CISPR TR 16-4-3:2004+AMD1:2006 CSV

Specification for radio disturbance and immunity measuring apparatus and methods - Part 4-3: Uncertainties, statistics and limit modelling - Statistical considerations in the determination of EMC compliance of mass-produced products

This Technical Report deals with statistical considerations in the determination of EMC compliance of mass-produced products. The reasons for such statistical considerations are: a) that the abatement of interference aims that the majority of the appliances to be approved shall not cause interference; b) that the CISPR limits should be suitable for the purpose of type approval of mass-produced appliances as well as approval of single-produced appliances; c that to ensure compliance of mass-produced appliances with the CISPR limits, statistical techniques have to be applied; d) that it is important for international trade that the limits shall be interpreted in the same way in every country; e) that the National Committees of the IEC which collaborate in the work of the CISPR should seek to secure the agreement of the competent authorities in their countries. Therefore, this part of CISPR 16 specifies requirements and provides guidance based on statistical techniques. EMC compliance of mass-produced appliances should be based on the application of statistical techniques that must reassure the consumer, with an 80 % degree of confidence, that 80 % of the appliances of a type being investigated comply with the emission or immunity requirements. This consolidated version consists of the second edition (2004) and its amendment 1 (2006). Therefore, no need to order amendment in addition to this publication.

General Information

• Status: Published
• Publication Date: 17-Jan-2007
• ICS: 01 - GENERALITIES. TERMINOLOGY. STANDARDIZATION. DOCUMENTATION
• ICS: 33.100.10 - Emission
• ICS: 33.100.20 - Immunity
• Technical Committee: A - Radio-interference measurements and statistical methods
• Current Stage: PPUB - Publication issued
• Start Date: 18-Jan-2007
• Completion Date: 18-Jan-2007

Standards Content (Sample)

CISPR 16-4-3
Edition 2.1 2007-01
TECHNICAL
REPORT

INTERNATIONAL SPECIAL COMMITTEE ON RADIO INTERFERENCE
Specification for radio disturbance and immunity measuring apparatus and
methods –
Part 4-3: Uncertainties, statistics and limit modelling – Statistical considerations
in the determination of EMC compliance of mass-produced products


CISPR 16-4-3/TR:2004+A1:2006(E)

---------------------- Page: 1 ----------------------
THIS PUBLICATION IS COPYRIGHT PROTECTED
Copyright © 2007 IEC, Geneva, Switzerland

All rights reserved. Unless otherwise specified, no part of this publication may be reproduced or utilized in any form
or by any means, electronic or mechanical, including photocopying and microfilm, without permission in writing from
either IEC or IEC's member National Committee in the country of the requester.
If you have any questions about IEC copyright or have an enquiry about obtaining additional rights to this publication,
please contact the address below or your local IEC member National Committee for further information.

IEC Central Office
3, rue de Varembé
CH-1211 Geneva 20
Switzerland
Email: [e-mail]
Web: www.iec.ch

About the IEC
The International Electrotechnical Commission (IEC) is the leading global organization that prepares and publishes
International Standards for all electrical, electronic and related technologies.

About IEC publications
The technical content of IEC publications is kept under constant review by the IEC. Please make sure that you have the
latest edition, a corrigenda or an amendment might have been published.
ƒ Catalogue of IEC publications: www.iec.ch/searchpub
The IEC on-line Catalogue enables you to search by a variety of criteria (reference number, text, technical committee,…).
It also gives information on projects, withdrawn and replaced publications.
ƒ IEC Just Published: www.iec.ch/online_news/justpub
Stay up to date on all new IEC publications. Just Published details twice a month all new publications released. Available
on-line and also by email.
ƒ Electropedia: www.electropedia.org
The world's leading online dictionary of electronic and electrical terms containing more than 20 000 terms and definitions
in English and French, with equivalent terms in additional languages. Also known as the International Electrotechnical
Vocabulary online.
ƒ Customer Service Centre: www.iec.ch/webstore/custserv
If you wish to give us your feedback on this publication or need further assistance, please visit the Customer Service
Centre FAQ or contact us:
Email: [e-mail]
Tel.: [phone]
Fax: [phone]

---------------------- Page: 2 ----------------------
CISPR 16-4-3
Edition 2.1 2007-01
TECHNICAL
REPORT

INTERNATIONAL SPECIAL COMMITTEE ON RADIO INTERFERENCE
Specification for radio disturbance and immunity measuring apparatus and
methods –
Part 4-3: Uncertainties, statistics and limit modelling – Statistical considerations
in the determination of EMC compliance of mass-produced products

INTERNATIONAL
ELECTROTECHNICAL
COMMISSION
PRICE CODE
CM
ICS 33.100.10; 33.100.20 ISBN 2-8318-8915-4

---------------------- Page: 3 ----------------------
– 2 – TR CISPR 16-4-3 © IEC:2004
+A1:2006(E)
CONTENTS

FOREWORD.3

1 Scope.5
2 Normative references.5
3 Terms, definitions and symbols .6
4 General requirements .6
4.1 Limits.6
4.2 Type testing approaches .6
5 Emission measurements.6
5.1 Test based on the non-central t-distribution.6
5.2 Test based on the binomial distribution .9
5.3 Test based on an additional acceptance limit .9
5.4 Additional sampling in case of non-compliance.10
5.5 Properties of the different methods that can be used .11
5.6 Compliance criteria and measurement instrumentation uncertainty.12
6 Immunity tests.12
6.1 Application of the CISPR 80 %/80 % rule to immunity tests .12
6.2 Application guidelines.12

Annex A (informative) Statistical considerations in the determination of limits of radio
interference .14
Annex B (informative) An analytical assessment of statistical parameters of radio
disturbance in the case of an incompletely defined sample .22
Annex C (informative) Test based on an additional acceptance limit .27
Annex D (informative) Estimation of the acceptance probability of a sample .31

Bibliography.36

---------------------- Page: 4 ----------------------
TR CISPR 16-4-3 © IEC:2004 – 3 –
+A1:2006(E)
INTERNATIONAL ELECTROTECHNICAL COMMISSION
____________

SPECIFICATION FOR RADIO DISTURBANCE
AND IMMUNITY MEASURING APPARATUS AND METHODS –

Part 4-3: Uncertainties, statistics and limit modelling –
Statistical considerations in the determination
of EMC compliance of mass-produced products


FOREWORD
1) The International Electrotechnical Commission (IEC) is a worldwide organization for standardization comprising
all national electrotechnical committees (IEC National Committees). The object of IEC is to promote
international co-operation on all questions concerning standardization in the electrical and electronic fields. To
this end and in addition to other activities, IEC publishes International Standards, Technical Specifications,
Technical Reports, Publicly Available Specifications (PAS) and Guides (hereafter referred to as “IEC
Publication(s)”). Their preparation is entrusted to technical committees; any IEC National Committee interested
in the subject dealt with may participate in this preparatory work. International, governmental and non-
governmental organizations liaising with the IEC also participate in this preparation. IEC collaborates closely
with the International Organization for Standardization (ISO) in accordance with conditions determined by
agreement between the two organizations.
2) The formal decisions or agreements of IEC on technical matters express, as nearly as possible, an international
consensus of opinion on the relevant subjects since each technical committee has representation from all
interested IEC National Committees.
3) IEC Publications have the form of recommendations for international use and are accepted by IEC National
Committees in that sense. While all reasonable efforts are made to ensure that the technical content of IEC
Publications is accurate, IEC cannot be held responsible for the way in which they are used or for any
misinterpretation by any end user.
4) In order to promote international uniformity, IEC National Committees undertake to apply IEC Publications
transparently to the maximum extent possible in their national and regional publications. Any divergence
between any IEC Publication and the corresponding national or regional publication shall be clearly indicated in
the latter.
5) IEC provides no marking procedure to indicate its approval and cannot be rendered responsible for any
equipment declared to be in conformity with an IEC Publication.
6) All users should ensure that they have the latest edition of this publication.
7) No liability shall attach to IEC or its directors, employees, servants or agents including individual experts and
members of its technical committees and IEC National Committees for any personal injury, property damage or
other damage of any nature whatsoever, whether direct or indirect, or for costs (including legal fees) and
expenses arising out of the publication, use of, or reliance upon, this IEC Publication or any other IEC
Publications.
8) Attention is drawn to the Normative references cited in this publication. Use of the referenced publications is
indispensable for the correct application of this publication.
9) Attention is drawn to the possibility that some of the elements of this IEC Publication may be the subject of
patent rights. IEC shall not be held responsible for identifying any or all such patent rights.
The main task of IEC technical committees is to prepare International Standards. However, a
technical committee may propose the publication of a technical report when it has collected
data of a different kind from that which is normally published as an International Standard, for
example "state of the art".
CISPR 16-4-3, which is a technical report, has been prepared by CISPR subcommittee A:
Radio interference measurements and statistical methods.
This second edition of CISPR 16-4-3 cancels and replaces the first edition published in 2003
and constitutes a technical revision. It includes a new mathematical approach for the
application of the 80%/80% rule, based on a method involving an additional acceptance limit.
The mathematical basis for this new method is also provided. Furthermore, an additional test
approach, based on the non-central t-distribution and using frequency sub-ranges has been
added as well, along with a description of the properties of all methods which are available at
this point in time.

---------------------- Page: 5 ----------------------
– 4 – TR CISPR 16-4-3 © IEC:2004
+A1:2006(E)
This consolidated version of CISPR 16-4-3 consists of the second edition (2004) [documents
CISPR/A/491/DTR + CISPR/A/492/DTR and CISPR/A/507/RVC + CISPR/A/508/RVC] and its
amendment 1 (2006) [documents CISPR/A/666/DTR and CISPR/A/691/RVC].
The technical content is therefore identical to the base edition and its amendment and has
been prepared for user convenience.
It bears the edition number 2.1.
A vertical line in the margin shows where the base publication has been modified by
amendment 1.
This publication has been drafted in accordance with the ISO/IEC Directives, Part 2.
The committee has decided that the contents of the base publication and its amendments will
remain unchanged until the maintenance result date indicated on the IEC web site under
"http://webstore.iec.ch" in the data related to the specific publication. At this date,
the publication will be
• reconfirmed,
• withdrawn,
• replaced by a revised edition, or
• amended.

A bilingual version of this publication may be issued at a later date.

---------------------- Page: 6 ----------------------
TR CISPR 16-4-3 © IEC:2004 – 5 –
+A1:2006(E)
SPECIFICATION FOR RADIO DISTURBANCE
AND IMMUNITY MEASURING APPARATUS AND METHODS –

Part 4-3: Uncertainties, statistics and limit modelling –
Statistical considerations in the determination
of EMC compliance of mass-produced products



1 Scope
This part of CISPR 16 deals with statistical considerations in the determination of EMC
compliance of mass-produced products.
The reasons for such statistical considerations are:
a) that the abatement of interference aims that the majority of the appliances to be approved
shall not cause interference;
b) that the CISPR limits should be suitable for the purpose of type approval of mass-
produced appliances as well as approval of single-produced appliances;
c) that to ensure compliance of mass-produced appliances with the CISPR limits, statistical
techniques have to be applied;
d) that it is important for international trade that the limits shall be interpreted in the same
way in every country;
e) that the National Committees of the IEC which collaborate in the work of the CISPR should
seek to secure the agreement of the competent authorities in their countries.
Therefore, this part of CISPR 16 specifies requirements and provides guidance based on
statistical techniques. EMC compliance of mass-produced appliances should be based on the
application of statistical techniques that must reassure the consumer, with an 80 % degree of
confidence, that 80 % of the appliances of a type being investigated comply with the emission
or immunity requirements. Clause 4 gives some general requirements for this so-called
80 %/80 % rule. Clause 5 gives more specific requirements for the application of the
80 %/80 % rule to emission tests. Clause 6 gives guidance on the application of the CISPR
80 %/80 % rule to immunity tests. The 80 %/80 % rule protects the consumer from non-
compliant appliances, but it says hardly anything about the probability that a batch of
appliances from which the sample has been taken will be accepted. This acceptance
probability is very important to the manufacturer. In Annex A, more information is given on
acceptance probability (manufacturer’s risk).
2 Normative references
The following referenced documents are indispensable for the application of this document.
For dated references, only the edition cited applies. For undated references, the latest edition
of the referenced document (including any amendments) applies.
IEC 60050-161:1990, International Electrotechnical Vocabulary (IEV) – Chapter 161:
Electromagnetic compatibility
Amendment 1 (1997)
Amendment 2 (1998)
CISPR 16-4-2, Specification for radio disturbance and immunity measuring apparatus
and methods – Part 4-2: Uncertainties, statistics and limit modelling – Uncertainty in EMC
measurements

---------------------- Page: 7 ----------------------
– 6 – TR CISPR 16-4-3 © IEC:2004
+A1:2006(E)
3 Terms, definitions and symbols
For the purpose of this document, the terms, definitions and symbols given in IEC 60050-161
apply.
4 General requirements
The following interpretation of CISPR limits and of methods of statistical sampling for
compliance of mass-produced appliances with these limits should be applied.
4.1 Limits
4.1.1 A CISPR limit is a limit that is recommended to national authorities for incorporation in
national standards, relevant legal regulations and official specifications. It is also
recommended that international organizations use these limits.
4.1.2 The significance of the limits for type-approved appliances shall be that, on a
statistical basis, at least 80 % of the mass-produced appliances comply with the limits with at
least 80 % confidence.
4.2 Type testing approaches
Type tests can be made using the following two approaches.
4.2.1 Use of a sample of appliances of the same type
When using this approach, the sample of appliances of the same type shall be evaluated
statistically in accordance with the methods described in Clause 5 (emission tests) and
Clause 6 (immunity tests).
Statistical assessment of compliance with limits shall be made according to the methods
described in Clauses 5 and 6 or in accordance with some other method that ensures
compliance with the requirements of clause 4.1.2.
4.2.2 Use of a single device with subsequent quality assurance testing
For simplicity, a type test can be performed initially on one item only. However, subsequent
tests from time to time on items taken at random from the production are necessary.
4.2.3 Withdrawal of the type approval
In the case of controversy involving the possible withdrawal of a type approval, withdrawal
shall be considered only after tests on an adequate sample in accordance with 4.2.1 above.
5 Emission measurements
Statistical assessment of compliance with emission limits shall be made according to one of
the three tests described below or to some other test that ensures compliance with the
requirements of 4.1.2.
5.1 Test based on the non-central t-distribution.
This test should be performed on a sample of not less than five items of the type, but if, in
exceptional circumstances, five items are not available, then a sample of three shall be used.
Compliance is judged from the following relationship:
x + kS ≤ L (1)
n n

---------------------- Page: 8 ----------------------
TR CISPR 16-4-3 © IEC:2004 – 7 –
+A1:2006(E)
where
x = arithmetic mean value of the levels of n items in the sample;
n
2
2
 S = ()x − x ()n −1; (2)
∑
n n
x = level of individual item;
k = the factor derived from tables of the non-central t-distribution with 80 % confidence that
80 % of the type is below the limit; the value of k depends on the sample size n and is
stated below:
N 3 4 5 6 7 8 9 10 11 12
k 2,04 1,69 1,52 1,42 1,35 1,30 1,27 1,24 1,21 1,20

L = the permissible limit;
the quantities x, x , S and L are expressed logarithmically dB(μV), dB(μV/m) or dB(pW);
n n
If one or some appliance of the sample can not be measured due to the insufficient sensitivity
of the test equipment, Annex B describes an approach to solve this situation.
5.1.1 Tests using sub-ranges
5.1.1.1 Introduction
The 80 %/80 % rule shall be used for the specific emission at a specific frequency or
frequency range at each EUT of the sample. Modern computer-controlled measurement
equipment usually scans the frequency range and measures a limited number of the highest
disturbances at certain frequencies of the whole emission spectrum. Because the level of the
disturbance at the same frequency or the frequency at the highest emission varies from EUT
to EUT, the measured frequencies of the highest disturbance levels usually vary from one
EUT to another in a sample. These measurement results cannot be used for the
80 %/80 % rule because one does not obtain measurement levels at approximately the same
frequency for each EUT to calculate the average and standard deviation of the EUT’s level.
For this reason, it is useful to divide the whole frequency range into defined sub-ranges,
which allow a statistical analysis of the emission spectrum in the whole frequency range by
taking the highest measured level in each sub-range.
For the application of the non-central t-distribution in the 80 %/80 % rule, it is necessary to
normalise the measured values. These normalised values allow the use of the 80 %/80 % rule
in the sub-ranges independently of variations of the limit in a sub-range.
The whole frequency range shall be divided on a logarithmic frequency axis into sub-ranges.
The border of the sub-ranges may correspond to changes in limits, if a product committee so
requires.
NOTE The division of the frequency range into sub-ranges is applicable only to the test based on the non-central
t-distribution.
5.1.1.2 Number of sub-ranges
It is suggested that the frequency range of the disturbance measurement method in question
is divided into a number of frequency sub-ranges. The span of each frequency sub-range
should decrease in a logarithmic way as a function of the frequency. For the different
disturbance measurement methods, the following number of sub-ranges is suggested:
– at least 8 sub-ranges in the frequency range of up to 30 MHz for the measurement of the
disturbance voltage;
– at least 4 sub-ranges in the frequency range from 30 MHz to 300 MHz for the
measurement of the disturbance power, and

---------------------- Page: 9 ----------------------
– 8 – TR CISPR 16-4-3 © IEC:2004
+A1:2006(E)
– about 8 sub-ranges in the frequency range from 30 MHz to 1000 MHz for the
measurement of disturbance field strength.
NOTE 1 The number of sub-ranges shall be determined such that the frequency dependence of the
disturbance’s characteristic can be estimated. This condition is fulfilled if the ratio of limit to average plus
standard deviation of the emission in the sub-ranges does not decrease when the number of sub-ranges is
reduced.
NOTE 2 The product committees should determine the number of sub-ranges depending on the disturbance
characteristics of the different products.
NOTE 3 The recommended number of sub-ranges is based on the investigations of samples of CISPR 14 and
CISPR 22 devices.
NOTE 4 The sub-range transition frequency can be calculated as follows:
⎛ f ⎞
i
upp
⎜ ⎟
log
⎜ ⎟
N f
⎝ low ⎠
f = f ×10
low
where
i = 1 … N is the index of the i-th sub-range transition frequency;
f , f are the lower and upper frequency of the frequency range;
low upp
N= is the number of frequency sub-ranges.
NOTE 5 For predominantly narrow band emission it is possible to select single narrow band emission by
preexamination for the use of the non-central t-distribution without using sub-ranges.
5.1.1.3 Normalization of the measured disturbance levels
The average value and the standard deviation of the measured values in a frequency sub-
range shall be compared to the limit. Because the limit may not be constant over the
frequency sub-range, it is necessary to normalize the measured values.
For normalization, the difference, d , between the measured level, x , and the limit level, L , shall
f f f
be determined at the specific frequency f that has the largest difference, using Equation (3).
The difference is negative as long as the measured value is below the limit.
d = x – L (3)
f f f
where
d = the gap to the limit at the specific frequency in dB;
f
x = the measured level in dB(μV or pW or μV/m);
f
L = the limit at the specific frequency in dB(μV or pW or μV/m).
f
5.1.1.4 Tests based on the non-central t-distribution with frequency sub-ranges
As a result of the measurement of all pieces of the sample for each sub-frequency range, the
average and the standard deviation of the gap d shall be calculated. The average of the gap
f
is
1
d = d (4)
∑
f f
n n
where
n = the number of items in the sample
d = the average gap in the sub-range
f

---------------------- Page: 10 ----------------------
TR CISPR 16-4-3 © IEC:2004 – 9 –
+A1:2006(E)
and the standard deviation is
1
2
S = (d − d ) (5)
df ∑ f f
n −1
n
where S = the standard deviation in the sub-range.
df
Compliance is judged from the following relationship:
d + k ⋅ S ≤ 0
f df
(6)
k: see 5.1 above.
5.2 Test based on the binomial distribution
This test should be performed on a sample of not less than seven items. Compliance is
judged from the condition that the number of appliances with an interference level above the
permissible limit may not exceed c in a sample of size n.
n 7 14 20 26 32
c 0 1 2 3 4

5.3 Test based on an additional acceptance limit
This test should be performed on a sample of not less than five items of a particular type, but
if, in exceptional circumstances, five items are not available, then a sample of at least three
shall be used. Details on this method are described in 5.5. Compliance is judged if every
measured disturbance level x satisfies the following relation:
i
x ≤ AL = L – σ ·k (7)
i max E
where
AL is the acceptance limit
L is the permissible limit
σ is the expected maximum standard deviation of the product, which is 2 times the
max
expected standard deviation, and which is determined by the product committee
using the procedure of 5.3.1 or alternatively the following conservative values for the
different types of disturbance measurements can be used:
disturbance voltage: σ = 6 dB*)
max
disturbance power: σ = 6 dB**)
max
1
disturbance field strength: σ = xx dB
max
NOTE 1 The values of 6 dB were determined by measurements of 130*) and 40**) different EUT types
(3 or 5 samples each). The value of 6 dB was estimated by comparing the tests using the non-central
t-distribution with the tests using the additional margin. Both tests give about the same percentage of
approvals.
NOTE 2 The disturbance field strength value is under consideration.
———————
1
 Under consideration

---------------------- Page: 11 ----------------------
– 10 – TR CISPR 16-4-3 © IEC:2004
+A1:2006(E)
k is the factor derived from tables of the normal distribution with 80 % confidence that
E
80 % of the type is below the limit; the value of k depends on the sample size n and
E
is stated below (see Annex C.1):
n 3 4 5 6
k 0,63 0,41 0,24 0,12

E
The quantities x, L, k and σ are expressed logarithmically as dB(μV), dB(μV/m)
E max
or dB(pW).
NOTE With σ = 6 dB the following additional acceptance limit will be calculated:
max
Sample size 3 4 5 6
additional acceptance limit [dB] 3,8 2,5 1,5 0,7

5.3.1 Estimation of the maximum expected standard deviation
The expected standard deviation of disturbance emission shall be determined by an efficient
number of samples of the product concerned. The following procedure is recommended:
On each investigated frequency or in each frequency sub-range in the sample being
investigated, the difference x between the measured maximum emission x and the limit L
min i
shall be determined
x = (x – L) (8)
min i max
The standard deviation S of the differences in a sub-range or investigated frequency of a
sub
sample shall be calculated
1
2
S = (x − x ) (9)
∑
sub min min
n −1 n
where n is the number of appliances in the sample.
The average standard deviation S over the sub-ranges shall be determined for each
sample
sample. The expected standard deviation S is the average over S of all samples.
sample
expect
The maximum expected standard deviation is two times the expected standard deviation.
NOTE The factor of two is chosen by comparison of the test methods using the additional margin and the non-
central t-distribution. Both test methods have, with the factor two, approximately the same rejection rate of
samples.
Product committees may verify the expected standard deviation of their products.
5.4 Additional sampling in case of non-compliance
Should the test on the sample result in non-compliance with the requirements in 5.1, 5.2 or
5.3, then a second sample may be tested and the results combined with those from the first
sample and compliance checked for the larger sample. For 5.3 this method is only applicable
to samples of 7 or less appliances.

---------------------- Page: 12 ----------------------
TR CISPR 16-4-3 © IEC:2004 – 11 –
+A1:2006(E)
5.5 Properties of the different methods that can be used
The possible four test methods for compliance evaluation of mass products are:
• using a single device,
• non-central t-distribution (see 5.1),
• binomial distribution (see 5.2) and
• the additional margin (see 5.3)
Each of these methods are based on different statistical methodologies, and therefore each of
the methods have different properties (advantages or disadvantages) when applied in practice
by manufacturers or authorities.
a) Using a single device
A test on a single device is used by manufacturers. The method requires that repetitive
testing of the product over time has to occur.
b) Non-central t-distribution:
The test is based on the non-central t-distribution and contains the condition of normal
distribution for the totality. As long as this condition is fulfilled, the test gives correct
results for the approval of a sample. But disapproval may be indicated without reason if
one or two measurements are far below the limit and the other measurement results are
near (but below) the limit.
If the failure is caused by measurement results far below the limit due to the large
standard deviation, alternatively the test with the additional margin may be used for the
failed sample. If the sample passes, the product is o.k.
In case of disapproval, it is possible to select further devices from the same product batch
and to combine all the failed and newly selected devices in a larger sample.
An advantage of this test method is that the sample can be relatively small.
c) Binomial distribution:
The test is based on the binomial distribution and contains no further condition of
distribution for the totality. The test gives correct results for the approval and disapproval
of a sample.
In case of disapproval, it is possible to select further devices from the same product batch
and to combine all the failed and newly selected devices in a larger sample.
The disadvantage of this test method is that the sample must have at least 7 devices.
d) Additional acceptance limit:
The test is based on the condition of normal distribution for the totality and the estimation
of the expected standard deviation. The test gives correct results for the approval of a
sample.
If the failure is caused by measurement results which are close to the limit, an additional
test on the sample based on the non-central t-distribution may be used for the failed
sample. If the sample passes the test, the product is o.k.
In case of disapproval, it is possible to select further devices from the same product batch
and to combine all the failed and newly selected devices in a larger sample. This method
is only applicable to samples with less than 7 devices.

---------------------- Page: 13 ----------------------
...