Introduction To Acceptance Sampling
Introduction To Acceptance Sampling
com
Copyright 2004, www.sixsigmatutorial.com. All rights reserved.
Reproduction or sharing without written permission is prohibited. 1 of 16
BASICS OF ACCEPTANCE SAMPLING
www.SixSigmaTutorial.com
Author
Mi chael G. Whi te, P.E., C.Q.E.
President, Qi2 (www.SixSigmaQi2.com)
qi 2@qual i tyi 2.com
Acceptance Sampling Tutorial www.SixSigmaTutorial.com
Copyright 2004, www.sixsigmatutorial.com. All rights reserved.
Reproduction or sharing without written permission is prohibited. 2 of 16
OPERATI NG CHARACTERISTIC ( OC) CURVE
The OC curve quantifies the and risks of an attribute sampling plan. Below is an ideal
OC curve (the bold line) for a situation in which we might want to accept all lots that
are, say, 1% defective and rej ect all l ots t hat are > 1% defective:
With this ideal (no risks) curve, all batches with < 1% defective incoming quality level
woul d have a probabilit y of accept ance (P
a
) of 1. 0. And, all lots with > 1%
defective would have a P
a
of 0. The P
a
is the probability that the sampling plan will accept the
lot; it is the long-run % of submitted lots that would be accepted when many lots of a stated
quality level are submitted for inspection. It is the probability of accepting lots from a
steady stream of product having a fraction defective p.
TYPICAL OC CURVE
Since there will always be some risks, a more typical looking OC curve l ooks more like the
one below. It is based on the Poisson distribution* (with the defective rate < 10%
and n is relatively large compared to N).
0 1 2 3
Lot % Defective
1.0
P
a
.5
0
Acceptance Sampling Tutorial www.SixSigmaTutorial.com
Copyright 2004, www.sixsigmatutorial.com. All rights reserved.
Reproduction or sharing without written permission is prohibited. 3 of 16
The AQL (Acceptance Quality Level), the maximum % defective t hat can be
consi dered sat i sfact or y as a process average for sampling inspection, here is 1%. Its
corresponding P
a
is about 89%. It should normally be at least that high.
The RQL (Rejectable Quality Level) is the % defective, here at 5%, t hat i s associ at ed wi t h
t he est abl i shed ri sk (whi ch i s usual l y st andar di zed at 10%) . It i s al so
known as t he Lot Tol erance Percent Defective (LTPD).
*The hypergeometric and binomial distns are also used.
The alpha risk is the probability of rejecting relatively good lots (at AQL). The beta risk is the
probability of accepting relatively bad lots (at LTPD/RQL). It is the probability of accepting
product of some stated undesirable quality; it is the value of P
a
at that stated quality level.
The OC curves are a means of quantifying alpha and beta risks for a given attribute sampling
plan. The P
a
value obtained assumes that the distribution of defectives among a lot is random
either the underlying process is in control, or the product was well mixed before being divided
into lots. The samples must be selected randomly from the entire lot. The alpha risk is 1 P
a
.
The shape of the OC curves is affected by the sample size (n) and accept number (c) parameters.
Increasing both the accept number and sample size will bring the curve closer to the ideal shape,
with better discrimination.
0 1 2 3 4 5 6
AQL Lot % Def ect i ve RQL
1.0
.9
P
a
.5
.1
0
Acceptance Sampling Tutorial www.SixSigmaTutorial.com
Copyright 2004, www.sixsigmatutorial.com. All rights reserved.
Reproduction or sharing without written permission is prohibited. 4 of 16
If the lot size N changes, the above curves change very little. However, the curves will change
quite a bit as sample size n changes. So, basing a sampling plan on a fixed percentage
sample size will yield greatly different risks. For consistent risk levels, it is better to fix the
sample size at n, even if the lot sizes N vary.
If n = 10 & c = 2, what is the alpha risk for a vendor running at p = .02? P
a
is about .55, so alpha
is about .45. What is the beta risk if the worst-case quality the customer will accept is 3%?
(about 15%). To lower alpha & beta, you can increase n & c.
1.0
.9
P
a
.5
.1
0
0 1 2 3 4 5 6
Lot % Def ect i ve
n = 10
c = 2
n = 50
c = 5
n = 20
c = 0
Acceptance Sampling Tutorial www.SixSigmaTutorial.com
Copyright 2004, www.sixsigmatutorial.com. All rights reserved.
Reproduction or sharing without written permission is prohibited. 5 of 16
Is c = 0 the best plan for the producer and the consumer?
At the 2.8% lot defect rate, both plans give the producer equal protection: P
a
= 11%, or
P
r ej .
= 89%. Which one gives better prot ection against rej ect ing relat ivel y good
lots, e. g. , at the .5% lot defect rate, and why? For (1), = about 8%; for (2), =
about 30%. (1) has a lower error so l ess chance of rejecti ng good lot s. With
(2), you will reject any lot of 500 if there is even 1 defect i n the sampl e, but it
will l ead to hi gher costs.
DISCRIMINATION IN ACCEPTANCE SAMPLI NG PLANS
Discrimination is the ability of a sampling plan to distinguish between relatively good levels of
quality and relatively bad levels of quality, i.e., having
A high P
a
(e.g., 95%, 1-) associated with a good level of quality P
1
(e.g., .5% or better)
A low P
a
(e.g., 10%, ) associated with a bad level of quality P
2
(e.g., 3% or worse)
1.0
.9
.8
.7
P
a
.6
.5
.4
.3
.2
.1
0
0 .5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Lot % Defect i ve
N = 500
n = 100
c = 0
N = 500
n = 150
c = 1
(2)
(1)
Acceptance Sampling Tutorial www.SixSigmaTutorial.com
Copyright 2004, www.sixsigmatutorial.com. All rights reserved.
Reproduction or sharing without written permission is prohibited. 6 of 16
The Operating Ratio is defined as
R =
2
1 1
p
p
p p
=
Example
R =
.03
.005
= 6.0
DESI GNI NG YOUR OWN SI NGLE ACCEPTANCE SAMPLI NG
PLAN
Derive a plan that comes as close as possible to satisfying two points on the OC curve
The two points are (p
1
, 1-) and (p
2
, )
The derived plan will contain an n and a c value
Example
Desired risk of .05 for a p
1
of .005, along with a desired risk of .05 for a p
2
of .03.
1. Determine R:
R =
2
1
p
p
=
.030
.005
= 6.0
2. Enter the Values of Operating Ratio Table with and and find the closest R to the
calculated value in step 1.
For = .05 and = .05, the closest table value is 5.67. This is acceptable since it is
slightly more discriminating than 6.0. Note the c value of 3 in the far left column.
3. Obtain the np
1
value in the far right column. Then calculate n from:
n =
1
1
n p
p
=
1.366
.005
= 273.2 274.
The acceptance sampling plan is n = 274, c = 3.
Acceptance Sampling Tutorial www.SixSigmaTutorial.com
Copyright 2004, www.sixsigmatutorial.com. All rights reserved.
Reproduction or sharing without written permission is prohibited. 7 of 16
OC CURVE
AQL BASED SAMPLI NG PLANS
ANSI/ASQ Z1.4-1993 (MIL-STD 105E was withdrawn in February 1995) sampling
plans are based on the use of AQL the percent defective that is considered
acceptable as a process average for t he purposes of acceptance sampling.
With these plans, it is not necessary to assume the provision of 100% screening
(with replacement of all defective units) of all rejected lots.
To enter any of these tables, you must first decide on the AQL to use (and
determine the sample size code letter); from the table you will get the
acceptance number (Ac) and the rejection number (Re) for the plan.
There are also separate tables to provide the AOQL for different values of
AQL and sample code letters
o More serious defects should have a lower AQL as the accept ance cri t eri on,
and l ess seri ous defect s can have higher AQLs using a
Classification of Defects system
o Tightened inspection should be used whenever the quality history is
unsatisfactory or when there are other good reasons for being suspicious
about quality keeping the beta risk down
o Reduced inspection can be used when the quality history is shown to be
good enough through Normal inspection
0 p
1
p
2
Lot Proportion Defective
1.0
1-
P
a
0
.95
.05
.005 .03
Acceptance Sampling Tutorial www.SixSigmaTutorial.com
Copyright 2004, www.sixsigmatutorial.com. All rights reserved.
Reproduction or sharing without written permission is prohibited. 8 of 16
o These plans are generally chosen to protect the producer under normal
conditions, i.e., not to reject submitted lots that are at the AQL or
better when there are no reasons to be suspicious about quality
DOUBLE SAMPLI NG PLANS
There is usually less sampling than for a single sampling plan
The OC curve is better than the c = 0 curve for the single sampling plan with a smaller
R (for better discrimination)
Inspect the first sample of n
1
pieces
If d
1
Ac
1
if Ac
1
< d
1
< Re
1
if d
1
Re
1
Inspect a 2
nd
sample of n
2
pieces
Accept the Lot
Reject the Lot
If d
1
+ d
2
< Ac
2
If d
1
+ d
2
Re
2
Acceptance Sampling Tutorial www.SixSigmaTutorial.com
Copyright 2004, www.sixsigmatutorial.com. All rights reserved.
Reproduction or sharing without written permission is prohibited. 9 of 16
DESI GNI NG YOUR OWN DOUBLE ACCEPTANCE SAMPLI NG
PLAN
Exercise
Desired risk of .05 for a p
1
of .008, along with a desired risk of .10 for a p
2
of .06. Use Table
4-4.
1. Determine R:
R =
2
1
p
p
=
.060
.008
= 7.5
2. Enter Duncans Double Sampling Tables and find the closest R to the calculated value
in step 1.
The closest table value is 7.54 in Plan Number 2. This is very close to 7.5. Note the c
1
value of 1 and the c
2
value of 2.
3. For Pa
1-
= .95, obtain the np
1
(pn
1
in the table) value. Then calculate n from:
n =
1
1
n p
p
=
.52
.008
= 65.
The acceptance sampling plan is n
1
= 65, c
1
= 1; n
2
= 65, c
2
= 2.
Acceptance Sampling Tutorial www.SixSigmaTutorial.com
Copyright 2004, www.sixsigmatutorial.com. All rights reserved.
Reproduction or sharing without written permission is prohibited. 10 of 16
VARI ABLES SAMPLI NG PLANS
ANSI/ASQ Z1.9 (MIL-STD-414 is withdrawn) sampling plans are based on t he use of
variabl e dat a (from an assumed normal distri buti on).
Act ual measurement s are made on the sampl es t he sample dat a is used to
calculate a statisti c, such as X, R, or S and then the calculated statistic is
compared to the criti cal val ue from a table.
Acceptance criteria must be applied separately to each quality characteristic (vs.
overall lot accept vs. reject for attribute sampling), so its more expensive than attribute
sampling for larger lots, s o i t s generally best to use only on key characteristics, with attribute
sampling on the rest.
Compared to attribute plans, these plans, for the same n, provide a greater quality
protection in judging a single qualit y characteristic, or for the same amount of risk, a
smaller n is OK.
The use of variable data can provide more information about the extent of
nonconformity How? (Hint: think frequency distribution and what you can do
with this information).
These sampling procedures are based on the assumption that the qualit y
characteristic is normally distributed (it is possible to use data transformation if
it is not).
Procedure:
To use any of these plans, you must first decide on the:
Inspection level (II is the default for Z1.9)
Method to use S or R, with variability (population sigma) known or unknown
(see Decision Tree)
AQL
Lot si ze
Determine the sample-size code letter from the table
Cal cul at e t he Q val ue
Ent er t he Mast er Table by sample size code letter and AQL, and look up the k value.
Acceptance Sampling Tutorial www.SixSigmaTutorial.com
Copyright 2004, www.sixsigmatutorial.com. All rights reserved.
Reproduction or sharing without written permission is prohibited. 11 of 16
Compare the k and Q values if the calculated Q value is than the critical k value
from the table, accept the lot. If not , rej ect i t .
Deci si on Tr ee for Sel ect i ng Type of Vari abl es Sampl i ng Pl an
Determine the AQL to use in the Master Table: Since there are standard AQLs used
in the Master Tables, you need to convert the AQL per table below:
For Specified
AQL Values:
Use this AQL
Value:
To .049 . 10
. 050-. 069 . 10
. 070-. 109 . 10
. 110-. 164 . 15
. 165-. 279 . 25
. 280-. 439 . 40
. 440-. 699 . 65
. 700-1. 09 1. 0
1. 10-1. 64 1. 5
1. 65-2. 79 2. 5
2. 80-4. 39 4. 0
4. 40-6. 99 6. 5
7. 00-10.9 10.0
Unknown
Standard Deviation
Method is used
Range Method
is used:
Variabil i t y () i s:
Known
Use Section B Plans
Use Section C Plans Use Section D Plans
Acceptance Sampling Tutorial www.SixSigmaTutorial.com
Copyright 2004, www.sixsigmatutorial.com. All rights reserved.
Reproduction or sharing without written permission is prohibited. 12 of 16
Determine the Sample-Size Code Letter to use in t he Mast er Tabl e:
ANSI/ASQ Z1.9 Table A.2
L O T
SIZE
GENERAL
I
INSPECTION
II
LEVELS
III
2-8 B B C
9-15 B B D
16-25 B C E
26-50 C D F
51-90 D E G
91-150 E F H
151-280 F G I
281-400 G H J
401-500 G I J
501-1200 H J K
1201-3200 I K L
3201-10000 J L M
10001-35000 K M N
Formulas for the Q Value:
For Section B Standard Deviation Method ( Unknown):
Q
U
=
USL X
S
or Q
L
=
X LSL
S
For Section C Range Method ( Unknown):
Q
U
=
USL X
R
or Q
L
=
X LSL
R
For Section D Standard Deviation Method ( Known):
Q
U
=
USL X
or Q
L
=
X LSL
Note: For 2-sided spec limits, an AQL can be assigned to both limits combined, or to each end
of the spec limit separatel y.
Acceptance Sampling Tutorial www.SixSigmaTutorial.com
Copyright 2004, www.sixsigmatutorial.com. All rights reserved.
Reproduction or sharing without written permission is prohibited. 13 of 16
Table B-1 Standard Deviation Method Master Table B-1 for Normal and Tightened Inspection
for Plans Based on Variability Unknown (Single Specification Limit)
Obtain the k Value from the Master Table:
Code Sample .10 .15 .25 .40 .65 1.00 1.50 2.50 4.00 6.50
Letter Size k k k k k k k k k k
B 3 1.12 .958 .765
C 4 1.46 1.34 1.17 1.01 .815
D 5 1.77 1.65 1.52 1.40 1.24 1.07 .874
E 7 2.22 2.13 2.00 1.88 1.75 1.62 1.50 1.33 1.15 .955
F 10 2.34 2.24 2.11 1.98 1.84 1.72 1.59 1.41 1.23 1.03
G 15 2.42 2.32 2.19 2.06 1.92 1.79 1.65 1.47 1.30 1.09
H 20 2.47 2.37 2.23 2.10 1.96 1.83 1.69 1.51 1.33 1.12
I 25 2.50 2.40 2.26 2.13 1.98 1.85 1.72 1.53 1.35 1.14
J 35 2.55 2.45 2.31 2.18 2.03 1.89 1.76 1.55 1.39 1.19
K 50 2.61 2.50 2.36 2.22 2.08 1.94 1.80 1.57 1.42 1.21
L 75 2.66 2.55 2.41 2.27 2.12 1.98 1.84 1.58 1.96 1.25
M 100 2.69 2.58 2.43 2.29 2.14 2.00 1.86 1.61 1.48 1.26
N 150 2.73 2.62 2.47 2.33 2.18 2.03 1.89 1.65 1.51 1.29
P 200 2.73 2.62 2.47 2.33 2.18 2.04 1.89 1.70 1.51 1.29
.15 .25 .40 .65 1.00 1.50 2.50 4.00 6.50
--
Acceptance Quality Levels (Normal Inspection)
AQL ( Ti ght ened Ins pect i on)
Acceptance Sampling Tutorial www.SixSigmaTutorial.com
Copyright 2004, www.sixsigmatutorial.com. All rights reserved.
Reproduction or sharing without written permission is prohibited. 14 of 16
DESI GNI NG YOUR OWN VARI ABLE SAMPLI NG PLAN
For a known population sigma:
1 2
2
(Z Z )
n
(Z Z )
+
=
For an unknown population sigma:
2
1 2
2
(Z Z )
k
n (1 )
2
(Z Z )
+
= +
Where
2 1
Z Z Z Z
k
Z Z
+
=
+
Exercise
What would be the variables sampling plan (sigma unknown) for the following conditions? High
P
a
( = .05) for a fraction non-conformi ng (P
1
) of . 005, wi t h a l ow P
a
( = . 05) for a
fract i on conformi ng (P
2
) of . 03.
Solution
Z
= Z
.05
= 1.645 Z
= Z
.05
= 1.645
Z
1
= Z
.005
= 2.576 Z
2
= Z
.03
= 1.881
1.881(1.645) 2.576(1.645)
k 2.23
1.645 1.645
+
= =
+
( )
2
2
1.645 1.645
2.23
n 1
2
2.576 1.881
+
= +
= 78.128 79
Note: with the same operating characteristics, an attribute sampling plan would require n = 274.
Example (Method: Population Sigma Known)
A lot of 1500 bobbins is submitted for inspection. Inspection level II, normal
inspection, with AQL = .65%, is to be used. The specified mi ni mum yi el d val ue for
t he t ensi l e st rengt h i s 25. 0 l bs. The variability is known to be 2.4 lbs.
The sample size code letter from Table A.2 is K. In Table D-2 (p. 86 in ANSI/ASQ
Z1.9-2003), for reduced inspection, the required sample size is 7 and the k value is
1.80. The 7 sample specimens tensile strengths are 25.7, 26.4, 26.1, 27.2, 25.8, 28.3,
and 27.4.
Acceptance Sampling Tutorial www.SixSigmaTutorial.com
Copyright 2004, www.sixsigmatutorial.com. All rights reserved.
Reproduction or sharing without written permission is prohibited. 15 of 16
L
X LSL 26.7 25.0
Q .63
2.4
= = =
Since Q
L
< k, the lot does not meet the acceptability criterion and should be rejected.
What are the alpha and beta risks for sampling letter K (p. 23 in ANSI/ASQ Z1.9-2003), for an
AQL of .65, for various incoming quality levels (P)?
P P
a
.25 99.5 .5
.50 96 4
.75 90 10 90
1.50 62 38 62
2.00 46 46
3.00 22 22
4.00 10 10
5.00 4.5 4.5
* All values are in Percentages
Why is the lot rejected even though none of the samples were out of spec? (Assuming this is a
representative sample, a larger, +/- 3 distribution would provide some product out of
specification; in this case, in the left tail of the distribution.)
Relatively
Good Quality
Marginal
Poor
Acceptance Sampling Tutorial www.SixSigmaTutorial.com
Copyright 2004, www.sixsigmatutorial.com. All rights reserved.
Reproduction or sharing without written permission is prohibited. 16 of 16
Si x Si gma Tr ai ni ng & Consul t i ng
If you like the information in this paper, and you need training, consider Qi2 for your needs.
We specialize in on-site work-shops. If you have two or more people for training, you are
probably at the break-even point (vs. sending them off somewhere for an off-the-shelf,
public course).
Qi2 offers both Black Belt and Green Belt workshops, in addition to Executive Overviews,
Champion training, and Lean Six Sigma. Obtain your certification from Qi2. Typically, the
Black Belt curriculum takes 4 weeks, and the Green Belt 2 weeks. The latter does not
include Design of Experiments or Statistical Process Control. You can see typical workshop
outlines at www.SixSigmaQi2.com.
We will customize the workshops (at no charge) to include examples in your
product/process terminology. Each attendee receives a comprehensive set of training/
reference materials for each week of training. The courses are designed to be practical and
flexible. The Black and Green Belt candidates will be given time in each workshop week to
apply the 6 steps and tools to their own projects just after the material is presented.
The course and the materials are organized by the 12 or so steps of DMAIC, so the
attendees will have a roadmap of 6 implementation that they experience in the workshop
and then further apply in between the separated training weeks.
Contact Qi2 at [email protected] for a quote for on-site training at your location, or for
an instructor bio and a list of references.