ATI,CHENNAI

CONTROL CHARTS

• A statistical tool to ensure that process is stable or in control

• A statistical tool to detect the presence of Assignable Causes in the process.
 ATI,CHENNAI
CONTROL CHARTS

Theory of Control Charts

For Normally distributed Data

If the process is stable then
Variation will be between Mean ± 3 x SD
 ATI,CHENNAI
CONTROL CHARTS

Control Charts

A Graphical Tool with three horizontal lines

1. Lower Control Limit (LCL)
2. Center Line (CL)
3. Upper Control Limit (UCL)

Control Chart
UCL

CL

LCL
 ATI,CHENNAI
CONTROL CHARTS

Control Charts

UCL & LCL are such that

if the value lies between UCL & LCL then the process is stable or in
control

Control Chart
UCL

CL

LCL
 ATI,CHENNAI
CONTROL CHARTS

Control Charts

For Normal Data

UCL =Mean + 3 x SD
CL = Mean
LCL = Mean – 3 x SD

Control Chart
UCL

CL

LCL
 ATI,CHENNAI
CONTROL CHARTS

Control Charts: Working

1. Calculate the Control Limits from past data

2. Plot the values in the chart
3. If the values are within the limits, the process is stable. Otherwise not.
 ATI,CHENNAI
CONTROL CHARTS

Types of Control Charts

Continuous Data
• Xbar & R Chart
• Individual X & Moving range Chart

Discrete Data
Control Chart for Defectives
• p chart
• np chart
Control Chart for Defects
• c chart
• u chart
 ATI,CHENNAI
CONTROL CHARTS

Control Charts: Continuous Data

If the data is Normal then

UCL = Mean + 3 x SD
CL = Mean
LCL = Mean – 3 x SD

But all continuous data are not Normal

Use Central Limit theorem
 ATI,CHENNAI
CONTROL CHARTS

Control Charts: Continuous Data

Central Limit Theorem

The averages of continuous data will be asymptotically normally
distributed even if the individual values are not Normally
distributed
 ATI,CHENNAI
CONTROL CHARTS

Central Limit Theorem : Example

Consider the following 9 sets of data

Group 1 2 3 4 5 6 7 8 9
1 2 6 4 2 2 2 5 4 6
2 2 6 4 3 4 2 5 4 4
3 2 6 4 3 4 3 5 5 4
4 2 6 4 3 4 4 5 6 4
5 2 6 4 4 6 4 5 6 2
Mean 2 6 4 3 4 3 5 5 4
 ATI,CHENNAI
CONTROL CHARTS

Central Limit Theorem : Example

 ATI,CHENNAI
CONTROL CHARTS

Central Limit Theorem : Example

Conclusion:
Instead of Individual values, as far as possible plot averages in control charts
Hence Xbar – R chart
 ATI,CHENNAI
CONTROL CHARTS

Xbar – R Chart: Methodology

• Decide on the Total Number of Samples N . (N >19)

• Decide on Sub Group Size n. (n > 3)
• Decide on Frequency of Sampling. ( eg: Once in a hour, Once in 2 hours,
Once in every 50 items, etc.)
 ATI,CHENNAI
CONTROL CHARTS

Xbar – R Chart: Rational Sub Grouping

Items within a sub groups should be taken Rationally not Randomly

Suppose Frequency of Sampling be Once in a hour & Sub Group Size n : 5
Then for a particular hour say 10:00 am

Random Sampling
Choose 5 pieces randomly from all those items processed in 9:00 – 10:00

Rational Sampling
Choose 5 consecutive pieces processed at 10:00 am
 ATI,CHENNAI
CONTROL CHARTS

Xbar – R Chart: Rational Sub Grouping

Random Sampling
• Variation within the items in a sub group will be maximum
• Variation between items in different subgroups will be minimum

Rational Sampling
• Variation within the items in a sub group will be minimum
• Variation between items in different sub groups will be maximum

For Control Charts: Sampling should be Rational

 ATI,CHENNAI
CONTROL CHARTS

Xbar – R Chart: Example

Process: Turning Characteristic: Inner diameter
Sample Size N: 9 Sub Group Size n: 4
Frequency of Sampling: Once in a Hour
Step 1: Collect Data
Sample No. Hour x1 x2 x3 x4
1 8:00 5.00 5.01 4.98 5.00
2 9:00 5.01 4.98 5.00 5.00
3 10:00 5.02 5.01 5.00 5.00
4 11:00 5.00 5.00 5.00 5.00
5 12:00 4.98 4.98 5.01 4.99
6 13:00 5.02 4.99 5.00 4.98
7 14:00 4.99 4.99 4.98 4.98
8 15:00 5.00 5.01 5.02 5.00
9 16:00 4.98 5.00 5.01 4.98
 ATI,CHENNAI
CONTROL CHARTS

Xbar – R Chart: Example

Step 2: Calculate Sub Group Mean & Range

Sample No. Hour x1 x2 x3 x4 Mean Range

1 8:00 5.00 5.01 4.98 5.00 4.998 0.03
2 9:00 5.01 4.98 5.00 5.00 4.998 0.03
3 10:00 5.02 5.01 5.00 5.00 5.008 0.02
4 11:00 5.00 5.00 5.00 5.00 5.00 0.00
5 12:00 4.98 4.98 5.01 4.99 4.990 0.03
6 13:00 5.02 4.99 5.00 4.98 4.998 0.04
7 14:00 4.99 4.99 4.98 4.98 4.985 0.01
8 15:00 5.00 5.01 5.02 5.00 5.008 0.02
9 16:00 4.98 5.00 5.01 4.98 4.993 0.03
 ATI,CHENNAI
CONTROL CHARTS

Xbar – R Chart: Example

Step 3: Calculate Control Limits for R Chart

Center Line
CL = Mean = Rbar = Sum of all Range Values / Total Number
of Values
= 0.21 / 9 = 0.0233

Upper Control Limit

UCL = Mean + 3 SD = D4 Rbar, For n = 4, D4 = 2.282
= 2.282 x 0.0233 = 0.053

Lower Control Limit

LCL = Mean - 3 SD = D3 Rbar, For n = 4, D3 = 0
= 0 x 0.0233 = 0.0
 ATI,CHENNAI
CONTROL CHARTS

Xbar – R Chart: Example

Step 4: Plot R values in R chart as shown below:

Step 5: If any value is beyond Control Limits, Do Homogenization

Homogenization:
Remove the out of control value
Recalculate the limits
 ATI,CHENNAI
CONTROL CHARTS

Xbar – R Chart: Example

Step 6: Calculate Control Limits for Xbar Chart

Center Line
CL = Mean = Xdoublebar = Sum of all Means / Total Number
of Values
= 44.975 / 9 = 4.997

Upper Control Limit

UCL = Mean + 3 SD = xdoublebar + A2Rbar, For n = 4, A2 = 0.729
= 4.997 + 0.729 x 0.0233 = 5.014

Lower Control Limit

LCL = Mean - 3 SD = xdoublebar - A2Rbar, For n = 4, A2 = 0.729

= 4.997 - 0.729 x 0.0233 = 4.98

 ATI,CHENNAI
CONTROL CHARTS

Xbar – R Chart: Example

Step 7: Plot mean values in xbar chart as shown below:

Step 8: If any value is beyond Control Limits, Do Homogenization

 ATI,CHENNAI
CONTROL CHARTS

Control Chart Constants

n A2 D3 D4
2 1.880 0.00 3.267
3 1.023 0.00 2.574
4 0.729 0.00 2.282
5 0.577 0.00 2.114
 ATI,CHENNAI
CONTROL CHARTS

Xbar – R Chart: Exercise

The following are the data on Time (in Minutes) to Process Transactions in a
BPO company. Construct an Xbar – R chart to monitor the process

Sample x1 x2 x3 Sample x1 x2 x3
Number Number
1 6.0 5.8 6.1 11 6.2 6.9 5.0
2 5.2 6.4 6.9 12 6.7 7.1 6.2
3 5.5 5.8 5.2 13 6.1 6.9 7.4
4 5.0 5.7 6.5 14 6.2 5.2 6.8
5 6.7 6.5 5.5 15 4.9 6.6 6.6
6 5.8 5.2 5.0 16 7.0 6.4 6.1
7 5.6 5.1 5.2 17 5.4 6.5 6.7
8 6.0 5.8 6.0 18 6.6 7.0 6.8
9 5.5 4.9 5.7 19 4.7 6.2 7.1
10 4.3 6.4 6.3 20 6.7 5.4 6.7
 ATI,CHENNAI
CONTROL CHARTS

Individual X & Moving Range Chart: Example

• For Short Run Process

• Not possible to collect data in Sub Groups

Process: Heat Treatment Characteristic: Case Depth

Sample Number Data

1 2.5
2 2.3
3 2.8
4 2.6
5 2.4
6 2.9
7 2.1
8 2.5
 ATI,CHENNAI
CONTROL CHARTS

Individual X & Moving Range Chart: Example

Step 2: Calculate Moving Range

Sample Number Date MR

1 2.5
2 2.3 0.2
3 2.8 0.5
4 2.6 0.2
5 2.4 0.2
6 2.9 0.5
7 2.1 0.8
8 2.5 0.4
 ATI,CHENNAI
CONTROL CHARTS

Individual X & Moving Range Chart: Example

Step 3: Calculate Control Limits for M R Chart

Center Line
CL = Mean = MRbar = Sum of all MR Values / Total Number
of Values
= 2.8 / 7 = 0.4

Upper Control Limit

UCL = Mean + 3 SD = D4 MRbar, For n = 2, D4 = 3.276
= 3.276 x 0.4 = 1.31

Lower Control Limit

LCL = Mean - 3 SD = D3 MRbar, For n = 2, D3 = 0
= 0 x 0.4 = 0.0
 ATI,CHENNAI
CONTROL CHARTS

Individual X & Moving Range Chart: Example

Step 4: Plot MR values in MR chart as shown below:

Step 5: If any value is beyond Control Limits, Do Homogenization

 ATI,CHENNAI
CONTROL CHARTS

Individual X & Moving Range Chart: Example

Step 6: Calculate Control Limits for Individual X Chart

Center Line
CL = Mean = Xbar = Sum of all Data / Total Number
of Values
= 20.1 / 8 = 2.512

Upper Control Limit

UCL = Mean + 3 SD = xbar + A2MRbar, For n = 1, A2 = 2.66
= 2.512 + 2.66 x 0.4 = 3.58

Lower Control Limit

LCL = Mean - 3 SD = xbar - A2MRbar, For n = 1, A2 = 2.66
= = 2.512 - 2.66 x 0.4 = 1.45
 ATI,CHENNAI
CONTROL CHARTS

Individual X & Moving Range Chart: Example

Step 7: Plot individual values in Individual X chart as shown below:

Step 8: If any value is beyond Control Limits, Do Homogenization

 ATI,CHENNAI
CONTROL CHARTS

Individual X & Moving Range Chart: Exercise

The data given below are surface Finish values of 30 jobs after chromium
plating. Construct an Individual X & Moving Range chart to monitor the
process?

0.078 0.079 0.077 0.076 0.074 0.072 0.069 0.075 0.078 0.077

0.075 0.078 0.08 0.081 0.08 0.079 0.082 0.073 0.078 0.074
0.072 0.075 0.068 0.073 0.074 0.081 0.076 0.08 0.074 0.07
 ATI,CHENNAI
CONTROL CHARTS

Control Charts for Defectives: np Chart

Used when sample size is constant

Example: Inspection results of video of the month shipment to customers for

10 consecutive days are given in table. The number of inspection
each day is constant and is equal to 1000. Construct np chart to
control the defectives?
Sample Number Number of Defectives

1 47
2 42
3 48
4 58
5 32
6 38
7 53
8 68
9 45
10 37
 ATI,CHENNAI
CONTROL CHARTS

np Chart : Calculation of Control Limits

CL = Mean
UCL = Mean + 3 SD
LCL = Mean – 3 SD
pbar = Sum of Defectives / total Number Inspected
= 468 /(1000*10) = 0.0468

Mean = npbar = 1000 x 0.0468 = 46.8

SD = npbar(1-pbar) = (1000 x (0.0468 x (1-0.0468))) = 6.67

CL = npbar = 1000 x 0.0468 = 46.8

UCL = Mean + 3 SD = 46.8 + 3 x 6.68 = 66.84
LCL = Mean - 3 SD = 46.8 - 3 x 6.68 = 26.76
 ATI,CHENNAI
CONTROL CHARTS

np chart: Example

Plot the number of defectives in np chart as shown below:

If any value is beyond Control Limits, Do Homogenization

 ATI,CHENNAI
CONTROL CHARTS

np Chart: Exercise

The following are the data on defectives in payment of dental insurance

claims. Control the dental insurance payment process with np chart. Sample
Size is 300
Sample Number of Sample Number of
Number Defectives Number Defectives

1 3 11 6
2 6 12 9
3 4 13 5
4 6 14 6
5 20 15 7
6 2 16 4
7 6 17 5
8 7 18 7
9 3 19 5
10 0 20 0
 ATI,CHENNAI
CONTROL CHARTS

Control Charts for Defectives: p Chart

Used when sample size is not constant

Example: The daily inspection results for electric carving knives are given
below. Construct a control chart to monitor the process ?
Sample Number Number Inspected Number of Defectives
1 500 5
2 550 6
3 700 8
4 625 9
5 700 7
6 550 8
7 450 10
8 600 6
9 475 9
10 650 6
 ATI,CHENNAI
CONTROL CHARTS

p Chart : Calculation of Control Limits

CL = Mean
UCL = Mean + 3 SD
LCL = Mean – 3 SD
pbar = Sum of Defectives / Total Number Inspected
= 74 / 5800 = 0.0128

Mean = pbar = 0.0128

SD = pbar(1-pbar) / ni
 ATI,CHENNAI
CONTROL CHARTS

p Chart : Calculation of Control Limits

Sample Number Number of p LCL UCL

Number Inspected Defectives
1 500 5 0.010 0 0.028
2 550 6 0.011 0 0.027
3 700 8 0.011 0 0.026
4 625 9 0.014 0 0.026
5 700 7 0.010 0 0.026
6 550 8 0.015 0 0.027
7 450 10 0.022 0 0.029
8 600 6 0.010 0 0.027
9 475 9 0.019 0 0.028
10 650 6 0.009 0 0.026
 ATI,CHENNAI
CONTROL CHARTS

p chart: Example

Plot the proportion of defectives in p chart as shown below:

If any value is beyond Control Limits, Do Homogenization

 ATI,CHENNAI
CONTROL CHARTS

p Chart: Exercise

Daily inspection results for the model 305 electric range assembly line are
given in the table. Construct a control chart to monitor the process?

Sample Number Number of Sample Number Number of

Number Inspected Defectives Number Inspected Defectives

1 171 31 11 181 38
2 167 6 12 115 33
3 170 8 13 165 26
4 135 13 14 189 15
5 137 26 15 165 16
6 170 30 16 170 35
7 45 3 17 175 12
8 155 11 18 167 6
9 195 30 19 141 50
10 180 36 20 159 26
 ATI,CHENNAI
CONTROL CHARTS

Control Charts for Defects: c Chart

Used when sample size is constant

Example: A leading bank has compiled the data in the table showing the
count of nonconformities for 1000 accounting transactions per day.
Construct a control chart to monitor the process?

Day Number of nonconformities

1 8
2 19
3 14
4 18
5 11
6 16
7 8
8 15
9 21
10 8
 ATI,CHENNAI
CONTROL CHARTS

c Chart : Calculation of Control Limits

CL = Mean
UCL = Mean + 3 SD
LCL = Mean – 3 SD
cbar = Sum of nonconformities / Total Number Inspected
= 138 /100 = 13.8

Mean = cbar = 13.8

SD = cbar = 13.8 = 3.71

CL = cbar = 13.8
UCL = Mean + 3 SD = 13.8 + 3 x 3.71 = 24.93
LCL = Mean - 3 SD = 13.8 - 3 x 3.71 = 2.67
 ATI,CHENNAI
CONTROL CHARTS

c chart: Example

Plot the number of nonconformities in c chart as shown below:

If any value is beyond Control Limits, Do Homogenization

 ATI,CHENNAI
CONTROL CHARTS

c Chart: Exercise

100 product labels are inspected every day for surface nonconformities. The
data for the past 20 days is given below. Construct a suitable control chart to
monitor the nonconformities
Day Number of Day Number of
Nonconformities Nonconformities

1 22 11 15
2 29 12 10
3 25 13 33
4 17 14 23
5 20 15 27
6 16 16 15
7 34 17 17
8 11 18 17
9 31 19 19
10 29 20 22
 ATI,CHENNAI
CONTROL CHARTS

Control Charts for Defects: u Chart

Used when sample size is not constant

Example: The inspection results for the surface finish of rolls of white paper
for 10 lots is given below. Construct a control chart to monitor the
process ?
Lot Number Number Inspected Number of Defects
1 10 45
2 10 51
3 10 36
4 9 48
5 10 42
6 10 5
7 10 33
8 8 27
9 8 31
10 8 22
 ATI,CHENNAI
CONTROL CHARTS

u Chart : Calculation of Control Limits

CL = Mean
UCL = Mean + 3 SD
LCL = Mean – 3 SD
ubar = Sum of Defects / Total Number Inspected
= 340 / 93 = 3.66

Mean =ubar = 3.66

SD = (ubar / ni)
 ATI,CHENNAI
CONTROL CHARTS

p Chart : Calculation of Control Limits

Lot Number Number of u LCL UCL

Number Inspected Defects
1 10 45 4.5 2.80 5.47
2 10 51 5.1 2.86 5.47
3 10 36 3.6 2.70 5.47
4 9 48 5.3 2.83 5.57
5 10 42 4.2 2.77 5.47
6 10 5 0.5 1.09 5.47
7 10 33 3.3 2.66 5.47
8 8 27 3.4 2.56 5.69
9 8 31 3.9 2.63 5.69
10 8 22 2.8 2.44 5.69
 ATI,CHENNAI
CONTROL CHARTS

u chart: Example

Plot the defects / unit (u) of in u chart as shown below:

If any value is beyond Control Limits, Do Homogenization

 ATI,CHENNAI
CONTROL CHARTS

u Chart: Exercise

Construct a suitable control chart for the data in the table for empty bottle
inspections of a soft – drink manufacturer?

Sample Number Number of Sample Number Number of

Number of bottles Defects Number of bottles Defects
inspected Inspected

1 40 45 11 52 55
2 40 40 12 52 74
3 40 33 13 52 43
4 40 43 14 52 61
5 40 62 15 40 43
6 52 79 16 40 32
7 52 60 17 40 45
8 52 50 18 40 33
9 52 73 19 40 50
10 52 54 20 52 28
 ATI,CHENNAI
CONTROL CHARTS

Some More Out of Control Cases

• 7 consecutive values in one side of center line

• 6 consecutive values steadily increasing or decreasing
• 2 out of 3 values in Mean ± 2 x SD
• 4 out of 5 values in Mean ± 1 x SD