Statistical Process

Control & Process

Capability
 What is SPC supposed to tell us?

• When you are doing something you

should not
SPC & Cpk

• When you are not doing something

you should
• When you are doing things right

SPC & Cpk

 How Does SPC tell us?

• Think of SPC as charting, similar to

the monitoring (temperature, heart
SPC & Cpk

rate, blood pressure) that a doctor

and nurse chart on a patient.

• SPC monitors the condition of your

job activity for the same reason.
When something is not looking
ordinary, action is taken
SPC & Cpk
 What good is it to an Operator?

There is a clear identification of when

actions are to be taken by empowered
SPC & Cpk

operators (i.e., requiring them to

identify causes and solve problems).

SPC & Cpk

 Why is the Customer interested?

• Tighter control of ‘critical to quality’

attributes (features) of your product
SPC & Cpk

will improve their own control.

• Rejects, late deliveries and product

returns can disappear.

SPC & Cpk

 Why Use Control Charts?

“Failure to use control charts to

analyze data is one of the best
SPC & Cpk

ways known to mankind to increase

costs, waste effort, and lower
morale.”
Dr. Donald J. Wheeler

SPC & Cpk

 Causes
Common Cause (i.e., Noise)
• Is present in every process
• Is produced by the process itself (the way we do business)
SPC & Cpk

• Can be removed and/or lessened but requires a fundamental

change in the process
• Requires Management intervention
Special Cause (i.e., Signals)
• Exists in many operations/processes
• Caused by unique disturbances or a series of them
• Can be removed/lessened by using basic process control to
identify opportunities for improvement in our existing process
• Requires Operator intervention

SPC & Cpk

 Control Chart Basics
• Control Chart Components
• UCL - Upper Control Limit
• LCL - Lower Control Limit
SPC & Cpk

• CL - Center Line (average of dataset)

• Control charts are designed to be used by

operators
• Clear and easy to see abnormal conditions
• Uses the premise that quality is everyone’s
responsibility
• Quality at the Source

• Processes are in control when variation lies

within UCL and LCL
SPC & Cpk
 Control Charts
 Purpose
 Separate common cause from special cause
variation
SPC & Cpk

 Communicates process performance over time

 Limits are statistically calculated based 3σ

 Determined by the process
 Independent of the design or customer
specifications

 A process is in control when

 It operates with common (natural) variation
 It is not influenced by special cause variation
SPC & Cpk
 The Basic Control Chart
Individual / Moving Range
• Use:
• Basic analysis tool in testing for stability
• When its inconvenient or impossible to obtain more than
SPC & Cpk

one measurement per sample

• Or when technology allows for easy measurement of
every unit a minimal cost
• Data availability is sparse
• Variation:
• Short Term: Represented by the variation from one unit
to the next (MR Chart)
• Long Term: Represented by a sequence of such events
(Individuals Chart)

Charts are based on a Subgroup Size of 1

SPC & Cpk

 Long Term Variation
SPC & Cpk

Short Term Variation

SPC & Cpk
 Control Chart Formulas
Upper Limit Lower Limit

Individuals UCLX = X + 3 MR LCLX = X - 3 MR

Chart d2 d2
SPC & Cpk

Formulas

X-bar Chart UCLX = X +A2 R LCLX = X - A2 R

Formulas

MR-Chart UCLMR = D4 MR LCLMR = D3 MR

Formulas

R-Chart UCLR = D4 R LCLR = D3 R

SPC & Cpk

 Interpreting Control Charts

We use the phase “Out of Control” when

a control chart rule has been broken.
SPC & Cpk

These rules are based on the probability

that a chart pattern would occur, if
nothing has changed in the process.

This means something unusual has

happened –
Question it – Go Check It Out !

SPC & Cpk

 8 Rules of Control Charts
1. One (1) point > 3 Standard Deviations from the Center Line
(Outside the UCL or LCL {3 Sigma Limit})
2. Nine (9) points in a row on same side of center line
SPC & Cpk

3. Six (6) points in a row, all increasing or all decreasing

4. Fourteen (14) points in a row, alternating up and down
5. Two (2) out of three (3) points > 2 standard deviations from
the center line (Same side)
6. Four (4) out of five (5) points > 1 standard deviations from the
center line (Same side)
7. Fifteen (15) points in a row within 1 standard deviation of the
center line (Either side)
8. Eight (8) points in a row within 1 standard deviation of the
center line (Either side)

SPC & Cpk

 Detecting Lack of Control

What do you do when you

SPC & Cpk

determine that a process is not

stable (it’s out of control)?

Check it out!

SPC & Cpk

SPC & Cpk

Process Capability
(% Non-Conforming, Capability Analysis Cp & Cpk)

SPC & Cpk

 Process Capability
Assesses the relationship
between natural variation of a
process and design
specifications
SPC & Cpk

 An indication of process performance with respect to

upper and lower design specifications

Application of Process Capability

 Design products that can be manufactured with
existing resources
 Identify process’ weaknesses
 Select and qualify new processes
 Link successive processes to maintain design
requirements

SPC & Cpk

 Process Capability Illustrated
Cp = 0.5 Cp = 1.0
Cpk = 0.5 Cpk = 0.5
DPM = 133,614 DPM = 66,810

σ Quality σ Quality
Level = 1.5 Level = 1.5
SPC & Cpk

Cp = 2.0
Cp = 1.0 Cpk = 1.0
Cpk = 1.0 DPM = 1,350
DPM = 2,700
σ Quality
σ Quality Level = 4
Level = 3

Cp = 2.0
Cpk = 2.0
DPM = .002 Six Sigma
(without Shift)
σ Quality
Level = 6

SPC & Cpk

 DPM & Quality Levels
Quality Level (sigma)
Off-Centering
(sigma)
3 3.5 4 4.5 5 5.5 6
0 1,350 233 32 3.4 0.29 0.017 0.002
0.25 3,577 666 99 12.8 1.02 0.1056 0.0063
SPC & Cpk

0.5 6,440 1,382 236 32 3.4 0.71 0.019

0.75 12,288 3,011 665 88.5 11 1.02 0.1
1 22,832 6,433 1,350 233 32 3.4 0.39
1.25 40,111 12,201 3,000 577 89 11 1
1.5 66,803 22,800 6,200 1,350 233 32 3.4
1.75 105,601 40,100 12,200 3,000 577 88 11
2 158,700 66,800 22,800 6,200 1,300 233 32

SPC & Cpk

 Steps for calculating
Cp, Cpku & CpkL
 Determine upper and lower spec limits
 Calculate the mean
 Determine or estimate the standard deviation
SPC & Cpk

 Calculate Cp
 Calculate Cpku & CpkL
 Cpk should be at least 1.0 or better yet, 1.33
 An acceptable Cpk value is dependent upon:
 How critical the measured process is to downstream processes
 How a non-conformance will impact on customer satisfaction
 Industry type must also be considered ( i.e. automotive vs.
general industry)
 A Cpk value of 2.0 indicates a 6σ process - a goal we
should attempt to attain!

SPC & Cpk

 Calculating Cp & Cpk
 Recall:
 Cp: Process Capability Index
 Cpk: Process capability considering location within the
upper and lower specification limits or tolerance
SPC & Cpk

 Cpk is the smaller of the Cpku or CpkL

CpkU = (USL - X)
3σ
Cp = (USL - LSL)
6σ
CpkL = (X - LSL)
3σ

SPC & Cpk

SPC & Cpk

SPC & Cpk