Control charts and process

capability
 contents
• Introduction
• Control chart for variables
• X bar R chart
• Control charts for Attributes
• p chart
• C chart
• U chart
Cp, Cpk
• Process capability index
Seven basic quality control tools
 Histograms
 Run Charts
 Pareto Charts
 Flow Charts
 Scatter Diagrams
 Cause and Effect Diagrams
 Control Charts
Control Chart
• Control charts are statistical tool, showing whether a
process is in control or not. It is a graphical tool for
monitoring the activities of an ongoing process also
referred as Shewhart control charts.
• Define Upper limit, lower limit and Center line
• Draw Chart
• Plot the data points into chart
• Interpret the control chart
…control chart
 Causes of Variations
• Special causes: Variation due to identifiable factors in the production
process. Examples of special causes include: wrong tool, wrong
production method, improper raw material, operator's skill, wrong die
etc. Control of process is achieved through the elimination of special
causes. According to Deming, only 15% of the problems are due to the
special causes. Special causes or also sometimes referred as Assignable
causes.
• Common causes: Variation inherent in the process. Improvement of
process is accomplished through the reduction of common causes and
improving the system. According to Deming, 85% of the problems are due
to the common causes.
Steps in constructing a control chart
 Decide what to measure or count
     Collect the sample data
     Plot the samples on a control chart
     Calculate and plot the control limits on the control chart
     Determine if the data is in control
     If non-random variation is present, discard the data (fix the
problem) and recalculate the control limits
     When data are with in the control limits we leave the process
assuming there are only chance causes present
A process is in control IF
• No sample points outside control limits
• Most points near process average or center line
• About equal number of points above and below the center line
• Sample point are distributed randomly
Control chart including causes
Any point outside the control limits
• A Run of 7 Points all above or All
below the central line - Stop the
production
• Quarantine and 100% check
• Adjust Process.
• Check 5 Consecutive samples
• Continue The Process.
• A Run of 7 Point Up or Down -
Instruction as above
Types of control charts
• Variable: continuous data. Things we can measure. Example
includes length, weight, time, temperature, diameter, etc.
• Attribute: discrete data. Things we count. Examples include
number or percent defective items in a lot, number of
defects per item etc
Control charts

Control charts for variables Control charts for attributes

X bar and R charts p, c, u charts

Control charts for the variable type of data
(X bar and R charts)
  𝑅𝑖= 𝑋 max ⁡(𝑖) − 𝑋 min ⁡(𝑖)
 

= Mean of i th sample
𝑅𝑖 − 𝑅𝑎𝑛𝑔𝑒 𝑜𝑓 𝑖𝑡h 𝑠𝑎𝑚𝑝𝑙𝑒
 
n =Sample size
 
=I th data

𝑖= 𝑔
 
∑ 𝑅𝑖
´ = 𝑖=1
𝑅
𝑖= 𝑔 𝑔
 
∑ 𝑋
´𝑖 ´ − 𝑀𝑒𝑎𝑛 𝑜𝑓 𝑔 𝑠𝑎𝑚𝑝𝑙𝑒𝑠
𝑅
 
´ = 𝑖=1
𝑋 =CL x́
𝑔
 
𝑈𝐶𝐿 ´ ´ 𝑈𝐶𝐿 ´
´ = 𝐷 4 𝑅 𝑈𝐶𝐿 𝐹𝑂𝑅 𝑅 𝐶𝐻𝐴𝑅𝑇
´𝑥 = 𝑋 + 𝐴 2 𝑅 𝑈𝐶𝐿 𝐹𝑂𝑅 𝑋 𝐵𝐴𝑅 𝐶𝐻𝐴𝑅𝑇
   
𝑅
 𝐿𝐶𝐿 ´𝑥 = 𝑋´ − 𝐴2 𝑅 ´ 𝐿𝐶𝐿 𝐹𝑂𝑅 𝑋 𝐵𝐴𝑅 𝐶𝐻𝐴𝑅𝑇  𝐿𝐶𝐿 ´ =𝐷 𝑅 ´ 𝐿𝐶𝐿 𝐹𝑂𝑅 𝑅 𝐶𝐻𝐴𝑅𝑇
𝑅 3

Example:
Mean values and ranges of data from 20 samples (sample size = 4) are shown in the table below:
   Mean of 20 samples
𝑖= 𝑔
  𝑅𝑖
∑
´ = 𝑖=1
𝑅 = 4.15 𝐶𝐿 𝑓𝑜𝑟 𝑅 𝑐h𝑎𝑟𝑡  
𝑔
 
=11.6-0.729x4.15=8.57
=11.6+0.729x4.15=14.63
X bar chart

Sample data at S.N 2, 16, and 18 are

slightly above the UCL. Efforts must
be made to find the special causes
and revised limits are advised to
calculate after deleting these data.
All the data are within the LCL and UCL in R Chart. Hence variability of the process data is not an issue to worry.
Control charts for Attribute type data (p, c, u
charts)
• p-charts calculates the percent defective in sample. p-charts are used when observations
can be placed in two categories such as yes or no, good or bad, pass or fail etc.

• c-charts counts the number of defects in an item. c-charts are used only when the
number of occurrence per unit of measure can be counted such as number of scratches,
cracks etc.

• u-chart counts the number of defect per sample. The u chart is used when it is not
possible to have a sample size of a fixed size.

For attribute control charts, the estimate of the variability of the process is a function of
the process average.
 P chart formulae

  𝑆𝑢𝑚 𝑜𝑓 𝑑𝑒𝑓𝑒𝑐𝑡𝑖𝑣𝑒𝑠 𝑖𝑛 𝑎𝑙𝑙 𝑠𝑎𝑚𝑝𝑙𝑒𝑠

´𝑝 = =𝑐𝑒𝑛𝑡𝑟𝑒 𝑙𝑖𝑛𝑒 𝑜𝑓 𝑝 𝑐h𝑎𝑟𝑡
𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑖𝑡𝑒𝑚𝑠 𝑖𝑛 𝑎𝑙𝑙 𝑠𝑎𝑚𝑝𝑙𝑒𝑠

 
´𝑝 (1− ´𝑝 )
´𝑝 (1 − ´𝑝 )
√
 

𝑈𝐶𝐿= 𝑝´ +3
√ 𝑛
𝐿𝐶𝐿= ´𝑝 −3
𝑛

𝑤h𝑒𝑟𝑒 𝑛 𝑖𝑠𝑡h𝑒 𝑠𝑎𝑚𝑝𝑙𝑒 𝑠𝑖𝑧𝑒 . 𝑆𝑎𝑚𝑝𝑙𝑒 𝑠𝑖𝑧𝑒 𝑖𝑛 𝑝 𝑐h𝑎𝑟𝑡 𝑚𝑢𝑠𝑡 𝑏𝑒 ≥ 50

 

𝑆𝑜𝑚𝑒𝑡𝑖𝑚𝑒𝑠
  𝐿𝐶𝐿𝑖𝑛 𝑝 𝑐h𝑎𝑟𝑡 𝑏𝑒𝑐𝑜𝑚𝑒𝑠 𝑛𝑒𝑔𝑎𝑡𝑖𝑣𝑒 . 𝐼𝑛𝑠𝑢𝑐h 𝑐𝑎𝑠𝑒𝑠 𝐿𝐶𝐿 𝑠h𝑜𝑢𝑙𝑑 𝑏𝑒 𝑡𝑎𝑘𝑒𝑛 𝑎𝑠 𝑧𝑒𝑟𝑜 .
Example: p-chart Data for defective CDs from 20 samples (sample size = 100) are shown in the table
below:

  ´𝑝 (1− ´𝑝 ) 0.051 𝑥(1− 0.051)

𝑈𝐶𝐿= 𝑝´ +3
√ 𝑛
=0.051+3
√ 100
=0.066
  ´𝑝 ( 1− ´𝑝 ) 0.051 𝑥 ( 1−0.051 )
𝐿𝐶𝐿= ´𝑝 −3
√ 𝑛 √
=0.051− 3
100
=0.036
Sample data at S.N 16 , 18, and 20  are above the UCL. Efforts must be made to find the special causes and
revised limits are advised to calculate after deleting these data. There is important observation that is clearly
visible from the data points that there is an increasing trend in the average proportion defectives beyond sample
number15 also, data show cyclic pattern. Process appears to be out of control and also there is a strong evidence
that data are not from independent source.
 c chart formulae
  𝑆𝑢𝑚 𝑜𝑓 𝑑𝑒𝑓𝑒𝑐𝑡𝑠 𝑖𝑛 𝑎𝑙𝑙 𝑠𝑎𝑚𝑝𝑙𝑒𝑠
𝑐´ = =𝐶𝐿 𝑜𝑓 𝑐 𝑐h𝑎𝑟𝑡
𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑖𝑡𝑒𝑚𝑠 𝑖𝑛 𝑎𝑙𝑙 𝑠𝑎𝑚𝑝𝑙𝑒𝑠

𝑈𝐶𝐿=
  ´𝑐 +3 √ 𝑐´  𝐿𝐶𝐿=𝑐
´ −3 √ 𝑐´
Example
Data for defects on TV set from 20 samples (sample size = 10) are shown in the table below:  
  𝑆𝑢𝑚𝑜𝑓 𝑑𝑒𝑓𝑒𝑐𝑡𝑠 𝑖𝑛 𝑎𝑙𝑙 𝑠𝑎𝑚𝑝𝑙𝑒𝑠 109
𝑐´ = = =𝐶𝐿𝑜𝑓 𝑐 𝑐h𝑎𝑟𝑡 =5.45
𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑖𝑡𝑒𝑚𝑠 𝑖𝑛𝑎𝑙𝑙 𝑠𝑎𝑚𝑝𝑙𝑒𝑠 20
 

 
None of the sample is out of
the LCL and UCL. But the chart
shows cyclic trend.

HW: Study U chart

Process capability
A process output is considered stable when it consists of only common-
cause variation and has the reproducibility over a long period of time.
Common-cause variation originates from the basic elements of a
manufacturing process. Which are 5 Ms:
• Machine,
•      Man (operator),
•      Material,
•      Method of work, and
•      Measurement system
unstable

stable
The process parameters can not be correctly estimated for an unstable process because of the following
reasons.
•     No well defined output distribution
•     Misleading decisions
•     No useful estimation of process capability
•     No useful purpose for process improvement

Process Capability
Prerequisites for process capability is to estimate process average and process standard deviation.
yes

Bilateral Specification
A process producing a
characteristic with a bilateral
specification meets the
minimum requirement of
capability when it is stable,
and has no more than 0.135
percent of its output for this
characteristic outside either
specification limit.
 Unilateral Specification

A  process  producing a characteristic with a unilateral specification  meets the minimum requirement of  capability  when it
is stable, and  has no more  than 0.135 percent of its output for this characteristic outside the single specification limit.
Why Processes Fail?
Measuring Process Capability

 
𝑇𝑜𝑙𝑒𝑟𝑎𝑛𝑐𝑒 𝑟𝑎𝑛𝑔𝑒
𝐶 𝑝=
𝑝𝑟𝑜𝑐𝑒𝑠𝑠 𝑟𝑎𝑛𝑔𝑒

 
𝑈𝑆𝐿 − 𝐿𝑆𝐿
𝐶 𝑝=
6𝜎
 Example: Net weight specification
=9± 0.5
Process mean=8.80

Process standard deviation =0.12

 
=
Process capability index
 

𝑋´ −𝐿𝑆𝐿 𝑈𝑆𝐿− 𝑋´
𝐶 𝑝𝑘=𝑀𝑖𝑛𝑖𝑚𝑢𝑚( ,
3𝜎 3𝜎 )
 Example: Net weight specification
=9± 0.5
Process mean=8.80

Process standard deviation =0.12

 
𝑋´ − 𝐿𝑆𝐿 𝑈𝑆𝐿 − 𝑋´
𝐶 𝑝𝑘 =𝑀𝑖𝑛𝑖𝑚𝑢𝑚
3𝜎
, (3𝜎 )
 
=0.83
control charts and process
capability analysis