6 SIGMA: THE SEARCH FOR AND CONTROL OF X

1
6 SIGMA THE SEARCH FOR AND CONTROL OF X

• Michael W. Piczak
• Dipl.T., B.Comm., MBA
• University of Buffalo
• May 2003

2
WHO IS USING IT?

• MOTOROLA
• GENERAL ELECTRIC
• TEXAS INSTRUMENTS
• SONY
• LOCKHEED-MARTIN
• RAYTHEON
• CAMCO
• BOART LONGYEAR
• ALLIED SIGNAL
• CRANE VALVES

3
GOALS OF 6 SIGMA

• Defect reduction
• Yield improvement
• Improved customer satisfaction
• Higher net income
• Cost reduction
• Process improvement and optimization
• Bottleneck relief

4
2 ROADS TO PROFITABILITY
5
FIXING BELIEF

• Method of tenacity
• Method of authority
• Method of reasoning
• Method of science

6
6 SIGMA IN GOLFING PERSPECTIVE

• 6 sigma quality levels amount to 3.4 defects per
  million or 99.999966 good.
• 2 sigma missing 6 putts per round
• 3 sigma missing 1 putt per round
• 4 sigma missing 1 putt every 9 rounds
• 5 sigma missing 1 putt every 2.33 years
• 6 sigma missing 1 putt every 163 years
• based on 100 rounds per year

7
PROBLEM SOLVING TOOLS AND MODELS

• MICHAEL W. PICZAK
• DIPL.T., B.COMM., MBA
• MARCH 2002

8
Traditional Problem Solving Models
9
Piczak, 1995
10
NARROW VS. BROAD CONCEPTUALIZATION

• AS QUALITY PROGRAM
• OR
• STAND ALONE OPERATIONS RESEARCH METHODOLOGY

11
AS QUALITY PROGRAM
12
Motorolas 6 Sigma Program
13
AS STAND ALONE OPERATIONS RESEARCH METHODOLOGY
14
WHO IS THE ENEMY?

• VARIATION

15
WHAT ARE WE REACHING FOR?
16
PRIMARY SOURCES OF VARIATION

• Inadequate design margin
• Unstable parts and material
• Insufficient process capability

17
WHERE TO FOCUS?

• For each product or process critical to quality
  (CTQ)
• Measure
• Analyze
• Improve
• Control

18
GENESIS OF 6 SIGMA
19
ELEMENT 1
20
PORTERS 5 FORCES MODEL
21
PEST MODEL
22
VOICE OF THE CUSTOMER

• 2 Brands of customers
• internal
• external

23
Voice of the customer
24
DESCRIBE THE PROCESS
25
IMPROVING THE PROCESS

• Elimination
• Simplification
• Combination
• Reuse
• Parallel processing
• Subcontracting

26
CRITICAL EXAMINATION
27
NO NEW PROBLEMS PLEASE

• Poka Yoke techniques
• guide pins
• templates
• limit switches
• limited computer screen fields
• checklists
• interconnects

28
GETTING BETTER?

• The need to measure in quantitative terms
  important
• ISO/QS9000 demands it in terms of quality and
  effectiveness
• customer satisfaction
• quality levels ( non-conformances, dpu, dpmo)
• cycle times
• die change times

29
ELEMENT 2 MEASUREMENT
30
OLD METRICS

• Measures of central tendency or typicality
  (mean, median, mode)
• Measures of dispersion (range, variance,
  standard deviation)

31
THE NORMAL DISTRIBUTION
32
A KEY FORMULA
33
NORMAL CURVE CHARACTERISTICS

• Continuous
• Symmetrical
• Tails asymptotic to zero
• Bell shaped
• Mean median mode
• Total area under curve 1

34
VISUALIZING VARIATION
35
(No Transcript)
36
PROCESS CAPABILITY
37
PROCESS CAPABILITY II
38
Cpk
39
CAPABILITY ST LT
40
Cp LONG TERM (LT)
41
ST to LT
42
NEW METRICS

• dpu
• dmpo
• A CAVEAT
• ABOUT DPMO

43
JURANS COSTS OF QUALITY
44
ELEMENT 3 QUALITY INITIATIVES
45
SDWTs
From Quality Progress, October 1996
46
LITERATURE IDENTIFIED BENEFITS

• Productivity ? 15 -250
• All employees can perform all tasks
• Costs ? 30
• Cycle time ? 50-90
• Inventory ? 66
• Rework due to engineering flaws ? 50

47
BENEFITS CONTD

• Late jobs ? 1000
• Quality ?
• Recurring defective product problems ? 10
• Return on investment/sales ?

48
BENEFITS CONTD

• Sales ? 830
• Operating statistics improved by 25-40
• Accounts receivable ? from 66 days to 51 days
• Corporate overhead ? from 100M to 24M
• Accidents ? 72

49
SHORT CYCLE MFG.

• SMED
• automated computerized inspection
• X and moving range control charts
• automated systems (MAPs/CAD/CAM/flexible mfg.,
  etc.)
• flexible, self directed work force

50
DFM

• group technology
• accessibility of different parts areas
• ease of workpiece handling
• ergonomic principles
• safety requirements
• appearance
• QFD

51
BENCHMARKING

• more than just organized tourism
• more than just a nice walk over at a friends
  plant
• not industrial espionage
• not a one way channel of communication
• inside and outside of your industry

52
THE ALCOA SEQUENCE
53
SPC

• using numbers to describe absence or presence of
  a phenomenon
• systematic gathering of data
• using a collection of analytics that promote
  common understanding and profound knowledge
  (Deming)
• emphasis is on measurement

54
STATISTICS

• Collecting
• Organizing
• Summarizing
• Analyzing
• Presenting

55
THE ANALYSTS DUTY

• start with a regularity, uniformity or curiosity
• identify all previously significant predictors
  of phemon in question
• theorize as to why independent variables (Xs)
  should be predictive of dependent variables (Y)
• tease out the key Xs that impact Y and control
  same

56
ELEMENT 4 IMPROVEMENT TOOLS
57
Voice of the customer
58
Flowchart For Classic Problem Solving
Is It Working?
NO
YES
Dont Mess With It!
Did You Mess With It?
YES
YOU IDIOT!
NO
Will it Blow Up In Your Hands?
Anyone Else Knows?
Youre F_at_D!
YES
YES
Can You Blame Someone Else?
NO
NO
NO
Hide It
Look The Other Way
Yes
NO PROBLEM!
59
PARETO ANALYSIS
60
HISTOGRAMS
61
CAUSE EFFECT DIAGRAMS CARDS
62
SYSTEMATIC CAUSES OF VARIATION

• Lack of preventative maintenance
• Worn tools
• Operator performance
• Environmental changes
• Sorting practices
• Materials
• Measurement System

63
RECOGNITION OF STATEMENT OF PROBLEM
CHOICE OF FACTORS (Xis), LEVELS, RANGES
SELECTION OF RESPONSE VARIABLE (Y)
CHOICE OF EXPERIMENTAL DESIGN
PERFORMING EXPERIMENT
STATISTICAL ANALYSIS OF DATA
CONCLUSIONS, RECOMMENDATIONS, NEXT STEPS
64
NASA DATA REGRESSION LINE
65
REGRESSION ANALYSIS
66
ALWAYS LOOK AT THE DATA FIRST
67
AS STAND ALONE OPERATIONS RESEARCH METHODOLOGY
68
THE HUNT FOR X
69
6 SIGMA OLD WINE OUT OF NEW BOTTLES?
70
 
DUPONTS PPI
71
3 KINDS OF STATISTICS

• Descriptive
• Inferential
• Predictive

72
(No Transcript)
73
GEs MAIC
74
(No Transcript)
75
JURANs 6 SIGMA PROGRAM
76
Xs, Ys, and Zs of DOE
77
RECOGNITION OF STATEMENT OF PROBLEM
CHOICE OF FACTORS (Xis), LEVELS, RANGES
SELECTION OF RESPONSE VARIABLE (Y)
CHOICE OF EXPERIMENTAL DESIGN
PERFORMING EXPERIMENT
STATISTICAL ANALYSIS OF DATA
CONCLUSIONS, RECOMMENDATIONS, NEXT STEPS
CLASSIC DOE
78
(No Transcript)
79
(No Transcript)
80
6 SIGMA TRAINING TERMINAL PERFORMANCE OBJECTIVES

• As a result of taking this program, the
  participant will be able to
• Appreciate the scope of 6 Sigma practices in
  context of other company initiatives
• Apply a variety of tools to solve problems

81
T.P.O.s CONTINUED...

• Participate as a contributing member of a
  continuous improvement or problem solving team
• Use Minitab as a data analysis tool

82
PEDAGOGICAL APPROACH

• Lecture
• Discussion, debate and argument
• Videos
• Hands-on exercises using general and company
  specific examples

83
A BASIC RESEARCH PARADIGM

• Select the operation to be analyzed
• Explicate key variables a priori
• Gather data on key Xs and Ys
• Enter data and edit same
• Verify data integrity via Look, Counts,
  Describe, Scatter Diagram
• Run Descriptives
• Generate other graphs charts of data
• GRRs, regressions, ANOVAs, DOEs,

84
THE JOURNEY

• Most companies presently at 3-4 sigma
• The move is toward 6 sigma (Cp 2)
• Literature has references to 12 sigma (Cp ?)
• Tinnerman-Pallnut is at 50 ppb on selected
  product lines

85
6 SIGMA THE SEARCH FOR AND CONTROL OF X

• Michael W. Piczak
• Dipl.T., B.Comm., MBA