QUALITY IMPROVEMENT

TOOLS
 QUALITY IMPROVEMENT
BASIC TOOLS
1. TALLY SHEET
2. HISTOGRAM
3. PARETO DIAGRAM
4. CAUSE AND EFFECT DIAGRAM
5. SCATTER DIAGRAM
6. SRATIFICATION
7. CONTROL CHART
1. TALLY SHEET
 THE TALLY SHEET
SIMPLE REGISTRATION OF DATA AT THE
WORK STATION -
Frequently it makes sense to register data
immediately at the work station in a quick
and simple way. It is a list in which stripes
have to be filled in hence it is called tally.
The test data or defects are filled in the
related column in a preprinted form
immediately after the test. The observed
or measured value need not be written
down but can be indicated by a stripe/line.
 TALLY SHEET no.
Date:
Type: manufacturer:
test point: place
number checked: checked by:
total number: remark:
code no:

Totals

1 2 6 13 10 16 20 17 12 16 20 17 11 8 5 6 2 1

Specification limit
Specification limit

15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
 Check Sheet
Product Date:

Manufacturing stage: final insp. Section:

Type of defect: scar. Incomplete Inspector’s name
crack, misshapen
Lot no.
Total no. inspected: 1525 Order no:
Remarks all items inspected

Type Check Subtotal

Surface scars 17
Cracks 11
In complete 26
Misshapen 3
Others 5
Total: 62

Total rejects 42
2.HISTOGRAM
 Histogram
When used in process capability studies,
histograms can display specification
limits to show what portion of the data
does not meet the specifications.

The histogram plots data in a frequency

distribution table. What distinguishes
the histogram from a check sheet is that
its data are grouped into rows so that
the identity of individual values is
removed.
 Histogram

After the raw data are collected, they

are grouped in value and frequency and
plotted in graphical form.
A histogram’s shape shows the nature
of the distribution of the data, as well
as central tendency and variability.
Specification limits can be used to
display the capability of the process.
 Histogram

Commonly used to present quality

improvement data, histograms work
best with small amounts of data that
vary considerably.
The histogram is a bar chart
showing a distribution of variables
Histogram
 Product indication: METAL BLOCK XB2 Date : 6.12.1982
Specification: 3.45+ 0.15mm date: 6.12.2000 Production stage
Production stage Checked
Checked by
Number checked (n): 100
Test values X max X min

3.36 3.46 3.48 3.50 3.42 3.43 3.51 3.48 344 3.50 3.56 3.42

3.48 3.56 3.50 3.32 3.47 3.40 3.46 3.56 3.54 3.30 3.56 3.30

3.41 3.37 3.47 3.47 3.45 3.44 3.50 3.47 3.46 3.46 3.50 3.37

3.55 3.52 3.44 3.50 3.45 3.49 3.48 3.46 3.52 3.46 3.55 3.44

3.48 3.48 3.32 3.40 3.52 3.34 3.46 3.43 3.30 3.46 3.52 3.30

3.59 3.63 3.59 3.47 3.38 3.52 3.45 3.42 3.31 3.46 3.63 3.31

3.40 3.54 3.46 3.51 3.40 3.50 3.68 3.60 3.46 3.52 3.68 3.40

3.48 3.50 3.56 3.50 3.52 3.46 3.49 3.46 3.52 3.56 3.56 3.46
3.52 3.48 3.46 3.45 3.46 3.54 3.51 3.48 3.49 3.41 3.54 3.41
3.41 3.45 3.34 3.44 3.47 3.40 3.46 3.48 3.54 3.47 3.54 3.34
 DIVISION IN CLASSES

Class Class Class

number limits mid-marks

1 3.275 - 3.325 3.30

2 3.325 - 3.375 3.35
3 3.375 - 3.425 3.40
4 3.425 - 3.475 3.45
5 3.475 - 3.525 3.50
6 3.525 - 3.575 3.55
7 3.575 - 3.625 3.60
8 3.625 - 3.675 3.65
9 3.675 - 3.725 3.70
 Frequency table

Class Class Class Number Total

limits mid-marks
number n = 100

1 3.275 - 3.325 3.30 3

2 3.325 - 3.375 3.35 3
3 3.375 - 3.425 3.40 9
4 3.425 - 3.475 3.45 32
5 3.475 - 3.525 3.50 38

6 3.525 - 3.575 3.55 10

7 3.575 - 3.625 3.60 3
8 3.625 - 3.675 3.65 1
9 3.675 - 3.725 3.70 1
 Specified tolerance
3.30 - 3.60
40

30

20

10

0
3.30 3.40 3.50 3.60 3.70
Thickness in mm

Specification value 3.45±0.15

HOW TO READ A HISTOGRAM
Histogram with an
isolated part.
Examine cause of this
irregularity.
E.g. Mixed
products of
different
machines,
different
Histogram

Histogram with an
interruption. Possibly
errors have been made
during measurement of
the preparation of the
histogram.

Check the measuring

equipment the reading
methods or change the
number of classes.
Histogram

Histogram with two

peaks.
This shows products
from two manufacturing
processes.

Study the subdivision of

the groups and make a
histogram per group.
Histogram

Distribution cut off at

the left.
Products have been
selected before hand.
COMPARISON WITH THE
TOLERANCE LIMITS
Spec limits Range (R) of the
R measured values
exactly in the
middle.

In this situation no
risk of rejects, due
to small variations
in the process.
 Histogram

Spec limits Although R is within

R the tolerance the
mean of the process
is too close to the
lower limit. A minor
change in the
production process
will result in
increased rejects.
 Histogram

Spec limits/ R R and the tolerance

are exactly the
same.

This situation
offers no room for a
change in the
process i.e. any
change in the
 Histogram

Spec limits The tolerance is

R wider than R.

It is possible to
reduce the specified
tolerance or the
accuracy of the
process.
 Histogram

Spec limits
The process mean
R
is too far to the left
and must be shifted
to the nominal
value.
R
The distribution is
too wide.
Change the
process to limit
distribution or
extend tolerance.
3.PARETO DIAGRAM
 PARETO DIAGRAM

 The Pareto diagram is named after

Vilfredo Pareto, a 19th century Italian
economist who postulated that a large
share of wealth is owned by a small
percentage of the population.

 This basic principle translates well into

quality problems–most quality problems
result from a small number of causes.
 PARETO DIAGRAM

 Quality experts often refer to the

principle as the 80/20 rule;

 that is 80% of the problems are

caused by 20% of the potential
sources.
 PARETO DIAGRAM

 A Pareto diagram puts data in a

hierarchical order which allows the most
significant problems to be corrected first.
 The Pareto analysis technique is used
primarily to identify and evaluate
nonconformities, although it can
summarize all types of data.
 It is perhaps the diagram most often
used in management presentations.
 PARETO DIAGRAM

To create a Pareto diagram, the

operator collects random data,
regroups the categories in order of
frequency, and creates a bar graph
based on the results.
 PARETO DIAGRAM

 This tool is a graphical picture of

the most frequent causes of a
particular problem.

 It shows where to put your initial

effort to get the most gain.
Sales Department Loss
(In millions of Rupees)
A 5
B 2.5
C 10
D 20
E 45
F 1.5
G 1
H 1.5
Others 0.5

Total 87.0
 Loss Percentage Cumulative
Sales Department
(In millions of %
of Rupees) total
E 5 51.7 51.7

D 2.5 23.0 74.7

C 11.5 86.2
10

A 20 5.7 91.9

B 45 2.9 94.8

F 1.5 1.7 96.5

H 1 1.7 98.2

G 1.5 1.2 99.4

Others 0.5 0.6 100

Total 87.0 100

 PARETO DIAGRAM

90 100%

80

70
Jan - May 19XX
60

50

40 50%

30

20

10
0
E D C A B F H G Div
 PARETO DIAGRAM PROCEDURE
 Make a list of possible defects or problems
according to cause, phenomenon, cost price,
etc.
 Collect these data over a certain length of time.
 Express the data in the correct units.
 Classify the groups in descending order of
Occurrence and draw the bar diagram.
 Draw a broken line indicating the cumulative
%, starting with the right hand top corner of
the first bar.
PARETO DIAGRAM
 200 100
Number of units investigated
180 90

160 80
Number of defective units

140 70

Cumulative percentage
120 60
PARETO DIAGRAM
100 50

80 40

60 30

40 20
A:Crack
20 10
B: Scratch
C: Stain
D: Strain
D B F A C E Others E: Gap
F: Pinhole
4.CAUSE AND EFFECT
DIAGRAM
 CAUSE AND EFFECT DIAGRAM

 A cause and effect diagram describes a

relationship between variables. The
undesirable outcome is shown as effect,
& related causes are shown as leading
to or potentially leading to said effect.

 This tool has one severe limitation,

however, in that users can overlook
important, complex interactions
between causes.
 CAUSE AND EFFECT DIAGRAM

 Thus if a problem is caused by a

combination of factors, it is difficult to
use this tool to depict and solve it.

 A fish bone diagram displays all

contributing factors and their
relationships to the outcome to identify
areas where data should be collected
and analyzed.
 Steps
 Determine the quality characteristic.
 Write the chosen quality characteristic on the
RH side of a sheet of paper & enclose the
characteristic in a square.
 Draw the backbone line from left to right.
 Primary causes which directly affect the quality
characteristic be drawn as big bones.
 Secondary causes are linked by medium sized
bones.
 Tertiary causes comes small bones.
 Assign an importance to each factor and mark
particularly important factors that seem to have
a significant affect on the quality characteristic.

 Procedure
 Decide on the quality characteristic.
 Find as many causes (which may affect
the quality characteristic) as possible.
 Sort out the relations among the causes
and make a C & E diagram by connecting
those elements.
 Assign an importance no. to each factors
and mark the particularly important
factors which seem to have significant
effect on the quality characteristic.
 Write in any necessary information.
 CAUSE AND EFFECT DIAGRAM

 The major areas of potential causes

are shown in the main bones, e.g.,
Materials, Methods, Men,
Measurements, Machines.
 Later, the sub-areas are
depicted. Thorough analysis of each
cause can eliminate causes one by one,
and the most probable root cause can
be selected for corrective action.
 Cause And Effect Diagram
 Quantitative information can also be used to
prioritize means for improvement whether it be
machine, design, or operator.

 Its most frequent use is to list the cause

of particular problems. The lines coming
off the core horizontal line are the main
causes and the lines coming off those are
sub causes.
 Solder Bath Conveyor Storage
Speed
Solder
Composition
Dust

Wave Height
Moisture
Temperature Humidity

Dross Inclination Quality

of wave
soldering
Baking Clinched
Type Lead length
PTH
Specific Quality Lead
Gravity Clearance Forma-
between tion/
Foaming lead dia & stress
Pre heat Action/ level hole dia
Flux
Temperature Solder Mask Tinning
type
PCB components
 PR 4 R&D New product Info
Production & Despatch Sales & Marketing
Emplo- (define frequency)
Timely Quality Nil failure at the Courtesy calls
yees time of installations
morale Reliable Delivery (define frequency
Product Minimal failure due
upgradable time feed back
Low product Qlty partnership
Development Cost effective
Attrition Define delivery time Buy back offers
(<10%) Meet customers for higher end m/c & Customer req. v/s
need for lower end m/c Suppliers Spec.
Career Prod Spec Right pricing
Development Transit damages Nil transit damages
Define response to Timely collections
Performance this Satisfied customer
management Introduction training
system AMC Renewal and
Product training
collection target
Invoicing to be met
Component level Accurate &
Training &
Supports all development training timely Response
Define for major
functions billing and time
Personality A/C, for metro and
collection
development remote.
Customer relation skills
Job/Repair Time
Functional training Define

Cust.Care Training Commercial & Accounts Customer support

 Health Spirit Encouragement
Relaxation Pride
Patience Fighting spirit
Rest
Carefulness Devotion
Amusement
Concentration Calmness
Time

Sleep Composure

Confidence
Depth
Defeat in a sports
Information power match
Theory Schedule
Motion
Study of Exercise
Planning
opponent Speed Quality
Analysis Cooperation Quantity
Rules
Observation Advice
Teamwork Form
Common Judgement of Function
sense situation Model Repetition
Strategy Experiences in matches Technique
 MACHINES Parts wearing
Poor alignments
Old
Wrong spare Machine
OPERATORS parts breakdown
Poor maintenance
Unaware
Shortage
Breakdown not reported
Poor training
POOR QUALITY
OF GLAZES
Variable quality
Wrong time
Wrong temperature
Cheaper brand
Incorrect settings
Damp weather
(storage) METHODS

MATERIALS
 MAN MACHINE
Dirty hands
Maintenance
Possible other
adjustment
Causes
Direction
Holes in gloves

EFFECT

measurement flatness
Description/procedures

calibration hardness

METHOD
MATERIAL

CAUSE AND EFFECT DIAGRAM

CAUSE AND EFFECT DIAGRAM
5. SCATTER DIAGRAM
 SCATTER DIAGRAM

 A scatter diagram shows how two

variables are related and is thus used to
test for cause and effect relationships.It
cannot prove that one variable causes
the change in the other, only that a
relationship exists and how strong it
is. In a scatter diagram, the horizontal
(x) axis represents the measurement
values of one variable, and the vertical
(y) axis represents the measurements
of the second variable.
 SCATTER DIAGRAM

 FIGURE shows part clearance

values on the x-axis and the
corresponding quantitative
measurement values on the y-axis.
 SCATTER DIAGRAM

 The scatter diagram shows the

pattern of relationship between
two variables that are thought
to be related.
 The closer the points hug a
diagonal
line the more closely there is a one
to one relationship.
 Y
cm SCATTER DIAGRAM

160

150

140

130

120
X
150 160 170 180 190 cm
 10
SCATTER DIAGRAM
9
Machine accuracy (% defective readings)
8

1 2 3 4 5 6 7

Height of water (metres)

 TYPES OF SCATTER
DIAGRAMS
Y

Y increases as X increases
Strong positive correlation

X
 SCATTER DIAGRAM
Y

Y may increase as X increases

weak positive correlation

X
 SCATTER DIAGRAM
Y

No relation between X and Y

No correlation

X
 SCATTER DIAGRAM
Y
Y may decrease as X
increases
weak negative correlation

X
 SCATTER DIAGRAM
Y

Y decreases as X
increases
Strong negative
correlation
X
ANALYSING SCATTER DIAGRAM

A scatter diagram should be analysed

carefully. Although it is possible to express
correlation in the first instance, complicated
calculations are not required. Sometimes the
correlation between two variables is so
evident that certain conclusions can be drawn
immediately. If the correlation is less
obvious, it will be necessary to analyse the
diagram carefully. In these cases there may
be other causes.
60

55

SCATTER DIAGRAM
50

45

850 900
 SCATTER DIAGRAM

Scatter diagram showing the relation

between the hardening temperature
and the hardness of a certain
component. There is an apparent
correlation whereby the hardness
assumes higher values if the hardening
temperature increases.
 In other words the hardness can be
controlled by controlling the hardening
temperature.
The scatter diagram enables us to visualize
this relation.
The scatter diagram becomes
particularly interesting when the
properties of a product can only be
established by destroying the product
( destructive testing).
If a substituting property can be found
showing a correlation with the desired
property, without the need to destroy the
product the method offers clear
advantages.
Example determining the correlation by
means of a scatter diagram.
By means of scatter diagrams we would
like to find out whether there is a
correlation between the heights of fathers
and those of their children.
 PROCEDURE
Count the total number of observation in
the graph.
Draw horizontal and vertical lines dividing
the number of points in two equal halves
across top/bottom and left and right resp.
This yields four sectors I, II, III and IV and
thus four groups of observations n1, n2, n3
and n4. Points lying on the center lines are
either distributed randomly or neglected.
SCATTER DIAGRAM
6. STRATIFICATION
 Types of stratification
By material
Consider variations in manufacturer, brand, size of consignment,
storage item and conditions.
By machine
look for variations in machine type, make, model, age position.
By operator
consider factors such as experiences, age, training, gender.
By inspection method
Classify by instruments used, inspection
procedure, inspector.
By time
Are the data affected by time of day, season,
position, in the operation lifecycle?
By weather
Look for changes in humidity and temperature.
Others
Look for significant strata all along the
production chain.
 Manufacturer, buyer, brand, place of production, purchase
1. By material
date, lot received, production lot, components, purity, size,
parts, time stored, storage place, etc.

2. By machine, Machine type , number, model, performance, and age: by

equipment, or tool factory, line, tool, and die.

3. By operator Individual, team, group, age, experiences, gender, etc.

4. By operating procedure Temperature, pressure, speed, rational frequency, line

and by operating speed,location of operation, illumination, air temperature,
conditions humidity,weather, operating procedure, etc.

5. By measurement and Instrument, measurement procedure, place of measurement.

inspection Person making the measurement, inspection tools, inspection
procedures, place of inspection, inspector, etc.

Time, morning, afternoon, night, day, week, month, period

6. By time
season, etc: just before starting and just after finishing the
operation.
7. By environment Air temperature, humidity, clear, cloudy, rainy, windy,
and weather snowy, rainy season, dry season, sound, illumination, etc.

New vs.. previous product, unit product vs.. continuously

8. Others
produced product, good product vs.. defective product,
packing method and transportation method etc.
7.Control Charts
 Control Charts

 A control chart displays statistically

determined upper and lower limits
drawn on either side of a process
average.
 This chart shows if the collected
data are within upper and lower
limits previously determined
through statistical calculations of
raw data from earlier trials.
 Control Charts

 The construction of a control chart

is based on statistical principles
and statistical distributions,
namely normal Distribution.

 When used in conjunction with a

manufacturing process, such
charts can indicate trends and
signal when a process is out of
control.
 Control Charts

The centerline of a control chart represents

an estimate of the process mean; the
upper and lower critical limits are also
indicated.
The process results are monitored over time
and should remain within the control
limits; if they do not, an investigation is
conducted for the causes and corrective
action is taken.
A control chart helps determine variability
so it can be reduced as much as is
economically justifiable.
 Control Charts

 In preparing a control chart, the mean

upper control limit (UCL) and the lower
control limit (LCL) of an approved
process and its data are calculated. A
blank control chart with mean UCL and
LCL with no data points is created; data
points are added as they are
statistically calculated from raw data.
 Control Charts
 The control chart is a line chart with
control limits. By mathematically
constructing control limits at 3 standard
deviations above and below the average,
one can determine what variation is due to
normal ongoing causes (common
causes)and what variation is produced by
unique events (special causes).

 By eliminating the special causes first and

then reducing common causes, quality can
be improved.
 CONTROL CHART

8.45

8.41 Control limit

8.37

Days
1 2 3 4 5 6 7 8 9 10 11 12

Mr. A walks faster because it is past 8.45

 Control Chart

Abnormal

Normal
v

Abnormal
 Control Chart

(a) Seven consecutive points

consecutive

Center line

Six consecutive points

(b)
Bias to
one side

Center line

Ten of eleven consecutive points on one side

of the center line

Control Chart
(c)
trend
Falling trend in a wave
Six consecutive raising point

Control Chart
 (d) Two of three consecutive points
Approaching the limits

JCL (3s lines)

Two of three consecutive points

Control Chart
 (c)
Periodicity

Period 1 wk 1 wk 1 wk

M T W R F S M T W R F S M TWR F SMT W F FSMTW

Control Chart
 Control Chart

a small wave varies the large periodic wave

Control Chart
QUALITY IMPROVEMENT
BASIC TOOLS - SUMMARY
1. TALLY SHEET
2. HISTOGRAM
3. PARETO DIAGRAM
4. CAUSE AND EFFECT DIAGRAM
5. SCATTER DIAGRAM
6. SRATIFICATION
7. CONTROL CHART
 APPLICATIONS SUMMARY
ACTIVITY TOOL TO BE USED

1 SEGGREGATING DEFECTS, PARETO

PRIOROTISING
2 IDENTIFYING EFFECTS & THEIR C&E DIAGRAM
ROOT CAUSES
3 COMPILING DATA & ANALYSIS CHECK SHEETS
FOR DECISION MAKING
4 STUDY PATTERN OF HISTOGRAM
OBSERVATIONS & PROVIDE
CONTROL LIMITS
5 STUDY CONSISTENCY & QUALITY CONTROL CHARTS
OF PROCESSES
6 TESTING HYPOTHETICAL SCATTER DIAGRAM
RELATIONSHIP BETWEEN
VARIABLES ORCHARECTERISTICS
7 ASCERTAIN DIFFRENCES STRATIFICATION
DIAGRAM