SPC for MS Excel V2 - BPI Consulting for MS Excel V3.0 Instructions.pdf · Business Process...

Business Process Improvement 281-304-9504

20314 Lakeland Falls www.spcforexcel.com

Cypress, TX 77433

Instruction Manual for SPC for MS Excel V3.0

Capability Analysis

LSL=60 USL=80Nominal=70

0

5

10

15

20

25

30

35

57 62 67 72 77 82 87

Measurement

Fre

qu

en

cy

Statistics

Cp=1.34

Cpk= 0.59

Cpu= 0.59 (3.84%)

Cpl= 2.09 (0%)

Est. Sigma= 2.49

Pp=1.31

Ppk= 0.57

Ppu= 0.57 (4.36%)

Ppl= 2.04 (0%)

Sigma= 2.55

Average=75.63

Count=96

No. Out of Spec=5 (5.21%)

Kurtosis=0.62

Skewness=0.71

Thank you for selecting our software package. This program was written by Dr. William H. McNeese

and is distributed by Business Process Improvement (Cypress, Texas). This program cannot be copied or

used unless under license with Business Process Improvement. Business Process Improvement is not

liable for any decisions made based on the use of this software package.

Requirements: This program is a Microsoft Excel® add-in. You must Microsoft Excel® for this program

to work. This program supports any version of Excel from 2000 on.

Business Process Improvement

20314 Lakeland Falls

Cypress, TX 77433

281-304-9504

www.spcforexcel.com

http://www.spcforexcel.com/

2 ©2007 Business Process Improvement

SPC for MS Excel V3.0

Table of Contents

Instructions Manual for SPC for MS Excel V3.0 ......................................................................................... 1 Installation..................................................................................................................................................... 4 Pareto Diagrams ............................................................................................................................................ 6 Histograms .................................................................................................................................................. 10 Attribute Control Charts ............................................................................................................................. 13

p Charts ................................................................................................................................................... 13 np Control Charts .................................................................................................................................... 16 c Control Charts ...................................................................................................................................... 18 u Control Charts ...................................................................................................................................... 20

Variable Control Charts .............................................................................................................................. 22

X -R Control Chart ................................................................................................................................. 22 Control Limit Option ............................................................................................................................... 26

X -s Control Charts ................................................................................................................................. 27 X-MR (Individuals) Control Charts ........................................................................................................ 28 Moving Average/Moving Range (MA/MR) Charts ................................................................................ 30 Table X-MR (Individuals) Chart............................................................................................................. 31 Run Charts .............................................................................................................................................. 32 Subgroup Maker: Make Subgroups from Column of Numbers .............................................................. 32

Process Capability ....................................................................................................................................... 33 Advanced Process Capability...................................................................................................................... 38 Scatter Diagram .......................................................................................................................................... 40 Updating Charts .......................................................................................................................................... 43 Changing Chart Options ............................................................................................................................. 43 Single Point Actions ................................................................................................................................... 44 All Points Action......................................................................................................................................... 46 Cause and Effect Diagram .......................................................................................................................... 47 Measurement Systems Analysis.................................................................................................................. 48

ANOVA Method ..................................................................................................................................... 53 Range Method for Gage R&R ................................................................................................................ 55 Bias – Independent Sample Method ....................................................................................................... 57 Bias – Control Chart Method .................................................................................................................. 59 Linearity .................................................................................................................................................. 61 Attribute Gage R&R ............................................................................................................................... 62

Transfer Charts to PowerPoint or Word ..................................................................................................... 65 Regression ................................................................................................................................................... 66

Changing the Variables in the Regression .............................................................................................. 69 Miscellaneous ............................................................................................................................................. 70

Descriptive Statistics ............................................................................................................................... 70 Confidence Interval Around a Mean ....................................................................................................... 71 Confidence Interval Around a Variance ................................................................................................. 73 Confidence Interval for the Difference in Two Means ........................................................................... 74 Confidence Interval for Multiple Processes ............................................................................................ 76 Paired Sample Comparison ..................................................................................................................... 78 Analysis of Means ................................................................................................................................... 79 Correlation Coefficients .......................................................................................................................... 81 Failure Mode and Effect Analysis .......................................................................................................... 82


Box and Whisker Plots ............................................................................................................................ 83 Sample Size Calculator ........................................................................................................................... 85 Side by Side Histogram .......................................................................................................................... 86 Plot Multiple Y Variables Against One X Variable................................................................................ 88

Select Cells.................................................................................................................................................. 89 Frequently Asked Questions ....................................................................................................................... 90

What out of control tests does the program use? .................................................................................... 90 Do all out of control tests apply to all the charts? ................................................................................... 90 How do I know if the chart has any out of control points? ..................................................................... 90 Can I remove the out of control points from the calculations? ............................................................... 90 Can I change the name of the worksheet tab containing the chart? ........................................................ 90 How come I can’t see the name of one of my charts in the list of charts to be updated? ....................... 90 How can I change the title or the x and y labels on an existing chart? ................................................... 91


Installation

The necessary files to run the program are installed when you run the installation program. The

installation file is the .exe program you downloaded or received on the CD. Running the setup.exe file

creates the following directory (there are slight differences for Excel 2007):

C:\Documents and Settings\{username}\Application Data\SPC_for_MS_Excel

The following program files are installed into this directory:

spcforexcelv3.xla

spcfor2007excelv3.xlam

installspcforexcelv3.exe

The installation process may leave uninstallation files such as unins001.exe and unins001.dat in this

directory. During operation, the program may save user preferences and other settings in one or more text

or binary files in this directory. The user is discouraged from altering any of these files and from storing

any work files in this directory.

The program also creates the following directory:

C:\Documents and Settings\{username}\My Documents\SPC for MS Excel

The following sample data files and instruction manual are installed into this directory:

Gage R&R Example Workbook.xls

SPC Example Data V3.xls

SPC for MS Excel V3.0 Instructions.pdf

The user is encouraged to use these files to learn how SPC for MS Excel works. If you also purchased the

PowerPoint training modules, they will be installed this directory as well.

The program is installed as an add-in. It will open whenever Excel is opened. In Excel 2000 to Excel

2003, SPC for Excel will appear on the Worksheet and Chart Menus next to Window. There is also a free

standing toolbar that can be placed anywhere in the window. Both are shown below.


In Excel 2007, SPC for Excel appears on the ribbon next to Home as shown below. The SPC Menu to the

right lists all the buttons.

The menu and toolbar allows you to access the various components of the program:

The data entry requirements to run each component of this program are given below. All the examples

are in the workbook SPC Example Data V3.0.xls and Gage R&R V3 Example Workbook.xls. This

instruction manual is intended to demonstrate how the program is used.

To learn more about SPC, please refer to one of the many books on the subject. The best reference is

probably Understanding Statistical Process Control by D. Wheeler and D. Chambers, SPC Inc., 1986 or

any of the later books by Dr. Wheeler.

You can also visit our website where many of these SPC tools are described in our past free e-zines. Go

to www.spcforexcel.com.



Pareto Diagrams

A Pareto diagram is a special type of bar chart that is used to separate the "vital few" from the "trivial

many." It is based on the 80/20 rule; e.g., 20% of our customers buy 80% of our products. The

horizontal (x) axis most often represents problems or causes of problems (the “categories”). The vertical

(y) axis most often represents frequency or cost. The problem or cause that occurs most frequently (or

costs the most) is listed first on the x axis. The second most frequently occurring problem or cause is

listed second and so on. A bar is generated for each cause or problem. The height of the bar is the

frequency with which that problem or cause occurred. A cumulative percentage line is sometimes added

to the Pareto diagram.

An example of a Pareto

diagram is shown to the

right. In this Pareto

diagram, the number of

return goods by product is

analyzed. The x axis is

the different types of

products. The y axis is

how often each product

has been returned. The

bars are arranged so the

first bar (for Product B)

has the largest frequency.

The other bars are then

arranged in decreasing

frequency.

The above Pareto diagram

indicates that product B has been returned more times (25) than any other product. To reduce the number

of returned goods, one would probably want to investigate why product B is returned so often. The

highest bars represent the “vital few.” The smaller bars represent the “trivial many,” such as for products

C and D.

This program will construct Pareto diagrams with or without a cumulative percentage line. If selected,

the calculations for the cumulative percentage line are completed and added to the Pareto diagram. The

program will not alter your worksheet. The data are copied from your worksheet to a hidden sheet.

Data Entry

The data entry requirements for the Pareto diagram are shown below. In all cases, it is recommended you

select a range in the worksheet. This helps save time when the input dialog box appears. The data is in

columns in the examples below but can also be in rows.

Option 1: Basic Pareto Diagram

For this option, the frequencies have already been totaled by category. For

example, suppose you are tracking returns by product name for four products: A,

B, C, and D. You collect data for a two-month period. You then total the number

of returns and enter the data into an Excel spreadsheet. To start the Pareto

program, highlight the product names as shown to the right (shaded area) and

Products Number of Returns

A 15

B 25

C 8

D 2

Pareto Diagram

25

15

8

2

50%

80%

96%

100%

0

5

10

15

20

25

30

35

40

45

50

B A C D

Products

Nu

mb

er

of

Re

turn

s

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Pe

rcen

t


select the Pareto Diagram option from the toolbar. You could also highlight both the product and number

of returns information. The data does not have to in adjacent columns. In this case, you would select the

category range, then hold down the Control key and select the frequency range. Then select the Pareto

Diagram option on the SPC toolbar. You will get the Pareto Diagram dialog which is described below

(after the data entry requirements for option 3). Once you fill the information in the dialog box and select

OK, you will get the Pareto diagram shown above (with the cumulative line option selected).

Option 2: Basic Pareto Diagram but Program Calculates the Totals

For this option, the frequencies have been totaled over some time period

but not overall. For example, suppose you are tracking the returns and

total the returns for each product by week. In this case, you would enter

the following data into an Excel spreadsheet. You would select the

products (shaded area) and then select the Pareto diagram option from

the toolbar. The program will automatically calculate the overall totals.

Option 3: Pareto Diagram Based on Data in One

Column Only

For this option, none of the frequencies have been

totaled. For example, you might be tracking each

individual returned product to discover the reason for

returns. In this case, you would enter data similar to the

data shown below into an Excel spreadsheet. To make a

Pareto diagram based on data in one column only, select

the range in the column to include in the Pareto. Then

select the Pareto diagram option on the toolbar.

Pareto Dialog Box

When you select the Pareto Diagram option (PD) on the SPC toolbar, you will get the form below. This

example is for Option 1: Basic Pareto Diagram above with the product names selected prior to the

selecting the Pareto diagram option from the SPC toolbar. There are two pages on the form: Input

Ranges/Chart Name and Options. The Input Ranges/Chart Name always comes up first. The entries on

both pages are given below. Selecting OK at the bottom of each page will run the program. Selecting

Cancel will cancel the program. The Switch Tabs button can be used to switch between the two pages.

Week Product Number of Returns

1 A 4 1 B 2 1 C 12 1 D 3 2 A 0 2 B 3 2 C 1 2 D 4 3 A 4 3 B 2 3 C 5 3 D 1 4 A 8 4 B 1 4 C 10 4 D 1

Date Product Returned

Reason

2/1/03 A Customer Did Not Need

2/4/03 A Broken

2/6/03 B Wrong Quantity

2/9/03 C Wrong Quantity

2/11/03 A Salesman Ordered Wrong

2/14/03 A Wrong Quantity

2/14/03 D Wrong Quantity

1/2/00 B Broken

2/23/03 A Customer Did Not Need

2/24/03 D Salesman Ordered Wrong




Input Ranges/Chart Name Page

Enter Category Range: This is the range

containing the categories (used for Options 1

and 2 above). The default value is what is

selected prior to selecting the Pareto Diagram

option on the toolbar.

Enter Frequency Range: This is the range

containing the frequencies for Options 1 and 2

above. The default value is the range next to

the categories (but the categories and

frequencies do not have to be adjacent.

Name of Chart: This is very important. This

will be the name of the worksheet tab that

contains the chart in your workbook.

Include Cumulative Line Select “Yes” to

include a cumulative line. The default value is

“No.”

Categories On: Selecting X axis puts the

categories on the x (horizontal) axis. Select Y axis places the categories on the Y axis. This is

helpful if the categories have long names. You cannot use a cumulative line if the categories are

on the Y axis.

Enter Pareto Diagram Title: The default title is “Pareto Diagram.” Enter the title you want to

appear above the chart.

Enter X-Axis (Category) Label: If there is a title in the cell about the first frequency selected, this

is the default entry. Otherwise, the label is left blank. Enter the category label you want for the

x-axis.

Enter Y-Axis (Frequency) Label: If there is a title in the cell above the frequency range, this is the

default entry. Otherwise, the label is left blank. Enter the frequency label you want for the y-

axis.

Data in: Select columns or rows depending on

how the data is entered into the spreadsheet.

The program selects one or the other

depending on the range selected prior to

selecting PD on the SPC toolbar.

Dates of Data Collection: Add the starting

date and ending dates of data collection.

These dates are optional. If entered, they will

appear in a dialog box in the lower left-hand

corner of the chart.

Options Page

Calculation Options: This is Option 2: Basic

Pareto Diagram but Program Calculates the

Totals. Select “Yes” if you want the program

to total the frequency results for the various

categories. “No” is the default value. Once

you select “Yes”, you must select the option

you want. Most of the time it will be “Sum,”


but there are other options including count, average, and standard deviation.

Pareto on One Column? This is Option 3: Pareto Diagram Based on One Column. Select “Yes”

if the data are in one column. The “Data Range” contains the worksheet range containing the

data. The default value is the range that is selected prior to PD being selected on the toolbar.

“Include Frequencies >= to” is used to determine what frequencies you want to include in the

chart. For example, if you enter 3, only those items that occur three or more times will be

included in the chart.

Examples

Below are the results for Options 2 and 3 using the data given above.

Option 2: Summing the results for each week

Option 3: Reasons for return in one column

Pareto Diagram

28

16

98

46%

72%

87%

100%

0

10

20

30

40

50

60

C Total A Total D Total B Total

Product

Nu

mb

er

of

Re

turn

s

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Pe

rcen

t

Pareto Diagram

6

2 2 2

50.0%

66.7%

83.3%

100.0%

0

2

4

6

8

10

12

Wrong Quantity Broken Customer Did Not Need Salesman Ordered Wrong

Reason

0.0%

10.0%

20.0%

30.0%

40.0%

50.0%

60.0%

70.0%

80.0%

90.0%

100.0%

Pe

rcen

t


Histograms

A histogram is a bar chart that provides a snapshot in time of the variation in a process. It tells us how

often a value or range of values occurred in a given time frame. A histogram will tell us the most

frequently occurring value (the mode), the overall range, and the shape of the distribution (e.g., bell-

shaped, skewed, bimodal, etc.). It is best to have 50 to 100 data points to construct a histogram, if

possible. This program will construct a histogram from the raw data. It will automatically determine the

number of classes (bars) as well as the class width. You have the opportunity to change the number of

classes. An example of a histogram is shown below.

Data Entry

Enter the data you want to use in the histogram into a worksheet. The data

can be in any number of rows and columns. Select the cells containing the

data for the histogram as shown to the right. Then select the histogram

option (H) from the SPC toolbar.

Histogram Dialog Box

81 77 75 74 77 73 77 74 76 75 79 74 74 79 73 75 75 74 75 80 80 79 72 78 73 74 74 73 75 74 77 75 75 72 75 74 76 75 74 74 78 75 76 76 78 77 78 75 74 76 77 76 72 73 79 82 73 75 74 79 77 73 72 75 73 73 76 76 76 75 74 72 76 76 76 74 79 79 75 81 77 74 77 71 84 74 79 70 77 74 73 77 76 74 81 75

Histogram of Yields

1 (1.0%)

0 (0.0%)

1 (1.0%)

3 (3.1%)

2 (2.1%)

8 (8.3%)

4 (4.2%)

11 (11.5%)

13 (13.5%)

17 (17.7%)

19 (19.8%)

10 (10.4%)

5 (5.2%)

1 (1.0%)1 (1.0%)

0

2

4

6

8

10

12

14

16

18

20

69.5 to

70.5

70.5 to

71.5

71.5 to

72.5

72.5 to

73.5

73.5 to

74.5

74.5 to

75.5

75.5 to

76.5

76.5 to

77.5

77.5 to

78.5

78.5 to

79.5

79.5 to

80.5

80.5 to

81.5

81.5 to

82.5

82.5 to

83.5

83.5 to

84.5

Measurement

Fre

qu

en

cy

Descriptive Stats

Mean=75.625

Standard Error=0.26

Median=75

Standard Deviation=2.547

Variance=6.489

Sum=7260

Count=96

Maximum=84

Mininum=70

Range=14

Kurtosis=0.6157

Skewness=0.7079


When you select the histogram option (H) on the SPC toolbar, you will get the dialog box shown to the

left. Each entry is discussed below.

Enter location of Values: This is the range containing the values for the histogram. The default

range is the range selected on the worksheet before selecting the histogram option on the toolbar.

Enter Histogram Title: This is the title that goes on the histogram chart. The default value is

“Histogram.”

Enter Y-Axis (Vertical Label): This is the vertical axis label. The default is “Frequency.”

Enter X-Axis (Horizontal Label): This is the horizontal axis label. The default is “Measurement.”

Name of Chart: This is very important. This will be the name of the worksheet tab that contains

the chart in your workbook.

Enter Number of Integers to Right of Decimal: This is the rounding that is used in the data. For

example, if the data contains whole numbers, this value is 0 (the default value). If the data has

one decimal point to the right of the data (as shown in the data above), this value is 1. It is used

to set the class boundaries.

Dates of Data Collection: Add the starting date and ending dates of data collection. These dates

are optional. If entered, they will appear in a dialog box in the lower left-hand corner of the chart.

Include Descriptive Statistics?” If you want the descriptive statistics on the chart, select “Yes.”

The descriptive statistics include the average, standard deviation, count, etc. The default value is

“Yes.” There is also the option to “Select Which to Include.” This option allows you to

determine which of the descriptive statistics you want to include. If you select this option, you

will see the dialog box below. Select which statistics you want to include. The statistics you

select will remain the same if you update the histogram. You can “Check All” or “Uncheck All”

if desired.


The number of classes (bars) on the histogram is determined automatically

by the program. It is set as the square root of the number of data

points in the range. Once the histogram is made, you can change

the number of classes. There is a button in the upper left hand

corner of the histogram chart that is used for this (you will see it

when the histogram is first made).

When you select this button on the chart, you will get the dialog

box to the right. There are essentially two options:

Enter the number of classes you want and select OK.

The chart will then be displayed.

Enter the class width and enter the lower bound. This

lets you set the starting point for the histogram (the lower

bound) and the width of each class. The number of

classes is set by these two values.

You also have the option to view the frequency distribution for the

histogram. This is done by selecting the button with the caption

“View/Hide Frequency Distribution.” This button appears the first time the histogram is made. An

example of a histogram with the frequency distribution added is shown below. Selecting the button again

hides the frequency distribution.

Change Number of Classes

View/Hide Frequency Distribution

Histogram of Yields

1 (1.0%)

0 (0.0%)

1 (1.0%)

3 (3.1%)

2 (2.1%)

8 (8.3%)

4 (4.2%)

11 (11.5%)

13 (13.5%)

17 (17.7%)

19 (19.8%)

10 (10.4%)

5 (5.2%)

1 (1.0%)1 (1.0%)

0

2

4

6

8

10

12

14

16

18

20

69.5 to

70.5

70.5 to

71.5

71.5 to

72.5

72.5 to

73.5

73.5 to

74.5

74.5 to

75.5

75.5 to

76.5

76.5 to

77.5

77.5 to

78.5

78.5 to

79.5

79.5 to

80.5

80.5 to

81.5

81.5 to

82.5

82.5 to

83.5

83.5 to

84.5

Measurement

Fre

qu

en

cy

Descriptive Stats

Mean=75.625

Standard Error=0.26

Median=75

Standard Deviation=2.547

Variance=6.489

Sum=7260

Count=96

Maximum=84

Mininum=70

Range=14

Kurtosis=0.6157

Skewness=0.7079

Classes Freq. Rel. Freq.

69.5 to 70.5 1 1.0%

70.5 to 71.5 1 1.0%

71.5 to 72.5 5 5.2%

72.5 to 73.5 10 10.4%

73.5 to 74.5 19 19.8%

74.5 to 75.5 17 17.7%

75.5 to 76.5 13 13.5%

76.5 to 77.5 11 11.5%

77.5 to 78.5 4 4.2%

78.5 to 79.5 8 8.3%

79.5 to 80.5 2 2.1%

80.5 to 81.5 3 3.1%

81.5 to 82.5 1 1.0%

82.5 to 83.5 0 0.0%

83.5 to 84.5 1 1.0%


Attribute Control Charts

This program handles p, np; c and u attribute control charts. The data

entry depends on the type of chart you are using. You access this

feature by selecting the attribute control chart option (ATT) on the SPC

toolbar. You will see the dialog box to the right. Select the type of

chart you want to make.

p Charts

A p control chart is used to examine the variation in the proportion (or percentage) of defective items in a

group of items. An item is defective if it fails to conform to some preset specification (operational

definition). The p control chart is used with "yes/no" attributes data. This means that there are only two

possible outcomes: either the item is defective or it is not defective. For example: either the phone is

answered or it is not answered.

An example of a p chart generated by this program is given below. In this example, the percentage of

telemarketing calls that result in an order each day is being examined. "n" is the subgroup size (the

number of telemarketing calls made each day). "np" is the number of "defective" items -- in this case, the

number of calls that result in an order. "p" is the proportion defective and is determined by p = np/n. For

example, on the first day there were 40 telemarketing calls made (n = 40). Of these, 5 resulted in an order

(np = 5). Thus, p = np/n = 5/40 = 0.125 or 12.5%. In the chart below, 12.5% is the point plotted on

2/1/2003.

The values of p are plotted

over time. The average

( p ), the upper control limit

(UCL) and the lower control

limit (LCL) are calculated

using the equations below.

The average is plotted as a

green solid line and the

control limits are plotted as

red dashed lines. The

control limits in this

example vary because the

subgroup size varies. The

values for the average and

control limits (based on the

average subgroup size, n )

are also printed on the chart

or in the title depending on

the option selected.

n

npp

n

)p1(p3pUCL

n

)p1(p3pLCL

p Control Chart

Avg=19.47

UCL=36.16

LCL=2.780%

5%

10%

15%

20%

25%

30%

35%

40%

45%

2/1/

2003

2/2/

2003

2/3/

2003

2/4/

2003

2/5/

2003

2/6/

2003

2/7/

2003

2/8/

2003

2/9/

2003

2/10

/200

3

2/11

/200

3

2/12

/200

3

2/13

/200

3

2/14

/200

3

2/15

/200

3

Subgroup Number

% D

efe

cti

ve


The above control limits are not valid for the “small np case.” This occurs when n p < 5 or n(1- p )< 5.

In this case, the program automatically calculates the control limits using the binomial distribution.

Data Entry

The p chart monitors the fraction or percentage of defective

items in a group of items. Subgroup number (like the date

shown to the right), subgroup size (n) and number

nonconforming (np) are required as shown in the example

below. After entering the data, highlight the subgroup

numbers (in the example these are the dates). Then select the

attribute control chart option (Att) from the SPC toolbar and

select the p control chart option.

p Chart Dialog Box

Once you select the p control chart option, you will get the

dialog box shown to the right. There are two pages for this

dialog box. Each page is discussed below. Selecting OK

at the bottom of each page will run the program. Selecting

Cancel will cancel the program. The Switch Tabs button

can be used to switch between the two pages.

Input Ranges/Chart Name/Labels Page

Range containing the subgroup identifiers: This is

the range containing the subgroup numbers (dates

in the above example). The default value is the

range selected on the worksheet prior to selecting

the attribute control option on the toolbar.

Range containing the n values: This is the range

containing the subgroup size (n). The default

value is the range next to the subgroups unless you selected multiple ranges using the control key.

Range containing the np values: This is the range containing the number non-conforming (np).

The default value is the range next to the n values unless you selected multiple ranges using the

control key.

Name of Chart: This is very important. Decide what you want to call the chart. This will be the

name of the sheet that contains the chart in your workbook.

Control Chart Title: This is the title that goes on the control chart. The default value is “p

Control Chart.”

Y-Axis Label: This is the vertical axis label. The default value is “% Defective.”

X-Axis Label: This is the horizontal axis label. The default value is “Subgroup Number”

Date Number of Telemarketing Calls (n)

Number that Result in an Order (np)

2/1/2003 40 5 2/2/2003 63 10 2/3/2003 47 12 2/4/2003 52 7 2/5/2003 34 3 2/6/2003 59 21 2/7/2003 36 12 2/8/2003 71 7 2/9/2003 53 11 2/10/2003 50 3 2/11/2003 41 12 2/12/2003 48 10 2/13/2003 67 5 2/14/2003 45 12 2/15/2003 54 18


Data in: Select columns or rows depending on how the data is entered into the spreadsheet. The

program selects one or the other depending on the range selected prior to selecting Att on the SPC

toolbar.

Control Limits/Other Options Page

Test for Control: There are two options:

points beyond the limits and the rules of

seven (seven in a row above or below the

average, or seven in a row trending up or

trending down).

Automatic Update of Limits?: This

determines if the control limits are

automatically updated when you add

additional data to the chart. Select “Yes” if

you want the control limits to automatically

update; no if you don’t want the limits to

automatically update. The default is yes.

Based Limits on Average n? Select no to

change the limits each time the subgroup size

changes. Select Yes to base the limits on the

average subgroup size. The default value is

No.

Print Average/Limits: Selecting “On Avg. and Limits” will print these on the lines in the chart.

Selecting “In Chart Title” will print the values in the chart title.

Target for Average: This is the target value for the variable. It is not required.

Use Percent for Format?: Select yes to format the chart as percent; no to format the chart as a

general number.



Rounding to Use in Titles: This the rounding to use for the average and control limits printed in

the title. The default value is determined by the program.


np Control Charts

A np control chart is used to monitor the variation in the number of defective items in a group of items.

With this chart, the subgroup size (n), the number of items in the group, must be the same each time. An

item is defective if it fails to conform to some preset specification (operational definition). The np control

chart is used with "yes/no" attributes data. This means that there are only two possible outcomes: either

the item is defective or it is not defective. For example: either the phone is answered or it is not

answered.

An example of a np chart generated by this program is shown below. In this example, the number of

defective invoices each day is being tracked. The control chart is developed by taking a random sample

of 100 invoices each day and determining the number that are defective. In this case, the subgroup size is

constant (100). np is the number of defective items. For example, on day one, there were 22 defective

invoices.

The values of np are plotted


( pn ), the upper control

limit (UCL), and the lower

control limit (LCL) are

calculated using the

equations below. k is the

number of subgroups used

in the calculations (k = 15 in

this chart). The average is

plotted as a solid green line

and the control limits are

plotted as red dashed lines.

The values for the average

and control limits, along

with the subgroup size, are

printed on the chart or in the

chart title depending on the

option selected.

k

nppn

n

pnp )p1(pn3pnUCL )p1(pn3pnLCL

The above control limits are not valid for the “small np case.” This occurs when n p < 5 or n(1- p )< 5.

In this case, the program automatically calculates the control limits using the binomial distribution.

np Control Chart

Avg=24.87

UCL=37.83

LCL=11.9

0

5

10

15

20

25

30

35

40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Subgroup Number

Nu

mb

er

De

fec

tiv

e


Data Entry

A np control chart monitors the number of defective items in a constant

subgroup size. The required data to enter into the spreadsheet are the

subgroup numbers and the number of defective items as shown to the right.

Select the subgroup numbers (shaded area). Then select the attribute control

chart option (Att) from the SPC toolbar and select the np control chart

option.

np Chart Dialog Box

After selecting the np chart option, you will get the two

page dialog box shown here. Only items not described

in the p chart dialog box are explained below. See page

14 for the p chart dialog box.


Subgroup size for np chart: Enter the constant

subgroup size. It is required.


All entries are explained in the p control chart

section. See page 14 for the p chart dialog box.

Day Number

Number of Defective Invoices (np)

1 22 2 33 3 24 4 20 5 18 6 24 7 24 8 29 9 18 10 27 11 31 12 26 13 31 14 24 15 22


c Control Charts

A c control chart is used to monitor the variation in the number of defects. A defect occurs when

something does not meet a preset specification (operational definition). A c control chart is used with

counting type attributes data (e.g., 0, 1, 2, 3). These are whole numbers. You cannot have 1/2 defect. In

addition, to use a c control chart, two other conditions must be true:

The opportunity for defects to occur must be large.

The actual number that occurs must be small.

With a c control chart, we are often looking at an area, not a group of items. For example, we may use a c

control chart to monitor injuries in a chemical plant. In this case, the subgroup is the plant. The

opportunities for injuries to occur is large; the actual number that occur is small relative to the

opportunity. With a c control chart, the area of opportunity for defects to occur must be constant.

An example of a c control

chart generated using this

program is given to the

right. In this example, the

number of returned goods

to a distributor is being

tracked. “c” is the

number of returned goods

each day. For the first

date, there were 20 goods

returned.

The values of c are plotted


( c ), the upper control

limit (UCL), and the

lower control limit (LCL)

are calculated using the

equations below. k is the

number of subgroups used in the calculations (k = 15 in the above chart). The average is plotted as a

solid green line and the control limits are plotted as red dashed lines. The values for the average and

control limits, along with the subgroup size, are printed on the chart or in the chart title depending on the

option selected.

k

cc

c3cUCL c3cLCL

These control limits are valid only if c > 3. The program will automatically use the Poisson Distribution

to determine the control limits if the average is less than 3.

c Control Chart

Avg=23.13

UCL=37.56

LCL=8.7

0

5

10

15

20

25

30

35

40

45

50

2/1/

2003

2/2/

2003

2/3/

2003

2/4/

2003

2/5/

2003

2/6/

2003

2/7/

2003

2/8/

2003

2/9/

2003

2/10

/200

3

2/11

/200

3

2/12

/200

3

2/13

/200

3

2/14

/200

3

2/15

/200

3

Subgroup Number

Nu

mb

er

of

De

fec

ts


Data Entry

The c control chart is used to monitor the variation in the number of defects in

a constant subgroup size. Require data include the subgroup number and the

number of defects. An example of a c chart data is given to the right. To

make a c chart, select the subgroup numbers in the worksheet (shaded area).

Then select the attribute control chart option (Att) from the SPC toolbar and

select the c control chart option.

c Chart Dialog Box

After selecting the c Chart option in the dialog box, you

will get a two page dialog box shown here. Only items

not described in the p chart dialog box are explained

below. See page 14 for the p chart dialog box.


Range containing c values: The range

containing the number of defects. The default

value is the column next to the subgroup

identifiers unless you selected multiple ranges

using the control key.




Day Number of

Returned Goods (c)

2/1/2003 20 2/2/2003 24 2/3/2003 14 2/4/2003 32 2/5/2003 28 2/6/2003 16 2/7/2003 19 2/8/2003 32 2/9/2003 27

2/10/2003 25 2/11/2003 24 2/12/2003 12 2/13/2003 17 2/14/2003 44 2/15/2003 13


u Control Charts

A u control chart is used to monitor the variation in the number of defects. A defect occurs when

something does not meet a preset specification (operational definition). A u control chart is used with

counting type attributes data (e.g., 0, 1, 2, 3). These are whole numbers. You can not have 1/2 defect. In

addition, to use a u control chart, two other conditions must be true:

• The opportunity for defects to occur must be large.

• The actual number that occur must be small.

With a u control chart, we are often looking at an area of opportunity for defects to occur. A u control is

similar to a c control chart, except that the area of opportunity for defects to occur is not constant.

An example of a u chart generated by this program is given below. In this example, radiators are being

checked for leaks. Each day, the number of radiators assembled is counted. This is the area of

opportunity for leaks to occur (n). The number of leaks found when two portions of the radiator were

assembled for the first time is also counted (c). For example, on the first day, 39 radiators were hooked

up. There were 14 leaks detected. In this case, u = c/n = 14/39 = .359

The u values are plotted


( u ), the upper control

limit (UCL) and the lower

control limit (LCL) are

calculated using the

equations below. There is

no LCL in this example

(it is negative). The

average is plotted as a

green solid line and the

control limits are plotted

as red dashed lines. The

values for the average and

control limits (based on

the average subgroup size,

n ) are printed on the

chart or in the chart title

depending on the option

selected.

n

cu

n

u3uUCL

n

u3uLCL

The control limits will be based on the actual subgroup size for each point. If the subgroup size varies,

the control limits will also vary (as shown in the above example). These control limits are valid only if

c > 3. The program will automatically use the Poisson Distribution to determine the control limits if the

average is less than 3.

u Control Chart

Avg=0.15

UCL=0.32

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

2-Jun 3-Jun 4-Jun 5-Jun 6-Jun 7-Jun 8-Jun 9-Jun 10-Jun 11-Jun 12-Jun 13-Jun 14-Jun 15-Jun 16-Jun

Subgroup Number

De

fects

pe

r In

sp

ecti

on

Un

it


Data Entry

A u control chart is used to monitor the number of defects in

a changing subgroup size. The required data to be entered

into the spreadsheet are the subgroup numbers, the subgroup

size, and the number of defects. To make a u chart, select the

subgroup numbers (shaded area), select the attribute control

chart option (Att) from the SPC toolbar, and then select the u

control chart option.

u Chart Dialog Box

Once you select the u chart option on the dialog box, you

will get the two page dialog box shown here. Only items not

described in the p or c chart dialog box are explained below.

See page 14 for the p chart dialog box.


Range containing n values: The range containing the

subgroup size. The default value is the column next

to the subgroup identifiers unless you selected

multiple ranges using the control key.

Range containing c values: The range containing the

number of defects. The default value is the column

next to n unless you selected multiple ranges using

the control key.

Inspection Unit: Enter the size of the inspection unit.

For example, you can have an inspection unit of 1

radiator, 10 radiators, 20 radiators, etc. This choice

impacts the scaling of the chart.




Date Number of Radiatiors

Assembled (n)

Number of Leaks (c)

2-Jun 39 14 3-Jun 45 4 4-Jun 46 5 5-Jun 48 13 6-Jun 40 6 7-Jun 58 2 8-Jun 50 4 9-Jun 50 11

10-Jun 50 8 11-Jun 50 10 12-Jun 32 3 13-Jun 50 11 14-Jun 33 1 15-Jun 50 3 16-Jun 50 6


Variable Control Charts

You access these control charts by selecting the variable control chart

option (Var) on the SPC toolbar. You will then get the dialog box

shown to the right. Select the option you want. The options are:

X -R control charts

X -s control charts

Moving average and moving range control charts

X-MR (individuals) control charts

Table X-MR (for making multiple individuals charts at once)

Run charts

There is also an option to make subgroups from data in a single

column. This option is described latter in this section.

The dialog boxes for all the options are very similar. The X -R control chart is used to show how the

program works.

X -R Control Chart

An X -R control chart is used to examine the variation in variables data. Variables data are

“measurements” (e.g., height, weight, time, dollars, density). This control chart is used when you have

lots of data and a method of rationally subgrouping the data. An example of an X -R chart is given on the

next page. In this example, sales per day are monitored. A subgroup is made up of the results for one

week. The subgroup size (n) in this case is 5.

The X -R chart is really two charts. The top chart is the X chart. This chart looks at the variation in

subgroup averages. The subgroup average is the average of the individual results in the subgroup. The

bottom chart is the range chart. The subgroup range is the largest result minus the smallest result in the

subgroup.

The values of X and the moving range are plotted over time. The average and control limits for both

charts are calculated using the equations below. The average is plotted as a green solid line and the

control limits are plotted as red dashed lines on both charts. For the equations below, k is the number of

subgroups. A2, D3, and D4, d2 are constants used in the calculations for charts. See the e-zines on our

website for more information about these constants.

X Chart Equations:

X

X

k UCL X A 2R LCL X A2R

Range Chart Equations:

R

R

k UCL D4R LCL D3R

ˆ R

d2

The values for the average and control limits are printed on the respective charts.


Xbar Chart

Avg=30.2

UCL=33.8

LCL=26.6

26

27

28

29

30

31

32

33

34

35

2/1/2003 2/8/2003 2/15/2003 2/22/2003 3/1/2003 3/8/2003 3/15/2003 3/22/2003 3/29/2003 4/5/2003 4/12/2003 4/19/2003 4/26/2003 5/3/2003 5/10/2003

Subgroup Number

Su

bg

rou

p A

vera

ge

R Chart

Avg=6.3

UCL=13.3

0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

2/1/2003 2/8/2003 2/15/2003 2/22/2003 3/1/2003 3/8/2003 3/15/2003 3/22/2003 3/29/2003 4/5/2003 4/12/2003 4/19/2003 4/26/2003 5/3/2003 5/10/2003

Subgroup Number

Su

bg

rou

p R

an

ge


Data Entry

The data input for this type of chart is shown

to the right. The subgroup identifiers (week of

in this example) are in the first column. In

this example, data is collected once a day

every weekday. The results for one week are

used to form the subgroup (n = 5). Select the

subgroup identifiers and data (shaded). Then

select the variable control chart option (Var)

from the SPC toolbar. You will get the dialog

box above. Select the X -R chart option and

select OK.

X -R Chart Dialog Box

You will get the three-page dialog box shown

to the right. Each page is discussed below.

Selecting OK at the bottom of each page will

run the program. Selecting Cancel will cancel

the program. The Next Page and Previous

page buttons can be used to switch between the

three pages.


Range Containing the Subgroup

Identifiers: This is the range that

contains the subgroup numbers. The

default value is the first column in the

range you selected prior to selecting

the variable control chart option in the

toolbar.

Range Containing the Data: This is

the range containing the data. The default range is the range you selected prior to selecting the

variable control chart option excluding the first column in the range.

Subgroup Size: This is the subgroup size. The default value is the number of columns minus one

in the range you selected prior to selecting the variable control chart option in the toolbar. THIS

VALUE DETERMINES WHAT DATA IS INCLUDED.



Xbar Chart Title and Labels

o Title: This is the title that goes on the control chart. The default title is Xbar Chart.

o Y-Axis Label: This is the vertical axis label. The default label is Subgroup Average.

o X-Axis Label: This is the horizontal axis label. The default label is Subgroup Number.

R Chart Title and Labels

o Title: This is the title that goes on the range chart. The default title is R Chart.

o Y-Axis Label: This is the vertical axis label. The default label is Subgroup Range.

Chart Options

Week of Monday Tuesday Wednesday Thursday Friday 2/1/2003 33.0 33.1 26.4 28.3 28.9 2/8/2003 30.0 30.1 29.2 31.5 28.4

2/15/2003 31.8 29.4 28.0 26.9 32.2 2/22/2003 28.5 34.0 33.6 29.7 33.4 3/1/2003 27.2 27.6 30.6 30.1 27.4 3/8/2003 30.5 30.5 28.1 37.7 28.7

3/15/2003 35.4 31.3 27.8 31.3 33.0 3/22/2003 33.6 33.3 26.4 32.4 34.1 3/29/2003 35.8 34.1 30.1 30.3 26.1 4/5/2003 30.4 32.6 32.5 25.2 32.1

4/12/2003 26.9 32.4 29.0 26.8 29.3 4/19/2003 28.0 28.2 25.5 31.1 34.4 4/26/2003 29.1 31.6 29.0 33.1 30.9 5/3/2003 26.4 30.8 34.0 27.0 31.7

5/10/2003 27.4 26.0 28.2 27.9 27.3


o Xbar Chart Only: only the X chart is constructed. This is the default option

o Xbar and R Charts – Different Worksheets: Both the X and range charts are constructed

but on different worksheets.

o Xbar and R Charts – Same Worksheet: Both the X and range charts are constructed on

the same worksheet.


program selects one or the other depending on the range selected prior to selecting the variable

control chart option (Var) on the SPC toolbar.

Options Page

Tests for Control: There are three

options for interpreting the charts for

control: points beyond the limits, the

rules of seven (seven in a row above or

below the average or trending up or

down) and the zone tests (zones A, B, C,

stratification, mixtures). The zone tests

are not applied to range or standard

deviation charts. If an out of control

situation is detected, the points on the

chart will be in red.

Target for Average: This is the target

value for the variable. It is not required.

Print Average/Limits: There are two

options.

o On Avg. and Limits: This option

prints the average and control

limits values on the lines. This is the default option.

o In Chart Title: This option prints the values in the control chart title.

Allow values below 0?: Sometimes, it is not possible for the variable to have values below 0. If

that is the case, select “No” for this option. The default value is “Yes.”

Generate New/Update Existing Capability Chart? Select Yes to do a process capability analysis.

The default option is No. See the Process Capability section in this manual for more information.



Rounding to Use in Titles: This is the rounding to use for the average and control limits printed in

the title. If the first cell in the range has been formatted, this format is used. If not, the value

entered here is used for rounding. The program will estimate the rounding in the data.


Control Limit Option

Automatic Update of Limits?: This

determines if the control limits are

automatically updated when you add

additional data to the chart. Select

“Yes” if you want the control limits to

automatically update; No if you don’t

want the limits to automatically

update. The default value is Yes.

The rest of this dialog box is used only if you

want to change the default way the program

calculates the control limits. The program uses

the equations given in this manual for the +/-

three sigma limits. There are two other

options you have:

Base control limits on +/- “x” sigma:

You can select the sigma limits you want to use in the control chart. The default value is 3. DO

NOT CHANGE ANYTHING IF YOU WANT THE PROGRAM TO USE THE STANDARD

CONTROL LIMIT EQUATIONS. You can also add two additional lines to the charts (above

and below the average). Any values entered for these additional lines are ignored if the 3 sigma

limits are being used. If you use any other value than 3 sigma for the control limits, the zone tests

for out of control points will not be applied since it is no longer valid.

Enter your own limits: You may also enter your own values for the X chart control limits. An

additional two lines can also be added to the chart. The values entered here must be between the

average and the UCL entered. The program will add them automatically to the area between the

average and the LCL.


X -s Control Charts

A X -s chart is very similar to the X -R chart. Instead of using the subgroup range, the X -s chart uses

the subgroup standard deviation to determine process variability. It is usually used when your subgroup

is greater than or equal to 10.

For the equations below, k is the number of subgroups. A3, B3, and B4, C4 are constants used in the

calculations for the charts.

X Chart Equations:

X

X

k UCL X A3s LCL X A3s

s Chart Equation:

s

s

k UCL B4s LCL B3s

ˆ s

c4

The data entry requirements and the dialog boxes for these two charts are the same as the X -R control

chart. Please refer to the instructions above for the X -R control charts (page 22). The X -s charts for the

same data as used above for the X -R charts are shown below.

Xbar Chart

Avg=30.2

UCL=33.9

LCL=26.5

26

27

28

29

30

31

32

33

34

35

2/1/2003 2/8/2003 2/15/2003 2/22/2003 3/1/2003 3/8/2003 3/15/2003 3/22/2003 3/29/2003 4/5/2003 4/12/2003 4/19/2003 4/26/2003 5/3/2003 5/10/2003

Subgroup Number

Su

bg

rou

p A

vera

ge

s Chart

Avg=2.6

UCL=5.4

0.0

1.0

2.0

3.0

4.0

5.0

6.0

2/1/2003 2/8/2003 2/15/2003 2/22/2003 3/1/2003 3/8/2003 3/15/2003 3/22/2003 3/29/2003 4/5/2003 4/12/2003 4/19/2003 4/26/2003 5/3/2003 5/10/2003

Subgroup Number

Su

bg

rou

p S

tan

dard

Devia

tio

n


X-MR (Individuals) Control Charts

An individuals control chart (with a moving range of two) is used to examine the variation in variables

data. Variables data are “measurements” (e.g., height, weight, time, dollars, density). This chart is used

when you have limited data (for example, one data point per day or per week). It is also useful when data

are difficult to obtain. To use this chart, the individual measurements should be normally distributed, i.e.,

a histogram of the individual measurements is bell-shaped.

An example of an individuals control chart is given below. In this example, the dollar value of accounts

receivable past due 90 days is being monitored. An individuals control chart is really two charts. The top

chart is the X chart where the individual result (accounts receivable past due 90 days for a week) is

plotted. For example, the first point corresponds to $110,000 in past due receivables for the first week

(2/6). The second point corresponds to $104,000 for the second week (2/13).

Individuals Chart

Avg=103.6

UCL=124.3

LCL=82.9

79

84

89

94

99

104

109

114

119

124

129

2/5/

2003

2/12

/200

3

2/19

/200

3

2/26

/200

3

3/5/

2003

3/12

/200

3

3/19

/200

3

3/26

/200

3

4/2/

2003

4/9/

2003

4/16

/200

3

4/23

/200

3

4/30

/200

3

5/7/

2003

5/14

/200

3

5/21

/200

3

5/28

/200

3

6/4/

2003

6/11

/200

3

6/18

/200

3

Sample Number

Resu

lt

Moving Range Chart

Avg=7.8

UCL=25.5

0

5

10

15

20

25

30

2/5/

2003

2/12

/200

3

2/19

/200

3

2/26

/200

3

3/5/

2003

3/12

/200

3

3/19

/200

3

3/26

/200

3

4/2/

2003

4/9/

2003

4/16

/200

3

4/23

/200

3

4/30

/200

3

5/7/

2003

5/14

/200

3

5/21

/200

3

5/28

/200

3

6/4/

2003

6/11

/200

3

6/18

/200

3

Sample Number

Mo

vin

g R

an

ge

The bottom chart is the moving range chart. The moving range between consecutive points is plotted on

this chart. For example, the range between accounts receivable past due for 90 days between the week of

2/6 and 2/13 is $110,000 - $104,000 = $6,000. There is no range corresponding to the first data point on

the X chart.

The values of X and the moving range are plotted over time. The average and control limits for both

charts are calculated using the equations below. The average is plotted as a green solid line and the

control limits are plotted as red dashed lines on both charts. For the equations below, k is the number of

samples (individual X values).


X Chart Equations:

k

XX

R66.2XUCL R66.2XLCL

Moving Range Chart Equations:

1

k

RR R27.3UCL NoneLCL

128.1

Rˆ

The values for the average and control limits are printed on the respective charts

Data Entry

The only data required for an individuals control chart are the sample number

and the result as shown to the right. Select the sample numbers (shaded area).

Then select the variable control chart option (Var) on the SPC toolbar and

select the X-MR (Individuals) Chart option. Select OK and you will get the

two-page dialog box for the individuals control chart. This dialog box is the

same as for the X -R control chart. Please refer to the instructions above for

the X -R control charts.

Week of Accounts

Receivable 2/5/2003 110

2/12/2003 104 2/19/2003 98 2/26/2003 112 3/5/2003 113

3/12/2003 100 3/19/2003 89 3/26/2003 113 4/2/2003 109 4/9/2003 105

4/16/2003 108 4/23/2003 95 4/30/2003 101 5/7/2003 98

5/14/2003 100 5/21/2003 105 5/28/2003 103 6/4/2003 99

6/11/2003 112 6/18/2003 98


Moving Average/Moving Range (MA/MR) Charts

A moving average/moving range (MA/MR) chart is very similar to the X -R chart. The data entry

requirements are the same. Please refer to the X -R chart directions (page 22). The only major difference

in constructing a MA/MR chart is in how the subgroups are formed. The MA/MR chart reuses data. For

example, the data for the X-MR chart above could be regrouped into subgroup sizes of three using a

MA/MR chart. The first subgroup for the MA/MR chart is formed using the first three results (for the

weeks of 2/5, 2/12 and 2/19. The second subgroup for the MA/MR chart uses the weeks of 2/12 and 2/19

and then adds in the week of 2/26. This continues for each of the remaining samples. You use a MA/MR

chart when you have infrequent data that is not normally distributed. For MA/MR Chart

For the X-MR Chart

Subgroup Number

1 2 3

Week of Accounts

Receivable 1 110 104 98 2/5/2003 110 2 104 98 112 2/12/2003 104 3 98 112 113 2/19/2003 98 4 112 113 100 2/26/2003 112 5 113 100 89 3/5/2003 113 6 100 89 113 3/12/2003 100 7 89 113 109 3/19/2003 89 8 113 109 105 3/26/2003 113 9 109 105 108 4/2/2003 109

10 105 108 95 4/9/2003 105 11 108 95 101 4/16/2003 108 12 95 101 98 4/23/2003 95 13 101 98 100 4/30/2003 101 14 98 100 105 5/7/2003 98 15 100 105 103 5/14/2003 100 16 105 103 99 5/21/2003 105 17 103 99 112 5/28/2003 103 18 99 112 98 6/4/2003 99

6/11/2003 112 6/18/2003 98

The charts below are the output from the MA/MR chart for this data.

Moving Averge Chart

Avg=103.4

UCL=115.8

LCL=91

89

94

99

104

109

114

119

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Subgroup Number

Mo

vin

g A

vera

ge

Moving Range Chart

Avg=12.1

UCL=31.2

0

5

10

15

20

25

30

35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Subgroup Number

Mo

vin

g R

an

ge


Table X-MR (Individuals) Chart

This option is used to generate multiple individual control charts at the same time. You can either

generate the charts one at a time (loops through the dialog box each time) or all at once using chart names

and labels that already are entered on the worksheet.

Generate Charts One at a Time

The data is entered as shown to the right. The sample numbers are in one

column. The results are in the adjacent columns. Select the sample numbers

(shaded area). Then select the variable control chart option (Var) on the SPC

toolbar and select the Table X-MR (Individuals) Chart option. Select OK and

you will get the two-page dialog box for the individuals control chart. Enter the

required information into the dialog box and select OK. The first individuals

chart will be constructed. The program then moves to the adjacent column and

shows the dialog box again. The program runs until it finds a blank cell for

sample 1.

Generate Charts All at Once

The data entry requirements are shown to the right. To use this option, the

unique name of the charts as well as the Y labels and X labels must be

entered into the worksheet as shown to the right. Select the sample numbers

(shaded area). Then select the variable control chart option (Var) on the SPC

toolbar and select the Table X-MR (Individuals) Chart option. Select OK

and you will get the two-page dialog box for the individuals control chart.

Select the second page. In the lower right hand side of the dialog box is

“Base Labels on Cell Locations and Run Automatically.” Select Yes. You

will then get the dialog box below.

Select the row containing the title (name of chart),

the Y labels, and the X labels. Then select OK.

This returns you to the individuals control chart

form. Select OK. This will generate all the charts

automatically.

Sample Result 1 Result 2

1 98.5 93.61

2 101.22 106.38

3 105.99 108.67

4 89.08 98.83

5 105.48 94.57

6 96.55 91.55

7 90.77 95.11

8 96.13 89.41

9 97.16 97.98

10 100.67 98.17

Name Chart 1 Chart 2

Y Label Y Y

X Label X X

Sample Result 1 Result 2

1 98.5 93.61

2 101.22 106.38

3 105.99 108.67

4 89.08 98.83

5 105.48 94.57

6 96.55 91.55

7 90.77 95.11

8 96.13 89.41

9 97.16 97.98

10 100.67 98.17


Run Charts

The data entry requirements and the dialog box for the run chart are essentially the same as the X-MR

chart. Please review to the information above on the X-MR chart (page 28).

Subgroup Maker: Make Subgroups from Column of Numbers

The program has the option to make subgroups from a single column of numbers. The data

must be in a single column as shown to the right.

Select the data you want to make into subgroups (the data only, not any sample

identifier).

Select the variable control chart option on the SPC toolbar

Select the “Make Subgroups from Single Column” option. You will get the dialog

box shown below.

The range listed is the range you selected

on the worksheet. You may change it here

if it is not correct.

Enter the subgroup size, not to exceed 25.

Select the Output Option you desire:

o First cell of output range on a

worksheet: enter the cell location

on the worksheet where you want

the subgroups formed.

o New worksheet: select this option

to put the subgroups on a new worksheet.

Select OK: The subgroups will be generated and

place based on your output option. You will then get

the Variables Chart dialog box to select the type of

chart you want to make and follow the instructions

for that chart.

Data

97.00

87.22

102.44

112.76

111.98

117.33

78.16

97.66

110.95

89.13

93.10

83.10

81.53

90.22

92.26

78.82

94.32

95.96

101.35

96.35

96.73

96.30

113.43

99.15

98.14

94.87

119.72

108.66

123.76

93.45


Process Capability

Process capability answers the question: Is the process capable of meeting specifications? Specifications

can be set by customers. Specifications could also be standards set by management for a process. For

example, the standard for days sales outstanding might be set by leadership to be less than 46 days. One

measure of process capability is the Cpk index. Another is Ppk. To determine the process capability, the

individual sample results should be normally distributed (the histogram is a bell shaped curve) and the

process should be in statistical control.

The value of Cpk is the minimum of two process capability indices. One process capability is Cpu, which

is the process capability based on the upper specification limit. The other is Cpl, which is the process

capability based on the lower specification limit. Algebraically, Cpk is defined as:

Cpk = Minimum (Cpu, Cpl)

'ˆ3

XUSLCpu

'ˆ3

LSLXCpl

where USL = upper specification limit and LSL = lower specification limit. Both Cpu and Cpl take into

account where the process is centered. The value of Cpk is the difference between the process average

( X ) and the nearest specification limit divided by three times the standard deviation ( '̂ ). This standard

deviation is the standard deviation estimated from a range or s chart. In determining Ppk, the standard

deviation is the actual standard deviation of the measurements.

Cpk values above 1.0 are desired. This means that essentially no product or service is being produced

above USL or below LSL. The figure below shows how the Cpk values are developed. If Cpk is less

than 1.0, this means that there is some product being produced out of specification.

X=

+3̂'X=

+2̂'X=

+1̂'X=

X=

-1̂'X=

-2̂'X=

-3̂'

LSL USL

X=

- LSL USL - X=

3̂' 3̂'

The process capability feature of this program includes Cpk and Ppk. The data used in the analysis can

either be data entered into a spreadsheet for this analysis alone or data that has been used for a control

chart previously.


An example of the process capability analysis performed by the program is shown below. The histogram

of the data is given along with a normal curve. The specification limits are added. The statistics are

given to the right. The percentages in parentheses give the % out of specification for that metric.

Capability Analysis

LSL=60 USL=80Nominal=70

0

5

10

15

20

25

30

35

57 62 67 72 77 82 87

Measurement

Fre

qu

en

cy

Statistics

Cp=1.34

Cpk= 0.59

Cpu= 0.59 (3.84%)

Cpl= 2.09 (0%)

Est. Sigma= 2.5

Pp=1.31

Ppk= 0.57

Ppu= 0.57 (4.36%)

Ppl= 2.04 (0%)

Sigma= 2.6

Average=75.6

Min=70

Max=84

Count=96

No. Out of Spec=5 (5.21%)

Kurtosis=0.62

Skewness=0.71

Sigma Level=1.77

DPMO=394117.9

The statistics include the following:

Cp: = (USL-LSL)/6 '̂ where '̂ is the estimated standard deviation from a range or s chart

Cpk: the minium of Cpu or Cpl

Cpu: the capability based on the USL = (USL- X )/3 '̂ where X is the overall average (the

number in parentheses is the theoretical % greater than the USL)

Cpl: the capability based on the LSL = ( X -LSL)/3 '̂ (the number in parentheses is the

theoretical % less than the LSL)

Est. Sigma = '̂

Pp: = (USL-LSL)/6s where s is the standard deviation of the measurements

Ppk: the minium of Ppu or Ppl

Ppu: the capability based on the USL = (USL- X )/3s where X is the overall average (the number

in parentheses is the theoretical % greater than the USL)

Ppl: the capability based on the LSL = ( X -LSL)/3s (the number in parentheses is the theoretical

% less than the LSL)

Sigma: = s

Average: = X

Count: = number of data points in the analysis

No. Out of Spec: = actual number out of specification (number in parentheses is the % out)


Kurtosis: a measure of the shape of the distribution. A positive value means that the distribution

has longer tails than a normal distribution; a negative value means that the distribution has shorter

tails. The normal distribution has kurtosis of 0.

Skewness: a measure of asymmetry. If skewness is 0, there is perfect symmetry (like the normal

distribution). A positive value means that the tail of the distribution is stretched on the side above

the mean. The negative values means it is stretch on the side below the mean.

Sigma Level: A statistical term that measures how much a process varies from perfection, based

on the number of defects per million units.

o One Sigma = 690,000 per million units

o Two Sigma = 308,000 per million units

o Three Sigma = 66,800 per million units

o Four Sigma = 6,210 per million units

o Five Sigma = 230 per million units

o Six Sigma = 3.4 per million units

DPMO: Defects per million opportunities

Data Entry

If you are just using data to determine process capability without using a

control chart, enter the data into the spreadsheet. An example is shown to

the right. Select the data to be used in the analysis and then select the

process capability option (Cpk) on the SPC toolbar. If you want to do a

process capability analysis for an existing chart, you do not have to select

anything on a worksheet prior to selecting the process capability option on

the SPC toolbar.

81 77 75 74 77 73 77 74 76 75 79 74 74 79 73 75 75 74 75 80 80 79 72 78 73 74 74 73 75 74 77 75 75 72 75 74 76 75 74 74 78 75 76 76 78 77 78 75 74 76 77 76 72 73 79 82 73 75 74 79 77 73 72 75 73 73 76 76 76 75 74 72 76 76 76 74 79 79 75 81 77 74 77 71 84 74 79 70 77 74 73 77 76 74 81 75


Process Capability Dialog Box

Once you select the process capability option,

you will get the two page dialog box shown.

Each page is discussed below. Selecting OK at

the bottom of the page will run the program.

Selecting Cancel will end the program. The

Switch Tabs button can be used to switch

between the two pages.

Input Ranges/Chart Name Page

Data or Existing chart?: Select the

option you want. “Data Only” is the

default option. When “Data Only” is

selected, the “Range Containing Values”

is enabled.

o Range Containing Values: This

is the range containing the values on which to do the process capability analysis. The

default value is the range selected on the spreadsheet before selecting the process

capability option.

If “Existing Chart” is selected, the “Select Existing Chart” list box is enabled and a list of

available charts is given in the list box. Select the chart you want to do the process capability

analysis on.


name of the sheet that contains the chart in your workbook. If you select the “Existing Chart”

option, the chart automatically will be the name of the existing chart worksheet with Cpk added.


program selects one or the other depending on the range selected prior to selecting the process

capability option on the SPC toolbar.

Specifications: Enter the upper specification limit (USL), the lower specification limit (LSL) and

the nominal, the target (if desired). Only one specification limit is required.

Add +/- 3 Sigma Limits: In addition to the specifications, you can add the +/- three sigma limits to

the chart. The default is No. If you select Yes, you can chose sigma to the estimated sigma from

the range chart or the calculated standard deviation of all the data.

Titles/Labels/Dates of Data Collection/Multiple Charts/Outliers Page

Capability Chart Title: This is the title

that goes on the chart. The default value

is “Capability Analysis.”

Y-Axis Label: This is the vertical axis

label. The default value is “Frequency.”

X-Axis Label: This is the horizontal axis

label. The default value is

“Measurement.”

Number of Decimal Places for Rounding:

This is the rounding to use for the values

in the titles on the chart.




More Than One Chart? Select “Yes” if you want to make multiple process capability charts by

looping through the dialog box. The program assumes that the next set of data for the process

capability analysis is adjacent to the current set. Use one row or one column of data if you are

selecting this option. “No” is the default value.

Remove Outliers? Select “Yes” if you want to remove outliers from the calculations. Enter the

number of standard deviations you want to remove outliers beyond (e.g., beyond +/- 6 sigma).

The default option is “No.”


Advanced Process Capability

This option is used to automatically generate multiple process capability analysis, to remove outliers,

adjust specification limits, and generate a summary process capability table.

Data Entry

The data entry requirements for this option are shown to

the right. There must be a row containing the unique

“Name of Chart.” This becomes the worksheet tab name.

In addition, there must be a row containing the LSL and/or

USL. The row containing the nominal value is optional.

Select the data in the column containing the data for the

first process capability chart (shaded). Then select the

advanced process capability (ACpk) from the SPC toolbar.

Advanced Process Capability Dialog Box

One you select the Advanced Process

Capability option from the toolbar, you get

the two page dialog box shown to the right.

Each page is discussed below. Selecting

OK at the bottom of the page will run the

program. Selecting Cancel will end the

program. The Switch Tabs button can be

used to switch between the two pages.

Input Page

Capability Results Table: If “Yes”

is selected, a table summarizing the

process capability for all charts will

be generated. An example is shown

at the end of this section.

Range Containing Values: This is the range of the data for the first process capability chart. The

default value is the selected area on the spreadsheet. The data must be in columns for this feature.

Select Row Containing Name: Select a cell in the row or the row itself that contains the unique

name of the chart. This name will be on the worksheet tab containing the process capability

chart.

Select Row Containing USL: Select a cell in the row or the row that contains the USL values.

Select Row Containing Nominal: Select a cell in the row or the row that contains the nominal

values.

Select Row Containing LSL: Select a cell in the row or the row that contains the LSL values.

Name of Chart Chart 1 Chart 2 Chart 3 Chart 4 Chart 5

LSL 70 68 75 60 65

Nominal 100 100 97.5 100 95

USL 130 132 120 140 125

90 92 96 109 111

92 112 101 102 99

109 82 117 103 90

89 91 91 116 117

92 109 105 83 95

97 107 112 76 105

112 85 97 108 75

95 99 101 115 94

95 108 105 108 94

102 99 85 112 101

108 97 89 99 107

101 93 111 83 106

116 89 104 105 102

110 118 102 98 99

81 87 117 114 107

94 101 97 116 95

91 103 106 100 99

108 97 106 108 87

95 87 84 116 105

100 96 113 114 112


Remove Outliers? Select “Yes” if you want to remove outliers from the calculations. Enter the

number of standard deviations you want to remove outliers beyond (e.g., beyond +/- 6 sigma).

The default option is “No.”

Reset Specifications Limits? Select “Yes” if you want the program to replace the existing

specification limits with new limits set at the value of +/- sigma you enter. This is useful if you

are trying to set specification limits, e.g., for prototype data.

Add +/- 3 Sigma Limits: In addition to the specifications, you can add the +/- three sigma limits to

the chart. The default is No. If you select Yes, you can chose sigma to the estimated sigma from

the range chart or the calculated standard deviation of all the data.

Labels and Date Page

Y-Axis Label: This is the vertical

axis label. The default value is

“Frequency.”

X-Axis Label: This is the horizontal

axis label. The default value is

“Measurement.”

Dates of Data Collection: Add the

starting date and ending dates of

data collection. These dates are

optional. If entered, they will

appear in a dialog box in the lower

left-hand corner of the chart.

Example of Process Capability Table Output

An example of the process capability table output is shown below. This is from the data in the example

workbook.

Name Cp Cpk Cpu Cpl Est. Sigma Pp Ppk Ppu Ppl Sigma Average Count Minimum Maximum Kurtosis Skewness LSL USL

Chart 1 1.07 1.03 1.11 1.03 9.33 1.09 1.05 1.13 1.05 9.17 98.85 20 81 116 -0.65 0.19 70 130

Chart 2 0.94 0.87 1.01 0.87 11.38 1.1 1.02 1.18 1.02 9.72 97.6 20 82 118 -0.53 0.37 68 132

Chart 3 0.67 0.54 0.54 0.8 11.24 0.77 0.62 0.62 0.92 9.73 101.95 20 84 117 -0.58 -0.28 75 120

Chart 4 1.29 1.16 1.16 1.43 10.31 1.13 1.01 1.01 1.25 11.78 104.25 20 76 116 0.74 -1.2 60 140

Chart 5 1.02 0.85 0.85 1.19 9.85 1.04 0.87 0.87 1.21 9.63 100 20 75 117 1.16 -0.68 65 125


Scatter Diagram

A scatter diagram is used to show the relationship between two kinds of data. It could be the relationship

between a cause and an effect, between one cause and another, or even between one cause and two others.

In the example file, there is data that relates steam usage in a plant to the atmospheric temperature. The

question being answered here is “Does the atmospheric temperature have an effect on steam usage in the

plant.” For 25 days, data were collected on steam usage and temperature. There are 25 sets of data point.

Each set of data points is charted on a scatter. The scatter diagram is shown below.

As can be seen in the

figure, there is a negative

correlation in the data.

This means that as

temperature decreases, the

steam usage in the plant

tends to increase. There

can be a positive

correlation, a negative

correlation, or no

correlation.

The program determines if

the relationship between

the two variables is

statistically significant at a

probability of 0.05. If

there is a significant

relationship, the resulting probability is given in the title (in this case, the round goes to p =0). The

equation is of the form:

y = b1x + b0

where y is the variable on the y axis, x is the variable on the x axis, b1 is the slope of the line and b0 is

where the line crosses the y axis. In the example above, the slope is -0.0871. This means that for each

unit increase in x (one degree of temperature in this case), the y value (steam usage in this case) decreases

by -0.0871.

The value of R2 in the chart is the % of variation in y that is explained by x. In this example, 75% of the

variation in steam usage is explained by the variation in temperature.

Scatter Diagram (Significant, p = 0)

y = -0.0871x + 14.098

R2 = 0.7525

6.3

7.3

8.3

9.3

10.3

11.3

12.3

13.3

28 38 48 58 68 78 88

Temperature (X)

Ste

am

Us

ag

e (

Y)


Data Entry

Enter the X values and the Y Value into a spreadsheet as shown to

the right. Select the X values and the Y values. If the X value

and Y values are adjacent, you can just select the X values. Then

select the scatter diagram option (SD) from the SPC toolbar. The

sample number entries can be used for point labels (see below).

Scatter Diagram Dialog Box

Once you select the scatter diagram option

from the SPC toolbar, you will see the two

page dialog box shown to the right. Each page

is discussed below. Selecting OK at the

bottom of the page will run the program.

Selecting Cancel will end the program. The

Switch Tabs button can be used to switch

between the two pages.

Input Data/Titles Page

X Values Range: This is the range

containing the X values. The default

value is the range you selected prior to

selecting the scatter diagram option in

the toolbar.

Y Values Range: This is the range containing the Y values. The default value is the range next to

the X values or the second range selected on the worksheet.



Scatter Diagram Title: This is the title that goes on the scatter diagram. The default value is

“Scatter Diagram.”

Y-Axis Label: This is the vertical axis label. The default value is the cell contents above the X

values range.

X-Axis Label: This is the horizontal axis label. The default value is the cell contents above the Y

values range.

Trend/Regression Options: Select the regression you want. The options are:

o Linear

o Logarithmic

o Polynomial (activates order or period option)

o Power

o Exponential



Sample

No.

Temperature

(X)

Steam

Usage (Y)

1 35.3 10.98

2 29.7 11.13

3 30.8 12.51

4 58.8 8.4

5 61.4 9.27

6 71.3 8.73

7 74.4 6.36

8 76.7 8.5

9 70.7 7.82

10 57.5 9.14

11 46.4 8.24

12 28.9 12.19

13 28.1 11.88

14 39.1 9.57

15 46.8 10.94



program selects one or the other depending on the range selected prior to selecting the scatter

diagram option on the SPC toolbar.

Rounding/Forecasting/Point Labels

Rounding to Use in Titles: This the

rounding to use for the probablity

printed in the title. The default value

is 4.

Options: There are three options to

consider:

o Intercept: check this box if

you want to force the y-

intercept through 0 or another

value

o Display equation on chart:

select this option to display

the equation on the chart.

o Display R-Square: select this

option to display R2 on the

chart.

Forecasting: You can forecast forward or backward by changing the 0 values in the appropriate

boxes.

Add Labels to Point: Select “Yes” if you want to add labels to the points.

o Point Range Label: This is the range containing the point label ranges and should be

equal to the number of points on the chart.

o Label Position: This determines where to put the labels (above, below, left, center, right)

relative to the point.


Updating Charts

All charts can be updated easily after new data has been entered into the spreadsheet. There is no need to

select anything on the worksheet. The program automatically checks to see what new data has been

entered. Once you have entered the new data, select the

update option from the SPC toolbar. You will get the

dialog shown to the right.

The dialog box lists all the available charts for updating

in the workbook. You may select multiple charts at

once. You can also check the “Update All Charts”

option to update all the charts in the workbook.

Caution on Updating Charts:

If you change the name on the worksheet tab containing your chart, you will not be able to update the

chart because the program can’t find the new name. There is a manual method that allows you to change

the name on the worksheet tab. See the section on “Frequently Asked Questions” in the manual.

Changing Chart Options

You make changes to the chart using the options button on the SPC

toolbar. When you select this button, you get the dialog box to the

right. Select the chart whose options you want to change. The

dialog box for that chart then is shown. You can change anything

except the name of the chart.

Note on chart title, y-axis label and x-axis label: To make

permanent changes to the chart title or the labels on either axis, you

must go through the options shown here. Making changes directly on the chart will not permanently

change the title and labels. If you make the changes on the chart and then update the chart, the program

will use the stored values.

Caution on Updating Charts:

If you change the name on the worksheet tab containing your chart, you will not be able to update the

chart because the program can’t find the new name. There is a manual method that allows you to change

the name on the worksheet tab. See the section on “Frequently Asked Questions” in the manual.


Single Point Actions

This option is used for action on a single point on a chart including:

Splitting control limits at a point

Removing the split at the point

Starting the chart at the point

Undo starting the chart at the point

Removing a point from the calculations

Adding a point back to the calculations

Adding or replacing a comment to the point

Deleting the existing comment for the point

You must select a point on the chart prior to selecting the single point

option from the menu. To select a single point:

Select the series by pointing the mouse at the series and

clicking

Select the point you want using the mouse.

Below is an example of a chart with one point selected (this is the c chart data from the example

workbook).

After selecting the point, select the single point action option (SP) from the SPC toolbar. The dialog box

above will appear. Select the option you want and then select OK. The average and limits are

recalculated based on your option and the chart is re-made. See the notes on the following page for more

information.


Notes on Single Point Actions

Some of the actions on this option will produce minor changes in your workbook. These changes along

with other issues on single point actions are described below.

Control limits can be split multiple times within a chart.

o This will produce a change to your worksheet. The cell containing the point will be in

italics. This will occur in the first column or row of data.

Starting the chart at a new point will produce a change on your worksheet. The cell containing

that point will be in bold. This will occur for in the first column or row of data.

If a point is removed from the calculations, it is still plotted. Its appearance will change - it will

just be outlined as shown to the right.

o This will also produce a change to your worksheet. The cell containing the point will be

shaded a light tan. The shading will occur in

the first column or row of data. The shading

used is shown to the right.

Comments added to or deleted from the points must

be added through this option. Adding or deleting

them manually to the chart will not store the result.

o This will also produce a change to your

worksheet. The comments will be added to

the cell containing that point as an Excel

comment.

c Control Chart

Avg=21.64

UCL=35.6

LCL=7.69

0

5

10

15

20

25

30

35

40

45

50

2/1/2003 2/2/2003 2/3/2003 2/4/2003 2/5/2003 2/6/2003 2/7/2003 2/8/2003 2/9/2003 2/10/2003 2/11/2003 2/12/2003 2/13/2003 2/14/2003 2/15/2003

Subgroup Number

Nu

mb

er

of

Defe

cts

c Control Chart

Avg=23.13

UCL=37.56

LCL=8.7

0

5

10

15

20

25

30

35

40

45

50

2/1/2003 2/2/2003 2/3/2003 2/4/2003 2/5/2003 2/6/2003 2/7/2003 2/8/2003 2/9/2003 2/10/2003 2/11/2003 2/12/2003 2/13/2003 2/14/2003 2/15/2003

Subgroup Number

Nu

mb

er

of

De

fec

ts

Point included in

calculations

Point not included in

calculations


All Points Action

When you select the All Points Action from the SPC toolbar, you will

see the dialog box to the right. With this option, you can:

remove or add back to the calculations all points beyond the

control limits.

Select the range to base control limits on.

A chart must be selected before this option can be used. If you have

the two charts on the same worksheet (e.g., X -R chart), you must

select the X chart. You cannot delete points by selecting the range

chart.

Once you have selected the chart, select the all points option (All) on the SPC toolbar. You will get the

dialog box to the right. Select the option you want.

If you select the option “Select Range to Base Control Limits On,” you will

see a box like the one to the right. This box will contain the subgroup

identifiers from your chart. Select the subgroup you want to include in the

calculation of averages and control limits. Selecting this option will set the

option to automatic update the averages and limits to No.

Note: When removing all points beyond the limits, the average and limits are

recalculated. It is possible that additional points will now be beyond the

limits. You may have to run this several times to remove all points beyond

the limits.


Cause and Effect Diagram

This program contains a cause and effect (fishbone) diagram. To use this feature, select the cause and

effect option (CE) on the SPC toolbar. A blank cause and effect diagram will be inserted into your

workbook.

Date:

Machines Methods Materials

Product out of

specifications

Measurement People

Cause and Effect Diagram DiagramPrepared by:

Environment

You can change the main category headings (measurement, people, etc). You can enter the problem or

goal at the head of the fishbone (see above). You will see a button that says “Add Item.” Select this

button and you will get the dialog box below.

Enter an item to add to the cause and effect diagram. For example, you might enter “Not calibrated.”

Then select OK. This item is then shown on the chart and you may move it to any location you want.

Date:

Machines Methods Materials

Product out of

specifications

Measurement People

Cause and Effect Diagram DiagramPrepared by:

Environment

Not calibrated


Measurement Systems Analysis

This Measurement Systems Analysis used by the program is based on the following two sources:

1. Measurement Systems Analysis, Third Edition, AIAG, May 2003 (www.aiag.org)

2. Evaluating the Measurement System by Donald Wheeler and Richard Lyday, SPC Press,

Knoxville, TN, 1989 (www.spcpress.com)

Both are excellent references for developing a better understanding of the measurement system. The

program has the following components:

1. Average and Range Method to generate the classical Gage R&R report with the following chart

options:

a. Averages charts – stacked and unstacked

b. Range charts – stacked and unstacked

c. Run chart by part

d. Scatter plots

e. Whiskers charts

f. Error charts

g. Normalized histograms

h. X-Y Plots

i. Range charts for each operator

j. Operator bias chart

k. Operator consistency chart

2. ANOVA Method that includes:

a. ANOVA table

b. Residuals plot

c. ANOVA Gage R&R report

3. Range Method for Gage R&R

4. Bias Method – Independent Sample Method

5. Bias Method – Control Chart Method (checks Stability also)

6. Linearity Method

7. Attribute Gage R&R that includes:

a. Effectiveness table (attribute Gage R&R report)

b. Crosstabulations

c. Kappa values

These are explained in detail on the following pages.

http://www.spcpress.com/


Average and Range Method

Set-up Data Entry

The example below uses the data from page 101 of Measurement System Analysis, Third Edition.

Suppose you have completed a Gage R&R with 3 appraisers, 10 parts, and 3 trials. The first step in using

the Gage R&R program is to setup the data entry page. Select the MSA icon on the SPC toolbar. You

will get the form shown below.

Select the first option to setup the data entry sheet based on the number of operators, parts, and trials and

then select OK. You will see the form below which has been filled in with the numbers for this example.

You must enter all the information. The number of decimal places in the measurement is VERY

IMPORTANT. It controls how the data is rounded and shown. Entering zero when you have two or

three decimal places in the data may lead to inaccurate results. After entering all the information, select

OK. This will generate the data entry sheet shown below. The number of trials and parts can range from

2 to 20; the number of appraisers can range from 1 to 25 (with one appraiser you will only get a Gage

R&R report based on the average and range method).

Enter the operator names in the upper right hand corner. The names will automatically appear in the first

column. NOTE: The program uses Microsoft Excel’s naming function to run. You cannot have

spaces or certain characters (e.g. /). Instead of using John Smith, use John_Smith. Enter the rest of

the information for Date, Gage Name, Gage Number, Gage Type, Product, Characteristic, Upper

Specification Limit, Lower Specification Limit, and Performed By. None of this information is required

to run the program with the possible exception of the specification limits. These are required if you are

basing the acceptability of the measurement system on the tolerances. You then enter the data from the

appraisers for each trial and each part. A completed data entry screen is shown on the next page.


Blank Data Entry Form: Gage R&R Study 2

Date: Operator 1: Enter Operator 1 Name Here

Gage Name: Operator 2: Enter Operator 2 Name Here

Gage Number: Operator 3: Enter Operator 3 Name Here

Gage Type:

Product:

Characteristic:

Upper Specification Limit:

Lower Specification Limit:

Performed By:

Operator Trial/Part 1 2 3 4 5 6 7 8 9 10

Enter Operator 1 Name Here 1









Completed Data Entry Form: Gage R&R Study 2

Date: 10/31/2005 Operator 1: Hal

Gage Name: Thickness Gage Operator 2: Beth

Gage Number: T-101 Operator 3: Loa

Gage Type: Thickness

Product: Widget

Characteristic: Thickness

Upper Specification Limit: 3

Lower Specification Limit: -3

Performed By: Bill

Operator Trial/Part 1 2 3 4 5 6 7 8 9 10

Hal 1 0.29 -0.56 1.34 0.47 -0.8 0.02 0.59 -0.31 2.26 -1.36

Hal 2 0.41 -0.68 1.17 0.5 -0.92 -0.11 0.75 -0.2 1.99 -1.25

Hal 3 0.64 -0.58 1.27 0.64 -0.84 -0.21 0.66 -0.17 2.01 -1.31

Beth 1 0.08 -0.47 1.19 0.01 -0.56 -0.2 0.47 -0.63 1.8 -1.68

Beth 2 0.25 -1.22 0.94 1.03 -1.2 0.22 0.55 0.08 2.12 -1.62

Beth 3 0.07 -0.68 1.34 0.2 -1.28 0.06 0.83 -0.34 2.19 -1.5

Loa 1 0.04 -1.38 0.88 0.14 -1.46 -0.29 0.02 -0.46 1.77 -1.49

Loa 2 -0.11 -1.13 1.09 0.2 -1.07 -0.67 0.01 -0.56 1.45 -1.77

Loa 3 -0.15 -0.96 0.67 0.11 -1.45 -0.49 0.21 -0.49 1.87 -2.16

Generating the Results

You are now ready to run the program to generate the results. To run the program, select the icon on the

SPC toolbar. You will get the form shown above when you have selected the “Run the analysis (have

entered data into the data entry sheet)” under the Average/Range Method.

You have two things to decide at this point. First, on the

left-hand side of the form is what to base acceptability of the

measurement on. You have the following three options:

1. Tolerances: Use this option if your parts have very

little variation or not representative of the total

variation in your production process.

2. Total Variation Based on Parts: Use this option if

your parts are representative of the total variation in

your process

3. Process Standard Deviation: Use this if your parts

are not representative of the total variation in your

process and if you have a good estimate of the

process standard deviation (e.g., from a control chart

kept on the process).


On the right-hand side, you have three options for additional charts that can be generated along with the

Gage R&R report. The options are:

1. All Charts: this will generate all the charts associated with the study (see the first page of this

instructional manual for the list or the figure below)

2. No Charts: only the Gage R&R report will be generated

3. Select Charts: only the charts you select will be generated

If you select “Select Charts”, you will get the dialog box below. You select the charts you want to

include in the output and then select OK. This returns you to the form above.

Once you have selected your options, select OK and the program will generate the Gage R&R report as

well as the charts you have selected (if any).

When finished, the program will display the Gage R&R report. The one generated from this data (using

the Total Variation Based on Parts Option) is shown below. The report contains all the information in a

classical Gage R&R study and bases the conclusion if the measurement system is acceptable based on one

of the three options selected below. Any charts that were selected are generated on separate worksheets

in the workbook. You can download a completed workbook with all the charts from our website

(www.spcforexcel.com). A summary of each chart is given below.

Stacked Averages Chart – the average of each appraiser on each part is plotted by appraiser

using the part number as the index. There is one line for each appraiser. This helps

determine how consistent the operators are. The overall average and control limits are also

plotted. If the parts represent the total (true) variation in the process, at least half of theses

points should be out of control. If this not the case, the measurement system does not have

the ability to distinguish between samples (poor resolution) or the parts do not reflect the total

variation in the process.

Unstacked Averages Chart – same as the stacked chart but the appraisers are plotted together,

not separately.

Stacked Range Chart – used to show the range of each operator’s trials on a part and includes

the average range and control limits. There is one line for each appraiser. The chart is used

to determine if the process is in control. If there are out of control points, the special causes

need to be found and eliminated. Care should be taken with interpreting the Gage R&R

results if there are special causes present. Special causes occur if there are points beyond the

control limits.

Unstacked Range Chart – same as the stacked range chart but the appraisers are plotted as

one line.

Run Chart by Part – plots the individual readings by part for all appraisers to help see if there

are any outliers and to see the variation in the individual parts.



Scatter Plot – plots the individual readings by part-by-appraiser to examine how consistent

the appraisers are, to look for part-appraiser interactions, and to look for outliers.

Whiskers Charts – plots the high, average, and low value by part for each appraiser to

examine how consistent the appraisers are, to look for part-appraiser interactions, and to look

for outliers (same items as for the scatter plot).

Error Charts – plots the error (observed value – average measurement of the part) by part-

appraiser to determine which operator may have bias and which operator has the most

variability.

Normalized Histograms – plots the normalized value (observed value – average measurement

of part) as a histogram to determine how the error is distributed by appraiser.

X-Y Plot – plots the average of the readings by each appraiser against the overall part

averages to examine consistency in linearity between appraisers.

Appraiser Charts – consists of three charts:

o Range charts for each appraiser to determine if each is in control

o Bias chart for all appraisers to determine if different appraisers display detectably

different average values for the parts.

o Consistency chart for all appraisers to determine if different appraisers display detectably

different standard deviations for the parts.

Gage Repeatability and Reproducibility Report

Gage Name: Thickness Gage Product: Widget Date: 10/31/05

Gage No. T-101 Characteristic: Thickness Performed by: Bill

Gage Type: Thickness USL: 3

LSL: -3

Rbar= 0.341667 XbarDiff= 0.444667 Rp= 3.511111

K1= 0.5908 K2= 0.5231 K3= 0.3146

Measurement Unit Analysis % Total Variation (TV) % Tolerance

Repeatability - Equipment Variation (EV)

EV = Rbar * K1 % EV = 100(EV/TV) % EV = 100(EV/(USL-LSL)/6)

= 0.20186 = 17.61% = 20.19%

Reproducibility - Appraiser Variation (AV)

AV= Sqrt((XbarDiff * K2)2 - (EV

2)/nr)) % AV = 100(AV/TV) % AV = 100(AV/(USL-LSL)/6)

= 0.22967 = 20.04% = 22.97%

Repeatability & Reproducibility (R & R)

R&R= sqrt(EV2 + AV

2) % R&R = 100(R&R/TV) % R&R = 100(R&R/(USL-LSL)/6)

= 0.30577 = 26.68% = 30.58%

Part Variation (PV)

PV= Rp * K3 % PV = 100(PV/TV) % PV = 100(PV/(USL-LSL)/6)

= 1.1046 = 96.38% = 110.46%

Total Variation (TV)

TV= sqrt(R&R2 + PV

2) ndc= 1.41(PV/R&R)

= 1.14614 = 5.093652

Conclusion

% R&R under 10% of Total Variation: Measurement system is acceptable

**** % R&R from 10% to 30% of Total Variation: Measurement system may be acceptable based the application

% R&R over 30% of Total Variation: Measurement system needs improvement


ANOVA Method

Set-up Data Entry

The set-up is the same as for the Average and Range Method. Please follow the instructions for set-up

data entry as well as data entry for the Average and Range Method.

Generating the Results

After the data has been entered into the worksheet, select the MSA icon on the SPC toolbar. You will get

the form below. Select the ANOVA Method option as shown in the form and then select OK.

The program generates two new worksheets. One (ANOVA Report) contains the Gage Repeatability and

Reproducibility ANOVA Method Report as shown below. The acceptability of the measurement system

is based on percent contribution, not on % Total Variation.

Gage Name: Thickness Gage Product: Widget Date: 10/31/05

Gage No. T-101 Characteristic: Thickness Performed by: Bill

Gage Type: Thickness USL: 3

LSL: -3

% TOTAL PERCENT

STD DEV. VARIATION CONTRIBUTION

Repeatability (EV) 0.199933 18.42% 3.39%

Reproducibility (AV) 0.226838 20.90% 4.37%

Appraiser by Part (INT) 0 0.00% 0.00%

GRR 0.302372 27.86% 7.76%

Part (PV) 1.042327 96.04% 92.24%

Number of distinct data categories (ndc)= 4

Total Variation (TV) = 1.0853

Gage Repeatability and Reproducibility ANOVA Method Report

% R&R under 10%: Measurement system is acceptable

The other worksheet (GRR ANOVA) contains the ANOVA table and the residuals plot as shown below.

The residuals chart plots the residual versus the average for each appraiser for each part. The residual is

the result minus that average. The points should be randomly scattered above and below zero.


ANOVA Results

Source df SS MS F Sig

Appraiser 2 3.16726222 1.58363111 34.44 0.0000

Part 9 88.3619344 9.81799272 213.517 0.0000

Appraiser by Part 18 0.35898222 0.01994346 0.434 0.9740

Equipment 60 2.75893333 0.04598222

Total 89 94.6471122

F values in italics are significant at alpha = 0.05 level

Source of

Variation

Estimate of

Variance Std. Dev.

% Total

Variation

%

Contribution

Equipment 0.03997328 0.19993318 EV = 1.02965588 18.42% 3.39%

Appraiser 0.05145526 0.22683752 AV = 1.16821324 20.90% 4.37%

Interaction 0 0 INT = 0 0.00% 0.00%

GRR 0.09142854 0.30237152 GRR = 1.55721334 27.86% 7.76%

Part 1.0864466 1.04232749 PV = 5.36798659 96.04% 92.24%

Total Variation 1.17787514 1.08529956 TV = 5.58929275 100.00%

ndc = 4.86051648 or 4

5.15 Std Dev.

Residuals versus Average Values

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

-2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5

Average Values

Resid

uals


Range Method for Gage R&R

This method provides a quick look at measurement variability and requires that you have an estimate of

the process standard deviation. This method typically uses two appraisers and five parts. Each appraiser

measures the part one time only. An example of the data input required is shown in the figure below and

is based on the data from page 98 of Measurement Systems Analysis, Third Edition. The data can be

anywhere on the worksheet.

Select the data as highlighted above. Do not select the column headings. To run the program, select the

icon on the SPC toolbar. You will see the first page of the form as usual. Select the second page tab

labeled “Range/Bias/Linearity/Attribute Gage R&R.”

Select the Range Gage R&R Method and select OK. The dialog below will appear.

The range that appears in the “Range containing part numbers and appraiser results’ is the range that is

selected on the worksheet. You can save some time by selecting the data before selecting the icon on the


toolbar. Enter the process standard deviation. For the example, this standard deviation is 0.0777. Then

select OK. A new worksheet will be added to the workbook with the following output.

Range Method for Gage R&R

Average Range (Rbar) = 0.07

Upper Control Limit = 0.22869

Lower Control Limits = None

In Control? Yes

GRR = Rbar/d2* = 0.058772

Process Standard Deviation = 0.0777

%GRR = 100(GRR/Process Standard Deviation = 75.6%

Conclusion: The measurement system needs improvement.

Range Chart

0

0.05

0.1

0.15

0.2

0.25

1 2 3 4 5

Part Number

Ran

ge

The results give the average range and control limits. The range chart is shown to check for special

causes of variation. The % GRR is calculated and the conclusion given based on the % GRR.


Bias – Independent Sample Method

This method determines if the measurement system is biased. It is done by using one

sample and determining its reference value. The one appraiser measures the sample ten

or more times. The data used in this example is from page 87 in Measurement Systems

Analysis, Third Edition. An example of how the data is entered in the spreadsheet is

given to the right. The data can be anywhere in the spreadsheet.

Select the data as highlighted to the right. Do not select the column heading or the trial

numbers – just the results of running the sample multiple times. To run the program,

select the icon on the SPC toolbar. You will see the first page of the form as usual.

Select the second page tab labeled “Range/Bias/Linearity/Attribute Gage R&R.” Select

the Bias (Independent Sample Method) option and select OK. You will then see the

dialog below.

The range that appears in the dialog box is the range selected on the

worksheet. You can change it here if it is not correct. Enter the

reference value (which is 6.00 in this example); enter the number of

decimal places (1), and alpha. The default value is 0.05 which gives

95% confidence limits.

The second page of the dialog box is output options. It is shown to

the left. These options are available for both bias reports. There are

two options for plotting the histogram – by midpoint or by class

width. By midpoint is the default value. The histogram can be

plotted on the same sheet as the analysis (recommended) or as a new

sheet in the workbook. There are two output options for the analysis

of bias study. You can place it in the worksheet where the data is or

as a new sheet (recommended). You do not have to go to this page if

you are satisfied with the default options.

Select OK and the results are generated. A new worksheet is added

with the results as shown on the next page. A histogram of the

results is included along with the numerical calculation results. The

numerical values include:

n = number of readings

Mean = average of the readings

Reference Value = reference value of the sample

Standard Deviation (s) = standard deviation of the readings

t statistic = the t value based on the degrees of freedom

df = degrees of freedom

t value (2 tailed) = t value from the t tables

Bias = average – reference value

Lower = lower confidence interval (based on alpha)

Upper = upper confidence interval (based on alpha)

A conclusion is also presented that states if you can assume the bias is zero.


n Mean

Reference

Value

Standard

Deviation

() t statistic df

t value (2

tailed) Bias Lower Upper

Measured Value 15 6.006667 6 0.21202 0.121781 14 2.144787 0.006667 5.889254 6.124079

There is no evidence that the average is significantly different than reference value. You may assume the bias is zero.

Alpha=0.05 Confidence Interval

Bias - Independent Sample Method

0

1

2

3

4

5

5.6 5.7 5.8 5.9 6.0 6.1 6.2 6.3 6.4

Measured Value

Fre

qu

en

cy


Bias – Control Chart Method

This method uses the control charts to determine if the measurement system is biased. It

can also be used to check the stability of the measurement system. You must have a

reference value for the sample. The control chart can be an individuals chart (X-MR),

X -R chart, or X -s chart. The example below uses the X-MR chart. An example of the

data entry requirements are shown to the right.

Select the data as highlighted to the right. Do not select the column heading but do select

the sample numbers. To run the program, select the icon on the SPC toolbar. You will

see the first page of the form as usual. Select the second page tab labeled

“Range/Bias/Linearity/Attribute Gage R&R.” Select the Bias (Control Chart Method)

option and select OK. You will then see the dialog below.

You have an option of selecting the type of chart

you are using. The program selects what it thinks you have based

on the selected area on the worksheet. You can change it if it is not

correct. The dialog box also shows the range containing the

subgroup identifiers (sample numbers) and the range containing the

measurements. The subgroup identifiers are always assumed to be

in the first column, followed by the data.

Enter the reference value, the number of decimal places in the data,

and alpha (default is 0.05). In this example, the reference value is

100.3.

The second page in the dialog box contains the Output Options.

These are the same as those given in the Bias – Independent Sample

Method above. Please refer to that section for instructions.

The third page of the dialog box contains the chart title and labels as shown below. You can use the

default ones or change them. You can also change the titles and labels after the results are generated.

Once you have entered the information into the dialog box, select OK. A new worksheet is added to the

workbook with the results as shown on the next page for this example.


The output contains the same information as given in the Bias – Independent Sample Method. It also

includes the control charts for the sample to help determine stability.

n Mean

Reference

Value

Standard

Deviation

() t statistic df

t value (2

tailed) Bias Lower Upper

Measured Value 0 101.045 100.3 0.566127 5.885146 19 2.093024 0.745 100.78 101.31

There is evidence that the average is significantly different than the reference value. The bias is not zero.

Alpha=0.05 Confidence Interval

Bias - Independent Sample Method

0

0.5

1

1.5

2

2.5

100.

1

100.

2

100.

3

100.

4

100.

5

100.

6

100.

7

100.

8

100.

9

101.

0

101.

1

101.

2

101.

3

101.

4

101.

5

101.

6

101.

7

101.

8

101.

9

102.

0

Measured Value

Fre

qu

en

cy

X Chart

Avg=101.05

UCL=102.98

UCL=99.11

97

98

99

100

101

102

103

104

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Sample Number

Sa

mp

le R

es

ult

Moving Range Chart

Avg=0.73

UCL=2.38

0

0.5

1

1.5

2

2.5

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Sample Number

Mo

vin

g R

an

ge


Linearity

Linearity is the difference of bias throughout the

measurement range. To determine linearity, the

samples you select must cover the expected operating

range of the measurement system. You should use at

least five samples that cover this range. One

appraiser should measure each of the parts at least ten

times. An example of the required data input is

shown to the right (page 95, Measurement Systems

Analysis, Third Edition). The data must be in

columns. The first row of data contains the part

number; the second row contains the reference values

for each part; and the remaining rows contain the

measurements.

Select the data as highlighted above. Do not select the first column. To

run the program, select the icon on the SPC toolbar. You will see the

first page of the form as usual. Select the second page tab labeled

“Range/Bias/Linearity/Attribute Gage R&R.” Select the Linearity

option and select OK. You will then see the dialog box to the left.

The range is the range selected on the worksheet before starting the

program. Enter the number of decimal places and alpha (default is

0.05). The “Rounding in Equations” determines how the best fit equation is displayed. The default value

is 3. Once you have entered the information, select OK and the program will generate a chart like the one

below.

Linearity

y = -0.132x + 0.737, R Sqr = 71.4%

-1

-0.5

0

0.5

1

1.5

2 3 4 5 6 7 8 9 10

Reference Values

Bia

s

Bias

Regression

Upper 95% CI

Lower 95% CI

Bias Average

Bias=0

The linearity is NOT acceptable. Ta=12.043, Tcritical=2.002

The bias is NOT the same for all reference values. Tb=10.158, Tcritical=2.002

In the upper left-hand side of the chart, the conclusion is given for linearity. In this example, there is a

prl0bem with linearity. The bias = 0 line (green on the chart) should be contained by the upper and lower

95% confidence intervals. The equation is the title is the best fit equation for the individual readings.

The R squared value gives the % of variation in the bias that is explained by the variation in reference

values.


Attribute Gage R&R

Set-up Data Entry

The attribute gage R&R component will generate the effectiveness table, crosstabulations and kappa

scores. The example below uses the data from page 127 of Measurement System Analysis, Third Edition.

To set-up the data entry, select the icon on the SPC toolbar and then:

Select the second page tab labeled “Range/Bias/Linearity/Attribute Gage R&R.”

Select the Attribute Gage R&R option.

Select the first option to setup the data entry sheet based on the number of operators, parts and

trials

Select OK.

You will then see the dialog box to the right. Enter the number of

appraisers, parts, and trials. Enter the pass value and fail value (e.g, 1

and 0, pass and fail, etc.). Then enter the name of the Gage R&R

study. The values in the dialog box to the right are for this example.

Then select OK. A new worksheet will be added containing the data

entry sheet as shown below for this example (first 13 parts shown).

Attribute Gage R&R Study

Date: Operator 1: Enter Operator 1 Name Here

Gage Name: Operator 2: Enter Operator 2 Name Here

Gage Number: Operator 3: Enter Operator 3 Name Here

Gage Type:

Product:

Characteristic:

Pass Value 1

Fail Value 0

Performed By:

Reference

Operator Trial/Part 1 2 3 4 5 6 7 8 9 10 11 12 13










Enter the operator names in the upper right hand corner. The names will automatically appear in the first

column. NOTE: The program uses Microsoft Excel’s naming function to run. You cannot have

spaces or certain characters (e.g. /). Instead of using John Smith, use John_Smith. Enter the rest of

the information for Date, Gage Name, Gage Number, Gage Type, Product, Characteristic, Pass Value,

Fail Value, and Performed By. None of this information is required to run the program except the Pass

Value and Fail Value. Then fill in the data as shown below.

Attribute Gage R&R Study

Date: Operator 1: A

Gage Name: Operator 2: B

Gage Number: Operator 3: C

Gage Type:

Product:

Characteristic:

Pass Value 1

Fail Value 0

Performed By:

Reference 1 1 0 0 0 1 1 1 0 1 1 0 1

Operator Trial/Part 1 2 3 4 5 6 7 8 9 10 11 12 13

A 1 1 1 0 0 0 1 1 1 0 1 1 0 1

A 2 1 1 0 0 0 1 1 1 0 1 1 0 1

A 3 1 1 0 0 0 0 1 1 0 1 1 0 1

B 1 1 1 0 0 0 1 1 1 0 1 1 0 1

B 2 1 1 0 0 0 1 1 1 0 1 1 0 1

B 3 1 1 0 0 0 0 1 1 0 1 1 0 1

C 1 1 1 0 0 0 1 1 1 0 1 1 0 1

C 2 1 1 0 0 0 0 0 1 0 1 1 1 1

C 3 1 1 0 0 0 0 1 1 0 1 1 0 1


To generate the results, select the icon on the SPC toolbar. You will see the first page of the form as

usual. Select the second page tab labeled “Range/Bias/Linearity/Attribute Gage R&R.” You will see the

form below. Select the Attribute Gage R&R option, followed by the Run the Attribute Gage R&R

analysis.

There are three options: crosstabulations, kappa values, and the effectiveness table. Select the options

you want and select OK. The program will generate the following results depending on options selected.

Crosstabulations

You will get a table like this one for each of the

combinations of appraisers. The kappa value is

given. If kappa is above 0.75 there is good

agreement between the appraisers. If it is less

than 0.40, there is poor agreement. The count

information is as follows:

A and B both rate as fail

A rates as fail and B rates as pass

A rates as pass and B rates as fail

A and B both rate as pass

These tables will help you determine how well

the appraisers agree with one another.

Kappa Values

The output for Kappa values is shown on the next page. It includes the kappa values for the appraisers as

well as the kappa value compared to the reference values (if there are any).

Effectiveness Table

The effectiveness table is shown on the next page. This table determines how effective each appraiser is.

Fail Pass Total Kappa

A Fail Count 44 6 50 0.86

Expected Count 15.7 34.3 50.0

Pass Count 3 97 100

Expected Count 31.3 68.7 100

Total Count 47 103 150

Expected Count 47.0 103.0

A * B Crosstabulation

B


Kappa Measures Output

Kappa Measures for the Appraisers

Kappa A B C

A - 0.86 0.78

B 0.86 - 0.79

C 0.78 0.79 -

There is good to excellent agreement since all kappa values are greater than 0.75

Kappa Values For Each Appraiser to Reference

A B C

Kappa 0.88 0.92 0.77

There is good to excellent agreement since all kappa values are greater than 0.75

Effectiveness Table

Attribute Gage R&R Effectiveness

Gage Name: Product: Date:

Gage No. Characteristic: Performed by:

Gage Type:

Source A B C A B C

Total Inspected 50 50 50 50 50 50

# Matched 42 45 40 42 45 40

False Negative (Appraiser biased toward rejection) 0 0 0

False Positive (Apprasier biased toward acceptance) 0 0 0

Mixed (Appraiser accepts and rejects the same part) 8 5 10

95% UCI 92.8% 96.7% 90.0% 92.8% 96.7% 90.0%

Calculated Score 84.0% 90.0% 80.0% 84.0% 90.0% 80.0%

95% LCI 70.9% 78.2% 66.3% 70.9% 78.2% 66.3%

Total Inspected 50 50

# in Agreement 39 39

95% UCI 88.5% 88.5%

Calculated Score 78.0% 78.0%

95% LCI 64.0% 64.0%

Notes:

(1) Appraiser agrees with him/herself on all trials

(2) Appraiser agrees on all trails with the known reference

(3) All appraisers agreed within and between themselves

(4) All appraisers agreed with and between themselves and agreed with the reference

(5) UCI and LCLI are the upper and lower confidence interval bounds respectively

System % Effectiveness Score3

System % Effectiveness Score vs Reference4

% Appraiser1

% Score vs Attribute2

If the calculated score for each appraiser falls within the confidence interval of the other appraisers, the

effectiveness of the appraisers is the same.


Transfer Charts to PowerPoint or Word

You can transfer one or more charts to PowerPoint by selecting “PP” on the SPC toolbar or to Word by

selecting “W.”

PowerPoint

To transfer to PowerPoint, a presentation must be opened. The program adds a slide to the presentation

and copies the chart over. You can transfer multiple charts at once by selecting multiple worksheets in

Excel before running this option.

Word

A Word document must be opened to transfer a chart. The chart is placed in the open document wherever

the cursor is. You can transfer multiple charts at once by selecting multiple worksheets in Excel before

running this option.


Regression

The program contains a multiple regression component under the miscellaneous button on the SPC

toolbar. This is used to determine which independent variables (the X’s) have a significant impact on the

dependent variable Y. An example of the data entry is shown below. In this example, we want to find

out if attendance (Y) at major league baseball games is impacted by the team batting average, the number

of home runs hit by the team, the team’s earned run average, winning percentage of total payroll. Data

for the 2005 baseball season is given below.

Team Average Home Runs

Earned Run Avg.

Winning % Total Payroll Attendance

Arizona Diamondbacks 0.256 191 4.84 0.475 $62,329,166 2,059,331

Atlanta Braves 0.265 184 3.98 0.556 $86,457,302 2,521,534

Baltimore Orioles 0.269 189 4.56 0.457 $73,914,333 2,624,804

Boston Red Sox 0.281 199 4.74 0.586 $123,505,125 2,813,354

Chicago Cubs 0.27 194 4.19 0.488 $87,032,933 3,100,262

Chicago White Sox 0.262 200 3.61 0.611 $75,178,000 2,342,834

Cincinnati Reds 0.261 222 5.15 0.451 $61,892,583 1,943,157

Cleveland Indians 0.271 207 3.61 0.574 $41,502,500 1,973,185

Colorado Rockies 0.267 150 5.13 0.414 $48,155,000 1,915,586

Detroit Tigers 0.272 168 4.51 0.438 $69,092,000 2,024,505

Florida Marlins 0.272 128 4.16 0.512 $60,408,834 1,823,388

Houston Astros 0.256 161 3.51 0.549 $76,779,000 2,762,472

Kansas City Royals 0.263 126 5.49 0.346 $36,881,000 1,371,181

Los Angeles Angels 0.27 147 3.68 0.586 $97,725,322 3,404,686

Los Angeles Dodgers 0.253 149 4.38 0.438 $83,039,000 3,603,680

Milwaukee Brewers 0.259 175 3.97 0.5 $39,934,833 2,211,023

Minnesota Twins 0.259 134 3.71 0.512 $56,186,000 2,013,453

New York Mets 0.258 175 3.76 0.512 $101,305,821 2,782,212

New York Yankees 0.276 229 4.52 0.586 $208,306,817 4,090,440

Oakland Athletics 0.262 155 3.69 0.543 $55,425,762 2,109,298

Philadelphia Phillies 0.27 167 4.21 0.543 $95,522,000 2,665,301

Pittsburgh Pirates 0.259 139 4.42 0.414 $38,133,000 1,794,237

San Diego Padres 0.257 130 4.13 0.506 $63,290,833 2,832,039

San Francisco Giants 0.261 128 4.33 0.463 $90,199,500 3,140,781

Seattle Mariners 0.256 130 4.49 0.426 $87,754,334 2,689,529

St. Louis Cardinals 0.27 170 3.49 0.617 $92,106,833 3,491,837

Tampa Bay Devil Rays 0.274 157 5.39 0.414 $29,679,067 1,124,189

Toronto Blue Jays 0.267 260 4.96 0.488 $55,849,000 2,486,925

Washington Nationals 0.265 136 4.06 0.494 $45,719,500 1,977,949

Texas Rangers 0.252 117 3.87 0.5 $48,581,500 2,692,123

To run the regression program, select the shaded area as shown. You MUST select the column labels

(data can be rows also). Then select the regression option under the MISC icon on the SPC toolbar. You

will get the dialog box shown below.


This two-page dialog box is used to enter the

information for the regression. The first page

contains the following:

Enter unique name for regression: This

is the name that is used on the

worksheet tab containing the results

and for updating regression results.

Enter range containing the Y values:

This is assumed to be the last

column/row in the selected range on the

worksheet.

Enter range containing X values: This

is assumed to be all except the Y values

above. You can select just one cell in

the data field and the program will

expand that range as the input range.

Data in: Select columns if the data is I columns; the default depends on the number of columns

and rows selected on the worksheet prior to selecting the SPC toolbar.

Set b0 = Zero: Select Yes is you want to force the best line through the origin. The default value

is No.

All Residual Data and Charts?: This option allows you to select all the options on the second

page of the dialog box (see below). The default value is No.

The second page of the dialog box is shown to

the right. Select the options you want for the

charts and data. These include:

Time sequence plot of residuals: the

difference between the actual and

predicted y over time.

Residuals versus the predicted Y

Residuals versus the X variables

Normal probability plot

Line fit plots for each X variables.

You can also print out the residuals and

standardized residuals on a worksheet.

The output from the regression program

(excluding the charts and data options shown on

the second page of the dialog box) is given

below.


Regression Output

ANOVA Table

df SS MS F Signif. F

Regression 5 9.79271E+12 1.95854E+12 14.54672883 1.33582E-06

Residual 24 3.23131E+12 1.34638E+11

Total 29 1.3024E+13

Coefficients

Coefficient Standard Error t Stat p Value 95% Lower 95% Upper

Intercept 8099237.433 2667080.801 3.03674243 0.005686025 2594653.242 13603821.62

AVG -18922620.41 12537980.63 -1.509223931 0.144295197 -44799740.41 6954499.592

HR 689.429779 2600.899626 0.265073582 0.793216208 -4678.56318 6057.422738

ERA -355703.5027 270091.4996 -1.316974074 0.200284332 -913144.9564 201737.9511

Winning % -863435.1508 2564965.184 -0.33662646 0.739324292 -6157263.068 4430392.766

Totalpayroll 0.016759643 0.00246773 6.791523367 5.03538E-07 0.0116665 0.021852787

Regression Statistics

Multiple R 0.867119393

R Square 0.751896042

Adjusted R Square 0.700207717

Standard Error 366930.4105

Observations 30

Durbin-Waston Statistic 1.655584043

Regression Summary for Attendance

Actual verus Predicted

790,000

1,290,000

1,790,000

2,290,000

2,790,000

3,290,000

3,790,000

4,290,000

4,790,000

790,000 1,290,000 1,790,000 2,290,000 2,790,000 3,290,000 3,790,000 4,290,000 4,790,000

Actual

Pre

dic

ted


Changing the Variables in the Regression

You have the option to change the variables included in the regression after you have run the initial

regression. There are often variables that are not significant and you may want to remove them from the

regression. You start this routine from the sheet with the regression summary (as shown above). Select

the Update icon on the SPC toolbar. You will get the dialog box below.

You have the following options:

Select variables from the listbox to

include

Remove any variable with a p value

>0.05

Remove any variable with a p value >

0.20

Remove the variable with the highest p

value

Remove the variable with a p value

greater than a value you enter

You also have the option here to change the

chart options. The default value is No. If you

select Yes, you get the dialog box to the right.

Select the option you want or check the “Select

All” box for all options.

You will not lose any previous regression data.

The new worksheets will be named the same as

the previous regression with a “1” after the

previous name.


Miscellaneous

Selecting the miscellaneous option (Misc) on

the SPC toolbar generates the dialog box

shown to the right. Select the option you

want. The various options are described

below.

Descriptive Statistics

This option displays certain statistics on a number of samples. For example,

suppose you have measured the moisture content of a powder over time and

have fifteen sample results. The data is entered into the spreadsheet as shown

to the right. Select the data only (shaded area), then select the miscellaneous

option (Misc) on the SPC toolbar and then the descriptive statistics option.

You will get the dialog below.

The input range is the range selected on the

spreadsheet. You can change this if it is

not correct. You have two output options:

on the current worksheet (select the cell

location) or a new worksheet. Select OK

and the program will generate the descriptive statistics. A check is

made to ensure that there is no data in the area selected for the output.

The output for the data in this example is shown below.

Descriptive Statistics

Mean 0.131333

Standard Error 0.006537

Median 0.13

Mode 0.14

Standard Deviation 0.025317

Sample Variance 0.000641

Kurtosis -0.0048

Skewness -0.29422

Range 0.09

Minimum 0.08

Maximum 0.17

Sum 1.97

Count 15

Observation % Moisture

1 0.08

2 0.13

3 0.12

4 0.17

5 0.1

6 0.16

7 0.14

8 0.14

9 0.13

10 0.14

11 0.10

12 0.17

13 0.14

14 0.12

15 0.13


Confidence Interval Around a Mean

This option constructs a confidence interval around a mean based on a number of observations (samples)

There are two cases: one with a known standard deviation and one without a known standard deviation.

The procedure is essentially the same for both. With the know standard deviation, you enter the value and

the program uses the normal distribution to set the confidence interval. When the standard deviation is

not known, the program calculates the standard deviation of the samples and uses the t distribution to set

the confidence interval.

Known Standard Deviation

As an example, consider a process that produces a powdered product whose moisture content is critical to

downstream applications. Too much moisture causes clumping problems in downstream equipment.

Control charts have been kept on moisture content for a long time, and the process is in statistical control.

The average has been estimated to be 0.15% moisture (from the center line on the X chart); the standard

deviation has been estimated to be 0.02 (from the R chart).

An engineer has made a process change that she believes will decrease the

average moisture content. After making the process change, fifteen samples

were collected from the process and measured for moisture content. The

results are given to the right. The question we want to answer is "Has the

process change significantly decreased the moisture content of the product?"

Select the shaded data as shown. Select the confidence interval around a mean

option after selecting the miscellaneous option (Misc) from the SPC toolbar.

You will get the dialog box below.

Enter data range (no headings): The default range is the range

selected on the spreadsheet. You can change this if necessary.

Alpha: The default value for alpha is 0.05. This represents the chance

of what we find out in this sample is not representative of the population. It is typically 0.05.

Enter the known standard deviation (if known): In

this example, it is known and is 0.02.

Output Options: You have two output options: on

the current worksheet (select the cell location) or a

new worksheet. Select OK and the program

generates the output shown below.


Mean 0.131333


Count 15

Degrees of Freedom 14

Alpha 0.05

t Value 1.959964

Upper Confidence Level 0.141455

Lower Confidence Level 0.121212

Since the range of the lower to upper confidence levels (0.121 to 0.141) does not include 0.15, you

conclude that the change has made a significant difference.

Observation % Moisture

1 0.08

2 0.13

3 0.12

4 0.17

5 0.1

6 0.16

7 0.14

8 0.14

9 0.13

10 0.14

11 0.10

12 0.17

13 0.14

14 0.12

15 0.13


Unknown Standard Deviation

The procedure for the unknown standard deviation is identical to the one for the known standard

deviation. When you get the dialog box, leave the box for the known standard deviation empty. For the

example data above, the resulting output is shown below.


Mean 0.131333


Count 15


Alpha 0.05

t Value 2.144787



The confidence interval for the average is between 0.117 and 0.145.


Confidence Interval Around a Variance

This option places a confidence interval around the variance (the square of the standard deviation) based

on a number of observations (samples). The chi-squared distribution is used.

For example, suppose you are interested in determining a 95% confidence interval

for the variance in bulk density of a powdered product. Ten observations are

pulled from the process and measured for bulk density in grams/cc. The results are

given to the right.

Select the confidence interval around a variance option after selecting the

miscellaneous option (Misc) from the SPC toolbar. You will get the dialog box

shown to the right below. This is the same used for the variance around a mean

(with the standard deviation box not used). See above for the description of each

item. Select the output option and OK to generate the output shown below.

The confidence level for the variance is 0.000165 to 0.001166.

Sample Number

Bulk Density

1 0.697

2 0.698

3 0.673

4 0.657

5 0.710

6 0.702

7 0.680

8 0.709

9 0.670

10 0.671

Confidence Interval Around a Variance

Mean 0.6867


Variance 0.00035

Count 10


Alpha 0.05




Confidence Interval for the Difference in Two Means

This option is used to determine the confidence interval for the difference in two means – that is, are two

processes operating at difference averages.

To compare the averages and variances of two processes, we take observations from each process.

Suppose we take n1 observations from process 1 and n

2 observations from process 2. We can then

calculate the sample statistics listed below.

Process 1 Process 2

Sample size n1 n2

Sample average Y–

1 Y

– 2

Sample standard deviation s1 s

2

Sample variance s1

2

s2

2

These sample statistics will be used to compare the variances and averages of the two processes. The

variances will be compared by taking the ratio of the two variances and using the F distribution to

determine if there is a significant difference between the two. The averages will be compared by

constructing a confidence interval around the difference in two averages. If the confidence interval

contains zero (i.e., the difference could be zero), we will conclude that there is no evidence that the two

processes are operating at different averages. If the confidence interval does not include zero, we will

conclude that there is evidence that the processes are operating at different averages.

For example, consider two batch reactors that make the "same" product.

Questions have arisen about whether the reactors really do make the same

product. One indication of the completeness of reaction is the residual

catalyst remaining after reaction. Ten observations were taken from each

reactor. The results are given to the right.

Select the cells in the shaded area for the two processes (do not have to be

adjacent on spreadsheet) and then select the confidence interval for the

difference in two means option after selecting the miscellaneous option

(Misc) from the SPC toolbar. You will get the dialog box shown below.

Enter range for variable 1: The default range is the first range

selected on the spreadsheet. You can change this if necessary.

Enter range for variable 2: The default range is the second range selected on the spreadsheet.

You can change this if necessary.

Alpha: The default value for alpha is 0.05. This

represents the chance of what we find out in this

sample is not representative of the population. It is

typically 0.05.

Enter the known standard deviation (if known): In

this example, it is known and is 0.02.

Reactor 1 Reactor

2

450 482

423 422

443 463

476 492

490 445

436 483

457 476

421 462

485 479

491 495


Output Options: You have two output options: on the current worksheet (select the cell location)

or a new worksheet. Select OK and the program generates the output shown below.

Difference Betwen Two Means Confidence Interval

Variable 1 Variable 2

Mean 457.2 469.9

Standard Deviation 26.93119 22.56078

Variance 725.2889 508.9889

Observations 10 10

Pooled Variance 617.1389

Variance Same? Yes

t Statistic 1.143134


Alpha 0.05

Critical t value 2.100922


Lower Confidence Level -36.0408

P(T<t) 0.26796

Conclusion No statistical difference in the means.

The program will provide the conclusion for you. In this case, there is no statistical difference in the

means. Note that the program also tells you if the variances in the two processes are the same. In this

case, the variances are the same also.


Confidence Interval for Multiple Processes

This option is used to determine if the averages and variation in multiple processes are the same or not.

You start with n samples from each process. You can have up to 25 samples per process.

For example, suppose you have four furnaces producing ethylene. You measure the % ethylene from

each furnace. You want to know if there a difference in the average results from each furnace. To

determine this you take seven samples from each furnace. The results are given below.

Furnace

1

Furnace

2

Furnace

3

Furnace

4

67 64 68 69

67 64 65 72

64 64 67 70

65 64 64 68

64 64.1 67 67

65 64 65 68

67 64 65 70

To run the program, select the data including the headings. Then select the Confidence Interval for

Multiple Processes. You will see the dialog box below.

Alpha: The default value for alpha is 0.05. This represents the chance of what we find out in this sample

is not representative of the population. It is typically 0.05. Select OK. The program will insert a new

worksheet with the output. The output from the data above is shown on the next page. The following are

calculated:

Mean: average for each process

Sigma: standard deviation of each process

Variance: variance of each process

Observations: number of observations per process

Average Sigma: average standard deviation

UCLs: Upper control limit for the s control chart

LCLs: Lower control limit for the s control chart

Pooled Variance: the estimated variance

Interval Alpha: the interval alpha (which is the alpha entered above divided by the number of

confidence intervals)

t value: the value of the t distribution


Furnace 1 Furnace 2

Furnace 3

Furnace 4

67 64 68 69

67 65 65 72

64 67 67 70

65 64 64 68

64 65 67 67

65 68 65 68

67 64 65 70

Mean 65.57143 65.28571 65.85714 69.14286

Sigma 1.397276 1.603567 1.46385 1.676163

Variance 1.952381 2.571429 2.142857 2.809524

Observations 7 7 7 7 Average Sigma 1.535214

UCLs 2.889273

LCLs 0.181155 Pooled Variance 2.369048

Interval Alpha 0.008

T Value 2.892495

Process Lower Confidence Upper Confidence

Furnace 1 -2.094 2.665432 No statistical difference in the means.

Furnace 2


Furnace 3

Furnace 1 -5.95115 -1.19171 Means are different

Furnace 4


Furnace 3


Furnace 4


Furnace 4

Each possible pair of averages is checked. The program tells you if the means are different (see italics

above).


Paired Sample Comparison

There are occasions when the same samples may be used in two different processes. For example, you

might be interested in comparing two test methods. To do this, you would mix a sample thoroughly and

then split it in two. Half of the sample would be run in one test, and the other half would be run in the

other test. The samples used are not independent. In this case, the method for comparing two processes

given above can not be used. The paired sample comparison method option must be used.

For example, suppose you are interested in comparing two test

methods for analyzing particle size in microns. One test method

involves the use of sieves. The other test method is a particle

analyzer that measures particle size in slurry. Ten samples are split

in half and run in each test method. The results are shown to the

right. The program is examining the difference between each

process for the same sample to see if the mean difference is

significantly different than 0.

Select the shaded areas (the data for the two tests; do not have to be

adjacent) and then select the paired sample comparison option after

selecting the miscellaneous option (Misc) from the SPC toolbar.

You will get the dialog box shown below. This is identical to the

dialog box for the confidence interval between two means. Please

see the above section for the description of the dialog box

entries.

Fill in the dialog box and select OK. The program will

generate the output below.

The program will provide the conclusion for you. In this

case, there is a difference between the two processes. The

confidence interval defined by the lower and upper

confidence limits do not contain zero.

Paired Sample Comparison

Mean Difference 5.2


Variance 27.73333

Observations 10


Alpha 0.05

t value 2.262157



Conclusion Means are different

Observation Number

Sieve Test

Particle Size Analyzer

1 245 242

2 196 190

3 205 208

4 226 215

5 213 203

6 234 235

7 229 220

8 192 193

9 233 224

10 204 195


Analysis of Means

Analysis of means is a graphical and statistical way of comparing k treatments means with the overall

mean. The method used in this program is described in the book Advanced Topics in Statistical Control

by Dr. Donald J. Wheeler (www.spcpress.com). The example is from the book.

The maximum number of treatments is 25. The maximum subgroup size for each treatment is also 25.

Suppose you are studying five different methods

(treatments) of applying a coating and are

measuring the weight of the coating. The data

entry requirement for this option is shown to the

right. The treatments are A – E. The weight for

each treatment is given in the column under the treatment letter.

To run the Analysis of Means, select the treatment headings and the data

(the shaded area above). Then select the Analysis of Means option after

selecting the miscellaneous option (Misc) from the SPC toolbar. The

dialog box to the right appears.

Select the range containing the treatments with the headings:

Select both the data and the treatment names.

Treatment Data in Columns/Row: The data can either be in

columns or rows.

Output Options: You have two output options: on the current

worksheet (select the cell location) or a new worksheet.

Title: The title of the averages chart. Default is “ANOM.”

Y-Axis Label: Default value is “Treatment Average”

X-Axis Label: Default value is “Treatment Number.”

Rounding to Use in Titles: This is the rounding that is used in the

chart titles for the averages and control limits.

Select OK and the output and charts on the next page are generated. The UDL and LDL are the upper

decision limit and the lower decision limits. Any points beyond these limits on the top chart represent

significant differences from the overall treatment average. The range chart is a classical range chart and

compares the variation within treatments to see if they are the same.

A B C D E

250 310 250 340 250

260 330 230 270 240

230 280 220 300 270

270 360 260 320 290

http://www.spcpress.com/


Treatment A B C D E

250 310 250 340 250

260 330 230 270 240

230 280 220 300 270

270 360 260 320 290

Average 252.5 320 240 307.5 262.5

Maximum 270 360 260 340 290

Minimum 230 280 220 270 240

Range 40 80 40 70 50

Treatment Variance 291.6666667 1133.333333 333.3333333 891.6666667 491.6666667

Treatment DF 3 3 3 3 3

Overall Average 276.5

Average Range 56

Est. V(X) N/A

d2* 2.096

Est SD(X) 26.71755725

Est SD(Xbar) 13.35877863


H 2.266666667

ANOM Upper Limit 306.7798982

ANOM Lower Limit 246.2201018

ANOM (Avg=277, UDL=307, LDL=246)

Avg

UDL

LDL

216

236

256

276

296

316

336

A B C D E

Treatment Number

Tre

atm

en

t A

vera

ge

Range Chart (Avg=56, UCL=128, LCL=None)

Avg

UCL

LCL0

20

40

60

80

100

120

140

A B C D E

Treatment Number

Tre

atm

en

t R

an

ge


Correlation Coefficients

The linear correlation coefficient, R, is a measure of the association between two

variables. The data to the right shows the lines picked per hour in a warehouse and

the overtime hours. The value of R will measure the degree of association

between these two variables. The maximum value for R is + 1. The minimum

value for R is - 1. In both these cases, all sample points fall on a straight line.

As R approaches +1 or -1, the stronger the correlation between x and y. The

square of this coefficient (R2) indicates the fraction of variation in y that is

associated with x.

To determine the correlation coefficient, select the data including the headings

(shaded area to the right) and then select the Correlation Coefficients options after

selecting the miscellaneous option (Misc) from the SPC toolbar. The dialog box

below appears.

Input Range: The input

range is the range selected

on the worksheet. If it is not

correct, you can change it.

Output Options: You have

two output options: on the current worksheet

(select the cell location) or a new worksheet.

Select OK and the output below is generated.

Lines Picked Hours Overtime

Lines Picked 1.00 0.92

Hours Overtime 0.92 1.00

The value of R is 0.92 indicating a strong relationship between lines picked and overtime hours.

In most cases, you will have more than two variables that you are examining. The program works the

same for multi-variables and will generate a correlation matrix as shown below (for an example using 6

variables, A – F.

A B C D E F

A 1.00 0.14 0.37 -0.03 -0.15 -0.04

B 0.14 1.00 0.30 0.08 -0.34 0.19

C 0.37 0.30 1.00 -0.17 -0.35 -0.05

D -0.03 0.08 -0.17 1.00 0.09 0.12

E -0.15 -0.34 -0.35 0.09 1.00 0.13

F -0.04 0.19 -0.05 0.12 0.13 1.00

Lines Picked

Hours Overtime

599 23.5

658 28.5

699 29

738 30.5

791 31.5

685 28

656 28

570 24.5

614 26

684 29.5

749 30

608 24.5

653 25.5

650 27

671 29

606 24

648 25.5

758 31

712 30

611 25.5

671 26

651 27


Failure Mode and Effect Analysis

A Failure Mode and Effects Analysis (FMEA) template is included in the program. To access the

template, select the FMEA option from the toolbar. The template below will be added to the workbook.

Date Prepared:

What is the

Potential Failure

Mode?

What are the Potential Effects of

the Failure Mode

Severi

ty

What are the Possible Causes of

the Failure?

Occu

rren

ce

How Do We Currently Prevent

Each Listed Cause of Failure from

Happening?

Dete

cti

on

S O D RPN

No. ResponsibilityO D

Improved

RPNWhat Needs to Be Done Status

What is the

Process?

Recommended Action Steps and Status Due

DateS

Failure Effects and Mode AnalysisPrepared by:

Add the text as you would in an Excel spreadsheet.


Box and Whisker Plots

A Box and Whisker plot is used to present a visual

representation of how data is spread out and how much

variation there is in the data. It focuses attention on the

median, the quartiles, and the minimum and maximum

values. For example, the data to the right shows how the

average monthly temperature for three cities varies. The Box

and Whisker plot can show this variation.

Select the data including the headings (the shaded area to the

right) and then select the Box and Whisker plots option after

selecting the miscellaneous option (Misc) from the SPC

toolbar. You will see the dialog box below.

Select the range: The default range is the range that

is selected on the worksheet.

Treatment Data in Columns/Rows: The data can be in

columns or rows.

Box and Whisker Title and Labels

o Title: The title that will appear on the chart

o Y and Axis Label: The label that will appear on the

y axis.

Selecting OK will generate the Box and Whisker plot shown on the

next page. The graph below shows the values for each part of the

plot. The maximum value is presented by the top line (whisker).

The top part of the box is the 75% quartile; the bottom part is the

25% quartile. The minimum is represented by the bottom line

(whisker).

Seattle, WA

San Antonio TX

New York, NY

Jan 46 61 38

Feb 51 66 40

Mar 54 74 50

Apr 58 80 61

May 64 85 72

June 70 92 80

July 74 95 85

Aug 74 95 84

Sept 69 89 76

Oct 60 82 65

Nov 52 72 54

Dec 46 64 43

Box and Whisker Chart

Ne

w Y

ork

, N

Y

San

An

ton

io T

X

Sea

ttle

, W

A

0

10

20

30

40

50

60

70

80

90

100

Su

bg

rou

p A

ve

rag

eMedian

75% quartile

Maximum

25% quartile

Minimum


Box and Whisker Chart

Ne

w Y

ork

, N

Y

Sa

n A

nto

nio

TX

Se

att

le, W

A

0

10

20

30

40

50

60

70

80

90

100

Su

bg

rou

p A

vera

ge


Sample Size Calculator

The sample size calculator is available under the MISC option on the

SPC toolbar. When you select this option, you will see the dialog

boxes to the right. The first page is for variables data. The second

page is for attributes data.

For variables data, enter the following:

Confidence interval

Standard deviation

Measurement error

Then select “Calculate Sample Size,” and the sample size needed

will appear in the bottom box.

For attributes data, enter the following:

Confidence interval

Estimated percentage of successes

Measurement error

Population size (if known)

Then select “Calculate Sample Size,” and the sample size needed

will appear in the bottom box.


Side by Side Histogram

You can develop side by side histograms to compare results. For example, you may have done a survey

that separates the male and female responses. A side by side histogram is shown below.

Histogram of Male/Female Responses

1

3

1

0

2

4

3 3 3

4

1

6

0

1

5

4

2 2

3

00

1

2

3

4

5

6

7

0-10 11-20 21-30 31-40 41-50 51-60 61-70 71-80 81-90 91-100

Rating

Fre

qu

en

cy

Female

Male

There are two options you have. The first option

is that you have already totaled the data by

groups. In this case, your data entry will look

the data to the right. The classes are in the first

column, followed by the data that you have

already totaled.

The second option is to let the program total the

data. In this case, you simply enter in the results

as shown to the left.

Class Female Male

0-10 1 1

11-20 3 6

21-30 1 0

31-40 0 1

41-50 2 5

51-60 4 4

61-70 3 2

71-80 3 2

81-90 3 3

91-100 4 0

Female Male

41 18

99 89

78 56

96 78

99 88

58 45

69 51

85 65

54 45

79 60

65 47

7 4

52 36

23 14

15 12

63 45

41 15

78 55

12 11

87 76

96 87

13 12

55 45

85 64


When you select the “Side by Side Histogram” option

under the Misc icon on the SPC toolbar, you will get the

form shown to the right. Each dialog box entry is

described below.

Enter range containing data including labels;

data must be in columns: Enter the worksheet

range containing the data including the column

headings (see examples above). The data must be

in columns. The default value is the range

selected on the worksheet before selecting the

Side by Side Histogram option.

Data: There are the two options explained above.

o Totaled by Class Already: The data has

already be summed by class.

o Not Totaled: The data has not been

totaled. The program will perform this

function based on the following two

values:

Lower Bound >: This value gives the starting point for the histogram. For

example, if you enter 0 here, the first class will involve values greater than 0 to

the class width

Class Width =: This is the width of one class.

Enter Histogram Title: Enter the title that will appear on the chart. The default title is

“Histogram.”

Enter Y-Axis (Vertical Label): Enter the label for the y axis. The default value is “Frequency.”

Enter X-Axis (Horizontal Label): Enter the label for the x axis. The default value is

“Measurement.”

Date of Data Collection: You can enter the dates of data collection. This is optional.


Plot Multiple Y Variables Against One X Variable

The program will plot multiple Y variables against one X variable. An example of this type of chart is

shown below.

Amps versus Time

-6

-4

-2

0

2

4

6

8

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 3 3.2 3.4 3.6 3.8 4 4.2

time (sec)

Y

Experimental Intensity (amps)

Theoretical Intensity (amps)

The data for this chart must be in columns with the X variable in

the first column. The Y variables must be in adjacent columns.

The data used to generate the above chart is shown to the right.

When you select the “Plot Multiple Y Variables Against One X

Variable” from the “Misc” icon on the SPC toolbar, you will get

the form shown below. Each entry is discussed below.

The first entry is for the range containing the data (including the

headings as shown to the right). The default value is the range

selected on the worksheet. Enter the title for the chart as well as

the axis labels. You can enter the dates of data collection

(optional).

time (sec)

Experimental Intensity (amps)

Theoretical Intensity (amps)

0 6 6

0.2 5.8 5.967131372

0.4 5.8 5.868885604

0.6 5.6 5.706339098

0.8 5.5 5.481272746

1 5 5.196152423

1.2 4.7 4.854101966

1.4 4.5 4.458868953

1.6 4 4.014783638

1.8 3.4 3.526711514

2 2.9 3

2.2 2.3 2.440419858

2.4 1.7 1.854101966

2.6 1 1.247470145

2.8 0.5 0.62717078

3 0 -9.64723E-16

3.2 -0.05 -0.62717078

3.4 -1.25 -1.247470145

3.6 -1.75 -1.854101966

3.8 -2.25 -2.440419858

4 -3 -3

4.2 -3.5 -3.526711514


Select Cells

You can save some time with some of the tools in this program by selecting the appropriate cells prior to

selecting the icon on the SPC toolbar. For example, if you are making a p control chart, it is beneficial to

select the subgroup identifiers before selecting the attribute control chart icon from the SPC toolbar. This

can be a little cumbersome if you have lots of subgroups. The Select Cells (SC) on the SPC toolbar helps

make this easier.

Selecting One Cell Only

If you have the cursor in one cell only (as shown by the

shaded cell in the table to the right) and you select the SC

icon on the SPC toolbar, the program will select the cells

directly below the shaded cell (including the shaded cell)

as shown the table below.

Date Number of

Telemarketing Calls (n)


2/1/2003 40 5

2/2/2003 63 10

2/3/2003 47 12

2/4/2003 52 7

2/5/2003 34 3

2/6/2003 59 21

Selecting More Than One Cell

If you select more than one cell (as shown to the right), all

adjacent cells (beginning with the first row and first

column) will be selected.

Date Number of



2/1/2003 40 5

2/2/2003 63 10

2/3/2003 47 12

2/4/2003 52 7

2/5/2003 34 3

2/6/2003 59 21

Date Number of



2/1/2003 40 5

2/2/2003 63 10

2/3/2003 47 12

2/4/2003 52 7

2/5/2003 34 3

2/6/2003 59 21

Date Number of



2/1/2003 40 5

2/2/2003 63 10

2/3/2003 47 12

2/4/2003 52 7

2/5/2003 34 3

2/6/2003 59 21


Frequently Asked Questions

What out of control tests does the program use?

The program uses the following tests to check if a control chart is out of control:

Points beyond the limits

Rule of seven tests

o Seven in a row above or below the average

o Seven in a row trending up or trending down

Zone Tests

o Zone A: 2 out of 3 consecutive points in zone A or beyond

o Zone B: 4 out of 5 consecutive points in zone B or beyond

o Zone C: 7 or more consecutive points in zone C or beyond

o Mixtures: 8 or more consecutive points fall outside of zone C on both sides the average

o Stratification: 15 or more consecutive points fall in zone C on either side of the average

Do all out of control tests apply to all the charts?

No. You will have the option to select which out of control charts you want to apply to your chart when

the dialog box for the control chart is shown. The out of control tests listed in the dialog box are the ones

that apply for your chart.

How do I know if the chart has any out of control points?

Out of control points are shown in red on the chart. The in-control points are in blue.

Can I remove the out of control points from the calculations?

Yes. You can remove the out of control point from the calculations by selecting the point and then

selecting “SP” from the SPC toolbar. This brings up a dialog box that allows you to remove the point

from calculations (or to add it back in). The point remains plotted on the graph but is outlined in color

only.

You can also remove all points beyond the control limits at once by selecting “All” from the SPC toolbar.

This brings up a dialog box that allows you to remove all the points beyond the limits at once (or add

them back in). This option does not remove the out of control points for the rule of seven tests or the

zone tests.

Can I change the name of the worksheet tab containing the chart?

If you do this, the program will not be able to find the chart if you want to update it or to make changes.

There is a manual method of doing this, but it is better if you just make a new chart.

How come I can’t see the name of one of my charts in the list of charts to be updated?

The most common cause of this problem is that the name on the worksheet tab containing the chart has

been changed. The program can no longer find the chart. The easiest solution is to make a new chart.


How can I change the title or the x and y labels on an existing chart?

To change the chart title or labels on an existing chart, you must go select the Options icon on the SPC

toolbar and change the title or labels there. If you make the changes directly on the chart, they will not be

saved if the chart is updated.

.

Any other questions?

Please feel free to e-mail us with your questions or suggestions for improvement ([email protected])

or visit our website (www.spcforexcel.com) for all the SPC information we have in our e-zines and

articles.

Enjoy the program!

mailto:[email protected]


Date post:	09-Mar-2018
Category:	Documents
Upload:	donhu
View:	216 times
Download:	0 times