0.166
or 16.6%
Probability
of a Defect
Example = 0.166
Z0
Z
Probability of a defect
Normal Distribution
Z 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
0.0 5.00E-01 4.96E-01 4.92E-01 4.88E-01 4.84E-01 4.80E-01 4.76E-01 4.72E-01 4.68E-01 4.64E-01
0.1 4.60E-01 4.56E-01 4.52E-01 4.48E-01 4.44E-01 4.40E-01 4.36E-01 4.33E-01 4.29E-01 4.25E-01
0.2 4.21E-01 4.17E-01 4.13E-01 4.09E-01 4.05E-01 4.01E-01 3.97E-01 3.94E-01 3.90E-01 3.86E-01
0.3 3.82E-01 3.78E-01 3.75E-01 3.71E-01 3.67E-01 3.63E-01 3.59E-01 3.56E-01 3.52E-01 3.48E-01
0.4 3.45E-01 3.41E-01 3.37E-01 3.34E-01 3.30E-01 3.26E-01 3.23E-01 3.19E-01 3.16E-01 3.12E-01
0.5 3.09E-01 3.05E-01 3.02E-01 2.98E-01 2.95E-01 2.91E-01 2.88E-01 2.84E-01 2.81E-01 2.78E-01
0.6 2.74E-01 2.71E-01 2.68E-01 2.64E-01 2.61E-01 2.58E-01 2.55E-01 2.51E-01 2.48E-01 2.45E-01
0.7 2.42E-01 2.39E-01 2.36E-01 2.33E-01 2.30E-01 2.27E-01 2.24E-01 2.21E-01 2.18E-01 2.15E-01
0.8 2.12E-01 2.09E-01 2.06E-01 2.03E-01 2.01E-01 1.98E-01 1.95E-01 1.92E-01 1.89E-01 1.87E-01
0.9 1.84E-01 1.81E-01 1.79E-01 1.76E-01 1.74E-01 1.71E-01 1.69E-01 1.66E-01 1.64E-01 1.61E-01
Z
Area under curve is the probability of a defect
Let us suppose that we calculate the
standard normal deviate for a given
performance limit and discover that Z =
0.97. The question becomes, “What
portion of the total area under the
normal curve lies beyond a Z value of
0.97?” Answering this question will
give us the probability of producing a
defect. Remember, the Z value is a
measure of process capability and is
often referred to as the “sigma of the
process,” not to be confused with the
process standard deviation.
Performance
Limit
6
103
For One-Sided Specs --
From Z table, for Z = 3.0, p(defect) = .00135
USL = 124
8
 = 100
Z =
|SL - |

|124 - 100|
8
= = 3.0
Z = |SL - |

Probability of a defect
6
104
ZUSL =
|USL - |

|124 - 100|
8
= = 3.0
USL = 124
8
 = 100
LSL = 76
ZLSL =
|LSL - |

|76 - 100|
8
= = 3.0
p(d)US for ZUSL = 3.0 is .00135 from table
p(d)LS for ZLSL = 3.0 is .00135 from table
From table, for p(d) = .0027, Z = 2.78
Z of Centered process
The file tra-pro
.mtw gives the
gain of the
transistors
taken at three
different
times.Analyze
each process.
TARGET = 100 ;
USL = 124 ; LSL = 76
6
105
(mean)

 (standard deviation)
Probability of defects: PUSL
USL
LSL
Target
Probability of defects: PLSL
|  -LSL |

ZLSL = PLSL
Table
| USL- |

ZUSL = PUSL
Table
PLSL +PUSL = PTotal
Z
Table
Z calculation with 2 limits and
process which is not centered .
What if I have two specifications (upper and lower) and
a process which is not centered?
Z of shifted process
6
106
Mode
Mean
Median
Standard Deviation
Zcalculated = ( MEAN- LSL)/
P(d) =((# of scores < LSL)/# of scores)
Ztable ( look up using the P(defect))
Class Exercise: Calculate the following for your assigned cricketer
Note: LSL = 44
Ramesh Siva David Vinod
16 52 54 34
9 69 64 49
100 40 65 49
56 45 45 39
40 67 65 40
50 72 43 48
9 70 65 47
45 45 78 39
20 51 65 46
105 69 82 49
6
107
Is the process capability adequate to meet the Specifications ?
Process potential = Tolerance = USL - LSL
index - CP Pro. Cap. 6
= 1 just capable (3 sigma process )
> 1 capable process (Cp = 2 for a 6 sigma process)
< 1 not capable of meeting the specifications.
Process performance index (used when process is not centered).
CPU = USL -  ; CPL =  - LSL ;
3  3 
CPK = Lower of CPU , CPL .
The inherent variability of a quality characteristic
that the process is capable of maintaining , when in
a state of statistical control.
Process capability = 6 Std. Devs
Process Capability
6
108
Process Capability Ratios
 The greater the design margin, the lower the Total Defects Per
Unit (TDU).
 Design margin is measured by the Process Capability Index (Cp)
(Maximum Allowable Range of Characteristic)
(Normal Variation of Process)
Cp =
LSL
Process Width
Design Width
USL
T
o
+3
-3
ZST = 3CP
Cp =
USL - LSL
±3 
6
109
Process Capability Ratios
Cpk = Cp (1 - k)
Where k is the percentage of the tolerance zone
consumed by the static mean shift
k =
T - 
(USL - LSL)/2
3.4 ppm
0 ppm
LSL USL
T
o 1
Cpk = Cp (1 - k)
k =
T - 
(USL - LSL)/2 4.5 lt
6 st
Example: Cp = 2 , k = .25
Cpk = 2 (1-0.25) = 1.5
6
110
Process Capability
6 is Cp=2
6
111
Process Capability
CP= 2
CP= 1
+61
- 61
USL
LSL
UCL
+31
LCL
- 31

T
UCL
+3
LCL
- 3
+3
- 3 USL
LSL

T
6
112
Process Capability
CPk= 1.5
T
4.5
USL
LSL
LCL UCL
+3
- 3
1.5

6
113
Process does not understand the design specification
If you do not want to attack any of the
above, forget ‘Quality Improvement’
Process Capability
1 For a given design specification select the process such that the
‘Process width’ is less than the ‘Design width’(For a 6 process
the process width should be half of Design width)
2 If the ‘process width’ is more than the ‘Design width’, improve
the process to reduce the ‘process width’, if the technology
permits.
3 If the technology does not permit(poor technology) go for a
better technology.
4 If both 2 & 3 are not possible, have a relook at the design to
increase the ‘Design width’, if competition permits.
6
114
Why Assess Measurement System
Measure
Process/Product
Knowledge Understanding
Analysis Improvement
You don’t know what you can’t measure !!
6
115
Is It Valid?
Am I measuring what
my customer thinks is
critical to quality (CTQ)?
Is It Reliable?
Is the measurement system
accurate, stable, repeatable,
reproducible and linear?
Do Specifications Exist?
What is the order of magnitude of my
measurement? What are the spec
limits for the measurement?
?
?
?
YES or NO?
What must I do to show this?
YES or NO?
What must I do to show this?
YES or NO?
Are the specs valid?
What must I do to show
this?
Three Questions To Ask About
Measurements &
Measurement Systems:
Answering These Questions Is The Heart of MEASURE Phase
Measurement System Analysis
6
116
Gage Repeatability is the variation in measurements
obtained when one operator uses the same gage or
measurement process for measuring the identical
characteristics of the same parts or items.
Possible Causes of Poor
Repeatability:
Equipment:
• Gage instrument needs
maintenance.
• The gage needs to be
more rigid.
• The clamping of part
needs improvement.
People:
• Environmental conditions
(lighting, noise)
• Physical conditions
(eyesight)
Repeatability
Gage repeatability
6
117
Gage Reproducibility is the variation in the
average of measurements made by different
operators using the same gage or measurement
process when measuring identical characteristics of
the same parts or items.
Possible causes of
poor reproducibility
 Measurement
procedure is not clear
 Operator is not
properly trained in
using and reading
gage
 For transactional
applications -
operational definitions
not established
Reproducibility
Mean of
the measurements
of Operator B
Mean of
the measurements
of Operator A
Gage Reproducibility
6
118
Gage Accuracy also referred as Bias is the
difference between the observed average of
measurements and the true average. Establishing
the true average is best determined by measuring
with the most accurate measuring equipment or
system available.-Viz Master gauge or Master
Measuring Equipment
How to Calculate (Continuous Data):
• Obtain true average of Sample Parts
• Compare with the operators observed values from the R & R study.
• To convert accuracy to a % of tolerance, multiply difference of true average vs. observed by 100 and
divide by tolerance.
How to Calculate (Discrete Data):
• Have an item with a known defect count inspected several times. Compare the observed defect count
with the actual defect count.
Possible causes of poor accuracy :
• Gage not properly calibrated
• Improper use of gage by operator
• Unclear procedures
• Human limitations
True
Average
Accuracy
Observed
Average
Gage Accuracy
6
119
Gage Stability
Time 2
Time 1
Gage Stability refers to the difference in the
average of at least two sets of measurements
obtained with the same gage or measurement
system on the same parts or items taken at
different times.
How to Calculate (Continuous Data):
• For gages normally used for relatively long periods of time without calibration
• Conduct a second Gage R&R study just prior to the time recalibration is due.
• Gage stability is the difference between the grand averages of the
measurements from the two studies.
How to Calculate (Discrete Data):
• Conduct a second Repeatability, Reproducibility and or Accuracy study. Look
for trends over time.
Possible causes of poor stability:
Continuous data
• Gage not being calibrated as frequently as needed
• If air gage, may need filter or regulator.
• If electronic gage, may need warm-up and stabilization.
Discrete data
• New operators
• Software changes
Gage Stability
6
120
Gage Linearity is the difference in the accuracy
values through the expected operating range
Within the same instrument.
True
Average
Observed
Average
(Low End)
Accuracy
(Low End)
Observed
Average
(High End)
Accuracy
(High End)
True
Average
LINEARITY = | Accuracy (low) - Accuracy ( high) |
How to Calculate:
• Conduct Accuracy study through the expected operating range. At least two
studies should be done, one at each end of range.
• Subtract the smaller accuracy value from the larger to determine linearity.
Possible causes of poor linearity:
• Gage not calibrated properly at both lower and upper end of operating range.
• Error in the minimum or maximum master.
• Worn gage.
• Internal gage characteristics.
Gage Linearity
6
121
REPEATABILITY:
GAGE:
The instrument used for
making measurements that
we want to validate
The % Variation Due to the Measurement Method
GAGE R & R
Does the same operator
get the same results
when measuring the
same part several times?
REPRODUCIBILITY: Do
different operators get
the same results when
measuring the same
part several times?
Why Gage R&R Study-Contd..
6
122
Measurement Systems Analysis
Match the definitions to the terms:
Repeatability
Reproducibility
Accuracy
Stability
Linearity
A. Difference in the average of two or more sets of
data taken obtained with the same gage on the
same parts taken at different times.
B. Variation in average measurements made by
different operators.
C. Difference in accuracy values through
expected operating range.
D. Variation in measurements obtained by one
operator measuring the same part with the same
gage.
E. Difference between the observed average and
the true average.
Review Questions
6
Gage R&R
Gage
Repeatability and Reproducibility
333
6
Upper Spec Limit
Lower Spec Limit
SCRAP !
SCRAP !
OBSERVED
PRODUCT / PROCESS QUALITY
LET’S MEASURE THIS PART
SEVERAL TIMES...
MEASUREMENT
QUALITY
LET’S PRODUCE AND MEASURE !
BEQI /M29EMD
Feb2002
334
6
Upper Spec Limit
Lower Spec Limit
OBSERVED
PRODUCT / PROCESS QUALITY
MEASUREMENT
QUALITY
Gage Error
Let’s reduce GAGE R&R variation !
BEQI /M29EMD
Feb2002
335
6
Upper Spec Limit
Lower Spec Limit
OBSERVED
PRODUCT / PROCESS QUALITY
MEASUREMENT
QUALITY
Gage Error
BEQI /M29EMD
Feb2002
336
6
Upper Spec Limit
Lower Spec Limit
OBSERVED
PRODUCT / PROCESS QUALITY
MEASUREMENT
QUALITY
Gage Error
LOOK!
Observed Capability has improved !
BEQI /M29EMD
Feb2002
337
6
OBSERVED
PRODUCT / PROCESS QUALITY
Upper Spec Limit
Lower Spec Limit
MEASUREMENT
QUALITY
Gage Error
Lower the Gage Error, the better the Capability!
This means... • fewer good parts rejected
• fewer bad parts accepted
BEQI /M29EMD
Feb2002
338
6
2 WAYS TO LOOK AT
MEASUREMENT SYSTEM
VARIATION
1- Gage R&R as % Contribution
2- Gage R&R as % Tolerance
BEQI /M29EMD
Feb2002
339
6
1- Gage R&R
as
% Contribution
BEQI /M29EMD
Feb2002
340
6
m
p
Upper Spec Limit
Lower Spec Limit
OBSERVED
PRODUCT / PROCESS QUALITY
MEASUREMENT
QUALITY
GAGE R&R % Contribution =
Gage Variance
Observed Process Variance
=
(m)²
(p)²
BEQI /M29EMD
Feb2002
341
6
Rules of Thumb
R&R % Contribution < 2%
GOOD !
2 % < R&R % Contribution < 7.7 %
RISK EVALUATION NEEDED !
R&R % Contribution > 7.7 %
INACCEPTABLE !
GAGE R&R % CONTRIBUTION
BEQI /M29EMD
Feb2002
342
6
2- Gage R&R
as % Tolerance
BEQI /M29EMD
Feb2002
343
6
Gage Error
m
Tolerance
GAGE R&R % Tolerance
Upper Spec Limit
Lower Spec Limit
OBSERVED
PRODUCT / PROCESS QUALITY
MEASUREMENT
QUALITY
=
Gage Error
Tolerance
=
5.15 m
USL - LSL
BEQI /M29EMD
Feb2002
344
6
Rules of Thumb
R&R % Tolerance < 8%
GOOD !
8 % < R&R % Tolerance < 30%
RISK EVALUATION NEEDED !
R&R % Tolerance > 30%
INACCEPTABLE !
GAGE R&R % TOLERANCE
BEQI /M29EMD
Feb2002
345
6
WARNING !
BE PRECISE ON THE GAGE % YOU ANNOUNCE!
GAGE R&R % Contribution?
or
GAGE R&R % Tolerance?
Calibration Sticker does not imply a good Gage R&R!
An accurate Gage does not guaranty a good Gage R&R!
BEQI /M29EMD
Feb2002
346
6
137
Tying it all Together
% Contribution %Tol Typical Conclusions:
Green Green
Repeatability and Reproducability are
acceptably small portions of the total
observed variation…Proceed.
Green
Yellow or
Red
Implement any quick or obvious
measurement system improvements
but project work can begin with the
current system. As the actual
process variation improves, the
measurement system may need
improvement.
Yellow
Green,
Yellow, or
Red
The variation in the measurement
system is not as small as we would
like to see. The system is marginal
and will need to be evaluated for
improvements before proceeding.
Red
Green,
Yellow, or
Red
The variation in the measurement
system is not an acceptably small
portion of the total observed variation.
Need to fix measurement before
proceeding with process
improvement.
Here is a guide to interpret your Gage
RR results:
6
138
• Can’t apply techniques of continuous gages.
• The gage can be assessed based on known samples.
• Attribute gage studies are performed by 2-3 people
measuring 20 or more parts.
• If all operators and known samples agree - receive a
“Percent Effectiveness Score “ of 100%.
• Use results to validate measurement or to prioritize
improvement needs.
How Do You Test the Reliability of Measurement
Systems That Give only Pass/Fail Data?
Gage R&R for Discrete Data.
6
139
Methodology
Analyze
6
140
Analyze
1.Graphical Analysis
2.Statistical Analysis
6
141
Multi Vari Analysis
(Graphical)
Example : Electroplating of several parts in a bath.
1 2 3 4 5 6 7 8 9 10
0
10
20
30
40
50
Location
Within a Part - Positional
Over Time
1 2 3 4 5 6 7 8 9
0
10
20
30
40
50
60
70
80
90
Part
Part
Between Parts
1 2 3
0
10
20
30
40
50
60
70
80
90
Time
Between
Time
Within
359
6
142
Multi Vari – An Example
Variability of Epi-Thickness
on the wafers
81%
8%
11% Position of wafer
Within wafer
Run to Run
T
R
C
L
B
Upper
Middle
Lower
360
6
143
Statistical analysis
Why ?
How ?
Difference
Difference
• To find out the vital few factors
• By finding out the “Statistically
Significant difference” each factor
(x) makes on the response (y)
361
6
144
Y = f (X1……XN)
Independent
variables
Dependant
variable
20% + 80% = 100%
Trivial Many
Variables
Vital Few
Variables
362
6
145
Statistical Analysis Tools
 t-test
 f- test
 ANOVA
 Chi-square ( 2) Test
 Correlation & Regression
 Multi vari Analysis
363
6
146
Methodology
Improve
6
147
• How to Reduce extraneous variation in a
process or product
by
i) PM Process Mapping
ii) CE Cause & effect diagram
iii) CNX Process input / output diagram
iv) SOP Standard Operating Procedure
Without Statistics!!
Improve without Statistics
PM / CE / CNX / SOP
6
148
Process Mapping
Why?
• Graphically outlines the sequence of a process
• Provides visual foundation for current situation and
analysis
• Aids in identifying bottlenecks, redundancies and
waste
• Look for non value adding steps
• Look for variables
6
149
How Process Mapping is to be done?
1. The start of the process and the end of the process should be well
defined.
2. Observe closely the process steps and flow.
3. Involve all the personnel who are carrying out the process and take their
opinion.
4. The process map should be as elaborate as possible; even minor
details should not be missed.
5. The correct sequence of the process should be indicated in the
process flow.
How Process Mapping can help in cycle time reduction project ! !
1. Look for time consuming, redundant, unnecessary movement & storage
delays.
2. Look for major choke points that create significant delays.
3. Check for all rework steps which may be present.
4. Look for in-efficient layouts, sequences or flows.
5. Look for redundant material handling/packaging/unpacking.
Analyze with following questions
1. What is the real purpose or function of the step
2. Does the step add value to the output?
3. Can the process step be eliminated, minimized or combined with another
value adding step
• Validate the process map by physically walking through the process.
Process Mapping contd...
6
150
START
PRINTING OF PRs
AFTER MRP
RECEIPT OF PRs
IN PURCHASE
DISTRIBUTION TO
PURCHASE CELLS
SORTING PRs
REPEAT/VENDOR
REQUEST FOR
DRAWINGS
PREPARATION OF
REPEAT ORDERS
PREPARATION
OF ENQUIRIES
POSTING /
FAXING OF
ENQUIRIES
RECEIPT OF
QUOTE
PREPN. OF COMP
STATEMENT
CLARIFICATION
FROM VENDOR /
INDENTOR.
NEGOTIATION
PRICE REVIEW
PREPARATION OF
PURCHASE
PROPOSALS
COMPILATION
OF DATA FOR
NYRO.
NYRO / SGRO.
FAXING DATA TO
NYRO.
CONFIRMATION
OF AVAIL
FROM NYRO.
VETTING OF
PRs BY FINANCE
REQUEST D&E
FOR END USE.
MAILING OF PR
EU s TO NYRO.
PRINTING OF
EU s.
RECEIPT OF
PO FROM
NYRO.
ENTER PO
DATA IN EDP
DATABASE.
POST FACTO
APPROVAL
FAX ADVANCE
COPY TO
VENDOR.
END
POST ORDER BY
REGISTERED
MAIL.
SEND TO
CENTRAL
REGISTRY.
PREPARE FOR
POSTING.
STAMP PO NO.
AND DATE.
ALLOTMENT OF
PO NO. & ENTRY
IN REGISTER.
SIGNATURE
ON PO BY
MGR / DGM.
VETTING BY
FINANCE
APPROVAL OF
PROPOSAL AS
PER DELEGATION
OF POWER.
ORDERING BY
NYRO.
NON NYRO
Process Mapping contd...
A typical Process Map
PR to PO
6
151
Use of Process Mapping for Defect Identification
Customer calls
Confirm
payment
type
Take order
details
Problem
areas
Deliver & collect
money, if non-credit
card customer
Call
Answered
?
Credit
card?
Despatch
order
Take card details
Enter customer
details in the
system
No
No
Yes
Yes
Process Mapping contd...
6
152
Versions of a Process
What you
think it is ….
What it really
is….
What it should
be….
What it could
be….
Process Mapping contd...
6
153
Cause & Effect Diagram
Output Y
C
N
X
Every variable on the diagram should be labeled as either:
C = Constant
N = Noise
X = Controlled variable or factor
Sources of Variation
C C
N X
X
People Material Machine
Environment
Method
Measurement
N X
N
X
X
C N
N
C
C
Make cause & effect diagram as detailed as possible
6
154
Process
Response/
Outputs
Constants or Fixed Variables
C
Y
X
Experimental
Factors
N
Noise Variables
Please Note:
C – These are non- experimental factors influencing the response Y which
should be addressed & made as a constant and controlled.
N - These are non- experimental factors influencing the response Y very
marginally and which are costly / difficult to address.
X – These are factors greatly influencing the response Y which have to be
experimented upon to optimize the response.
CNX Diagram
6
155
S O P
Standard Operating Procedures are rules that we define to
ensure that we have consistent processes in everything we do.
• Based on good judgment
• Common sense
• Engineering/Process knowledge
Incorporate in documents like process sheets, instruction sheets,
work instructions of ISO 9000, process standards etc.
Make sure we have defined processes and that the rules are being
obeyed by all, by
constant monitoring.
6
156
Class Exercise #1
(Exercise from Air Academy)
“Wiffle”
“Solid”
1) Ensure Pin Position and Stop Angle Pin are at
Position #3. Cup is at Position #1. Hook is at
Position #4.
2) Use 1 Rubber Band. Pull the Arm to 177o and
Shoot the Rubber Ball.
3) Disconnect Rubber Band between shots.
4) Have someone measure the Distance in cms.
5) Each person takes 5 shots in less than 60 sec.
6) Record Distances in spaces on
next page.
Form Your Team. Each Participant Must Shoot
Catapult 5 Times Using the Following SOP:
4
Pin Position
1 Rubber Band
Stop Angle
1
3
3
Start Angle 177
1
6
157
Catapult Data
Before PM / CE / CNX / SOP
Longest _______
Shortest _______
_________________
Range _______
Record
What Have We Learnt?
Enter values in Minitab- Calculate mean, range and standard deviation
Save data in ‘A’ drive
Target : 500 cms. Tolerance +/- 25 cms. Calculate Z
Class Exercise # 1 Contd….
Shot #1 _____ _____ _____ _____ _____ _____
Shot #2 _____ _____ _____ _____ _____ _____
Shot #3 _____ _____ _____ _____ _____ _____
Shot #4 _____ _____ _____ _____ _____ _____
Shot #5 _____ _____ _____ _____ _____ _____
Distance cms
Participant 1 2 3 4 5 6
Enter the data in file S1  Catapult-Project.mtw and save
6
158
Cause & Effect Diagram for Catapult
With Control - Noise - Experimental Parameters
Machine
Men
Environment
Measurement
Methods
Material
Distance
Travelled
by
the
ball
Class Exercise # 1contd...
6
159
Optimum Launch
Distance
in cms
Noise Variables (N)
Constants( C )
Experimental
Factors (X)
Stability
Clamp
Skill
of
operator
Wind
Friction
of
arm
Elasticity
Start Angle
Stop Angle
Pin pos
Type
of
tape
Band
length
Ball
weight
&
type
Hook pos
Cup pos
Temperature
Eyesight
of
inspector
Angle
marking
Class Exercise # 1 contd...
CNX Diagram for the Catapult
Response (Y)
6
160
a) Clamp the catapult to the table
b) Have a suitable block to fix the
start angle at 177 degree.
c) Use a better measurement system
(by having a carbon impression
at the first pitch of the ball).
Class Exercise # 2
Repeat the exercise (same team, same
catapult) by following the Improved SOP
6
161
Clamp catapult
to table
with clamps
Set pin pos at 3
Set stop
Angle pin at 3
Set cup pos at 1
Set Hook
pos at 4
Install one
rubber band
Have a block
to set arm
at 177 degree
Pull back the
arm to
177 degree
Place the
rubber ball
in the cup
Spread carbon
sheet in front of
catapult
Record the
distance in
record sheet
Measure the
distance
traveled by ball
Release the
arm
Repeat the
process
Class Exercise # 2 contd...
Process Mapping
Disconnect
and reconnect
the rubber band
6
162
4
Pin Position
1 Rubber Band
Stop Angle
1
3
3
Start Angle 177
1
Class Exercise # II Contd...
6
163
Catapult Data
After PM / CE / CNX / SOP
Record
Distance cms
Enter in Minitab calculate mean,range & standard deviation.
Save data in ‘A’ drive
What Have We Learnt?
Has the range reduced?
Has the std dev reduced?
Have the same specification as in exercise 1 &
calculate sigma value
Compare sigma values and draw the inference
Class Exercise # II Contd...
Shot #1 _____ _____ _____ _____ _____ _____
Shot #2 _____ _____ _____ _____ _____ _____
Shot #3 _____ _____ _____ _____ _____ _____
Shot #4 _____ _____ _____ _____ _____ _____
Shot #5 _____ _____ _____ _____ _____ _____
Participant 1 2 3 4 5 6
6
164
Catapult project progress
Y = Traverse distance of ball 500+/-25 cms.
Metric Current SOP Improved
SOP
Target
Mean
Std. Dev.
Cp
Cpk
Z
Histogram
Run chart
500
4.16
2
6
6
165
Lessons Learnt
Initial Condition
Response
USL
USL
Apply PM/CE/CNX/SOP
2nd Condition
Response
Now: Identify X’s that will
Reduce the variance, Shift the mean
What is this telling us?
6
166
What about the “Xs”??
1. In both the previous exercises the experimental factors
were fixed as follows:
a. Start angle: 177°
b. Stop angle pos.: 3
c. Pin pos.: 3
d. Hook pos.: 4
e. Cup pos.: 1
2. By experimenting on the above five factors we can fix
them to get the exact distance required within the
practical range.
3. This method of experimenting will be dealt with later in
the ‘Improve’ using Statistics (DoE) stage.
6
167
DoE
an
Overview
6
168
Quality Improvement
is
Problem with Spread
Accurate but not Precise
Desired
Current
Situation
LSL USL
T
Shrinking the variation
&
Problem with Centering
Shifting the mean to the target
d
Precise but not Accurate
USL
LSL
Current
Situation
Desired
6
169
How Do You
Get a
“Knowledge
Equation”?
What Kind of KNOWLEDGE We
Need To Improve Our
Products & Processes?
Product
or
Process
X2
Y
X3
X4
X1
Y =
Cos(X1/X2)
Ln(X3)
- X4 + 200
What Can You Do With This Equation?
• Set X’s To Assure That Your Y Hits Target
• Trade-Off X’s To Reduce Cost But Still Hit Target Y
• Find Tolerances on X’s To Keep Y in Spec
CTQ - Important to
Your Customer
6
170
Y
X
Y
X
One Factor Experiment
6
171
One-Factor-At-a-Time Experimentation (OFAT)
Problem: Improve Gas Mileage
Approach: Evaluate Gas Type and Timing Setting
(2) G1 Gives Best MPG. Hold Gas Type G1 Constant
(3) Evaluate MPG for Timing T1 and T2. Determine Best Value
One-At-A-Time:
(1) Set Timing at T1, Evaluate MPG of Gas Types G1 and G2
Gas Type Timing Set-Up MPG
T1
T1
G1
G2
30
20
Gas Type Timing Set-Up MPG
T2
T1
G1
G1
30
25
Best Value
G1, T1
6
172
Design of Experiment
Problem: Improve Gas Mileage
Approach: Evaluate Gas Type and Timing Setting
Gas Type & Timing
Interact: Effect of
One Factor Depends
on the Level of The
Other Factor
Design of Experiment:
Gas Type Timing Set-Up MPG
T2
T1
G1
G1
30
25
G2
G2
T1
T2
20
45
Full
Factorial
Design:
50
40
30
20
T1 T2
G2
G1
MPG
Missed This Point In
One-At-A-Time
6
173
Strategy of Experimentation
Plan:
What is the Objective of your Experiment?
– What questions do you expect to be answered when the experimentation & analysis is complete?
What is Your Primary Response?
– What Y & other Y’s should be measured?
What are the Potential X’s?
– What is the current capability of each Y after the implementing SOPs on Control and Noise type X’s?
What Experimental Plans best fit your problem?
Plan the Roles & Responsibilities of all involved in the DoE.
Do:
Assess the ease of implementing the Experimental Plan.
Implement the Experimental Plan.
Observe each experimental run.
– Take notes & collect data on background Noise variables in addition to collecting data on your Y’s.
Study:
Analyze your data graphically and statistically.
– Prepare graphs that illustrate the affect of each X on each Y.
– Compare the outcome of the analysis to your initial theories. What did you learn?
Confirm what you think you have learned.
Act:
Act on the results.
– Have you made sufficient improvement given the new knowledge?
– Plan additional experimentation to further your knowledge if necessary.
– Communicate your findings and implement new SOPs where appropriate.
6
174
Methodology
Control
6
175
Statistical Process Control:
Understanding Variation
• Developed by Walter A. Shewhart in 1924.
• Patterns data for statistical test, leading to product /
process behavior information.
• Facilitates underlying “Cause System” understanding.
• A graphical representation of product and/or process
performance.
• Assignable (special) cause detection -- central tendency
and/or variability impactors.
• Serves as a probability based decision making tool.
• Performance change detector.
• Assess ability to predict performance based on sample
data.
• Points out actionable areas, with known degrees of risk
and confidence.
6
176
Statistical Process Control
Control charts: CC
A plot of values of ‘location’ and ‘dispersion’ over time, used to identify
assignable variations .
Control limits:CL
UCL and LCL on control charts within which all observations must fall for the
process to be in control.
 They refer to process control ,stability.
 Inherent to the process.
Specification limits: SL
The designed USL and LSL for a process variable or a product.
They refer to product acceptability.
When the process has been brought in-control (using SPC ),
quality can be further improved by redesigning the process to
reduce the chance causes . This is called ‘Breakthrough ’.
6
177
Common cause (Chance cause)
 variability within control limits
 contributed by many factors
 is of low magnitude
 stable over time
 part of the manufacturing system
2. Special cause (Assignable cause)
 variability exceeds control limits
 contributed by few factors
 higher than acceptable levels
 unusual occurrence
 external to the system
Using SPC we eliminate the Assignable causes of variation.
Types of variability in a process.
6
178
1. Common cause (Chance cause)
 affects all individual values
 always present
 random variation inherent in the process
 not easily correctable
 distribution of data is normal
2. Special cause (Assignable cause)
 source of variation intermittent
 assignable to some event
 non-random pattern
 easily correctable
 distribution –skewed, shifted or abnormal
Types of variability in a process…contd.
6
179
Two Types of Causes
Common Causes:
Those causes that are inherently part of the
process (or system) hour after hour, day after day,
and affect everyone working in the process. The
variation that results from the consistent operation
of the process as designed.
Assignable Causes:
Those causes that are not part of the process (or
system) all of the time, or do not affect everyone,
but arise because of specific circumstances. The
variation that is the result of special causes in the
operation of the process.
6
180
Specification
Target
Target
Performance
Time
Understanding Variation
….We’ll
use these
simple
graphs
to illustrate
some
important
points.
6
181
On Target, Minimum Variation
Average =
Target
Performance
Time
6
182
Average =
Target
Performance
Time
Average Is On Target…
…But
There Is
More
Variation
6
183
Target
Performance
Time
Average
Off Target, Minimum Variation
6
184
?
Average =
Target
Performance
Time
The area of usual
process performance
Indication of a special occurrence
Assignable Cause
6
185
Run Chart Example
Dr. Thomas W. Nolan came one day to talk to Dr. W. Edwards Deming
and brought with him a chart that his son Patrick, age 11, had made for a
school project. The chart showed, day by day, the time of arrival of his
school bus. He recognized assignable causes of variation. Patrick had
kept a record of the daily time of arrival of the school bus that came to
carry him off to school, and had plotted the points in time order. He
identified assignable causes of delay on two days. Think what a good
start in life Patrick had, understanding common and assignable causes of
variation - at age 11!. He had recognized, without calculation, assignable
causes of delay on two days, and had shown his explanation of these
delays.
Time of arrival of
school bus, by
Patrick Nolan, 11
0800
0830
0900
Date
(Source: Dr. W. Edwards Deming’s 4 day management seminar)
6
186
Four Possible States For A Process
1) Stable & Capable
Average
Lower Spec.
Upper Spec.
6
187
2) Stable and Incapable
Average
Upper Spec.
Lower Spec.
6
188
3) Unstable & Currently Capable
Average
Upper Spec.
Lower Spec.
6
189
4) Unstable and Incapable
Upper Spec.
Lower Spec.
Average
6
190
Summary
Four Process Conditions
IS THE PROCESS IN STATISTICAL CONTROL?
IS
THE
PROCESS
CAPABLE
OF
MEETING
REQUIREMENTS?
YES NO
IDEAL STATE
ACTION:
ATTAIN CONTROL
ACTION:
IMPROVE
CAPABILITY
(TECHNOLOGY)
ACTION:
1st: CONTROL
THEN: CAPABILITY
6
191
The Control Chart
The control chart is a process analysis technique. It
is used to quickly detect the occurrence of
assignable causes or process shifts so that a
successful investigation of the process and proper
corrective action may be taken. This proactive
approach assists in keeping the process on target
with minimum variation.
Average
Performance
Time
UCL
LCL
Control Charts show the distribution over time
6
192
Components of the Control Chart
Center Line
UCL
LCL
SAMPLE NUMBER
Region of Nonrandom Variation
Region of Nonrandom Variation
Inherent Process + Measurement Variation Spread
+3
-3
X-Bar
To interpret these charts, we use the decision limits and look for
identifiable patterns that indicate non-randomness in process behavior.
UCL and LCL on control charts within which all
observation must fall for the process to be in control
•They refer to process control, stability
•Inherent to the process
6
193
Establishing control charts
 Collection of data on the selected characteristic
 Select the appropriate chart
 Choose the center line. ( targeted value or calculated from
data)
 Calculation of +/- 3S trial control limits
 Plot the data points
 Find assignable cause points ( points beyond the control limits)
 Remove them and recalculate the control limits
 ( now only chance causes are present )
 Monitor process by plotting the instant values with the newly fixed
control limits
 If rule violation occurs, find the cause and take corrective action
 Improve process and review control limits
6
194
Individual X and
Moving Range
Average & Range
(X-bar and R)
Average & StandardDeviation
(X-bar and S)
Exponentially
Weighted
Moving Average
Type of Chart When do you need it?
• When production is low volume or cycle time
to build product is long, or shift and drift
are a problem
• When production is high volume, sample size
is 3-8; allows process mean and variability
to be viewed and assessed together
• When production is high volume, sample size
is >8; allows process mean and variability
to be viewed and assessed together
• When a small shift needs to be detected, or
process is continuous and mixing can occur
(e.g., in a chemical plant, wire drawing )
What Kind of Control Chart You Need?
Variable Data?
• When you want to know the fraction of defective
units; sample size is variable and usually > 50
• When you want to know the number of defective
units; sample size is constant and usually > 50
• When you want to know the number of defects;
sample size is constant
• When you want to know the number of
defects per unit; sample size is variable
p
np
c
u
Attribute Data?
6
195
Reading control charts
For an unstable process – take action ?
Identify rule violations indicating assignable causes of variation.
Identify the assignable cause(s) as to source(s).
Remove and prevent recurrence.
Adjust/correct for assignable cause if it can not be removed and prevented.
Action to adjust centering
Action to reduce variability
For a stable process - take no action.
 Most of the plotted points occur near the center line
 A few of the points occur near the control limits
 Only an occasional rare point occurs beyond the control limits
( 2.7 out of 1000)
 The plotted points occur in a random manner with no clustering,
trending, or other departure from a random distribution.
 Histogram is normal
6
196
Rule 1 A single data point above U.C.L. or below L.C.L.
U.C.L.
+2 S random
+1 S random
Target
-1 S random
-2 S random
L.C.L.
Rule 2 Nine consecutive data points all above the target, or nine consecutive
data points all below the target.
U.C.L.
+2 S random
+1 S random
Target
-1 S random
-2 S random
L.C.L.
Rules to test presence of Assignable causes
6
197
Rule 3 Six consecutive data points declining in value, or six consecutive
data points increasing in value.
U.C.L.
+2 S random
+1 S random
Target
-1 S random
-2 S random
L.C.L.
Rule 4 Intentionally not shown since it is not as critical.
(14 points in a row, alternating up and down)
Rule 5 Three consecutive data points of which two are at least +/- 2 S random
from the target.
U.C.L.
+2 S random
+1 S random
Target
-1 S random
-2 S random
L.C.L.
6
198
Rule 5 Five consecutive data points of which four are at least +/- 1 Srandom
from the target.
U.C.L.
+2 Srandom
+1 Srandom
Target
-1 Srandom
-2 Srandom
L.C.L.
Rules to test presence of Assignable causes
UCL = Upper Control Limit
LCL = Lower Control limit
Target = Target or Process Average
S random = Estimated Standard deviation, often from subgroup ranges,
representing inherent common cause variation.
6
199
Tables
Table of Constants for X –R, and I-MR charts for sub-
group size = n
2 1.880 1.128 0 3.267
A2 d2 D3 D4
3 1.023 1.693 0 2.574
4 0.729 2.059 0 2.282
5 0.577 2.326 0 2.114
6 0.483 2.534 0 2.004
7 0.419 2.704 0.076 1.924
8 0.373 2.847 0.136 1.864
9 0.337 2.970 0.184 1.816
10 0.308 3.078 0.223 1.777
E2
2.660
1.772
1.457
1.290
1.184
1.109
1.054
1.010
0.975
n
6
200
u
u - 3 u/n
u + 3 u/n
u
c
c - 3 c
c + 3 c
c
p
p - 3 p(1-p )/n
p + 3 p (1-p)/n
p
np
np - 3 np (1 – p )
np + 3 np (1 - p)
np
x
x - 2.66R
x + 2.66R
I or x
R
D3R
D4R
R or MR
x
x – A2R
x + A2R
x
CL
LCL
UCL
Chart
 





List of formulae for Control Lines
x = Average of subgroup averages (x)
total no. rejected total no. of defects total no. of defects
total no. inspected total no. of subgroups total units inspected
UCL – Upper Control Limit
LCL – Lower Control Limit
CL – Centre Line
p = c = u =

6
201
Make the items in a
subgroup as alike as
possible
Make the subgroups
themselves as
different as possible
The observations in a logical subgroup is free from assignable causes of variation.
In production process the rational subgroup could be ‘manufactured at the same
time’.
In service area the rational subgroup could be ‘similar transactions by the same
person’.
Need to apply Rationale and Knowledge
of the process to determine Subgroup size
Production Sequence
••••••••••••••••••••••••••••••••••••••
Sampling Windows n= 5
Production
Unit
Rational Subgroups
Instant time method (widely used)
gives minimum variation within sub groups
maximum variation between sub groups
6
202
Attribute Charts
• You want to track the causal variable on
a transactional project where data is
discrete.
Common Choices:
• p chart - To track fraction defective
• u chart - To track defects per unit
• Or skip this and go back to the
Individuals and Moving Range control
charts for variable data
6
203
A claim of process
capability without
process consistency
(no assignable causes)
illustrated on a
control chart
is not a
valid claim!
6
204
Some Misconceptions About
Control Charts
Although control charts have been used
for many years in a variety of situations,
there are a number of misconceptions
concerning their uses. The following
misconceptions are summarized from a
presentation by Michael Flynn (Flynn,
1983):
6
205
1. Control charts are tools for production workers
to tell them when to adjust their processes.
Control charts are tools for understanding variation. An
operator reacting to an out of control situation is one of
many possible uses of control charts, but certainly not
the most important in many organizations.
( Associates in Process Improvement)
6
206
2. Control charts are only for the production or
manufacturing operations.
Control charts should be used to understand variation in
all of the important processes in an organization. These
include employee relations, safety, accounting, planning,
lead times, maintenance, engineering, research, customer
service, quality control, finance, complaints,
environmental monitoring data.
( Modified from Associates in Process Improvement)
6
207
3. Control limits are boundaries beyond which
we do not want to go.
Control limits have nothing to do with what we want.
The limits just define the boundaries for common
causes of variation. Often we want a process to go out
of control, if for example, the change results in higher
yield or fewer errors on purchase orders.
( Associates in Process Improvement)
6
208
4. Control limits are boundaries within which
the process can vary by chance.
A better statement would be that control limits are
boundaries on the process within which sample results
can vary due to common causes when the process does
not change at all.
( Associates in Process Improvement)
6
209
5. The process can go back and forth -- in
control, out of control, and then back in
control
The calculated statistic for different subgroups will vary.
If a special cause results in a shift in the process, the
subgroup points will still vary, but now some may be
within the control limits and some outside.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 1516 17 18 19 20
( Associates in Process Improvement)
S
t
a
t
i
s
t
i
c
6
210
6. Control charts can only be used to track
processes over time.
Defining subgroups by time periods is the most
common way to develop control charts. But there are
many other possibilities such as by clerk, customer,
city, supplier, spindle, filler head, material lot,
instrument number, and so on. Control charts are
appropriate for all of these groupings of data.
( Modified from Associates in Process Improvement)
6
211
The control limits are calculated using the same
method every time “Tight” control limits indicate that
the common cause variation in the process is relatively
small.
( Associates in Process Improvement)
7. It is harder to hold narrow control limits
than wide ones.
6
212
8. Two-sigma control limits will result in
“tighter ” control than the traditional
three-sigma limits.
Using limits other than Shewhart’s three-sigma control
limits will likely result in higher costs due to over and
under reaction to assignable causes. For a stable
process, reacting to all points outside of a two-sigma
limit will result in increased variation in the
output of the process.
( Associates in Process Improvement)
6
213
9. Assignable causes are always an indication
of a problem or of poorer quality, and
Shewart called special causes assignable causes, i.e.
the variation could be “assigned ” to a particular
cause. Variation in the right direction can certainly be
good.
( Associates in Process Improvement)
10. It is not necessary to investigate assignable
causes that result in better quality
6
214
Poka-yoke = Mistake Proofing
From the Japanese words:
yokeru - To Avoid
poka - Inadvertent Errors /
Unintentional Error
 The idea is to respect the intelligence of workers by taking
over repetitive tasks that depend on vigilance or memory.
 Promotes Creativity and Value-Added Activities
 Reduces Ego problem among employees.
This is based on the principle that the defects can be
prevented by controlling the performance of a process
so that it cannot produce the defects even when a
mistake is committed by humans.
6
215
Common Types of Human Error
1) Forgetfulness (Not Concentrating)
2) Errors Due to Misunderstanding (Jump to Conclusions)
3) Errors in Identification (View Incorrectly...Too Far Away)
4) Errors Made by Untrained Workers
5) Willful Errors (Ignore Rules)
6) Inadvertent Errors (Distraction, Fatigue)
7) Errors Due to Slowness (Delay in Judgment)
8) Errors Due to Lack of Standards (Written & Visual)
9) Surprise Errors (Machine Not Capable, Malfunctions)
10) Intentional Errors (Sabotage - Least Common)
6
216
Red Flag
A condition in the process
which commonly provokes
errors
Red
Flag
Error Defect
6
217
1) Frequent Adjustments
2) Constant Equipment Changes
3) Dimensionality / Specification /
Critical Condition
4) Many Parts / Mixed Parts
5) Multiple Steps
6) Lack of, OR Ineffective,
Standards
Some Red Flag Conditions Are:
7) Rapid Repetition
8) Volume
a. Sudden change
b. Quantity push vs Quality
9) Environmental Conditions:
a. Material / Process Handling
b. Housekeeping
c. Foreign Matter
d. Poor Lighting
e. Other
10) Other?
6
218
Common Mistake Proofing Devices
Using wisdom and ingenuity to create devices that allow
you to do your job 100% defect free 100% of the time.
HOME
• Automated Shut-Offs on
Electric Coffee Pots
• Ground Fault Circuit Breakers
for Bathroom or Outside
Electric Circuits
• Child-Proof Caps on
Medications
• Butane Lighters with Safety
Buttons
Retail
• Tamper-Proof Packaging
• Bar Coding at Checkout
OFFICE
• Spell Check in Word
Processing Software
• Questioning “Do you want to
delete?” After Depressing
the “Delete” Button on Your
Computer
FACTORY
• Dual Palm Buttons and Other
Guards on Machinery
• Bar Coding
6
219
Common Mistake Proofing Devices…Contd.
Must be taller than the line to go on ride
Amusement park rides
use this technique to
avoid mistake of having
young children stand in
line for rides they are not
able to enjoy
When converting from leaded to
unleaded gas, the tank openings
and nozzles were smaller to
eliminate potential of putting
wrong gas in tank
6
220
Common Mistake Proofing Devices…Contd.
On Airplane Lavatory Doors, in
order to turn on the lights, the
passenger must lock the door
which automatically activates
the OCCUPIED sign.
LOCK
OCCUPIED
6
221
Rely on getting
information to where
it can be used to
prevent a mistake.
Common Mistake Proofing Devices…Contd.
6
222
Common Mistake Proofing Devices…Contd.
Poka Yoke / Mistake Proofing - Processing Errors -
Before Improvement:
It was possible to insert the
chassis in the jig backwards.
Correct operation depended
on the workers vigilance.
After Improvement:
A guide pin was added,
keyed to an asymmetrical
feature of the chassis. This
completely eliminates the
danger of backwards
processing.
Description of Process: A chassis was placed in a jig for machining
Problem: Chassis set backwards in jig
Solution: Additional guide pin taking advantage of asymmetry
Key Improvement: Jig modified to guarantee correct positioning
Guide
Pin
additional guide pin
6
223
Poka Yoke / Mistake Proofing - Missing Parts -
Before Improvement:
Detection of missing screw holes
depended on the vigilance of the
operators at further processes along
the line. However, defects got through
to final assembly often and the tuners
could not be mounted in the TVs.
After Improvement:
Rods for detecting the presence of the
screw holes were mounted on the jig for
inspecting the tuner assemblies. The
tuner cannot be set in position for
quality inspection unless the screw
holes are in the proper place. Defective
tuners are now detected before being
sent on.
Description of Process: Mounting brackets are added to the chassis of the TV
tuner for later attachment to the rest of the TV
assembly.
Problem: Sometimes the brackets are missing from the chassis.
Solution: Modify functional test test mounting to incorporate bracket check
Key Improvement: Jig modified to identify defective parts
inspection jig
Common Mistake Proofing Devices…Contd.
6
224
Common Mistake Proofing Devices…Contd.
Poka Yoke / Mistake Proofing - Omitted Processing -
Before Improvement:
The operator counted the holes as
they were drilled. However, the
operator sometimes made errors,
and the products with the wrong
number of holes
were produced.
After Improvement:
A counter was mounted on the drill
press to detect each hole as it is
drilled. Along with this, a limit
switch was mounted on the jig to
detect when a part was removed
before the proper number of holes
was drilled.
Description of Process: A number of holes are drilled in each workpiece
Problem: Incorrect number of holes drilled
Solution: An automatic counter to keep track of the number of holes
drilled
Key Improvement: Tool modified to guarantee correct processing count
limit switch for
detecting workpieces
workpiece
counter
correct
defective
buzzer
jig
6
225
Advantages of Poka-Yoke
• Simple – No formal training programs required.
• Inexpensive – No high cost investment required.
• Gives Prompt Feedback.
• Eliminates many inspection operations.
• Reduces operator dependence.
• Promotes Creativity and Value-Adding Activities.
• Results in Defect-Free Work.
• Requires immediate action when problems arise.
• Provides 100% inspection internal to the
operation.
6
226
GOOD Detects error before it continues to the
next operation
BETTER Allows for detection while error is being
made.
BEST Makes it impossible for errors to occur.
SUPERIOR Makes it impossible for errors to occur
with productivity gain.
Levels of Mistake-Proof Processes
6
227
368
“The Right Quality and Uniformity
are Foundations of
Commerce,Prosperity and Peace”
Deming
6
228
Lean Concepts
6
229
Defining Lean
Lean is:
“A systematic approach to identifying and
eliminating waste (non-value-added
activities) through continuous improvement
by flowing the product at the pull of the
customer in pursuit of perfection.”
— The MEP Lean Network
6
230
Reduced Lead Time
“One of the most noteworthy accomplishments
in keeping the price of Ford products low is
the gradual shortening of the production cycle.
The longer an article is in the process of
manufacture and the more it is moved about,
the greater is its ultimate cost.”
— Henry Ford, 1926
6
231
Definition of Value-Added
Value-Added
Any activity that
increases the market
form or function of
the product or
service. (These are
things the customer is
willing to pay for.)
Non-Value-Added
Any activity that does
not add market form
or function or is not
necessary. (These
activities should be
eliminated,
simplified, reduced,
or integrated.)
6
232
Lean = Eliminating Waste
Typically 95% of all lead time is non-value-added.
Value-Added Non-Value-Added
• Overproduction
• Waiting
• Transportation
• Non-value-added processing
• Excess inventory
• Defects
• Excess motion
• Underutilized people
6
233
Eight Wastes
6
234
Overproduction
• Making more than is required by the next process
• Making earlier than is required by the next process
• Making faster than is required by the next process
• Causes of overproduction:
– Just-in-case logic
– Misuse of automation
– Long process setup
– Unlevel scheduling
– Unbalanced workload
– Over engineered
– Redundant inspections
6
235
Inventory Waste
• Any supply in excess of a one-piece flow through
your manufacturing process
• Causes of excess inventory:
– Need for buffer against inefficiencies and unexpected
problems
– Product complexity
– Unleveled scheduling
– Poor market forecast
– Unbalanced workload
– Misunderstood communications
– Unreliable shipments by suppliers
6
236
Defects
• Inspection and repair of material in inventory
• Causes of defects:
– Weak process control
– Poor quality
– Unbalanced inventory level
– Deficient planned maintenance
– Inadequate education, training,
or work instructions
– Product design
– Customer needs not understood
6
237
Processing Waste
• Effort that adds no value to the product or service
from the customers’ viewpoint
• Causes of processing waste:
– Product changes without process changes
– Just-in-case logic
– True customer requirements not clearly defined
– Over-processing to accommodate downtime
– Lack of communication
– Redundant approvals
– Extra copies or excessive information
6
238
Waiting Waste
• Idle time created when waiting for…?
• Causes of waiting waste:
– Unbalanced workload
– Unplanned maintenance
– Long process setup times
– Misuses of automation
– Upstream quality problems
– Unlevel scheduling
6
239
People Waste
• The waste of not using people’s mental,
creative, and physical abilities
• Causes of people waste:
– Old guard thinking, politics, the business
culture
– Poor hiring practices
– Low or no investment in training
– Low pay, high turnover strategy
6
240
Motion Waste
• Any movement of people or machines that
does not add value to the product or service
• Causes of motion waste:
– Poor people or machine effectiveness
– Inconsistent work methods
– Unfavorable facility or cell layout
– Poor workplace organization and housekeeping
– Extra “busy” movements while waiting
6
241
Waste of Transportation
• Transporting parts and materials around the plant
• Causes of transportation waste:
– Poor plant layout
– Poor understanding of the process flow for production
– Large batch sizes, long lead times, and large storage
areas
6
242
Elements of a 5S Program
Sort — Perform “Sort
Through and Sort Out,” by
placing a red tag on all
unneeded items and moving
them to a temporary holding
area. Within a predetermined
time the red tag items are
disposed, sold, moved or
given away. “When in doubt,
throw it out!”
Set in Order — Identify the
best location for remaining
items, relocate out of place
items, set inventory limits,
and install temporary
location indicators.
Shine — Clean everything,
inside and out. Continue to
inspect items by cleaning
them and to prevent dirt,
grime, and contamination
from occurring.
Standardize — Create the
rules for maintaining and
controlling the first three S’s
and use visual controls.
Sustain — Ensure adherence
to the 5S standards through
communication, training, and
self-discipline.
6
243
Visual Controls
Simple signals that provide an immediate
understanding of a situation or condition.
They are efficient, self-regulating, and
worker-managed.
Examples:
• Kanban cards
• Color-coded dies, tools, pallets
• Lines on the floor to delineate storage areas,
walkways, work areas, etc.
• Andon lights
6
244
Plant Layout
Brake
Screw
Machine
Raw Stock QC Rec. Ship
Shear
QC
Assembly
Parts Stock
Stamp
Mill Lathe Drill
Finish
Weld Grind
6
245
CSSBB
(Certified Six Sigma Black Belt)
of ASQ(American Society for Quality)
www.asq.org
• Associate member fee : $74 per year
• Student member fee : $25 per year
• Exam fee : $270
• Exams twice a year
• Bengaluru is Exam center
• Two completed projects with signed affidavits
• QCFI conducts 60 hrs. of coaching classes in ten consecutive
Sundays- Fee Rs.20000 only
‘Develop your Career and Grow your Organization’
6
246
6
Science
Art
Magic
Six
Sigma
The Six Sigma Methodology can help us
move from Magic and Art to Science
6
248
Current State
Desired State
It’s Great to Have a Philosophy. . . But We
Need a Strategy !!
Don’t Change Your Strategy in the Middle
of the Maze
6
249
Strategic Business Policy
“Commitment to achieve excellence in all
our business processes, through the
application of ‘Six Sigma’ break through
methodology, by the year 2004”
Key Performance Indicators (KPI)
• Customer Satisfaction
• Cost reduction
• Cycle time improvement
• Yield improvement
• Design improvement
6
250
What is the
need?
When is the
need felt?
Where is the
need felt?
Why is the
need felt?
How is the situation
handled now?
AC should be silent Sound sleep At night
In the
bedroom
To remain
fresh next
morning
Uses a ceiling fan that
makes a lot of noise
AC should be efficient Good cooling At night
In the
bedroom
It gets very hot
in May - June
Uses a ceiling fan that
is not so effective in
summer
AC Should not cost
much
Affordability N / A N / A
Limited
finance
N / A
What customermeant
Household
member
1
VOC Table
Sl.
No.
Who is the
Customer?
What customersaid
(Voice of Customer)
Example of Air Conditioner
Customer CTQ’s
6 Some Plain Talk About Six Sigma
World Class Quality
Mikel J. Harry, Ph.D.
Chief Executive Officer
Six Sigma Academy, Inc.
Phoenix, Arizona
Today, focusing on the customer is absolutely essential. Of course, we all recognize this. But do we really internalize the idea?
Do we really believe that such a focus has the potential to drive business growth and impact the level of prosperity which we
should come to expect?
Closely linked to the idea of customer satisfaction is the concept of operational excellence -- the kingpin of success. Without a
focus on excellence, it becomes easy to accept the position of second or third best. Being the best means embracing change
and reaching out for new and higher standards of performance. Only then can one break the chains of complacency and pave
the way for breakthrough. The attainment of excellence is no longer a lofty goal or ideal, it is now a fundamental requirement --
the ante for entering the game of business.
As most of us already know, Japan has assumed this perspective and steadily acted upon it. The result of their focus has been
staggering, as evidenced by their superb products and services, not to mention their tremendous position in the marketplace.
Hence, smart money says we must view the idea of operational excellence as a major force in today's marketplace. For those
who might doubt such an assertion, just ask a customer what they think. After all, what the customer thinks provides us with a
strong indicator of what will be purchased -- and from whom.
Of course, it is widely recognized that the operational performance of an organization is largely determined by the capability of its
processes. Another way of looking at this would be to say that our performance as a company is governed by the quality of our
processes -- high quality processes delivers high quality products, at the lowest possible cost, on time. Therefore, a focus on
operational excellence (in everything we do) translates to a focus on process quality. Of course, we can not focus on what we do
not measure and if we do not measure, we can not improve.
Trying to improve something when you don't have a standard to measure against, is like playing a sports game without knowing
the score. Can you imagine setting out on cross-country trip in an automobile without a fuel gauge? Just think of the personal
grief, cost, and inconvenience which might result. Would you allow yourself to be placed in such a situation?
6
As we will discover, the measurement and improvement of our processes is absolutely essential if we are to achieve operational excellence and the
ideals of total quality. To do this, we must bear in mind the old axiom -- let the product do the talking. In other words, the quality of our products can
tell us how capable our processes really are. To measure product quality is to measure process quality because the two are correlated. It is very
important to recognize that the inverse of this also holds true.
Perhaps we should now set the stage for our ensuing discussion by examining what some have called the "chain of causation:"
Our survival is dependent upon growing the business.
Our business growth is largely determined by customer satisfaction.
Customer Satisfaction is governed by quality, price, and delivery.
Quality, price, and delivery is controlled by process capability.
Our process capability is greatly limited by variation.
Process variation leads to an increase in defects, cost, and cycle time.
To eliminate variation, we must apply the right knowledge.
In order to apply the right knowledge, we must first acquire it.
To acquire new knowledge means that we must have the will to survive.
If you can't express something in the form of numbers you don't really know much about it. And if you don't know much about it, you can't control it.
And if you can't control it, you're at the mercy of chance. And if you're at the mercy of chance, why bother with it? Hence, we must learn the
language of numbers.
Such thinking represents a business philosophy -- a way of guiding our company. From this perspective, we will focus our discussion on an all
embracing standard and methodology called "Six Sigma." As we shall see, Six Sigma can be used to measure the quality of our work processes -- in
any field, from assembling a motor car to inspiring a classroom full of students.
Over the years, this writer has been asked a great many questions about Six Sigma -- most of which have been quite simple, practical, and straight
forward. For the sake of simplicity and reading ease, we shall complement such questions with simple, practical, and straight forward answers. Of
course, we recognize that more sophisticated and detailed answers do exist, as many students of Six Sigma will testify. However, there is no point in
hooking-up a fire hose when all we want is a glass of water.
QUESTION: What is Six Sigma?
ANSWER: Six Sigma is several things. First, it is a statistical measurement. It tells us how good our products, services, and processes really are.
The Six Sigma method allows us to draw comparisons to other similar or dissimilar products, services, and processes. In this manner, we can see
how far ahead or behind we are. Most importantly, we can see where we need to go and what we must do to get there. In other words, Six Sigma
helps us to establish our course and gauge our pace in the race for total customer satisfaction.
Some Plain Talk About Six Sigma…Contd.
Sid Sigma educational and problem solving power point- Six Sigma.ppt
Sid Sigma educational and problem solving power point- Six Sigma.ppt
Sid Sigma educational and problem solving power point- Six Sigma.ppt
Sid Sigma educational and problem solving power point- Six Sigma.ppt
Sid Sigma educational and problem solving power point- Six Sigma.ppt
Sid Sigma educational and problem solving power point- Six Sigma.ppt

More Related Content

Similar to Sid Sigma educational and problem solving power point- Six Sigma.ppt

Six sigma seminar
Six sigma seminarSix sigma seminar
Six sigma seminar
jairajputana
 
LSSBB Training Material Version.pdf
LSSBB Training Material Version.pdfLSSBB Training Material Version.pdf
LSSBB Training Material Version.pdf
Ranjit708960
 
1 introduction-to-six-sigma-458-k-ppt4941
1 introduction-to-six-sigma-458-k-ppt49411 introduction-to-six-sigma-458-k-ppt4941
1 introduction-to-six-sigma-458-k-ppt4941
Shahzad Danish
 
Six sigma introduction
Six sigma introductionSix sigma introduction
Six sigma introduction
Md Khairul Shaleh Md Asari
 
Basic overview six sigma
Basic overview six sigmaBasic overview six sigma
Basic overview six sigma
Khushmeet Khushi
 
6 sigma
6 sigma6 sigma
6 sigma
Shabbar Ali
 
Six Sigma
Six SigmaSix Sigma
Six Sigma
Tushar Makvana
 
Lean Six Sigma Introduction free web class.pptx
Lean Six Sigma Introduction free web class.pptxLean Six Sigma Introduction free web class.pptx
Lean Six Sigma Introduction free web class.pptx
Daniel397407
 
Six sigma
Six sigmaSix sigma
Chapter 4 six sigma
Chapter 4 six sigmaChapter 4 six sigma
Chapter 4 six sigma
Prof.Mayur Modi
 
Sixsigmaintrorerevisedii 13198308548378-phpapp02-111028144148-phpapp02
Sixsigmaintrorerevisedii 13198308548378-phpapp02-111028144148-phpapp02Sixsigmaintrorerevisedii 13198308548378-phpapp02-111028144148-phpapp02
Sixsigmaintrorerevisedii 13198308548378-phpapp02-111028144148-phpapp02
satyajit pandey
 
Yellow belt training 68 s
Yellow belt training 68 sYellow belt training 68 s
Yellow belt training 68 s
Rachit Gaur
 
Six Sigma Introduction
Six Sigma IntroductionSix Sigma Introduction
Six Sigma Introduction
Abhishek Kumar
 
Six sigma - pimg
Six sigma - pimgSix sigma - pimg
Six sigma - pimg
KULDEEP MATHUR
 
Presentation introduction to six sigma
Presentation   introduction to six sigmaPresentation   introduction to six sigma
Presentation introduction to six sigma
jpanchpor
 
Six sigma
Six sigmaSix sigma
Six sigma
Kumar Kunal
 
SIX SIGMA
SIX SIGMA SIX SIGMA
SIX SIGMA
Kumar Kunal
 
Lean six sigma Yellow Belt Complete training
Lean six sigma Yellow Belt Complete trainingLean six sigma Yellow Belt Complete training
Lean six sigma Yellow Belt Complete training
Ankit Sharma
 
Six Sigma
Six SigmaSix Sigma
Knowledge Sharing Session on Six Sigma
Knowledge Sharing Session on Six SigmaKnowledge Sharing Session on Six Sigma
Knowledge Sharing Session on Six Sigma
DrGPurushothaman
 

Similar to Sid Sigma educational and problem solving power point- Six Sigma.ppt (20)

Six sigma seminar
Six sigma seminarSix sigma seminar
Six sigma seminar
 
LSSBB Training Material Version.pdf
LSSBB Training Material Version.pdfLSSBB Training Material Version.pdf
LSSBB Training Material Version.pdf
 
1 introduction-to-six-sigma-458-k-ppt4941
1 introduction-to-six-sigma-458-k-ppt49411 introduction-to-six-sigma-458-k-ppt4941
1 introduction-to-six-sigma-458-k-ppt4941
 
Six sigma introduction
Six sigma introductionSix sigma introduction
Six sigma introduction
 
Basic overview six sigma
Basic overview six sigmaBasic overview six sigma
Basic overview six sigma
 
6 sigma
6 sigma6 sigma
6 sigma
 
Six Sigma
Six SigmaSix Sigma
Six Sigma
 
Lean Six Sigma Introduction free web class.pptx
Lean Six Sigma Introduction free web class.pptxLean Six Sigma Introduction free web class.pptx
Lean Six Sigma Introduction free web class.pptx
 
Six sigma
Six sigmaSix sigma
Six sigma
 
Chapter 4 six sigma
Chapter 4 six sigmaChapter 4 six sigma
Chapter 4 six sigma
 
Sixsigmaintrorerevisedii 13198308548378-phpapp02-111028144148-phpapp02
Sixsigmaintrorerevisedii 13198308548378-phpapp02-111028144148-phpapp02Sixsigmaintrorerevisedii 13198308548378-phpapp02-111028144148-phpapp02
Sixsigmaintrorerevisedii 13198308548378-phpapp02-111028144148-phpapp02
 
Yellow belt training 68 s
Yellow belt training 68 sYellow belt training 68 s
Yellow belt training 68 s
 
Six Sigma Introduction
Six Sigma IntroductionSix Sigma Introduction
Six Sigma Introduction
 
Six sigma - pimg
Six sigma - pimgSix sigma - pimg
Six sigma - pimg
 
Presentation introduction to six sigma
Presentation   introduction to six sigmaPresentation   introduction to six sigma
Presentation introduction to six sigma
 
Six sigma
Six sigmaSix sigma
Six sigma
 
SIX SIGMA
SIX SIGMA SIX SIGMA
SIX SIGMA
 
Lean six sigma Yellow Belt Complete training
Lean six sigma Yellow Belt Complete trainingLean six sigma Yellow Belt Complete training
Lean six sigma Yellow Belt Complete training
 
Six Sigma
Six SigmaSix Sigma
Six Sigma
 
Knowledge Sharing Session on Six Sigma
Knowledge Sharing Session on Six SigmaKnowledge Sharing Session on Six Sigma
Knowledge Sharing Session on Six Sigma
 

Recently uploaded

Training on CSPro and step by steps.pptx
Training on CSPro and step by steps.pptxTraining on CSPro and step by steps.pptx
Training on CSPro and step by steps.pptx
lenjisoHussein
 
Celebrity Girls Call Noida 9873940964 Unlimited Short Providing Girls Service...
Celebrity Girls Call Noida 9873940964 Unlimited Short Providing Girls Service...Celebrity Girls Call Noida 9873940964 Unlimited Short Providing Girls Service...
Celebrity Girls Call Noida 9873940964 Unlimited Short Providing Girls Service...
ginni singh$A17
 
BDSM Girls Call Delhi 🎈🔥9711199171 🔥💋🎈 Provide Best And Top Girl Service And ...
BDSM Girls Call Delhi 🎈🔥9711199171 🔥💋🎈 Provide Best And Top Girl Service And ...BDSM Girls Call Delhi 🎈🔥9711199171 🔥💋🎈 Provide Best And Top Girl Service And ...
BDSM Girls Call Delhi 🎈🔥9711199171 🔥💋🎈 Provide Best And Top Girl Service And ...
fatima shekh$A17
 
Semantic Web and organizational data .pptx
Semantic Web and organizational data .pptxSemantic Web and organizational data .pptx
Semantic Web and organizational data .pptx
Kanchana Weerasinghe
 
Harnessing Wild and Untamed (Publicly Available) Data for the Cost efficient ...
Harnessing Wild and Untamed (Publicly Available) Data for the Cost efficient ...Harnessing Wild and Untamed (Publicly Available) Data for the Cost efficient ...
Harnessing Wild and Untamed (Publicly Available) Data for the Cost efficient ...
weiwchu
 
FINAL PROJECT WORK PORTFOLIO MANAGEMENT (2) hhh (1) (2) (5) (1) (1).pdf
FINAL PROJECT WORK PORTFOLIO MANAGEMENT (2)  hhh (1) (2) (5) (1) (1).pdfFINAL PROJECT WORK PORTFOLIO MANAGEMENT (2)  hhh (1) (2) (5) (1) (1).pdf
FINAL PROJECT WORK PORTFOLIO MANAGEMENT (2) hhh (1) (2) (5) (1) (1).pdf
bala krishna
 
Oracle Database Desupported Features on 23ai (Part B)
Oracle Database Desupported Features on 23ai (Part B)Oracle Database Desupported Features on 23ai (Part B)
Oracle Database Desupported Features on 23ai (Part B)
Alireza Kamrani
 
DataScienceConcept_Kanchana_Weerasinghe.pptx
DataScienceConcept_Kanchana_Weerasinghe.pptxDataScienceConcept_Kanchana_Weerasinghe.pptx
DataScienceConcept_Kanchana_Weerasinghe.pptx
Kanchana Weerasinghe
 
🚂🚘 Premium Girls Call Guwahati 🛵🚡000XX00000 💃 Choose Best And Top Girl Servi...
🚂🚘 Premium Girls Call Guwahati  🛵🚡000XX00000 💃 Choose Best And Top Girl Servi...🚂🚘 Premium Girls Call Guwahati  🛵🚡000XX00000 💃 Choose Best And Top Girl Servi...
🚂🚘 Premium Girls Call Guwahati 🛵🚡000XX00000 💃 Choose Best And Top Girl Servi...
kuldeepsharmaks8120
 
Histology of Muscle types histology o.ppt
Histology of Muscle types histology o.pptHistology of Muscle types histology o.ppt
Histology of Muscle types histology o.ppt
SamanArshad11
 
Experience, Excellence & Commitment are the characteristics that describe Fla...
Experience, Excellence & Commitment are the characteristics that describe Fla...Experience, Excellence & Commitment are the characteristics that describe Fla...
Experience, Excellence & Commitment are the characteristics that describe Fla...
kittycrispy617
 
High Girls Call Nagpur 000XX00000 Provide Best And Top Girl Service And No1 i...
High Girls Call Nagpur 000XX00000 Provide Best And Top Girl Service And No1 i...High Girls Call Nagpur 000XX00000 Provide Best And Top Girl Service And No1 i...
High Girls Call Nagpur 000XX00000 Provide Best And Top Girl Service And No1 i...
saadkhan1485265
 
Potential Uses of the Floyd-Warshall Algorithm as appropriate
Potential Uses of the Floyd-Warshall Algorithm as appropriatePotential Uses of the Floyd-Warshall Algorithm as appropriate
Potential Uses of the Floyd-Warshall Algorithm as appropriate
huseindihon
 
社内勉強会資料_TransNeXt: Robust Foveal Visual Perception for Vision Transformers
社内勉強会資料_TransNeXt: Robust Foveal Visual Perception for Vision Transformers社内勉強会資料_TransNeXt: Robust Foveal Visual Perception for Vision Transformers
社内勉強会資料_TransNeXt: Robust Foveal Visual Perception for Vision Transformers
NABLAS株式会社
 
VVIP Girls Call Noida 9873940964 Provide Best And Top Girl Service And No1 in...
VVIP Girls Call Noida 9873940964 Provide Best And Top Girl Service And No1 in...VVIP Girls Call Noida 9873940964 Provide Best And Top Girl Service And No1 in...
VVIP Girls Call Noida 9873940964 Provide Best And Top Girl Service And No1 in...
Ak47
 
New Girls Call Noida 9873940964 Unlimited Short Providing Girls Service Avail...
New Girls Call Noida 9873940964 Unlimited Short Providing Girls Service Avail...New Girls Call Noida 9873940964 Unlimited Short Providing Girls Service Avail...
New Girls Call Noida 9873940964 Unlimited Short Providing Girls Service Avail...
kinni singh$A17
 
Celebrity Girls Call Andheri 9930245274 Unlimited Short Providing Girls Servi...
Celebrity Girls Call Andheri 9930245274 Unlimited Short Providing Girls Servi...Celebrity Girls Call Andheri 9930245274 Unlimited Short Providing Girls Servi...
Celebrity Girls Call Andheri 9930245274 Unlimited Short Providing Girls Servi...
revolutionary575
 
New Girls Call Delhi 🎈🔥9711199171 🔥💋🎈 Provide Best And Top Girl Service And N...
New Girls Call Delhi 🎈🔥9711199171 🔥💋🎈 Provide Best And Top Girl Service And N...New Girls Call Delhi 🎈🔥9711199171 🔥💋🎈 Provide Best And Top Girl Service And N...
New Girls Call Delhi 🎈🔥9711199171 🔥💋🎈 Provide Best And Top Girl Service And N...
tanupasswan6
 
High Profile Girls Call Delhi 🛵🚡9711199171 💃 Choose Best And Top Girl Service...
High Profile Girls Call Delhi 🛵🚡9711199171 💃 Choose Best And Top Girl Service...High Profile Girls Call Delhi 🛵🚡9711199171 💃 Choose Best And Top Girl Service...
High Profile Girls Call Delhi 🛵🚡9711199171 💃 Choose Best And Top Girl Service...
kinni singh$A17
 
DU degree offer diploma Transcript
DU degree offer diploma TranscriptDU degree offer diploma Transcript
DU degree offer diploma Transcript
uapta
 

Recently uploaded (20)

Training on CSPro and step by steps.pptx
Training on CSPro and step by steps.pptxTraining on CSPro and step by steps.pptx
Training on CSPro and step by steps.pptx
 
Celebrity Girls Call Noida 9873940964 Unlimited Short Providing Girls Service...
Celebrity Girls Call Noida 9873940964 Unlimited Short Providing Girls Service...Celebrity Girls Call Noida 9873940964 Unlimited Short Providing Girls Service...
Celebrity Girls Call Noida 9873940964 Unlimited Short Providing Girls Service...
 
BDSM Girls Call Delhi 🎈🔥9711199171 🔥💋🎈 Provide Best And Top Girl Service And ...
BDSM Girls Call Delhi 🎈🔥9711199171 🔥💋🎈 Provide Best And Top Girl Service And ...BDSM Girls Call Delhi 🎈🔥9711199171 🔥💋🎈 Provide Best And Top Girl Service And ...
BDSM Girls Call Delhi 🎈🔥9711199171 🔥💋🎈 Provide Best And Top Girl Service And ...
 
Semantic Web and organizational data .pptx
Semantic Web and organizational data .pptxSemantic Web and organizational data .pptx
Semantic Web and organizational data .pptx
 
Harnessing Wild and Untamed (Publicly Available) Data for the Cost efficient ...
Harnessing Wild and Untamed (Publicly Available) Data for the Cost efficient ...Harnessing Wild and Untamed (Publicly Available) Data for the Cost efficient ...
Harnessing Wild and Untamed (Publicly Available) Data for the Cost efficient ...
 
FINAL PROJECT WORK PORTFOLIO MANAGEMENT (2) hhh (1) (2) (5) (1) (1).pdf
FINAL PROJECT WORK PORTFOLIO MANAGEMENT (2)  hhh (1) (2) (5) (1) (1).pdfFINAL PROJECT WORK PORTFOLIO MANAGEMENT (2)  hhh (1) (2) (5) (1) (1).pdf
FINAL PROJECT WORK PORTFOLIO MANAGEMENT (2) hhh (1) (2) (5) (1) (1).pdf
 
Oracle Database Desupported Features on 23ai (Part B)
Oracle Database Desupported Features on 23ai (Part B)Oracle Database Desupported Features on 23ai (Part B)
Oracle Database Desupported Features on 23ai (Part B)
 
DataScienceConcept_Kanchana_Weerasinghe.pptx
DataScienceConcept_Kanchana_Weerasinghe.pptxDataScienceConcept_Kanchana_Weerasinghe.pptx
DataScienceConcept_Kanchana_Weerasinghe.pptx
 
🚂🚘 Premium Girls Call Guwahati 🛵🚡000XX00000 💃 Choose Best And Top Girl Servi...
🚂🚘 Premium Girls Call Guwahati  🛵🚡000XX00000 💃 Choose Best And Top Girl Servi...🚂🚘 Premium Girls Call Guwahati  🛵🚡000XX00000 💃 Choose Best And Top Girl Servi...
🚂🚘 Premium Girls Call Guwahati 🛵🚡000XX00000 💃 Choose Best And Top Girl Servi...
 
Histology of Muscle types histology o.ppt
Histology of Muscle types histology o.pptHistology of Muscle types histology o.ppt
Histology of Muscle types histology o.ppt
 
Experience, Excellence & Commitment are the characteristics that describe Fla...
Experience, Excellence & Commitment are the characteristics that describe Fla...Experience, Excellence & Commitment are the characteristics that describe Fla...
Experience, Excellence & Commitment are the characteristics that describe Fla...
 
High Girls Call Nagpur 000XX00000 Provide Best And Top Girl Service And No1 i...
High Girls Call Nagpur 000XX00000 Provide Best And Top Girl Service And No1 i...High Girls Call Nagpur 000XX00000 Provide Best And Top Girl Service And No1 i...
High Girls Call Nagpur 000XX00000 Provide Best And Top Girl Service And No1 i...
 
Potential Uses of the Floyd-Warshall Algorithm as appropriate
Potential Uses of the Floyd-Warshall Algorithm as appropriatePotential Uses of the Floyd-Warshall Algorithm as appropriate
Potential Uses of the Floyd-Warshall Algorithm as appropriate
 
社内勉強会資料_TransNeXt: Robust Foveal Visual Perception for Vision Transformers
社内勉強会資料_TransNeXt: Robust Foveal Visual Perception for Vision Transformers社内勉強会資料_TransNeXt: Robust Foveal Visual Perception for Vision Transformers
社内勉強会資料_TransNeXt: Robust Foveal Visual Perception for Vision Transformers
 
VVIP Girls Call Noida 9873940964 Provide Best And Top Girl Service And No1 in...
VVIP Girls Call Noida 9873940964 Provide Best And Top Girl Service And No1 in...VVIP Girls Call Noida 9873940964 Provide Best And Top Girl Service And No1 in...
VVIP Girls Call Noida 9873940964 Provide Best And Top Girl Service And No1 in...
 
New Girls Call Noida 9873940964 Unlimited Short Providing Girls Service Avail...
New Girls Call Noida 9873940964 Unlimited Short Providing Girls Service Avail...New Girls Call Noida 9873940964 Unlimited Short Providing Girls Service Avail...
New Girls Call Noida 9873940964 Unlimited Short Providing Girls Service Avail...
 
Celebrity Girls Call Andheri 9930245274 Unlimited Short Providing Girls Servi...
Celebrity Girls Call Andheri 9930245274 Unlimited Short Providing Girls Servi...Celebrity Girls Call Andheri 9930245274 Unlimited Short Providing Girls Servi...
Celebrity Girls Call Andheri 9930245274 Unlimited Short Providing Girls Servi...
 
New Girls Call Delhi 🎈🔥9711199171 🔥💋🎈 Provide Best And Top Girl Service And N...
New Girls Call Delhi 🎈🔥9711199171 🔥💋🎈 Provide Best And Top Girl Service And N...New Girls Call Delhi 🎈🔥9711199171 🔥💋🎈 Provide Best And Top Girl Service And N...
New Girls Call Delhi 🎈🔥9711199171 🔥💋🎈 Provide Best And Top Girl Service And N...
 
High Profile Girls Call Delhi 🛵🚡9711199171 💃 Choose Best And Top Girl Service...
High Profile Girls Call Delhi 🛵🚡9711199171 💃 Choose Best And Top Girl Service...High Profile Girls Call Delhi 🛵🚡9711199171 💃 Choose Best And Top Girl Service...
High Profile Girls Call Delhi 🛵🚡9711199171 💃 Choose Best And Top Girl Service...
 
DU degree offer diploma Transcript
DU degree offer diploma TranscriptDU degree offer diploma Transcript
DU degree offer diploma Transcript
 

Sid Sigma educational and problem solving power point- Six Sigma.ppt

  • 2. 6 2 Emerging Paradigm • Rapid Pace of technology change • Shrinking Life cycle
  • 3. 6 3 How to cope with the Emerging Paradigm? Lean Thinking!
  • 4. 6 4 Lean Thinking ‘Lean’ is for increasing the speed by eliminating wastes, where as ‘SIX SIGMA’ is for improving Quality Combined together they make our Operations excellent.
  • 5. 6 5 WR RR WW RW BUSINESS EXCELLENCE Strategic Excellence Operations Excellence
  • 6. 6 6 Strategy Excellence (Doing the right things) 1.Corporate Level Strategy • What businesses to go into • Merger and Acquisition • Diversification • Allocation of Resources • Such plans concerning matters of vital, long term and continuing importance to organization (After studying the Economical, Technological, Social and Political facets)
  • 7. 6 7 Strategy Excellence (Doing the right things) 2. Business Level Strategy (SBU/Unit) • Competing in a given business or Product area through OPERATIONS EXCELLENCE
  • 8. 6 8 Operations Excellence (Doing things right) Includes : • Marketing Excellence • D & E Excellence • Manufacturing & Service Excellence
  • 9. 6 9 SIX SIGMA - A means to realize the philosophy and values associated with TQM
  • 10. 6 10 - the ultimate aspiration • SIX SIGMA - the threshold of excellence • ZERO DEFECTS Dr. Mikel J. Harry
  • 11. 6 11 Missing Strategy for QI Tools & Techniques Strategy Discipline Philosophy & Vision Quality Vision existed for years (ISO, “O defects”…) What was missing -- Six Sigma! Many tools existed for years (SPC,DOE...)
  • 12. 6 12 What Is Six Sigma ?  IT IS A STATISTICAL MEASUREMENT (HOW GOOD ?) - Dr. Mikel J. Harry  IT IS A BUSINESS STRATEGY (COMPETITIVE EDGE).  IT IS A PHILOSOPHY. (WORKING SMARTER AND NOT HARDER).
  • 13. 6 13 Example: OFFICE IN-PUNCH time Historic data of IN-PUNCH timings for the last 50 weeks. 50 Weekly Average values Target time = 9.00hrs USL = 9.18 hrs LSL = 8.42 hrs DATA: Understanding Six Sigma
  • 14. 6 14 906 902 910 914 918 846 842 850 854 858   = 06 USL LSL = 900 = 900  T 3
  • 15. 6 15 T  906 902 900 910 846 842 850 854 858 838 834 = 900 = 852 USL LSL
  • 16. 6 16 61 T  1 USL LSL 1 = 03 = 900 = 900 905 907 909 853 851 855 857 859 901 903 842 918
  • 17. 6 17   = 06 USL LSL 3 = 900 909 851 1 USL 842 1= 03 61 918 = 900 LSL 906 902 910 914 918 846 842 850 854 858 T 
  • 18. 6 18 Quality Improvement is Problem with Spread Accurate but not Precise Desired Current Situation LSL USL T Shrinking the variation & Problem with Centering Shifting the mean to the target d Precise but not Accurate USL LSL Current Situation Desired
  • 19. 6 19 Sigma Scale (Distribution Shifted 1.5) 2 3 4 5 6 308,537 66,807 6,210 233 3.4 Z Defects PPM 7.5 0.001
  • 20. 6 20 Six Sigma Centered Short Term Zst = 6.0 DPPMst = .002 Zlt = 4.5 DPPMlt = 3.4 USL LSL 3.4 ppm LSL USL 4.5 T  .000 ppm Six Sigma Shifted 1.5 Long Term Short Term / Long Term USL LSL ± 6 .001 ppm .001 ppm USL T  7.5
  • 21. 6 21 SIGMA SPELLING 3 1.5 misspelled words per page in a book 4 1 misspelled word per 30 pages in a book 5 1 misspelled word in a set of encyclopedias 6 1 misspelled word in all of the books contained in a small library Practical Meaning
  • 22. 6 22 Understanding The Difference 4 Capability: Defect Dots = 1849 6 Capability: Defect Dots = 1
  • 23. 6 23 d 1 2 3 4 5 6 7 8 9 10 a           b           c           d           Units = 10 Opp/unit = 4 DPO = 0.25 TOP = 40 Defects = 10 Z = 2.17 DPO  Probability of Defect c b a Z calculation for Discrete Data
  • 24. 6 24 1 2 3 4 5 6 7 8 9 10 a           b           c           d           Units = 10 Opp/unit = 4 DPO = 0.025 TOP = 40 Defect = 1 Z = 3.46 DPO  Probability of Defect d c b a Z calculation for Discrete Data
  • 26. 6 26 SIX SIGMA WHAT IS DIFFERENT ? • SIX SIGMA ORGANIZATION • DISCIPLINED BREAKTHROUGH METHODOLOGY • UNIQUE TRAINING METHODOLOGY • RECOGNITION SYSTEM
  • 27. 6 27 The Six Sigma 0rgn. Grand Champion Master Black Belts Full Time Six Sigma Black Belts Full Time Six Sigma Green Belts Part Time Deployment Teams Champions
  • 28. 6 28 Role of Champion Create the vision of Six Sigma for the company Select high-impact projects Review regularly to see that Black Belts / Green Belts complete the project. Support development of “statistical thinking” Ask Black Belts / Green Belts many questions to ensure that they are properly focused. Realize the gains by supporting Six Sigma projects through allocation of resources and removal of road blocks. Hold the ground by implementing Black Belt / Green Belt recommendation.       
  • 29. 6 29 Role of MBB Assist champion in the identification of projects. Develop and deliver training to Black Belts / Green Belts. Coach and support Black Belts in project work. Participate in project reviews conducted by the champion to offer technical expertise. Facilitate sharing of best practices across the company. * * * * *
  • 30. 6 30 Role of Black Belt Be the change agent of the company-(full time at least for 2 years) Be the Six Sigma Break though strategy expert and enthusiast. Support Green Belts in addition to doing own’s projects. Influence without direct authority. Identify barriers and stimulate champion thinking.     
  • 31. 6 31 Role of Green Belt Do projects on a part time basis, while performing the regular duties Continue to learn and practice the six sigma methods after project completion.  
  • 32. 6 32 SIX SIGMA WHAT IS DIFFERENT ? • SIX SIGMA ORGANIZATION • DISCIPLINED BREAKTHROUGH METHODOLOGY • UNIQUE TRAINING METHODOLOGY • RECOGNITION SYSTEM
  • 33. 6 33 2 to 3 Sigma = 5X improvement 3 to 4 Sigma = 10X improvement 4 to 5 Sigma = 27X improvement 5 to 6 Sigma = 70X improvement So... 3 to 6 Sigma = 19600X improvement!!! Breakthrough Improvement
  • 34. 6 34 Sweet Fruit Design for Six Sigma Bulk of Fruit Process Characterization and Optimization Low Hanging Fruit Seven Basic Tools Ground Fruit Logic and Intuition 3 Level Achieving Six Sigma is like reaching for the fruit at the top of a tree...it gets progressively harder to do! 4 Level 5 Level
  • 35. 6 35 Product = ƒ (design, manufacturing process) Focus on this is DMADV Focus on this is DMAIC The benefits can only be realized if the ‘design’ and ‘manufacturing Process’ can play together Six Sigma Methodologies
  • 38. 6 38 Design Opportunity Difficult to see/predict Easy to fix Easy to see Costly to fix Defects are: $ Research Design Prototype Production Customer Most DMAIC Six Sigma effort is here. Cost to Correct Quality Moving upstream (DFSS) increases return on investment
  • 39. 6 39 SIX SIGMA WHAT IS DIFFERENT ? • SIX SIGMA ORGANIZATION • DISCIPLINED BREAKTHROUGH METHODOLOGY • UNIQUE TRAINING METHODOLOGY • RECOGNITION SYSTEM
  • 40. 6 40 Six Sigma Training- What is different? • Six sigma Training is project based • The projects are linked to the bottom line • The computer is used extensively in Six Sigma training • Six Sigma training methods involve roadmaps to explain the application of improvement tools • Six Sigma training focuses on how to apply the tools to improve processes and not on the tools specifically • Six Sigma instructors rely on a variety of media to deliver material
  • 41. 6 41 Which factors have the greatest effect on results? Are All These Factors Vital ? 4 Pin Position Rubber Band Stop Angle 1 3 3 Start Angle 1 “Wiffle” “Solid” Ball type Cup position Hook position
  • 42. 6 42 SIX SIGMA WHAT IS DIFFERENT ? • SIX SIGMA ORGANIZATION • DISCIPLINED BREAKTHROUGH METHODOLOGY • UNIQUE TRAINING METHODOLOGY • RECOGNITION SYSTEM
  • 43. 6 43 Recognition System •No body gets promoted to an executive position at GE without Six Sigma Training. •“Get only your best people in the Six Sigma program and give them the options”. - Jack Welch
  • 44. 6 44 Latent Defects. ‘BILL SMITH’ an Engineer at ‘MOTOROLA’ was studying the correlation between a product’s life and how often it had been repaired during the manufacturing process. In 1985 he presented a paper that concluded that even if an item had been found defective and corrected during the manufacturing process, other defects were bound to be missed. These would then show up in the product’s early days with the customer. When, however a product was manufactured error- free, it rarely failed during the aforementioned early days.
  • 45. 6 45 0% 5% 10% 15% 20% 25% 3 4 5 6 7 Cost of Poor Quality (% of Revenues) Sigma level Cost of Poor Quality Versus Sigma Level - Thomas Pyzdek
  • 46. 6 Find and Control the Critical X’s  Y  Dependent  Output  Effect  Symptom  Monitor  X1 . . . XN  Independent  Input-Process  Cause  Problem  Control Putting Science into Everything We Do! f (X) Y= Establishing Y=f(X) The Framework for Six Sigma
  • 47. 6 The Impact of Added Inspection Escaping PPM Number of Consecutive Inspectors The Y axis represents the undetected defects-per-million defects. Each curve represents the inspection efficiency per independent inspection. 99% 90% 80% 70% Example: If the likelihood of detecting the defect is 70% and we have 10 consecutive inspectors with this level of capability, we would expect about 6 escaping defects out of every 1,000,000 defects produced. 1 10 100 1000 10000 100000 1000000 1 2 3 4 5 6 7 8 9 10 11 6 3.4 ppm Inspection is an expensive and time consuming way to get to Six Sigma ! Focus must be on doing it right the first time...
  • 48. 6 Unmasking the Hidden Factory “Theoretical Cycle Time: The back-to-back process time required for a single unit to complete all stages of a task without waiting, stopping, or setups.” T = test A = analysis F = fix Product A F A F Step 1 T Step 2 Floor Space Floor Space Floor Space T Value Added Non-Value Added The Hidden Factory
  • 49. 6 49 Project Documentation Why document? To preserve the project details To ensure recipe is not lost and problem does not come back To share the knowledge with others To prevent re-inventing the wheel Also an ISO 9000 requirement     
  • 50. 6 50 Project Monitor Project Documentation….Contd. Contents of the project binder  How the project was selected (brief summary)  Team members and contacts  Discussions held with the various relevant depts. regarding the projects  All forms of data collected  R-0 form  Presentation details after characterization  Presentation details after optimization  Relevant office orders if any, issued for implementing the recommendation  Financial savings details  Many more details beyond the presentation materials.
  • 51. 6 51 Chain of Causation PROCESS VARIATION LEADS TO AN INCREASE IN DEFECTS, COSTS AND CYCLE TIME. - Dr. Mikel J. Harry OUR SURVIVAL IS DEPENDENT UPON GROWING OUR BUSINESS. OUR BUSINESS GROWTH IS LARGELY DETERMINED BY CUSTOMER SATISFACTION. PROCESS CAPABILITY IS GREATLY LIMITED BY VARIATION. Q, P & D IS CONTROLLED BY PROCESS CAPABILITY. CUSTOMER SATISFACTION IS GOVERNED BY Q,P & D.
  • 52. 6 52  A Vision  An Organization  Champions  Black Belts ,Green Belts and the Teams  Master Black Belts  Training and Application  BB/GB must have a project  Entire company must be involved and trained  Business relevance is key to project selection  Review Mechanisms  Champions and operational managers responsibility  A Rigorous Methodology  DMAIC  DMADV  Recognition System  Supplier Involvement  Communication  Six Sigma website Operationalizing Six Sigma is the Key to Success To summarize the elements of this strategy: Summary 6
  • 54. 6 54 • Has a clearly defined problem statement * Is clearly linked to the customer * Is relevant to the business and will have positive impact * Relates directly to the GB’s job responsibilities, Gs&Os * Has a clearly defined and measurable defect (Y) * Is manageable in scope A Well-Defined Project: Problem Statement Definition looks easy - difficult for many
  • 55. 6 55 Good Example In the last 3 months, 12% of our customers are late, by over 45 days in paying their bills. This represents 20% of our outstanding receivables & negatively affects our operating cash flow Poor Example Our customers are angry with us and thus delay paying their bills What When Consequence Magnitude Defining a Problem Statement
  • 56. 6 56  Are you aware of any problem that a customer is having with the products/services your organization offers?  Is the Quality from competitive products/services better?  Do you have a persistent problem that you have attempted to fix in the past with limited success?  Are your cycle times too long in any process?  Are your costs too high in any process?  Do you have regulatory/compliance problem? Any of these questions can be addressed through a project based on Six Sigma thinking Project Selection Checklist
  • 57. 6 57 General problem: Electric motor reliability is poor Project scope: Reduce variation in brush hardness Electric motor issue Brush wear issue Brush hardness variability Project Focussing
  • 58. 6 58 Y1 – Electric motor reliability depends on: X1 – Motor reliability X2 – Controller reliability X3 – Mechanical mounting integrity X4 – Specific application or use Y3 – Brush reliability depends on: X1 – Assembly stack up issues X2 – Brush brittleness issues X3 – Contamination of the brush assembly X4 – Spring rate or condition issues X5 – Brush dimensional issues X6 – Brush hardness issues Y2 – The reliability of the motor itself depends on the reliability of the: X1 – Stator X2 – Rotor X3 – Brush X4 - Housing Y4 – Brush hardness issues depend on: X1 – Mean brush hardness X2 – Variation in brush hardness Y1 = X1+X2+X3+X4 Y2 = X1+X2+X3+X4 Y3 = X1+X2+X3+X4 +X5 +X6 Y4 = X1+ X2 Project Focussing (contd.)
  • 59. 6 59 Airplane Factory 3.4 What is our factory’s quality level? Dents 0 40 Airplane Wings Controls Fuselage Sheet Metal Wheels Altimeter Throttle Doors Windows Size Pressure Tread Cracks Calibration Grips Intermittent Air leaks Falls off Cracks Size 29 1080 40 3 40 1 16 280 5 480 40 1 40 2 160 0 160 0 40 3 40 0 40 5 40 1 80 2 80 0 40 1 40 0 40 2 320 8 120 7 120 3 160 1 80 Defects during plane assembly 1 120 0 0 5 40 2 40 0 0 0 0 7 280 40 2 1 # of opportunities Total # of defects Total opportunities Roll-up Example 3.4
  • 61. 6 61 Functions – (Individuals) a) What for your function exists?(Role) b) What is Quality in your function ?  due date, cycle time, errors etc. c) Who are your “customers”? d) What are your customers’ major complaints/expectations (Some times perennial complaints, other times sporadic). (If you have records on these pl. go & collect the same) e) Can we convert the complaints into measurable form? (pl. try) 15 minutes.
  • 62. 6 62 Products (Concerned Individuals) a) List out the end Products your external customer receives. - 10 minutes b) List out the complaints you have received on these Products (you may pl. go back to your work place and gather information from your records) - 30 minutes
  • 63. 6 63 Manufacturing Processes (Groups) a) List out the Manufacturing Processes - 10 minutes b) Make Process maps for all the above processes - 15 minutes c) Make it more detailed by walking through the Processes - 30 minutes d) Identify the steps which are yielding less (lot of rework and also scrap) (you may pl. go back and refer your records) - 30 minutes
  • 64. 6 64 Transactional Processes (Intra or Inter -Functional) (individuals or groups) a) Write the process map. - (10 minutes) b) Walk through the process and make the above process more detailed and realistic. (You may pl. go back to your places and sit with your people.) - (30minutes)
  • 65. 6 65 Typical Good Projects 1. Reduction of variation of Hfe of NPN transistors (BC547, BF 494) at low currents and operating currents. 2. To improve the wire bonding process in IC division. (focused on BEL 1895 device) 3. Defect reduction in C1Bellow Assembly (Vac interrupters) 4. Reduction in rejection of sealing flange 7051 04850102 of V1 tubes 5. Minimisation of Insertion and Return loss in Multiplexers. 6. CT reduction from Sanction to Award of contracts in Services. 7. Reduction of cycle time in Internal Mailing system. (Same day delivery) 8. Cycle time reduction from enquiry to quotation for spares in HF division.
  • 67. 6 67  We don’t know what we don’t know.  If we can’t express what we know in the form of numbers, we really don’t know much about it.  If we don’t know much about it, we can’t control it.  If we can’t control it, we are at the mercy of chance. Measurements Are the Foundation for Six Sigma 0 20 40 60 80 100 1st Qtr 2nd Qtr 3rd Qtr 4th Qtr East West North
  • 68. 6 68 Data: A collection of any number of related observations on one or more variables. Raw Data: Information before it is arranged or analyzed by statistical methods. Data Array: The arrangement of raw data by observations in either ascending or descending order. Data Point: A single observation from a data set. What Is Data?
  • 69. 6 69 Population: A collection of all the elements we are studying and about which we are trying to draw conclusions. Sample: A collection of some, but not all, of the elements of the population under study, used to describe the population. Representative Sample: A sample that contains the relevant characteristics of the population in the same proportions as they are included in that population. Advantages of sampling: a. costs less , b. takes less time , c. case of destructive testing, The wolf did not have to eat the whole ox to know the meat is tough. Population And Sample
  • 70. 6 70 Statistics is a collection of techniques useful for making decisions about a process or population based on an analysis of the information contained in a sample from that process or population. When summarizing large data it is useful to distribute data into small groups or classes. Frequency distribution shows the number of observations in the data set that fall into each distinct classes. Histogram is a graph of the observed frequencies that fall into the different classes. Statistics means never having to say you are certain ! What Is Statistics?
  • 71. 6 71 Discrete Data: (Attribute Data) 1. Data that take discrete values. 2. Always expressed in whole numbers. 3. Countable data. 4. Requires large sample size to characterize product or feature. e.g. No. of defects, good / bad , go-nogo etc. Probability distributions applicable are binomial , Poisson etc. Continuous Data: (Variable Data) 1. Data that has continuous values 2. Can be expressed easily in fractions or precise increments. 3. Measurable data. e.g. Resistance of coil , thickness of coating ,Voltage etc. Probability distributions applicable are normal, exponential etc. Data : Discrete and Continuous
  • 72. 6 72 A measure of location indicating the value of a typical middle point of a distribution. e.g. Mean , Median , Mode etc. 1. Mean: A central tendency measure representing the arithmetic average of a set of observations. 2. Median: If the data set is arranged in order of magnitude ,the middle value that divides the data set into halves is called median. Quartiles : Values that divide data into four equal parts ( Q1, Q2 and Q3). Percentiles : Values that divide data into 100 equal parts . 3. Mode: The value most often repeated in the data set. It is represented by the highest point in the distribution curve of a data set. Measures of Central Tendency
  • 73. 6 73 Dispersion: The spread or variability in a set of data. Measure of Dispersion: A measure describing how the observations in a data set or distribution are scattered or spread out. Measures the precision of a process. 1. Range: The distance between the highest and lowest values in a data set. 2. Deviation: The difference of a data point from the mean . The sum of deviations of a data set from their mean is zero. 3. Mean Absolute Deviation: Average of the absolute deviations from the mean . Measures of Dispersion
  • 74. 6 74 4. Sum of squares ( SS) : Sum of the squared deviations from the mean. 5. Variance: The average of the squared deviations from the mean. 6. Standard Deviation: The root mean square deviation. The positive square root of the variance . Widely used measure of dispersion. 7. Inter-quartile range (Q3-Q1) / 2 Measures of Dispersion
  • 75. 6 75 Parameters And Statistics File : wt-ss4.mtw. 1. Find mean, median and mode for weight data. 2. Find variance and Standard deviation also. 3. Check that sum of the deviations of all observations from its mean is zero. Measure Statistics Parameters (of sample) (of population) Mean x  Sum of squares ( x - x )2 ( x -  )2 Variance s2 = ( x- x )2 2 = ( x -  )2 n - 1 N Std.Dev. s = (x - x )2  = ( x -  )2 n - 1 N No. of observations n N
  • 76. 6 76 Symmetrical: A characteristic of a distribution in which each half is the mirror image of the other half. Skewness: The extent to which a distribution of data points is concentrated at one end or the other; the lack of symmetry. Measure of the degree of asymmetry. Positively skewed - curve that tails off toward the high end or right. Negatively skewed - curve that tails off toward the low end or left. Kurtosis: The degree of peaked ness of a distribution of points. Other Characteristics of a Distribution
  • 78. 6 78  = s = ( X i - X ) 2 S i = 1 n n - 1 ? If we have 4 numbers and you know the average and 3 of the numbers, you can derive the 4th (You don’t have that 4th DF) Degrees of freedom (n-1) is employed to derive an unbiased estimator of the population standard deviation. (n-1) is the number of independent contrasts out of n observations. Obs. value contrast 1 6 2 4 2 3 6 -2 4 1 5 The fourth contrast (6-1) has no independent existence because its value is known from adding the three contrasts (2-2+5) = 5. Thus the degrees of freedom for four observations is three. Degrees of Freedom
  • 79. 6 79 The Nature of Variation 5 4 3 2 1 5 4 3 2 1 Mean is Centered ( it is On Target), but there is Large Variation Precise but not Accurate Accurate but not Precise Mean is not Centered ( it is Off Target); the Variation is Small Accuracy refers to the closeness of the average of the Measurements to the target value. Precision refers to the closeness among the individual measurements
  • 80. 6 80 Understanding Process Variation 1.233 1.235 1.237 1.239 1.241 1.243 1.245 1.247 USL T LSL Recognize that the process width is independent of the design width. In other words, the inherent precision of a process is not determined by the design specifications. Centered... but large variation!
  • 81. 6 81 Understanding Process Centering USL T LSL Recognize that the process center () is independent of the design center (T). In other words, the ability of a process to repeat any given centering condition is independent of the design specifications. 1.233 1.235 1.239 1.241 1.243 1.245 1.247 1.237  Increase in nonconformance due to shift in process centering 5 4 3 2 1 Small variation... but not centered! Without deviating from the norm, progress is not possible!
  • 82. 6 82 What’s the probability of making it to the Limo before he walks into the lamp post ? - Parking lot is perfectly rectangular - The end of the limo just touches the fence at the end of the parking spot Probability The Blindfold Pub Wind direction, straight out of Northwest,10 MPH Region of better light He walks at 50 ft per minute. 30 ft 60 ft 10 ft 50 ft 100 ft Limo 15 ft x 6 ft, corner parking spot Perimeter=1ft
  • 83. 6 83 The Blindfold... Use this page if you wish to sketch out his path in order to determine the answer Pub Probability…Contd. Don’t use statistics as a drunken man uses lamp post , for support rather than illumination.
  • 84. 6 84 Understanding the Histogram A total of 1,000 parts were produced. 22 parts were greater than the USL. 31 parts were less than the LSL. The probability of defect is 53/1000 = .053 The process yield is 1 - .053 = .947, or 94.7% 300 1.247 Height of the bar corresponds to the number of times a measurement was observed within the given interval width (on the X axis). Total width of the histogram is an estimate of the process capability. This is the range over which the process operates most of the time. 0 50 100 150 200 250 1.233 1.235 1.237 1.239 1.241 1.243 1.245 USL LSL Probability…Contd. Process Capability Report
  • 85. 6 85 DPU and DPO = defect = 1 unit = 1 opportunity 1 unit = 4 opportunities DPU = 1 dpo = 0.25 DPU = 2 dpo = 0.5 DPU = 3 dpo = 0.75
  • 86. 6 86 Unit of Product 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 1 • 2 • 3 • • • 4 • • • • • 5 • 6 • 7 • • • 8 • • • 9 • 10 • 11 • • 12 • • • 13 • • • 14 • • 15 • • • • 16 • • • 17 • • 18 • 19 • • 20 • • • • 21 • 22 • 23 • 24 • • 25 • • 26 • 27 • • 28 • • • 29 30 • Class Exercise: PCB Production Each PCB Has 10 Opportunities for a Defect. We Have Produced a Lot of 60 PCBs and Inspected for Defects: DPU and DPO…Contd.. = Defect = Opportunity • =
  • 87. 6 87 Do The Following  How many total defects?  How many total units?  What’s the average Defects/Unit?  How many Total Opportunities/unit?  What’s the Defects/Total Opportunities?  Fill in the observed ... defects/unit observed 0 1 2 3 4 5 DPU and DPO…Contd..
  • 88. 6 88 Nature Of The Problem 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 1 • 2 • 3 • • • 4 • • • • • 5 • 6 • 7 • • • 8 • • • 9 • 10 • 11 • • 12 • • • 13 • • • 14 • • 15 • • • • 16 • • • 17 • • 18 • 19 • • 20 • • • • 21 • 22 • 23 • 24 • • 25 • • 26 • 27 • • 28 • • • 29 30 • 1) The defects are randomly distributed 2) 60 defects were observed out of 60 units produced 3) The defects-per-unit is 1.0 4) There are 10 opportunities for defect per unit of product Given the facts, what is the likelihood of producing a part with zero defects? In turn, this guarantees no rework or repair. DPU and DPO…Contd..
  • 89. 6 89 Given the facts, what is the likelihood of producing a part with zero defects? In turn, this guarantees no rework or repair. In other words.... What is the probability of 1 unit having all 10 of its opportunities as good ones. Note: If an opportunity has a 10% chance of being bad, then it has a 90% chance of being good. DPU and DPO…Contd..
  • 90. 6 90 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 1 • 2 • 3 • • • 4 • • • • • 5 • 6 • 7 • • • 8 • • • 9 • 10 • 11 • • 12 • • • 13 • • • 14 • • 15 • • • • 16 • • • 17 • • 18 • 19 • • 20 • • • • 21 • 22 • 23 • 24 • • 25 • • 26 • 27 • • 28 • • • 29 30 • = Defect = Opportunity = Unit of Product 1-.10 =.90 .10 Probability the Opportunity is Defective = Defects per Opportunity(DPO) .9010 = .34867844 Thus, the likelihood that any given unit of product will contain zero defects is 34.87% Defects: 60 Production: 60 Units PCB Part Probability the Opportunity is not Defective DPU and DPO…Contd..
  • 91. 6 91 A given process has two operations. Each operation has a first- time yield of 99 percent. The rolled-throughput yield equals: There is a 98% probability that any given unit of product could pass through both operations defect free. Extending The Concept Process Centered Op 1 99% Process Centered Op 2 99% 98% Output x = YRT = e-DPU YFT = Rolled-Throughput Yield Classical First-Time Yield s U Rolled-Throughput Yield S = good parts of the final step, U = total parts of final step
  • 92. 6 92 The hidden operation 37 Units are guaranteed to pass because they contain no defects. A total of 63 units contain one or more defects. Of these, some will be repaired while others will become scrap. In this case, 53 were repaired and 10 were scrap. 37 Units d = 0 dpu = 1 100 Units Submitted d = defects u = units The Hidden Operation Defect Number Type Observed A 30 B 10 C 0 D 20 E 40 Total 100 90 Units Passed (Final yield) Verify Not OK Scrap 10 Units Operation Rework Scrap Y .tp Throughput yield (Units with Zero defects) 63 Units d  1 53 Units Repaired
  • 93. 6 93 Rolled-Throughput Yield YIELD DECREASES WHEN COMPLEXITY INCREASES DIE BONDING WIRE BONDING PROCESS MAP 6  Process/Product 4  Process/Product 100 P(d) = 6.2 10-3 Yield = (1-p)400 Yield = 8,3 % 400 COMPLEXITY Process Steps or CTQ’s QUALITY P(d) = 3.4 10-6 Yield = (1-p)400 Yield = 99.8 % P(d) = 6.2 10-3 Yield = (1-p)100 Yield = 53 % P(d) = 3.4 10-6 Yield = (1-p)100 Yield = 99.96% 4  6  DICING MOULDING TESTING
  • 95. 6 95 Why Six Sigma? Yields thru Multiple Steps/Parts/Processes Zst (distribution shifted 1.5) # of parts, steps, or processes 3 4 5 6 1 93.32% 99.38% 99.9767% 99.99966% 5 70.77% 96.93% 99.88% 99.9983% 10 50.09% 93.96% 99.77% 99.997% 20 25.09% 88.29% 99.54% 99.993% 50 3.15% 73.24% 98.84% 99.983% 100 53.64% 97.70% 99.966% 200 28.77% 95.45% 99.932% 500 4.44% 89.02% 99.830% 1000 0.20% 79.24% 99.660% 2000 62.79% 99.322% 10000 9.76% 96.656% Yields thru Multiple Steps/Parts/Processes Zst (distribution shifted 1.5) # of parts, steps, or processes 3 4 5 6 1 93.32% 99.38% 99.9767% 99.99966% 5 70.77% 96.93% 99.88% 99.9983% 10 50.09% 93.96% 99.77% 99.997% 20 25.09% 88.29% 99.54% 99.993% 50 3.15% 73.24% 98.84% 99.983% 100 53.64% 97.70% 99.966% 200 28.77% 95.45% 99.932% 500 4.44% 89.02% 99.830% 1000 0.20% 79.24% 99.660% 2000 62.79% 99.322% 10000 9.76% 96.656% Reduce Parts/Steps Improve Sigma per Part/Step Rolled-Throughput Yield
  • 96. 6 96 Probability Models •Binomial Distribution •Poisson Distribution •Normal Distribution Three Models That Have Frequent Application for Products & Processes: Probability & Statistics: If you have knowledge of a Population, Probability is a tool you can use to determine how likely you are to draw a certain Sample If you have knowledge of a Sample, Statistics is a tool you can use to draw conclusions about an unknown Population
  • 97. 6 97 Normal Distribution  Distribution of a continuous random variable.  Gaussian distribution.  Fits into actual observed natural phenomena e.g. human characteristics, process outputs etc.  Unimodal and bell shaped.  Mean lies at the center of the curve & is the highest point.  Symmetrical about the mean.  Median and mode coincides with the mean.  The two tails extent indefinitely and thus never touch the horizontal axis.  Area under the curve = 1.  Only mean and sigma required to make predictions.  Shape is related to frequency distribution & histogram.
  • 98. 6 98 The Focus of Improvement Leverage variables which control the Mean Leverage variables which control the Standard Deviation Y = f ( X 1 , ... , X N )   Scale of Y Characterization Some Xs might affect the mean, some might affect the spread () some might affect both. Very Low Probability of Defects Very Low Probability of Defects LSL USL Excellent Process Capability Very High Probability of Defects Very High Probability of Defects LSL USL Poor Process Capability Capability Low Z High Z
  • 99. 6 99 Standard Deviation.  Point of Inflection 1 T USL p(defect) 3 The distance between the point of inflection and the mean constitutes the size of a standard deviation. If three such deviations can be fit between the target value and the specification limit, we would say the process has “three sigma capability.” Z = USL -  
  • 100. 6 100 -3 -2 -1 0 +1 +2 +3 Z 1 and -1 PROPERTIES OF SD 1. 68.27 % of the observations are included between + 3. 99.73 % of the observations are included between +3 2. 95.45 % of the observations are included between +2 and -2  and -3 Properties of Standard Deviation
  • 101. 6 101 Using Z as a Measure of Capability As variation decreases, the standard deviation (s) gets smaller and capability increases, which in turn decreases the probability of a defect. s z USL Xbar 3 Capability USL 6 Capability Xbar Z = SL - Xbar s Z = 6 Z = 3
  • 102. 6 102 What Is Probability Of a Defect when Z = 0.97? Area to the right of Z0 = 1.66x10-1 Probability = 0.166 or 16.6% Probability of a Defect Example = 0.166 Z0 Z Probability of a defect Normal Distribution Z 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.0 5.00E-01 4.96E-01 4.92E-01 4.88E-01 4.84E-01 4.80E-01 4.76E-01 4.72E-01 4.68E-01 4.64E-01 0.1 4.60E-01 4.56E-01 4.52E-01 4.48E-01 4.44E-01 4.40E-01 4.36E-01 4.33E-01 4.29E-01 4.25E-01 0.2 4.21E-01 4.17E-01 4.13E-01 4.09E-01 4.05E-01 4.01E-01 3.97E-01 3.94E-01 3.90E-01 3.86E-01 0.3 3.82E-01 3.78E-01 3.75E-01 3.71E-01 3.67E-01 3.63E-01 3.59E-01 3.56E-01 3.52E-01 3.48E-01 0.4 3.45E-01 3.41E-01 3.37E-01 3.34E-01 3.30E-01 3.26E-01 3.23E-01 3.19E-01 3.16E-01 3.12E-01 0.5 3.09E-01 3.05E-01 3.02E-01 2.98E-01 2.95E-01 2.91E-01 2.88E-01 2.84E-01 2.81E-01 2.78E-01 0.6 2.74E-01 2.71E-01 2.68E-01 2.64E-01 2.61E-01 2.58E-01 2.55E-01 2.51E-01 2.48E-01 2.45E-01 0.7 2.42E-01 2.39E-01 2.36E-01 2.33E-01 2.30E-01 2.27E-01 2.24E-01 2.21E-01 2.18E-01 2.15E-01 0.8 2.12E-01 2.09E-01 2.06E-01 2.03E-01 2.01E-01 1.98E-01 1.95E-01 1.92E-01 1.89E-01 1.87E-01 0.9 1.84E-01 1.81E-01 1.79E-01 1.76E-01 1.74E-01 1.71E-01 1.69E-01 1.66E-01 1.64E-01 1.61E-01 Z Area under curve is the probability of a defect Let us suppose that we calculate the standard normal deviate for a given performance limit and discover that Z = 0.97. The question becomes, “What portion of the total area under the normal curve lies beyond a Z value of 0.97?” Answering this question will give us the probability of producing a defect. Remember, the Z value is a measure of process capability and is often referred to as the “sigma of the process,” not to be confused with the process standard deviation. Performance Limit
  • 103. 6 103 For One-Sided Specs -- From Z table, for Z = 3.0, p(defect) = .00135 USL = 124 8  = 100 Z = |SL - |  |124 - 100| 8 = = 3.0 Z = |SL - |  Probability of a defect
  • 104. 6 104 ZUSL = |USL - |  |124 - 100| 8 = = 3.0 USL = 124 8  = 100 LSL = 76 ZLSL = |LSL - |  |76 - 100| 8 = = 3.0 p(d)US for ZUSL = 3.0 is .00135 from table p(d)LS for ZLSL = 3.0 is .00135 from table From table, for p(d) = .0027, Z = 2.78 Z of Centered process The file tra-pro .mtw gives the gain of the transistors taken at three different times.Analyze each process. TARGET = 100 ; USL = 124 ; LSL = 76
  • 105. 6 105 (mean)   (standard deviation) Probability of defects: PUSL USL LSL Target Probability of defects: PLSL |  -LSL |  ZLSL = PLSL Table | USL- |  ZUSL = PUSL Table PLSL +PUSL = PTotal Z Table Z calculation with 2 limits and process which is not centered . What if I have two specifications (upper and lower) and a process which is not centered? Z of shifted process
  • 106. 6 106 Mode Mean Median Standard Deviation Zcalculated = ( MEAN- LSL)/ P(d) =((# of scores < LSL)/# of scores) Ztable ( look up using the P(defect)) Class Exercise: Calculate the following for your assigned cricketer Note: LSL = 44 Ramesh Siva David Vinod 16 52 54 34 9 69 64 49 100 40 65 49 56 45 45 39 40 67 65 40 50 72 43 48 9 70 65 47 45 45 78 39 20 51 65 46 105 69 82 49
  • 107. 6 107 Is the process capability adequate to meet the Specifications ? Process potential = Tolerance = USL - LSL index - CP Pro. Cap. 6 = 1 just capable (3 sigma process ) > 1 capable process (Cp = 2 for a 6 sigma process) < 1 not capable of meeting the specifications. Process performance index (used when process is not centered). CPU = USL -  ; CPL =  - LSL ; 3  3  CPK = Lower of CPU , CPL . The inherent variability of a quality characteristic that the process is capable of maintaining , when in a state of statistical control. Process capability = 6 Std. Devs Process Capability
  • 108. 6 108 Process Capability Ratios  The greater the design margin, the lower the Total Defects Per Unit (TDU).  Design margin is measured by the Process Capability Index (Cp) (Maximum Allowable Range of Characteristic) (Normal Variation of Process) Cp = LSL Process Width Design Width USL T o +3 -3 ZST = 3CP Cp = USL - LSL ±3 
  • 109. 6 109 Process Capability Ratios Cpk = Cp (1 - k) Where k is the percentage of the tolerance zone consumed by the static mean shift k = T -  (USL - LSL)/2 3.4 ppm 0 ppm LSL USL T o 1 Cpk = Cp (1 - k) k = T -  (USL - LSL)/2 4.5 lt 6 st Example: Cp = 2 , k = .25 Cpk = 2 (1-0.25) = 1.5
  • 111. 6 111 Process Capability CP= 2 CP= 1 +61 - 61 USL LSL UCL +31 LCL - 31  T UCL +3 LCL - 3 +3 - 3 USL LSL  T
  • 113. 6 113 Process does not understand the design specification If you do not want to attack any of the above, forget ‘Quality Improvement’ Process Capability 1 For a given design specification select the process such that the ‘Process width’ is less than the ‘Design width’(For a 6 process the process width should be half of Design width) 2 If the ‘process width’ is more than the ‘Design width’, improve the process to reduce the ‘process width’, if the technology permits. 3 If the technology does not permit(poor technology) go for a better technology. 4 If both 2 & 3 are not possible, have a relook at the design to increase the ‘Design width’, if competition permits.
  • 114. 6 114 Why Assess Measurement System Measure Process/Product Knowledge Understanding Analysis Improvement You don’t know what you can’t measure !!
  • 115. 6 115 Is It Valid? Am I measuring what my customer thinks is critical to quality (CTQ)? Is It Reliable? Is the measurement system accurate, stable, repeatable, reproducible and linear? Do Specifications Exist? What is the order of magnitude of my measurement? What are the spec limits for the measurement? ? ? ? YES or NO? What must I do to show this? YES or NO? What must I do to show this? YES or NO? Are the specs valid? What must I do to show this? Three Questions To Ask About Measurements & Measurement Systems: Answering These Questions Is The Heart of MEASURE Phase Measurement System Analysis
  • 116. 6 116 Gage Repeatability is the variation in measurements obtained when one operator uses the same gage or measurement process for measuring the identical characteristics of the same parts or items. Possible Causes of Poor Repeatability: Equipment: • Gage instrument needs maintenance. • The gage needs to be more rigid. • The clamping of part needs improvement. People: • Environmental conditions (lighting, noise) • Physical conditions (eyesight) Repeatability Gage repeatability
  • 117. 6 117 Gage Reproducibility is the variation in the average of measurements made by different operators using the same gage or measurement process when measuring identical characteristics of the same parts or items. Possible causes of poor reproducibility  Measurement procedure is not clear  Operator is not properly trained in using and reading gage  For transactional applications - operational definitions not established Reproducibility Mean of the measurements of Operator B Mean of the measurements of Operator A Gage Reproducibility
  • 118. 6 118 Gage Accuracy also referred as Bias is the difference between the observed average of measurements and the true average. Establishing the true average is best determined by measuring with the most accurate measuring equipment or system available.-Viz Master gauge or Master Measuring Equipment How to Calculate (Continuous Data): • Obtain true average of Sample Parts • Compare with the operators observed values from the R & R study. • To convert accuracy to a % of tolerance, multiply difference of true average vs. observed by 100 and divide by tolerance. How to Calculate (Discrete Data): • Have an item with a known defect count inspected several times. Compare the observed defect count with the actual defect count. Possible causes of poor accuracy : • Gage not properly calibrated • Improper use of gage by operator • Unclear procedures • Human limitations True Average Accuracy Observed Average Gage Accuracy
  • 119. 6 119 Gage Stability Time 2 Time 1 Gage Stability refers to the difference in the average of at least two sets of measurements obtained with the same gage or measurement system on the same parts or items taken at different times. How to Calculate (Continuous Data): • For gages normally used for relatively long periods of time without calibration • Conduct a second Gage R&R study just prior to the time recalibration is due. • Gage stability is the difference between the grand averages of the measurements from the two studies. How to Calculate (Discrete Data): • Conduct a second Repeatability, Reproducibility and or Accuracy study. Look for trends over time. Possible causes of poor stability: Continuous data • Gage not being calibrated as frequently as needed • If air gage, may need filter or regulator. • If electronic gage, may need warm-up and stabilization. Discrete data • New operators • Software changes Gage Stability
  • 120. 6 120 Gage Linearity is the difference in the accuracy values through the expected operating range Within the same instrument. True Average Observed Average (Low End) Accuracy (Low End) Observed Average (High End) Accuracy (High End) True Average LINEARITY = | Accuracy (low) - Accuracy ( high) | How to Calculate: • Conduct Accuracy study through the expected operating range. At least two studies should be done, one at each end of range. • Subtract the smaller accuracy value from the larger to determine linearity. Possible causes of poor linearity: • Gage not calibrated properly at both lower and upper end of operating range. • Error in the minimum or maximum master. • Worn gage. • Internal gage characteristics. Gage Linearity
  • 121. 6 121 REPEATABILITY: GAGE: The instrument used for making measurements that we want to validate The % Variation Due to the Measurement Method GAGE R & R Does the same operator get the same results when measuring the same part several times? REPRODUCIBILITY: Do different operators get the same results when measuring the same part several times? Why Gage R&R Study-Contd..
  • 122. 6 122 Measurement Systems Analysis Match the definitions to the terms: Repeatability Reproducibility Accuracy Stability Linearity A. Difference in the average of two or more sets of data taken obtained with the same gage on the same parts taken at different times. B. Variation in average measurements made by different operators. C. Difference in accuracy values through expected operating range. D. Variation in measurements obtained by one operator measuring the same part with the same gage. E. Difference between the observed average and the true average. Review Questions
  • 123. 6 Gage R&R Gage Repeatability and Reproducibility 333
  • 124. 6 Upper Spec Limit Lower Spec Limit SCRAP ! SCRAP ! OBSERVED PRODUCT / PROCESS QUALITY LET’S MEASURE THIS PART SEVERAL TIMES... MEASUREMENT QUALITY LET’S PRODUCE AND MEASURE ! BEQI /M29EMD Feb2002 334
  • 125. 6 Upper Spec Limit Lower Spec Limit OBSERVED PRODUCT / PROCESS QUALITY MEASUREMENT QUALITY Gage Error Let’s reduce GAGE R&R variation ! BEQI /M29EMD Feb2002 335
  • 126. 6 Upper Spec Limit Lower Spec Limit OBSERVED PRODUCT / PROCESS QUALITY MEASUREMENT QUALITY Gage Error BEQI /M29EMD Feb2002 336
  • 127. 6 Upper Spec Limit Lower Spec Limit OBSERVED PRODUCT / PROCESS QUALITY MEASUREMENT QUALITY Gage Error LOOK! Observed Capability has improved ! BEQI /M29EMD Feb2002 337
  • 128. 6 OBSERVED PRODUCT / PROCESS QUALITY Upper Spec Limit Lower Spec Limit MEASUREMENT QUALITY Gage Error Lower the Gage Error, the better the Capability! This means... • fewer good parts rejected • fewer bad parts accepted BEQI /M29EMD Feb2002 338
  • 129. 6 2 WAYS TO LOOK AT MEASUREMENT SYSTEM VARIATION 1- Gage R&R as % Contribution 2- Gage R&R as % Tolerance BEQI /M29EMD Feb2002 339
  • 130. 6 1- Gage R&R as % Contribution BEQI /M29EMD Feb2002 340
  • 131. 6 m p Upper Spec Limit Lower Spec Limit OBSERVED PRODUCT / PROCESS QUALITY MEASUREMENT QUALITY GAGE R&R % Contribution = Gage Variance Observed Process Variance = (m)² (p)² BEQI /M29EMD Feb2002 341
  • 132. 6 Rules of Thumb R&R % Contribution < 2% GOOD ! 2 % < R&R % Contribution < 7.7 % RISK EVALUATION NEEDED ! R&R % Contribution > 7.7 % INACCEPTABLE ! GAGE R&R % CONTRIBUTION BEQI /M29EMD Feb2002 342
  • 133. 6 2- Gage R&R as % Tolerance BEQI /M29EMD Feb2002 343
  • 134. 6 Gage Error m Tolerance GAGE R&R % Tolerance Upper Spec Limit Lower Spec Limit OBSERVED PRODUCT / PROCESS QUALITY MEASUREMENT QUALITY = Gage Error Tolerance = 5.15 m USL - LSL BEQI /M29EMD Feb2002 344
  • 135. 6 Rules of Thumb R&R % Tolerance < 8% GOOD ! 8 % < R&R % Tolerance < 30% RISK EVALUATION NEEDED ! R&R % Tolerance > 30% INACCEPTABLE ! GAGE R&R % TOLERANCE BEQI /M29EMD Feb2002 345
  • 136. 6 WARNING ! BE PRECISE ON THE GAGE % YOU ANNOUNCE! GAGE R&R % Contribution? or GAGE R&R % Tolerance? Calibration Sticker does not imply a good Gage R&R! An accurate Gage does not guaranty a good Gage R&R! BEQI /M29EMD Feb2002 346
  • 137. 6 137 Tying it all Together % Contribution %Tol Typical Conclusions: Green Green Repeatability and Reproducability are acceptably small portions of the total observed variation…Proceed. Green Yellow or Red Implement any quick or obvious measurement system improvements but project work can begin with the current system. As the actual process variation improves, the measurement system may need improvement. Yellow Green, Yellow, or Red The variation in the measurement system is not as small as we would like to see. The system is marginal and will need to be evaluated for improvements before proceeding. Red Green, Yellow, or Red The variation in the measurement system is not an acceptably small portion of the total observed variation. Need to fix measurement before proceeding with process improvement. Here is a guide to interpret your Gage RR results:
  • 138. 6 138 • Can’t apply techniques of continuous gages. • The gage can be assessed based on known samples. • Attribute gage studies are performed by 2-3 people measuring 20 or more parts. • If all operators and known samples agree - receive a “Percent Effectiveness Score “ of 100%. • Use results to validate measurement or to prioritize improvement needs. How Do You Test the Reliability of Measurement Systems That Give only Pass/Fail Data? Gage R&R for Discrete Data.
  • 141. 6 141 Multi Vari Analysis (Graphical) Example : Electroplating of several parts in a bath. 1 2 3 4 5 6 7 8 9 10 0 10 20 30 40 50 Location Within a Part - Positional Over Time 1 2 3 4 5 6 7 8 9 0 10 20 30 40 50 60 70 80 90 Part Part Between Parts 1 2 3 0 10 20 30 40 50 60 70 80 90 Time Between Time Within 359
  • 142. 6 142 Multi Vari – An Example Variability of Epi-Thickness on the wafers 81% 8% 11% Position of wafer Within wafer Run to Run T R C L B Upper Middle Lower 360
  • 143. 6 143 Statistical analysis Why ? How ? Difference Difference • To find out the vital few factors • By finding out the “Statistically Significant difference” each factor (x) makes on the response (y) 361
  • 144. 6 144 Y = f (X1……XN) Independent variables Dependant variable 20% + 80% = 100% Trivial Many Variables Vital Few Variables 362
  • 145. 6 145 Statistical Analysis Tools  t-test  f- test  ANOVA  Chi-square ( 2) Test  Correlation & Regression  Multi vari Analysis 363
  • 147. 6 147 • How to Reduce extraneous variation in a process or product by i) PM Process Mapping ii) CE Cause & effect diagram iii) CNX Process input / output diagram iv) SOP Standard Operating Procedure Without Statistics!! Improve without Statistics PM / CE / CNX / SOP
  • 148. 6 148 Process Mapping Why? • Graphically outlines the sequence of a process • Provides visual foundation for current situation and analysis • Aids in identifying bottlenecks, redundancies and waste • Look for non value adding steps • Look for variables
  • 149. 6 149 How Process Mapping is to be done? 1. The start of the process and the end of the process should be well defined. 2. Observe closely the process steps and flow. 3. Involve all the personnel who are carrying out the process and take their opinion. 4. The process map should be as elaborate as possible; even minor details should not be missed. 5. The correct sequence of the process should be indicated in the process flow. How Process Mapping can help in cycle time reduction project ! ! 1. Look for time consuming, redundant, unnecessary movement & storage delays. 2. Look for major choke points that create significant delays. 3. Check for all rework steps which may be present. 4. Look for in-efficient layouts, sequences or flows. 5. Look for redundant material handling/packaging/unpacking. Analyze with following questions 1. What is the real purpose or function of the step 2. Does the step add value to the output? 3. Can the process step be eliminated, minimized or combined with another value adding step • Validate the process map by physically walking through the process. Process Mapping contd...
  • 150. 6 150 START PRINTING OF PRs AFTER MRP RECEIPT OF PRs IN PURCHASE DISTRIBUTION TO PURCHASE CELLS SORTING PRs REPEAT/VENDOR REQUEST FOR DRAWINGS PREPARATION OF REPEAT ORDERS PREPARATION OF ENQUIRIES POSTING / FAXING OF ENQUIRIES RECEIPT OF QUOTE PREPN. OF COMP STATEMENT CLARIFICATION FROM VENDOR / INDENTOR. NEGOTIATION PRICE REVIEW PREPARATION OF PURCHASE PROPOSALS COMPILATION OF DATA FOR NYRO. NYRO / SGRO. FAXING DATA TO NYRO. CONFIRMATION OF AVAIL FROM NYRO. VETTING OF PRs BY FINANCE REQUEST D&E FOR END USE. MAILING OF PR EU s TO NYRO. PRINTING OF EU s. RECEIPT OF PO FROM NYRO. ENTER PO DATA IN EDP DATABASE. POST FACTO APPROVAL FAX ADVANCE COPY TO VENDOR. END POST ORDER BY REGISTERED MAIL. SEND TO CENTRAL REGISTRY. PREPARE FOR POSTING. STAMP PO NO. AND DATE. ALLOTMENT OF PO NO. & ENTRY IN REGISTER. SIGNATURE ON PO BY MGR / DGM. VETTING BY FINANCE APPROVAL OF PROPOSAL AS PER DELEGATION OF POWER. ORDERING BY NYRO. NON NYRO Process Mapping contd... A typical Process Map PR to PO
  • 151. 6 151 Use of Process Mapping for Defect Identification Customer calls Confirm payment type Take order details Problem areas Deliver & collect money, if non-credit card customer Call Answered ? Credit card? Despatch order Take card details Enter customer details in the system No No Yes Yes Process Mapping contd...
  • 152. 6 152 Versions of a Process What you think it is …. What it really is…. What it should be…. What it could be…. Process Mapping contd...
  • 153. 6 153 Cause & Effect Diagram Output Y C N X Every variable on the diagram should be labeled as either: C = Constant N = Noise X = Controlled variable or factor Sources of Variation C C N X X People Material Machine Environment Method Measurement N X N X X C N N C C Make cause & effect diagram as detailed as possible
  • 154. 6 154 Process Response/ Outputs Constants or Fixed Variables C Y X Experimental Factors N Noise Variables Please Note: C – These are non- experimental factors influencing the response Y which should be addressed & made as a constant and controlled. N - These are non- experimental factors influencing the response Y very marginally and which are costly / difficult to address. X – These are factors greatly influencing the response Y which have to be experimented upon to optimize the response. CNX Diagram
  • 155. 6 155 S O P Standard Operating Procedures are rules that we define to ensure that we have consistent processes in everything we do. • Based on good judgment • Common sense • Engineering/Process knowledge Incorporate in documents like process sheets, instruction sheets, work instructions of ISO 9000, process standards etc. Make sure we have defined processes and that the rules are being obeyed by all, by constant monitoring.
  • 156. 6 156 Class Exercise #1 (Exercise from Air Academy) “Wiffle” “Solid” 1) Ensure Pin Position and Stop Angle Pin are at Position #3. Cup is at Position #1. Hook is at Position #4. 2) Use 1 Rubber Band. Pull the Arm to 177o and Shoot the Rubber Ball. 3) Disconnect Rubber Band between shots. 4) Have someone measure the Distance in cms. 5) Each person takes 5 shots in less than 60 sec. 6) Record Distances in spaces on next page. Form Your Team. Each Participant Must Shoot Catapult 5 Times Using the Following SOP: 4 Pin Position 1 Rubber Band Stop Angle 1 3 3 Start Angle 177 1
  • 157. 6 157 Catapult Data Before PM / CE / CNX / SOP Longest _______ Shortest _______ _________________ Range _______ Record What Have We Learnt? Enter values in Minitab- Calculate mean, range and standard deviation Save data in ‘A’ drive Target : 500 cms. Tolerance +/- 25 cms. Calculate Z Class Exercise # 1 Contd…. Shot #1 _____ _____ _____ _____ _____ _____ Shot #2 _____ _____ _____ _____ _____ _____ Shot #3 _____ _____ _____ _____ _____ _____ Shot #4 _____ _____ _____ _____ _____ _____ Shot #5 _____ _____ _____ _____ _____ _____ Distance cms Participant 1 2 3 4 5 6 Enter the data in file S1 Catapult-Project.mtw and save
  • 158. 6 158 Cause & Effect Diagram for Catapult With Control - Noise - Experimental Parameters Machine Men Environment Measurement Methods Material Distance Travelled by the ball Class Exercise # 1contd...
  • 159. 6 159 Optimum Launch Distance in cms Noise Variables (N) Constants( C ) Experimental Factors (X) Stability Clamp Skill of operator Wind Friction of arm Elasticity Start Angle Stop Angle Pin pos Type of tape Band length Ball weight & type Hook pos Cup pos Temperature Eyesight of inspector Angle marking Class Exercise # 1 contd... CNX Diagram for the Catapult Response (Y)
  • 160. 6 160 a) Clamp the catapult to the table b) Have a suitable block to fix the start angle at 177 degree. c) Use a better measurement system (by having a carbon impression at the first pitch of the ball). Class Exercise # 2 Repeat the exercise (same team, same catapult) by following the Improved SOP
  • 161. 6 161 Clamp catapult to table with clamps Set pin pos at 3 Set stop Angle pin at 3 Set cup pos at 1 Set Hook pos at 4 Install one rubber band Have a block to set arm at 177 degree Pull back the arm to 177 degree Place the rubber ball in the cup Spread carbon sheet in front of catapult Record the distance in record sheet Measure the distance traveled by ball Release the arm Repeat the process Class Exercise # 2 contd... Process Mapping Disconnect and reconnect the rubber band
  • 162. 6 162 4 Pin Position 1 Rubber Band Stop Angle 1 3 3 Start Angle 177 1 Class Exercise # II Contd...
  • 163. 6 163 Catapult Data After PM / CE / CNX / SOP Record Distance cms Enter in Minitab calculate mean,range & standard deviation. Save data in ‘A’ drive What Have We Learnt? Has the range reduced? Has the std dev reduced? Have the same specification as in exercise 1 & calculate sigma value Compare sigma values and draw the inference Class Exercise # II Contd... Shot #1 _____ _____ _____ _____ _____ _____ Shot #2 _____ _____ _____ _____ _____ _____ Shot #3 _____ _____ _____ _____ _____ _____ Shot #4 _____ _____ _____ _____ _____ _____ Shot #5 _____ _____ _____ _____ _____ _____ Participant 1 2 3 4 5 6
  • 164. 6 164 Catapult project progress Y = Traverse distance of ball 500+/-25 cms. Metric Current SOP Improved SOP Target Mean Std. Dev. Cp Cpk Z Histogram Run chart 500 4.16 2 6
  • 165. 6 165 Lessons Learnt Initial Condition Response USL USL Apply PM/CE/CNX/SOP 2nd Condition Response Now: Identify X’s that will Reduce the variance, Shift the mean What is this telling us?
  • 166. 6 166 What about the “Xs”?? 1. In both the previous exercises the experimental factors were fixed as follows: a. Start angle: 177° b. Stop angle pos.: 3 c. Pin pos.: 3 d. Hook pos.: 4 e. Cup pos.: 1 2. By experimenting on the above five factors we can fix them to get the exact distance required within the practical range. 3. This method of experimenting will be dealt with later in the ‘Improve’ using Statistics (DoE) stage.
  • 168. 6 168 Quality Improvement is Problem with Spread Accurate but not Precise Desired Current Situation LSL USL T Shrinking the variation & Problem with Centering Shifting the mean to the target d Precise but not Accurate USL LSL Current Situation Desired
  • 169. 6 169 How Do You Get a “Knowledge Equation”? What Kind of KNOWLEDGE We Need To Improve Our Products & Processes? Product or Process X2 Y X3 X4 X1 Y = Cos(X1/X2) Ln(X3) - X4 + 200 What Can You Do With This Equation? • Set X’s To Assure That Your Y Hits Target • Trade-Off X’s To Reduce Cost But Still Hit Target Y • Find Tolerances on X’s To Keep Y in Spec CTQ - Important to Your Customer
  • 171. 6 171 One-Factor-At-a-Time Experimentation (OFAT) Problem: Improve Gas Mileage Approach: Evaluate Gas Type and Timing Setting (2) G1 Gives Best MPG. Hold Gas Type G1 Constant (3) Evaluate MPG for Timing T1 and T2. Determine Best Value One-At-A-Time: (1) Set Timing at T1, Evaluate MPG of Gas Types G1 and G2 Gas Type Timing Set-Up MPG T1 T1 G1 G2 30 20 Gas Type Timing Set-Up MPG T2 T1 G1 G1 30 25 Best Value G1, T1
  • 172. 6 172 Design of Experiment Problem: Improve Gas Mileage Approach: Evaluate Gas Type and Timing Setting Gas Type & Timing Interact: Effect of One Factor Depends on the Level of The Other Factor Design of Experiment: Gas Type Timing Set-Up MPG T2 T1 G1 G1 30 25 G2 G2 T1 T2 20 45 Full Factorial Design: 50 40 30 20 T1 T2 G2 G1 MPG Missed This Point In One-At-A-Time
  • 173. 6 173 Strategy of Experimentation Plan: What is the Objective of your Experiment? – What questions do you expect to be answered when the experimentation & analysis is complete? What is Your Primary Response? – What Y & other Y’s should be measured? What are the Potential X’s? – What is the current capability of each Y after the implementing SOPs on Control and Noise type X’s? What Experimental Plans best fit your problem? Plan the Roles & Responsibilities of all involved in the DoE. Do: Assess the ease of implementing the Experimental Plan. Implement the Experimental Plan. Observe each experimental run. – Take notes & collect data on background Noise variables in addition to collecting data on your Y’s. Study: Analyze your data graphically and statistically. – Prepare graphs that illustrate the affect of each X on each Y. – Compare the outcome of the analysis to your initial theories. What did you learn? Confirm what you think you have learned. Act: Act on the results. – Have you made sufficient improvement given the new knowledge? – Plan additional experimentation to further your knowledge if necessary. – Communicate your findings and implement new SOPs where appropriate.
  • 175. 6 175 Statistical Process Control: Understanding Variation • Developed by Walter A. Shewhart in 1924. • Patterns data for statistical test, leading to product / process behavior information. • Facilitates underlying “Cause System” understanding. • A graphical representation of product and/or process performance. • Assignable (special) cause detection -- central tendency and/or variability impactors. • Serves as a probability based decision making tool. • Performance change detector. • Assess ability to predict performance based on sample data. • Points out actionable areas, with known degrees of risk and confidence.
  • 176. 6 176 Statistical Process Control Control charts: CC A plot of values of ‘location’ and ‘dispersion’ over time, used to identify assignable variations . Control limits:CL UCL and LCL on control charts within which all observations must fall for the process to be in control.  They refer to process control ,stability.  Inherent to the process. Specification limits: SL The designed USL and LSL for a process variable or a product. They refer to product acceptability. When the process has been brought in-control (using SPC ), quality can be further improved by redesigning the process to reduce the chance causes . This is called ‘Breakthrough ’.
  • 177. 6 177 Common cause (Chance cause)  variability within control limits  contributed by many factors  is of low magnitude  stable over time  part of the manufacturing system 2. Special cause (Assignable cause)  variability exceeds control limits  contributed by few factors  higher than acceptable levels  unusual occurrence  external to the system Using SPC we eliminate the Assignable causes of variation. Types of variability in a process.
  • 178. 6 178 1. Common cause (Chance cause)  affects all individual values  always present  random variation inherent in the process  not easily correctable  distribution of data is normal 2. Special cause (Assignable cause)  source of variation intermittent  assignable to some event  non-random pattern  easily correctable  distribution –skewed, shifted or abnormal Types of variability in a process…contd.
  • 179. 6 179 Two Types of Causes Common Causes: Those causes that are inherently part of the process (or system) hour after hour, day after day, and affect everyone working in the process. The variation that results from the consistent operation of the process as designed. Assignable Causes: Those causes that are not part of the process (or system) all of the time, or do not affect everyone, but arise because of specific circumstances. The variation that is the result of special causes in the operation of the process.
  • 181. 6 181 On Target, Minimum Variation Average = Target Performance Time
  • 182. 6 182 Average = Target Performance Time Average Is On Target… …But There Is More Variation
  • 184. 6 184 ? Average = Target Performance Time The area of usual process performance Indication of a special occurrence Assignable Cause
  • 185. 6 185 Run Chart Example Dr. Thomas W. Nolan came one day to talk to Dr. W. Edwards Deming and brought with him a chart that his son Patrick, age 11, had made for a school project. The chart showed, day by day, the time of arrival of his school bus. He recognized assignable causes of variation. Patrick had kept a record of the daily time of arrival of the school bus that came to carry him off to school, and had plotted the points in time order. He identified assignable causes of delay on two days. Think what a good start in life Patrick had, understanding common and assignable causes of variation - at age 11!. He had recognized, without calculation, assignable causes of delay on two days, and had shown his explanation of these delays. Time of arrival of school bus, by Patrick Nolan, 11 0800 0830 0900 Date (Source: Dr. W. Edwards Deming’s 4 day management seminar)
  • 186. 6 186 Four Possible States For A Process 1) Stable & Capable Average Lower Spec. Upper Spec.
  • 187. 6 187 2) Stable and Incapable Average Upper Spec. Lower Spec.
  • 188. 6 188 3) Unstable & Currently Capable Average Upper Spec. Lower Spec.
  • 189. 6 189 4) Unstable and Incapable Upper Spec. Lower Spec. Average
  • 190. 6 190 Summary Four Process Conditions IS THE PROCESS IN STATISTICAL CONTROL? IS THE PROCESS CAPABLE OF MEETING REQUIREMENTS? YES NO IDEAL STATE ACTION: ATTAIN CONTROL ACTION: IMPROVE CAPABILITY (TECHNOLOGY) ACTION: 1st: CONTROL THEN: CAPABILITY
  • 191. 6 191 The Control Chart The control chart is a process analysis technique. It is used to quickly detect the occurrence of assignable causes or process shifts so that a successful investigation of the process and proper corrective action may be taken. This proactive approach assists in keeping the process on target with minimum variation. Average Performance Time UCL LCL Control Charts show the distribution over time
  • 192. 6 192 Components of the Control Chart Center Line UCL LCL SAMPLE NUMBER Region of Nonrandom Variation Region of Nonrandom Variation Inherent Process + Measurement Variation Spread +3 -3 X-Bar To interpret these charts, we use the decision limits and look for identifiable patterns that indicate non-randomness in process behavior. UCL and LCL on control charts within which all observation must fall for the process to be in control •They refer to process control, stability •Inherent to the process
  • 193. 6 193 Establishing control charts  Collection of data on the selected characteristic  Select the appropriate chart  Choose the center line. ( targeted value or calculated from data)  Calculation of +/- 3S trial control limits  Plot the data points  Find assignable cause points ( points beyond the control limits)  Remove them and recalculate the control limits  ( now only chance causes are present )  Monitor process by plotting the instant values with the newly fixed control limits  If rule violation occurs, find the cause and take corrective action  Improve process and review control limits
  • 194. 6 194 Individual X and Moving Range Average & Range (X-bar and R) Average & StandardDeviation (X-bar and S) Exponentially Weighted Moving Average Type of Chart When do you need it? • When production is low volume or cycle time to build product is long, or shift and drift are a problem • When production is high volume, sample size is 3-8; allows process mean and variability to be viewed and assessed together • When production is high volume, sample size is >8; allows process mean and variability to be viewed and assessed together • When a small shift needs to be detected, or process is continuous and mixing can occur (e.g., in a chemical plant, wire drawing ) What Kind of Control Chart You Need? Variable Data? • When you want to know the fraction of defective units; sample size is variable and usually > 50 • When you want to know the number of defective units; sample size is constant and usually > 50 • When you want to know the number of defects; sample size is constant • When you want to know the number of defects per unit; sample size is variable p np c u Attribute Data?
  • 195. 6 195 Reading control charts For an unstable process – take action ? Identify rule violations indicating assignable causes of variation. Identify the assignable cause(s) as to source(s). Remove and prevent recurrence. Adjust/correct for assignable cause if it can not be removed and prevented. Action to adjust centering Action to reduce variability For a stable process - take no action.  Most of the plotted points occur near the center line  A few of the points occur near the control limits  Only an occasional rare point occurs beyond the control limits ( 2.7 out of 1000)  The plotted points occur in a random manner with no clustering, trending, or other departure from a random distribution.  Histogram is normal
  • 196. 6 196 Rule 1 A single data point above U.C.L. or below L.C.L. U.C.L. +2 S random +1 S random Target -1 S random -2 S random L.C.L. Rule 2 Nine consecutive data points all above the target, or nine consecutive data points all below the target. U.C.L. +2 S random +1 S random Target -1 S random -2 S random L.C.L. Rules to test presence of Assignable causes
  • 197. 6 197 Rule 3 Six consecutive data points declining in value, or six consecutive data points increasing in value. U.C.L. +2 S random +1 S random Target -1 S random -2 S random L.C.L. Rule 4 Intentionally not shown since it is not as critical. (14 points in a row, alternating up and down) Rule 5 Three consecutive data points of which two are at least +/- 2 S random from the target. U.C.L. +2 S random +1 S random Target -1 S random -2 S random L.C.L.
  • 198. 6 198 Rule 5 Five consecutive data points of which four are at least +/- 1 Srandom from the target. U.C.L. +2 Srandom +1 Srandom Target -1 Srandom -2 Srandom L.C.L. Rules to test presence of Assignable causes UCL = Upper Control Limit LCL = Lower Control limit Target = Target or Process Average S random = Estimated Standard deviation, often from subgroup ranges, representing inherent common cause variation.
  • 199. 6 199 Tables Table of Constants for X –R, and I-MR charts for sub- group size = n 2 1.880 1.128 0 3.267 A2 d2 D3 D4 3 1.023 1.693 0 2.574 4 0.729 2.059 0 2.282 5 0.577 2.326 0 2.114 6 0.483 2.534 0 2.004 7 0.419 2.704 0.076 1.924 8 0.373 2.847 0.136 1.864 9 0.337 2.970 0.184 1.816 10 0.308 3.078 0.223 1.777 E2 2.660 1.772 1.457 1.290 1.184 1.109 1.054 1.010 0.975 n
  • 200. 6 200 u u - 3 u/n u + 3 u/n u c c - 3 c c + 3 c c p p - 3 p(1-p )/n p + 3 p (1-p)/n p np np - 3 np (1 – p ) np + 3 np (1 - p) np x x - 2.66R x + 2.66R I or x R D3R D4R R or MR x x – A2R x + A2R x CL LCL UCL Chart        List of formulae for Control Lines x = Average of subgroup averages (x) total no. rejected total no. of defects total no. of defects total no. inspected total no. of subgroups total units inspected UCL – Upper Control Limit LCL – Lower Control Limit CL – Centre Line p = c = u = 
  • 201. 6 201 Make the items in a subgroup as alike as possible Make the subgroups themselves as different as possible The observations in a logical subgroup is free from assignable causes of variation. In production process the rational subgroup could be ‘manufactured at the same time’. In service area the rational subgroup could be ‘similar transactions by the same person’. Need to apply Rationale and Knowledge of the process to determine Subgroup size Production Sequence •••••••••••••••••••••••••••••••••••••• Sampling Windows n= 5 Production Unit Rational Subgroups Instant time method (widely used) gives minimum variation within sub groups maximum variation between sub groups
  • 202. 6 202 Attribute Charts • You want to track the causal variable on a transactional project where data is discrete. Common Choices: • p chart - To track fraction defective • u chart - To track defects per unit • Or skip this and go back to the Individuals and Moving Range control charts for variable data
  • 203. 6 203 A claim of process capability without process consistency (no assignable causes) illustrated on a control chart is not a valid claim!
  • 204. 6 204 Some Misconceptions About Control Charts Although control charts have been used for many years in a variety of situations, there are a number of misconceptions concerning their uses. The following misconceptions are summarized from a presentation by Michael Flynn (Flynn, 1983):
  • 205. 6 205 1. Control charts are tools for production workers to tell them when to adjust their processes. Control charts are tools for understanding variation. An operator reacting to an out of control situation is one of many possible uses of control charts, but certainly not the most important in many organizations. ( Associates in Process Improvement)
  • 206. 6 206 2. Control charts are only for the production or manufacturing operations. Control charts should be used to understand variation in all of the important processes in an organization. These include employee relations, safety, accounting, planning, lead times, maintenance, engineering, research, customer service, quality control, finance, complaints, environmental monitoring data. ( Modified from Associates in Process Improvement)
  • 207. 6 207 3. Control limits are boundaries beyond which we do not want to go. Control limits have nothing to do with what we want. The limits just define the boundaries for common causes of variation. Often we want a process to go out of control, if for example, the change results in higher yield or fewer errors on purchase orders. ( Associates in Process Improvement)
  • 208. 6 208 4. Control limits are boundaries within which the process can vary by chance. A better statement would be that control limits are boundaries on the process within which sample results can vary due to common causes when the process does not change at all. ( Associates in Process Improvement)
  • 209. 6 209 5. The process can go back and forth -- in control, out of control, and then back in control The calculated statistic for different subgroups will vary. If a special cause results in a shift in the process, the subgroup points will still vary, but now some may be within the control limits and some outside. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 1516 17 18 19 20 ( Associates in Process Improvement) S t a t i s t i c
  • 210. 6 210 6. Control charts can only be used to track processes over time. Defining subgroups by time periods is the most common way to develop control charts. But there are many other possibilities such as by clerk, customer, city, supplier, spindle, filler head, material lot, instrument number, and so on. Control charts are appropriate for all of these groupings of data. ( Modified from Associates in Process Improvement)
  • 211. 6 211 The control limits are calculated using the same method every time “Tight” control limits indicate that the common cause variation in the process is relatively small. ( Associates in Process Improvement) 7. It is harder to hold narrow control limits than wide ones.
  • 212. 6 212 8. Two-sigma control limits will result in “tighter ” control than the traditional three-sigma limits. Using limits other than Shewhart’s three-sigma control limits will likely result in higher costs due to over and under reaction to assignable causes. For a stable process, reacting to all points outside of a two-sigma limit will result in increased variation in the output of the process. ( Associates in Process Improvement)
  • 213. 6 213 9. Assignable causes are always an indication of a problem or of poorer quality, and Shewart called special causes assignable causes, i.e. the variation could be “assigned ” to a particular cause. Variation in the right direction can certainly be good. ( Associates in Process Improvement) 10. It is not necessary to investigate assignable causes that result in better quality
  • 214. 6 214 Poka-yoke = Mistake Proofing From the Japanese words: yokeru - To Avoid poka - Inadvertent Errors / Unintentional Error  The idea is to respect the intelligence of workers by taking over repetitive tasks that depend on vigilance or memory.  Promotes Creativity and Value-Added Activities  Reduces Ego problem among employees. This is based on the principle that the defects can be prevented by controlling the performance of a process so that it cannot produce the defects even when a mistake is committed by humans.
  • 215. 6 215 Common Types of Human Error 1) Forgetfulness (Not Concentrating) 2) Errors Due to Misunderstanding (Jump to Conclusions) 3) Errors in Identification (View Incorrectly...Too Far Away) 4) Errors Made by Untrained Workers 5) Willful Errors (Ignore Rules) 6) Inadvertent Errors (Distraction, Fatigue) 7) Errors Due to Slowness (Delay in Judgment) 8) Errors Due to Lack of Standards (Written & Visual) 9) Surprise Errors (Machine Not Capable, Malfunctions) 10) Intentional Errors (Sabotage - Least Common)
  • 216. 6 216 Red Flag A condition in the process which commonly provokes errors Red Flag Error Defect
  • 217. 6 217 1) Frequent Adjustments 2) Constant Equipment Changes 3) Dimensionality / Specification / Critical Condition 4) Many Parts / Mixed Parts 5) Multiple Steps 6) Lack of, OR Ineffective, Standards Some Red Flag Conditions Are: 7) Rapid Repetition 8) Volume a. Sudden change b. Quantity push vs Quality 9) Environmental Conditions: a. Material / Process Handling b. Housekeeping c. Foreign Matter d. Poor Lighting e. Other 10) Other?
  • 218. 6 218 Common Mistake Proofing Devices Using wisdom and ingenuity to create devices that allow you to do your job 100% defect free 100% of the time. HOME • Automated Shut-Offs on Electric Coffee Pots • Ground Fault Circuit Breakers for Bathroom or Outside Electric Circuits • Child-Proof Caps on Medications • Butane Lighters with Safety Buttons Retail • Tamper-Proof Packaging • Bar Coding at Checkout OFFICE • Spell Check in Word Processing Software • Questioning “Do you want to delete?” After Depressing the “Delete” Button on Your Computer FACTORY • Dual Palm Buttons and Other Guards on Machinery • Bar Coding
  • 219. 6 219 Common Mistake Proofing Devices…Contd. Must be taller than the line to go on ride Amusement park rides use this technique to avoid mistake of having young children stand in line for rides they are not able to enjoy When converting from leaded to unleaded gas, the tank openings and nozzles were smaller to eliminate potential of putting wrong gas in tank
  • 220. 6 220 Common Mistake Proofing Devices…Contd. On Airplane Lavatory Doors, in order to turn on the lights, the passenger must lock the door which automatically activates the OCCUPIED sign. LOCK OCCUPIED
  • 221. 6 221 Rely on getting information to where it can be used to prevent a mistake. Common Mistake Proofing Devices…Contd.
  • 222. 6 222 Common Mistake Proofing Devices…Contd. Poka Yoke / Mistake Proofing - Processing Errors - Before Improvement: It was possible to insert the chassis in the jig backwards. Correct operation depended on the workers vigilance. After Improvement: A guide pin was added, keyed to an asymmetrical feature of the chassis. This completely eliminates the danger of backwards processing. Description of Process: A chassis was placed in a jig for machining Problem: Chassis set backwards in jig Solution: Additional guide pin taking advantage of asymmetry Key Improvement: Jig modified to guarantee correct positioning Guide Pin additional guide pin
  • 223. 6 223 Poka Yoke / Mistake Proofing - Missing Parts - Before Improvement: Detection of missing screw holes depended on the vigilance of the operators at further processes along the line. However, defects got through to final assembly often and the tuners could not be mounted in the TVs. After Improvement: Rods for detecting the presence of the screw holes were mounted on the jig for inspecting the tuner assemblies. The tuner cannot be set in position for quality inspection unless the screw holes are in the proper place. Defective tuners are now detected before being sent on. Description of Process: Mounting brackets are added to the chassis of the TV tuner for later attachment to the rest of the TV assembly. Problem: Sometimes the brackets are missing from the chassis. Solution: Modify functional test test mounting to incorporate bracket check Key Improvement: Jig modified to identify defective parts inspection jig Common Mistake Proofing Devices…Contd.
  • 224. 6 224 Common Mistake Proofing Devices…Contd. Poka Yoke / Mistake Proofing - Omitted Processing - Before Improvement: The operator counted the holes as they were drilled. However, the operator sometimes made errors, and the products with the wrong number of holes were produced. After Improvement: A counter was mounted on the drill press to detect each hole as it is drilled. Along with this, a limit switch was mounted on the jig to detect when a part was removed before the proper number of holes was drilled. Description of Process: A number of holes are drilled in each workpiece Problem: Incorrect number of holes drilled Solution: An automatic counter to keep track of the number of holes drilled Key Improvement: Tool modified to guarantee correct processing count limit switch for detecting workpieces workpiece counter correct defective buzzer jig
  • 225. 6 225 Advantages of Poka-Yoke • Simple – No formal training programs required. • Inexpensive – No high cost investment required. • Gives Prompt Feedback. • Eliminates many inspection operations. • Reduces operator dependence. • Promotes Creativity and Value-Adding Activities. • Results in Defect-Free Work. • Requires immediate action when problems arise. • Provides 100% inspection internal to the operation.
  • 226. 6 226 GOOD Detects error before it continues to the next operation BETTER Allows for detection while error is being made. BEST Makes it impossible for errors to occur. SUPERIOR Makes it impossible for errors to occur with productivity gain. Levels of Mistake-Proof Processes
  • 227. 6 227 368 “The Right Quality and Uniformity are Foundations of Commerce,Prosperity and Peace” Deming
  • 229. 6 229 Defining Lean Lean is: “A systematic approach to identifying and eliminating waste (non-value-added activities) through continuous improvement by flowing the product at the pull of the customer in pursuit of perfection.” — The MEP Lean Network
  • 230. 6 230 Reduced Lead Time “One of the most noteworthy accomplishments in keeping the price of Ford products low is the gradual shortening of the production cycle. The longer an article is in the process of manufacture and the more it is moved about, the greater is its ultimate cost.” — Henry Ford, 1926
  • 231. 6 231 Definition of Value-Added Value-Added Any activity that increases the market form or function of the product or service. (These are things the customer is willing to pay for.) Non-Value-Added Any activity that does not add market form or function or is not necessary. (These activities should be eliminated, simplified, reduced, or integrated.)
  • 232. 6 232 Lean = Eliminating Waste Typically 95% of all lead time is non-value-added. Value-Added Non-Value-Added • Overproduction • Waiting • Transportation • Non-value-added processing • Excess inventory • Defects • Excess motion • Underutilized people
  • 234. 6 234 Overproduction • Making more than is required by the next process • Making earlier than is required by the next process • Making faster than is required by the next process • Causes of overproduction: – Just-in-case logic – Misuse of automation – Long process setup – Unlevel scheduling – Unbalanced workload – Over engineered – Redundant inspections
  • 235. 6 235 Inventory Waste • Any supply in excess of a one-piece flow through your manufacturing process • Causes of excess inventory: – Need for buffer against inefficiencies and unexpected problems – Product complexity – Unleveled scheduling – Poor market forecast – Unbalanced workload – Misunderstood communications – Unreliable shipments by suppliers
  • 236. 6 236 Defects • Inspection and repair of material in inventory • Causes of defects: – Weak process control – Poor quality – Unbalanced inventory level – Deficient planned maintenance – Inadequate education, training, or work instructions – Product design – Customer needs not understood
  • 237. 6 237 Processing Waste • Effort that adds no value to the product or service from the customers’ viewpoint • Causes of processing waste: – Product changes without process changes – Just-in-case logic – True customer requirements not clearly defined – Over-processing to accommodate downtime – Lack of communication – Redundant approvals – Extra copies or excessive information
  • 238. 6 238 Waiting Waste • Idle time created when waiting for…? • Causes of waiting waste: – Unbalanced workload – Unplanned maintenance – Long process setup times – Misuses of automation – Upstream quality problems – Unlevel scheduling
  • 239. 6 239 People Waste • The waste of not using people’s mental, creative, and physical abilities • Causes of people waste: – Old guard thinking, politics, the business culture – Poor hiring practices – Low or no investment in training – Low pay, high turnover strategy
  • 240. 6 240 Motion Waste • Any movement of people or machines that does not add value to the product or service • Causes of motion waste: – Poor people or machine effectiveness – Inconsistent work methods – Unfavorable facility or cell layout – Poor workplace organization and housekeeping – Extra “busy” movements while waiting
  • 241. 6 241 Waste of Transportation • Transporting parts and materials around the plant • Causes of transportation waste: – Poor plant layout – Poor understanding of the process flow for production – Large batch sizes, long lead times, and large storage areas
  • 242. 6 242 Elements of a 5S Program Sort — Perform “Sort Through and Sort Out,” by placing a red tag on all unneeded items and moving them to a temporary holding area. Within a predetermined time the red tag items are disposed, sold, moved or given away. “When in doubt, throw it out!” Set in Order — Identify the best location for remaining items, relocate out of place items, set inventory limits, and install temporary location indicators. Shine — Clean everything, inside and out. Continue to inspect items by cleaning them and to prevent dirt, grime, and contamination from occurring. Standardize — Create the rules for maintaining and controlling the first three S’s and use visual controls. Sustain — Ensure adherence to the 5S standards through communication, training, and self-discipline.
  • 243. 6 243 Visual Controls Simple signals that provide an immediate understanding of a situation or condition. They are efficient, self-regulating, and worker-managed. Examples: • Kanban cards • Color-coded dies, tools, pallets • Lines on the floor to delineate storage areas, walkways, work areas, etc. • Andon lights
  • 244. 6 244 Plant Layout Brake Screw Machine Raw Stock QC Rec. Ship Shear QC Assembly Parts Stock Stamp Mill Lathe Drill Finish Weld Grind
  • 245. 6 245 CSSBB (Certified Six Sigma Black Belt) of ASQ(American Society for Quality) www.asq.org • Associate member fee : $74 per year • Student member fee : $25 per year • Exam fee : $270 • Exams twice a year • Bengaluru is Exam center • Two completed projects with signed affidavits • QCFI conducts 60 hrs. of coaching classes in ten consecutive Sundays- Fee Rs.20000 only ‘Develop your Career and Grow your Organization’
  • 247. 6 Science Art Magic Six Sigma The Six Sigma Methodology can help us move from Magic and Art to Science
  • 248. 6 248 Current State Desired State It’s Great to Have a Philosophy. . . But We Need a Strategy !! Don’t Change Your Strategy in the Middle of the Maze
  • 249. 6 249 Strategic Business Policy “Commitment to achieve excellence in all our business processes, through the application of ‘Six Sigma’ break through methodology, by the year 2004” Key Performance Indicators (KPI) • Customer Satisfaction • Cost reduction • Cycle time improvement • Yield improvement • Design improvement
  • 250. 6 250 What is the need? When is the need felt? Where is the need felt? Why is the need felt? How is the situation handled now? AC should be silent Sound sleep At night In the bedroom To remain fresh next morning Uses a ceiling fan that makes a lot of noise AC should be efficient Good cooling At night In the bedroom It gets very hot in May - June Uses a ceiling fan that is not so effective in summer AC Should not cost much Affordability N / A N / A Limited finance N / A What customermeant Household member 1 VOC Table Sl. No. Who is the Customer? What customersaid (Voice of Customer) Example of Air Conditioner Customer CTQ’s
  • 251. 6 Some Plain Talk About Six Sigma World Class Quality Mikel J. Harry, Ph.D. Chief Executive Officer Six Sigma Academy, Inc. Phoenix, Arizona Today, focusing on the customer is absolutely essential. Of course, we all recognize this. But do we really internalize the idea? Do we really believe that such a focus has the potential to drive business growth and impact the level of prosperity which we should come to expect? Closely linked to the idea of customer satisfaction is the concept of operational excellence -- the kingpin of success. Without a focus on excellence, it becomes easy to accept the position of second or third best. Being the best means embracing change and reaching out for new and higher standards of performance. Only then can one break the chains of complacency and pave the way for breakthrough. The attainment of excellence is no longer a lofty goal or ideal, it is now a fundamental requirement -- the ante for entering the game of business. As most of us already know, Japan has assumed this perspective and steadily acted upon it. The result of their focus has been staggering, as evidenced by their superb products and services, not to mention their tremendous position in the marketplace. Hence, smart money says we must view the idea of operational excellence as a major force in today's marketplace. For those who might doubt such an assertion, just ask a customer what they think. After all, what the customer thinks provides us with a strong indicator of what will be purchased -- and from whom. Of course, it is widely recognized that the operational performance of an organization is largely determined by the capability of its processes. Another way of looking at this would be to say that our performance as a company is governed by the quality of our processes -- high quality processes delivers high quality products, at the lowest possible cost, on time. Therefore, a focus on operational excellence (in everything we do) translates to a focus on process quality. Of course, we can not focus on what we do not measure and if we do not measure, we can not improve. Trying to improve something when you don't have a standard to measure against, is like playing a sports game without knowing the score. Can you imagine setting out on cross-country trip in an automobile without a fuel gauge? Just think of the personal grief, cost, and inconvenience which might result. Would you allow yourself to be placed in such a situation?
  • 252. 6 As we will discover, the measurement and improvement of our processes is absolutely essential if we are to achieve operational excellence and the ideals of total quality. To do this, we must bear in mind the old axiom -- let the product do the talking. In other words, the quality of our products can tell us how capable our processes really are. To measure product quality is to measure process quality because the two are correlated. It is very important to recognize that the inverse of this also holds true. Perhaps we should now set the stage for our ensuing discussion by examining what some have called the "chain of causation:" Our survival is dependent upon growing the business. Our business growth is largely determined by customer satisfaction. Customer Satisfaction is governed by quality, price, and delivery. Quality, price, and delivery is controlled by process capability. Our process capability is greatly limited by variation. Process variation leads to an increase in defects, cost, and cycle time. To eliminate variation, we must apply the right knowledge. In order to apply the right knowledge, we must first acquire it. To acquire new knowledge means that we must have the will to survive. If you can't express something in the form of numbers you don't really know much about it. And if you don't know much about it, you can't control it. And if you can't control it, you're at the mercy of chance. And if you're at the mercy of chance, why bother with it? Hence, we must learn the language of numbers. Such thinking represents a business philosophy -- a way of guiding our company. From this perspective, we will focus our discussion on an all embracing standard and methodology called "Six Sigma." As we shall see, Six Sigma can be used to measure the quality of our work processes -- in any field, from assembling a motor car to inspiring a classroom full of students. Over the years, this writer has been asked a great many questions about Six Sigma -- most of which have been quite simple, practical, and straight forward. For the sake of simplicity and reading ease, we shall complement such questions with simple, practical, and straight forward answers. Of course, we recognize that more sophisticated and detailed answers do exist, as many students of Six Sigma will testify. However, there is no point in hooking-up a fire hose when all we want is a glass of water. QUESTION: What is Six Sigma? ANSWER: Six Sigma is several things. First, it is a statistical measurement. It tells us how good our products, services, and processes really are. The Six Sigma method allows us to draw comparisons to other similar or dissimilar products, services, and processes. In this manner, we can see how far ahead or behind we are. Most importantly, we can see where we need to go and what we must do to get there. In other words, Six Sigma helps us to establish our course and gauge our pace in the race for total customer satisfaction. Some Plain Talk About Six Sigma…Contd.