Statistical Process Control

Statistics : An estimation of a parameter as of the population mean or variance, obtained

from a sample deals with designing for collection and analysis of data
Process : A system of operation in the production of sometime
Control : Dominating influence over the regulate, operate
SPC : is the practice of using statistical methods such as control charts and capability
analysis to monitor and control a process

SPC allows user to take appropriate actions so that process remains in a state of
statistical control
It also allows user to improve the processes ability to produce output that meets
customer expectations
SPC is to monitor product quality and maintain processes to fixed targets
SPC refers to using statistical techniques for measuring and improving quality of
processes and includes to other techniques, such as sampling plans
Experimental design
Variation reduction
Process capability analysis
Process improvement plans

A methodology for monitoring quality of manufacturing quality of manufacturing and

service delivery processes to help identify and eliminate unwanted causes of variation
Used to determine whether a process is performing as expected
SPC is built on straight forward statistical parameters mean and standard deviation
Its use depends on statistical thinking not on statistical knowledge
Involving inspecting the output from a process
Quality characteristics are measured and charted
Helpful in identifying in-process variation
SPC 10 % is Statistics and 90 % Product and process knowledge
SPC is used to monitor the consistency of processes used to manufacture a product as
designed. Control chart is a primary tool used for SPC
It aims to get and keep processes under control
SPC can ensure that the product is being manufactured as designed and intended
SPC will not improve poorly designed products reliability

Page 1 of 8

G Aswath

Statistical Process Control

SPC benefits

Provides surveillance and feed back fro keeping processes in control

Signals when problem with process has occurred
Detects assignable causes of variation
Accomplishes process characterization
Reduces need for inspection
Monitors process quality
Provides mechanism to make process changes and track effects of those changes
Once a process is stable (assignable causes of variation have been eliminated),
provides process capability analysis with comparison to the product tolerance
Eliminates non conformity
Minimizes quality cost
Eliminates in house and end customer complaints
Total customer satisfaction

Monitoring of SPC Station

1.
2.
3.
4.
5.

Keep the work station clean

Keep the gauges and instruments in the respective places
Only current sample should be on the SPC station
Buffer should be maintained
Ensure the setting when the plotted point is near the X median line and within the range
chart
6. Trend run periodicity, bigger scatter and lower scatter should be observed only in the X
median chart, not in the range chart
7. Look at the range chart first, for the stability of the process, then to the X median, Rchart
8. Collect the last 5 samples for plotting
9. In case of trend run and periodicity, operator has to reset once in see any changes, if no
changes operator should inform the concerned for investigating and continue the
process
10. Resetting should be done in the beginning of the shift and when tool is changed, power
failure noticed
11. In case of abnormalities machine should be stopped strictly followed the instructions as
per action plan circle code re-plotting and entering at the batch of control chart

Documents on the SPC station

1.
2.
3.
4.
5.
6.
7.

Blank control chart

Running control chart
Previously filled up control chart
General instructions sheet
Corrective action plan
Standard codes
SPC characteristic sheet

Page 2 of 8

G Aswath

Statistical Process Control

Control limits are fixed based on the capability of the process

Samples produced consecutively are collected directly from the machine
Samples are measured and readings are entered in SPC chart
Sampling to be done at the beginning of the shift to continue the production
Sampling to be retaken immediately after re-adjust of tool/setting at commencement
of shift
Production to be continued only when X bar and S bar, X mean, P value are with in the
control limits
Operation of SPC chart ensures self assured production
Charts clearly tell the quality produced in all shifts
Charts indicate clearly the behavior of the process
Eliminates the necessity of line inspection

Why SPC

To monitor the consistency of processes used to manufacture a product as designed

It aims to get and keep processes under control
It can ensure that the product is being manufactured as designed and intended
To assure the quality of parts to customer
With out re-inspection for the dimensions

Variation
Less Variation = higher quality, i.e. the better the consistency, the lower the variability
Statistical methods offer the only way to understand and measure variability
1.
2.
3.
4.

Variation exists in all process

Variation is inevitable
Variation is predictable
Variation is measurable

No two pieces from the process are exactly alike (the difference between pieces to piece is
the variation)
Effective use of process control measurement data lies in understanding the concept
of variation
A stable process has the same normal distribution at all times and is in control
A stable process still has variation
Any process that is not stable is called an unstable or out of control processes
Kinds of instability

excursions
shifts
drifts
cycles
chaos

Page 3 of 8

G Aswath

Statistical Process Control

Random variation
Non Random Variation
Types of variation
1. Common or Chance or Random causes of variation
2. Assignable or Special Causes of variation
1. Common or Chance or Random causes of variation
We cannot identify and Unavoidable
Process has not changed
E.g. Slight difference in process variables like diameter, weight, serve time, temperature
Variation inherent in the process
Common causes distribution can be characterized by
Location (Centering)
Spread (Range or standard deviation)
Shape (the pattern of variation whether it is symmetrical or packed etc)
Common causes of variation: - is a natural cause of variation in the system
It is the result of complex interactions of variations in material, tools, machines,
information, workers and the environment
E.g.

Machine vibration
Temp. Fluctuations
Slight variation in raw material
Human variation in RM
Human variation in setting control dials

2. Assignable or Special Causes of variation

Can be identified and eliminated
E.g.
Poor employee training
Worn tool
Machine needing repair
Variability caused by an extraordinary event that impacts the performance measure of
process output
Special causes are recognized for the chances in the shape, spared or location of process
distribution
Special cause variation: Arises from external resources that are not inherent in the process, appear sporadically and
disrupt the random pattern of common causes and causing the process to change

Page 4 of 8

G Aswath

Statistical Process Control

Any thing that causes variation that is not part of the stable process is called special causes,
assignable causes or unnatural causes
E.g.
Batch of defective raw material
Faulty setup
Human error
Incorrect recipe
Blown gasket
Earth quake

Tools for control

i. Histogram
ii. Control chart
iii. Normal plot

---------- Bell shape

---------- no out of control
---------- Straight line

#
1
2
3
4
5
6
7
8

Common causes
Few in numbers
Variation is low
Part of the process
Constant variation
Predictable
Statistics applicable
Management controllable
Reduction lead to improvement
Eg. Pressure variation
Environment variation
9 Any one chance causes results in only a
minute amount of variation. How ever
many of the causes act simultaneously so
that the total amount of chance variation is
substantial
10 Cannot be economical
11 Process is stable when it is operating only
under chance causes

Special or assignable causes

Plenty in numbers
Variation is high
Visitor to the process
Fluctuating variation
Un-Predictable
Statistics cannot apply
Operator controllable
Eliminates lead to maintenance
Eg. Wrong setting
Wrong master
Any one special causes can result in a
large amount of variation
Economical justified

Improving stable process

1. centering the target
2. Reducing common causes variation
Reducing variation in an unstable processes
1.
2.
3.
4.

Do not ignore special causes

Do quickly detect special causes variation
Do stop production until the process is fixed (reaction)
Do identify and permanently eliminate special causes (prevention)

Page 5 of 8

G Aswath

Statistical Process Control

Four steps to improve an unstable process

1. Detect the special causes variation
2. Identify the special causes
3. Fix the process remove the special causes or
- compensate for the special causes
4. Prevent the special causes from occurring again
Type I error or alpha errors
Occur when a point falls outside
the control limits even through no
special causes is operating
The result is a witch-hunt for
special causes and adjustment of
things here and there
The tampering usually distorts a
stable process as well as wasting
time and energy
Chance of type-I error is 0.26%
Type II or beta error
Occur when you miss a special cause because the chart is not sensitive enough to detect it
In this case, you will go along unaware that the problem exists and thus unable to root it
out

Data

It is a set of values of input or outcome of a process

Based on data, we decide our action plans using statistical methods to solve problems
Data is a guide for actions
The origin of the data must be clearly recorded.
The data should clearly reflects fats
If data is not carefully and systematically recorded, It can not be analysis and put into
use.

1. Variable data (measured data)

Data which arises from measurement (usually occurs on a continuous scale)
Ex : Temperature, weight, volume, physical dimensions, etc.,
2. Attribute data or discrete data (non measured data)
The data that is classified as conforming or nonconforming to specification (Ok or
Not ok)
Ex:
Good
Bad
Defective
Non defective
Scratch
No scratch
Page 6 of 8

G Aswath

Statistical Process Control

With attributes data, there is a need to develop specific descriptions.

descriptions, which are called operational definitions.

These

For variables data, the standard deviation and average is an important measurement

Attribute data generates attributes charts include p, np, c and u charts. Variable data
generates Variables charts include Xbar-R charts, Xbar-s charts, individuals charts and
moving average and moving range charts.

There are two kinds of attributes data: yes/no type of data and counting data. p and
np control charts are used with yes/no type data; c and u charts are used with counting
type data.

For one item, there are only two possible outcomes: either it passes or it fails some
preset specification. Each item inspected is either defective (i.e., it does not meet the
specifications) or is not defective (i.e., it meets specifications). Examples of the yes/no
attributes data are:

mail delivery: is it on time or not on time?
phone answered: is it answered or not answered?


invoice correct: is it correct or not correct?

stock item: is it in stock or not in stock?

cycle count: is it correct or not correct?

product : in-spec or out of spec?

supplier: material received on-time or not on-time?

With counting data, you count the number of defects. A defect occurs when something
does not meet a preset specification. It does not mean that the item itself is defective.
For example, a television set can have a scratched cabinet (a defect) but still work
properly. When looking at counting data, you end up with whole numbers such as 0, 1,
2, 3; you can't have half of a defect.

To be considered counting data, the opportunity for defects to occur must be large; the
actual number that occurs must be small. For example, the opportunity for customer
complaints to occur is large. However, the number that actually occurs is small. Thus,
the number of customer complaints is an example of counting type data. Other
examples are:
o number of mistakes in picking
o number of items shipped incorrectly
o number of accidents for delivery trucks

When working with attributes data, you have to have a clear understanding of whether
the item you are looking at is defective or not (yes/no type data) or whether it should be
counted as a defect (counting type data). In order to know whether a shipment was on
time or to count the number of on-time shipments, you have to have a definition of what
"on time" means. Is "on time" anywhere from 1:55 p.m. to 2:05 p.m., anytime before 2:00
p.m., or anytime between 2:00 p.m. and 2:15 p.m.? This clear understanding of a quality
expectation is called an operational definition.

Page 7 of 8

G Aswath

Statistical Process Control

According to Dr. W. Edwards Deming, an operational definition includes:

A written statement (and/or a series of examples) of criteria or guidelines to be
applied to an object or to a group.
A test of the object or group for conformance with the guidelines that includes
specifics such as how to sample, how to test, and how to measure.
A decision: yes, the object or the group did meet the guidelines; no, the object or
group did not meet the guidelines; or the number of times the object or group did
not meet the guidelines.

Using an invoice error example, the written statement may read "An invoice error is an
incorrect shipping amount or a wrong price." The test could be to:
compare every invoice to the packing list to check for incorrect shipping amounts
and,
Compare every invoice to a price schedule to check for wrong prices.

Based on these guidelines and a test for conformance with these guidelines, you could
make a decision as to whether an invoice is defective or how many defects an invoice
contains.

Variables control charts are based on variables data. Variables data consist of
observations made from a continuum.
That is, the observation can be measured to any decimal place you want if your
measurement system allows it.
Some examples of variables data are contact time with a customer, sales dollars, amount
of time to make a delivery, height, weight, and costs.
In dealing with variables data, the average and standard deviation are very important
parameters. One must understand what is meant by these terms.
The average (also called the mean) is probably well understood by most. It represents a
"typical" value. For example, the average temperature for the day based on the past is
often given on weather reports. It represents a typical temperature for the time of year.
The average is calculated by adding up the results you have and dividing by the
number of results. For example, suppose the last five customer complaints took 5, 6, 2,
3, and 8 days to close. The average is determined by adding up these five numbers and
dividing by 5. The average is denoted by and in this case is;

5 + 6 + 2 + 3 + 8 24
=
= 4 .8
n
5
5
While the average is understood by most, few understand the standard deviation,
denoted by the letter s.
The standard deviation can be thought of as an average distance (the standard) that each
individual point is away from the mean.
The equation for the standard deviation is given below.
X=

s=

(X X )

n 1
We will be using control charts to estimate what our process average is and what the
process standard deviation is. For these two numbers to have any meaning, the process
must be in statistical control.
Page 8 of 8

G Aswath