Chapter 10

Quality Control
Chapter 10: Learning Objectives
You should be able to:
LO 10.1 Explain the need for quality control
LO 10.2 Discuss the basic issues of inspection
LO 10.3 List and briefly explain the elements of the control process
LO 10.4 Explain how control charts are used to monitor a process, and
the concepts that underlie their use
LO 10.5 Use and interpret control charts
LO 10.6 Perform run tests to check for non-randomness in process
output
LO 10.7 Assess process capability

What is Quality Control?

 Quality Control
 A process that evaluates output relative to a standard and takes
corrective action when output doesn’t meet standards
 If results are acceptable no further action is required
 Unacceptable results call for correction action
 Inspection alone is not is generally not sufficient to achieve a
reasonable level of quality
 Most organization rely upon some inspection and a great deal of
process control to achieve an acceptable level of quality.

Phases of Quality Assurance

1
Inspection
 Inspection
 An appraisal activity that compares goods or services to a standard
 Inspection issues:
• How much to inspect and how often
• At what points in the process to inspect
• Whether to inspect in a centralized or on-site location
• Whether to inspect attributes or variables

How Much to Inspect

Where to Inspect in the Process

 Typical Inspection Points:
 Raw materials and purchased parts
 Finished products
 Before a costly operation
 Before an irreversible process
 Before a covering process
2
Centralized vs. On-Site Inspection
 Effects on cost and level of disruption are a major issue in selecting
centralized vs. on-site inspection
 Centralized
 Specialized tests that may best be completed in a lab
 More specialized testing equipment
 More favorable testing environment
 On-Site
 Quicker decisions are rendered
 Avoid introduction of extraneous factors
 Quality at the source

Statistical Process Control (SPC)

 Quality control seeks
 Quality of Conformance
 A product or service conforms to specifications
 A tool used to help in this process:
 SPC
 Statistical evaluation of the output of a process
 Helps us to decide if a process is “in control” or if corrective
action is needed

Process Variability
 Two basic questions: concerning variability:
• Issue of Process Control
 Are the variations random? If nonrandom variation is
present, the process is said to be unstable.
• Issue of Process Capability
 Given a stable process, is the inherent variability of the process
within a range that conforms to performance criteria?
Variation

3
 Variation
 Random (common cause) variation:
 Natural variation in the output of a process, created by countless
minor factors
 Assignable (special cause) variation:
 A variation whose cause can be identified.
 A nonrandom variation

Sampling and Sampling Distribution

 SPC involves periodically taking samples of process output and computing
sample statistics:
 Sample means
 The number of occurrences of some outcome
 Sample statistics are used to judge the randomness of process variation

Control Process
 Sampling and corrective action are only a part of the control process
 Steps required for effective control:
 Define: What is to be controlled?
 Measure: How will measurement be accomplished?
 Compare: There must be a standard of comparison
 Evaluate: Establish a definition of out of control
 Correct: Uncover the cause of nonrandom variability and fix it
 Monitor: Verify that the problem has been eliminated
Control Charts:

4
The Voice of the Process
 Control Chart
 A time ordered plot of representative sample statistics obtained from
an ongoing process (e.g. sample means), used to distinguish between
random and nonrandom variability
 Control limits
 The dividing lines between random and nonrandom deviations
from the mean of the distribution
 Upper and lower control limits define the range of acceptable
variation

Control Chart

Errors
 Type I error

 Concluding a process is not in control when it actually is.

 The probability of rejecting the null hypothesis when the null
hypothesis is true.
 Manufacturer’s Risk
 Type II error
 Concluding a process is in control when it is not.
 The probability of failing to reject the null hypothesis when the
null hypothesis is false.
 Consumer’s Risk

5
Observations from Sample Distribution

Control Charts for Variables

 Variables generate data that are measured
 Mean control charts
 Used to monitor the central tendency of a process.
 “x- bar” charts
 Range control charts
 Used to monitor the process dispersion
 R charts
X-Bar chart: Control Limits

X-Bar Chart is used to monitor the central tendency of a process

k

∑x
k

x= i =1
i
∑R i

k R= i =1
k
where where
x = Average of sample means R = Average of sample ranges
x i = mean of sample i Ri = Range of sample i
k = number of samples

x − chart Control Limits

UCLx = x + A2 R

LCLx = x − A2 R
6
where
A2 = a control chart factor based on sample size, n
Range Chart: Control Limits
 Used to monitor process dispersion
R Chart Control Limits
UCLR = D4 R
LCLR = D3 R
where
D3 = a control chart factor based on sample size, n
D4 = a control chart factor based on sample size, n

Factors for Three Sigma Control Charts

Mean and Range Charts

7
 Example: X-bar and R charts

Samples of five observations have been taken and are shown below:

Sample
1 2 3 4 5 6
79.2 80.5 79.6 78.9 80.5 79.7
78.8 78.7 79.6 79.4 79.6 80.6
80.0 81.0 80.4 79.7 80.4 80.5
78.4 80.4 80.3 79.4 80.8 80.0
81.0 80.1 80.8 80.6 78.8 81.1

Determine upper and lower control limits. Round all intermediate values to a
maximum of three decimals. Round Mean and Range chart control limits to
three decimals.

Samples
1 2 3 4 5 6
79.2 80.5 79.6 78.9 80.5 79.7
78.8 78.7 79.6 79.4 79.6 80.6
80.0 81.0 80.4 79.7 80.4 80.5
78.4 80.4 80.3 79.4 80.8 80.0
81.0 80.1 80.8 80.6 78.8 81.1
Mean 79.48 80.14 80.14 79.60 80.02 80.38
Range 2.6 2.3 1.2 1.7 2.0 1.4

From Table 10.3 with n = 5: A2 = 0.58; D3 = 0; D4 = 2.11.

X-bar Chart:
Upper Control Limit
(UCL) =

Lower Control Limit

(LCL) =

R Chart:
Upper Control Limit (UCL) =
Lower Control Limit (LCL) = .000
8
 81.500
Mean Chart 81.043
80.500

79.960 UC
79.500 L

78.877
78.500
1 2 3 4 5 6

5.000
Range Chart
4.000 3.939

3.000

2.000 UCL
1.867
1.000

0.000 0.000
1 2 3 4 5 6

All sample means are within the control limits on the Mean chart.
All sample ranges are within the control limits on the Range chart.
The process is in control.

Using Mean and Range Charts

 To determine initial control limits:
 Obtain 20 to 25 samples
 Compute appropriate sample statistics
 Establish preliminary control limits
 Determine if any points fall outside of the control limits
 If you find no out-of-control signals, assume the process is in
control
 If you find an out-of-control signal, search for and correct the
assignable cause of variation

9
  Resume the process and collect another set of observations on which
to base control limits
 Plot the data on the control chart and check for out-of-control signals

Control Charts for Attributes

 Attributes generate data that are counted.
 p-Chart
 Control chart used to monitor the proportion of defectives in a
process
 c-Chart
 Control chart used to monitor the number of defects per unit
Use a p-chart:
 When observations can be placed into two categories.
 Good or bad
 Pass or fail
 Operate or don’t operate
 When the data consists of multiple samples of several observations
each

p-chart Control Limit

Total number of defectives
p=
Total number of observations
p (1 − p )
σˆ p =
n
UCL p = p + z (σˆ p )
LCL p = p − z (σˆ p )

10
Example – Use of P chart

Four samples of 200 credit card statements each were taken and
the number with errors recorded as shown below:

Sample 1 2 3 4
# with 4 2 5 9
errors

a. Determine the fraction (proportion) defective for each sample.

Round all values to a maximum of four decimals.

n = 200 per sample

Sample 1 2 3 4
# with errors 4 2 5 9
Prop. defective
(# with errors/200)
.020 .010 .025 .045

b. If the true fraction (proportion) defective is unknown, estimate

the fraction defective. Round all values to a maximum of four
decimals.

(.020 + .010 + .025 + .045)/4 = .025

Or:
Total number of defects = 20

11
 total number of defectives 20
p= = = .025
total number of observations 4(200)

c. Estimate the mean and the standard distribution of the sampling

distribution. Round all values to a maximum of four decimals.

Mean = .025
p (1 − p ) .025(1 − .025)
Std. dev. = = = .011
n 200

d. Determine control limits for an alpha risk of .03. Round all p-

Chart control limits to four decimals.

An alpha risk of .03 infers that there is .03/2 = .015 in each tail.
Using Appendix B, Table A, find the z value that corresponds to
.5000 – .0150 = .4850:
z = 2.17.
UCL: .025 + 2.17(.011) = .025 + .0239 = .0489
LCL: .025 – 2.17(.011) = .025 – .0239 = .0011

You can find the Z value corresponding to alpha using TI-83/84 as

follows:
Given 0.015 area in each tail, use inverse normal function as
follows:
On calculator press 2ND then VARS to access invNorm, then input
area: 0.015
µ:0
σ: 1.0

12
 Then press PASTE, so invorm (0.015, 0, 1) = -2.17, this is the
negative value of Z for the lower limit and for the upper limit
the Z values is +2.17.

f. Using the control limits from part d, determine if the process is

in control.

.0510
p-Chart .0470
.0410

.0310
.0250 UCL
.0210 LCL
Prop.
.0110

.0030
.0010
1 2 3 4

All points fall within the limits. The process is in control.

g. Assume that the long-run fraction (proportion) defective = 2% =
.02. Determine the mean and standard deviation of the sampling
distribution. Round all intermediate values to a maximum of
four decimals. Round all p-Chart control limits to four decimals.

Mean = .02
p (1 − p ) .02(1 − .02)
Std. dev. = = = .0099
n 200

h. Construct a control chart for the process, assuming a fraction

(proportion) defective of 2% (.02) using two-sigma control limits.
Round all intermediate values to a maximum of four decimals.
Round all p-Chart control limits to four decimals.
13
 p (1 − p ) .02(1 − .02)
p±2 = .02 ± 2 = .02 ± .0198
n 200

UCL = .0398
LCL = .0002

.0500
p-Chart
.0400 .0398

.0300
UCL
.0200 .0200 LCL
Prop.
.0100

.0000 .0002
1 2 3 4

Sample 4 proportion defective is above the upper control limit.

The process is not in control.

The C-chart

The C-chart is used to monitor the number of defect per part. Used
only when the number of occurrences per unit of measure can be
counted; non-occurrences cannot be counted. Examples include:

 Scratches, chips, dents, or errors per item

 Cracks or faults per unit of distance
 Breaks or Tears per unit of area
 Bacteria or pollutants per unit of volume
 Calls, complaints, failures per unit of time

UCL c = c + z c 14

LCL c = c − z c
 Example- Use of C chart

The postmaster of a small town has recorded the number of complaints per
day. Determine three-sigma control limits using the data below. Round all
intermediate values to a maximum of three decimals. Round all c-Chart
control limits to three decimals.

Day 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Comp. 4 10 14 8 9 6 5 12 13 7 6 4 2 10
110
c= = 7.857
14

Control limits: c ± 3 c = 7.857 ± 3 7.857 = 7.857 ± 8.409

UCL = 16.266
LCL = -0.552 = 0 (cannot be negative)

20.000
c-Chart
16.266
15.000

10.000 UCL
LCL
7.857
Comp.
5.000

0.000 0.000
1 2 3 4 5 6 7 8 9 10 11 12 13

All points are within the limits. The process is in control.

15
Managerial Considerations
 At what points in the process to use control charts
 What size samples to take
 Sample frequency
 What type of control chart to use
 Variables
 Attributes

Run Tests
 Even if a process appears to be in control, the data may still not
reflect a random process
 Analysts often supplement control charts with a run test
 Run test
 A test for patterns in a sequence
 Run
 Sequence of observations with a certain characteristic
Nonrandom Patterns

Process Capability

 Once a process has been determined to be stable, it is necessary to

determine if the process is capable of producing output that is
within an acceptable range.

 Tolerances or specifications
 Range of acceptable values established by engineering design or
customer requirements
 Process variability
 Natural or inherent variability in a process
 Process capability

16
  The inherent variability of process output (process width)
relative to the variation allowed by the design specification
(specification width)

Process Capability (Cp) and Cpk

If the process mean is centered at its target, we can use Cp to calculate process
capability as follows:

UTL - LTL
Cp =
6σ
where
UTL = upper tolerance (specification) limit
LTL = lower tolerance(specification) limit

If the process average is not centered at its target, we must use Cpk as follows:

C pk = min{C pu , C pl }
 UTL − x x − LTL 
= min  , 
 3σ 3σ 
Example. A filling process has specification (tolerance) limits of 12 ± .25 oz.
If the mean of process output is 12 oz and standard deviation of output is 0.12
oz, determine:

a- The process capability Cp and indicate the status of the process.

Cp = (12.25 – 11.75)/ (6*.12) = 0.69

This process is not capable in producing units within tolerance limits. The
value of Cp must be a minimum of 1.3 for a capable process.

17
 b- Calculate the capability if the process standard deviation is reduced by
40%.

New standard deviation = .12 (1-0.40) = 0.072

Cp = (12.25 – 11.75)/ (6*.072) = 1.16

It is more capable but it needs to be over 1.3 for capable process.

c- Find Cpk if the mean of process is set at 11.95 and process standard
deviation is reduced to 0.05. Assume tolerance is still 12 ± .25.

Cpk = minimum {(12.25 -11.95)/(3*0.05) , (11.95-11.75)/(3*0.05)}

= minimum (2, 1.33) = 1.33  Process is capable

Example- A manufacturing process has three steps. Step 1 has 5% defect,

step 2 has 3% defect, and step 3 has 2% defect. If we have an order for 1000
units, how many units must go through the first step to have 1000 units for a
customer?

Let X = number of original units

(1-.05)(1-.03)(1-.02)*X = 1000  .90307*X = 1000  X= 1107.33 round up

to 1108 units needed in the first step.

Improving Process Capability

 Simplify
 Standardize
 Mistake-proof
 Upgrade equipment
 Automate
Taguchi Loss Function
18
Limitations of Capability Measures
 There are several risks of using capability measures:
 The process may not be stable
 The process output may not be normally distributed
 The process not centered but Cp is used

Operations Strategy
 Quality is a primary consideration for nearly all customers
 Achieving and maintaining quality standards is of strategic
importance to all business organizations
 Product and service design
 Increase capability in order to move from extensive use of
control charts and inspection to achieve desired quality
outcomes

19