X=LSL = 24.0 X=USL = 26.

5
ZL = -2 ZU = +3

Yield s = 0.5

Defects Defects

-2s Target +3s

Mean
X = 25

Process Capability
Short Term Long Term
Performance Performance For Continuous and Discrete Data
Lower Spec Limit Upper Spec Limit
Considers LSL USL

Centring
Cpk Ppk Rejects Rejects
Acceptable
Quality Products
Does Not Lower Control Limit Target Upper Control Limit
Consider LCL UCL

Centring Cp Pp -3s
VOC = USP - LSL
+3s

VOP = UCL - LCL = 6s

Variation
• Don’t worry the rope is over half an inch thick on
average.
average

• It is not the average I’m worried about!

 Process Capability - Definition
• Process Capability is a measurement of how the process
is performing with respect to a desired outcome
Allowed Variation determined by the customer Voice of Customer
A
Actual
lVVariation
i i d
determined by the process
db h V i off P
Voice Process

Lower Spec Limit Upper Spec Limit Lower Spec Limit Upper Spec Limit
LSL USL LSL USL

Rejects Rejects Rejects Rejects

Acceptable Acceptable
Quality Products Quality Products
Lower Control Limit Target Upper Control Limit
LCL Target UCL
LCL UCL
-3s +3s -3s +3s
VOC = USP - LSL VOC = USP - LSL

VOP = UCL - LCL = 6s VOP = UCL - LCL = 6s

GOOD CAPABILITY POOR CAPABILITY

Short Term (Cp) vs Long Term (Pp) Capability

• Over an extended period we expect to see more variation

Changes in process

Time

This process is short term Capable but not long term Capable

Cp 9 LSL
Pp 8
uct Charactteristic

LCL
3s
-3s
Vallue of Produ

UCL
+3s USL
Short term Sustained long
1 cycle term
performance performance

Time
 Process Capability Ratios (Cp, Pp) & Index (Cpk, Ppk)

A centred process An off-centred process

USL - LSL X - LSL
Cp = TL =
6sshort term 3s
USL - LSL USL - X
Pp = TU =
6 long term
6s 3
3s
Min (TL, TU)short term
Cpk =
3sshort term
Mi (TL, TU)long term
Min
Ppk =
3slong term
X X
Lower Spec Limit Upper Spec Limit Lower Spec Limit Upper Spec Limit
LSL USL LSL USL

Rejects Rejects Rejects Rejects

Acceptable Acceptable
Quality Products Quality Products
LCL Target UCL LCL UCL
Target
-3s +3s -3s +3s
VOC = USP - LSL TL TU
VOP = UCL - LCL = 6s
Note: Standard Deviation long term is usually referred to as Sigma or σ
Summary of Capability Metrics

Short Term Long Term

Performance Performance

Considers
Centring
Cpk Ppk
Does Not
Consider
Centring Cp Pp
 Test Your Understanding

Sketch the histogram and specification

limits for the data sets below A

2.0
B

1.5 F

Cpk C
1.0
D E D
0.5 B C
E
A
0.0
0.0 0.5 1.0 1.5 2.0
F
Cp
 Calculating Sigma Score (Z)
Point of interest • For normally distributed data the
X
Z = +3 standard deviation ((s)) and mean ((Xbar))
can be used to estimate the probability
of defects
Yield
• Point of interest X can be an USL or LSL
Defects
• Example with X = 25, s =0.5,
X=USL=26.5 then Z = (26.5-25)/0.5 = 3.
• Z scores can be converted to
+1s +2s +3s probabilities in Minitab or Excel
Target fx NORMSDIST(Z). Note Excel returns the
Mean
X
“left
left tail
tail” or yield of the distribution. If you
enter Z=3 you get 0.99865. The defects are
(X – X) 1-0.99865 = 0.00135 or 0.135% (defects = 1-
Z= yield).
s
(Z = the number of standard deviations
from the mean. Z = 3Cp if using USL
and LSL)
 Calculating the Total Defect Rate (Zbench)
X=LSL = 24.0 X=USL = 26.5 • The total defect rate expected from a
ZL = -2 ZU = +3 process can be estimated by
p y converting
g
both upper and lower Z scores to
probabilities and adding them :-
Yield s = 0.5
NORMSDIST(3)=0.99865
( ) Æ P=1-0.99865 =
0.00135. NORMSDIST(-2)=0.02275 (left tail).
Defects Defects
Total probability PT = 0.00135+0.02275 =
0.0241 or 2.41%.
• The total sigma score called Zbench can
-2s Target +3s be found from the total yield of the
Mean
X = 25 process YT =1 - PT. This represents the
overall process capability:-
(X – X) If total probability PT = 0.0241 then YT =
0.9759. Using Excel fx NORMSINV(0.9759) =
Z=
s 1 9756
1.9756.
• We say the SIGMA of this process is ~2!

(Z = th
the number
b off standard
t d d deviations
d i ti
from the mean. Z = 3Cp if using USL
and LSL)
Cp/Pp Relationships

N.B. Yields and PPM’s are calculated using defects at both tails of the
Distribution assuming a Centred Process
Capability for Discrete/Attribute Data
• Discrete data has clear boundaries between adjoining
values includes names
names, categories
categories, counts and rank
orders e.g. dates, colours, defects, defectives
• The data will either be Defectives (e(e.g.
g pass/fail or on
time or late) = ‘Binomial’ (0 or 1) or the data will be
Defects ((e.g.
g scratches or number of omissions)) =
‘Poisson’.
• Capability can be calculated directly from Binomial or
Poisson distributions using Minitab or other statistical
software packages, however you can convert the data to
continuous and then use the previous method
method. e e.g.
g
convert defects to probabilities (%ages / 100).
Capability for Non-Normal Data

• Normal capability analysis with

non-normall ddata
t wouldld give
i
misleading results
• Non-normal capability
p y can be
found from software packages
like Minitab or by transforming
the data to a distribution which
is Normal.
• Try taking logs (either Log10 or
L ) or b
Ln) by raising
i i ththe d
data
t tto a
power (called a Box-Cox
transformation) e.g. data2.
Process Capability Summary
• All processes have variation
• Process
P C
Capability
bilit iis a measurementt off h
how th
the process
is performing with respect to a desired outcome
• Capability is defined as the voice of the customer over
the voice of the process
• Long term capability is not the same as short term
capability
• Covert discrete to continuous and non-normal
non normal data to
normal before analysing capability or use specialist
software
• The overall capability of a process can be defined by it’s
Sigma value.