Quality Control

Chapter 5- Control
Charts for Variables
PowerPoints created
by
Kahlil Al-Hatab

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Outline
 Variation
 Concept of Test of Hypothesis
 The Control Chart Method
 State of Introduction Control
 Specifications
 Process Capability
 Different Control Charts
Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Learning Objectives
When you have completed this chapter you should:
 Know the three categories of variation and their sources.
 Understand the concept of the control chart method.
 Know the purpose of variable control charts.
 Know how to select the quality characteristics, the rational
subgroup and the method of taking samples
 Be able to calculate the central value, trial control limits
and the revised control limits for Xbar and R charts;
compare an R chart with an s chart;
 Be able to explain what is meant by a process in control and
the various out-of-control patterns; and the advantages that
accrue;
Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Learning Objectives
When you have completed this chapter you should:
 Explain what is meant by a process out of control and
the various out-of-control patterns;
 Know the difference between individual measurements and
averages; control limits and specifications.
 Know the different situations between the process spread
and specifications and what can be done to correct the
undesirable situation.
 Be able to calculate process capability.
 Be able to identify the different types of variable control
charts and the reasons for their use.
Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Variation
 The variation concept is a law of nature in that no
two natural items in any category are the same.
 The variation may be quite large and easily
noticeable
 The variation may be very small. It may appear that
items are identical; however,
however precision instruments will
show difference
 The ability to measure variation is necessary before it
can be controlled.

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Categories of Variation
 There are three categories of variation in piece part
production:

1. Within-piece variation: e.g. surface roughness

2. Piece-to-piece variation: Among pieces produced at

the same time. e.g. dimensions

3. Time-to-time variation: Difference in product

produced at different times of the day. Different
outcomes e.g. morning & afternoon, tool wear,
workers tired.
Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Sources of Variation
Sources of Variation in production processes:

Operators Methods Measurement

Materials Instruments

INPUTS PROCESS OUTPUTS

Tools Human Inspection

Machines Environment Performance

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Sources of Variation
1. Equipment: There is a certain capability or precision within
which the equipment operates
1. Tool wear

2. Machine vibration

3. Electrical fluctuations for welding.

2. Material: Because variation occurs in the finished product,

it must also occur in the raw material (which was someone
else’s finished product).
1. Tensile strength

2. Ductility

3. Thickness

4. Porosity etc.

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Sources of Variation
3. Environment:
1. Temperature
2. Light
3. Radiation
4. Humidity etc.
4. Operator: Method, SOP followed, Motivation
level, Training
1. Personal problem
2. Physical problem etc.
 As our equipment has become more automated,
the operator’s effect on variation has lessened.
Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Sources of Variation

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Chance and Assignable Causes of Variation
 Variation may be due to chance causes (common/random
causes) or assignable causes (special causes).

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Concept of Chance & Assignable Cause

 As long as these sources of variation fluctuate in a natural or

expected manner, a stable pattern of many chance causes
(random causes) of variation develops, which are small - OK ✓
 Chance causes are inevitable (unavoidable).
 If causes of variation large in magnitude; can be identified,
classified as assignable causes of variation. If present, process
variation is excessive (beyond expected natural variation).
 State of statistical control (SSC) – chance
 State of out of control (SOC) – assignable cause
 Body temperature - 36.5oC ~ 37.5oC.

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
Chance and Assignable Causes of Variation

 A process is operating with only

chance causes of variation present
is said to be in statistical control.
The process is stable and
predictable.
 A process that is operating in the
presence of assignable causes is
said to be out of control.

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
Concept of Test of Hypothesis

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
Concept of Test of Hypothesis

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Concept of Test of Hypothesis
 The power is the probability of correctly rejecting a null
hypothesis that is false. Obviously, tests with high powers are the
most desirable.
 Steps in Hypothesis Testing: The following steps summarize the
hypothesis-testing procedure:
Step 1: Formulate the null and alternative hypotheses.
Step 2: Determine the test statistic.
Step 3: Determine the rejection region of the null hypothesis based on
a chosen level of significance α.
Step 4: Make the decision. If the test statistic lies in the rejection
region, reject the null hypothesis. Otherwise, do not reject the null
hypothesis.

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Concept of Test of Hypothesis

 A control chart is like a sideways hypothesis test

 Detects a shift in the process
 Heads-off costly errors by detecting trends

2
UCL

0 CL
 
2 2

LCL

0 2
Sample Number
2-Sided Hypothesis Test Sideways Hypothesis Test Shewhart Control Chart

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Test of Hypothesis

 A statistical hypothesis is a statement about the value

of a parameter from a probability distribution.

 Ex. Test of Hypothesis on the Mean

 Say that a process is in-control if its’ mean is 0.
 In a test of hypothesis, use a sample of data from the process
to see if it has a mean of 0 .

 Formally stated:
 H0:  = 0 (Process is in-control)

HA:  ≠ 0 (Process is out-of-control)

02/16/21 IENG 486 Statistical Quality & Process Control 18

 Test of Hypothesis on Mean (Variance
Known)

 State the Hypothesis

 H0:  = 0
 H1:  ≠ 0
 Take random sample from process and compute appropriate test
statistic x  x 
z0  0
 0

x  n

 Pick a Type I Error level ( and find the critical value z/2
 Reject H0 if |z0| > z/2
 2  2
z
 z 2 z 2

02/16/21 IENG 486 Statistical Quality & Process Control 19

 UCL and LCL are Equivalent to the Test
of Hypothesis

 Reject H0 if: x  0
z0   z 2
 n
 Case 1:
x  0
x  0
 z 2
 n
 Case 2:
 x   0  z 2  
n  UCL

x  0
0  x
 z 2
 n
 For 3-sigma limits z/2 =
3 x   0  z 2  n   LCL
02/16/21 IENG 486 Statistical Quality & Process Control 20
 Two Types of Errors May Occur When
Testing a Hypothesis

 Type I Error - 
 Reject H0 when we shouldn't

Analogous to false alarm on control chart, i.e.,
 point lays outside control limits but process is truly in-control

 Type II Error - 
 Fail to reject H0 when we should
 Analogous to insensitivity of control chart to problems, i.e.,
 point does not lay outside control limits but process is never-the-
less out-of-control

02/16/21 IENG 486 Statistical Quality & Process Control 21

 Choice of Control Limits:
Trade-off Between Wide or Narrow Control Limits

 Moving limits further from the center line

 Decreases risk of false alarm, BUT increases risk of insensitivity

x UCL

CL

LCL

 Moving limits closer to the center line Sample

 Decreases risk of insensitivity, BUT increases risk of false alarm

UCL
CL
LCL

Sample

02/16/21 IENG 486 Statistical Quality & Process Control 22

 Consequences of Incorrect Control
Limits

 NOT GOOD:
 A control chart that never finds anything wrong
with process, but the process produces bad
product

 NOT GOOD:
 Too many false alarms destroys the operating
personnel’s confidence in the control chart, and
they stop using it

02/16/21 IENG 486 Statistical Quality & Process Control 23

 Differences in Viewpoint Between Test of Hypothesis &
Control Charts

Hypothesis Test Control Chart

Checks for the validity of Detect departures from
assumptions. assumed state of statistical
(ex.: is the actual process mean control
what we think it is?)
Tests for sustained shift Detects shifts that are
(ex.: have we actually reduced the short lived
variation like we think we have?)
Detects steady drifts
Detects trends

02/16/21 IENG 486 Statistical Quality & Process Control 24

 Example: Part Dimension

 When process in-control, a dimension is normally distributed

with mean 30 and std dev 1. Sample size is 5. Find control
limits for an x-bar chart with a false alarm rate of 0.0027.
 r.v. x - dimension of part

x ~ N   30,  1
 r.v. x - sample mean dimension of part

x ~ N    30, X  n 1 5 

02/16/21 IENG 486 Statistical Quality & Process Control 25

 Distribution of x vs. Distribution of x

Distribution of
individual
measurements x :
N   ,  Distribution of
sample mean x :

N  , x   n 
UCL    3 x

CL  

LCL    3 x

02/16/21 IENG 486 Statistical Quality & Process Control 26

 Ex. Part Dimension
Cont'd

 Find UCL: P  x  UCL   0.0027 / 2  0.00135

 UCL   
P  z    0.00135
 2  n 
 UCL     
P  z    1  0.00135  0.99865    z 
 2  n   2
UCL  
z  3 
 n
3  UCL    3   n 
 The control limits are:
2

CL    30
UCL    3   
n  30  3 1  
5  31.34

LCL    3   
n  30  3 1  
5  28.66

02/16/21 IENG 486 Statistical Quality & Process Control 27

 Ex. Modified Part Limits

 Consider an in-control process. A process

measurement has mean 30 and std dev 1 and
n = 5.
 Design a control chart with prob. of false alarm =
0.005

 If the control limits are not 3-Sigma, they are called

"probability limits".

02/16/21 IENG 486 Statistical Quality & Process Control 28

 The Control Chart Method

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 The Control Chart Method
 Removing special Statistical Quality Control and Improvement
causes of Improving Process Capability and Performance
variation
Continually Improve the System
 Preparation for:
 hypothesis tests
Characterize Stable Process Capability
 control charts

 process
improvement Head Off Shifts in Location, Spread
 Managing the Time

process with Identify Special Causes - Bad (Remove)

control charts
Identify Special Causes - Good (Incorporate)
 Process
Improvement Reduce Variability

 Process Center the Process

Stabilization LSL 0 USL

 Confidence in
“When to Act”

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 The Control Chart Method

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 The Control Chart Method
 A point that plots within the control limits indicates the process is in
control. No action is necessary.
 A point that plots outside the control limits is evidence that the
process is out of control. Investigation and corrective action are
required to find and eliminate assignable cause(s).
 There is a close connection between control charts and hypothesis
testing,
 Example: Process is monitored by average
flow width
 Sample of 5 wafers
 Process mean is 1.5 microns
 Process standard deviation is 0.15 microns
 Note that all plotted points fall inside the
control limits. Process is considered to be in
statistical control

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 The Control Chart Method

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Types of Process Variability
 Stationary and uncorrelated  data vary around a fixed
mean in a stable or predictable manner
 Stationary and auto-correlated  successive
observations are dependent with tendency to move in long
runs on either side of mean
 Nonstationary  process drifts without any sense of a
stable or fixed mean

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 More Basic Principles
 Control charts may be used to estimate process
parameters, which are used to determine capability
 Two general types of control charts
 Variables (Chapter 5)
Continuous scale of measurement
Quality characteristic described by central
tendency and a measure of variability
 Attributes (Chapter 9)
Conforming/nonconforming
Counts
 Control chart design includes selection of sample size,
control limits, and sampling frequency.

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Control Charts
Variable 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)

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Control Charts

Continuous Numerical Categorical or Discrete

Data Control Numerical Data
Charts

Variables Attributes
Charts Charts

R X P C
Chart Chart Chart Chart

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Control Charts for Variables
 The control chart shows whether or not a process is in a stable state. It
is a statistical tool that distinguishes between natural and unnatural
variation.
 Unnatural variation is the result of assignable causes. It usually, but
not always, requires corrective action by people close to the process,
such as operators, technicians, clerks, maintenance workers, and first-
line supervisors.
 The control chart is used to keep a continuing record of a particular quality
characteristic. It is a picture of the process over time. When the chart is
completed, it is replaced by a fresh chart, and the completed chart is stored
in an office file. The chart is used to improve the process quality, to
determine the process capability, to help determine effective specifications,
to determine when to leave the process alone and when to make
adjustments, and to investigate causes of unacceptable or marginal quality
Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Control Charts

Figure 5-1 Example of a control chart

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Objectives of Variable Control Charts
1) For quality improvement: A variable control chart is an excellent
technique for achieving quality improvement.
2) To determine the process capability: True process capability can
be achieved only after substantial quality improvement has been
achieved. During the quality improvement cycle, the control chart
will indicate that no further improvement is possible without a large
dollar expenditure. At that point the true process capability is
obtained.
3) For decisions regarding product specifications: Once the true
process capability is obtained, effective specifications can be
determined.
4) For current decisions on the production process: process in-
control or Process out of control.
Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Objectives of Variable Control Charts
5) For current decisions on recently produced items:
Control charts for variables should be established to
achieve a particular purpose. These purposes are
frequently dependent on each other. For example,
quality improvement is needed prior to determining the
true process capability, which is needed prior to
determining effective specifications.
6) Control charts are effective in defect prevention.
7) Control charts prevent unnecessary process
adjustment.
8) Control charts provide diagnostic information.

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Control Chart Techniques
Procedure for establishing a pair of control charts for the
average Xbar and the range R:
1. Select the quality characteristic
2. Choose the rational subgroup
3. Collect the data
4. Determine the trial center line and control limits
5. Establish the revised central line and control limits
6. Achieve the objective

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Quality Characteristic
The Quality characteristic must be measurable.
It can expressed in terms of the seven basic units:
1.Length

2.Mass

3.Time

4.Electrical current

5.Temperature

6.Substance

7.Luminosity

as appropriate.
Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Rational Subgroup
 A rational subgroup is one in which the variation within a group is
due only to chance causes.
 Within-subgroup variation is used to determine the control limits.
Variation between subgroups is used to evaluate long-term stability.
 Subgroups/Samples should be selected so that if assignable causes
are present:
 Chance for differences between samples is maximized
 Chance for differences within a sample is minimized
 Use consecutive units of production
 Keep sample size small so that:
 New events won’t occur during sampling
 Inspection is not too expensive
 But size is large enough that x is normally distributed
Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Rational Subgroup
 The rational subgroup concept means that subgroups or samples
should be selected so that if assignable causes are present, chance
for differences between subgroups will be maximized, while
chance for difference due to assignable causes within a subgroup
will be minimized.
 Two general approaches for constructing rational subgroups:
1. Sample consists of units produced at the same time 
consecutive units - primary purpose is to detect process shifts.
2. Sample consists of units that are representative of all units
produced since last sample  random sample of all process
output over sampling interval.
 Often used to make decisions about acceptance of product.
 Effective at detecting shifts to out-of-control state and back
into in-control state between samples
 Care must be taken because we can often make any process
appear to be in statistical control just by stretching out the
interval between observations in the sample.
Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Rational Subgroup
 The manner in which we sample the process deserves our
careful attention. This means that separate control charts
may have to be kept for different operators, machines, or
vendors. Lots from which samples are chosen should be
homogeneous.
 There are two schemes for selecting the subgroup samples:

1. The Instant-time Method: select subgroup samples from

product or service produced at one instant of time or as
close to that instant as possible
2. The Period-of-time Method: select from product or service
produced over a period of time that is representative of all
the products or services.
Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Rational Subgroup
 In comparing the two schemes, the instant-time method will
have a minimum variation within a subgroup and a maximum
variation among subgroups.
 The period-of-time method will have a maximum variation
within a subgroup and a minimum variation among subgroups.
 The first scheme is the most commonly used, since it provides a
particular time reference for determining assignable causes. It
also provides a more sensitive measure of changes in the
process average. Because all the values are close together, the
variation will most likely be due to chance causes and thereby
meet the rational subgroup criteria.
 The second scheme provides better overall results and will
provide a more accurate picture of the quality.

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Rational Subgroup
 As the subgroup size (n) increases, the control
limits become closer to the central value, which
make the control chart more sensitive to small
variations in the process average.
 As the subgroup size (n) increases, the inspection
cost per subgroup increases.
 When destructive testing is used and the item is
expensive, a small subgroup size (n) is required.

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Rational Subgroup
 From a statistical basis a distribution of
subgroup averages are nearly normal for
groups of 4 or more even when samples are
taken from a non-normal distribution.
 When a subgroup size of 10 or more is
used, the s-chart should be used instead of
the R-chart.
 See Table 6-1 for sample sizes.
Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Rational Subgroup
Subgroup Size

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Rational Subgroup
 Symptoms of Two Common Sampling Mistakes
 Data points hug centerline – Stratification
 Indicating sample averages are not normally distributed
x Chart
UCL

CL

LCL

 Data points hug control limits – Mixing

 Indicating sample averages are not normally distributed
x Chart
UCL

CL

LCL

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
Stratification – Sample Averages Hug Centerline
 Quality Characteristic:
 Amount of liquid filled into a container.
 Machine:
 4 heads fill 4 containers simultaneously.
 Each head has a slightly different mean
 Sample: select 4 bottles, 1 bottle ea. from heads 1, 2, 3, and 4
 Symptom: data points hug centerline indicating sample averages are not normally distributed

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
Stratification – Why Does It Happen?
 Each head has a different mean.
 Why do points hug centerline?
 Ans.: Estimate of  is wrong
 So … the Control Charts limits are too wide
 And … the data points hug centerline because the scaling to detect a shift
in process mean is off

sample 1 sample 2 sample 3 sample 4

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
Mixing – Sample Averages-Hug Control Limits
 Quality Characteristic:
Dimension of a part

 2 Similar Machines:
 Old one: produces 40% of parts
New one: produces 60% of parts
Old and new machines have different means
 Sample size = 4:
 All parts are mixed together
 Symptom: Data points hug control limits

samples 1 & 2 samples 3 & 4

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
What to do if there is evidence of stratification or mixing

 Examine your sampling procedure

 Make separate control charts for each

filling head (or machine)

 Problem: It may not be obvious that

samples are stratified or mixed

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
Why Monitor Both Process Mean and Process Variability?

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Data Collection
 Data collection can be accomplished using the type of
figure shown in Figure 6-2, wherein the data are recorded
in a vertical fashion.
 It can also be collected using the method in Table 6-2,
wherein the data are recorded in a horizontal fashion.
 It is necessary to collect a minimum of 25 subgroups of
data.
 A run chart can be used to analyze the data in the
development stage of a product or prior to a state of
statistical control.
Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Data Collection

Figure 5-1 Example of a method of reporting inspection results

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Run Chart
 It does not have control limits but can be used to analyze the data,
especially in the development stage of a product or prior to a state of
statistical control. The data points are plotted by order of
production, as shown in the figure.
 A run chart showing individuals observations in each sample, called
a tolerance chart or tier diagram may reveal patterns or unusual
observations in the data.

Figure 6-4 Run Chart for data of Table 6-2

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Suppose a quality characteristic is x ~ N(, ) and we know  and 

 If x1, x2, …, xn is a random sample of size n then:

and n
x  1
n x
i 1
i

x ~ N x   ,
 Recall that the probability is  that either:
n 

or


x  UCL    z 2
n

x  LCL    z 2
n
Besterfield: Quality Control, 8 ed..
th
© 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 which is equivalent to:

   
P    z 2  x    z 2   1
 n n
P  LCL  x  UCL   1  

 Where LCL and UCL are the lower and upper control limits,
respectively

 In practice, one must estimate mean and s from data coming

from an in-control process

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Statistics of the Range
 R – the range – is a sample statistic
 If x1, x2, …, xn is a random sample of size n from a normal
distribution then one can estimate  using the range:

ˆ  R d 2
 where d2 is a function of n and can be found in Table B of the
Appendix.
 Can get a better estimate for  if using more than one sample
 Compute Ri for each of m samples where i = 1, …, m
 Then use the sample average of Ri
m
R   Ri ˆ  R d 2
i 1
Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Operation of Charts
Operation of Charts: Use of
control chart for monitoring
future production, once a set of
reliable limits are established, is
called phase II of control chart.

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Trial Central Lines
 Calculate the subgroup mean and range using:

 Central Lines are obtained using:

g g

X i R i
X  i 1
and R  i 1
g g
where
X  average of subgroup averages
X i  average of the ith subgroup
g  number of subgroups
R  average of subgroup ranges
Ri  range of the ith subgroup
Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Trial Control Limits
 Obtain the trial control limits for each chart. Trial control
limits are established at ±3 standard deviations from the central
value

UCLX  X  3 X UCLR  R  3 R
LCLX  X  3 X LCLR  R  3 R
where
UCL=upper control limit
LCL=lower control limit
 X  population standard deviation of the subgroup averages
 R  population standard deviation of the range

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Trial Control Limits
In practice calculations are simplified by using the
following equations where A2, D3 and D4 are factors that
vary with the subgroup size and are found in Table B of
the Appendix.
UCLX  X  A2 R UCLR  D4 R
LCLX  X  A2 R LCLR  D3 R

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Example Problem 6-1

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Trial Control Limits

Figure 5-5 Xbar and R chart for preliminary data with trial control limits
Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
Revision of Control Limits & Center Lines
 Effective use of control charts requires periodic review
and revision of control limits and center lines
 Sometimes users replace the center line on the x chart
with a target value.
 When R chart is out of control, out-of-control points are
often eliminated to re-compute a revised value of which R
is used to determine new limits and center line on R
chart and new limits on chart.x
 Good control can be briefly described as that which has
no out-of-control points, no long runs on either side of
the central line, and no unusual patterns of variation.

Besterfield: Quality Control, 8th ed.. 72

© 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Revised Central Lines
X new 
 XX d
and R new 
 RR d

g  gd g  gd
where
X d  discarded subgroup averages
g d  number of discarded subgroups
Rd  discarded subgroup ranges
R0
X 0  X new R0  R new and 0 
d2

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Standard Values
R0
X 0  X new R0  R new and  0 
d2

UCLX  X 0  A 0 UCLR  D2 0

LCLX  X 0  A 0 LCLR  D1 0

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
Figure 5-6 Trial control limits and revised control limits for Xbar and R charts
Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
Example 6.2 The Hard Bake Process

Besterfield: Quality Control, 8th ed.. 76

© 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
Besterfield: Quality Control, 8th ed.. 77
© 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
Besterfield: Quality Control, 8th ed.. 78
© 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
Besterfield: Quality Control, 8th ed.. 79
© 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
Besterfield: Quality Control, 8th ed.. 80
© 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Limits for both charts become narrower after
revised
 Revised limits used to report / plot future sub-
groups
 For effective use – chart must be displayed and
easily seen control

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Final comments

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Achieving objective
 Initiate control charts results in quality improvement
 Less variation in sub-group averages
 Reduction in variation of range
 Reduce frequency of inspection monitoring purpose.

How Control chart Helps In Q.I.

 Psychological effect to do better.
 May be purchasing changed material supplier to ensure
consistent quality.
 Improvements must be from investigation of assignable
causes.

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Achieve the Objective

Figure 5-7 Continuing use of control charts, showing improved quality

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Sample Standard Deviation
Control Chart
For subgroup sizes ≥10, an s chart is more accurate
than an R Chart. Trial control limits are given by:

 
g g
i 1 i
s Xi the standard deviation
s X  i 1
is
g g
UCLX  X  A3 s UCLs  B4 s
LCLX  X  A3 s LCLs  B3 s

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Revised Limits for s chart

X0  X new 
X  Xd
g  gd

s0  snew 
s  s d
0 
s0
g  gd c4
UCLX  X 0  A 0 UCLs  B6 0
LCLX  X 0  A 0 LCLs  B5 0
where
sd  discarded subgroup averages
c4 , A, B5 , B6  factors found in Table B

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Example Problem 6-3

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Example Problem 6-3

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
Besterfield: Quality Control, 8th ed.. 89
© 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
Besterfield: Quality Control, 8th ed.. 90
© 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
Besterfield: Quality Control, 8th ed.. 91
© 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 X-Bar & Sigma-Charts

Limits can also be generated from historical data:

UCL  μ  Aσ UCL  B6 o

CL  μ CL  c4 o
LCL  μ  Aσ
LCL  B5 o
02/16/21 IENG 486: Statistical Quality & Process Control 92
 X-Bar & Sigma-Charts
The thickness of the magnetic coating on audio tapes is an important
characteristic. Random samples of size 4 are selected, and the thickness is
measured using an optical instrument. Table 7-3 shows the mean X and standard
deviation s for 20 samples. The specifications are 38 ±4.5 micro-meters (μm). If a
coating thickness is less than the specifications call for, that tape can be used for a
different purpose by running it through another coating operation.

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 X-Bar & Sigma-Charts
Solution The standard deviation chart must first be constructed. The center-line of
the s-chart is The control limits for the s-chart are

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 X-Bar & Sigma-Charts
The control limits for the s-chart are;

(b) Assuming special causes

for the out-of-control points,
determine the revised control
limits.
Solution In this case, the
revised control limits will be
the same as the trial control
limits because we believe that
no special causes are present
in the system.

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 X-Bar & Sigma-Charts
(c) Assuming the thickness of the coating to be
normally distributed, what proportion of the product
will not meet specifications?
Solution The process standard deviation may be
estimated as
To find the proportion of the output that does not meet specifications, the standard
normal values at the upper and lower specification limits (USL and LSL) must be
found. At the lower specification limit we get

The area below the LSL, found by using the standard normal table in Appendix A-
3, is 0.2389. Similarly, the standard normal value at the upper specification limit is

From Appendix A-3, the area above the USL is 0.1539. Hence, the proportion of
product not meeting specifications is 0.2389 + 0.1539 = 0.3928

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
Charts for Variable Subgroup Size

Used when the sample size is not the same

 Different control limits for each subgroup
 As n increases, limits become narrower
 As n decreases, limits become wider apart
 Difficult to interpret and explain
 To be avoided

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Standardized Control Charts
 When the sample size varies, the control limits on an
Xbar- and an R-chart will change.
 With fluctuating control limits, the rules for identifying
out-of-control conditions we discussed become difficult
to apply.
 One way to overcome this drawback is to use a
standardized control chart. When we standardize a
statistic, we subtract its mean from its value and divide
this value by its standard deviation.
 The standardized values then represent the deviation
from the mean in units of standard deviation.

Besterfield: Quality Control, 8th ed.. 98

© 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
Besterfield: Quality Control, 8th ed.. 99
© 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
Besterfield: Quality Control, 8th ed.. 100
© 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
Besterfield: Quality Control, 8th ed.. 101
© 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
Besterfield: Quality Control, 8th ed.. 102
© 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
Besterfield: Quality Control, 8th ed.. 103
© 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
104
Control Limits for a Given Target or Standard

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
Control Limits for a Given Target or Standard

as tabulated in Appendix B

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 State of Control
Process in Control
 When special causes have been eliminated from the process to the extent that
the points plotted on the control chart remain within the control limits, the
process is in a state of control.
 Further improvement through changing basic process, system
 When a process is in control, there occurs a natural pattern of variation,
which is illustrated by the control chart in Figure 6 -9.
 What are the characteristics of process in control? (natural pattern of
variation) :

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 State of Control
Figure 5-9 Natural pattern of
variation of a control chart.

A natural pattern is one in

which:
No identifiable arrangement of
the plotted points exists.
No points fall outside the control
limits,
The majority of the points are
near the centerline, and
Few points are close to the
control limits.
Natural patterns are indicative of
a process that is in control; that
is, they demonstrate the presence
of a stable system of common
causes.
Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Errors in Making Inferences from Control Charts
Types of errors:
 Type I, occurs when looking for a
special cause of variation when in
reality a common cause is present.
It is the sum of the two tail areas outside the control limits.
 Type II, occurs when assuming
that a common cause of variation is
present when in reality there is a
special cause.
 A type I error occurs when we
infer that a process is out of control
when it is really in control.
The placement of the control limits influences these two errors.
 A type II error occurs when we As the control limits are placed further out from the center line,
the probability of a type I error decreases, but the probability of a
infer that a process is in control type II error increases, when all other conditions remain the same,
and vice versa. An increase in the sample size may lead to
when it is really out of control. reducing both errors.
Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 State of Control

Type I error in control charts. Type II error in control charts.

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Example 6-2
A control chart is to be constructed for the average
breaking strength of nylon fibers. Samples of size 5 are
randomly chosen from the process. The process mean and
standard deviation are estimated to be 120 kg and 8 kg,
respectively.
a)If the control limits are placed three standard deviations
from the process mean, what is the probability of a type I
error?
b)If the process mean shifts to 125 kg, what is the
probability of concluding that the process is in control and
hence making a type II error on the first sample plotted
after the shift?
Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Example 6-2
Similarly, the Z-value at
the lower control limit is
3.00. For these Z-values
in the standard normal
table in Appendix A-3,
each tail area is found to
be 0.0013. The
probability of a type I
error, as shown by the
shaded tail areas in Figure
6-5a, is therefore 0.0026.

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Example 6-2
The process mean shifts to 125kg. Assuming that the process standard
deviation is the same as before, the distribution of the sample means
is shown in Figure 6-5b. We find the standardized normal value at
the upper and lower control limits as:
From the standard normal table in Appendix A-3,
the tail area above the upper control limit is
0.0548.

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 State of Control

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 State of Control
When the process is in control:
1. Individual units of the product or service
will be more uniform
2. Since the product is more uniform, fewer
samples are needed to judge the quality
3. The process capability or spread of the
process is easily attained from 6ơ
4. Trouble can be anticipated before it occurs

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 State of Control
When the process is in control:
5. The % of product that falls within any
pair of values is more predictable
6. It allows the consumer to use the
producer’s data
7. It is an indication that the operator is
performing satisfactorily

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Common
Causes

Special
Causes

45
Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 State of Out -of-Control

Figure 5-11 Frequency Distribution of subgroup averages with control limits

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 State of Out Control
 When a point (subgroup value) falls outside its
control limits, the process is out of control.

 Out of control means a change in the process due

to a special cause. A process can also be
considered out of control even when the points
fall inside the 3ơ limits

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 State of Out of Control
 It is not natural for seven or more consecutive
points to be above or below the central line.
 Also when 10 out of 11 points or 12 out of 14
points are located on one side of the central line,
it is unnatural.
 Six points in a row are steadily increasing or
decreasing indicate an out of control situation

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Out-of-Control Condition
1. Change or jump in level.

Many causes can bring

about a sudden change
Trend or steady change in level
2. jump)
(or

Trends represent changes

that steadily increase or
decrease
2. Recurring cycles
a repetitive periodic behavior in the system
• An improvement in operator skill (downward trend)
• A decrease in operator skill due to fatigue, (upward trend)
• A gradual improvement in the homogeneity of incoming
material

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Out-of-Control Condition
4. Two populations (also called mixture)

5. Mistakes

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Patterns in Control Charts

Figure 5-12 Some unnatural runs-process out of control

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Patterns in Control Charts

Figure 5-13 Simplified rule for out-of-control pattern

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Out-of-Control Patterns

Change or jump in level Trend or steady change in level

Recurring cycles Two populations

Besterfield: Quality Control, 8th ed.. © 2009 Pearson Education, Upper Saddle River, NJ 07458.
All rights reserved
 Operating Characteristic
(OC) Curve
• An operating characteristic (OC) curve is a measure of goodness of a control
chart's ability to detect changes in process parameters.

 Ability of the x and R charts to detect shifts (sensitivity) is

described by OC curves

 For x chart; say we know 

 Mean shifts from
0 (in-control value) to
1 = 0 +k (out-of-control value)

 The probability of NOT detecting the shift on the first sample


after shift is   P LCL  x  UCL       k
1 0 
 LCL  1 UCL  1 
  P z 
02/16/21
  n  n 
IENG 486: Statistical Quality & Process Control 126
 Ex. Probability of NOT
Detecting Shift
 A 3-sigma x chart is used to monitor a normally distributed
quality characteristic. The process std dev is 1.2 and the
sample size is 5. The process mean is in-control at 22.
 Find the probability that a shift to 24.4 is not detected on the first
sample after the shift.

02/16/21 IENG 486: Statistical Quality & Process Control 127

 OC Curve for x Chart
 Plot of  vs. shift size (in std dev units) for various sample
sizes n OC Curve for x-bar chart with 3-sigma limits

1.00
0.90
0.80
0.70
0.60
Beta

0.50 n=20
0.40 n=5 n=2
n=1
0.30
0.20
0.10
0.00

 x chart not effective for small shift sizes, i.e., k  1.5

 Performance gets better for larger n and L or larger shifts

02/16/21 IENG 486: Statistical Quality & Process Control 128

 OC curve for R Chart
 Uses distribution of relative range r.v., i.e.,
W R 
 Suppose
 0 - in-control std dev
 1 - out-of-control std dev

 OC curve for R chart plots  vs. ratio of in-control to out-of-

control standard deviation for various sample sizes
 That is, plot β vs. 

 R chart not very effective for detecting shifts for small sample
sizes
(see Fig. 5-14 in text)

02/16/21 IENG 486: Statistical Quality & Process Control 129

 Probability of Detecting Shift
for Subsequent Samples
 After the shift has occurred:
 P(NOT detecting shift ON 1st sample)
  0.07078
 P(DETECTING shift ON 1st sample)

  1  shift
P(DETECTING  ON0.93
2nd sample)


 1
P(DETECTING  ON
shift  rth0.066
sample)

 P(DETECTING
r 1 shift BY 2 sample)
nd
  1  
 P(DETECTING shift BY rth sample)
 1      1    0.93  0.066  0.996

 1  
r
i 1
 i 1

02/16/21 IENG 486: Statistical Quality & Process Control 130

 Average Run Length (ARL)
 An alternative measure of the performance of a control chart, in addition to the OC curve, is the
average run length(ARL).
 This denotes the number of samples, on average, required to detect an out-of-control signal.
 Expected number of samples taken before shift is detected is called the Average Run Length (ARL)

 For a process in control, we prefer the ARL to be large because an observation plotting outside the
control limits represents a false alarm.
 it is desirable for the ARL to be small because we want to detect
 For an out-of-control process,
1 the

ARL   r   1   
Β is the
r 1
out-of-control condition as soon as possible
probability

r 1 1   of a type II
error

02/16/21 IENG 486: Statistical Quality & Process Control 131

 Performance of Any
Shewhart Control Chart
 In-Control ARL:
 Average number of points plotted on control chart before a false alarm
occurs
(ideally, should be large)
1
ARL0 

 Out-of-Control ARL:
 Average number of points, after the process goes out-of-control,
before the control chart detects it
(ideally, should be small)
1
ARL1 
1 

02/16/21 IENG 486: Statistical Quality & Process Control 132

 ARL Curve for x Chart
 Plot of ARL1 vs. shift size (in sd units) for various sample sizes n:
ARL for x-bar chart with 3-sigma limits

20.00
18.00
16.00
ARL to detect shift

14.00
12.00 n=20
10.00 n=4
n=2 n=1
8.00
6.00
4.00
2.00
0.00
0.0
0.2
0.3
0.5
0.6
0.8

1.2
1.4

1.7

2.0
2.1

2.6
2.7
2.9
3.0
0.9
1.1

1.5

1.8

2.3
2.4
 Average Time to Signal, (ATS):
k (shift size)

 Number of time periods that occur until signal is generated on control chart

ATS   ARL   h 
 h - time interval between samples

02/16/21 IENG 486: Statistical Quality & Process Control 133