Population Variance

Variance and Standard deviation tell us an average distance of any

observation in the date set from the mean of the distribution. To calculate the
population variance, we divide the sum of the squares, distances between the
mean and each item in the population by the total no. of items in the
population. By squaring each distance, we make each number positive and at
the same time, assign more weight to the larger deviation (Deviation is the
distance between the mean and a value).
2 = (X-)2
N

= X2 - 2
N

2 = Population variance
X = Item/observation
= Population mean
N = Total No. of items in the population
= Sum of all the values (X-)2 or all the values X2

PROCESS CAPABILITY
A predictable pattern of statistically stable behavior (most frequently a bell
shaped curve) where the chance causes of variation are compared to the
engineering specifications. A capable process is a process whose spread on
the bell shaped curve is narrower than the tolerance range.

LSL

MIN

USL

MAX

Figure 1.a Comparison Of Process Spread to Tolerance Range

Where USL and LSL are defined as upper specification and lower
specification limits respectively.
It is often necessary to compare the process variation with engineering or
specification tolerances to judge the suitability of the process. Process
capability addresses this issue.
Capability Applications:
Evaluation of new equipment
Reviewing tolerances based on the inherent variability of the process.
Assigning equipment to product (more capable equipment to tougher
jobs)
Routine process performance audits.
The effects of adjustments during processing

A process capability study includes three steps

Planning for data collection
Collecting data
Plotting and analyzing results
To determine process capability an estimation of sigma is necessary
R = r
d2

or

i = (X-)2
(n-1)

R is an estimate of process capability sigma. I is a measure of total data

sigma.
Common capability terms are capability ratio and performance ratio.
CR =

6 R
(USL-LSL)

PR = 6 I
(USL-LSL)

More widely used capability terms are capability index and performance
index.
C p = (USL-LSL)
6 R
As a rule of thumb:
C R < 0.75 Capable
C R = 0.75 to 1.00 Capable with tight control
C R > 1.00 Incapable
C p > 1.33 Capable
C p = 1.00 to 1.33 Capable with tight Control
C p < 1.00 Incapable

Pp = (USL-LSL)
6 I

Customers frequently require Cpk where Cpk is the ratio giving the smallest
answer between:
Cpk = USL-
3 R

or

-LSL
3 R

Customers may also request Ppk where Ppk is the ratio giving the smallest
answer between:
Cpk = USL-
or
-LSL
3 i
3 i
The objective of the process quality control is to establish a state of control
over the manufacturing process and then maintaining that state of control
through time. Actions that change or adjust the process are frequently the
result of some form of capability study. When the natural process limits are
compared with the specification range, any of the following possible courses
of action may result:
Do nothing: If the process limits fall well within the specification limits,
no action may be required.
Change the specifications: The specification limits may be unrealistic. In
some cases, specifications may be set tighter than necessary. Discuss the
situation with the final customer to see if the specifications may be
relaxed or modified.
Center the process: When the process spread is approximately the same
as the specification spread, an adjustment to the centering of the
processes may bring the bulk of the product within specifications.
Reduce Variability: This is often the most difficult option to achieve. It
may be possible to partition the variation (stream to stream, within piece,
batch to batch, etc) and work for the largest offender first. For a
complicated process, an experimental design may be used to identify the
leading source of variation.
Accept the Losses: In some cases, management must be content with a
high loss rate (at least temporarily). Some centering and reduction in
variation may be possible but the principal emphasis is on handling the
scrap and rework efficiently.

Table 1 Failure Rates for Cp values

Cp
0.33
0.67
1.00
1.10
1.20
1.30
1.40
1.50
1.60
1.67

PPM
317,500
45,500
2,700
967
318
96
27
6.8
1.6
0.6

PPM Parts Per Million of Nonconformance

For a two tailed Specification
Normally distributed
Centered on

Sample standard Deviation

S = 2 = (X-)2
N-1
= X2 - N2
N-1
N-1
Point Estimate:
A point estimate is a single number that is used to estimate the unknown
population parameter.
An internal estimate is a range of values used to estimate a population
parameter.
2 = (X-)2
N-1
Using a division of N-1 gives us an unbiased estimation of 2 .
Standard Error of the mean:
=
N
= Standard deviation of the population
10

68.3 %

20

99.5 %

30

99.7 %