“Six-Sigma” Quality Programs

Principles
• Most products and services have attributes or
characteristics that are “critical” to perceived quality.
• For each of these characteristics, our customer have
certain expectations about acceptable levels (or
tolerance limits) of performance. Output that does not
conform to these levels is considered defective.
• The goal of six sigma is to significantly reduce the
chance of producing outputs that don’t meet our
customer’s expectations for acceptable performance.
 A six-sigma variance
reduction program
• Reduces the likelihood of defects occurring
• Reduces the need for testing and inspection
• Reduces work in process inventories
• Shortens cycle time
• Increases customer satisfaction
With process output centered between tolerance
limits, located +/- 3 standard deviations (3 )
from process mean, there is a 0.27% chance
output will be defective.
Process Output Distribution

Chance output Chance output

is below LTL = 0.135% exceeds UTL = 0.135%

Value of Critical
Quality Characteristic

Lower +3 Upper

Tolerance Tolerance
Level (LTL) Level (UTL)
 Defects
per
million
opportun
-ities

Source: Harry, M. (1998). Six Sigma: A Breakthrough Strategy for Profitability.

Quality Progress 31(5), 60-64.
Implications of variance reduction on defects
produced

tolerance limits Left tail Right tail Total (PPM)

3 0.1349967% 0.1349967% 0.2699934% 2699.93445
4 0.0031686% 0.0031686% 0.0063372% 63.37207
5 0.0000287% 0.0000287% 0.0000574% 0.57421
6 0.0000001% 0.0000001% 0.0000002% 0.00198
Suppose we keep the same tolerance levels but cut the
process standard deviation in half, so the distance from
the mean to the tolerance limits is now +/- 6 standard
deviations (6 ). The chances of defective output falls to
0.0000002% (or 0.002 parts per million).

Process Output Distribution

Chance output Chance output

is below LTL = 0.0000001% exceeds UTL = 0.0000001%

Lower +6 Upper

Tolerance Tolerance
Level (LTL) Level (UTL)
 Six Sigma Road Map
(DMAIC Methodology)
Define: Identify critical to quality (CTQ) product
characteristics
Measure: Map related processes, establish tolerances
for the CTQ attributes, and establish
baseline performance.
Analyze: Identify process variables that affect
outcomes and the root causes for defects.
Improve: Establish procedures to reduce defects
Control: Sustain the improvement efforts over time.
 Six-Sigma Tool Kit
A Japanese Samurai carried seven tools into battle. After World
War II the Japanese discovered quality as a philosophy for
economic recovery. It is natural that they sought seven tools with
which to begin their economic rejuvenation. According to Ishikawa
(1982) the seven tools chosen were:

Check Sheets Run Charts

Pareto Diagrams Histograms
Cause and Effect Control Charts
Diagrams Scatter Graphs
CHECK SHEETS (or tally sheets)
 Simple tool to record frequency of an event: Use it to
gather data to begin to detect patterns. Often this is
the first step in most problem solving cycles.
 HINTS FOR USE:
 Decide what events to record
 Decide on a time period
 Develop a form that is easy to use

MJII – p. 31
 A Check Sheet
Number of Customer Complaints.

Problem Month

1 2 3 Total

Total
 Flow Charts
• To understand problems in a process, you must
first understand the process.
• The easiest way to understand a process is to
draw a picture of it.
• Also useful for standardizing procedures and
training new employees.
• Neat software: Visio, Edge Diagrammer
MJII- p.56
STANDARD PROCESS FLOW
CHART SYMBOLS

OPERATION: an action that transforms the workpiece

INSPECTION: an assessment of the output from an operation

TRANSPORT OR MOVE: often includes distance

STORAGE

DELAY
D
 FLOW CHART ENHANCEMENTS

• OPERATION RATES, or the number of units that

can be processed per unit time -- useful for
identifying “bottlenecks.”
• FAIL POINTS, or steps in the process where
errors that lead to dissatisfaction can occur.
• LINES OF VISIBILITY, or the parts of the
process that can be observed by a customer.
 PARETO CHARTS
 A vertical bar graph that helps us prioritize
problems
 Portrays attribute data (frequency of each type
of defect or problem) gathered from check
sheets.

MJII p.95
 Problem Number of
(most recent Customer
Percent of Total
6 months) Complaints
Print Quality 23 0.38%
Excess 13 0.21%
Flash
Type Error 6 0.10%
Late 6 0.10%
Shipment
Misplace 4 0.07%
Print
Bad Count 4 0.07%
Rough Edge 3 0.05%
Missing Ring 2 0.03%
Total
 CREATING A PARETO CHART
WITH EXCEL
 Select data containing category labels, frequency, and
cumulative percent
 Press chart wizard button
 Choose "custom types" tab
 Select Line-Column on 2 Axes
 Presto!!
 CAUSE & EFFECT
DIAGRAMS
 This tool helps you portray the various causes affecting a
process. Begin by stating the problem or "EFFECT" that
concerns you.
 Commonly there are many causes for the effect of interest.
These often fall into four different categories: Methods,
Materials, Machinery, and People.
 Use brainstorming techniques to help identify possible causes.
Use check sheets to confirm your intuition. Classify the major
causes into categories.
 For each cause, ask "Why does this happen?" Add ribs for each
explanation. To find the root cause, look for common or
repeated causes.

MJII – p. 23
Cause and Effect Diagram

Fill in the details for what makes a good cup of coffee

 HISTOGRAMS
HISTOGRAMS portray
the distribution of MJII – p. 66
sample data. Cure Time Experiment Histogram
We use them to visualize
the variation in the 20
18
data. 16
14

Frequency
We can use them to 12
10
detect irregular 8
6
patterns of variation 4
2
that suggest a 0 24.0000
26.2248
28.4497
30.6745
32.8993
35.1242
37.3490
39.5738
41.7987
44.0235
46.2483
48.4732
More
systematic cause.
Review your memory jogger for
tips on how to construct Cure Time
histograms (number of
classes and class width)
 Graphically shows the following:

1.center (i.e., the location) of the data;

2.spread (i.e., the scale) of the data;
3.skewness of the data;
4.presence of outliers; and
5.presence of multiple modes in the data.
Limitations
1. lack of information about the state of control of the process.
2. time-dependent or time-related trends are not captured.
EDA with histograms
1.How well is the histogram centered? The centering of
the data provides information on the process aim
about some mean or nominal value.

2.How wide is the histogram? Looking at histogram width

defines the variability of the process about the aim.

3.What is the shape of the histogram? Remember that the

data is expected to form a normal or bell-shaped curve.
Any significant change or anomaly usually
indicates that there is something going on in the process
which is causing the quality problem.
Truncated
Appears to end sharply or abruptly at one
end. Indicates possible sorting or
inspection of non-conforming parts.

SAW-TOOTHED
Also commonly referred to as a comb
distribution, appears as an alternating jagged
pattern. Often indicates a measuring
problem -- gage not sensitive enough for
readings.

BI-MODAL
Distribution appears to have two peaks.
May indicate that data from more than
process are mixed together materials may
come from two separate vendors samples
may have come from two separate
machines.
 SCATTER GRAPHS
These charts portray the
relationship between Scatter Graph, Platen Pressure vs
two variables. Use Defects/Batch
them to explore 9
hypotheses you 8

Defects/Batch
7
generate about 6
causes and effects. 5
4
If a relationship 3
2
shows some 1
0
correlation (not
0 10 20 30 40 50 60 70
necessarily linear -- Platen Pressure
don't forget quadratic,
exponential, Here it looks like platen pressure and defects/batch show a high degree of
correlation. Hmm …
logarithmic, etc)
check it out
mathematically. MJII p. 145
 Scatter (X,Y) Graphs
A’s points are close to the line; B’s are scattered

A – Strong Correlation B – Low Correlation