STATISTICAL PROCESS

CONTROL (SPC)
NANI KURNIATI, PhD

DEPARTMENT OF INDUSTRIAL AND SYSTEM ENGINEERING

INSTITUT TEKNOLOGI SEPULUH NOPEMBER (ITS)

Industrial Engineering ITS

CONCEPT OF VARIANCE

Industrial Engineering ITS

 Meanings of variance in Q
Industrial Engineering ITS

Variance may change the

parameter of product
distribution e.g
• mean,
• standar deviation,
• mean and standard
deviation
 Data analysis
Industrial Engineering ITS

• Suppose the Q characteristics of product (or

service) inspected and measured
• Data collected need to be analyzed to
recognize its pattern
• The pattern could be done numerically and
graphically, with important parameter of
data is variance
• Therefore, describing variance is important
for quality analysis (as the modern
definitions of quality)
 Describing variance: graphical
Industrial Engineering ITS

Stem-and-Leaf Display
Data 1: Health Insurance Claims

Claim Days Claim Days Claim Days Claim Days

1 48 11 35 21 37 31 16
2 41 12 34 22 43 32 22
3 35 13 36 23 17 33 33
4 36 14 42 24 26 34 30
5 37 15 43 25 28 35 24
6 26 16 36 26 27 36 23
7 36 17 56 27 45 37 22
8 46 18 32 28 33 38 30
9 35 19 46 29 22 39 31
10 47 20 30 30 27 40 17
 Describing variation : graphical
Industrial Engineering ITS

Dot plot / run chart

Data 1: Health Insurance Claims
 Describing variation : graphical
Industrial Engineering ITS

Histogram
Example 2: Metal Thickness in Silicon Wafers

Layer Thickness on Semiconductor Wafers

438 450 487 451 452 441 444 461 432 471
413 450 430 437 465 444 471 453 431 458
444 450 446 444 466 458 471 452 455 445
468 459 450 453 473 454 458 438 447 463
445 466 456 434 471 437 459 445 454 423
472 470 433 454 464 443 449 435 435 451
474 457 455 448 478 465 462 454 425 440
454 441 459 435 446 435 460 428 449 442
455 450 423 432 459 444 445 454 449 441
449 445 455 441 464 457 437 434 452 439
 Describing variation : graphical
Industrial Engineering ITS

Histogram
 Describing variation : graphical
Industrial Engineering ITS

Pareto Chart
 Describing variation : graphical
Industrial Engineering ITS

Scatter Diagram
 Describing variation : numerical
Industrial Engineering ITS

Sample mean
x1 + x2 + x3... + xn i =1
x i
x= =
n n
n

Sample variance
 ( xi − x )2
s 2 = i =1
n −1

Sample standard deviation i =1

( xi − x )2
s=
n −1
 Statistics Descriptive: numerical
Industrial Engineering ITS

Box plot
Data 3: hole diameter
Hole diameters (in mm) in wing leading edge ribs
120.5 120.4 120.7
120.9 120.2 121.1
120.3 120.1 120.9
121.3 120.5 120.8
 Statistics Descriptive: numerical
Industrial Engineering ITS

Box plot
Data 4: quality index

13
 Basic SPC Tools
Industrial Engineering ITS

1. Histogram or stem-leaf plot

2. Check sheet
3. Pareto chart
4. Cause and effect diagram
5. Defect concentration diagram
or flow chart
or operation process chart (process mapping)
6. Scatter diagram
7. Control chart

14
 Describing variance and SPC
Industrial Engineering ITS

Among those seven tools, we already

discussed the following three tools:

1. Histogram / stem leaf diagram

2. Pareto chart
3. Scatter diagram / run chart
ANOTHER TOOLS OF 7

Industrial Engineering ITS

Industrial Engineering ITS

Check Sheet
Industrial Engineering ITS

Cause-and-Effect Diagram
 Defect Concentration Diagram
Industrial Engineering ITS
Industrial Engineering ITS

Flow process chart

THE LAST TOOLS : CONTROL CHART

Industrial Engineering ITS

 Basic principles of CC
Industrial Engineering ITS

• The control chart is a graphical display of a quality

characteristic that has been measured or computed from a
sample versus the sample number or time.
• contains :
– a center line that represents the average value of the quality
characteristic
– Two other horizontal lines, called the upper control limit (UCL) and
the lower control limit (LCL), are also shown on the chart.
 Process status
Industrial Engineering ITS

• As long as the points plot within the control limits, the

process is assumed to be in control (in statistical
control).
– No action is necessary.
• If a point that plots outside of the control limits is
interpreted as evidence that the process is out of
control.
– investigation and corrective action are required to find and
eliminate the assignable cause or causes responsible for this
behavior.
• Moreover, even all the points plot inside the limits, if
they behave in a systematic or nonrandom manner →
indication that the process is out of control.
 Example : photolithography to silicon wafer
Industrial Engineering ITS

• In semiconductor manufacturing, an important fabrication

step is photolithography, in which a light-sensitive
photoresist material is applied to the silicon wafer, the
circuit pattern is exposed on the resist typically through the
use of high-intensity UV light, and the unwanted resist
material is removed through a developing process.
• After the resist pattern is defined, the underlying material
is removed by either wet chemical or plasma etching.
• It is fairly typical to follow development with a hard-bake
process to increase resist adherence and etch resistance.
• An important quality characteristic in hard bake is the flow
width (in microns) of the resist, a measure of how much it
expands due to the baking process.
 Photolithography Example
Industrial Engineering ITS

sample wafers
number
𝑥ҧ𝑖
• Suppose that flow width (i)
1
1
1,3235
2
1,4128
3
1,6744
4
1,4573
5
1,6914 1,5119
R
0,3679
can be controlled at a 2
3
1,4314
1,4284
1,3592
1,4871
1,6075
1,4932
1,4666
1,4324
1,6109
1,5674
1,4951
1,4817
0,2517
0,139
mean of 1.5 microns, 4
5
1,5028
1,5604
1,6352
1,2735
1,3841
1,5265
1,2831
1,4363
1,5507
1,6441
1,4712
1,4882
0,3521
0,3706
and it is known that the 6
7
1,5955
1,6274
1,5451
1,5064
1,3574
1,8366
1,3281
1,4177
1,4198
1,5144
1,4492
1,5805
0,2674
0,4189
standard deviation is 8
9
1,419
1,3884
1,4303
1,7277
1,6637
1,5355
1,6067
1,5176
1,5519
1,3688
1,5343
1,5076
0,2447
0,3589
0.15 microns 10
11
1,4039
1,4158
1,6697
1,7667
1,5089
1,4278
1,4627
1,5928
1,522
1,4181
1,5134
1,5242
0,2658
0,3509
• Every hour, a sample 12
13
1,5821
1,2856
1,3355
1,4106
1,5777
1,4447
1,3908
1,6398
1,7559
1,1928
1,5284
1,3947
0,4204
0,447
(subgroup) of five wafers 14
15
1,4951
1,3589
1,4036
1,2863
1,5893
1,5996
1,6458
1,2497
1,4969
1,5471
1,5261
1,4083
0,2422
0,3499
is taken, the average flow 16 1,5747 1,5301 1,5171 1,1839 1,8662 1,5344 0,6823
17 1,368 1,7269 1,3957 1,5014 1,4449 1,4874 0,3589
width computed 18 1,4163 1,3861 1,3057 1,621 1,5573 1,4573 0,3153
19 1,5796 1,4185 1,6541 1,5116 1,7247 1,5777 0,3062
• Mean and standard 20 1,7106 1,4412 1,2361 1,382 1,7601 1,5060 0,524
21 1,4371 1,5051 1,3485 1,567 1,488 1,4691 0,2185
deviation of the sample 22 1,4738 1,5936 1,6583 1,4973 1,472 1,5390 0,1863
23 1,5917 1,4333 1,5551 1,5295 1,6866 1,5592 0,2533
average  x = 1.5056 24 1,6399 1,5243 1,5705 1,5563 1,553 1,5688 0,1156
25 1,5797 1,3663 1,624 1,3732 1,6887 1,5264 0,3224
 0.15
x = = = 0.0671 ෍ 𝑥ҧ𝑖 =
37,6402 8,1302
= ෍ 𝑅𝑖

n 5 𝑥Ӗ = 1,5056 0,32521 = 𝑅ധ
 General Model : Shewhart Control Chart
Industrial Engineering ITS

Let w = a sample statistic that measures some quality characteristic of

interest (w could mean, range, standard deviation, etc)

suppose mean of w is  w
standard deviation of w is  w

Then use the statistics sample to construct the CC

The general model of CC is

Control chart design

encompasses selection of
• sample size,
• control limits,
• and sampling frequency

we specified a sample size of five measurements, three-sigma control

limits, and the sampling frequency to be every hour
 1. Selection of sample size
Industrial Engineering ITS

• Examine control chart design from an economic

point of view, considering explicitly
– the cost of sampling,
– losses from allowing defective product to be
produced,
– the costs of investigating out-of-control signals that
are really false alarms.
• increasing the sample size will decrease the
probability of type II error → enhancing the
chart’s ability to detect an out-of-control state
 2. Selection of control limit
Industrial Engineering ITS

• the control limits as a multiple of the standard deviation (L)

• When the control limits farther from the center line

– decrease the risk of a type I error
– increase the risk of a type II error
• When the control limits closer to the center line
– The risk of type I error is increased
– The risk of type II error is decreased

• Note :
– H0 : process in control or no assignable cause
– type I error : the risk of a point falling beyond the control
limits, indicating an out-of-control condition when no
assignable cause is present)
– type II error (the risk of a point falling between the control
limits when the process is really out of control)
 Choosing L
Industrial Engineering ITS

• The outer limit (L=3) = action limits,

when a point plots outside of
this limit, a search for an assignable
cause is made and corrective action
is taken
• The inner limits,(L=2) = warning limits, If a points fall
between the warning limits and the control limits, or very
close to the warning limit, we should be suspicious that the
process may not be operating properly
→ increase the sensitivity of the control chart; that is, it can
allow the control chart to signal a shift in the process more
quickly.
 Choosing probability limit
Industrial Engineering ITS

• the control limits as probability limit

• For 0,001 probability limit (or the total risk of making a type
I error is 0.002)
• 1- 0,001 = 0,999 → Z=3,09
UCL = 1.5 + 3.09(0.0671) = 1.7073
LCL = 1.5 − 3.09(0.0671) = 1.2927
 Normal probability distribution
Industrial Engineering ITS

• 3 means
– an incorrect out-of-control signal or false alarm will
be generated in only 27 out of 10,000 points.
– the probability that a point taken when the process
is in control will exceed the three-sigma limits in
two direction is 0.00270
 3. Selection of sample frequency
Industrial Engineering ITS

• Specify both the sample size and the frequency of sampling.

• →increase the sampling frequency and/or the sample size so
that information about the process can be obtained quickly →
adaptive or variable sampling interval (or variable sample size,
etc.) schemes.
• The most desirable : take large samples very frequently
• keep in mind : the size of the shift want to detect.
• For large process shift → use smaller sample sizes
• Allocating sampling effort : take small samples at short intervals
or larger samples at longer intervals.
• For high-volume manufacturing processes / many types of
assignable causes can occur → tends to favor smaller, more
frequent samples
• As automatic sensing and measurement technology → possible
to greatly increase sampling frequencies → single unit tested for
on-line process control
 Using ARL and ATS
Industrial Engineering ITS

• The average run length is the average number of points that

must be plotted before a point indicates an out-of-control
condition
ARL = 1 / p
ARL = 1 / probability that any point exceeds the control limits
• Consider the Shewhart CC with 3 limits.
• When the process is in control, we have noted that p = 0.0027
and the in-control is ARL= 1/p = 1/0.0027 = 370 → even if the
process remains in control, an out-of-control signal will be
generated every 370 samples, on the average.
• The average time to signal is time needed for false alarm
ATS = ARL.h
• For the hard-bake process : sampling every hour → we will have
a false alarm about every 370 hours on the average.
 Using Rationale Subgroup concepts
Industrial Engineering ITS

• The rational subgroup concept means that subgroups or

samples should be selected so that if assignable causes
are present, chance for differences between subgroups
will be maximized, while chance for difference due to
assignable causes within a subgroup will be minimized.
• Two general approaches for constructing rational
subgroups:
1. Sample consists of units produced at the same time −
consecutive units
– Primary purpose is to detect process shifts
2. Sample consists of units that are representative of all units
produced since last sample − random sample of all process
output over sampling interval
– Often used to make decisions about acceptance of product
– Effective at detecting shifts to out-of-control state and back
into in-control state between samples
– Care must be taken because we can often make any process
appear to be in statistical control just by stretching out the
interval between observations in the sample.
34
ANALYSIS OF CONTROL CHART

Industrial Engineering ITS

 Pattern on Control Chart
Industrial Engineering ITS

• Analysis a run (arrangement of points or sequence of

observations of the same type) → important
measure of nonrandom behavior on a control chart,
• Analysis the pattern recognition → recognizing
systematic or nonrandom patterns on the control
chart and identifying the reason for this behavior.
• requires experience and knowledge about the
process

• out-of-control condition when :

– one or more points fall beyond the control limits
– the plotted points exhibit some nonrandom pattern of
behavior.

36
 Detecting nonrandom pattern
Industrial Engineering ITS

By The Western Electric Statistical Quality Control

Handbook (1956) → consider as zone rules for control
charts
The process is out of control if either
1. one point plots outside the three-sigma control limits,
2. two out of three consecutive points plot beyond the
two-sigma warning limits,
3. four out of five consecutive
points plot at a distance
of one-sigma or beyond
from the center line, or
4. eight consecutive points
plot on one side of
the center line.
 Detecting nonrandom pattern
Industrial Engineering ITS
CONTROL CHART APPLICATIONS

Industrial Engineering ITS

 phase I
Industrial Engineering ITS

1. set of process data is gathered (m = 20 or 25 subgroups)

and analyzed all at once in a retrospective analysis,
2. constructing trial control limits are calculated based on
the m subgroups,
3. the data plotted on the control charts
4. Points that are outside the control limits are investigated,
looking for potential assignable causes need to eliminate.
That points are then excluded and a new set of revised
control limits are calculated.
5. Then new data are collected and compared to these
revised limits
6. Sometimes this type of analysis will require several cycles
7. Eventually the process is stabilized, and a clean set of data
that represents in-control process performance is
obtained for use in phase II.
 phase II
Industrial Engineering ITS

• Phase II begins after we have a “clean” set of process

data gathered under stable conditions and
representative of in-control process performance.
• In phase II, we usually assume that the process is
reasonably stable. Often, the assignable causes that
occur in phase II result in smaller process shifts,
because (it is hoped) most of the really ugly sources
of variability have been systematically removed
during phase I.
• Use the control chart to monitor the process by
comparing the sample statistic for each successive
sample as it is drawn from the process to the control
limits.
END OF SLIDE

Industrial Engineering ITS