SPC or Control Charts Control Chart
SPC or Control Charts Control Chart
1
Control Chart General Formula Process Control Charts
• UCL = Process Average + 3 Standard DeviaGons Control Charts show sample data ploYed on a graph with Center Line
(CL), Upper Control Limit (UCL), and Lower Control Limit (LCL).
• LCL = Process Average - 3 Standard DeviaGons
X
UCL
+ 3σ
Process Average
- 3σ
LCL
TIME
2.50±0.05
2.45 2.55
Questions:
- What does the X control chart look like?
µ-3σ µ+3σ
- How does control chart relate to the tolerances? Measured Variation
2
6 Sigma Capability Sources of VariaGons
• Six sigma capability assumes the process is capable • Common causes of varia4on
of producing output where the mean +/- 6 standard – Random causes that we cannot
idenGfy
deviaGons fall within the design specificaGons – Unavoidable
• The result: only 3.4 defects for every million – Cause slight differences in process
variables like diameter, weight,
produced service Gme, temperature, etc.
• Six sigma capability means smaller variaGon and • Assignable causes of
therefore higher quality varia4on
– Causes can be idenGfied and
eliminated
– Typical causes are poor employee
training, worn tool, machine
needing repair, etc.
3
Warning Limits Ac4on Limits
• (100-95.5) 4.5% of the (correct) values are • There is a probability of only (100-99.7) 0.3 %
outside the warning limits. that a (correct) measurement is outside the
• This is not very unlikely. acGon limits (3 out of 1000 measurements)
• Therefore this is only for warning, no • Therefore the process should be stopped
immediate acGon required. immediately and searched for errors
date date
06
06
06
06
06
06
06
06
06
06
06
06
06
06
06
06
06
06
06
06
06
06
06
06
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
13
14
16
20
13
14
16
20
22
27
22
27
12
15
19
26
12
15
19
26
23
23
21
21
4
Out-of-control Situa4on 3 Out-of-control Situa4on 4
• 7 successive values on one side of the • 7 successive increasing or decreasing values
central line Not so criGcal as 1 and 2 Not so criGcal as 1 and 2
concentration concentration
date date
06
06
06
06
06
06
06
06
06
06
06
06
06
06
06
06
06
06
06
06
06
06
06
06
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
13
14
16
20
13
14
16
20
22
27
22
27
12
15
19
26
12
15
19
26
23
23
21
21
Review: Types of Data Types of Control Charts
Variables Attributes
• Control chart for variables are used to monitor
þ Characteristics that can þ Defect-related characteristics
take any real value
characterisGcs that can be measured, e.g. length,
þ Classify products as either weight, height, speed, volume, diameter, Gme, etc.
þ May be in whole or in good or bad or count defects
fractional numbers þ Categorical or discrete
þ Continuous random random variables • Control charts for aAributes are used to monitor
variables þ Items are either good or bad, characterisGcs that have discrete values and can be
þ Measures dimensions such acceptable or unacceptable counted, e.g. % defecGve, number of flaws in a
as weight, speed, height, or þ Does not address degree of shirt, number of broken eggs in a box, etc.
strength
failure
þ Falls within an acceptable
range
5
Control Charts for Variables Control Charts for Variables
• Mean (x-bar) charts
• The salt content, thickness, or crispness of a bag of – Tracks the central
potato chips. tendency (the average
• The number of chocolate chips in a container of value observed) over
chocolate-chip ice cream. Gme or locaGon effects
• The diameter of a bearing, or the center of a gear. • Range (R) charts:
• The waiGng Gme at a fast-food restaurant or at an – Tracks the spread of the
airport check-in counter. distribuGon or dispersion
• The internal temperature of a “rare” steak when it over Gme (esGmates the
leaves a restaurant’s kitchen. observed variaGon)
31
UCL UCL
(x-chart detects (x-chart does not
x-chart shift in central x-chart detect the increase
tendency) in dispersion)
LCL LCL
UCL UCL
(R-chart does not (R-chart detects
R-chart detect change in R-chart increase in
mean) dispersion)
LCL LCL
6
X-chart Formula Mean chart Example 1
For x-Charts when we know σ
A quality control inspector at the Cocoa Fizz son drink company has taken
Upper control limit (UCL) = x + zσx samples with four observa4ons each of the volume of boYles filled. If the
standard devia4on of the boYling operaGon is .2 ounces, use the data
Lower control limit (LCL) = x - zσx below to develop control charts with limits of 3 standard deviaGons for the
16 oz. boYling operaGon.
where x = mean of the sample means or a target
value set for the process
Time 1 Time 2 Time 3
z = number of normal standard deviations Observation 1 15.8 16.1 16.0
(2 for 95.5% confidence, 3 for 99.73%) Observation 2 16.0 16.0 15.9
σ = population, system, or process deviation Observation 3 15.8 15.8 15.9
σx = standard deviation of the sample means
Observation 4 15.9 15.9 15.8
= σ/ n
n = sample size
n Center line (x-double bar):
7
Mean chart Example 2 Mean chart Example 2
The weight of boxes of Oat flakes within large Example: Hour 1
Sample Weight of
producGon lot are sampled each hour. To set Number Oat Flakes
control limits that include 99.73% of the sample 1 17
2 13
means, samples of nine boxes are randomly 3 16
selected and weighed. Standard deviaGon of the 4 18
n=9
system is 1. The nine boxes for hour 1 are as 5 17
6 16
follows: 7 15
17oz. 13oz. 16oz. 18oz. 17oz. 16oz. 15oz. 17oz. 16oz. 8 17
9 16
Mean 16.1
Summary: For Hours 1 through 12, the results are: Control Chart
Variation due
for sample of 9 Out of
boxes to assignable
Hour Mean Hour Mean Given: control causes
1 16.1 7 15.2 X = 16
σ = 1 ounce 17 = UCL
2 16.8 8 16.4
n= 9 Variation due to
3 15.5 9 16.3 16 = Mean
z = 3 (99.73% confidence) natural causes
4 16.5 10 14.8
5 16.5 11 14.2 15 = LCL
6 16.4 12 17.3 UCLx = x + zσx = 16 + 3(1/3) = 17 ozs
Variation due
The average mean = 16 ounces LCL = x - zσ = 16 - 3(1/3) = 15 ozs | | | | | | | | | | | |
to assignable
x x 1 2 3 4 5 6 7 8 9 10 11 12 Out of causes
Sample number control
AlternaGve Method for the X-bar Chart: Step 1: Calculate the Range of Each
R - bar and the A2 Factor Sample and Average Range
Factor for x-Chart Factors for R-Chart
Sample Size
Time 1 Time 2 Time 3
(n)
A2 D3 D4
Use this method when sigma 2 1.88 0.00 3.27 Observation 1 15.8 16.1 16.0
for the process distribu4on 3 1.02 0.00 2.57 Observation 2 16.0 16.0 15.9
is not known. 4 0.73 0.00 2.28
Observation 3 15.8 15.8 15.9
5 0.58 0.00 2.11
6 0.48 0.00 2.00 Observation 4 15.9 15.9 15.8
7 0.42 0.08 1.92
The table of coefficients for 8 0.37 0.14 1.86
Sample ranges 0.2 0.3 0.2
(R)
calcula4ng the control lines 9 0.34 0.18 1.82
10 0.31 0.22 1.78
are A2, D4, and D3 is
provided.
11 0.29 0.26 1.74
0.2 + 0.3 + 0.2
12
13
0.27
0.25
0.28
0.31
1.72
1.69
R = = .233
14 0.24 0.33 1.67 3
15 0.22 0.35 1.65
8
Average Range (R – bar) Formula R - bar Example 1
For x-Charts when we don t know σ
n Center line:
9
R - bar Example 2: Range Chart
Process average x = 12 ounces • Synonyms are R-chart or Precision chart.
Average range R = .25
Sample size n = 5 • Absolute difference between the highest and
lowest value of mulGple analyses
UCLx = x + A2R UCL = 12.144 • Repeatability Precision check
= 12 + (.58)(.25)
= 12 + .144 • Range Chart has only upper limits
= 12.144 ounces Mean = 12
concentration
= 11.857 ounces
sample-# 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Data:
(n)
A2 D3 D4
UCLR = D4 R 2
3
1.88
1.02
0.00
0.00
3.27
2.57
0.2 + 0.3 + 0.2 4 0.73 0.00 2.28
LCLR = D3 R CL = R =
3
= .233 5 0.58 0.00 2.11
6 0.48 0.00 2.00
where 7 0.42 0.08 1.92
UCL = D4R = 2.28(.233) = .53 8 0.37 0.14 1.86
D3 = a control chart factor based on sample size, n 9 0.34 0.18 1.82
LCL = D3R = 0.0(.233) = 0.0 10 0.31 0.22 1.78
D4 = a control chart factor based on sample size, n 11 0.29 0.26 1.74
12 0.27 0.28 1.72
13 0.25 0.31 1.69
14 0.24 0.33 1.67
15 0.22 0.35 1.65
LCLR = D 3R LCL = 0
= (0)(5.3)
= 0 pounds
10
Control Chart for AYributes P - chart
þ Fraction, proportion, or percent defectives
• AAributes generate data that are counted. þ a chart of the percent defective in each
(Good/bad, yes/no, acceptable/unacceptable) sample set
– p-Chart þ Use P-Charts for quality characteristics that
are discrete and involve decisions such as:
• Control chart used to monitor the proporGon of þ yes/no
defecGves in a process (Percent defecGve)
þ good/bad
– c-Chart þ pass/fail
• Control chart used to monitor the number of defects þ operate/not
per unit (Number of defects)
þ Ex: percent of leaking tubes in a box,
percent of broken eggs in a carton, etc.
C - chart C – chart
þ Observations are attributes whose defects
per unit of output can be counted • Example of occurrences and units of measure include:
þ The number counted is a small part of the – Scratches, chips, dents, or errors per item
possible occurrences – Cracks or faults per unit of distance (e.g. meters, miles)
– Breaks or tears, per unit of area (e.g. square yard, sqm.)
þ A chart of the number of defects per unit in
each sample set, for discrete defects when – Bacteria or pollutants per unit of volume (e.g. gallon, cubic
there can be more than one defect per unit foot, cubic yard)
– Calls, complaints, failures, equipment breakdown, or
þ Remember: Use only C – chart when the crimes per unit of Gme (e.g. hour, day, month, year)
number of occurrences per unit of measure – Number of flaws or stains in a carpet
can be counted; non-occurrences cannot be
counted. – Number of complaints per customer at a hotel
– Number of typographical errors on a page
11
Step 1:
Calculate the Percent defecGve of Each Sample and the
Step 2: Calculate the Standard DeviaGon of
Overall Percent DefecGve (P-Bar) P.
n Center line (p bar):
CL = p = .09
n Control limits for ±3σ limits:
p=
80 = .04 (.04)(1 - .04)
(100)(20) σp^= = .02
100
12
P – chart Example 2 P – chart Example 2
UCLp = p + ^zσp = .04 + 3(.02) = .10 UCLp = p + ^zσp = .04 + 3(.02) = .10
Possible assignable
LCLp = p - zσ
^
p = .04 - 3(.02) = 0 LCLp = p - zσ
^
p = .04 - 3(.02) = 0
causes present
Fraction defective
.08 – .08 –
.07 – .07 –
.06 – .06 –
.05 – .05 –
.04 – p = 0.04 .04 – p = 0.04
.03 – .03 –
.02 – .02 –
.01 – LCLp = 0.00 .01 – LCLp = 0.00
.00 – | | | | | | | | | | .00 – | | | | | | | | | |
2 4 6 8 10 12 14 16 18 20 2 4 6 8 10 12 14 16 18 20
Sample number Sample number
n Center line (c bar):
#complaints 22
CL = = = 2.2
# of samples 10
# of Complaints
13
C – chart Example 2 C – chart Example 2
Red Top Cab Company receives several c = 54 complaints/9 days = 6 complaints/day
complaints per day about the behavior of its
drivers. Over a 9-day period (where days are the UCLc = c + 3 c UCLc = 13.35
Number of complaints
units of measure), the owner received the =6+3 6
14 –
12 –
following number of calls from irate passengers: = 13.35 10 –
3, 0, 8, 9, 6, 7, 4, 9, 8, for a total of 54 8 –
complaints. 6 – c= 6
LCLc = c - 3 c 4 –
=6-3 6 2 – LCLc = 0
Compute using 99.73% control limits. =0 0 – | |
1 2
|
3
|
4
|
5
|
6
|
7
|
8
|
9
Day
14
Assignment: Prac/ce Exercise
Problem:
Samples were taken in 10 days from a produc/on lot of single served meat
packs isolated from an inspec/on in a commissary with each day consis/ng
of 4 samples each. The opera/ons manager determines that the standard
devia/on of the process is 1 . Use the data below to develop control charts
with limits of 3 standard devia/ons.
1
To Do:
X - Chart
2
R-bar Chart
Range Chart