Statistical

Process Control

Quality & Productivity Society of Pakistan

Contents

 Quality & TQM

 Basic Statistics

 Seven QC Tools

 Control Charts

 Process Capability Analysis

CUSTOMERS
“Anyone who thinks customers
are not important should try
doing without them for a
week”

Source : Unknown
Types of Customers

 External Customers
 Final Customers/End-Users
 Internal Customers
Types of Customers

“The next operation as customer”

- Kaoru Ishikawa
Types of Customers

Exercise

 External Customers - 3 main customers

(describe type)
 Internal Customers - 3 main customers
Sources of Variation in
Production Processes
What is Quality ?
Quality

Fitness for Use

(Juran 1988)
Quality in goods
 Performance
 Features
 Durability
 Reliability
 Conformance
 Serviceability
 Aesthetics
 Perceived Quality
Quality in Services
 Tangibles
 Reliability
 Responsiveness
 Competence
 Courtesy
 Security
 Access; Communication &Understanding the
Customer
 What Is Quality?
The Experts Say...

 Conformance to requirements (Philip B. Crosby)

 Zero Defects (Philip B. Crosby)
 Fitness for use (Joseph M. Juran)
Reduced variation (W. Edwards Deming)
Quality Control Evolution

Evolution

TQM
Quality
Assurance
Quality
Control

Foreman
Operator
1900 1918 1920 1940 1980
Total Quality
Management
Total Quality Management

Achieve customer satisfaction through

continually improving all work process and
participation of employees.
 Total Quality Management
Elements
 Leadership
 Employee Involvement

 Product/Process Excellence

 Customer Focus

5-8
Major Contributors to the
development of TQM
 Dr Edwards Deming
 Dr Joseph Juran
 Philip Crosby
 Armand Feigenbaum
 Prof. Kaori Ishikawa
 Genichi Taguchi
 Musaaki Imai
 It’s not the tip
of the iceberg
It’s what you
that’s the problem….can’t see…
Variation results in cost

Waste
Rejects Customer Returns

2-3% Testing Costs Inspection Costs

Rework Recalls

20-40%
Complaint Customer
Handling (invisible costs) Allowances
Excessive Field
Service Costs Unused Incorrectly
Capacity Completed Sales
Excessive Lost goodwill Order
Overtime Planning
Pricing or Overdue Delays Time with Delays
Billing Errors Receivables Employee Expediti Dissatisfied Excess Inventory
Turnover ng Costs Customer
Development Unmeasured Late Incorrect
Cost of Failed Productivity Paperwor Orders
Products k Shipped
Basic Statistics
Population
Any well-defined group of individuals
whose characteristics are to be studied.
Students of a college
Books in Library
Shirts in Market
Fishes in Lake
Sample

Part of the population which is to be

studied.
Variable

Characteristics of the individuals of a

population or sample which varies from
individual to individual.

Marks obtained by Student

Height of Students
Temperature of Person
Dimensions of Product
Statistics

Statistics are numericals in any field of

study.

Statistics deals with techniques or

methods for collecting, analysing and
drawing conclusions from data.
ACCURATE & PRECISE
VERY CLOSE TOGETHER (LOW
VARIATION) AND CENTERED ON
TARGET (TRUE VALUE)

THE GOAL OF ANY PROCESS

PRECISE AND ACCURATE
 Target first,
variation then.
1 2

3
Variation first,
target then.
1

2
Which pilot do you want to fly
with?

A-1

A-4

B-4
B-3
B-2
B-1

A-2

A-3
Quality Engineering Terminology

Specifications
Quality characteristics being measured are
often compared to standards or
specifications.
 Nominal or target value

 Upper Specification Limit (USL)

 Lower Specification Limit (LSL)

Quality Engineering Terminology

 When a component or product does not

meet specifications, they are considered to
be nonconforming.
 A nonconforming product is considered
defective if it has one or more defects.
 Defects are nonconformities that may
seriously affect the safe or effective use of
the product.
 Types of Data
Variables "Things we measure"
• Length • Height • Diameter
• Weight • Volume • Tensile Strength
• Time • Temperature • Strength of Solution

Attributes “Things we count”

• Number or percent of defective items in a lot.
• Number of defects per item.
• Types of defects.
• Value assigned to defects
(minor=1, major=5, critical=10)

5
 Averages
 Mean, median and mode

Weekly rent paid by 15 students sharing accommodation, 1998 (£)

45 35 51 45 49
51 40 42 46 36
37 42 47 49 42

 Mean (or average)

add observations and divide by number of

observations 657/15 = 43.8

x= ∑x
n
Averages… 2
 Median – the middle observation.
 Arrange the observations and find the middle
one (n+1)/2th observation
 35 36 37 40 42 42 42 45 45 46 47 49 49 51 51

the 8th observation (15+1)/2 is 45

 Mode – the most frequent observation

in this case 42
MEASURES OF DISPERSION

The dispersion is defined as the scatter or

spread of the values from one another
or from some common value.
MEASURES OF DISPERSION

68.26%

95.46%

99.73%

-3σ -2σ -1σ µ 1σ 2σ 3σ

MEASURES OF DISPERSION

ALTERNATIVE TO CENTRAL TENDENCY

 RANGE (R): HIGHEST – LOWEST [Max – Min]

 VARIANCE: How data is spread out, about the mean.

___

s2 =
∑ (x − x ) 2

n −1
 STANDARD DEVIATION: Positive Square Root of
Variance.
___

S=
∑ (x − x ) 2

n −1
 Spreads
 Standard Deviation (SD) calculated as below
 calculate residuals – individual observation minus
mean square and sum these
 divide by number of observations minus 1 [gives
Variance]
 take square root for Standard Deviation
 2
 ∑( i
Y − Y )
SD = 
 n −1



 example peoples heights (cm)

 190 185 182 208 186 187 189 179 183 191 179

 mean 187.18

 SD 8.02
STATISTICAL PROCESS CONTROL

The statistical process control allows the

analysis of the current trend of the
production, in order to detect possible
deviations from the desired target,
independently on the deviation of the
single object.
 WHAT IS SPC ?

It is important to note that the

SPC is not the cure for Quality
and Production problems. SPC
will only help leading to the
discovery of problems and
identifying the type and degree
of corrective action required.
CONTROL LOOP

INPUT OUTPUT
PROCESS

MEASUREMENT
ADJUST

DECIDE ON
FIX? STATISTICS

ID GAPS

EXAMINE
Selection of improvement steps
(1) Select a theme

(2) Grasp current situation

Grasp “status to be
(3) attained”

(4) Analyze causes

(5) Propose solution

Implement solutions and

(6) evaluate results

(7) Follow-up & Standardize

(8) Review
Seven QC Tools
QC tools
QC tools (7 QC Tools, New 7 QC Tools) used in
solving (or improving) various types of problems
that occur in workshops.
Whether in identifying causes of problems or in
working out their countermeasures, effective use of
QC techniques can produce good results quickly
and efficiently.
It is important to get used to the use of
7 QC Tools. You are encouraged to
collect actual data and practice using
them.
Use of QC tools
Fact
In QC-style problem-solving
activity facts are grasped
based on data and analyzed
Collect data . Check sheet scientifically. Judgments are
made based on facts to take
concrete actions

Process data . Use of QC tools In a situation where several

factors exert influence in a
complex manner, QC tools
. Adding skills and
are indispensable to
Judgeme
experience correctly grasp cause-and-
nt
effect relationships in order
to arrive at objective
Countermeasures and
actions judgments
 Benefits of using QC tools

Cat egor i es
Check sheet
Gr aphs
Par et o di agr am
Scat t er di agr am
Hi st ogr am
Cont r ol char t
Cause and ef f ect di agr am
Af f i ni t y char t
Li nkage char t
Syst em di agr am
Mat r i x di agr am
PDPC
Ar r ow di r agr am
Fl ow char t s
Br ai n st or mi ng
Br ai n wr i t i ng
Tool

STEPS

Sel ect a t heme ○ ○ ○

○ ○ ○ ○ ○ ○ ○
Shape a vi si on ○ ○ ○ ○ ○ ○
Assess t he si t uat i on ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
Anal yze causes ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
Devi se sol ut i ons ○ ○ ○ ○ ○ ○ ○ ○ ○
I mpl ement and eval uat e r esul t○s ○ ○ ○ ○ ○ ○ ○ ○ ○
Fol l ow- up ○ ○ ○ ○ ○ ○ ○
Revi ew ○ ○ ○ ○ ○ ○
Benefits of using QC tools

2. The situation can be grasped correctly, rather than

based on experience or intuition
3. Objective judgment can be made
4. The overall picture can be grasped
5. Problem points and shortcomings become clear
so that action can be taken
6. Problems can be shared
Problem solving and QC tools
Select a theme
- Define focus areas - Look at the control situation - Look at trends and habits -Process capability
150
100 30
100 133.0
140
80
80 132.0 130 25
60 131.0
60 120
20
40 40 130.0 110

20 20 129.0 100 15

0 0 128.0 90
10
配線 ミス

誤組付 け

加工部品不良

組 付 け破 損

圧着不良

購入部品不良

そ の他

127.0 80
70 5
126.0
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 60
1月 2月 3月 4月 5月 6月 0
9.78 9.83 9.88 9.93 9.98 10.03 10.08 10.13 10.18 10.23

Pareto diagram Control chart Line chart Histogram

Get hold of a vision Get hold of the current situation Analyze the factors
- 3 factors of targets - Activity Cause/result relationship, Take data
plan 　
● ：号機
1 ▲ ：号機
2
7月14日 7月15日 7月16日 7月17日 7月18日 7月19日 1号機 2号機
130 スケジュール表 ●▲ ●● ●● ●● ●●▲ ●● ●●●
合計

縫製 14 2 16
項目 担当者 ４月 ５月 ６月 ７月 ８月 ９月 １０月 １１月 １２月 １月 ２月 ３月
120 テーマ選定 福島 仕上がり
●●▲ ● ▲ ●● ● ●● ●●●
11 2 13

●●▲ ●●●●● ●●● ●▲ ● ●●●●●●●●●●

110 あるべき姿の把握 秋田 汚れ
●● ●
25 2 27

●●●▲ ●▲ ●● ●▲ ●● ●● ●▲ ●●●
現状の把握 青森 キズ 19 6 25
100 ●▲ ● ●▲ ●

▲ ●● ●● ●● ▲ ● ●● ▲ ●●
原因の解析 福岡 その他 11 3 14

90 解決策の選定 山口
合計 26 11 11 11 13 23 80 15 95

80

70
効果の把握

フォローアップ
静岡

三重 cause and effect diagram Check

目
sheet
レビュー 愛知

標
View at things in layers
60
：計画 ：実績
1月
2月
3月
4月
5月
6月
7月
8月
9月

1月
2月
3月
10月
11月
12月

Brain writing Affinity chart Line chart Gantt chart Confirm interrelations

Follow-up and review Confirm the effect Look at changes over time
- What, how much and until what time? 130 133.0
120 132.0
● ：号機
1 ▲ ：号機
2
7月14日 7月15日 7月16日 7月17日 7月18日 7月19日 1号機 2号機 合計 133.0 131.0
Solutions
110
●▲ ●● ●● ●● ●●▲ ●● ●●●
縫製 14 2 16 132.0
100 130.0
●●▲ ● ▲ ●● ● ●● ●●● 131.0
仕上がり 11 2 13
130.0
129.0
●●▲ ●●●●● ●●● ●▲ ● ●●●●●●●●●●
90
is effective

汚れ 27
25 2
128.0
Measures

129.0

proposal
●● ●

●●●▲ ●▲●● ●▲ ●● ●● ●▲ ●●● 80

128.0
キズ
●▲ ● ●▲ ●
19 6 25 127.0
▲ ●● ●● ●● ▲● ●● ▲ ●●
127.0 70
その他 11 3 14
126.0
126.0
合計 26 11 11 11 13 23 80 15 95 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 60 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
3月

8月

2月
1月
2月

4月
5月
6月
7月

9月

1月

3月
月
月
月
11
10

12

Line chart Control chart (for

Check sheet Control chart
analysis)
Seven QC Tools
 Stratification Basic processing performed when collecting data

 Pareto Diagram To identify the current status and issues

 Cause and Effect Diagram To identify the cause and effect relationship
 Histogram To see the distribution of data
 Scatter Diagram To identify the relationship between two things
 Check Sheet To record data collection
 Control Chart To find anomalies and identify the current status
 Graph / Flow Charts To find anomalies and identify the current status
 New Seven QC Tools

 Affinity
Chart Grasp current situation and problems
 Linkage Chart Sort out relationships in the situation
 System chart Systematic sorting of the situation
 Matrix diagram Grasp a relationship between two matters
 PDPC methods Risk management based on forecasting
 Arrow diagram Plan progress
 Matrix data analysis Correlation analysis
Stratification
Stratification means to “divide the whole
into smaller portions according to certain
criteria.” In case of quality control,
stratification generally means to divide data
into several groups according to common
factors or tendencies (e.g., type of defect
and cause of defect).
Dividing into groups “fosters understanding
of a situation.” This represents the basic
principle of quality control.
Example usage
Item Method of Stratification
Hour, a.m., p.m., immediately after start of work,
Elapse of time
shift, daytime, nighttime, day, week, month

Variations among Worker, age, male, female, years of experience,

workers shift, team, newly employed, experienced worker

Processing method, work method, working

Variations among
conditions (temperature, pressure, and speed),
work methods
temperature
Variations among Measurement tool, person performing
measurement/ measurement, method of measurement, inspector,
inspection methods sampling, place of inspection
Pareto Diagram

A Pareto diagram is a combination of bar and line graphs of

accumulated data, where data associated with a problem (e.g.,
a defect found, mechanical failure, or a
complaint from a customer) are divided （件） （％）
１６０ １００

into smaller groups by cause or by ｎ＝１６０

１２０
phenomenon and sorted, for example,
by the number of occurrences or the ８０ ５０

amount of money involved. (The name ４０

“Pareto” came from an Italian ０ ０

mathematician who created the diagram.) Ａ Ｂ Ｃ Ｄ Ｅ

When is it used and what results will be obtained?

Which is the most serious problem among many problems? It is mainly

used to prioritize action.

Usage Results
•Allows clarification of
•Used to identify a problem.
•Used to identify the cause of a important tasks.
problem. •Allows identification of a
•Used to review the effects of an starting point (which task
action to be taken. to start with).
•Used to prioritize actions. •Allows projection of the
effects of a measure to be
[Used during phases to monitor
the situation, analyze causes, and taken.
review effectiveness of an action.]
Example usage of Pareto
Diagram
(1) Assessment using (2) Confirmation of Effect
Pareto diagram (Comparison)
(prioritization)
Frequently used to
•To identify a course of check the effect of an
action to be emphasized improvement.
using a variety of data.
Details of A
Improv
ed!

B C D I J K L
A
W X Y Z X Y W Z
Ｘ Ｙ Ｚ Ｘ Ｙ Ｗ Ｚ
Cause and Effect Diagram
A cause and effect diagram is “a fish-bone diagram
that presents a systematic representation of the
relationship between the effect (result) and affecting
factors (causes).
Solving a problem in a scientific manner requires
clarification of a cause and effect relationship, where
the effect (e.g., the result of work) varies according
to factors (e.g., facilities and machines used, method
of work, workers, and materials and parts used). To
obtain a good work result, we must identify the
effects of various factors and develop measures to
improve the result accordingly.
 Cause and Effect Diagram
Name of big bone factor
mini bone small bone

characteristics
big bone

medium

(result)
bone back bone

factors (causes)
When is it used and what results will be obtained?

A cause and effect diagram is mainly used to study the cause of a

certain matter. As mentioned above, the use of a cause and effect
diagram allows clarification of a causal relation for efficient problem-
solving. It is also effective in assessing measures developed and can be
applied to other fields according to your needs.

Usage Results
•Used when clarifying a cause and •Can obtain a clear overall
effect relationship. picture of causal relation. (A
•[Used during a phase to change in the cause triggers
analyzecauses.] a variation in the result.)
•Used to develop •Can clarify the cause and
effect relationship.
•countermeasures.
•
•[Used during a phase to plan Can list up all causes to
countermeasures.] identify important causes.
•Can determine the direction of
action (countermeasure).
Histogram
Articles produced with the
same conditions may vary in
terms of quality specification range
characteristics.
A histogram is used to

Y axis (no. of occurrences)

judge whether such variations
are normal or abnormal.
First, the range of data
variations are divided into
several sections with a given
interval, and the number of
data in each section is
counted to produce a
range of variation
frequency table. Graphical X axis
representation of this table is (measured values)

a histogram.
 When is it used and what results will be
obtained?
A histogram is mainly used to analyze a process by
examining the location of the mean value in the graph or
degree of variations, to find a problem point that needs to
be improved. Its other applications are listed in the table
below. Usage Results
[Used during phases to monitor the Can identify the location of the
situation, analyze causes, and review mean (central) value or degree of
effectiveness of an action.] variations.
Used to assess the actual conditions. Can find out the scope of a defect
by inserting standard values.
Used to analyze a process to identify a
problem point that needs to be improved Can identify the condition of
by finding the location of the mean value distribution (e.g., whether there is
or degree of variations in the graph. an isolated, extreme value).
Used to examine that the target quality is
maintained throughout the process.
Histogram--Example No. 1

Data sheet of lengths of cut steel wire [Specification: 255±5cm] (n=100)

№ 1 2 3 4 5 6 7 8 9 10
1 255 259 257 254 253 254 253 257 258 252
2 253 256 255 255 256 255 257 255 256 258
3 257 255 256 251 255 253 255 256 254 256
4 257 255 257 254 254 260 258 253 260 255
5 255 252 255 253 253 258 253 259 255 257
6 253 257 258 256 253 254 255 254 257 253
7 255 254 253 255 257 252 254 256 255 255
8 254 254 254 254 255 255 257 255 253 254
9 258 256 253 256 255 254 255 256 256 256
10 256 254 255 257 254 254 259 253 258 254
Ｓ 253 252 253 251 253 252 253 253 253 252
Ｌ 258 259 258 257 257 260 259 259 260 258
(Unit;cm)
 Histogram--Example No.2

(Frequency Distribution Table Cutting Length of Steel Wire)

(Standard: 255± 5cm)
Central Valee of
№ Section Each Section Frequency Marking No. of Occurrences

1 250.5- 251.5 251 1

2 251.5- 252.5 252 3
3 252.5- 253.5 253 15
4 253.5- 254.5 254 19
5 254.5- 255.5 255 24
6 255.5- 256.5 256 14
7 256.5- 257.5 257 12
8 257.5- 258.5 258 7
9 258.5- 259.5 259 3
10 259.5- 260.5 260 2
Total 100
Histogram--Example No.3

Standard Standard
Lower Limit Upper Limit
Products
Standard Value

25 X
Standard Central
N=100 =255.19
20

15

10

0
250 252 254 256 258 260
[Histogram of Cutting Length of Steel Wire]
Interpretation of Data Depicted in
Histogram
Name Description Example Cause

A peak in the A so-called “normal

center, gradually distribution.”
declining in both Means that this
General directions. particular process
Shape Almost symmetric. is stable.

The average value Possible causes

(peak) is off- include the
centered. standard value
The shape of inserted off the
distribution shows center or the
Trailing Type a relatively steep component of an
Type e incline on one side impurity close to 0
and a moderate (zero). The stability
slope on the other. of the process is
Asymmetric. the same as that
described for the
General Shape.
 Name Description Example Cause

Less number of data This shape indicates

around the center of the overlapping of two
distribution. different distributions,
Two peaks, one on when there is a
variation between two
each side. machines or two
Twin-peak workers performing
the same task, often
Shape caused by one of them
doing the task in a
wrong way.

Small variations in the Caused by the same

number of data around reason described
the center of above, but with less
Plateau distribution, forming a variation.
Shape plateau.
 Name Description Example Cause

The average value is Distribution A portion of distribution

extremely off-centered, where defects depicted by dashed
showing a steep seem to be lines in the diagram
decline on one side excluded. has been removed for
and a moderate slope some reason.
on the other. For example, when
Asymmetric. defective products are
found during an
Precipitous inspection before
Shape shipping and removed
from the lot, the results
of an acceptance
inspection performed
on that lot by the
customer will show this
shape of distribution.
 Name Description Example Cause

The otherwise normal This shape appears

histogram shows an when a small amount
of data from a different
“isolated island” either distribution has been
on the right or left side. accidentally included.
It will be necessary to
Isolated 　 examine the data
Island Shape history to find
anomalies in the
process, errors in
measurement, or the
inclusion of data from
another process.

The every other It will be necessary to

section (vertical bar) check if the width of
each section has been
shows the number of
Gapped data smaller than the
determined by
multiplying the unit
Teeth Shape one next to it, forming (scale) of
(or Teeth of a gapped-teeth or measurement with an
integer, or if the
Comb teeth-of-a-comb person who performed
Shape) shape. the measurement has
read the scale in a
certain deviant
manner.
 Scatter Diagram
A scatter diagram is used to regression line
“examine the relationship
between the two, paired,
interrelated data types, ”
such as “height and weight .
of a person.” . . . .
. . . . . .
A scatter diagram provides a . . . . . . .

Abrasion
means to find whether or not . .. . . .
these two data types are . . . . .
interrelated. It is also used . . . .
to determine how closely . . .
they are related to identify a
problem point that should be
controlled or improved. Number of Rotations
When is it used and what results will be obtained?

Used to a assess relationship between 2 data matters

Usage Results
[Used during phases to monitor the situation, analyze Can identify cause and
causes and review effectiveness of an action.] effect relation.
Used to identify a relationship between two matters.
(Can understand the
Used to identify a relationship between two matters and relationship between two
establish countermeasures based on their cause and
effect relation. results.)
Example Usage
•Relationship between thermal treatment temperature of
a steel material and its tensile strengths.
•Relationship between visit made by a salesman and
volume of sales.
•Relationship between the number of persons visiting a
department store and volume of sales
 Various Forms of Scatter Diagram
The table below shows some examples of scatter diagram’s usage. If, for example,
there is a relationship where “an increase in the number of rotations (x) causes an
increase in abrasion (y),” there exists “positive correlation.” If, on the other hand, the
existence of a relationship where “an increase in the number of rotations (x) causes a
decline in abrasion (y)” indicates that there is “negative correlation.”

・ ・ ・ ・
・・ ・ ・・・
・ ・・・・・ ・・ ・ ・・
・・・ ・・ ・ ・ ・・・ ・・
・ ・・・・・ ・ ・・・・・
・・・・ ・ ・・・ ・・
・ ・・・ ・・ ・・・
・・・ ・ ・ ・
・ ・
Where there is a positive Where there is a negative
correlation correlation
・ ・ ・・
・ ・ ・ ・ ・ ・・ ・・ ・
・ ・ ・ ・・ ・ ・ ・ ・・
・ ・・ ・ ・ ・ ・ ・ ・ ・・ ・
・ ・ ・ ・ ・ ・ ・ ・ ・ ・ ・
・ ・ ・ ・ ・ ・ ・・ ・ ・
・ ・ ・ ・ ・ ・ ・・
・ ・ ・ ・・ ・
・ ・
Where there is no Where there is a non-
correlation linear correlation
Check Sheet
A check sheet is “a sheet designed in
advance to allow easy collection and
aggregation of data.” By just entering check
marks on a check sheet, data can be
collected to extract necessary information, or
a thorough inspection can be performed in an
efficient manner, eliminating a possibility of
skipping any of the required inspection items.
A check sheet is also effective in performing
stratification (categorization).
Example Usage of Check Sheet

A check sheet used to identify defects

Date
Defect
6/10 6/11 6/12 6/13 6/14 Total

Vertical 34
Scratch

Scratch 11

Dent 37
When is it used and what results will be obtained?

Usage Results
• Used to collect data. •Ensures collection of
• Used when performing a thorough required data.
inspection • Allows a thorough
• Used to identify the actual inspection
condition of a situation. of all check items.
(Used during phases to monitor the • Can understand
situation, analyze causes, review tendencies and variations.
effectiveness of an action, perform • Can record required data.
standardization, and implement a
selected control measure.)
Control Chart
A control chart is used to examine a
process to see if it is stable or to
maintain the stability of a process. This
method is often used to analyze a
process. To do so, a chart is created
from data collected for a certain period of
time, and dots plotted on the chart are
examined to see how they are distributed
or if they are within the established
● ●
●
● control limit. After some actions are
Ｘ ● ● ●
● taken to control and standardize various
● ●

factors, this method is also used to

●
● ● ●
examine if a process has stabilized by
these actions, and if so, to keep the
● ● ● ●
Ｒ ● ●

process stabilized.
X - R Control Chart
When is it used and what results will be obtained?

Usage Results
•[Used
Used toduring
collect phases
data. to monitor Can identify a change caused
•byEnsures
elapsecollection
of time. of
•the situation,
Used analyzea causes,
when performing thorough required data.
inspection
review effectiveness of an Can judge the process if it is in
•itsAllows
•action,
Used toperform
identify the actual condition of
Standardization normal a thorough
state or inspection
there are
a situation. some
and implement a selected of allanomalies
check items.by examining
(Used during phases to monitor the the dots plotted on the chart.
control measure.]
situation, analyze causes, review • Can understand tendencies
effectiveness of an action, perform In
and the example x(-)-R control
variations.
Used to observe
standardization, and a changea
implement chart, x(-)” represents the
• Can record required data.
caused by elapse
selected control of time.
measure.) central value, while “R”
indicates the range.

Control Chart for Managerial Purposes: Extends the line indicating the control limit used for analytical
purposes to plot data obtained daily to keep a process in a good state.

* Control Chart for Analytical Purposes: Examines a process if it is in a controlled state by

collecting data for a certain period of time. If the process is not controlled, a survey is performed to identify
its cause and develop countermeasures.
Major Application
Out of specification:
It is necessary to investigate the cause

N=5 UCL=5.780
5.8 ●

●
● ● CL=5.400
X 5.4
● ● ● ● ●
● ● ● ●
● ● ● ● ●
● ● ●
5.2 ●
LCL=5.020

1.0 ×
×
× × ×
×
R 0.5 ×
×
×
× ×
×
× ×

× × × × ×
× × ×
0
0 5 10 15 20
X-R Control Chart
Graph
A graph is “a graphical representation of data, which
allows a person to understand the meaning of these
data at a glance.”
Unprocessed data simply represent a list of
numbers, and finding certain tendencies or
magnitude of situation from these numbers is
difficult, sometimes resulting in an interpretational
error.
A graph is a effective means to monitor or judge the
situation, allowing quick and precise understanding
of the current or actual situation.
A graph is a visual and summarized representation
of data that need to be quickly and precisely
conveyed to others.
When is it used and what results will be obtained?

A graph, although it is listed as one of the QC tools, is

commonly used in our daily life and is the most familiar means
of assessing a situation.

Usage Usage Results

Used to observe A graphs is the most
changes in a time- frequently used tool
sequential order (line among QC 7 tools.
graph) Can recognize
Used to compare size changes in a time-
(bar graph) sequential order,
Used to observe Ratios ratios, and size.
( pie graph, column
graph)
 Example usage of Graph

Bar Graph of Sales Band Chart of Expenses

(Yen million)
Survey Period:2000.12
Presented by:M/K
500
(Yen million)
　0 100 200 300 400 500 600 700 800
400

S
Before Chemicals Oils Electricity
a
300 Taking (170)
l (430) (200)
Actions
e
s (Total:Yen 8 million)
200

After Chemicals Oils

100 Taking
(240) (150)
Actions
0 (Total:Yen 4.95 million)
Electricity
Iwate Tokyo Osaka Shizuoka
(105)
Control Charts
The History of Control Charts

 Developed in the 1920’s

 Dr. Walter Shewhart, then an employee of
Bell Laboratories developed the control chart
to separate the special causes of variation
from the common causes of variation.
Statistical Process Control (SPC)
A methodology for monitoring a process to
identify special causes of variation and
signal the need to take corrective action
when appropriate
 SPC relies on control charts
Common
Causes

Special
Causes
COMMON CAUSE

 RANDOM VARIATION
 SUM OF MANY SMALL VARIANCES

 SYSTEM-RELATED

 80% OF PROCESS VARIATION

 RESPONSIBILITY OF MANAGEMENT

 WRONGLY ATTRIBUTED TO LINE

EMPLOYEES
SPECIAL CAUSES

 ASSIGNABLE

 20% OF PROCESS VARIATION

 IDENTIFIABLE TO SPECIFIC
CONDITIONS
 OVERCOME BY REMOVAL, TRAINING,
EXPERIENCE and/or COACHING
(2) Assignable causes variation:

425 425 425

(a) Location (b) Spread (c) Shape

Out of control In control

(assignable causes present) (no assignable causes)
Histograms do not
take into account
changes over time.

Control charts
can tell us when a
process changes
Control Chart Applications
Establish state of statistical control
Monitor a process and signal when
it goes out of control
Determine process capability
Capability Versus Control

Control
Capability In Control Out of Control

Capable IDEAL

Not Capable
Commonly Used Control Charts
 Variables data
 x-bar and R-charts
 x-bar and s-charts

 Charts for individuals (x-charts)

 Attribute data
 For “defectives” (p-chart, np-chart)
 For “defects” (c-chart, u-chart)
Developing Control Charts
1. Prepare
 Choose measurement
 Determine how to collect data, sample size,
and frequency of sampling
 Set up an initial control chart
2. Collect Data
 Record data
 Calculate appropriate statistics
 Plot statistics on chart
Next Steps
1. Determine trial control limits
 Center line (process average)
 Compute UCL, LCL
2. Analyze and interpret results
 Determine if in control
 Eliminate out-of-control points
 Recompute control limits as necessary
Typical Out-of-Control Patterns
 Point outside control limits
 Sudden shift in process average

 Cycles

 Trends

 Hugging the center line

 Hugging the control limits

 Instability
Shift in Process Average
Identifying Potential Shifts
Cycles
Trend
Final Steps

1. Use as a problem-solving tool

 Continue to collect and plot data
 Take corrective action when
necessary
2. Compute process capability
Process Capability Calculations
Special Variables Control Charts
x-barand s charts
x-chart for individuals
Charts for Attributes
 Fraction nonconforming (p-chart)
 Fixed sample size
 Variable sample size

 np-chart for number nonconforming

 Charts for defects

 c-chart
 u-chart
Control Chart Selection
Quality Characteristic
variable attribute
defective defect
no
n>1? x and MR
yes constant
yes constant
p or sampling
sample
np unit?
n>=10 or no size?
x and R
computer? yes no
no
yes
p-chart with c u
x and s variable sample
size
Types of Shewhart Control Charts
Control Charts for Variables Data
X and R charts: for sample averages and ranges.
X and s charts: for sample means and standard deviations.
Md and R charts: for sample medians and ranges.
X charts: for individual measures; uses moving ranges.

Control Charts for Attributes Data

p charts: proportion of units nonconforming.
np charts: number of units nonconforming.
c charts: number of nonconformities.
u charts: number of nonconformities per unit.

5
The Central Limit Theorem

Suppose a population has a mean (µ )

and a standard deviation (σ )

The Central Limit Theorem states

The distribution of sample means ( X )

will be approximately normal.
Its mean X = µ ,
and its standard deviation σ X = σ / n

5
Central Limit Theorem Illustrated
99.7% of all
sample means

Sample
(Basis for specification limits) means
Population,
Individual
items

µ -3σ x µ µ +3σx

5
Control Charts

 Logic Behind Control Charts

 Consider measurement of variables data
 We know that a sample average typically varies from the
population average.
 The problem is to determine if any variation from a
specified population average is
 Is simply random variation
 Or is because the population average is not as specified
 We therefore establish limits on how different we’ll allow
the sample average (or whatever other summary measure)
to be before we conclude the specification is not being met.
 Control Limits Set via Sampling Theory
Control Charts

 The Good News:

 We don’t need to go back to the statistics books
and tables
 Simple-to-use tables and formulae have been
developed for creating control charts
 Formulae and tables for variables data
 Formulae only for attributes data
Process Control Chart Factors
Control Limit UCL Factor LCL Factor Factor for
Factor for for Ranges for Ranges Estimating
Sample (Range (Range
(Subgroup) Averages Sigma
Size (Mean Charts) Charts) Charts) ( = R/d2)
(n) (A2) (D4) (D3) (d2)

2 1.880 3.267 0 1.128

3 1.023 2.575 0 1.693
4 0.729 2.282 0 2.059
5 0.577 2.115 0 2.326
6 0.483 2.004 0 2.534
7 0.419 1.924 0.076 2.704
8 0.373 1.864 0.136 2.847
9 0.337 1.816 0.184 2.970
10 0.308 1.777 0.223 3.078

5
Control Charts
 Process Overview
 First, develop sampling plan:
 Number of observations per sample
 Frequency of sampling
 Stage 1 sampling:
 Conduct initial periodic sampling
 Determine control limits
 Perform calculations
 Decide whether in control or not
 Stage 2 sampling (only if Stage 1 is successful):
 Continue operating with periodic sampling
 Perform calculations
 Decide whether in control (each sample)
 SPC: Control Limits

µ +3σx UCL

µ -3σx LCL
 SPC: Control Limits
In control Out of control
Process Process center
is stable has shifted

UCL •
µ +3σx •
•
• •
µ
• •
• • •
•
µ -3σx LCL
 X and R Charts
Sample Number
1 2 3 4 25

Values
4 7 6 7
Select 25 small samples
(in this case, n=4) 6 3 9 6
Find X and R of each 5 8 8 6
sample.
5 6 9 5
The X chart is used to Sum 20 24 32 24 28 Total
control the process mean.
X 5 6 8 6 7 150
The R chart is used to R
control process variation. 2 5 3 2 3 75
 X and R Charts
Sample Number
n A2 D4 D3 d2 1 2 3 4 25
4 7 6 7

Values
2 1.880 3.267 0 1.128 6 3 9 6
3 1.023 2.575 0 1.693 5 8 8 6
4 0.729 2.282 0 2.059 5 6 9 5
Sum 20 24 32 24 28 Total
X 5 6 8 6 7 150
R 2 5 3 2 3 75
 X and R Charts
Sample Number
n A2 D4 D3 d2 1 2 3 4 25
4 7 6 7

Values
2 1.880 3.267 0 1.128 6 3 9 6
3 1.023 2.575 0 1.693 5 8 8 6
4 0.729 2.282 0 2.059 5 6 9 5
Sum 20 24 32 24 28 Total
X 5 6 8 6 7 150
X = 150 / 25 = 6 R 2 5 3 2 3 75
–
–
R = 75 / 25 = 3
–
A2R = 0.729(3) = 2.2
–
UCLX = X– + A2R
– – = 6 + 2.2 = 8.2
LCLX = X
– - A2R
– – = 6 - 2.2 = 3.8
UCLR = D4R
– = 2.282(3) = 6.8
– = 0(3) = 0
LCLR = D3R
 X and R Charts
Sample Number
n A2 D4 D3 d2 1 2 3 4 25
4 7 6 7

Values
2 1.880 3.267 0 1.128 6 3 9 6
3 1.023 2.575 0 1.693 5 8 8 6
4 0.729 2.282 0 2.059 5 6 9 5
Sum 20 24 32 24 28 Total
X = 150 / 25 = 6 X 5 6 8 6 7 150
–
– = 75 / 25 = 3 R 2 5 3 2 3 75
R
– R = 0.729(3) = 2.2
A UCL X– = 8.2
2 –
–
Mean
UCL – = X + A R = 6 + 2.2 = 8.2 X = 6.0
X 2
–
– – LCL X– = 3.8
LCLX = X - A2R = 6 - 2.2 = 3.8
–
– – UCL R = 6.8
UCLR = D4R = 2.282(3) = 6.8
Range

– –
LCLR = D3R R = 3.0
– = 0(3) = 0
LCL R = 0
p Chart
Sample number
1 2 3 4 25 Total
n 50 50 50 50 50 1250
#def 2 4 0 3 2 50
p .04 .08 0 .06 .04 1.00
 p Chart
– Σ #def
p= = 50/1250 = .04
Σn Sample number

3 σP = 3 p(1-p) 1 2 3 4 25 Total
n n 50 50 50 50 50 1250
=3 .04(.96) #def 2 4 0 3 2 50
50
p .04 .08 0 .06 .04 1.00
= 0.083

UCL P = p + 3 σP
–

= .04 + .083 = .123

–
UCL P = p - 3 σP

= .04 - .083 = 0
can't be negative
 p Chart
– Σ #def
p= = 50/1250 = .04
Σn Sample number

3 σP = 3 p(1-p) 1 2 3 4 25 Total
n n 50 50 50 50 50 1250
=3 .04(.96) #def 2 4 0 3 2 50
50
p .04 .08 0 .06 .04 1.00
= 0.083

UCL P = p + 3 σP
–
UCL P = 0.123
= .04 + .083 = .123 •
•
–
UCL P = p - 3 σP –
• p = 0.04
= .04 - .083 = 0 LCL P = 0
can't be negative
•
Hotel Suite Inspection -
Defects Discovered
Day Defects Day Defects Day Defects

1 2 10 4 19 1
2 0 11 2 20 1
3 3 12 1 21 2
4 1 13 2 22 1
5 2 14 3 23 0
6 3 15 1 24 3
7 1 16 3 25 0
8 0 17 2 26 1
9 0 18 0
Total 39
c Chart for Hotel Suite
Inspection
Number of defects
5 UCL = 5.16

4
3
2
c = 1.50
1
0 LCL = 0
5 10 15 20 25 Day
CONTROL CHARTS
WHY INSPECTION DOESN’T WORK

 IN ORDER TO CONSISTENTLY SHIP QUALITY

PRODUCT TO THE CUSTOMER YOU HAVE TO
MONITOR THE PROCESS NOT THE PRODUCT
 INSPECTION (if you’re lucky) FINDS DEFECTS
AFTER THE FACT
 THIS RESULTS IN C.O.P.Q. COSTS THAT COULD
HAVE BEEN DETECTED OR AVOIDED MUCH
EARLIER IN THE PROCESS
CONTROL CHARTS
THE BASICS

CONTROL CHART

Upper Control
Limit
Y (results)

X (Grand Average or
(Expected Result)

Lower Control
Limit

X (observations)
CONTROL CHARTS
VARIATION
 CONTROLCHARTS DISTINGUISHES
BETWEEN:
 NATURAL VARIATION (COMMON CAUSE)
 UNNATURAL VARIATION (SPECIAL CAUSE)

UNNATURAL VARIATION
UCL

Average NATURAL VARIATION

LCL
UNNATURAL VARIATION
CONTROL CHARTS
XBAR - R CHART STEPS (1)
 DETERMINE SAMPLE SIZE (n=2-6)
 DETERMINE FREQUENCY OF SAMPLING
 COLLECT 20-25 DATA SETS
 AVERAGE EACH SAMPLE (X-bar)
 RANGE FOR EACH SAMPLE (R)
 AVERAGE OF SAMPLE AVERAGES =
 X-double bar
 AVERAGE SAMPLE RANGES =
 R-bar
CONTROL CHARTS
XBAR - R CHART STEPS (2)

X CONTROL LIMITS: -
BAR

UCL = XDBAR + (A2)(RBAR) -

LCL = XDBAR - (A2)(RBAR)
R CONTROL LIMITS:
- UCL = (D4)(RBAR) -
LCL = (D3)(RBAR)
Determining if your control Chart is “Out of
Control”
 Control Chart

Upper Control
Limit
Zone “A”
2 sigma limit
Zone “B”
Y (results)

1 sigma limit
Zone “C”
Average
Zone “C”
1 sigma limit
Zone “B”
2 sigma limit
Zone “A” Lower Control
Limit

X (observations)
 Control Charts
 Tests for Assignable (special) causes
Test 1 One point beyond 3 sigma
Test 2 Nine points in a row on one side of the centerline
Test 3 Six points in a row steadily increasing or decreasing
Test 4 Fourteen points in a row alternating up and down
Test 5 Two out of three points in a row beyond 2 sigma
Test 6 Four out of five points in a row beyond 1 sigma
Test 7 Fifteen points in a row within I sigma of the
centerline
Test 8 Eight points in a row on both sides of the centerline,
all beyond 1 sigma
CONTROL CHARTS
INTERPRETATION
 SPECIAL: ANY POINT ABOVE UCL OR
BELOW LCL
 RUN: > 7 CONSECUTIVE PTS ABOVE OR
BELOW CENTERLINE
 1-IN-20: MORE THAN 1 POINT IN 20
CONSECUTIVE POINTS CLOSE TO
UCL OR LCL
 TREND: 5-7 CONSECUTIVE POINTS IN
ONE DIRECTION (UP OR DOWN)
CONTROL CHARTS
IN CONTROL w/ CHANCE VARIATION
Control Chart - Chance Variation

UCL
Y (results)

Ave.

LCL

X (observations)
CONTROL CHARTS
LACK OF VARIABILITY
Control Chart - Lack of Variability

UCL
Y (results)

Ave.

LCL

X (observations)
CONTROL CHARTS
TRENDS
Control Chart - Trend

UCL
Y (results)

Ave.

LCL

X (observations)
CONTROL CHARTS
SHIFTS IN PROCESS LEVELS
Control Chart - Shifts in Process Level

UCL
Y (results)

Ave.

LCL

X (observations)
CONTROL CHARTS
RECURRING CYCLES
Control Chart - Recurring Cycles

UCL
Y (results)

Ave.

LCL

X (observations)
CONTROL CHARTS
POINTS NEAR OR OUTSIDE LIMITS

Control Chart - Points Near or Outside

Control Limits

UCL
Y (results)

Ave.

LCL

X (observations)
CONTROL CHARTS
ATTRIBUTE CHARTS
 TRACKS CHARACTERISTICS
- SHORT OR TALL; PASS OR FAIL
 ONE CHART PER PROCESS
 FOLLOW TRENDS AND CYCLES

 EVALUATE ANY PROCESS CHANGE

 CONSISTS OF SEVERAL SUBGROUPS

(a.k.a. - LOTS)
- SUBGROUP SIZE > 50
CONTROL CHARTS
ATTRIBUTE CHART TYPES

p chart = Proportion Defective

 np chart = Number Defective

 c chart = Number of nonconformities within a

constant sample size
 u chart = Number of nonconformities within a
varying sample size
CONTROL CHARTS
np CHART EXAMPLE

np Chart

25
20
# of Defects

UCL
15
10 c

5
0
11

13

15

17

19

21
1

Serial Number
CONTROL CHARTS
RISKS
 RISK 1: FALSE ALARM -
REJECT GOOD LOT - CALL
PROCESS OUT OF CONTROL WHEN IN
CONTROL
 RISK 2: NO DETECTION OF PROBLEM
- SHIP BAD LOT -
CALL PROCESS IN CONTROL WHEN
OUT OF CONTROL
Process Capability
Analysis
Process Capability Analysis

 Differs Fundamentally from Control Charting

 Focuses on improvement, not control
 Variables, not attributes, data involved
 Capability studies address range of individual
outputs
 Control charting addresses range of sample
measures
 Assumes Normal Distribution
 Remember the Empirical Rule?
 Inherent capability (6σ x ) is compared to
specifications
 Requires Process First to be In Control
Process Capability:
Normal Curve

2σ
(68%)

µ
4 σ (95.5%)
6 σ (99.7%)

5
 Process Capability
Process Capability (PC) is the range in which "all" output
can be produced.

Definition:
PC = 6σ

µ
6 σ (99.7%)
5
Process Capability Chart
Process output
distribution

Output Output
out of spec out of spec

5.010

4.90 4.95 5.00 5.05 5.10 5.15

cm
X
Tolerance band
LSL USL
Inherent capability (6σ )

5
 Process Capability
This process is
CAPABLE of
producing all good
output.

➤ Control the process.

Lower Upper
Spec Spec
Limit Limit
This process is
× NOT CAPABLE.
CAPABLE

➤ INSPECT - Sort out

the defectives

5
 Process Capability

Process Capability:
Cp = Design Spec Width / Process Width
Cp = (USL-LSL) / 6σ
Cp should be a large as possible

Process Capability Ratio:

Cr = 1/Cp * 100
Indicates percent of design spec. used by process
variability
Cr should be as small as possible
 Process Capability

Process Capability Index (account for Mean Shifts):

Cpk = Cp * (1-k)
where k = Process Shift / (Design Spec Width/2)

Or
Cpk = Min (Cpl, Cpu)
Cpl = (X - LSL)/3σ
Cpu = (USL - X)/3σ
 Process Capability

Cpk Meaning
Negative. Process Mean outside Spec Limits
0 - 1.0 Portion of process spread falls Outside Specs
> 1.0 Process spread falls within Spec Limits

Six Sigma Cpk = 1.5

 Process Capability

Process Capability Ratio: Cp = Design Spec

Width/Process Width
Cp = (USL-LSL)/6σ
Process Capability Index (account for Mean Shifts):
Cpk = Cp (1-k)
where k = Process Shift/(Design Spec Width/2 )
Or, Minimum of
• (X - LSL)/3σ
• (USL - X)/3σ
 Process Capability Ratios
(Desired Performance) / (Actual Performance)
Note that average
performance is not
centered between the The shaded areas
This curve is the spec limits represent the
distribution of data percentage of off-
from the process spec
production

Voice of Customer

Voice of Process
Target rule:
Cp - Cpk ≤ 0.33

Variation rule:
Cp ≥ 1.33
Process Capability Index
Index Cpk compares the spread and location
of the process, relative to the
specifications.
–

{
Upper Spec Limit - X
Cpk = the smaller of: OR – 3σ
X - Lower Spec Limit
3σ
Alternate Form

Zmin –

{
Upper Spec Limit - X
Cpk =
3 OR
σ
Where Zmin is the smaller of: –
X - Lower Spec Limit
σ
5
Process Capability: C pk Variations
(a) (b) (c)

Cpk = 1.0 Cpk = 1.33 Cpk = 3.0

LSL USL LSL USL LSL USL

(d) (e) (f)

Cpk = 1.0 Cpk = 0.60 Cpk = 0.80

LSL USL LSL USL LSL USL

5
PROCESS CAPABILITY MEASUREMENT

Process Capability is computed as :

6 σ = 6 S = 6 R / d
Process Capability Index Cp = U – L / 6 σ
Cpk = U – X/3σ
If : Cp > 1.6 Process is Excellent
Cp > 1.3 Process is Good
Cp > 1.0 Process is Satisfactory
Cp < 1.0 Process is Poor
Sources of Variation in
Production Processes
CONTROL LOOP

INPUT OUTPUT
PROCESS

MEASUREMENT
ADJUST

DECIDE ON
FIX? STATISTICS

ID GAPS

EXAMINE
 Control

Capability In Control Out of Control

Capable Ideal

Not Capable
Contents

 Quality & TQM

 Basic Statistics

 Seven QC Tools

 Control Charts

 Process Capability Analysis

Thank You.