What is Statistical Process Control (SPC)?

SPC is a set of tools used to monitor and gauge the performance of a particular process over time. Pr
process variability over time and judging whether actual process outcomes are varying in a normal or abn
used to assess how well process outcomes are conforming to desired specifications. When a process is
meet desired specifications, employees must take actions that will improve the process.

About The Darden SPC Calculator

The Darden SPC calculator has been designed to aid in the design, construction, and interpretation of
The calculator is accessible via the GMOE course web page- Class 9 SPCcalculatorwith Silko Exhibit 5.x
behind SPC are detailed below. For additional background, the reader is referred to the reading "Managin
Capability (UVA-OM-1103), which is in the GMOE course packet.

Constructing Control Charts

Process control charts are relatively simple to construct and easy to understand. They allow for the ide
over time is normal or abnormal. A process with normal variability is said to be in control; statistically, th
time. Thus, the operations manager can expect process outcomes to continue over time in a predictable
The control chart provides no information about how well the process is performing relative to desired spe
characterized as abnormal, the process mean and variance are thought to be under the influences of spe
process. Control charts enable the operations manger to judge when a process is performing abnormally
and how to deal with them.
The first step in constructing a process control chart is to decide what to measure. This requires sound
understanding of the different dimensions of quality. This measure might be the size of a part being produ
operation, the number of daily transactions for a bank teller, or the error rate in a hospital pharmacy, etc.
attribute data.
Once the metric to be monitored has been identified, data are collected over a specified time period. T
average of the observed data and one to display the variation in the data. Both charts are plotted as a tim
control limits, i.e., limits within which a stable process should remain.
Since measuring each and every product or event within the process can be time consuming and cost
periodically. The size of each sample (n) will typically be from three to five items but can be larger; multip
groups, each with three or more items) should be taken at set intervals over a time horizon (hourly, shifts,
displayed using two charts: the Xbar chart, which displays the mean (Xbar) of each sample and the R (ran
each sample (the range is simply the highest value minus the lowest value within a given sample group).
The SPC Calculator generates control charts for samples data.

Calculating Capability Ratios

Whereas Control Charts will help determine if a process is performing with variation that is stable over
process outcomes will meet specifications (customer needs). The SPC Calculator allows you to interpret
be careful not to confuse Control Charts with Capability Ratios.
process over time. Process control charts are useful for tracking
ng in a normal or abnormal manner. Process capability ratios are
. When a process is producing products or services that do not
ess.

and interpretation of process control charts and capability ratios.

rwith Silko Exhibit 5.xlsx. The general principles and assumptions
the reading "Managing Variability: Process Control and Process

They allow for the identification of whether a processs variation

ontrol; statistically, the processs mean and variance are stable over
time in a predictable manner with the same mean and variance.
elative to desired specification targets. When process variability is
the influences of specific factors new and unexpected to the
erforming abnormally and then investigate the nature of the factors

e. This requires sound managerial judgment and a thorough

e of a part being produced in a factory, the fill level for a packaging
spital pharmacy, etc. The process outcome may be variable data or

ecified time period. Two charts are constructed: one to display the
ts are plotted as a time series and compared with upper and lower

consuming and costly, samples of several items can also be taken

can be larger; multiple groups of samples (e.g., at least 20 sample
horizon (hourly, shifts, daily, weekly, etc). These data are then
sample and the R (range) chart, which displays the variation within
given sample group).

ion that is stable over time, capability ratios tells you how well the
lows you to interpret capability ratios for a process, but you should
SPC Calculator

General Instructions (specific instructions are at the top of each worksheet)

1) Proceed through each worksheet tab in order, i.e. (1) Input Data; then (2) Xbar Chart & R Chart, (3) Control Interpretation

2) Enter data in yellow-shaded cells where requested

3) In Worksheet (1) Input Data, you can enter data collected as samples groups over time

In Worksheet (1) Input Data, enter the number of standard deviations to be used as control limits (normally 3) and the sa

4) Worksheet (2) Xbar Chart & R Charts contains these charts for the data you have entered

5) Use Worksheet (3) Control Interpretation, for tips to interpret the control charts in Worksheet (2)

6) Use Worksheet (4) Capability Ratios, to determine measures of process capability relative to customer specifications

Notes

By default, the worksheets are protected in order to prevent inadvertently changing the embedded calculations.

To turn protection off for a given worksheet, go to ToolsProtectionUnprotect Sheet. There is no password.
orksheet)

Chart, (3) Control Interpretation and (4) Capability Ratios

ol limits (normally 3) and the sample size

e to customer specifications

dded calculations.

e is no password.
Enter your data here if you are using samples of data collected over time
- This template should be used for data collected in sample sizes of 3 to 9 items
- For sample sizes less than n = 9, fill in the left-most columns (n = 1, 2, etc.) first
- Do not enter zero or press the space bar for blanks

Sample # n=1 n=2 n=3 n=4 n=5 n=6 n=7 n=8 n=9
X
1 0.4772 0.4804 0.4779 0.4719 0.4777
2 0.4756 0.4787 0.4733 0.4742 0.4734
3 0.4756 0.4773 0.4722 0.4760 0.4745
4 0.4744 0.4780 0.4754 0.4774 0.4774
5 0.4802 0.4726 0.4748 0.4758 0.4744
6 0.4783 0.4807 0.4791 0.4762 0.4757
7 0.4747 0.4766 0.4753 0.4758 0.4767
8 0.4788 0.4750 0.4784 0.4769 0.4762
9 0.4757 0.4747 0.4741 0.4746 0.4747
10 0.4713 0.4730 0.4710 0.4705 0.4727
11 0.4716 0.4730 0.4752 0.4735 0.4751
12 0.4746 0.4727 0.4763 0.4734 0.4730
13 0.4749 0.4762 0.4778 0.4787 0.4771
14 0.4771 0.4767 0.4785 0.4772 0.4765
15 0.4771 0.4758 0.4769 0.4770 0.4771
16 0.4767 0.4769 0.4770 0.4794 0.4786
17
18
19
20
21
22
23
24
25
26
27
 28
29
30

Chart for Picking Constants D3 and

Calculations D4 based on Sample Size n

Variable Excel Name Value Interpretation n D3 D4

X X_2bar 0.4758 average of sample averages 2 0 3.267
X Sigma_X_bar 0.0010 standard deviation of sample averages 3 0 2.575
R R
R_bar 0.0039 average of sample ranges 4 0 2.282
s Sigma 0.0023 standard deviation of all process data 5 0 2.115
Number of standard
deviations for Control 3
Limits in X-bar chart Three is the default value 6 0 2.004
n sample size 5 7 0.076 1.924
 Sample Sample
Average Range
R
0.4770 0.0085
0.4750 0.0054
0.4751 0.0051
0.4765 0.0036
0.4756 0.0076
0.4780 0.0050
0.4758 0.0020
0.4771 0.0038
0.4748 0.0016
0.4717 0.0025
0.4737 0.0036
0.4740 0.0036
0.4769 0.0038
0.4772 0.0020
0.4768 0.0013
0.4777 0.0027
Xbar Chart and R Chart UCL Center LCL
Xbar chart 0.48 0.48 0.47
R chart 0.01 0.00 0.00

Xbar Chart
0.4800

0.4780

0.4760
X
b
a 0.4740 Xb
r ar

0.4720

0.4700

0.4680
0 2 4 6 8 10 12 14 16 18

Sample

R Chart

0.0090
 R Chart

0.0090

0.0080

0.0070

0.0060 R

Ce
0.0050 nte
R r

0.0040

0.0030

0.0020

0.0010

0.0000
0 2 4 6 8 10 12 14 16 18

Sample

See sheet "3) Control Interpretation" for explanation of dotted lines

Xb
ar
 R

Ce
nte
r

18
Control Interpretation:
Once you have constructed the control charts, you need to interpret each chart's pattern to determine whether the process is "in control" or "out of control."
Below are several common indicators of being out of control, i.e. something has changed in the process being studied. Each Chart is judged separately.

Divide the control chart into equal-size zones (A, B, and C) above and below the center line as shown below. These zones are equal to 1, 2 and 3 standard
deviations from the "center line" Then apply the following tests:
"Out of control" conditions occurs if any one of the following situations is found:
1 Any occurrence of a point outside of the either control limits: UCL or LCL
2 Two out of three points in Zone A on the same side of the center line
3 Four out of five points in Zone B or beyond on the same side of the center line
4 Nine consecutive points on the same side of the center line
5 Six consecutive points increasing or decreasing
6 Fourteen consecutive points alternating up and down
7 Fifteen consecutive points in either Zone C (above or below the center line)

A
B Data
C Center

C UCL
LCL
B
A

1 2 3 4 5 6 7
Observatio Data Center UCL LCL
0 0 3 -3
1 1.5 0 3 -3
2 3.5 0 3 -3
3 2 0 3 -3
4 -0.7 0 3 -3
5 -2.7 0 3 -3
6 -2.2 0 3 -3
7 1.3 0 3 -3
8 0.7 0 3 -3
9 1.2 0 3 -3
10 2.2 0 3 -3
11 1.5 0 3 -3
12 -2.3 0 3 -3
13 -0.3 0 3 -3
14 -1.2 0 3 -3
15 -1.5 0 3 -3
16 -1.8 0 3 -3
17 -2.7 0 3 -3
18 -1.5 0 3 -3
19 -0.4 0 3 -3
20 -0.7 0 3 -3
21 0.1 0 3 -3
22 0.4 0 3 -3
23 0.7 0 3 -3
24 1.4 0 3 -3
25 1.5 0 3 -3
26 1.8 0 3 -3
27 -2.312 0 3 -3
28 0.073 0 3 -3
29 -1.796 0 3 -3
30 1.303 0 3 -3
31 -2.203 0 3 -3
32 1.244 0 3 -3
33 -2.173 0 3 -3
34 1.566 0 3 -3
35 -1.047 0 3 -3
36 0.772 0 3 -3
37 -2.428 0 3 -3
38 2.385 0 3 -3
39 -0.517 0 3 -3
40 2.391 0 3 -3
41 0.266 0 3 -3
42 0.340 0 3 -3
43 0.791 0 3 -3
44 0.128 0 3 -3
45 0.596 0 3 -3
46 0.149 0 3 -3
47 -0.974 0 3 -3
48 0.866 0 3 -3
49 -0.335 0 3 -3
50 0.760 0 3 -3
51 -0.410 0 3 -3
52 0.268 0 3 -3
53 0.084 0 3 -3
54 -0.463 0 3 -3
55 0.672 0 3 -3
rpret each chart's pattern to determine whether the process is "in control" or "out of control."
e. something has changed in the process being studied. Each Chart is judged separately.

ve and below the center line as shown below. These zones are equal to 1, 2 and 3 standard

Data
Center
UCL
LCL
Capability Ratios
Capability Calculations: require 1) an estimate of the process mean and standard deviation and 2) the customer's specification limits

1) Determine the standard deviation of the process 2) Input the customer's specification limits:

You may use an estimate derived from control chart data, or enter another estimate: Upper Specification Limit (USL): 0.4800
Lower Specification Limit (LSL): 0.4750
If you wish to use process data from the control charts, type "yes" here: yes

Alternately, enter your estimate of the process standard deviation: 0.00 0.00 (Process Sigma)

The following estimated process standard deviation sigma will be used: 0.00

Interpreting Capability Ratios: C P and CPK Measures

USL LSL 0.37
Cp Capability ratios compare the customer's expectations (embodied in the specification limits) with the process' natural
6 variation performance. It is assumed that the process is "in control". (The natural process variation performance is
measured as plus and minus 3 standard deviations from the process mean). However, each ratio differs slightly. C P
LSL 0.12 measures the "ideal" process capability, i.e. the process capability if the process mean is centered on the mid-point of
C pl the specification limits. C PK takes into account the possibility that the process is actually not centered at the mid-point
3 of the specification, or can also be used for "one-sided" specification limits.

USL 0.62
For either CP or CPK, a value less than one indicates an excessive defect rate; a value of exactly one indicates that the
C pu
3 normal process variation of plus/minus 3 standard deviations is barely meeting the specifications zone; and a value of

C pk MinC pu , C pl
greater than one indicates that the process variation is smaller than the range of the specification limits (i.e. this is
0.12 good, and the larger the CP or CPK the better)

Thus CPK tells you whether the process is currently capable; C P tells you what the process could be capable of if it
were centered. Note that if CPK < CP, then the process is not centered.

0
0
0

Page 17
Silko Terminal Block Data (Using NORMDIST)

Below LSL

X 0.4750 LSL
Mean 0.4758 X_2 Bar
Standard Deviation 0.0023 Sigma

% of Defects below LSL 36.03%

Above USL

X 0.4800 USL
Mean 0.4758 X_2 Bar
Standard Deviation 0.0023 Sigma

% of Defects above USL 3.17%