Lecture on control chart

Real-life Problems

Problem 1:
Suppose a new bottle-making machine has just been installed. The machine is set to produce
41-ounce bottles, but variation in the weights of these bottles is expected. This expected
variation is due to a number of factors, namely, the temperature of the glass, the temperature
of the molds and the composition of the glass mixture. The quality control took a sample of
five bottles every hour and recorded the weight of each of the five bottles she selected. The
data for the samples taken at 8, 9, 10 and 11 A.M. are given below. How is a mean chart
constructed?

Solution
Weights of five bottles selected at random (in ounces)
Time Bottle Range S (S.D)
1 2 3 4 5 X (R)
8 A.M 41 43 42 41 43 42 2 1.00
9 A.M 39 40 40 39 42 40 3 1.22
10 A.M 41 44 43 46 41 43 5 2.12
11 A.M 38 39 40 39 39 39 2 0.71
Total=∑ 164 12 5.05

164 12 5.05
Here x= = 41, R= = 3, S= = 1.26
4 4 4

Here  and  are not known. So, the control limits are as follows:

(a) LCL = x − A1S = 41 − 1.596  1.26

= 41 − 2.011 = 38.989

CL = x = 41

𝑈𝐶𝐿 = 𝑥̄̄ + 𝐴1 𝑆̄ = 41 + 2.011

= 43.011

(b) LCL = x − A2 R = 41 − .577  3

= 39.269

1
 CL = x = 41

𝑈𝐶𝐿 = 𝑥̄̄ + 𝐴2 𝑅̄ = 42.731

For the data given in example 1, construct a range chart.

Solution:

We know R = 3.

From the table of D3 and D4 , we know that for n = 5, D3 = 0, D4 = 2.115.

 LCL = D3 R = 0(3) = 0)
CL = 3
UCL = D4 R = (2.115)(3) = 6.345

Range chart for the weights of bottles

Comment: All the ranges lie within the control limits. So, the process is under
control.

So, we can portray these limits in the following charts

Mean chart for the weights of Bottles, here we see that all the points are within
the control limits. So, the system is in control.
From Range chart, we see that the two points are outside the control limits. So,
the process is out of control.

2
 Problem 2: For the following data construct Mean chart and comment:
Sample Measurements of individual items S R
Number X1 X2 X3 X4 X
1 20 22 16 21 19.75 2.630 5
2 24 18 20 22 21.00 2.582 6
3 19 15 21 22 19.25 3.096 7
4 15 20 22 16 18.25 3.304 7
5 21 19 23 15 19.50 3.416 8
6 14 16 21 23 18.50 4.203 9
7 23 14 15 20 18.00 4.243 9
8 17 19 22 24 20.50 3.109 7
9 18 20 23 21 20.50 2.082 5
10 24 18 15 22 19.75 4.031 8
11 22 17 16 20 18.75 2.754 6
12 24 18 19 20 20.25 2.630 6
13 22 17 21 15 18.75 3.304 7
14 14 20 24 16 18.50 4.435 10
15 21 24 16 21 20.50 3.317 8
Total=∑ 291.75 49.136 108

Solution

291.75 108 49.135

x= = 19.45, R = = 7.20 , S = = 3.276
15 15 15
Case (i) LCL = x − A1S Case (ii) LCL = x − A2 R
= 19.45 − (1.880)(3.276) = 19.45 − (0.729)(7.20)
= 19.45 − 6.159 = 19.45 − 5.2488
= 13.291 = 14.20

CL = x = 19.45 CL = x = 19.45

UCL = x + A1S = 25.609 UCL = x + A2 R = 24.670

Comment: The process is in control. Comment: The process is in control.

R Chart: Another control chart for variables is referred to as a range chart.

Logically it shows the variation in the sample ranges.
Control limits are:
LCL = D3 R

3
 CL = R
UCL = D4 R

Problem 3:
The table below gives the life (in hours) of 10 samples of 6 electric bulbs each
drawn at an interval of half an hour from a production process. Construct
control charts based on the mean. the values of these statistics are given below
the repetitive samples.
1 2 3 4 5 6 7 8 9 10
620 501 673 646 494 634 619 630 482 706
687 585 701 626 984 755 710 723 790 524
666 524 686 572 659 625 664 614 533 626
689 585 567 628 643 582 693 535 612 503
738 653 619 631 660 683 770 550 497 661
686 668 660 743 640 555 534 570 499 754
X (Mean) 681 586 651 641 680 639 665 604 569 629
R (Range) 118 167 134 171 490 200 236 188 309 251

Find whether the production process is in a state of statistical control

Mean chart: UCL = X + A2 R=634.5+(.483)(214.6)=738.15 and

LCL = X − A2 R=634.5-(.483)(214.6)=530.085
From the chart it is clear that no points lie outside the control limits. Therefore,
the process is well within the statistical control.

Range chart: UCL = R + D4 R=2.004 x 214.6=430.06 and

LCL = D3 R=0 x 214.6=0

From the chart it is clear that no points lie outside the control limits. Therefore,
the process is well within the statistical control.

4
Problem 4:
For the following data, construct percent defective chart or p − chart.
Lot size Number of Proportion defective or fraction
(n) defective defective
(np ) ( p)
100 5 .05
100 6 .06
100 7 .07
100 4 .04
100 8 .08
Total=∑ .30

Solution:

.30
Here p = = .06
5
p (1 − p )
 LCL = p − 3
n
.06  .94
= .06 − 3
100
= .06 − .0712
= −.0112 = 0

CL = .06

p (1 − p )
UCL = p + 3
n
= .1312

Percent defective or p chart reveals that process is in control.

np Charts or charts for number of defectives or charts for the number of

defectives.

LCL = np − 3 np (1 − p )
CL = np
UCL = np + 3 np (1 − p )

5
Example 5:
For the data in example 4, construct np chart.
Solution:
LCL = np − 3 np (1 − p )
= 100  .06 − 3 100  .06  .94
= 6 − 7.125
= −1.125
=0

CL = np = 6

UCL = np + 3 np (1 − p )
= 6 + 7.125
= 13.125

The process is under control.

Problem 6:
A quality control inspector checks a sample of defective tuners and counts the
number of defects per tuner which are given as 8, 5, 6, 4, 3, 8, 8, 10, 9 and 9.
How is a C − chart constructed?

Solution:
The formula for the control limits is

LCL = C − 3 C
CL = C

UCL = C + 3 C
Average
Where C =
Number of defects
8 + 5 + 6 + 4 + 3 + 8 + 8 + 10 + 9
C= =7
10

The UCL = C + 3 C = 7 + 7.94 = 14.94

6
LCL = 7 − 7.94 = −.94 = 0
CL = 7

Problem 7:
The following are the figures for the number of defectives of 10 lots Each
containing 100 items: 12, 10, 6, 8, 9, 9, 7, 10, 11, 8, draw the control chart for
fraction defective (p chart), plot the points on it. Comment on the state of
control of the process.
Solution
Lot No.: 1 2 3 4 5 6 7 8 9 10
Number of defectives (d): 12 10 6 8 9 9 7 10 11 8
Fraction defectives (d/n): .12 .10 .06 .08 .09 .09 .07 .10 .11 .08
 d = 90
 d / n = 0.90
p = 90 /1000 = .09

.09(1-.09)
UCL = p + 3 p (1 − p ) / n =.09+3 =0.176
100
LCL = p − 3 p (1 − p ) / n =0.004 .
Graph: p and control limits.
Comments: From the chart it is clear that no points lie outside the control
limits. Therefore, the process is well within the statistical control.

7
Problem 8:
The prices of cloth out of different rolls of equal length contained the following
number of defectives: 3, 4, 2, 0, 5, 7, 6, 3, 4 and 7. Find whether the production
process is in a state of statistical control (c-chart).

Solution:
36
The total number of defects=36, as c = , the central lies at 3.6.
10
Standard error of the number of defects = c = 1.897.

c  3 c = 3.6  3 3.6 = 3.6  3(1.896) .

UCL=9.291; LCL=-1.091. Since the limit cannot be negative, therefore LCL=0.
Graph:
Comments Since all points lie within the control limits, it may be concluded
that the process is under statistical control.

Problem 9:
In welding of seams, defects pinholes, cracks, cold laps, etc. A record was
making of the number of defects found in one seam each hour and is given
below
1.12.83 8 A.M. 2 12 A.M. 6
9 A.M. 4 1 P.M. 4
10 A.M. 7 2 P.M. 9
11 A.M. 3 3 P.M. 9
12 A.M. 1 3.12.83 8 A.M. 6
1 P.M. 4 9 A.M. 4
2 P.M. 8 10 A.M. 3
3 P.M. 9 11 A.M. 9
2.12.83 8 A.M. 5 12 A.M. 7
9 A. M. 3 1 P.M. 4
10 A.M. 7 2 P.M. 7
11 A.M. 11 3 P.M. 12
Draw the control chart for number of defects and give your comments.