SPC – Attribute Control Charts

10-2 Quality Control

Basics of Statistical
Process Control
 Statistical Process Control
(SPC)
 monitoring production process to
detect and prevent poor quality UCL
 Sample
 subset of items produced to use for
inspection
 Control Charts LCL

 process is within statistical control

limits

Copyright 2006 John Wiley & Sons, Inc. 4-2

10-3 Quality Control

Variability

 Random  Non-Random
 common causes  special causes
 inherent in a process  due to identifiable factors
 can be eliminated only  can be modified through
through improvements in operator or management
the system action

Copyright 2006 John Wiley & Sons, Inc. 4-3

10-4 Quality Control

SPC in TQM

 SPC
 tool for identifying problems and
make improvements
 contributes to the TQM goal of
continuous improvements

Copyright 2006 John Wiley & Sons, Inc. 4-4

10-5 Quality Control

 Attribute Quality Measures

 a product characteristic that can be evaluated with a
discrete response
 good – bad; yes - no
 Variable
 a product characteristic that is continuous and can
be measured
 weight - length

Copyright 2006 John Wiley & Sons, Inc. 4-5

10-6 Quality Control

Applying SPC to Service

 Nature of defect is different in services

 Service defect is a failure to meet customer
requirements
 Monitor times, customer satisfaction

Copyright 2006 John Wiley & Sons, Inc. 4-6

10-7 Quality Control

 Hospitals
Applying SPC to
timeliness andService
 (cont.)
quickness of care, staff responses to requests, accuracy of lab
tests, cleanliness, courtesy, accuracy of paperwork, speed of admittance and
checkouts
 Grocery Stores
 waiting time to check out, frequency of out-of-stock items, quality of food
items, cleanliness, customer complaints, checkout register errors
 Airlines
 flight delays, lost luggage and luggage handling, waiting time at ticket
counters and check-in, agent and flight attendant courtesy, accurate flight
information, passenger cabin cleanliness and maintenance

Copyright 2006 John Wiley & Sons, Inc. 4-7

10-8 Quality Control

Applying SPC to
Service (cont.)
 Fast-Food Restaurants
 waiting time for service, customer complaints, cleanliness,
food quality, order accuracy, employee courtesy
 Catalogue-Order Companies
 order accuracy, operator knowledge and courtesy,
packaging, delivery time, phone order waiting time
 Insurance Companies
 billing accuracy, timeliness of claims processing, agent
availability and response time

Copyright 2006 John Wiley & Sons, Inc. 4-8

10-9 Quality Control

Where to Use Control Charts

 Process has a tendency to go out of control

 Process is particularly harmful and costly if it goes
out of control
 Examples
 at the beginning of a process because it is a waste of time and
money to begin production process with bad supplies
 before a costly or irreversible point, after which product is
difficult to rework or correct
 before and after assembly or painting operations that might
cover defects
 before the outgoing final product or service is delivered

Copyright 2006 John Wiley & Sons, Inc. 4-9

10-10 Quality Control

Control Charts
 A graph that establishes control
limits of a process
 Control limits
 upper and lower bands of a
control chart
 Types of charts
 Attributes
 p-chart
 c-chart

 Variables
 range (R-chart)
 mean (x bar – chart)

Copyright 2006 John Wiley & Sons, Inc. 4-10

10-11 Quality Control

Process Control
Chart
Out of control
Upper
control
limit

Process
average

Lower
control
limit

1 2 3 4 5 6 7 8 9 10
Sample number
Copyright 2006 John Wiley & Sons, Inc. 4-11
10-12 Quality Control

Normal Distribution

95%
99.74%
-3 -2 -1 =0 1 2 3

Copyright 2006 John Wiley & Sons, Inc. 4-12

10-13 Quality Control

A Process Is in
Control If …

1. … no sample points outside limits

2. … most points near process average
3. … about equal number of points above
and below centerline
4. … points appear randomly distributed

Copyright 2006 John Wiley & Sons, Inc. 4-13

 Types of Control Charts
 Attribute charts
 Monitor fraction of defective units
 Monitor number of defects

 Difference between “defective unit” and a

“defect?”
 A defective unit is a unit that is either defective.
 A defect is flaw on a given unit of a product.
 A unit can have many defects.
 A defective unit may be defined as, for example, a unit
that has 3 or more defects.
 Types of Control Charts
 Variables charts
 Monitors continuous quality characteristics.
 Continuous values (variables data) can
theoretically assume an infinite number of
values in some interval.
 Time
 Weight
 Ounces
 Diameter
10-16 Quality Control
 Types of Control

Control Chart Monitors

Attribute control charts
p chart Process fraction defective
c chart number of defects
u chart defects per unit
Variables control charts
X-bar chart Process mean
R chart (Range Chart) Process variability
 Use of p-Charts
 When observations can be placed into
two categories.
 Good or bad
 Pass or fail
 Operate or doesn’t operate
 p-chart
 A company that makes light bulbs wants to
monitor the fraction of defective bulbs.
 The company decides to select a random
sample of 100 bulbs in each day over a 5
day period
 Each of the 100 bulbs are tested to
determine if they light up.
 p-chart
 If a bulb does not light up, the bulb is
defective.
 The company wants to
 Estimate the percentage of defective bulbs and
 Determine if the percentage of defective bulbs
is increasing over time.
 A p chart is the appropriate tool for providing
the company with this information.
 Notation
 Sample size = n = 100
 Number of samples (subgroups) = k = 5
 X = number of defective bulbs in a sample
 p = sample fraction defective = ???
 p-bar = estimated process fraction defective
 P = process fraction defective (unknown)
 p-bar is an estimate of P
Inspection Results

Day n X
1 100 20
2 100 5
3 100 30
4 100 35
5 100 24
 Compute p and p-bar

Day n X p=X/n
1 100 20 0.20
2 100 5 0.05
3 100 30 0.30
4 100 35 0.35
5 100 24 0.24
Sum 1.14
p-bar 0.23
 p-bar
(Estimated Process Fraction Defective)

p - bar  p 
 p
k

p
 p 1.14
  0.23
k 5
 p-Chart Control Limits

p (1  p )
UCL  p  3
n

p (1  p )
LCL  p  3
n
 p-Chart - Control Limits

.23(1  .23)
UCL  .23  3  .2426
100
.23(1  .23)
LCL  .23  3  .2173
100
 p-Chart for Bulbs

0.4

0.3 LCL
UCL
0.2
p

p-bar
0.1 p

0
1 2 3 4 5
Day
 Interpretation
 The estimated fraction of defective bulbs
produced is .23.
 On Day 2, p was below the LCL.
 This means that a special cause occurred on
that day to cause the process to go out of
control.
 The special cause shifted the process fraction
defective downward.
 This special cause was therefore favorable
and should be ???
 Interpretation
 After Day 2, the special cause lost its impact
because on Day 4, the process appears to be
back in control and at old fraction defective of .
23.
 Until the special cause is identified and made
part of the process, the process will be
unstable and unpredictable.
 It is therefore impossible to obtain a statistical
valid estimate of the process fraction defective
because it can change from day to day.
 Trend Within Control Limits
Process fractions defective
is shifting (trending) upward

P = process
fraction
P defective
P
Sampling P
Distribution P

UCL

p-Chart
LCL
 Applications
 Think of an application of a p-chart in:
 Sales
 Shipping department

 Law
 Use of c-Charts
 When we are interested in monitoring
number of defects on a given unit of
product or service.
 Scratches, chips, dents on an airplane wing
 Errors on an invoice
 Pot holes on a 5-mile section of highway
 Complaints received per day
 Opportunity for a defect must be infinite.
 Probability of a defect on any one location
or any one point in time must be small.
 c-Chart
c-chart notation:
c = number of defects
k = number of samples
c - bar  c  estimated mean number of defects
 c-Chart
 A car company wants to monitor the number
of paint defects on a certain new model of one
of its cars.
• Each day one car in inspected.
• The results after 5 days are shown on the
next slide.
 c-Chart

Day c c-bar LCL UCL

1 5 6.6 0 14.307
2 2 6.6 0 14.307
3 8 6.6 0 14.307
4 7 6.6 0 14.307
5 11 6.6 0 14.307
Sum c 33
c-bar 6.60
c-Chart - Mean

c
 c
k

c
 c 33
  6.6
k 5
c-Chart – Control Limits

LCL  c  3 c
UCL  c  3 c
c-Chart – Control Limits
LCL  c  3 c

 6.6  3 6.6

 1.107 or 0

UCL  c  3 c

 6.6  3 6.6

 14.307
 c-Chart for Number of Paint Defects

16
14
c, number of defects

12
c
10
LCL
8
UCL
6
c-bar
4
2
0
1 2 3 4 5
Car
 Conclusion
 Process shows upward trend.
 Even though trend is within the control limits,
the process is out of control.
 Mean is shifting upward
 This is due to an unfavorable special cause.
 Must identify special cause and eliminate it
from process.
 Who is responsible for finding and eliminating
special cause?
 Mini Case
 Think of an application of a c-chart bank.
 u-Chart
 With a c chart, the sample size is one
unit.
 A u-chart is like a c-chart, except that the
sample size is greater than one unit.
 As a result, a u-chart tracks the number
of defects per unit.
 A c-chart monitors the number of defects
on one unit.
 u-Chart
A car company monitors the number of
paint defects per car by taking a sample
of 5 cars each day over the next 6 days.
 The results are shown on next side.
 u-Chart

Day n c u=c/n u-bar LCL UCL

1 5 45 9.0 10.5 6.18 14.888
2 5 58 11.6 10.5 6.18 14.888
3 5 48 9.6 10.5 6.18 14.888
4 5 53 10.6 10.5 6.18 14.888
5 5 68 13.6 10.5 6.18 14.888
6 5 44 8.8 10.5 6.18 14.888
Sum u 63.2
u-bar 10.5
 u-Chart

u 
 u
k

u 
 u 63.2
  10.5
k 6
 u-Chart

u
LCL  u  3
n
u
UCL  u  3
n
 u-Chart
u
LCL  u  3
n
10.5
 10.5  3  6.18
5
u
UCL  u  3
n
10.5
 10.5  3  14.89
5
 u-Chart

u-Chart
Number of Paint Defects Per Car

16
c, number of defects

14
12 LCL
10
UCL
8
u
6
4 u-bar

2
0
1 2 3 4 5 6

Car
 Conclusion
 The process appears stable.
 We can therefore get a statistically valid
estimate the process mean number of defects
per car.
 Our estimate of the mean number of paint
defects per car is 10.5, the center line on the
control chart.
 Thus, we expect each car to have, on
average, 10.5 paint defects.
 Conclusion
 Although the process is stable, the number of
defects per car is too high.
 Deming calls this a stable process for the
production of defective product.
 Important take away:
 A stable process (process in control) is not
necessarily a good process because it can be
in control at the wrong level.
 A stable process is predictable, but this doesn’t
mean that what is being predicted is favorable.
 Mini Case
 Who is responsible for improving this
process?

 What is required to improve the process?

 Sample Size

Control Chart When To Use Sample Size

p-Chart Monitor the At least 50

proportion of
defectives in a process
c-Chart Monitor the number 1
of defects
u-chart Monitor the number >1
of defects per unit
 In Practice
 You need 25 to 30 samples before computing
initial control limits.
 When a special cause occurs, you should
eliminate that sample and re-compute control
limits if
 Special cause is identified
 Eliminated or made part of process
 To identify special causes, workers must keep
log sheet, where they record any changes
they make to the process.
 Tracking Improvements

UCL UCL
UCL

LCL
LCL
Additional improvements
LCL Process centered made to the process
Process not centered and stable
and not stable
10-55 Quality Control

Control Charts for Variables

 Use x-bar charts to monitor the changes in

the mean of a process (central tendencies)
 Use R-bar charts to monitor the dispersion or
variability of the process
 System can show acceptable central
tendencies but unacceptable variability or
 System can show acceptable variability but
unacceptable central tendencies

© Wiley 2010 55
10-56 Quality Control

Constructing an X-bar Chart: A quality control inspector at the Cocoa Fizz soft drink company has taken
three samples with four observations each of the volume of bottles filled, use the below data to develop
control charts with limits of 3 standard deviations for the 16 oz. bottling operation.

Time 1 Time 2 Time 3 Center line and

Observation 1 16.1 16.0 control limit
formulas
Observation 2 16.0 16.0 15.9
x1  x 2  ...x n
Observation 3 15.8 15.8 15.9 x ,
k
Observation 4 15.9 15.9 15.8 where (k) is the # of sample means and (n)

Sample 15.875 15.975 15.9

means (X-bar)
Sample 0.2 0.3 0.2
ranges (R)

© Wiley 2010 56
10-57 Quality Control

Solution and Control Chart (x-bar)

 Center line (x-double bar):

15.875  15.975  15.9

x  15.92
3

© Wiley 2010 57
10-58 Quality Control
Factor for x-Chart Factors for R-Chart
Sample Size
(n)
A2 D3 D4
2 1.88 0.00 3.27
3 1.02 0.00 2.57
4 0.73 0.00 2.28
5 0.58 0.00 2.11
6 0.48 0.00 2.00
7 0.42 0.08 1.92
8 0.37 0.14 1.86
9 0.34 0.18 1.82
10 0.31 0.22 1.78
11 0.29 0.26 1.74
12 0.27 0.28 1.72
13 0.25 0.31 1.69
14 0.24 0.33 1.67
15 0.22 0.35 1.65
10-59
Method for the X-bar Chart Using
Quality Control

R-bar and the A2 Factor

 Use this method when sigma for the process

distribution is not know
 Control limits solution:
0.2  0.3  0.2
R  .233
3

UCL x  x  A 2 R  15.92   0.73 .233  16.09

LCL x  x  A 2 R  15.92   0.73 .233  15.75

© Wiley 2010 59
10-60 Quality Control

X-Bar Control Chart

© Wiley 2010 60
10-61 Quality Control

Control Chart for Range (R)

 Center Line and Control Limit  Factors for three sigma control limits
Factor for x-Chart Factors for R-Chart
formulas: Sample Size
(n) A2 D3 D4
2 1.88 0.00 3.27
0.2  0.3  0.2 3 1.02 0.00 2.57
R  .233 4 0.73 0.00 2.28
3
5 0.58 0.00 2.11
6 0.48 0.00 2.00
UCLR  D4 R  2.28(.233)  .53 7 0.42 0.08 1.92
8 0.37 0.14 1.86
LCLR  D3 R  0.0(.233)  0.0 9 0.34 0.18 1.82
10 0.31 0.22 1.78
11 0.29 0.26 1.74
12 0.27 0.28 1.72
13 0.25 0.31 1.69
14 0.24 0.33 1.67
© Wiley 2010 15 0.22
61 0.35 1.65
10-62 Quality Control

R-Bar Control Chart

© Wiley 2010 62
10-63 Quality Control
A good number of samples of the items coming out of the machine are collected at the
random at the different intervals of the times of their quality characteristics (say
diameter or length

  X bar R
1 6.5 2
2 7 3
3 8 3
4 8.5 2
5 8 3
10-64 Quality Control

Mean values and ranges of data from 20 samples (sample size = 4) are shown
in the table below:

S.N Mean of Sample Range

1 10 4
2 15 4
3 12 5
4 11 4
5 9 5
6 11 6
7 11 4
8 9 4
9 10 4
10 11 6
11 12 5
12 13 4
13 12 4
14 12 3
15 11 3
16 15 4
17 12 4
18 15 3
19 11 3
20 10 4
10-65 Quality Control

Calculate X and R chart and discuss the process under control

mean UCL LCL

11.6 14.62535 8.57465