CHECK Improvement

DO improvement results
Improvement D C
activities
P A
PLAN improvement ACT to standardize or
replan

Management
strategy
Initiate Standardization
Focus on improvement
vital few

KNOW the STANDARD

ACT to improve the
standard or its use A S
Daily work
C D
CHECK the work DO the work according
against the standard to the standard

Focussing on the
Vital Few Successful QC Circle Activities

1
 ACT to improve the standard or its use Know the STANDARD

SDCA Cycle of A S
Daily work
C D
CHECK the work against the standard Do the work according
to the standard

The SDCA Cycle

CHECK Improvement results
DO improvement
D C
PDCA Cycle of
improvement
work P A
PLAN improvement ACT to standardize or replan

KNOW the STANDARD

ACT to improve the
Standard or its use A S
Daily work
C D
CHECK the work DO the work according to
against the standard the standard

The SDCA/PDCA Improvement Cycle

2
 Seven Problem Solving Disciplines
S Definition Observation Analysis Actions Check Standardisation Conclusion
T (Identify & define) (Investigate (Find root causes) (Establish and (Ensure (Hold Gains) (Review process
E Features) implement effectiveness) & future plans)
P remedies)
S

 To collect problems in work  To investigate  To identity all  To identity  To check the  To ensure that the  To review
P
area. the time, place, factors which important counter process and the gains of problem solving
U
type and influence measures. results of improvements are process for
R  To select and define a
symptoms of the problems and implementation held for all time in future
P priority problem.  To systematically
problem. find out root for monitoring future. improvements in
O implement
causes. extent of work and
S counter
improvements. problem solving
E measures.
process.

 Discuss in groups work area  Observe the  Establish the  Establish the  Check the trend  Incorporate the  Review the
corrective and and the level of measures in the problem solving
problems. features of the cause effect
preventive
problem. relationship measures. the improvement operation process for
 Identify problems in inputs, against target standards future
K  Workout
outputs and the work  To collect the  Find out feasibility & improvements in
E
Y effectiveness.  Check the  Train the people
process. data and the important root work place
A  Select and Process in revised
required facts causes having
C prioritise counter standards  Analyse the
 Prioritise based on urgency,
T the highest measures.  Check & ensure
importance and control. steps followed &
I impact on the  Chalkout that no adverse  Cotinuously track
V activities, time tools used in
I problem side effects are the effectiveness
 Define the problem schedule and group to improve
T resources present in the to hold on to the
I statement. the future
E  Assign next and previous gains problem solving
S responsibility processes capacity.
 Setup control
point
 Carry out the
plan

3
 Systematic Problem Solving
and Deming Cycle

ACT PLAN

CHECK DO

Plan : Setup the Objectives and Means

Do : Put the Plan into Practice
Check : Observe the Results and the Process
Act : Standardise if Results are Satisfactory otherwise
Re-plan and Follow the Cycle again

4
 Basic Steps of Problem Solving

DEFINITION
PLAN OBSERVATION
ANALYSIS

DO ACTION

CHECK CHECK

STANDARDISATION
ACT
CONCLUSION

5
 Steps of Problem Solving

1. Definition –
Identifying and defining the problem
2. Observation – Investing
the features of the problem
3. Analysis – Finding
the root causes
4. Actions –
Establishing and implementing
remedies (countermeasures)
5. Check – Ensuring
the effectiveness of remedies
6. Standardisation – Holding
on the gains
6
Benefits of Systematic Problem Solving Process

• It helps avoid jumping to conclusion

• Helps avoid decisions based on opinions and feeling

• Helps the group in focusing their attention

• It provides a logic for remedies

• Sets a platform to involve all concerned

• Makes implementation simple

7
 Summary

• QC story and the Basic Seven Step Problem

Solving Process.

• Concept of PDCA and its adaptation to

problem solving in specific work situations.

8
 Technical Analysis and Statistical Analysis

If you see female Japanese assembly line workers without any technical background whatsoever
making suggestions even engineers haven’t been able to think of, you will ask what makes it possible
for those women to acquire their technical knowledge. The answer is statistical methods.
We have two methods of analyzing and eliminating trouble in the manufacturing shop. One is by
technological analysis; the other is by statistical analysis. QC uses statistical methods to analyze and
improve the quality of products.
In the color TV factory of our company, a female employee of the Quality Assurance Section found that
the failure rate of TV tuners differs depending on the type of TV, even when the same type of TV tuner is
installed in the sets.
She thought that there must be some reason for this difference in failure rates of TV tuners. Therefore,
she drew diagrams which showed the relation between the failure rate of tuner and the length of the
shaft, the temperature of the set, the diameter of the tuner knob, size of the cabinet and so on. At last,
she discovered a correlation between the failure rate and the distance from tuner to speaker; in other
words, the failure rate of the turner is quite low when the tuner is attached far from the speaker. On the
other hand, when the tuner is attached near the speaker, the set doesn’t work well.

9
 Technical Analysis and Statistical Analysis (Contd.)

Such a conclusion would be hard to draw through technical analysis alone. But by accumulating market
data we can discover such a phenomenon. We call this the law of large numbers.
Therefore, you may understand that we have two ways to find out the cause of defectives, one is by the
use of analysis based on technology and the other is through statistics.
One of QC’s specialties is the use of statistical methods to eliminate trouble in the shop.
In order to find causes for a defect, you don’t need sophisticated technical expertise. What you need to
do is analyze data. And quality control circles have learned to use statistical tools; this is what makes
the circles so successful. Statistical analysis can be used to solve problems not only in manufacturing
but also in sales, accounting, personnel management and service.
Spurred on by this method, QC circles are in demand among various fields including manufacturing,
construction, financing, restaurants and department stores. The same thing may produce varying
degrees of results. Companies introducing the method seem to get better results in their work.
– Source : Hajime Karatsu

10
The Seven QC Tools

11
 The Seven QC Tools

Tool Function
1. Check sheet / : Easy data collection
Data Sheet / Tally Sheet
2. Histogram : Grasp shape of distribution of quality
3. Pareto Diagram : Narrow problem area
4. Cause and Effect : Assess factors which seem to cause
Diagram quality characteristic
5. Stratification : Verify effect of assumed discrete
causes on quality characteristic
6. Scatter diagram : Verify effects of assumed continuous
causes on quality characteristic
7. Graph / Control : Grasp dynamic change and confirm
chart observation of standards
12
Pareto Chart

13
 Pareto Analysis

1897 Italian Economist V. Pareto

Formula showing ‘Distribution of income as uneven.
1907 U. S. Economist M. C. Lorenz
Pareto The largest share of income or wealth is held by a very
and small number of people.
Lorenz
Dr. J. M. Juran applied the theory to classify problems of quality
into the

‘Vital Few and Trivial Many’

14
Step 1 Decide what problems are to be inspected and how to
collect the data.
Step 2 Design a data tally sheet listing the items and record
their totals.
Step 3 Fill out Tally sheet and calculate total.
Step 4 List items; their individual total, cumulative totals,
percentage of overall total and cumulative percentage.
Step 5 Arrange the items in the order of quantity and fill out
the data sheet.
Step 6 Mark abscissa and ordinate.
Step 7 Construct a bar diagram.
Step 8 Draw the cumulative curve.
Step 9 Write any necessary item on the diagram.

15
 Pareto Charts

Record of defects
Date : Number inspected : N – 2165
Type of Number of Percent Percent of
Defect defects defects composition

Caulking 198 9.1% 47.6

Fitting 25 1.2% 6.0
Connecting 103 4.8% 24.7
Improper torque 18 0.8% 4.3
Gapping 72 3.3% 17.3
Total 416 19.2% 99.9

16
 Number of defects

200
400
(Pieces)

Caulking

Connecting

Gapping

17
Fitting
Pareto Diagram

Improper
torque
25
50
75
100%

Percent of composition
Check Sheet
or
Tally Sheet

18
 Check Sheet

Purpose
To systematically record and compile data to study patterns and trends
by team members.
A check sheet helps to count and accumulate data.
What does it do ?
• Simple and easy-to-understand presentation of data.
• Helpful to analyse performance of key processes.
• Helps to make decision making process easier by the team
members. Check sheet brings clarity and forces agreement on
the definition of each condition or event by team members.
• Brings clarity in the pattern of data.

19
Procedure to use it
1 Bring clarity of understanding in the events or conditions under
study.
Example : If the event is ‘late payment’, agree on definition of late.
2 Data collection – decide responsibility, duration and sources.
(Be careful about ‘stratification of data’)
3 Checksheet form is to be simple to use, easy to understand and
complete in all respects.
(refer the model check sheet form enclosed)
4. Ensure consistent and accurate data collection. Adhere to the
check sheet form and write legibly.

20
 Check Sheet Form (Example)
1. Project : Delay in railway Form No. XYZ/5.1.95/Rev. 1
Reservation
2. Location : Booking counter,
Central Railway 4. Shift : All
3. Recorder : Harish Samant 5. Date – 5 / 2 / 94 to 28 /
2 / 94
7. Dates 9. Row
5/2/94 7/2/94 11/2/94 18/2/94 21/2/94 25/2/94 28/2/94
6. Reasons Total
a. Initial setting time 9 4 6 6 3 12 12 52
b. Terminal Break Down
2 7 2 4 5 8 3 31

c. Non-uniform
processing speed 7 3 1 2 2 4 5 24
(manual)
8. Column Total 18 14 9 12 10 24 20 10. Grand
Total 107

1. Name of the project 2. Location of data collection 3. Name of person collecting data
4. Important identifiers 5. Period of observation 6. Defect/Reason names
7. Columns with observation dates 8. Totals for each column 9. Totals for each row
10. Grand total of all columns or all rows

21
Suggestion
Managers and/or team members can do their part to help the data collector (s) do their job well by simply
showing an interest in the project. Ask the collector (s) how the project is working out. Show your
support – tell the data collector (s) you think it is most important to collect the information.
Above all – act on the data as quickly as possible.

Check Sheet
Keyboard Errors in class Assignment
March
Mistakes
1 2 3 Total
Centering 8
Spelling 23
Punctuation 40
Missed Paragraph 4
Wrong numbers 10
Wrong page numbers 4
Tables 13
Total 34 35 33 102

22
Production Process distribution tally-sheet

Production Process distribution tally-sheet

Product : All
Components Inner Section : Production section 3 Date _____________ Manager
Dimension Measurement
Range : ± 0.05 Section Head
Inspector’s name : ______________________ Leader

No. Measure Occurrence checks

ment
5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 Time
1 -0.07
2 -0.06
3 -0.05
4 -0.04 4
5 0.03 7
6 0.02 15
7 0.01 37
8 +0 45
9 +0.01 49
10 +0.02 31
11 +0.03 11
12 +0.04 1
13 +0.05
14 +0.06
15 +0.07
Description Total Production Figure 14,370 items Grand Total 200

23
Stratification

24
 Stratification

Purpose
Stratification breaks down data into meaningful categories or
classifications and plots each category. It also helps us to
analyse and interpret the data according to various sources.

Application
Scatter diagrams, Control charts, Pareto Analysis, and most of
the other visual presentations of data.
When the same products are made on several machines or by
several operators, the data if classified according to machine or
operator, presents a clear picture of the variation contributed by
different sources. This makes the task of process control easier.

25
Principle
Stratification is a method of identifying the source of variation of
gathered data, classifying the data according to various factors.
It is similar to taking stones, mud, sand and water in a bottle and
after shaking the bottle for a while if we allow the bottle to rest
for sometime, we will find layers of water, mud, stones and sand
getting separated and settled according to its heaviness. Similar
to this event, data when received from various sources are
mixed up. They give meaningful information only when the data
is separated according to its various sources. There is no
general principle as to which factor should be stratified, but it is
essential to check all the production conditions.

26
Only Notes

27
Only Notes

28
Cause and Effect Diagram
or
Fish Bone Diagram
or
Ishikawa Diagram

29
 Cause and Effect Diagram

Introduction
To improve processes, one must continuously strive to obtain
more information about processes and their output. Cause and
Effect diagram is a valuable tool to accomplish this purpose. In
1943, this tool was first introduced by Dr. Kaoru Ishikawa, at the
University of Tokyo. He used, this tool to explain how various
factors could be stored and related while explaining to a group
of engineers from Kawasaki Steel Works. Today, it is known as
Ishikawa diagram or Fishbone diagram (since the completed
diagram resembles the skeleton of a fish).

30
Purpose
Find and cure causes and NOT symptoms. It is a method for
analyzing process dispersion. It relates causes and effects. This
tool has been applied widely by people at all levels. Following
are the main objectives of application:
• Helps a team to focus on the problem.
• Snapshot of the collective knowledge.
• Consensus of team around the problem.
• Building foundation for resulting solution.
• Focuses on causes and not symptoms.
• Excellent means to facilitate brain-storming session.

31
 Types of Causes and Effect Diagram

• Dispersion analysis
• Production process classification
• Cause enumeration

32
 Format of a Cause and Effect Diagram

Main Cause Main Cause A Level 2

Cause
Level 3
Cause Level 1
Cause
Problem to
be resolved
(Effect)

Main Cause Main Cause

Above diagram depicts the basic format of Cause and Effect diagram.
There is a hierarchical relationship of the effect to the main causes and
their subsequent relationship to the sub-causes.

33
Steps towards constructing Cause and Effect Diagram
1. Select the most appropriate Cause and Effect format. From the 3
types of major formats explained later.
2. Generate the causes needed to build a Cause and Effect diagram
Use :
• Brainstorming – without previous preparation
• Checksheet – based on data collected prior to meeting
1. Construct the Cause and Effect diagram
a) Place the problem statement in a box on the right hand side of
the writing surface.
* Allow plenty of space. Use a flipchart sheet, or a large white board.
A paper surface is preferred since the final Cause & Effect Diagram
can be moved.

34
 Causes
“Bones”
(Major Cause categories)
Effect

Late pizza
deliveries on
Fridays &
Saturdays

Tip Make sure everyone agrees on the problem statement.

Include as much information as possible on the “what”,
“where”, “when”, and “how much” of the problem. Use data to
specify the problem.
b) Draw major cause categories or steps in the production or
service process. Connect them to the “backbone” of the
fishbone chart.

35
 Machinery /
Equipment People

Late pizza
deliveries on
Fridays &
Saturdays

Methods Materials

Illustration note: In a Process Classification Type format, replace the major “bone” categories with :
“order Taking”, “Preparation”, “Cooking”, and “Delivery”.
* Be flexible in the major cause “bones” that are used. In a Production Process
the traditional categories are:
Machines (equipment), Methods (how work is done), Materials (components or
raw materials), and people (the human element). In a service process the
traditional methods are : policies (higher-level decision rules), procedures
(steps in a task), plant (equipment and space), and people. In both types of
processes, Environment (buildings, logistics and space) and Measurement
(calibration and data collection) are frequently used. There is no perfect set or
number of categories. Make them fit the problem.

36
 Machinery / Equipment
People
Unreliable cars People don’t
Ovens Drivers show up Late pizza
too small get lost deliveries on
Poor handling Fridays &
of large Poor Run out of Saturdays
orders dispatching ingredients
Methods Materials

c) Place the brainstormed or data-based causes in the

appropriate category.
 In brainstorming, possible causes can be placed in a major cause
category as each is generated, or only after the entire list has been
created. Either works well but brainstorming the whole list first
maintains the creative flow of ideas without being constrained by the
major cause categories or where the ideas fit in each “bone”.
 Some causes seem to fit in more than one category. Ideally each
cause should be in only one category, but some of the “people”
causes may legitimately belong in two places. Place them in both
categories and see how they work out in the end.

37
Tip If ideas are slow in coming, use the major cause categories as
catalysts, eg. “What in ‘materials’ is causing . . . . .?”
d) Ask repeatedly of each cause listed on the “bones”, either”
H “Why does it happen?” For example, under “Run out of ingredients” this question
would lead to more basic causes such as “Inaccurate ordering”, “Poor use of
space”, and so on.

Dispersion Analysis Vs. Process Classification

The process for constructing the diagram are same however there are differences in their
application. For process classification method, one has to identify the flow of the process
requiring improvement. Afterwards, the key quality characteristics are to be listed at each of the
steps which require improvement.

Cause enumeration
The cause enumeration method involves brainstorming all possible causes and listing them in
the order they are offered. Once, the brainstorming has exhausted itself, the team begins the
process of grouping the causes as it did for the dispersion analysis diagram. The end result
looks alike.

38
Suggestion : Process can be enhanced dramatically by using affinity
diagram process.
Hints and Creative Uses
1. Consider the big picture
2. Facilitation
3. Review and embellishment
4. Broad bases participation
5. Working toward the desired result

39
 The Benefits and Weaknesses of each Cause-
and-Effect Diagram
Diagram type : Dispersion analysis
Key benefits :
1. Helps organize and relate factors.
2. Provides a structure for brainstorming.
3. Involves everyone.
4. It’s fun.

Potential drawbacks :
1. Might be difficult to facilitate if developed in true brain storming
fashion.
2. Might become very complex; requires dedication and patience.

40
Diagram type : Process classification

Key benefits :
1. Provides a solid sequential view of the process and the factors
that influence each step.

2. Might help determine functional ownership for the work to be done

in improvement.

Potential drawbacks :
1. It is sometimes difficult to identify or demonstrate
interrelationships.

41
Diagram type : Cause enumeration

Key benefits :
1. Easy to facilitate.

2. Provides in-depth list of all possible causes.

Potential drawbacks :
1. The added step of creating an affinity diagram might add time to
the process.

2. The final diagram might be difficult to draw because of the random

output of the brainstorming session.

42
 Cause and Effect Diagram for Wobbling

Workers Materials
Over
Experienced
Content
Experienced
Training
Personality
Under
Knowledge Experienced

Wobble
Measurement Off-centre
Adjustment Off-centre
Errors
G.Axle Cover
Measuring Training Cover hole
Tool Metal drill Axle stop
G. Axle bearing
Inspector F.Axle
Judgement uneven Cover
Cover
Plating
Experience
Punch width F.Cover

Inspection Tools

43
Scatter Diagram

44
 Scatter Diagram
Measuring relationships between variables

Why use it ?
To study and identify the possible relationship between the
changes observed in two different sets of variables.
What does it do ?
• Supplies the data to confirm a hypothesis that two variables
are related
• Provides both a visual and statistical means to test the
strength of a potential relationship
• Provides a good follow-up to a Cause & Effect Diagram to find
out if there is more than just a consensus connection between
causes and the effect.

45
How to do it ?
1. Collect 50-100 paired samples of data that you think may be
related and construct a data sheet

Average Average
Session Rating Experience of
Course (on a 1-5 scale) Training Team
(days)
1 4.2 220
2 3.7 270
3 4.3 270
* * *
* * *
* * *
40 3.9 625
46
Theory : There is a possible relationship between the number of days of experience
the training team has received and the ratings of course sessions.
2. Draw the horizontal (x axis) and vertical (y axis) lines of the diagram
* The measurement scales generally increase as you move up the vertical
axis and to the right on the horizontal axis.

4.5
4.4

4.3 Dependent variable (“effect”)

4.2

4.1
4.0

3.9 * Looking for relationships

3.8 not cause & effect

3.7 Independent variable (“cause”)

3.6

3.5

150 200 250 300 350 400 450 500 550 600 650

47
3. Plot the data on the diagram
* If values are repeated, circle that point as many times as appropriate.

Overall rating of the Session 4.5

4.4
4.3
4.2
4.0 y axis
4.1
430 x axis
4
3.9
3.8
3.7
3.6
3.5
3.4
100 150 200 250 300 350 400 450 500 550 600 650

Average Experience of Training Team (in days)

48
4. Interpret the data
* There are many levels of analysis that can be applied to scatter
Diagram data. Any basic statistical process control text, like Kaoru
Ishikawa’s Guide to Quality Control, describes additional correlation
tests. It is important to note that all of the examples in this chapter are
based on straight line correlations. There are a number of non linear
patterns that can be routinely encountered, eg. y = o, y = x. These types
of analysis are beyond the scope of this book.
* The following five illustrations show the various patterns and
meanings that Scatter Diagrams can have. The example used is the
training session assessment previously shown. The pattern have been
altered for illustrative purposes. Pattern # 3 is the actual sample.

Tip The Scatter Diagram does not predict cause and effect relationships.
It only shows the strength of the relationship between two variables. The
stronger the relationship, the greater the likelihood that change in one
variable will affect change in another variable.

49
 Scatter Diagrams

n = 30 r = 0.9 n = 30 r = 0.6 n = 30 r = 0.0

y y y
6 6 6

4 4 4

2 2 2

0 0 0
0 2 4 6 x 0 2 4 6 x 0 2 4 6 x

Fig 1. Positive Correlation Fig 2. Positive Correlation Fig 3. No Correlation

May Be Present

n = 30 r = 0.9 n = 30 r = 0.6 n = 30

y y y
6 6 6

4 4 4

2 2 2

0 0 0
0 2 4 6 x 0 2 4 6 x 0 2 4 6 x

Fig 4. Negative Correlation Fig 5. Negative Correlation Fig 6

May Be Present

50
Histogram

51
 Histogram – The Basics

A histogram may be described by its central tendency and its spread

or variation. The measures commonly used to describe central
tendency are
1. Mean : the sample values summed
and divided by the number observed; the
arithmetic average; the centre of gravity for the
distribution.
2. Median : the centre value when all
observed values are placed in the order of size; the
fiftieth percentile value; the value dividing the
distribution exactly in half.
3. Mode : the most frequently
occurring value; the highest bar in a histogram.

52
The measures commonly used to describe spread or variation are

1. Variance : the deviations from the

mean squared, summed, and divided by one less
than the sample size.
2. Standard : the positive square root
of the variance.
deviation
3. Range : the difference between
the largest and the smallest value in the sample.
For distributional analysis, the mean and the standard deviation are the
characteristics normally used. For the variables control chart, the mean
and range are used.
Using the algebraic symbol X to represent any single value, the
following equations are used for calculation:

53
 sum of sample values X
The mean = X = ———————————— = ———
Sample size n

The sum of the squared deviations from the mean

The variance = S2 = —————————————————————————
The sample size minus 1

 (X – X)2 X2 – (X)2 / n nX2 – (X)2

= —————— = ———————— = ———————
n–1 n–1 n (n – 1)

The standard deviation = s = The square root of the variance

X2 – (X)2 / n
= —————————
n–1
The range = R = Largest minus smallest value

= X L — XS

54
 Frequency Table
Class Class Boundaries Mid Frequency Frequency
No. Value Tally

1 3.275 — 3.325 3.3 3

2 3.325 — 3.375 3.35 3

3 3.375 — 3.425 3.4 9

4 3.425 — 3.475 3.45 32

5 3.475 — 3.525 3.5 38

6 3.525 — 3.575 3.55 10

7 3.575 — 3.625 3.6 3

8 3.625 — 3.675 3.65 1

9 3.675 — 3.725 3.7 1

55
 Histogram
40
35
30
25
20
15
10
5
0
0 3.3 3.35 3.4 3.45 3.5 3.55 3.6 3.65 3.7

Class Boundaries (Mid-Values)

56
 a) General type b) Comb type

c) Positively skew type d) Left–hand precipice type

57
e) Plateau type f) Twin–peak type

g) Isolated–peak type

58
 A Procedure for Constructing a Histogram

A histogram is constructed to show the general variation pattern inherent

in data. The shape can be altered by varying the class size and class width.
To have some consistency in construction, the following guidelines are
recommended.
Preconstruction Steps
1. Determine the range of the data by subtracting the smallest
observed measurement from the largest and designate it as R.
Example 1
Largest observed measurement = 1.1185 inches
Smallest observed measurement = 1.1030 inches
R = 1.1185 inches – 1.1030 inches = 0.0155 inch

59
2. Record the measurement unit (MU) used. This is usually controlled by the measuring
instrument least count.
Example 1
MU = 0.0001 inch
3. Determine the number of classes and the class width.
a. The number of classes, K, is dependent on the total measurements, n.
b. Obtain the preferred number of classes, K, from the graph show in the Fig.
Example 1
n = 50
k = 7.7 or 8 classes
c. Calculate the class width by dividing the range, R, by the preferred number of
classes, k.
R 0.0155 inch
–– = ––––––––––– = 0.0019375 inch
k 8
d. The class width selected should be an odd-numbered multiple of the
measurement unit, MU. This value should preferably be close to the class width
calculated in step 3c.
Example 1
MU = 0.0001 inch
Class width = 0.0019 inch or 0.0021 inch (odd-numbered multiple is 19 or 21,
respectively)

60
 20

18
n : Number of
16 Measurements

14

K 12

10

8 Fig : Graph to
determine k
6

0
1 2 3 4 5 678 10 20 50 200 500
1,000
n

61
4. Establish the class midpoints and class limits.
a. The first class midpoint should be located near the largest
observed measurement. If possible, it should also be a convenient
increment.
Example 1
First class midpoint = 1.1185 inches
b. Always make the class widths equal in size.
c. Express the class limits in terms which are one-half unit beyond
the accuracy of the original measurement unit. This avoids plotting
an observed measurement on a class limit. If the procedure under
step 3 is followed correctly, this this will be done automatically .
Example 1
First class midpoint = 1.1185 inches

62
The class width is 0.0019 inch. Limits for the first class would be
1.1185 ± 0.0019
2
1.1185 ± 0.00095
Midpoints and class limits for the entire data set are
Class Limits
Midpoints Lower Upper
1.1185 1.11755 1.11945
1.1166 1.11565 1.11755
1.1147 1.11375 1.11565
1.1128 1.11185 1.11375
1.1109 1.10995 1.11185
1.1090 1.10805 1.10995
1.1071 1.10615 1.10805
1.1052 1.10425 1.10615
1.1033 1.10235 1.10425

63
This results in one more class than originally determined in step 3b (eight) due to dropping
0.0000375 inch. Using 0.0021 inch would give the
following:
Class Limits
Midpoints Lower Upper
1.1185 1.11745 1.11955
1.1164 1.11535 1.11745
1.1143 1.11325 1.11535
1.1122 1.11115 1.11325
1.1101 1.10905 1.11115
1.1080 1.10695 1.10905
1.1059 1.10485 1.10695
1.1038 1.10275 1.10485

64
 Constructing the Histogram

1. Determine the axes for the graph.

a. The frequency scale on the vertical axis should slightly exceed the
largest class frequency.
b. The measurement scale along the horizontal axis should be at
regular intervals which are independent on the class width.

2. Draw the graph.

a. Mark off classes.
b. Draw rectangles with heights corresponding to the measurement
frequencies in that class.

3. Title the histogram.

a. Give an overall title.
b. Identify each axis.

65
 Analysis with the Histogram
After constructing the histogram, draw in the specifications. It will then quickly
answer three questions:
 Is the product consistent with previous checks ?
 Is the product well centred ?
 Does it appear that the product is meeting engineering specifications ?
Histograms often reveal, without elaborate analysis, much information about the
process or product under observation and, because they are easily understood by
operating personnel, can materially aid in making improvements. The histogram may
be used for such purposes as
 Assessing material strengths
 Evaluating processes
 Indicating the necessity for corrective action
 Measuring the effects of corrective action
 Determining machine capabilities
 Portraying life characteristics
 Comparing operators
 Comparing materials
 Comparing vendors
 Comparing products
66
 Analysis with the Histogram (Contd.)

The histogram is helpful in making an analysis, but its use is limited because
 It requires many measurements
 It does not take time into consideration
 It does not separate the two kinds of variation-variation due to chance
and variation due to assignable factors
 It does not show trends
Essentially, a histogram is a postmortem picture in that it portrays a situation that has
already occurred. For illustrating variation patterns or for indicating shift or improvement,
the histogram is unexcelled. But one precaution should be observed; a histogram will hide
more than it reveals if control is not exercised.

67
 Interpretation of Histograms
Minimum Maximum Minimum Maximum
Limit Limit Limit Limit
1 4

2 5

3 6

68
 Control Chart
1. What is a Control Chart ?
The quality of a product manufactured in a process is inevitably
accompanied by dispersion. Various Causes of such dispersion exist and
they can be classified into the following two types.
1 Chance
Dispersion by chance is unavoidable and inevitably occurs in a
process, even if the operation is carried out using standardized raw
materials and methods. It is impossible to avoid the chance cause
technically and economically at present.
2 Assignable
Dispersion from an assignable cause is unusual and meaningful in that
it is avoidable and cannot be overlooked eg. Neglecting various
standards or application of improper standards.
In order to control a process it is necessary to eliminate assignable
causes and take action to prevent their recurrence, while tolerating
dispersion by chance.

69
A control chart was first used in 1924 by W.A. Shewhart, who belonged
to the Bell Telephone Laboratories, with a view to classifying an
abnormal process by distinguishing variations due to chance from
those due to assignable causes. A process chart consists of a central
line, a pair of control limits, one each, allocated at the top and the
bottom of the chart and characteristic values plotted on the chart
which represent the states of a process. If all these values occur on a
smooth line within the control limits, the process is regarded as being
in the controlled state. However, if they occur outside the control limits
or are arranged in a peculiar form, the process is judged to be in the
abnormal state. An example is shown in the figure.

Fig. Examples of Control Charts

Upper control
Limit
Central line

Lower control
Limit
Control chart for controlled state

70
 Control chart for uncontrolled state

To make a control chart it is necessary to classify the process by type

of raw materials, machine and line and further to classify these data
into small groups such as time or shift.
There are various types of control chart, according to the
characteristic values or purpose. However, in any type of control chart
the control limit is calculated by the formula :
(average value) ± 3 x (standard deviation)
Therefore such a chart is called a three sigma control chart.
2. Type of Control Chart
Control charts vary according to the indiscrete value and the discrete
values. The types of chart prescribed by JIS are shown in the table 1 and
it’s control limit line is shown in table 2

71
Table 1 Types of Control Chart

Characteristic Name
Value

Indiscreet Value X – R chart (average value and range)

X chart (measured value)

Discreet value pn chart (number of defective units)

p chart (fraction defective)
c chart (number of defects)
u chart (number of defects per unit)

72
73
74