PROCESS CONTROL

Starts with measuring an important

variable. This can be a
Product attribute
Diameter of a metal component, weight
of a bag of potato chip
Process Attribute
Temperature in a restaurants oven,
length of waiting time in a ticket booth,
pressure applied in a moulding process

CONTROL LIMITS ARE BASED

ON SAMPLING DISTRIBUTION

STATISTICAL PROCESS
CONTROL (SPC)
A statistical process control involves testing a random
sample of output from a process to determine whether
the process is producing items within a pre-selected
range
SPC uses statistical tools to observe the performance
of the production process in order to detect significant
variations before they result in the production of a substandard article.
SPC is about monitoring consistency and repeatability
of a process

MAJOR OBJECTIVES
OF SPC
Quickly detect the occurrence of assignable
causes of process so that investigation of the
process and corrective action may be undertaken
before non-conforming units are manufactured
Reducing variability in the process

WHY QUALITY PROBLEMS?

Variation due to two reasons
Common Cause or random variation
Assignable or Special cause or
controllable variation

STATISTICAL PROCESS
CONTROL TOOLS (SPC)
Variation in output is due to:
Common causes (also known as natural variation)
Inherent variation present in every process
Variations are created by combined influences of
countless minor factors
Causes may difficult to distinguish or wholly
unidentifiable
Resulting degree of variation is minor
Assignable causes (known as special variation)
Variations due to specific causes
A process subject to assignable variation is out of control

STATISTICAL PROCESS
CONTROL TOOLS (SPC)
Control (or in control or stable)
A process that exhibits only common cause
variation is said to be in control or stable
A process is said to be out of control when it
exhibits assignable variation
Examples: less experienced worker has replaced
an experienced worker, machine malfunctioning,
change of machine settings

Natural and Assignable Variation

Process Control: Three Types of Process Outputs

RELATIONSHIP BETWEEN POPULATION AND

SAMPLING DISTRIBUTION

CONTROL CHARTS
A control chart is a time ordered plot of sample statistics

Sometimes called the voice of the process

Graphical display of a quality characteristics (for example, level of beer in

each bottle in a bottling plant)

Distinguish between random and non-random variability

Chart contains a center line and two limits

Upper control limit

Lower control limit
Control limits are statistical limits that reflect the extent to which sample
statistics (means, ranges) can vary due to randomness

If the process is in control, all sample points will fall between them

As long as points fall within control limits the process is in statistical

control

However, any point outside limits investigate the assignable causes

CONTROL CHARTS
If all the points plot inside the limits, but
behave in a nonrandom manner
indication that process is out of control and
needs investigation

A CONTROL CHART

IN STATISTICAL CONTROL
A process that is operating with only
chance cause of variation present is said
to be in statistical control
If the process is in control, all the plotted
points should have an essentially
random pattern

REASONS FOR POPULARITY

OF CONTROL CHARTS
Proven technique for improving
productivity
Effective in defect prevention
Prevents unnecessary process adjustment
Provides diagnostic information

Provides information about process

capability

STATISTICAL PROCESS
CONTROL TOOLS
Control Charts for variables (Characteristics that are expressed on a
numerical scale: density, weight, diameter, resistance, length, time,
volume)
X-bar Chart and R-Chart
X-bar chart for process average
R-chart for process variability
Control Charts for attributes (characteristic that cant be measured
on a numerical scale: smell of cologne acceptable or not
acceptable, color of a fabric acceptable or not)
p-chart and c-chart
p-charts for percent defective in a sample
c-charts for counts (e.g. # of defects)

SPC TOOLS
Control Charts for variables (X-chart, R-Chart)
Variables data are measured on continuous scale
Length, width, weight
Temperature
Voltage
Viscosity
Amount of time needed to complete a task

R-CHART CONTROL
LIMITS

D3, D4 = constants that provide 3 standard deviations (3) limits

for a given sample size

X-BAR CHART LIMITS

A2 = constant to provide three sigma limits for the sample mean

VARIABLE CONTROL CHART CONSTANTS

Sample Size

X-bar chart

R-chart

R-chart

n
2
3
4
5
6
7
8
9
10

A2
1.880
1.023
0.729
0.577
0.483
0.419
0.373
0.337
0.308

D3
0
0
0
0
0
0.076
0.136
0.184
0.223

D4
3.267
2.574
2.282
2.114
2.004
1.924
1.864
1.816
1.777

Source: Grant and Leavenworth

STEPS IN DEVELOPING XBAR AND R-CHART

Collect data on the variable measured (time,
weight, diameter). Collect at least 20-25 samples
randomly. Sample size should be of 4 to 5 units.
Compute range for each sample, and average Rbar
Calculate the UCL and LCL
Plot the sample ranges. If all are in control,
process is in statistical control
Calculate UCL and LCL for x-bar chart
Plot the sample means. If all are in control,
process is in statistical control.

ZONES FOR
IDENTIFICATION OF
NONRANDOM PATTERN

CONTROL CHART
PATTERNS

CONTROL CHART
PATTERNS

PATTERN RECOGNITION
IN CONTROL CHARTS
Recognizing non-random patterns on the
control chart
One point plots outside 3 limits
Two or three consecutive points plot beyond 2 limits
Four out of 5 consecutive points plot at a distance of 1 or
beyond from the centre line
Eight consecutive points on one side of centre line

P-CHART
Control charts for attributes
p-chart measures % defective items or proportion defective
items in a sample

Total # defects from all samples

p-bar = ---------------------------------------# samples Sample size
Appropriate when data consists of two categories of items
Good or bad, pass or fail
Examples: # bad light bulbs and good light bulbs in a
given lot
# of bad glass bottles and good glass bottles

P-CHART LIMITS

C-CHART
Appropriate when number of defects are counted
because not possible to compute proportion
defective
Examples
Number of accidents per day
Number of crimes committed in a month
Blemishes on a desk
Complaints in a day
Typo errors in a chapter of the text book
# customer invoice errors

C-CHART LIMITS

C-bar = average no. of defects per unit = Total number of defects

No of samples

PROCESS CAPABILITY
Specifications: A range of values imposed by
designers of the product or service based on
customer requirements
Control limits and based on production process,
and they reflect process variability due to
randomness alone
Process variability: Natural or inherent variability
in a process due to randomness
Process capability: The inherent variability of
process output relative to the variation allowed by
the design specifications

MEASURES OF
PROCESS CAPABILITY
Measures of Process Capability
Process Capability Ratio
Process Capability Index

PROCESS CAPABILITY
RATIO
Cp = (Upper Spec Lower Spec) / 6
If Cp < 1, process range > tolerance range

Process not capable of producing within design

specifications
If Cp = 1, Tolerance range and process range are
same
If Cp > 1, Tolerance range > process range
A desirable situation
Ideally Cp > 1.33

PROCESS CAPABILITY

PROCESS CAPABILITY

PROCESS CAPABILITY
Cp does not take into account where
the process mean is located relative
to the specifications
Cp simply measures the spread of the
specifications relative to the six sigma
spread in the process

PROCESS CAPABILITY
INDEX

Cpk measures if the process mean has shifted away from the design target, and which
Direction it has shifted --- that is, if it is off center. Ideally, Cpk = 1.33

Generally, if Cp = Cpk, the process is centered at the midpoint of the specifications

When Cpk < Cp, the process is off center

PROCESS CAPABILITY
INDEX
Cpk = 1, indicates the process variation is
within the upper and lower control limits
Cpk>1, process is capable of meeting
design specifications
Cpk<1, process has moved closer to one of
the upper or lower design specifications,
and it generate defects
When, Cpk = Cp, process mean is centered
on the design target

PROCESS CAPABILITY
(SEQUENTIAL STEPS)
1.

Calculate Cpk to check centrality

2.

Calculate Cp to check whether the process variation are

within design specifications