Control Charts

+3σ UCL
A
+2σ
B
+1σ
C
0 CL
C
-1σ
B
-2σ
A
-3σ LCL
Control Charts
An Introduction to Statistical Process Control
55 Special Cause Variation Region UCL = Upper Control Limit
3
50
Common Cause Variation Region
2
45
1
40
Values
99.7%
95%
68%
35
30
-1
25
-2
20
-3
15 LCL = Lower Control Limit
Special Cause Variation Region
Course Content
• Prerequisites
• Course Objectives
• What is SPC?
• Control Chart Basics
• Out of Control Conditions
• SPC vs. SQC
• Individuals and Moving Range Chart
• Central Limit Theorem
• X-bar and Range Charts
• Advanced Control Charts
• Attribute Charts
• Final Points
• Reference Section
http://www.biz-pi.com Control Charts

Prerequisites
• Learners should be familiar with the

following concepts prior to taking this course
– Variation
– Mean and Standard Deviation
– Histograms
– Normal Distributions
– Cp and Cpk
• Capability Course is available on BPI
website if you need to review these topics

Course Objectives
• Upon completion of this course, participants

should be able to:
– Understand the basics of creating variable and
attribute control charts
– Understand the concepts of advanced control
charting
– Identify an out of control condition
– Identify which control chart to use with each
process
– Calculate control limits for any control chart

Statistical Process Control (SPC)
• SPC is the application of statistical

methods to identify and control the
special cause of variation in a process
• SPC is a preventative tool to:
– Assess the consistency of a process
– Monitor a process to determine when it has
changed
– Reduce variation in a process

Run Chart
• Graph that displays observed data in a time

sequence
50
45
40
Values
35
30
25
20
Time

Control Charts
• Run chart with calculated control limits

– 3 standard deviations above and below average
• Help distinguish process variation due to
assignable or “common” causes from those
due to unassignable or “special” causes
• Used to detect whether a process is
statistically stable

History
• Invented by Walter A. Shewhart while

working for Bell Labs in the 1920s
• W. Edwards Deming became the foremost
champion of Shewhart's work
– long career as an industrial consultant in Japan,
spread use of the control charts throughout the
1940s and 1950s

Special vs. Common Cause
• Special Cause Variation

– Data points outside of control limits
– Trend or shift pattern within limits
– Can be removed from a process
• Common Cause Variation
– Noise within the system, typical, expected
– Data points randomly occurring within the
control limits
– Always exists in a process, but can be reduced

Control Chart Basics
55 Special Cause Variation Region

UCL
50
45
3σ
40
Values
35
CL
30
25 3σ
20 Common Cause Variation Region
15 LCL
10
Observation #4 Time
Value = 25

Special vs. Common Cause
Common Cause Special Cause

Real Life Examples
• Process: Driving to Work

• Average Time: 12 minutes
• Standard Deviation: 2.5 minutes
• Common Causes
– Wind speed, miss one green light, driving speed, number of cars on
road, time when leaving house, rainy weather
• Special Causes
– Stop for school bus crossing, traffic accident, pulled over for
speeding, poor weather conditions, car mechanical problems,
construction detour, stoplights not working properly, train crossing

Control Limits
• Centerline (CL) = average value of observations

• Upper Control Limit (UCL) ~ 3 standard
deviations ABOVE the centerline
• Lower Control Limit (LCL) ~ 3 standard
deviation BELOW the centerline
• Control limits are set when process is “in
control” or “stable”
– Fixed at baseline value
– Adjusted for improvements
– Never widened
• Control limits are not specification limits
Stability
• A process is considered “in control” or

“stable” when the data does not show any
out of control conditions on the control
chart
• Stability means predictability
Day 5
Day 4
Day 3
Day 2
Day 1
Control Limit Zones
+3σ UCL
A
+2σ
B
+1σ
C
0 CL
C
-1σ
B
-2σ
A
LCL
-3σ

Histogram vs. Control Chart
55 Special Cause Variation Region UCL = Upper Control Limit
3
50
2
45
1
40
Values
99.7%
95%
68%
35
30
-1
25
-2
20
-3
15 LCL = Lower Control Limit

Out of Control (OOC) Conditions

Out of Control Conditions
• Outside control limits is easiest to identify

• Patterns highlight other out of control
conditions
– Trends (increasing/decreasing points)
– Shifts (data jumps higher or lower than normal)
– Inconsistencies (not random, more or less
variation than history)

Western Electric Rules for Control
Any point outside control limits

7 consecutive points on same
UCL side of centerline
A 7 consecutive points increasing
B or decreasing
C 2 of 3 points in same zone A or
±3σ ±2σ ±1σ
C beyond
B 4 of 5 points in same zone B or
A beyond
LCL
14 consecutive points alternating
up and down
14 consecutive points in either
zone C

Nelson Tests for Control
 Any point outside control limits

 9 consecutive points on same
side of centerline
 6 consecutive points
UCL
A increasing or decreasing
B  2 of 3 points in same zone A or
beyond
±2σ
C
±3σ
6S 4S ±1σ
2S  4 of 5 points in same zone B or
C
beyond
B
 14 consecutive points alternating
A up and down
LCL
 15 consecutive points in either
zone C
 8 points in a row outside zone
C, same side of centerline
False Alarm Rates are the Key
Nelson False Alarm Rate

• Any point outside control limits .0027
• 9 consecutive points on same side of centerline Approx .003
• 6 consecutive points increasing or decreasing Approx .003
• 2 of 3 points in same zone A or beyond .00305
• 4 of 5 points in same zone B or beyond .0043
• 14 consecutive points alternating up and down Approx .004
• 15 consecutive points in either zone C Approx .003
• 8 points in a row outside zone C, either side of Approx .003
centerline
The Nelson tests are designed so that the false alarm rates for all
tests are approximately the same. The Western Electric rules do not
have this property.

9 Consecutive on Same Side
• If process is stable and normal, 50% of data will

be above centerline, 50% will be below
centerline
• Probability of getting 9 straight on one side,
same as flipping coin and getting 9 straight
heads
(.5)10 = 0.5 * 0.5 * 0.5 * 0.5 * 0.5 * 0.5 * 0.5 * 0.5 * 0.5 * 0.5 = .001
(.5)9 = 0.5 * 0.5 * 0.5 * 0.5 * 0.5 * 0.5 * 0.5 * 0.5 * 0.5 = .002
(.5)8 = 0.5 * 0.5 * 0.5 * 0.5 * 0.5 * 0.5 * 0.5 * 0.5 = .004

Nelson Test #1
+3σ UCL
A
+2σ
B
+1σ
C
0 CL
C
-1σ
B
-2σ
A
-3σ LCL
Rule 1: Any point outside control limits

Nelson Test #2
+3σ UCL
A
+2σ
B
+1σ
C
0 CL
C
-1σ
B
-2σ
A
-3σ LCL
Rule 2: 9 consecutive points on same side of centerline

Nelson Test #3
+3σ UCL
A
+2σ
B
+1σ
C
0 CL
C
-1σ
B
-2σ
A
-3σ LCL
Rule 3: 6 consecutive points increasing or decreasing

Nelson Test #4
+3σ UCL
A
+2σ
B
+1σ
C
0 CL
C
-1σ
B
-2σ
A
-3σ LCL
Rule 4: 2 of 3 points in same zone A or beyond

Nelson Test #5
+3σ UCL
A
+2σ
B
+1σ
C
0 CL
C
-1σ
B
-2σ
A
-3σ LCL
Rule 5: 4 of 5 points in same zone B or beyond

Nelson Test #6
+3σ UCL
A
+2σ
B
+1σ
C
0 CL
C
-1σ
B
-2σ
A
-3σ LCL
Rule 6: 14 consecutive points alternating up and down

Nelson Test #7
+3σ UCL
A
+2σ
B
+1σ
C
0 CL
C
-1σ
B
-2σ
A
-3σ LCL
Rule 7: 15 consecutive points in either zone C

Nelson Test #8
+3σ UCL
A
+2σ
B
+1σ
C
0 CL
C
-1σ
B
-2σ
A
-3σ LCL
Rule 8: 8 points in a row outside zone C, either side

Nelson Tests in Minitab TM
Stat > Control Charts > Xbar-R

SPC vs. SQC

Inputs and Outputs
INPUTS OUTPUTS
ROUTE
GAS USED
LEAVE TIME
TIRE PRESSURE PROCESS TOTAL TIME

GAS GRADE
SPEED ACCIDENT FREE

Evolution of control charts
SPC SQC
Statistical Statistical
Process Quality
Control Control
ROUTE
GAS USED
LEAVE TIME
TIRE PRESSURE PROCESS TOTAL TIME

GAS GRADE
SPEED ACCIDENT FREE

Individuals Chart

Individuals and Moving Range Chart
• Most common type of control chart

– Each individual value plotted over time
– Difference from previous value to current
value plotted on Moving Range chart
– Moving Range average used to calculate
control limits for individual readings

Individuals and MR example
VARIATION
AVERAGE

Individuals and MR UCL and LCL
UCLX  X  E2 R
LCLX  X  E2 R
UCLMR  D4 R
LCL MR  D3 R
UCLMR
E2 , D3 and D4 are constants in table at end of section

Individuals Chart
UCL
6:55 PM
0.204
CL
9:35 PM
0.169
LCL

Moving Range Chart
UCL
4:05 PM 10:15 PM
Current: 0.204 Current: 0.176
Prior: 0.172 Prior: 0.202
Moving Moving
Range Range
0.032 0.026
CL
LCL
Moving Range = Current Data Value – Prior Data Value

Non-Normal Data
• Individuals control charts rely on assumption

that data is normally-distributed
• If data does not pass normality test, what
can be done?
– Transform the data (difficult)
– Use central limit theorem to normalize the data
• X-bar and R chart does this for you

Central Limit Theorem

• No matter what the shape of the original

distribution (non-normal), the sampling
distribution of the mean approaches a
normal distribution
– normal distribution is approached very quickly
as subgroup size increases

Subgroup Size = 1
60
50
40
30
20
10
0
90 105 120 135 150 165
Number of Samples = 200

Subgroup Size = 2 Subgroup Size = 3

35 30
30 25
25
20
20
15
15
10
10
5 5
0 0
100 110 120 130 140 150 115 120 125 130 135 140
Number of Samples = 100 Number of Samples = 66
Subgroup Size = 4 Subgroup Size = 5

25 20
20
15
15
10
10
5
5
0 0
123 126 129 132 135 138 129 130 131 132 133 134
Number of Samples = 50 Number of Samples = 40

Rational Subgroups
• In order to use Central Limit Theorem, must

define the rational subgroups for data set
• Should be grouped in a way to:
– maximize the chance of detecting shifts in
process average (between subgroups)
– minimize the variation (range) between samples
within the subgroup (within subgroups)
• Don’t subgroup different machine setups,
different time of day, material differences,
operators, or any other variables

Central Limit Theorem Summary
• If you analyze the average of a subgroup,

not the actual data values, it will be normally-
distributed
• X-bar and R Chart applies Central Limit
Theorem
– Any data set can use the X-bar and R chart,
regardless of data distribution (even non-
normal)

X-bar and R Charts

Benefits of X-bar and R chart
• Can be used with any distribution of data

• Separates variation into two groups, for
ease of investigation
– Between subgroup (X-bar chart)
– Within subgroup (Range chart)
• Quicker identification of out of control
conditions than Individuals chart

X-bar and R example
AVERAGE
VARIATION
Between Subgroup Variation Within Subgroup Variation

X-bar & R UCL and LCL
UCLX  X  A2 R
LCLX  X  A2 R
UCLR  D4 R
UCLRR  D3 R
LCL
A2 , D3 and D4 are constants in table at end of section

Average (X-bar) Chart
6:55 PM
UCL
45
43
48
45
50
Avg = 46.2
CL
9:35 PM
44
48
43
42
45 LCL
Avg = 44.4

Range Chart
6:55 PM
45 UCL
43
48
45
50
Range = 7
CL
9:35 PM
44
48
43
42
45
LCL
Range = 6
Range = Max of Data Subgroup – Min of Data Subgroup

Normalized Charts
• If operating in a low volume situation, it is usually

more practical to use one control chart for
several parts or transactions
– Data must be “normalized” before it is plotted to
account for different spec limits of the data
• Process may have different limits or targets,
want to know how close process is to target
– Forecasting (Actual v. Predicted), product differences
from same process, etc
– Instead of plotting actual value, plot difference from
nominal, target or expected

Key Points for Normalized Charts
• Data should be from the same process

• Data should have similar variation
• Requires more data points to setup than
traditional charts
• Uses same calculations as traditional charts,
except use normalized values

Example Normalized Chart
Part B Part C
Part A Target = 22 Target = 10
Target = 6
3:31
Target 3:45
22 Target
Reading 10
3:15 23.2 Reading
Target Diff = 1.2 8.9
6 Diff = -1.1
Reading
4.4
Diff = -1.6

Adjust Sampling
• As processes improve, reduce sampling

frequency and/or subgroup size
– Reduce as Cpk increases
– More efficient use of resources
– If process goes out of control, increase sampling
frequency until process stabilized
• Concerns with reduced sampling
– Less sensitive to small shifts in process
– Can be harder to setup process for random
sampling, than capturing all measurements

Steps to Creating Control Charts
• Determine which process characteristic needs to be

controlled
• Determine where in the process the control chart should be
implemented
• Determine type of data needed to control process
• Choose the correct type of control chart
• Collect and calculate subgroup data
• Calculate centerlines and control limits
• Plot the data on the chart
• Interpret the chart for out of control conditions
• Improve process based upon analysis of control chart

Advanced Control Charts

Small Shift Example
1σ Process Shift
Process Mean = 11, σ = 1
Process Mean = 10, σ = 1
INDIVIDUALS CHART STILL SHOWS PROCESS IN CONTROL

10 DAYS AFTER SHIFT
Advanced Control Charts
• Typical control charts don’t use prior

readings, only current data point
– Make small shifts hard to detect (typically less
than 1.5σ)
• Use prior readings in calculation of current
data point to detect small shifts
• Two alternative charts are CUSUM and
EWMA

CUSUM
• Cumulative Sum
• Data converted to difference from mean, and
added to previous value Difference from
Mean
Mean(μ==10
10)
Date Value Value-mean New Value

-2.56 = -2.01 + -0.55
1-Oct 9.45 -0.55 -0.55
2-Oct 7.99 -2.01 -2.56
3-Oct 9.29 -0.71 -3.27
-1.61 = 1.66 + -3.27
4-Oct 11.66 1.66 -1.61

CUSUM Chart detects shift earlier!
1σ Process Shift
CUSUM CHART DETECTS SMALL SHIFT

ONLY 6 DAYS AFTER SHIFT (4 DAYS EARLIER)
How to set control limits for CUSUM
Table for obtaining CUSUM limits similar to Shewhart control charts
k 0.25 0.5 0.75 1.0 1.25 1.5

Mean Shift 0.5 1.0 1.5 2.0 2.5 3.0
Decision (h) 8.01 4.77 3.34 2.52 1.99 1.61
Example: To detect a mean shift of 1.5 standard

deviations, you would set k = 0.75, and the limits on the
CUSUM chart would be computed using h*σ (3.34σ),
which would be 3.34(1) = ± 3.34
Prior example used shift of 1σ, k=0.5 and h=4.77, so

limits were set at ±4.77
EWMA
• Exponentially Weighted Moving Average

• Good for detecting small shifts in processes
• Better at detecting large shifts than CUSUM
• Uses prior readings in calculation of current
data point
– the further away, the less influence
• Best if used with data from individuals chart
(subgroup size = 1)

How to setup EWMA chart
• Determine λ (between 0 and 1)

– λ is the proportion of current value used for calculating
newest value
– Recommend λ = 0.10, 0.20 or 0.40 (use smaller λ
values to detect smaller shifts)
• Calculate new z values using λ = 0.10
zi = λ*xi + (1 – λ) * zi-1 (where i = sample number)
9.7495 = (0.9*9.945)
Sample Date x z
+ (0.1*7.99)
1 1-Oct 9.45 9.945
2 2-Oct 7.99 9.7495
3 3-Oct 9.29 9.7035 9.899 = (0.9*9.7035)
4 4-Oct 11.66 9.899 + (0.1*11.6)
EWMA Chart detects shift earlier!
1σ Process Shift
λ = 0.10
EWMA CHART DETECTS SMALL SHIFT

2 DAYS EARLIER THAN INDIVIDUALS CHART
EWMA UCL and LCL
UCLz = u + Lσ
√ λ
2-λ
LCLz = u - Lσ
√ λ
2-λ
EWMA Chart uses L for determining limits, typically set to 2.7

EWMA Chart detects shift even earlier!
1σ Process Shift
λ = 0.05
EWMA CHART NOW DETECTS SMALL SHIFT

SIMILAR TO CUSUM CHART
Summary of CUSUM and EWMA
• CUSUM
– Uses current reading difference from mean, and
previous CUSUM value, equally weighted
• EWMA
– Uses weighted current reading and weighted
previous EWMA value
– Weight of each value is determined by user,
based on needs of chart

Disadvantages to CUSUM and EWMA
• Cannot quickly detect large shifts in process,

like traditional control charts
• Use CUSUM and EWMA for detecting small
shifts, Individuals and X-bar charts for larger
shifts
– Ideally, use both within the process, to detect
both small and large shifts, with limits set to
±3.5σ

Attribute Control Charts

Attribute Control Charts
• p chart
• np chart
• c chart
• u chart

p chart
• Plots the percentage of defectives within a

sample
3 red defective parts out of 20 parts
p = 3 / 20 = 15% defect rate
• Use when sample size varies

– Control limits adjust according to sample size

p chart UCL and LCL
√
p (1 - p)
UCLp = p + 3 n
√
p (1 - p)
LCLp = p 3 n
p Chart has adjusting control limits, based on sample size

p chart
8:00 PM
Parts Sampled = 214
11:15 AM Number Defective = 12
Parts Sampled = 415 Percent Defective = 5.61%
Number Defective = 18 Limit Width = 4.7%
Percent Defective = 4.34% Defective Average = 7%
Limit Width = 3% LCL = 2.3%
Defective Average = 7%
UCL = 11.7%
LCL = 4%
UCL = 10%

np chart
• Plots the number of defective parts in a

sample
• Requires the same sample size each time
• Easy to use, since no calculations required

np chart UCL and LCL
UCLnp = np + 3 √ np (1 - p)
LCLnp = np 3√ np (1 - p)

np Chart
4:05 PM 9:35 PM
Sample Size = 750 Sample Size = 750
Total Defects = 35 Total Defects = 29
This specific chart uses 750 samples each time

u chart
• Plots the quantity of defects per part in a

sample
• Each part can have more than one defect
• Use when sample size varies
11 total defects found on 6 documents

u = 11/6 = 1.833 defects per document
whereas p = 4/6 = 67% defect rate
u chart UCL and LCL
√
u
UCLu = u + 3
n
√
u
LCLu = u 3
n
u Chart has adjusting control limits, based on sample size

u Chart
12:35 PM
Number of Parts Sampled = 74
5:25 PM
Number of Defects Found = 55
Number of Parts Sampled = 120
Defects per part = 0.74
Number of Defects Found = 70
Limit Width = 0.29
Defects per part = 0.58
Defects per unit average = 0.681
Limit Width = 0.23
LCL = 0.393
Defects per unit average = 0.681
UCL = 0.969
LCL = 0.455
UCL = 0.907

c chart
• Plots the quantity of defects in a sample

• Each part can have more than one defect
• Requires same number of parts within each
sample
11 total defects found on 6 documents

c = 11 defects per sample
c chart UCL and LCL
UCLc = c + 3
√c
LCLc = c 3
√c

c Chart
1:50 PM
Sample 1: 7
Sample 2: 9
Sample 3: 0
Sample 4: 8
Sample 5: 11
Total Defects = 35
6:55 PM
Sample 1: 3
Sample 2: 5
Sample 3: 10
Sample 4: 0
Sample 5: 3
Total Defects = 21
This specific chart uses 5 samples each time

Attribute Selection Chart
Data Type
Attribute Defectives Count of
Charts (Pass/Fail) Defects
Sample Size
Constant
np c
Varies
p u
Final Points

Control Chart Implementation Phases
• Phase 1: Control Charts on few process outputs (key

quality characteristics)
• Phase 2: Expansion of charts, including numerous attribute
control charts
• Phase 3: Some control charts on few process inputs (key
process parameters), some converted to X-bar and R
charts
• Phase 4: Attribute control charts replaced with variable
charts, many charts deleted due to non-criticality or
ineffectiveness
• Phase 5: Most charts are X-bar and R charts on key
process parameters, some advanced charts (CUSUM,
EWMA, etc)

What if you react to common cause?
Adjusting Process
To Target of 200
Next Reading
After Each Reading
would be = 229 if
adjusted +17,
actually only 212 No Adjustment,
Random Variation
Actual Reading=183,
Adjust process +17
to get back to 200
Adjusting process back to center will increase overall variation

Other Key Points
• Create a control chart and histogram before performing any

data analysis or calculations
• Use at least 20 data points before calculating control limits
• Moving Range charts are optional, and not required with
Individuals charts
• Spec limits may be displayed on Individuals charts, but can
lead to complacency, never on X-bar and R charts
• Apply SPC to the inputs of a process whenever possible
• Use X-bar and Standard Deviation (S) chart instead of X-
bar and R chart when subgroup size > 5
• Special causes are indications of potential problems, they
cannot guarantee that a problem exists
• Document any out of control condition observed on a chart,
to show evidence that it was seen, and some investigation
took place
Common Pitfalls of SPC Failure
• Lack of commitment from management

– Resistance to change from reactive to preventative
• Lack of training and education in SPC
– Misinterpretation of control charts
– Lack of focus and/or maintenance on control charts
• Inadequate measurement system in place (poor
Gage R&R)

Reference Section

Decision Tree for Control Charts
What type of data:

Attribute or Variable?
Attribute Variable
How are defects counted:
How large are the subgroups?
Defectives (Y/N), or Count of Defects?
Defectives Count 1 2 to 5 5 or more
Constant Constant
Sample Size? Sample Size?
Individuals and X-bar and X-bar and
Moving Range Range Std Dev
Yes No Yes No
np chart P chart c chart u chart
(number (proportion (defects per (defects per
defective) defective) sample) unit)

Shewhart Control Chart Constants
n E2 D4 D3 A2
2 2.66 3.27 0.00 1.88
3 1.77 2.57 0.00 1.02
4 1.46 2.28 0.00 0.73
5 1.29 2.11 0.00 0.58
6 1.18 2.00 0.00 0.48
7 1.11 1.92 0.08 0.42
8 1.05 1.86 0.14 0.37
9 1.01 1.82 0.18 0.34
n is the subgroup size
Additional Resources
Business Performance Improvement

http://www.biz-pi.com

Control Charts

Uploaded by

Copyright:

Available Formats

Control Charts

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Control Charts

Uploaded by

Copyright:

Available Formats

+3σ UCL

55 Special Cause Variation Region UCL = Upper Control Limit

http://www.biz-pi.com Control Charts

• Learners should be familiar with the

http://www.biz-pi.com Control Charts

• Upon completion of this course, participants

http://www.biz-pi.com Control Charts

• SPC is the application of statistical

http://www.biz-pi.com Control Charts

• Graph that displays observed data in a time

http://www.biz-pi.com Control Charts

• Run chart with calculated control limits

http://www.biz-pi.com Control Charts

• Invented by Walter A. Shewhart while

http://www.biz-pi.com Control Charts

• Special Cause Variation

http://www.biz-pi.com Control Charts

55 Special Cause Variation Region

http://www.biz-pi.com Control Charts

Common Cause Special Cause

http://www.biz-pi.com Control Charts

• Process: Driving to Work

http://www.biz-pi.com Control Charts

• Centerline (CL) = average value of observations

• A process is considered “in control” or

http://www.biz-pi.com Control Charts

55 Special Cause Variation Region UCL = Upper Control Limit

http://www.biz-pi.com Control Charts

http://www.biz-pi.com Control Charts

• Outside control limits is easiest to identify

http://www.biz-pi.com Control Charts

Any point outside control limits

http://www.biz-pi.com Control Charts

 Any point outside control limits

Nelson False Alarm Rate

http://www.biz-pi.com Control Charts

• If process is stable and normal, 50% of data will

http://www.biz-pi.com Control Charts

Rule 1: Any point outside control limits

http://www.biz-pi.com Control Charts

Rule 2: 9 consecutive points on same side of centerline

http://www.biz-pi.com Control Charts

Rule 3: 6 consecutive points increasing or decreasing

http://www.biz-pi.com Control Charts

Rule 4: 2 of 3 points in same zone A or beyond

http://www.biz-pi.com Control Charts

Rule 5: 4 of 5 points in same zone B or beyond

http://www.biz-pi.com Control Charts

Rule 6: 14 consecutive points alternating up and down

http://www.biz-pi.com Control Charts

Rule 7: 15 consecutive points in either zone C

http://www.biz-pi.com Control Charts

Rule 8: 8 points in a row outside zone C, either side

http://www.biz-pi.com Control Charts

Stat > Control Charts > Xbar-R

http://www.biz-pi.com Control Charts

http://www.biz-pi.com Control Charts

TIRE PRESSURE PROCESS TOTAL TIME

http://www.biz-pi.com Control Charts