10

Quality Control

McGraw-Hill/Irwin Copyright © 2007 by The McGraw-Hill Companies, Inc. All

 Learning Objectives
 List and briefly explain the elements of the
control process.
 Explain how control charts are used to
monitor a process, and the concepts that
underlie their use.
 Use and interpret control charts.
 Use run tests to check for nonrandomness
in process output.
 Assess process capability.
10-2
 Phases of Quality Assurance
Figure 10.1

Inspection and
Inspection of lots corrective Quality built
before/after action during into the
production production process

Acceptance Process Continuous

sampling control improvement

The least The most

progressive progressive

10-3
 Inspection
Figure 10.2

 How Much/How Often

 Where/When
 Centralized vs. On-site

Inputs Transformation Outputs

Acceptance Process Acceptance

sampling control sampling

10-4
 Inspection Costs
Figure 10.3
Cost

Total Cost
Cost of
inspection

Cost of
passing
defectives
Optimal
Amount of Inspection

10-5
 Where to Inspect in the Process

 Raw materials and purchased parts

 Finished products
 Before a costly operation
 Before an irreversible process
 Before a covering process

10-6
 Examples of Inspection Points
Table 10.1
Type of Inspection Characteristics
business points
Fast Food Cashier Accuracy
Counter area Appearance, productivity
Eating area Cleanliness
Building Appearance
Kitchen Health regulations
Hotel/motel Parking lot Safe, well lighted
Accounting Accuracy, timeliness
Building Appearance, safety
Main desk Waiting times
Supermarket Cashiers Accuracy, courtesy
Deliveries Quality, quantity
10-7
 Statistical Control

 Statistical Process Control:

Statistical evaluation of the output of a
process during production
 Quality of Conformance:
A product or service conforms to
specifications

10-8
 Control Chart
 Control Chart
 Purpose: to monitor process output to see
if it is random
 A time ordered plot representative sample
statistics obtained from an on going
process (e.g. sample means)
 Upper and lower control limits define the
range of acceptable variation

10-9
 Control Chart
Figure 10.4

Abnormal variation Out of

due to assignable sources control
UCL

Mean
Normal variation
due to chance
LCL
Abnormal variation
due to assignable sources

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Sample number

10-10
 Statistical Process Control
 The essence of statistical process
control is to assure that the output of a
process is random so that future output
will be random.

10-11
 Statistical Process Control

 The Control Process

 Define
 Measure
 Compare
 Evaluate
 Correct
 Monitor results

10-12
 Statistical Process Control

 Variations and Control

 Random variation: Natural variations in the
output of a process, created by countless
minor factors
 Assignable variation: A variation whose
source can be identified

10-13
 Sampling Distribution
Figure 10.5

Sampling
distribution

Process
distribution

Mean

10-14
 Normal Distribution
Figure 10.6

σ = Standard deviation

−3σ −2σ + 2σ + 3σ
Mean
95.44%

99.74%

10-15
 Control Limits
Figure 10.7
Sampling
distribution

Process
distribution

Mean

Lower Upper
control control
limit limit

10-16
Control Limits are based on the
Normal Curve

x
µ
z
-3 -2 -1 0 1 2 3

Standard
Standard
deviation
deviation
units
unitsoror“z”
“z”
units.
units.
 Control Limits

We establish the Upper Control Limits (UCL) and

the Lower Control Limits (LCL) with plus or
minus 3 standard deviations from some x-bar or
mean value. Based on this we can expect 99.7%
of our sample observations to fall within these
limits.

99.7%
x
LCL UCL
 SPC Errors
 Type I error
 Concluding a process is not in control
when it actually is.
 Type II error
 Concluding a process is in control when it
is not.

10-19
 Type I and Type II Errors
Table 10.2

In control Out of control

In control No Error Type I error

(producers risk)
Out of Type II Error No error
control (consumers risk)

10-20
 Type I Error
Figure 10.8

α /2 α /2

Mean

α = Probability LCL UCL

of Type I error

10-21
 Observations from Sample
Figure 10.9 Distribution
UCL

LCL

1 2 3 4
Sample number

10-22
 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
10-23
 Mean and Range Charts
Figure 10.10A

(process mean is
shifting upward)
Sampling
Distribution

UCL

x-Chart Detects shift

LCL

UCL

Does not
R-chart
detect shift
LCL

10-24
 Mean and Range Charts
Figure 10.10B

Sampling
Distribution (process variability is increasin

UCL

Does not
x-Chart
LCL
reveal increase

UCL

R-chart Reveals increase

LCL

10-25
Statistical Process Control Chart

Exhibit S9.21

10-26
 Control Chart Decision Rules

Source: Bertrand L. Hansen, Quality Control: Theory and Applications © 1963, p. 65.
Reprinted by permission of Pearson Education, Inc., Upper Saddle River, NJ. Exhibit S9.22

10-27
 Control Chart Decision Rules

Source: Bertrand L. Hansen, Quality Control: Theory and Applications © 1963, p. 65.
Reprinted by permission of Pearson Education, Inc., Upper Saddle River, NJ. Exhibit S9.22 (cont’d)

10-28
Changes in Mean and Variation
of Sample Mean Distributions

Exhibit S9.23

10-29
X and R Charts for Variables SPC

Exhibit S9.24

10-30
 Control Chart for Attributes

 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

Attributes generate data that are counted.

10-31
 Use of p-Charts
Table 10.4

 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

10-32
 Use of c-Charts
Table 10.4

 Use only when the number of

occurrences per unit of measure can be
counted; non-occurrences cannot be
counted.
 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

10-33
 Use of Control Charts
 At what point in the process to use
control charts
 What size samples to take
 What type of control chart to use
 Variables
 Attributes

10-34
 Run Tests
 Run test – a test for randomness
 Any sort of pattern in the data would
suggest a non-random process
 All points are within the control limits -
the process may not be random

10-35
 Nonrandom Patterns in Control
charts
 Trend
 Cycles
 Bias
 Mean shift
 Too much dispersion

10-36
 Counting Runs

Figure 10.12 Counting Above/Below Median Runs (7 runs)

B A A B A B B B A A B

Figure 10.13 Counting Up/Down Runs (8 runs)

U U D U D U D U U D

10-37
 NonRandom Variation
 Managers should have response plans to
investigate cause
 May be false alarm (Type I error)
 May be assignable variation

10-38
 Process Capability
 Tolerances or specifications
 Range of acceptable values established by engineering design
or customer requirements
 Process variability
 Natural variability in a process—measured by std devn of
process
 Process capability
 Process variability relative to specification--- determination of
whether variability inherent in the output of process that is in
control falls within the design specifications for the product
output
 Control Limits
Statistical limits that reflect the extent to which sample statistics
like mean,range can vary due to randomness alone

10-39
 Properties of a Normal
Distribution
 The distribution is bilaterally symmetrical.
 68.3 percent of the distribution lies between
plus and minus one standard deviation from
the mean.
 95.4 percent of the distribution lies between
plus and minus two standard deviations from
the mean.
 99.7 percent of the distribution lies between
plus and minus three standard deviations
from the mean.
10-40
Areas under the Normal Distribution
Curve Corresponding to Different
Numbers of
Standard Deviations from the Mean

Exhibit S9.19

10-41
 Statistical Process Control
(cont’d)
 Process Capability (Study)
 Comparing inherent variation in a process to the
customer’s requirements (the specifications) to
determine whether the process can produce
what the customer requires
 Collect data on the process while the process is
operating without known causes of variation.
 Compare the customer’s requirements to the inherent
variation of the process.
 If the customer’s specifications fall within the three standard
deviations for the process, some predictable percentage of
the time, the process will produce output that will not meet
the customer’s needs.
10-42
Capability Study

Exhibit S9.20

10-43
 Process Capability
Figure 10.15
Lower Upper
Specification Specification

A. Process variability
matches specifications
Lower Upper
Specification Specification

B. Process variability
Lower Upper
well within specifications Specification Specification

C. Process variability
exceeds specifications

10-44
 In case c,options are
Redesign process to achieve desired output
Use a new process
Ensure 100% inspection
Change specification

10-45
 Process Capability Ratio
If the process is centered use Cp
specification width
Process capability ratio, Cp =
process width

Cp = Upper specification – lower specification

6σ
If the process is not centered use Cpk
 X − LTL UTL - X 
C pk = min  or 

 3σ 3σ 

10-46
 Minimum process capability required=1.
 Good measure=1.33
 For cp=1,DPM=2700 and for
cp=1.3330,DPM=30
 For a Six sigma programme,Cp=2 –Process
variability is so small that design tolerance is
six SD above and below the process mean
 Cp=2 means Specification
width=(6SD+6SD)/6SD
 Fig 10.16

10-47
 Process variability of 6 standard deviations refers to +/- 3 standard deviations of the process from mean

Eg length of time to perform a service is 10 mts and acceptable variation is +/-1 minute.If process SD=0.5 Mts

+/- 3 SD=1.5 min
Design tolerance=1 min
Hence not capable

10-48
 The Cereal Box Example

 We are the maker of this cereal. Consumer reports has

just published an article that shows that we frequently
have less than 15 ounces of cereal in a box.
 Let’s assume that the government says that we must be
within ± 5 percent of the weight advertised on the box.
 Upper Tolerance Limit = 16 + .05(16) = 16.8 ounces
 Lower Tolerance Limit = 16 – .05(16) = 15.2 ounces
 We go out and buy 1,000 boxes of cereal and find that
they weight an average of 15.875 ounces with a standard
deviation of .529 ounces.

10-49
 Cereal Box Process Capability
 Specification or Tolerance
Limits
 Upper Spec = 16.8 oz
 Lower Spec = 15.2 oz
 X − LTL UTL − X 
 Observed WeightC pk = Min  ; 
 Mean = 15.875 oz  3σ 3σ 
 Std Dev = .529 oz 15.875 − 15.2 16.8 − 15.875 
C pk = Min  ; 
 3(.529 ) 3(.529 ) 
C pk = Min{.4253; .5829}

C pk = .4253

10-50
 What does a Cpk of .4253 mean?

 An index that shows how well the units

being produced fit within the
specification limits.
 This is a process that will produce a
relatively high number of defects.
 Many companies look for a Cpk of 1.3
or better… 6-Sigma company wants
2.0!

10-51
Limitations of Capability Indexes
1. Process may not be stable
2. Process output may not be normally
distributed
3. Process not centered but Cp is used

10-52
 Example 8
Standard Machine
Machine Deviation Capability Cp
A 0.13 0.78 0.80/0.78 = 1.03
B 0.08 0.48 0.80/0.48 = 1.67
C 0.16 0.96 0.80/0.96 = 0.83

Cp > 1.33 is desirable

Cp = 1.00 process is barely capable
Cp < 1.00 process is not capable

10-53
 3 Sigma and 6 Sigma Quality

Lower Upper
specification specification

1350 ppm 1350 ppm

1.7 ppm 1.7 ppm

Process
mean
+/- 3 Sigma

+/- 6 Sigma

10-54
 Graph of the normal distribution, which underlies the
statistical assumptions of the Six Sigma model. The Greek
letter σ (sigma) marks the distance on the horizontal axis
between the mean, µ, and the curve's inflection point. The
greater this distance, the greater is the spread of values
encountered. For the curve shown above, µ = 0 and σ = 1.
The upper and lower specification limits (USL, LSL) are at a
distance of 6σ from the mean. Because of the properties of
the normal distribution, values lying that far away from the
mean are extremely unlikely. Even if the mean were to
move right or left by 1.5σ at some point in the future (1.5
sigma shift), there is still a good safety cushion. This is why
Six Sigma aims to have processes where the mean is at
least 6σ away from the nearest specification limit.

10-55
10-56
The Goal of Six Sigma

Exhibit S9.29

10-57
Impact of 1.5σ Shift on 3σ Process

Exhibit S9.30A

10-58
Impact of 1.5σ Shift on 6σ Process

Exhibit S9.30B

10-59
Defect Rates for Different Levels of Sigma
(σ)
Assuming a 1.5 Shift in Actual Mean from
Design Mean

Exhibit S9.31

10-60
Improving Process Capability
 Simplify
 Standardize
 Mistake-proof
 Upgrade equipment
 Automate

10-61
 Taguchi Loss Function
Figure 10.17

Traditional
cost function
Cost

Taguchi
cost function

Lower Target Upper

spec spec

10-62
 Advanced Quality Tools
 Affinity diagrams
 Used to structure and clarify ideas by organizing
them according to their affinity, or similarity, to
each other.
 Interrelationship digraph
 Helps to sort out cause-and-effect relationships
when there are a large number of interrelated
issues that need to be better understood.

10-63
 Affinity
Diagram of
Team Learning
Objectives Exhibit S9.9

10-64
Relations Diagram for Late Hospital
Discharge

Exhibit S9.10

10-65
 Advanced Quality Tools
(cont’d)
 Tree diagram
 Helps determine ways to meet objectives by
breaking down a main goal into subgoals and
actions and identify the strategy to be taken.
 Matrix diagram
 Used to organize information that can be compared
on a variety of characteristics in order to make a
comparison, selection, or choice.
 Arranges elements of a problem or event in rows
and columns on a chart that shows relationships
among each pair of elements.
10-66
Basic Structure of a Tree Diagram

Exhibit S9.11

10-67
Matrix Diagram

Exhibit S9.12

10-68