COMPLIANCE OF SMOKE ALARM DEVICES TO EN 14604 -

STATISTICAL ANALYSIS OF MARKET SURVEILLANCE RESULTS

Benot Stockbroeckx
Head of Division Laboratory

Proceedings AUBE 14, 15th Intern. Conf. on Automatic Fire Detection,

Duisburg, October 2014, Vol. 1, pp. 159-166.

Copyright all rights reserved

ANPI asbl - Parc scientifique Fleming

Granbonpr 1 - 1348 Louvain-la-Neuve
Belgique
info@anpi.be - www.anpi.eu ISBN D 2014/1381/93
ABSTRACT

60 different references of smoke alarm devices have been tested in the frame of market
surveillance processes from several countries. The samples have been tested at ANPI
according to EN14604 for 6 requirements: battery removal indication, marking and
data, directional dependence, initial sensitivity, fire sensitivity, and sound output. The
results have been statistically analysed along 2 axes: requirements and claimed
certification scheme. The results provide the occurrence of non-compliances to the
requirements. It appears that 33% of the sampled products are not compliant for at
least 1 requirement; that 19% of the products have a problem with fire detection; and
that products which claim to comply with at least one voluntary mark certification yield
significant better results by a factor of 2 to 4 than these claiming CE only.

Copyright all rights reserved ANPI 2 DOC ANPI 2014

ABSTRACT .................................................................................................................................2
1 INTRODUCTION .....................................................................................................................5
2 METHODOLOGY ....................................................................................................................5
3 PARAMETERS ........................................................................................................................6
4 RESULTS STATISTICAL ANALYSIS......................................................................................6
Requirement axis analysis results .......................................................................................6
Certification scheme axis analysis results ...........................................................................7
Full analysis results .............................................................................................................8
5 CONCLUSION AND INTERPRETATION ...............................................................................8
REFERENCES ...........................................................................................................................9

Copyright all rights reserved ANPI 3 DOC ANPI 2014

Copyright all rights reserved ANPI 4 DOC ANPI 2014
1 INTRODUCTION parameters used for the statistical
analysis. The results of this analysis
Several countries or regions in are discussed in the third section.
Europe have imposed or will soon Finally, the last section presents the
impose by law the installation of conclusion and an explanation of the
smoke alarm devices in private results.
homes. This has stimulated the
activation of market surveillance For obvious confidentiality reasons,
processes for these products. the product brands and references,
the considered countries, and the
Smoke alarm devices are within the test results will not be disclosed in
scope of the construction product this paper.
regulation (CPR)[1]. An EC certificate
of constancy of performance has to
be issued by a notified body prior to 2 METHODOLOGY
selling the product on the European
market. The harmonized standard for Some Member States independently
initial type testing and CE marking involved ANPI in their market
process is EN 14604:2005 + surveillance process. The sampling
AC:2008[2]. The surveillance process has been achieved by the
required by the CE-CPR scheme is an corresponding authorities within their
annual audit of the quality system of territory. Each country has decided
the factory. No profound inspection on a reduced test protocol based on
of the product is required. Beyond the EN 14604 standard. The choice
this mandatory CE marking process amongst all the requirements has
exist several quality marks across been made by an educated guess.
Europe whose application must The focus has been on the core
remain voluntary. These marks are functionality of the devices and on
supposed to be an added value to requirements for an appropriate use
the CE marking, distinguishing the of them:
most qualitative products on the battery removal indication
market. For all these marks one of (4.13), marking and data
the added requirements as compared (4.19), directional dependence
to the CE-CPR scheme is the (5.3),
inspection of the product itself: initial sensitivity (5.4),
sampling for comparison with the fire sensitivity (5.15), and
recorded technical file to detect sound output (5.17).
potential undeclared modifications
and periodical testing for some major The objectives and requirements of
requirements of the standard. these 6 clauses of the standard are
summarized here.
According to CPR requirements, the - battery removal indication: the
member states have to carry out a user has to be warned when trying
survey of the market by sampling to install his device without a battery
products and performing selected inside;
laboratory tests. ANPI laboratory has - marking and data: a set of
been asked by a number of countries indications must appear on the
to be involved in this process. The product (recommended date for
collected data provides a significant device replacement, recommended
amount of results reflecting the battery type, recommendation to test
reality of the market. the device after battery replacement,
etc.) and an accompanying
The first section of this paper documentation must for example
presents how the results have been guide the user during the installation
gathered and a short description of and the maintenance of the device;
the test objectives and requirements. - directional dependence: the
The second section describes the sensitivity to smoke must be the

Copyright all rights reserved ANPI 5 DOC ANPI 2014

same over 360 around the device quality marks are well known
(the dependence to the direction of European marks. The number of
arrival of the smoke must be products shown as not compliant
limited); with at least 1 requirement has been
- initial sensitivity: the calculated for both families of
manufacturer must be able to products. This provides an overview
calibrate all the produced samples to of the impact of the certification
the same sensitivity; during a normal scheme on the compliance of the
certification process the sensitivity of devices. Finally, a cross analysis
20 samples of the same model is combining these 2 axes has been
measured and all the results must be done. This cross analysis provides an
similar (the set has been reduced to image of the distribution of the non-
5 samples in the frame of the market compliances according to these 2
surveillance processes); orthogonal axes.
- fire sensitivity: the devices must
be able to produce an alarm when a
fire is developing before the amount 4 RESULTS STATISTICAL
of smoke exceeds a high level; ANALYSIS
4 types of fire are used: smouldering
pyrolysis wood fire (TF2), glowing 60 different references of smoke
smouldering cotton fire (TF3), alarm devices sampled by the
flaming plastics (polyurethane) fire Member States have been tested
(TF4) and flaming liquid (n-heptane) during the past 2 years in our
fire (TF5); this is of course to test laboratory. All of them include an
the major core functionality of the optical sensor. Products using the
device; ionization detection principle have
- sound output: when producing not been tested.
an alarm the sound level must be
sufficiently high but not too high and Requirement axis analysis results
the frequency of the sound must be
low enough.

ANPI laboratory is ISO 17025

accredited for all these tests.

3 PARAMETERS

The collected data have been

statistically analysed. The analysis
has been oriented along 2 orthogonal
(independent) axes. The first axis is
made up of the 6 different
requirements. The occurrence of the
non-compliances on each of these Fig. 1. Occurrence of the non-compliances for
each considered requirement and for the 6
requirements has been calculated. requirements altogether.
This provides an overview of the
flaws of the products available on the
market. The second axis is made of
the certification scheme claimed by
the supplier of the device on his
label. For one part of the products,
CE is the only certification scheme
claimed by the manufacturer. For the
other part of the products, voluntary
quality mark certificate is claimed in
addition to the CE. The encountered

Copyright all rights reserved ANPI 6 DOC ANPI 2014

 the same sensitivity by the
manufacturer. The sensitivity of the
products is not sufficiently isotropic
for 6% of the references. Finally, the
battery removal indication is not
satisfactory in 2% of the cases.

Fig. 2. Occurrence of the non-compliances for

each of the 4 type fires and for the 4 fires
altogether.

Fig. 3. Occurrence of the non-compliances for

The non-compliances for each of the each of the 2 sound parameters and for the 2
different considered requirements parameters altogether.
(first axis analysis) have been
counted. The results are expressed Certification scheme axis
as percentages in Figure 1. A first analysis results
result shows that 33% of the
products present at least one non- The second axis of this analysis
compliance. Most of these rely on investigates the occurrence of
the marking on the product or the unsatisfactory products as a function
instruction provided to the user. 19% of the claimed certification scheme.
of the products have a problem with About half of the references claim a
fire detection. voluntary mark certification while
those of the other half do not. The
The distribution of the non- occurrence of the non-compliances
compliances amongst the 4 types of for these 2 families of products is
fire is given in Figure 2. This depicted in Figure 4. It appears from
indicates a statistical gradual this analysis that 48% of the
difficulty of detection. TF5 (heptane products sold without any voluntary
flaming liquid fire) is the most mark above CE are not compliant
difficult fire to detect for an optical with at least one of the 6 considered
sensor. requirements. The result for the
other family is significantly different:
The third requirement giving a 21% of the products claiming a
relatively high occurrence of non- certificate of at least one voluntary
compliance (14%) concerns the mark are not compliant with at
alarm sound output. The distribution least one of the 6 considered
of the non-compliances amongst the requirements.
two sound parameters is shown in
Figure 3. The sound level is too low It has to be noted that 5% of the
for 8% of the sampled products. The references do not claim CE
sound frequency exceeds 3.5 kHz for certification at all. Two situations can
10% of the sampled products, but it be observed.
remains very close to this limit. 1. The product has been put on the
European market (has left the
10% of the sampled products are not factory) before August 1st 2008, the
compliant with the initial sensitivity date CE certification according to EN
requirements. It means that the five 14604 became mandatory, and is
samples with the same product still on the shelves.
reference have not been tuned to 2. The product is not legally put on
the market.

Copyright all rights reserved ANPI 7 DOC ANPI 2014

 Fig. 5. Occurrence of the non-compliances to
the 6 requirements, comparison for each of
the 2 kinds of certification schemes.

Figure 5 provides a quantitative

Fig. 4. Occurrence of the non-compliances for
each of the 2 kinds of certification schemes comparison of the two kinds of
and for the 2 kinds altogether. certification schemes. When paying
attention to the core functions of the
Full analysis results smoke alarm devices (fire sensitivity
and sound output), the voluntary
Counting the number of non- marks reduce by a factor of 3 the
compliances for each combination of occurrence of the non-compliances.
requirement and certification scheme
provides the qualitative map
depicted in Table 1 below. The disk 5 CONCLUSION AND
surface is proportional to the INTERPRETATION
occurrence of non-compliances for
each combination. The improvement When the European consumer buys a
offered by the voluntary marks is smoke alarm device which does not
obvious. claim any voluntary mark the risk to
get a product not complying with
4.13 Batt. Remov. Indic.
EN14604 is 1 in 2 and the risk to
purchase a product having problems
4.19 Marking and data
with fire detection is 3 in 10.
When buying a product claiming a
voluntary mark certification these
5.3 Direct. Depend.
risks become 1 in 5 and 1 in 10
respectively.
Considering the 6 requirements
5.4 Init. Senstiv.

presented in this paper, the risk

5.15 Fire sensitiv.
presented by the voluntary mark
certified products is reduced by a
5.17 Sound output
factor of 2 to 4 when compared to
CE-only products.
Global result

Models without Models with Even if the performance of the

All models voluntary mark voluntary mark voluntary marks is not 100% (0%
Table 1. Qualitative overview of the non-compliance), the improvement
distribution of the non-compliances for the 2
kinds of certification schemes. as compared to pure CE marking is
significant.

The reason for such a result probably

lies in the certification scheme
applied by these voluntary marks.
The scheme required by the CPR is
not able to detect small
modifications, having a strong impact
on the product reliability, which

Copyright all rights reserved ANPI 8 DOC ANPI 2014

have not been declared. The REFERENCES
voluntary marks compensate this
lack with a stronger market [1] "Regulation (EU) No 305/2011 of
surveillance scheme. the European Parliament and of the
Council of 9 March 2011 laying down
In this case study of smoke alarm harmonised conditions for the
devices, it can be observed that CE marketing of construction products
marking according to CPR is not a and repealing Council Directive
quality mark and the voluntary 89/106/EEC ", Official Journal of the
marks provide a significant added European Union, 2011, L 88, , April
value towards reliability and quality. 4, pp 5-43

[2] CEN TC72, "EN14604: Smoke

alarm devices", Brussels, Belgium,
July 2005 and October 2008.
.

Copyright all rights reserved ANPI 9 DOC ANPI 2014