Control Phase

Control Charts
Ashish Shrivastava
Assistant Professor
JECRC University

Causes of Variation

Variability is part of any process

Can be due to men,

materials, environment

methods,

machine,

Can be categorized in to two

Special causes
Assignable causes or
sporadic problem
Not inherent in the process
Will not affect all items
Can be due to use of a
wrong tool, improper
raw material, operator
error etc.

Common causes
Chance causes or chronic
problem
Inherent in the process
Affects all items, as it is
part of process design
Exists as long as the
process is not changed
Can be due to vibration of a
machine, fluctuation etc.

Introduction to Control Charts

Images obtained from:

http://pediatriccardiology.uchicago.edu/MP/ECG/ECG-Normal.jpg
http://research.vet.upenn.edu/Portals/75/third%20degree%20AV%20block.png

Introduction to Control Charts

A Control chart is a
Graphical tool
For monitoring the activity of an ongoing
process
It is called as Shewhart chart
Values of quality characteristic is plotted along
ordinate
Sample number or subgroups (in order of
time) is plotted along abscissa
Example for quality characteristic includes
Average length, Average service time etc.

Introduction to Control Charts Contd.

Image of Control chart obtained from: http://sixsigmaindonesia.com/blog/?p=56

Introduction to Control Charts Contd..

A Control chart has
A centre line, represents the average value of the
characteristic
It indicates, where the process is centred
UCL and LCL, used to make decision regarding the
process
If the points plot within the control limits and do not
exhibit any identifiable pattern, then process is in
statistical control
If a point plots outside the control limits or if an
identifiable random pattern exists process is out of
statistical control

Control Charts - Rules

Rule 2

Rule 1

Two out of three

consecutive points fall
beyond the 2 limit (in
zone A or beyond), on the
same side of the centerline

Any single data point falls

outside the 3 limit from
the centerline (i.e., any
point that falls outside
Zone A, beyond either the
upper or lower control limit)

Control Charts - Rules

Rule 3
Four out of five consecutive
points fall beyond the 1
limit (in zone B or beyond),
on the same side of the
centerline

Rule 4
Eight consecutive points
fall on the same side of the
centerline (in zone C or
beyond)

Control Charts - Patterns

Control Charts - Patterns

Control Charts - Patterns

Image obtained from:

http://www.spcforexcel.com/files/images/xbarchartbag.png

Difference between Control vs

Tolerance vs Specification Limits
Control limits are derived for
process that is derived
based on sampling and may
include both special and
common cause variation
NTL is derived from the
natural process variability (i.e
common causes). It is based
on individual items
Specification limits are
determined externally,
for example by
customers or designers
There is no mathematical or
statistical relationship
between the control limits and
the specification limits

Mathematics behind Control

Charts

Statistical Basis for Control Charts

Centre line is based on the mean of
a process and is found by taking the
average of the values in sample
Role of Normal distribution
Value of the statistic plotted on a control chart
are assumed to have an approximately
normal distribution
Any sample from population distribution that is
uni-modal and symmetric, the central limit
theorem states that if the plotted statistic is a
sample average, it tend to have a normal
distribution

Statistical Basis Contd..

Statistical Basis Contd..

99.994%
99.73%
95.44%
68.26%

3
4

3
4

Statistical Basis Contd..

Why 3 limits?
Chosen in such a way that probability of the sample points
falling between them is approximately 1, if the process is
in statistical control
Normal distribution theory states that a sample statistic will fall
within the limits 99.74% of the time, if the process is in control

If the control limits are calculated from current

data, then it tells whether the process is presently
in control or not
If the control limits are calculated from previous
data, based on a process that was in control, it helps
to find whether the process has drifted out of control

Statistical Basis Contd..

An assumption in performance properties is that the
underlying distribution of quality characteristic is normal
If underlying distribution is not normal, sampling distributions can
be derived and exact probability limits obtained

Burr (1967) notes the usual normal theory control

limits are very robust to normality assumption
Schilling and Nelson (1976) indicate that in most cases,
samples of size 4 or 5 are sufficient to ensure reasonable
robustness to normality assumption for X chart
Sampling distribution of R is not symmetric, thus
symmetric 3-sigma limits are an approximation and -risk
is not 0.0027. R chart is more sensitive to departures from
normality than X chart.
Assumptions of normality and independence are not a
primary concern

Preliminary decisions
Type of chart
Quality characteristic
Variable control chart:
A single component has usually several quality characteristic
Not feasible to maintain control chart for all as decision-making
becomes difficult
Involves selecting a few vital characteristics and selection is
based on those that cause more non-conforming items and
increase cost

Attribute control chart: Nonconforming items / nonconformity

What will be considered a nonconformity should be defined
It depends on product, functional use and customer needs

Preliminary decisions contd.

Sample size
Variable
Size is normally between 4 and 10
Larger the sample size, better the chance of detecting small shifts
Based on factors like cost of inspection or cost of shipping a
nonconforming item to the customer etc.

Attribute
Should be large enough to allow non conformities or
non conforming items to be observed in the sample
Eg. Non conformance rate is 2.5%, while sample size is 25, average
no. of non conforming per sample is 0.625
Small sample size leads to misleading inference

Frequency of sampling
Depends on production rate, cost of sampling etc.
As process is brought to control, frequency of sampling
decreases

Preliminary decisions Contd..

Choice of measuring instruments
Accuracy of the instruments directly influences the
quality of the data calculated
Should be calibrated and tested
Characteristic being controlled, the desired degree of
measurement precision, both impacts the selection

Design of data recording forms

Recording forms should be designed based on the
control chart to be used
Includes sample no., date, time, raw values of data,
part name, lot number, operator name, machine,
gauge, unit of measurement, specification etc.

Control chart - Working

Collect samples and get the measurement for variable/attribute

quality characteristic, which is of interest for the observed process

Calculate the mean, range, standard deviation for each sample as per
the type of control chart being used

Calculate the mean of sample means or sample ranges or

sample standard deviation as per the requirement

Use the suitable formulae to get the control limits

Plot all the sample means and sample ranges in separate chart
along with its associated control limits

Delete the out of control points for which remedial action has been
taken to remove special causes and the remaining samples are
used to obtain revised limits

Errors in making inferences

Making inferences from a

control chart is similar to testing
a hypothesis
Control limits are the critical
points that separate the rejection
and acceptance region
If a sample values fall above
UCL or below LCL, then we
reject null hypothesis or say that
process is out of control
Type I error
Inferring that a process is out of
control, when it is actually in control
Probability of Type I error is given
by

Type II error
Inferring that a process is in control,
when it is really out of control
All points may fall within limit, but
process mean has changed

Variable Control Charts

Image obtained from: http://www.technotraits.com/wp-content/uploads/2008/08/error2.jpg

Variable Control Charts

Types of charts
Xbar-R chart (Mean Range chart)
Xbar-s chart (Mean Std. deviation chart)
I-MR chart (Individual value-moving range
chart)
Cumulative Sum (CUSUM) Chart

Mostly, a pair of charts are drawn

Xbar-R chart is commonly used

Formulae for Xbar R Chart

Image obtained from: http://sarisawaged.tripod.com/sitebuildercontent/sitebuilderpictures/Contro5.gif

Example: Xbar R Chart

Time taken to process a
transaction
is
the
quality characteristic
5 customers are randomly
chosen every 1 hour from
morning 8 AM
Time taken for 9 different
are shown below
Develop an Xbar and R
chart and plot the values.
What comments can you
make about the process?

Photograph obtained from: http://newsimg.bbc.co.uk/media/images/42461000/jpg/_42461230_gazasalaryafp416.jpg

Example: Xbar R Chart Contd.

Example: Xbar R Chart Contd.

Example contd..

Example contd.

Example contd..

Refer Exhibit 9A.6, Page No. 398

In this case, D3 = 0 for n = 5 and
hence LCL = 0

Example contd..

Example contd..

Control chart for Mean and Range

Contd..
Plot the values of the range on the chart and find
whether points are in statistical control
An R chart is analysed before X-bar chart to determine
out of control situations, as R chart reflect process
variability, which should be brought to control.
If R chart shows out of control, then the X-bar chart is
meaningless

Delete the out of control points for which

remedial action has been taken to remove
special causes and the remaining samples are
used to obtain revised limits

Control chart for Mean and Range

Contd..
A point of interest is about the point that falls below
the LCL, when LCL is greater than zero
These points are desirable because they indicate
unusually small variability, within the sample and might
be due to special causes
This condition helps in further reducing our
process variability
Why two charts?
X bar chart monitors the mean between sample values
R chart monitors the variation within sample.

Control chart for Mean and Standard

deviation (X-bar and s chart)
R-chart is easy to
construct and use, but a
standard deviation chart
(s-chart) is preferable for
large sample sizes (> 10)
s

Reason is that range

accounts only for the s maximum and
minimum
sample values and is less
effective for large
samples, where as
sample standard
deviation serves as a
better measure of process
variability

Example
The advertised weight of frozen food packages
is 16oz. Random samples of size 8 are selected
from the output and weighed. The sample mean
and standard deviation are calculated.
Information on 25 such samples yields the
following:

25

I1

X I 398

25

I1

SI 3.00

Develop the suitable control limits.

Find the 1 control limits for each chart

Factors for variable control chart

Example contd.
Data given
No. of samples (g) = 25 and
Sample size (n) = 8

a.
Centre Line of the X-bar chart and s chart =
25 XI
X

398/ 25 15.92

I1

25 SI

I1

3/ 25 0.12

Control limits of the X-bar and s chart =

From table, A3 = 1.099, B4 = 1. 815, B3 = 0.185
UCLX X A3

15.92 1.099 X 0.12 16.052

UCL B
S

LCLX X A3 S 15.92 1.099 X 0.12 15.788

1.815 0.12 0.2178

LCLS B3S 0.815 0.12 0.0222

Example contd.
b.
One sigma control limits for s chart
Distance between CL and UCL for s chart is (0.2178
0.12) = 0.0978
= 0.12 +/- 2*(0.0978 / 3) = (0.0548, 0.1852)

One sigma control limits for X-bar chart

Distance between CL and UCL for X-bar chart is
(16.052 15.92) = 0.1379
= 15.92 +/- 2*(0.1379 / 3) = (15.828, 16.012)

Attribute control charts

Associated Terminologies
Nonconformity
A quality
characteristic that
does not meet its
stipulated
specifications
Eg. Thickness of a
plate: 5 +/- 0.1, but
if it is 5.2, then it is
just a nonconformity
Also called as
defect

Nonconforming unit
A unit that has one or
more nonconformities
such that the unit is
unable to meet the
intended standards
and is unable to
function as required
Eg. A plate having
both thickness and
length failing to meet
specifications
Also called as defective
unit

Types of attribute control chart

c Chart

u Chart

np Chart

p Chart

Types of Attribute Control Charts

Types of charts
Based on binomial distribution
Proportion of non conforming items (p-chart)
No. of non conforming items (np-chart)
Used in situation where the size of the sample is always
constant

Based on Poisson distribution

Total number of non conformities items (c-chart)
Non conformities per unit chart (u-chart)
Used in situation where the size of the sample varies from
sample to sample

Based on weights
Chart for Demerit per unit (U-chart)
Deals with combining nonconformities on a weighted basis,
based on the severity of the non conformity

Advantages and Disadvantages

A variable quality characteristic can be measured
as an attribute, because of limited time, money,
worker availability etc., if attribute information is
alone sufficient
Advantages
For a given product, if variable charts are used for each of its
quality characteristic, no. of charts constructed will be very
high and in such case, attribute control chart provides overall
quality information
Can be used to summarize information about several
components of a product
Attributes are encountered at all levels of an organization
(company, plant, department, operator level etc), but variable
charts are used at lower only (machine level)
Helps in going from the general to more focussed level

Advantages and Disadvantages Contd..

Disadvantages
Attribute information does not indicate the degree to
which specification are met or not met, but variable
provide more information a process and performance
Variable chart can forewarn, when the process goes out
of control, while attribute does not detect a lack of control,
until the process parameters are changed
When specification limit are equal or tighter than the
inherent variability of the process, attribute chart indicates
out of control process
Attribute chart requires larger sample size, to ensure
adequate protection against the certain level of
process changes

If no historical information is available, attribute

control charts are used first and once problem area
is identified, it is replaced by a variable control chart

Attribute control charts for

nonconforming items

Proportion of non conforming

items chart (p-chart)
Assumptions
Sample must be identical
Samples are independent
For a given sample, proportion of non conforming is
given by
X
P
N
where x = no .of nonconforming item in sample, n =
sample size
Probability of obtaining a non conforming item must be
constant

p-chart Contd..
Average of these individual sample proportions non
conforming is used as the centre line
g = no. of samples
G
G
X
xi = no. of non conforming items
I
PI
CL
P

Variance is calculated from

VAR( P)

P(1 P)
N

I1
G

I1

NG

p-chart Contd..
Control limits are given by

UCL P P 3.
LCL P P 3.

P(1 P)
N
P(1 P)
N

Control limits can also be placed depending on

the confidence interval
3 sigma control limits represent limits drawn at a confidence
level of 99.64%

p-chart Contd..
If sample size varies, control limits are
given by

Example of Constructing a p-Chart:

Required Data

A leading biscuit manufacturer in

India in recent times have got
customer complaints related to overbaking of biscuits. A control chart is
established to track the occurrence
of this problem and eliminate the
same by identifying the root cause
of the problem. Random samples of
100 biscuits are taken every 2 hours
in a shift and the records for 15
samples are shown in the Table
Identify the control limits
Comment about this biscuit making
process and list out possible
causes for the same

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

Sample
size (n)
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100

No. of overbaked
biscuits (x)
4
2
5
3
6
4
3
7
1
2
3
2
2
8
3

Statistical Process Control

Formulas: p-Chart
Given: p =

Total N um ber of D efectives

Total N um ber of O bservations

Sp =

p (1 - p)
N

Compute control limits:

UCL = p + z Sp
LCL = p - z Sp

Example of Constructing a p-chart

Step 1. Calculate the
sample proportions,
p (these are what
can be plotted on
the p-chart) for each
sample

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

n
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100

Defectives
4
2
5
3
6
4
3
7
1
2
3
2
2
8
3

p
0.04
0.02
0.05
0.03
0.06
0.04
0.03
0.07
0.01
0.02
0.03
0.02
0.02
0.08
0.03

Example of Constructing a p-chart

2. Calculate the average of the sample proportions

p=

55

= 0.036 1500

3. Calculate the standard deviation of the

sample proportion

S
p

= p (1 - p) =
N

.036(1 - .036) = .0188

100

Example of Constructing a p-chart

Step 4. Calculate the control limits
UCL = p + z S p
LCL = p - z Sp

.036 3(.0188)

UCL = 0.0924
LCL = -0.0204 (or 0)

Example of Constructing a p-Chart

Step 5. Plot the individual sample proportions, the
average of the proportions, and the control limits
0.1

UCL

0.09
0.08
0.07
0.06

0.05
0.04

CL

0.03
0.02
0.01
0
1

Observation

10

11

12

13

14

15

LCL

Example of Constructing a p-Chart

Inferences
All the plotted points lie within the UCL and
LCL
Hence the process is in control
No special causes are present

Causes for over-baked/burnt biscuits

Poor temperature control in the oven
Wrong temperature setting
Operator fails to remove the biscuit tray in time
Uneven heating within the oven

Chart for No. of nonconforming items (np-chart)

Based on the count of no. of nonconforming items
in sample
Operator may feel easy to use this chart
Assumptions used in p-chart is applicable for this
chart also
Drawback of np-chart
If sample size changes, the centre line and control limit
changes, and hence inference becomes difficult
G

Construction
Centre line is given by
th
xi = no. of non-conforming for i sample
g = no. of samples
n = sample size

CLNP N P

X
I

I1

Standard deviation of number of nonconforming is

UCL and LCL is given by
UCL NP N P 3. N P (1 P )
LCL NP N P 3. N P (1 P )

NP N P (1 P )

Attribute control charts

for nonconformities

Control chart for the number of

nonconformities (c-chart)
A c-chart is used to track the total no. of non
conformities in a sample of constant size

Example: Count no. of defects (scratches, chips etc.)

in each chair of a sample of 100 chairs

Cbar is the average number of counts in each

subgroup/sample.

If the LCL is less than zero, by convention it will set to

equal zero, because a negative count is impossible.

c-chart contd.
Assumptions
Occurrences of non conformities must be independent of
each other
Each sample should have an equal likelihood of the
occurrence of non conformities
Occurrence of nonconformities is assumed to follow a
Poisson distribution
Poisson distribution is well suited to model the number of
events that happen over a specified amount of time,
space or volume

c-chart Example

The number of scratch marks for a particular piece of

furniture is recorded for samples of size 10. The
results for 10 samples are shown in the table below:
Sample

10

No. of scratch marks

12

17

Construct a suitable attribute control chart.

Revise the control limit assuming special causes for point
outside the limits

Image obtained from: http://c487397.r97.cf2.rackcdn.com/wp-content/uploads/2011/01/scratch.jpg

Example contd.
Sample

10

CBAR

No. of scratch marks

12

17

7.7

UCL

LCL

16.025

-0.625 OR 0

Number of scratch marks

c- chart
18
16
14
12
10
8
6
4
2
0
1

10

Sample No.

From this figure, it can be found that the point number 7 is out of control.
Assuming special causes and eliminating it, the control chart is revised.
Revised control limits are as follows:
Cbar = 6.67
UCL 14.413
LCL = -1.079 or 0

Control chart for the number of

nonconformities per unit (u-chart)

Used, when the sample size changes

The control limits change as the sample size varies, but centre
line remains same
When sample size varies, the no. of non conformities per unit for
the ith sample is given by ui = ci / ni
th

ci is the no. of non conformities in the i sample

ni is the size of ith sample and it may not be a integer

Average no. of nonconformities per unit and the control limits are
given by

CL

I 1
G

I 1

UCL U U 3.
LCL U U 3.

As the sample sizes increases, control limits draws closer

If ni = 1, all formula will be equal to that of c - chart

U
NI
U
NI

Application in Services

Applying SPC to service

Hospitals
timeliness and 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 of-stock items,
quality of food items, cleanliness, customer complaints, checkout
register errors

Airlines
flight delays, lost luggage and luggage handling, waiting time a
at ticket counters and check-in, agent and flight attendant
courtesy, accurate flight information, passenger cabin
cleanliness and maintenance

Summary

Thank you!!