BME-102: OPERATIONS MANAGEMENT WITH TQM

LESSON 3:
2ND SEMESTER | A.Y. 202Y – 202Y
LECTURER: MS./SIR

PROCESS CAPABILITY

 “Process is In Control or Out of Control” is different

from “Process Capable or Not”
 In process capability, we compare the variation of a
process (x̄ chart, r chart, p chart, and c chart) to
the variations allowed by the design specifications
(or tolerances)
o The design specifications are what the
designers have in mind, what customers have
in mind when product designed.

Given:

𝑥̿ = 16 ounces  We don’t want a process to be wide (picture below).

 We don’t like that our mean is 16 but it is no longer
σ = 0.5 ounces
15.5 and 16.5 but already 13 and 18.
 (Purple circle) is the sigma.

 If, for example, we made a sample (x) and it

happens to be 17.5.
— Black = process limits (UCL and LCL)  Is the process in control? Yes. Because UCL is 18
— Orange = tolerances/specification limits (USL and and LCL is 13.
LSL)  But is the process capable? Is it still doing what we
— If we get a sample of 17.3 ounces. The process is though it should be? No.
still in control but not already in accordance with the
specifications.
— The distance from mean to upper control limit (UCL
is 3-σ because that is what most companies work
with.

Example:

Maker of Fruit Loops

 The mean of the original specification is 16.

 (Purple box) is our original specification, but the
Example:
curve is how the process is running now.
 Note that we want the process to be narrow (picture  We have a customer, International Tennis
below). Federation (ITF). We are working for a supplier of a
sports equipment for upcoming Grand Slam
tournaments, so they approach us with a
requirement which is an order of tennis balls.
  As a supplier of the tennis balls, we need to
understand the requirements of your customer and
one of the critical requirements is the diameter of
the ball.
 So they gave us specifications that the tennis ball
should range anywhere between 6.56 to 6.86 cm in
terms of diameter.
o Lower Specification Limit = 6.56 cm
o Upper Specification Limit = 6.86 cm
B

 As long as we stay within the specification limits,

our product is going to get accepted. — When specification limits and control limits coincide.
 But if we go beyond the specification limits on either We know that 99.73% follows specifications—that
side whether it's beyond us USL or below LSL is is why it is much better than A.
going to get rejected. — This is much better than A because only 27 boxes
 If we are operating a process represented by normal are outside the specification unlike A that has more
curve which extends beyond the specification limits, than 27 boxes outside the tolerances.
we are incurring some loss. — We are 99.73% sure.
 As there is an acceptance region, that's defined only — We know the 99.73% fall within ± 3-σ.
between the lower and upper specification limits. — If we have 10,000 products, only 27 of those 10,000
 Whatever goes beyond is going to get rejected. that are outside specification  (1 – 0.9973)
— Green shade are those that don’t follow
specification.

— Process can be in control but there are still

unacceptable items.
o It is “can be in control” instead of “is in control”
— Process is capable of always conforming to
because there are unacceptable items that are specifications.
outside the tolerances (green shade). It is in
— This is much better than A and B, as this is the
control because it is still within the process
best—much narrower.
range. — This is not only in control but also capable.
o More than 27 boxes that do not follow — Inside the normal curve, 99.73% of the 10,000
specifications. products are there. While the 0.0027% are may be
o But those items outside the specification are outside the control limits but still within the
still sellable. specification limits.

2|Page
 — Those outside of 99.73% are out of control but still  When we are taking into account 3-σ above the
capable because it is within the specification limits. mean and 3-σ below the mean—adding the two
— If ever there are items that will fall out of the becomes 6-σ.
specification limits, but the probability is very small.  Value of sigma and How many sigma is different.
o x̄ ± 3σ

𝐂𝑷 = 1
 Capability Ratio of 1
D
o Process is capable.
o Tolerance range and Process range are the
same.
o Process range is equal to specification range.
o In this case is equal to 6-σ which means we
can go 3-σ above the mean and 3-σ below the
mean.
o We are assuming a centered process.

— The best among the four.

— Narrower, the better.
— Production can be out of control but still within the
specification limits.
— This normal curve is tighter, which is better.

Interpreting CP Values

Process Capability Ratio 𝐂𝑷 < 1

 It is a ratio of the specification width which is
 Capability Ratio of Less Than 1
provided by the customer divided by the process
o Process is not capable.
width which is our capability to deliver.
o Process range is wider than tolerance range.
o Tolerance Range  specification
o In this case, the |USL – LSL| or the
o Process Range  6-σ
specification width must be less than the
o 6-σ = UCL – LCL
denominator which is 6-σ.
 What the process capability ratio takes into account
o This means that we have defects in the
is only the width of the tolerance and process.
process—we are going beyond the
Whether what of them are wider or not.
specification limits or we are incurring waste.
 It does not take into account the movement of the
process mean because it does not consider the 𝑥̿ in
the computation.
 It is not enough to know what is wider (either the
process range or specification range) because we
also should know how the process mean shifts to
the right or left.
 A process can be capable but out of control.

Tolerance Range USL − LSL

𝐂𝑷 = =
Process Range 6σ
𝛔 = 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛
𝐂𝑷 > 1
 6-σ because in case of normal distribution, 99.73%
of data points fall within +3-σ or –3-σ.  Capability Ratio of Greater Than 1
 We are achieving nearly 99.73% of the target. o Process is capable.
o Tolerance range is wider.
3|Page
 o This means that we are operating well within — Process range is wider.
the specification limit. — Capability ratio is less than 1
o We are pretty much under control and we are
not even touching the boundaries as we were  Process is not capable.
doing in case of CP = 1.

2nd EXAMPLE

ln a GE insurance claims process, 𝑥̿ = 210.0 minutes, and

σ = 0.516 minutes.

The design specification to meet customer expectations

is 210 ± 3 minutes. So the Upper Specification is 213
minutes and the lower specification is 207 minutes. The
OM manager wants to compute the process capability
ratio.
1ST EXAMPLE
Given:

𝑥̿ = 210.0 minutes

σ = 0.516 minutes

USL = 210 + 3 = 213 minutes

LSL = 210 – 3 = 207 minutes

USL − LSL
— Tolerance range is wider. 𝐂𝑷 =
6σ
Given:
213 − 207
17.5 − 16 𝐂𝑷 =
𝛔= 6 (0.516)
3
𝐂𝑷 = 1.9
σ = 0.5 ounces
 Process is capable
USL − LSL
𝐂𝑷 =
6σ
19 − 13 Process Capability Index
𝐂𝑷 =
6 (0.5) Process Mean Shifts

𝐂𝑷 = 2  Let’s assume that there is a shift in the process for

whatever reason over a period of time.
 Process is capable  Our process which used to be within the
specification limits is now either Below the lower
specification limits or is Above the upper
specification limit.

4|Page
 σ = 0.0005 inch

USL = 0.25 + 0.0005 = 0.251 inch

LSL = 0.25 – 0.0005 = 0.249 inch

𝑈𝑆𝐿 − 𝑥̿ 𝑥̿ − 𝐿𝑆𝐿
𝐂𝑷𝑲 = 𝑀𝑖𝑛𝑖𝑚𝑢𝑚 𝑜𝑓 [ , ]
3𝜎 3𝜎
0.251 − 0.25 0.25 − 0.249
𝐂𝑷𝑲 = 𝑀𝑖𝑛. 𝑜𝑓 [ , ]
3 (0.0005) 3 (0.0005)
 In the example above, the customer is willing to buy
our product only as long as we stay within the
𝐂𝑷𝑲 = (0.67, 0.67)
specification limits.
 So the shift that we observe here is actually a loss
𝐂𝑷𝑲 = 0.67
because all the products which are beyond the  Process is not capable.
specification limits whether it be below the lower
specification or above the upper specification are
going to get rejected.
 Capability Ratio cannot erase this kind of process Interpreting CPK Values
shift (below or above) because CP only takes into
account 2 widths and it only computes a ratio. C𝑃𝐾 < 0
 Even if your process is completely outside their
specification limits, it will still be able to give you a  CPK = Negative Number
numeric response but the problem is in that case o Process mean has gone beyond one of the
will not be doing any good to the process. specification limits.
 The solution comes in the form of Process Capability o Process does not meet specifications.
Index (CPK). o Either it has gone below the lower
specification limits or above this upper
specification limits.
Process Capability Index o The process will not produce within the
specified toleranc
USL − 𝑥̿ 𝑥̿ − LSL
𝐂𝑷𝑲 = Minimum of [ , ]
3σ 3σ
𝐱̿ = 𝑚𝑒𝑎𝑛 𝑜𝑓 𝑠𝑎𝑚𝑝𝑙𝑒 𝑚𝑒𝑎𝑛𝑠
𝛔 = 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛

 The answer in CPK is the minimum—it is either of the

2 figures.
 We take the minimum because we want to make
sure that we give importance to the specification,
which is critical.
 What the capability index covers that the capability
ratio does not is the movement of the process mean 1ST EXAMPLE
(to the right r to the left).

1ST EXAMPLE

You are the process manager and developed a machine

to cut insoles.

The insoles cannot be more than ± 0.001 of an inch from

the required thickness of 0.25 inches. We want to know
if we have to replace the machine.

Given:

𝑥̿ = 0.25 inch

5|Page
 — 99.73% of your product is in the normal curve, C𝑃𝐾 = 0
which are mostly outside the specification limits.
— The process is not capable because the mean was  CPK = Zero
moved from 16 to 19. o Process mean is overlapping with one of the
— The normal curve all outside the specification limits. specification limits.
o Process does not meet specifications.

𝑈𝑆𝐿 − 𝑥̿ 𝑥̿ − 𝐿𝑆𝐿
𝐂𝑷𝑲 = 𝑀𝑖𝑛𝑖𝑚𝑢𝑚 𝑜𝑓 [ , ]
3𝜎 3𝜎
17 − 19 19 − 15
𝐂𝑷𝑲 = 𝑀𝑖𝑛. 𝑜𝑓 [ , ]
3𝜎 3𝜎
𝐂𝑷𝑲 = (− , +)
𝐂𝑷𝑲 = −
 Process is not capable.

2nd EXAMPLE
1ST EXAMPLE

— 14 is the mean of the designer (tolerance mean).

— This mean that when we start the production, we
also want the mean of the process to be 14. — 50% of the product are not capable because it does
— However, in this case, when we started the not meet the specifications.
production, the mean shifted to the right (17). — USL = x̄ (process mean)

17 − 16.5
17 − 14.5 𝛔=
𝛔= 3
3
𝛔 = 0.17
𝛔 = 0.83 𝑈𝑆𝐿 − 𝑥̿ 𝑥̿ − 𝐿𝑆𝐿
𝐂𝑷𝑲 = 𝑀𝑖𝑛𝑖𝑚𝑢𝑚 𝑜𝑓 [ , ]
𝑈𝑆𝐿− 𝑥̿ 𝑥̿ −𝐿𝑆𝐿 3𝜎 3𝜎
𝐂𝑷𝑲 = 𝑀𝑖𝑛𝑖𝑚𝑢𝑚 𝑜𝑓 [ , ]
3𝜎 3𝜎
17 − 17 17 − 15
15 − 17 17 − 13 𝐂𝑷𝑲 = 𝑀𝑖𝑛. 𝑜𝑓 [ , ]
𝐂𝑷𝑲 = 𝑀𝑖𝑛. 𝑜𝑓 [ , ] 3 (0.17) 3 (0.17)
3 (0.83) 3 (0.83)
𝐂𝑷𝑲 = (0, 3.92)
𝐂𝑷𝑲 = (−0.8 , 1.6)
𝐂𝑷𝑲 = 0
𝐂𝑷𝑲 = −0.8
 Process is not capable.
 Process is not capable.
o The process mean is the same with either the USL
or LSL.

6|Page
 o If the process mean is on top of the USL, 50% of 𝐂𝑷𝑲 = (1.5, 0.17)
the product is capable.
𝐂𝑷𝑲 = 0.17
 Process is not capable.
0 < C𝑃𝐾 < 1
o The process mean has shifted to the left.
 CPK = between 0 and 1
o Process mean is within the specification limits
but some part of the process is outside the
specification limits.
o Process does not meet specifications.
o Mean of the process lies between the
specification limits but some part of the
process has started going beyond the
specification limit either below the lower
specification or above the upper specification.
o We are not running highly a capable process.
o It is deteriorating but the mean is still within
the specification limits.

C𝑃𝐾 = 1
 CPK = 1
o Process is operating at borderline.
o Process meet specifications.
o The width of the specification is equal to the
width of the process.
o CP = 1, which means CPK = 1 because the
process is perfectly centered but the
1ST EXAMPLE
processes operating at the borderline.
Food served at a restaurant should be between 39°C and o The process is operating at a 3-σ level.
49°C when it is delivered to the customer. The process
used to keep the food at the correct temperature has a
process standard deviation of 2°C and the mean value for
these temperature is 40. What is the capability of the
process?

Given:

𝑥̿ = 40°C

σ = 2°C

USL = 49°C

LSL = 39°C
1st EXAMPLE
UCL = 40 + 3 (2) = 46 (x̄ ± 3σ)

LCL = 40 – 3 (2) = 34 (x̄ ± 3σ)

𝑈𝑆𝐿 − 𝑥̿ 𝑥̿ − 𝐿𝑆𝐿
𝐂𝑷𝑲 = 𝑀𝑖𝑛𝑖𝑚𝑢𝑚 𝑜𝑓 [ , ]
3𝜎 3𝜎
49 − 40 40 − 39
𝐂𝑷𝑲 = 𝑀𝑖𝑛. 𝑜𝑓 [ , ]
3 (2) 3 (2)
7|Page
 (1) Process is perfectly centered and it is well within their
specification limits.

(2) Process is not centered but it is still within the

specification limits. The mean has shifted a little bit
— The process mean and designer mean is the same. but the process is still within the specification limits.

16 − 13
𝛔=
3
𝛔=1
𝑈𝑆𝐿 − 𝑥̿ 𝑥̿ − 𝐿𝑆𝐿 (3) Process is not centered but it is still within the
𝐂𝑷𝑲 = 𝑀𝑖𝑛𝑖𝑚𝑢𝑚 𝑜𝑓 [ , ]
3𝜎 3𝜎 specification limits. The mean has shifted on the
other side.
19 − 16 16 − 13
𝐂𝑷𝑲 = 𝑀𝑖𝑛. 𝑜𝑓 [ , ]
3 (1) 3 (1)
𝐂𝑷𝑲 = (1, 1)
𝐂𝑷𝑲 = 1
 Process is capable.

o 99.73% of the production follows the specification

1st EXAMPLE
limits.
Sam's Long Life 75-watt light bulbs are designed to have
a life of 1,125 hours with tolerances of ± 180 hours. The
process that makes light bulbs has a mean of 1050 hours
C𝑃𝐾 > 1 with a standard deviation of 40 hours. Compute the
process capability ratio and the process capability index,
 CPK greater than 1
and comment on the overall capability of the process.
o Process is well within in the specification
limits. Given:
o Process is better than specification requires.
o The process becomes increasingly target 𝑥̿ = 1,050
oriented with fewer defects.
σ = 40

USL = 1,125 + 180 = 1,305 hours

LSL = 1,125 – 180 = 945 hours

UCL = 1,050 + 3 (40) = 1,170 (x̄ ± 3σ)

LCL = 1,050 – 930 = 34 (x̄ ± 3σ)

USL − LSL
There could be multiple possibilities: 𝐂𝑷 =
6σ

8|Page
 1,305 − 945 tolerances of ± 0.033 inch). The CNC process that
𝐂𝑷 = manufactures these parts has a mean of 1.281 inches and
6 (40) a standard deviation of 0.01 inches. Compute the process
capability ratio and the process capability index, and
𝐂𝑷 = 1.5
comment on the overall capability of the process to meet
 Light bulb production (process) is capable. the design specifications.

Given:

𝑈𝑆𝐿 − 𝑥̿ 𝑥̿ − 𝐿𝑆𝐿 𝑥̿ = 1.281 inches

𝐂𝑷𝑲 = 𝑀𝑖𝑛𝑖𝑚𝑢𝑚 𝑜𝑓 [ , ]
3𝜎 3𝜎 σ = 0.01

1,305 − 1,050 1,050 − 945 USL = 1.275 + 0.033 = 1.308 inches

𝐂𝑷𝑲 = 𝑀𝑖𝑛. 𝑜𝑓 [ , ]
3 (40) 3 (40) LSL = 1.275 – 0.033 = 1.242 inches

𝐂𝑷𝑲 = (2.13, 0.88) UCL = 1.281 + 3 (0.01) = 1.31 (x̄ ± 3σ)

𝐂𝑷𝑲 = 0.88 LCL = 1.281 – 3 (0.01) = 1.25 (x̄ ± 3σ)

 Light bulb production (process) is not capable.

Conclusion: Production process is not capable. USL − LSL

𝐂𝑷 =
o When one of the measures show not capable, then 6σ
the conclusion is that the production is not 1.31 − 1.24
capable. 𝐂𝑷 =
o It will only be capable if both measures show 6 (0.01)
capable.
𝐂𝑷 = 1.2
 Aircraft part production (process) is capable.

𝑈𝑆𝐿 − 𝑥̿ 𝑥̿ − 𝐿𝑆𝐿
𝐂𝑷𝑲 = 𝑀𝑖𝑛𝑖𝑚𝑢𝑚 𝑜𝑓 [ , ]
3𝜎 3𝜎
1.308 − 1.281 1.281 − 1.242
𝐂𝑷𝑲 = 𝑀𝑖𝑛. 𝑜𝑓 [ , ]
3 (0.01) 3 (0.01)

𝐂𝑷𝑲 = (0.9, 1.3)

— Process is not capable.
— The process mean shifted the left of the design 𝐂𝑷𝑲 = 0.9
mean.
— Some part of the process is outside the specification  Aircraft part production (process) is not capable.
limits (below the LSL).
Conclusion: Production process is not capable.
— LCL is less than LSL.
— We should worry because we want the light bulbs to
light up for 945 to 1,305 hours with a mean of 1,125
hours. However, since the process mean shifted t
the left, the minimum hours the light bulb could last
is only 930 hours instead of 945 hours.

2nd EXAMPLE

Elon Corporation manufactures parts for an aircraft

company. It uses a computerized numerical controlled
(CNC) machining center to produce a specific part that
has a design (nominal target of 1.275 inches with

9|Page
— Process is not capable because the process has
shifted from the design mean.
— The process mean shifted the right of the design
mean.
— The process is slightly off-center from the desired
target and closer to the lower specification limit.
— The process is capable of meeting the design
specifications, but it could benefit from being more
centered around the target value.

10 | P a g e