An American National Standard

Designation: E 603 – 01

Standard Guide for

Room Fire Experiments1
This standard is issued under the fixed designation E 603; the number immediately following the designation indicates the year of
original adoption or, in the case of revision, the year of last revision. A number in parentheses indicates the year of last reapproval. A
superscript epsilon (e) indicates an editorial change since the last revision or reapproval.

INTRODUCTION

This guide has been written to assist those planning to conduct full-scale compartment fire
experiments. There are many issues that should be resolved before such an experimental program is
initiated, and this guide is written with the objective of identifying some of these issues and presenting
considerations that will affect each choice of procedure.
This guide deals with any or all stages of fire growth in a compartment. Whether it is a single- or
multi-room experiment, observations can be made from ignition to flashover or beyond full-room
involvement.
One major reason for conducting research on room fires is to learn about the room fire buildup
process so the results of standard fire test methods can be related to performance in full-scale room
fires, allowing the further refinement of these test methods or development of new ones.
Another reason concerns computer fire modeling. Full-scale tests can generate data needed for
modeling. Comparisons of modeling with full-scale test results can serve to validate the model.
The various results among room fire tests reflect different experimental conditions. The intent of this
guide is to identify these conditions and discuss their effects so meaningful comparisons can be made
among the room fire experiments conducted by various organizations.

1. Scope 1.4 The rationale for conducting room fire experiments

1.1 This guide addresses means of conducting full-scale fire according to this guide is shown in 1.5-1.8
experiments that evaluate the fire-test-response characteristics 1.5 Room fire experiments are a means of generating input
of materials, products, or assemblies under actual fire condi- data for computer fire models and for providing output data
tions. with which to compare modeling results.
1.2 It is intended as a guide for the design of the experiment 1.6 One of the major reasons for conducting room fire
and for the use and interpretation of its results. The guide is experiments is as an experimental means of assessing the
also useful for establishing laboratory conditions that simulate potential fire hazard associated with the use of a material or
a given set of fire conditions to the greatest extent possible. product in a particular application. This should be borne in
1.3 This guide allows users to obtain fire-test-response mind when designing nonstandard experiments.
characteristics of materials, products, or assemblies, which are 1.7 A rationale for conducting room fire experiments is the
useful data for describing or appraising their fire performance case when smaller-scale fire tests inadequately represent end-
under actual fire conditions. use applications.
1.3.1 The results of experiments conducted in accordance 1.8 A further rationale for conducting room fire experiments
with this guide are also useful elements for making regulatory is to verify the results obtained with smaller scale tests, to
decisions regarding fire safety requirements. The use for understand the scaling parameters for such tests.
regulatory purposes of data obtained from experiments con- 1.9 This standard is used to measure and describe the
ducted using this guide requires that certain conditions and response of materials, products, or assemblies to heat and
criteria be specified by the regulating authority. flame under controlled conditions, but does not by itself
incorporate all factors required for fire hazard or fire risk
assessment of the materials, products, or assemblies under
1
This guide is under the jurisdiction of ASTM Committee E05 on Fire Standards actual fire conditions
and is the direct responsibility of Subcommittee E05.13 on Large Scale Fire Tests.
Current edition approved October 10, 2001. Published November 2001. Origi-
nally published as E 603 – 77. Last previous edition E 603 – 98a.

Copyright © ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959, United States.

1
 E 603 – 01
1.10 This standard does not purport to address all of the ISO/IEC Guide 52 Glossary of Fire Terms and Definitions9
safety concerns, if any, associated with its use. It is the ISO 3261 Fire Tests—Vocabulary9
responsibility of the user of this standard to establish appro- ISO 9705 Fire Tests—Full Scale Room Fire Tests for
priate safety and health practices and determine the applica- Surface Products9
bility of regulatory limitations prior to use. 2.6 NFPA Standard:
NFPA 265 Methods of Fire Tests for Evaluating Room Fire
2. Referenced Documents Growth Contribution of Textile Wall Coverings10
2.1 ASTM Standards:
D 4442 Test Methods for Direct Moisture Content Measure- 3. Terminology
ment of Wood and Wood-Base Materials2 3.1 Definitions—For definitions of terms used in this guide
D 4444 Test Methods for Use and Calibration of Hand-Held and associated with fire issues, refer to the terminology
Moisture Meters2 contained in Terminology E 176, ISO/IEC Guide 52, and ISO
D 5424 Test Method for Smoke Obscuration of Insulating 3261. In case of conflict, the terminology in Terminology
Materials Contained in Electrical or Optical Fiber Cables E 176 shall prevail.
When Burning in a Vertical Cable Tray Configuration3 3.1.1 heat release rate, n—the heat evolved from the
D 5537 Test Method for Heat Release, Flame Spread and specimen, per unit of time.
Mass Loss Testing of Insulating Materials Contained in 3.1.2 oxygen consumption principle, n—the expression of
Electrical or Optical Fiber Cables When Burning in a the relationship between the mass of oxygen consumed during
Vertical Cable Tray Configuration3 combustion and the heat released.
E 176 Terminology of Fire Standards4 3.1.3 smoke obscuration, n—reduction of light transmission
E 800 Guide for Measurement of Gases Present or Gener- by smoke, as measured by light attenuation.
ated During Fires4 3.2 Definitions of Terms Specific to This Standard:
E 906 Test Method for Heat and Visible Smoke Release 3.2.1 full-scale test, n—a test in which the product(s) to be
Rates for Materials and Products4 tested is utilized in the same size as in its end use.
E 1321 Test Method for Determining Material Ignition and 3.2.1.1 Discussion—In practical applications, this term is
Flame Spread Properties4 usually applied to tests where the item to be tested is larger
E 1354 Test Method for Heat and Visible Smoke Release than would fit in a bench-scale test.
Rates for Materials and Products Using an Oxygen Con- 3.2.2 total heat released, n—integrated value of the rate of
sumption Calorimeter4 heat release, for a specified time period.
E 1355 Guide for Evaluating the Predictive Capability of
Deterministic Fire Models4 4. Summary of Guide
E 1537 Test Method for Fire Testing of Real Scale Uphol- 4.1 This guide does not define a standard room fire test. It
stered Furniture 4 does, however, set down many of the considerations for such a
E 1590 Test Method for Fire Testing of Mattresses4 test, for example, room size and shape, ventilation, specimen
Proposed ASTM Room Fire Test for Interior Finish Mate- description, ignition source, instrumentation, and safety con-
rials5 siderations that must be decided on in the design of a room fire
2.2 UL Standards: experiment. It discusses performance criteria for the particular
UL 1715 Room Corner Test6 array of finishing and furnishing products that comprise the
UL Subject 1040 Large Scale Open Corner Test6 room. The behavior of any particular product in the room
2.3 ICBO Standards: depends on the other products and materials present and how
Uniform Building Code Standard UBC 8-2 Standard Test they are arranged in relation to one another.
Method for Evaluating Room Fire Growth Contribution of 4.2 Whether a particular arrangement simulates the evalua-
Textile Wallcoverings7 tion desired depends on the size and location of the ignition
Uniform Building Code Standard UBC 26-3 Room Fire Test source. It is therefore important that the ignition source
Standard for Interior of Foam Plastic Systems7 simulate, insofar as possible, an initiating fire for the desired
2.4 FM Standard: scenario.
FM 4880 Large Scale Open Building Corner Test8 4.3 The main criterion suggested in this guide for evaluating
2.5 ISO Standards: fire performance is based on the time to flashover as indicated
by the time at which the radiation flux at the center of the floor
exceeds 20 kW/m2. Other suggested indicators of flashover
2
Annual Book of ASTM Standards, Vol 04.10. include an average upper air temperature in excess of 600°C
3
Annual Book of ASTM Standards, Vol 10.02.
4
Annual Book of ASTM Standards, Vol 04.07. and the ignition of a cotton indicator. Other possible perfor-
5
Discontinued—See 1982 Annual Book of ASTM Standards, Part 18, pp. mance criteria include the total amount or rate of smoke and
1618 – 1638.
6
Available from Underwriters Laboratories, Inc., 333 Pfingsten Rd., Northbrook,
IL 60062.
7 9
Available from International Conference of Building Officials, 5360 Workman Available from International Organization for Standardization, P.O. Box 56,
Mill Rd. Whittier, CA 90601. CH-1211, Geneva 20, Switzerland.
8 10
Available from Factory Mutual Research Corporation, 1151 Boston-Providence Available from National Fire Protection Association, Batterymarch Park,
Turnpike, P.O. Box 9102, Norwood, MA 02662. Quincy, MA 02269.

2
 E 603 – 01
heat production, extent of the flame spread for a low-energy 6. Experimental Choices
ignition source, and size of the primary ignition source 6.1 General—The complete program for any series of
required to produce flashover. full-scale compartment fire experiments usually involves many
4.3.1 Where multi-room experiments are being conducted, different considerations and possible simulations. This guide
flashover may not be an appropriate performance criteria. In reflects the current state of knowledge and suggests choices for
fact, the experiments may have to be conducted beyond geometry, ignition sources, and instrumentation.
flashover. Post-flashover is usually required in the test room in 6.2 Compartment Design:
order to observe high levels of toxic gases and smoke in remote 6.2.1 Ventilation:
rooms or flame spread in adjoining surface areas. Other 6.2.1.1 Experiments with ventilation-controlled fires in
performance criteria could be the levels of combustion prod- model rooms (1),11 where the fire has become large or reaches
ucts that impair visibility and cause incapacitation or lethality the point of flashover, show that the compartment geometry
in remote rooms. and dimension influence the burning rate. An important rela-
4.4 Primary ignition sources include gas burners, wood tionship is the following:
cribs, waste containers, and pools of liquid fuel. Waste con- ṁ 5 kA=H (1)
tainers and wood cribs have the advantage of presenting a solid
fuel fire with some feedback effects and a luminous flame that
where:
appears to simulate the burning of furniture. However, the gas ṁ = mass loss rate (kg/s),
burner is the best choice for most fire experiments because of A = area of the ventilation opening (m2),
its reproducibility. The placement of the ignition source de- H = height of the ventilation opening (m), and
pends on the desired effect on the target material. k = a proportionality constant, the value of which is
4.5 The instrumentation for measuring burning rate, heat approximately 0.09 kg/m5/2 s.
release rate, heat flux, temperature, upper layer depth, air This equation is an empirical relationship resulting from the
velocity, flame spread, smoke, and gas concentration is dis- classic ventilation-controlled wood crib fires that Kawagoe (2)
cussed, along with suggested locations. A minimum level of studied. Other experiments by Hagglund (3) reveal that flash-
instrumentation is also suggested. over was not observed for A=H below 0.8 m5/2. Hagglund
4.6 A typical compartment size is 2.4 by 3.7 m (8 by 12 ft), conducted experiments on wood cribs in a compartment
with a 2.4-m (8-ft) high ceiling. A standard-size doorway (0.80 measuring 2.9 by 3.75 by 3.7-m high. These studies suggest
by 2.0-m high) should be located in one wall, probably in one that a limiting burning rate that depends on the ventilation must
of the shorter ones. The top of the doorway should be at least be exceeded before flashover occurs. The correlation is useful
0.4 m (16 in.) down from the ceiling to partially contain smoke as a guideline for the occurrence of flashover.
and hot gases. 6.2.1.2 However, later studies show that the rate of burning
becomes independent of ventilation at flashover. Also, a single
4.7 Insofar as possible, the construction details of the wall
item with a large enough burning rate can induce flashover.
and ceiling, as well as any enclosed insulation, should dupli-
Among other parameters, ventilation plays an important role in
cate the room being simulated. Boundary surfaces that do not
fire severity. Drysdale (4) explores many of these parameters in
form the specimen should also be constructed of materials detail.
consistent with the room being simulated (see 6.2.3). 6.2.1.3 Ventilation should be continuous in a multi-room
4.8 The safety of observers and the crew extinguishing the test facility. The doors may be either open or partially closed.
fire is emphasized strongly in this guide. One can install a typical heating ventilation and air condition-
4.9 The analysis of data should include a comparison of the ing (HVAC) duct system if the compartments are closed.
critical times, heat fluxes, temperatures, heat release rate, and 6.2.2 Size and Shape of Compartment:
smoke generation in the room with ignition, flame spread, and 6.2.2.1 The geometry of the compartment in conjunction
smoke properties of the specimen materials. This would aid in with the thermal properties of the wall and ceiling materials has
the development or modification of small-scale tests and would substantial influence on the behavior of a confined fire,
provide useful information for assisting in the development of particularly by affecting flow patterns, and hence the mixing
analytical room fire models. and combustion characteristics of the fire. Thus, the compart-
ment size, shape, and openings should be chosen to simulate
5. Significance and Use the nature or type of compartment or facility in which the
subject material, product, or system is expected to be used in
5.1 This guide provides assistance for planning room fire actual service. If there is a range of sizes, account should be
tests. The object of each experiment is to evaluate the role of taken of the fact that for a given ignition exposure, the smaller
a material, product, or system in the fire growth within one or compartment sizes will usually provide the most severe fire
more compartments. development conditions. However, it has been found that room
5.2 The relationship between laboratory fire test methods size (if the floor area lies between 8.7 and 11.4 m2 and one of
and actual room fires can be investigated by the use of
full-scale and reduced-scale experiments. This guide is aimed
at establishing a basis for conducting full-scale experiments for 11
The boldface numbers in parentheses refer to the list of references at the end
the study of room fire growth. of this guide.

3
 E 603 – 01
the room floor dimensions is between 2.4 and 3.7 m) has little lost to the bounding surfaces and transfer of thermal energy
effect on heat development if the heat release rate is below 600 due to the net flow of hot gas from the room through natural
kW (5). The compartment should preferably be designed to be ventilation or forced ventilation systems. Heat transfer to a
symmetrical and as simple as possible for ease of analysis. The bounding surface in the presence of flames occurs mainly by
proposed ASTM Room Fire Test for Interior Finish Materials is radiation and convection. The amount of radiant energy im-
based on a 2.4 by 3.7-m (8 by 12-ft) room with a 2.4-m (8-ft) pinging on a surface depends on the radiative properties of the
high ceiling. It has one standard-size doorway left fully open. exposure fire and of the surrounding surfaces. The convective
The space between the top of the door and the ceiling is critical heat transfer rate is determined by the geometry of the
because of the trapping of smoke and hot gases. It is 0.4 m (16 bounding surface and the magnitude and turbulence associated
in.) in the proposed ASTM room. The room dimensions may be with the gas flow in the compartment. Heat transfer, which
chosen to simulate some particular applications. However, if affects the magnitude of heat flux acting on the bounding
there are no constraints, it would probably be better to remain surface, is related directly or indirectly to both the size and
within the dimensions of the proposed ASTM room for shape of the compartment involved even though radiative
possible comparison with other single compartment tests. Also, properties of the materials contained in bounding surfaces are
the proposed ASTM room is already setup and instrumented in unrelated to geometrical issues. Consequently, the geometry,
many commercial testing laboratories. The room should be thermal and radiative properties, and degradation characteris-
located inside a larger, carefully ventilated enclosure to ensure tics of the compartment surfaces should be considered care-
minimum interference from drafts or wind currents. Ref (6) fully when conducting compartment fire experiments.
shows how doorway size and room geometry affect fire 6.2.3.2 The thermal inertia (product of thermal conductivity,
growth. In order to measure many of the properties that are density, and heat capacity, krc) of the materials forming the
required from room-sized tests, a canopy hood and exhaust linings of a fire compartment (bounding materials) directly
duct are required. These are usually placed either in the room affects their surface temperature, and its corresponding rise, the
itself, or more commonly, just outside the doorway (see Fig. 1). rate of heat dissipated into the internal surface, and the room
6.2.2.2 Examples of tests that use full rooms are Test gas temperature. The influence of the wall materials on the
Methods D 5424, D 5537, E 1537, and E 1590, and UL 1715, temperature distribution in the gas is also a function of the
UBC 8-2 and 26-3, and ISO 9705, as well as the proposed radiative properties of the gas and the gas velocity. Relevant
ASTM room test. nondimensional parameters which account for this coupled
6.2.2.3 In a multi-room test, it is critical to duplicate the size interaction have been published (9). If the thermal inertia is
and location of corridors and remote rooms. If flame spread low (good insulation), the surface temperature rises more
along walls is being observed, it may not matter if the corridor rapidly, the rate of heat transfer decreases, and the radiation
has a closed end; it does matter when the flame spread on the emitted from the upper walls and ceiling to both the fire itself
floor is important. It has been shown that closing the corridor
and the lower part of the compartment increases. The emissive
has very important effects on gas flow and decay of gases (7,
power of surfaces and their temperatures are coupled through
8).
the radiative transfer equation. Bounding surfaces consisting of
6.2.3 Thermal and Radiative Properties of Compartment
materials with good insulating properties will produce substan-
Linings:
tially higher gas temperatures in the room than when poor
6.2.3.1 The fire gas temperature and heat flux levels in the
insulators are used for lining the enclosed space. The effect of
fire compartment depend on the heat balance of the compart-
compartment thermal properties on the time-temperature curve
ment (heat released during the combustion process and heat
has been analyzed mathematically in the post-flashover regime
with numerical methods (10-12). Full-scale studies demon-
strate the effect of compartment wall properties on the fire
intensity (13-15). Typical thermal property values of some
samples of common materials are given in Table 1 (16) as
guidance.
6.2.3.3 The radiative characteristics of the bounding sur-
faces influence the compartment gas temperatures, particularly
during the pre-flashover stages of compartment fires, but this
effect decreases with time (10). Bounding surfaces having a
greater absorptivity result in a lower gas temperature in the fire
compartment. However, the surface absorptivity effect is pro-
nounced when good thermal conducting materials are used on
the walls, ceiling, and floor and is of minor practical impor-
tance for the compartment lined with high-insulation materials.
6.2.3.4 Since the severity of a fire in its early stages will
depend on the heat exchange with the bounding surfaces of the
room, it is important that construction details, such as the
wallboard thickness, type, size, and spacing of the studs and
FIG. 1 Canopy Hood and Exhaust Duct joists, and insulation, if any, in the wall and ceiling cavities, be

4
 E 603 – 01
representative of the construction that is being simulated. For Method E 906, or the LIFT apparatus, Test Method E 1321,
those areas of the interior surface not being tested, a suitable reflect various aspects of the fire performance in a room fire.
inert material may be a ceramic fiberboard that has thermal Data such as heat release, smoke release, ignitability, flame
properties similar to those of gypsum board. (Tran and Jans- spread, etc. may assist in interpretation of the results of the
sens (15) have demonstrated that ceramic fiberboard is a very room fire experiment. The ignition times, flame spread distance
good insulator and can increase the severity of the test.) and rate, and heat release rates depend on many factors, such
Gypsum and ceramic fiberboard give different results, and the as the incident heat flux on the specimen and the type of flame.
results must not be intermixed. Gypsum is the material of Hence, the exposure conditions during the room fire experi-
choice for normal tests. ments should be described. If possible, the bench-scale fire
6.2.3.5 During the course of a compartment fire experiment, tests should be performed on specimens that have the same
the disintegration or cracking, if any, of the materials lining the thickness as the material used in the room temperature for
compartment will affect the behavior of the confined fire. thicknesses up to 50 mm (2 in.).
Vertical pressure gradients developed in the presence of the fire 6.3.2.4 The location of items of furniture in terms of their
will cause smoke and hot gases to leak to the outside and cool distance from the wall, corner, and other furniture items should
air to be drawn into the compartment through the cracks in the be identified in terms of their distance from the different walls,
compartment walls or specimens. corners, and any other furniture items specified. For each
6.3 Specimens: furniture item to be tested, the horizontal and vertical exposed
6.3.1 General: areas, total weight, and moisture content should also be
6.3.1.1 In the room fire experiment, all of the combustible described. It would also be helpful to indicate the material
products in the room can be considered to be part of the composition, if known. The test standards addressing specific
specimen. When some of these products are combined to form items, such as Test Method E 1537, for upholstered furniture or
an item of furnishing or a wall, the combination becomes the Test Method E 1590, for mattresses, give details of the loca-
specimen. In fact, the walls, ceiling, floor, and all of the tions to be used.
furnishings constitute a configured specimen whose properties 6.3.2.5 The ambient temperature and humidity of the room
include the physical and chemical properties of the items and and the time these conditions have been maintained prior to the
their location. experiment should be recorded.
6.3.1.2 The following paragraphs deal with recommenda- 6.3.3 Selection—The choice of the specimen is based on the
tions for the description and selection of specimens for the objective of the room fire experiment, which may be one of
room fire experiments to ensure that the important variables three types: (1) a demonstration experiment, (2) a comparison
will be considered, and to provide a basis of comparison of theory and experiment, or (3) a determination of the fire
between experiments conducted at different laboratories. performance of a particular product.
6.3.2 Description—As much information as possible should 6.3.3.1 In the demonstration experiment, the room should
be secured and reported for the materials, products, and be finished and furnished in the most realistic way possible.
assemblies in order to provide the necessary information on the Observations and measurements should be aimed at uncover-
room fire specimen. Along with a description of the ventilation ing the important phenomena involved in the simulated room
conditions and ignition source, the data are intended to provide fire and at establishing possible levels of temperature, gas
the input necessary to estimate the degree of involvement of concentration, and times of occurrence, etc.
the various combustibles and the maximum rise in the upper air 6.3.3.2 In the second type of room fire experiment, the
temperature that could potentially be attained. emphasis is on the ease of description so that calculated values
6.3.2.1 The specimen should be divided into components can be checked against the experimental results. The number of
classified either as finishing materials, wall and floor coverings, products in any given experiment should be minimized for
or furniture. simplicity of description. However, products covering a large
6.3.2.2 The location of the material, product, or assembly to range of properties should be selected for the tests so that the
be tested as a lining should be specified as in one or more of the prediction formulas developed do not have limited applicabil-
following zones: (1) ceiling, (2) upper half of wall, (3) lower ity.
half of wall, (4) floor, or (5) fraction of a zone, for screening 6.3.3.3 In the third type of experiment, to evaluate fire
purposes. Both combustible and noncombustible components performance, the location of the comparison product in the
are to be taken into account. The test standards addressing room should be based on its intended use (that is, a ceiling,
specific items, such as NFPA 265, give details of the locations wall, floor, wall covering, or item of furniture). Because of
to be used. heat-trapping effects, the ceiling material should cover the
6.3.2.3 The chemical composition, generic or brand name of complete room ceiling. While it may not be necessary to cover
the lining material, and any involved adhesive interfaces, the entire wall area with the wall product, the area covered by
description of exposed area, thickness, density, moisture con- a wall product must be large enough to contain all wall areas
tent, and fire properties of each component should be detailed. exposed during the experiment and extend beyond the end of
If possible, the thermal conductivity and specific heat should any expected flame spread. In general, other materials in the
also be listed. Some fundamental fire properties of the material room should be noncombustible, or at least of low heat release,
as determined by accepted test methods such as the cone and should remain the same from experiment to experiment.
calorimeter, Test Method E 1354, the OSU calorimeter, Test Because of its widespread use and low heat release, gypsum

5
 E 603 – 01
board is often used, but the board must be replaced between poured onto items of furnishings. The size is strongly depen-
experiments in those areas in which it was exposed to fire. An dent on the degree of fire buildup required for the experiment
alternative is ceramic fiberboard. and the combustibility of the materials used in the experiment.
6.3.3.4 The experimenter may occasionally want to evaluate When choosing an ignition source for a particular experiment,
the outcome of the most severe ignition source and product the characteristics of the product to be tested (size and heat
orientations. It would be prudent for a caveat to be added to the production capability) should be taken into account, so as to
conclusions of the experimental report stating that other make a reasonable selection.
ignition source strengths and material orientations were not 6.4.2.1 Gas burner flames have the following characteris-
considered and therefore could not be evaluated on the basis of tics: (1) they are reproducible; (2) they are well-defined (that is,
the subject experiments. their heat production rate is determined readily from the gas
6.3.3.5 Unless special considerations apply, the relative flow rates); (3) they can be varied with time to represent the
sizes of the product to be tested and of the ignition source burning of different items of furniture or be maintained
should be such that only a fraction of the product to be tested constant to facilitate analytical studies; (4) their burning rates
should be consumed, if the product to be tested has good are not influenced by heat feedback (unless controlled artifi-
enough fire performance. cially); (5) the radiation properties of the flames are different
6.3.4 General Considerations: than those of the product simulated; and (6) gas flames do not
6.3.4.1 The distinction between materials located on the resemble what is seen in real fires.
upper and lower walls is made because heat conduction losses 6.4.2.2 Differences between diffusion and premixed burners
occur primarily through the upper walls and ceiling. Increasing
should be recognized. For example, the flames from a pre-
the insulation in these areas increases the rate of temperature
mixed burner will be shorter and have lower emissivities. In
rise in the room and the maximum temperature that will be
order to avoid locally high velocities, the gas can be delivered
reached.
through a large-area diffusing surface, such as a porous plate or
6.3.4.2 The spacings between the items of furniture, along
a layer of sand.
with the ignitability of the furniture, determine the probability
and time of flame spread between them. When two or more 6.4.2.3 Liquid fuel pool fires have the following character-
items of furniture are burning, their separation distance deter- istics: (1) their rate of fuel production is determined readily
mines whether the flames will merge. Furthermore, the heat from their rate of mass loss or the flow rate necessary to
transfer between them will enhance their separate burning rates maintain a constant depth in the pool; (2) they have an
so that larger flames will result. The proximity of the burning interaction with the fire environment that can be quantified by
item of furniture to the wall and corner causes an increase in their change in heat production rate; (3) they are reproducible
flame height with an attendant increase in air temperature and under the same exposure conditions; (4) their radiation char-
the probability of the flame jumping between the item and the acteristics can be controlled by the choice of fuel; (5) the effect
wall. of feedback is not quantitatively the same as that for furnish-
6.3.4.3 In addition to its toxic effect and visibility problems, ings; and (6) they lack visual realism unless they are intended
smoke is a factor in the heat radiative exchange between the to represent liquid fuel spills. A variation of the liquid pool fire
upper and lower portions of the room. The height of the is obtained by supplying the liquid fuel in a matrix of sand in
furniture items or wall covering material will determine the order to vary its burning rate.
probability of their ignition by the hot air layer in the upper part 6.4.2.4 The solid fuels that have been used as ignition
of the room. Horizontal and vertical surface areas are therefore sources for room fire experiments have included primarily
specified separately because of the difference in heat transfer waste containers and wood cribs, with the latter having the
from flames to surfaces with these orientations. These differ- longest history. Stick size, type of wood and spacing, as well as
ences lead to different heat release rates and flame spread total mass have a large effect on the burning rate of the wood
characteristics. cribs. The use of the above two types of solid fuels is
6.4 Ignition Sources: emphasized in this guide because they have been used the most
6.4.1 General—The choice of a primary ignition source in a up to the present time. However, the reproducibility and
compartment fire experiment is a critical item. This guide precisely known heat output of a gas burner makes it a likely
presents a list of the important considerations for the choice. candidate for replacement of the cribs and waste containers for
There will always be compromises on the size, location, type of standard room fire experiments when detailed heat balances
fuel, time of burning, type of burning, and other factors. This must be obtained from the experiments. Waste container and
discussion will present some of the important considerations wood crib fires have the following advantages: (1) they provide
and various choices that can be made. the best visual simulation of the burning of furniture; (2) their
6.4.2 Type and Size—The complete character of the ignition interaction with the environment of the fire room is perhaps
source should be determined, including weight, material iden- closer to, though not the same as, that of the burning furniture;
tification, morphology, dimensions, and all other physical and and (3) their radiation characteristics more nearly match those
chemical characteristics that are necessary to repeat each of the furniture fire. Waste containers and wood cribs have the
ignition scenario. Typical ignition sources may be solid, liquid, following disadvantages: (1) their reproducibility is not as
or gaseous fuels and include wastebaskets, furniture items, good as that of gas burners and (2) the ratio of their heat release
wood cribs, gas burners, liquid pool fires, and liquid fuels rates to their measured mass loss rates vary throughout the test.

6
 E 603 – 01
6.4.2.5 Both the National Institute of Standards and Tech- 2.4 by 3.7-m (8 by 12-ft) enclosure with a 2.4-m (8-ft) ceiling
nology (NIST, formerly the National Bureau of Standards and a standard-size door opening, a 1.4-kg (3-lb) waste
(NBS)) and the University of California, Berkeley, laboratories container located in one corner was sufficient to cause full
have used plastic waste containers as ignition sources in involvement of the room when it was lined with a high flame
compartment fire experiments (17, 18). The combustibles spread foam, but not with low flame spread foams. An
within these waste containers have been plastic-coated paper overstuffed chair with cotton padding was found to have a
milk cartons, paper tissues, carbon paper, paper towels, or kraft burning rate equivalent to a 6.4-kg (14-lb) crib (18). In that
wrapping paper, or some combination thereof. Plasticized same study, when two 1.8 by 2.4-m (4 by 8-ft) panels of the
paper milk cartons make a relatively intense fire, as shown by specimen material formed one corner of a 3.0 by 3.0-m (10 by
burning rate, plume temperature, and heat flux. The milk 10-ft) experimental room, the 6.4-kg crib located in the same
cartons represent a combination of a cellulosic and a corner was able to cause flashover with plywood and particle
hydrocarbon-based polymeric material with a high surface-to- board but not with the wood fiber insulating board. A report by
volume ratio comparable to the contents of a typical waste Quintiere and McCaffrey (26) illustrates the effect of various
container in an American home. cribs and ventilation sizes on fire intensity.
6.4.2.6 If an ignition source is kept small, so that it does not 6.4.2.10 Relatively large-dimension open corner room con-
cause flashover by itself, it can then be used to determine the figurations, consisting of one long wall, one short or end wall,
effect of furniture or wall, ceiling, or floor covering on fire and an included horizontally oriented ceiling, have been
development in that compartment. The maximum size of an standardized and codified as proprietary test methods by
ignition source that should be chosen is thus dependent on the Underwriters’ Laboratories and by Factory Mutual Research
size, shape, and ventilation of the compartment as well as the Corporation (UL 1040 and FM 4880).
location and burning characteristics of the ignition source 6.4.3 Location—The location of the ignition source is one
itself. The size of the ignition source also depends on the of the most important considerations when conducting com-
scenario to be investigated. partment experiments. Fig. 2 shows schematically how the
6.4.2.7 It has been determined that the rate of heat release in flame height from a given ignition source increases when
a ventilation-controlled fire is proportional to A=H (6.2.1). placed against a wall and in the corner. Also, the flame height
For a typical fire scenario, the ignition source heat release rate is strongly dependent on the proximity of the ignition source to
should be less than 15 % of that estimated to produce flashover each wall. The distance from the wall might be set at 25 or 50
in the burn room. The size of the ignition source should not mm (1 or 2 in.). See Babrauskas (27) and Thomas, et al. (28)
repress the contribution of the product that is being tested. for more detail.
When using gas burners, or a pool fire, the flow rate can be 6.4.3.1 The simplest model of combustion of an ignition
adjusted so that it does not cause flashover by itself. Other source is given by Lie (29), wherein he notes that, although the
items used as ignition sources, such as furniture, can be tested combustion process is determined by a large number of
in calorimeters to determine the heat release rate prior to actual parameters, all of these parameters are related to the three
testing. essential elements of fuel, heat, and air. The estimated height of
6.4.2.8 Ignition sources are characterized by the following flames has been the subject of many studies (30-33), but in
categories: (1) total fuel content; (2) type of fuel content; (3) general it can be simplified to a relationship in which flame
rate of fuel release as a function of time; (4) rate of heat release height is governed by the entrainment of air into the flame
as a function of time; (5) height of flame for given position plume. If the access of air to the flame is blocked from one
(that is, corners, wall, etc.); (6) direct use of convective and side, such as would occur by placing the ignition source against
radiative heat flux; and (7) time of burning. These character- a wall, then one would expect a higher flame for the same rate
istics can be determined for a variety of ignition sources, and of gaseous fuel leaving the source. This analogy can be
the compartment experiment can be initiated with the appro- extended further to an ignition source in a corner in which the
priate source. Then, if a given ignition source does not lead to two walls block air access from two sides. This gives the
full room involvement with a given wall lining or to burning in longest flame extension compared to either the free-burning
a piece of furniture, when the intent is to determine the source or that against a wall. Because of this result, the normal
threshold size of the ignition source required to produce practice is to make the corner the standard location for the
flashover, the intensity of the ignition source can be increased ignition source for the room fire experiments when the lining
for the next experiment. Typical heat release rates as a function material is intended to become involved first. The ignition
of time for larger sources have been reported (19, 20). For source should be placed directly in contact with an item of
more data, the user should refer to Gross (21), Babrauskas, et furnishing if that is to be involved first. In the case of a liquid
al. (22), Holmlund (23), and Ahonen, et al. (24). fuel ignition source, it may be desirable to pour the fuel
6.4.2.9 The designer of any experiment would be wise to directly onto the item of furnishing.
explore the effect of a variety of ignition sources in the 6.4.3.2 The imposition of a ceiling on the flame plume of an
experimental arrangement. A 9.1-kg (20-lb) wood crib might ignition source has a very special effect on the combustion of
cause full room involvement in a very small compartment, the fuel. Entrainment of ambient air into the fire plume is
while a 22.7-kg (50-lb) crib would be necessary for a larger decreased sharply when the plume turns the corner and
compartment lined with identical material. In one set of becomes a ceiling jet. Such decreased entrainment leads to an
Underwriters’ Laboratories full-scale room burns (25) using a increase in flame lengths since more flame surface is necessary

7
 E 603 – 01

(a) In the center of the test area (b) At one wall (c) At the corner
FIG. 2 Schematic Diagrams of Ignition Sources (Note the change of flame height shown)

to consume the fuel vapor delivered by the plume. This special not contribute significantly to a fire. These experiments are
feature of ceiling jets has been discussed by Alpert (34, 35). conducted with the same ignition source, and the following
The net result of the interaction of the flame plume and the parameters should be recorded as a function of time: ceiling
ceiling is shown schematically in Fig. 3, in which the flame temperatures, mass loss rate, heat flux as a function of distance,
plume is represented as having a height L2 above the ceiling and observed flame height.
line in the absence of a ceiling, but spreading a distance L1 6.5 Instrumentation:
under a ceiling. It has been noted by P. H. Thomas that the ratio 6.5.1 Calibration:
of L1 to L2 may be as large as 6 or 7:1, but this does not appear 6.5.1.1 Instruments must be calibrated carefully with stan-
to have been measured systematically for a range of fuels, dard sources both before and after the room fire test. Among
ceiling materials, and boundary configurations. In any event, these are the load cells or weighing platforms, heat flux gages
the use of ignition sources that produce flame heights substan- or radiometers, smoke meters, flow or velocity transducers, gas
tially higher than the compartment ceiling may lead to flash- burners, and gas composition analyzers. Most such calibrations
over and sustained involvement with only minimal combustion can be conducted before the instrument is installed. Small,
of the specimen material. portable wind tunnels have been used successfully to calibrate
6.4.4 Burning Characteristics—A description should be anemometers. A portable black body is suitable only for
given of calibration experiments performed with linings that do calibrating narrow-angle heat flux transducers. Most of the heat

FIG. 3 Flame Plume Under a Ceiling

8
 E 603 – 01
flux transducers used in room tests are of wide-angle type. vective fraction as a result of calibration constant changes.
Methods have been developed at NIST (36) and at the Swedish Additional information is contained in the literature (40-42).
National Institute (37) for calibrating wide-angle heat flux 6.5.3.2 The importance of frequent cleaning, blackening,
transducers. and recalibration should be stressed. The buildup of deposits
6.5.1.2 Even if the instrumentation has been calibrated, on the foil of the Gardon gage will reduce sensitivity, and,
reliability can still be a problem. A small portion of the unfortunately, this can occur during an experiment. The
full-scale room should be set up and instrumented successively Schmidt-Boelter gage is less affected by deposits, generally
with the various transducers for this reason. A reliability check more rugged, and somewhat more accurate because of the
could be obtained by using gas burners in such a full-scale smaller surface temperature rise for a given flux. (This tem-
segment. perature is measured with a thermopile rather than a thermo-
6.5.1.3 Quantities such as heat release, temperature, and couple.)
heat flux cannot be measured with high precision in a fire 6.5.3.3 A better resolution of the source of the incident flux
environment. It is important not to report them with excessive is obtainable from narrow-angle radiometers mounted outside
significant figures. the enclosure. These measurements can be compared directly
6.5.2 Heat Release Rate—It is usually extremely important with those from wide-angle heat flux gages by estimating
in a room-scale test to know the heat release rate of the fire overall radiating surface areas. Wide-angle (up to 180°) radi-
throughout the experiment. Before the fire has spread signifi- ometers with air-cooled windows are nearly impossible to
cantly beyond the initial item, load cell or platform transducers construct; those obtainable have effective view angles of less
can be used to determine the mass-time history of the primary than 120°, making it difficult to interpret the signal.
burning specimen. Mass loss rate measurements cannot be 6.5.3.4 Spectral and angular effects on the transmission of
converted directly to heat release rates due to the unknown heat the radiometer window, soot deposits on the window, and
of combustion of the volatiles and the unknown completeness reradiation from the heated window to the foil make radiom-
of the combustion reaction. Heat release rates are normally eters unsuitable for use in room fire experiments. Another
determined, fairly accurately and continuously, throughout a method of heat flux measurement is obtained by imbedding one
test by measuring the oxygen concentration and calculating the or more fine wire thermocouples, often with one at the exposed
heat release rate by using the oxygen consumption principle face. Numerical solution of the transient thermal conduction
(38). This requires the installation of a hood and an exhaust equations or the inverse heat conduction equations for the
duct for collecting all of the combustion products leaving the measured surface temperature time history then yields the net
fire room. Moreover, it requires measurement of the oxygen heat flux to the exposed surface. Temperatures at several depths
concentration, differential pressure in the duct, and temperature within the material can be used to double-check the numerical
in the duct (the latter ones to determine mass flow rate in the result. What makes this method of heat flux measurement
duct). Increased accuracy of heat release measurements is particularly desirable is that portions of the furniture, walls,
obtained if carbon monoxide, carbon dioxide, and water are and ceiling in the room satisfy all of the conditions necessary
also measured in the exhaust duct. All exhaust duct measure- for solution of the conduction equations. Imbedded thermo-
ments should be made at a location at which the stream is well couples at several wall and ceiling locations, especially near
mixed. This will occur at approximately six to ten pipe the ignition point, are also desirable as a means of determining
diameters downstream of the exhaust duct (see 6.6). The gas the fire energy losses other than by convection and radiation
concentrations, along with the mass flow rate in the duct, can through room openings.
be used to calculate the heat release rate as described in Ref 6.5.3.5 There are four main areas in which sets of heat flux
(38). Further details can be found in the corresponding test gages should be located during a compartment fire experiment.
methods and in a textbook addressing the subject of heat One location is as close as possible to the product or specimen
release in fires (39). initially ignited. Such a heat flux measurement will enable the
6.5.3 Heat Flux: radiative environment of the burning fuel to be determined, and
6.5.3.1 While knowing the total energy output of the fire is this information is useful for evaluating the flammability of
useful for evaluating a product’s performance, determining the materials. A second heat flux gage location should be at any
distribution of energy flux within the compartment is necessary fuel specimen likely to become involved as a result of the
in order to explain how fire spread occurs. The heat flux gages gradual fire spread to contiguous or nearby materials. This
used in room fire tests should measure the total flux over a 2p measurement will be useful for evaluating the ignitability of
solid angle. Water-cooled Gardon-type and Schmidt-Boelter- fuel in the fire environment. Heat flux gages more remote from
type gages with black receiving surfaces are by far the most the primary fire but still within the compartment constitute a
reliable and accurate of all flux transducers. With this type of third group, which can be used to determine when general
gage, the flux incident on a surface from all sources can be room involvement occurs. Such measurements should be made
measured. Reradiative or convective heat transfer from a near the floor level. Finally, heat flux gages mounted outside
surface near the gage can be estimated from surface tempera- the compartment and viewing door or window openings will
ture measurements. Thus the calculation of the net flux to the respond when the compartment fire becomes an active threat to
surface is allowed. Caution must be exercised, when using other building areas.
Gardon-type gages, to make measurements with a large con- 6.5.4 Temperature:

9
 E 603 – 01
6.5.4.1 Gas temperature measurements at locations through- Benedict (43) and Ower and Pankhurst (44) are excellent
out the compartment are obtained easily by the installation of sources of information on temperature, pressure, and air flow
fine wire (preferably 0.31-mm (0.005-in.) diameter), exposed- measurements. Ducted air supplies or returns need to be
bead thermocouples. Such measurements will be relatively monitored in the same way as other flows. Pressure differen-
accurate and useful for thermocouple locations in the moving tials should also be measured between the compartment, air
smoke layer both within the compartment and in the openings. ducts, and exterior side of the compartment.
Radiation errors can be significant for thermocouples in view 6.5.5.1 The traditional methods of measuring the air veloc-
of flames or within optically thin flame zones. The magnitude ity used to be the hot wire or hot foil thermoanemometer for
of the radiation error can be determined by measuring the low-velocity cold air inflow in the lower part of the doorway
temperature with different size wires or welded junctions and and the pitot tube for the higher-velocity high-temperature
extrapolating to zero diameter wire or junction to determine the outflow near the top of the doorway. Two problems that must
true temperature. One method for reducing these radiation be considered are the effects of thermal radiation on the
errors is by the use of aspirated thermocouples (that is, thermoanemometer and sooting of the pitot tubes.
thermocouples recessed slightly from one end of a tube, the
6.5.6 Fire Propagation:
other end of which is connected to a pump drawing approxi-
mately 8 dm3/min). The resulting air-flow over the thermo- 6.5.6.1 While the grid of gas, surface, and imbedded ther-
couple bead, approximately 5 m/s, is sufficiently high to allow mocouples will provide a detailed plot of pyrolysis and
accurate temperature measurement based on the thermocouple probably reaction zones, a photographic record of the fire is
voltage alone, even within flame zones. Because of the essential to establish flame spread rates and burning area
difficulty associated with their installation and maintenance clearly. In general, tests should be videotaped. The usefulness
throughout the experiment, as well as their expense, aspirated of the photographic record depends primarily on an accurate
thermocouples would normally be limited to critical locations. time synchronization with all other experiment measurements.
Errors are also introduced when the number of connections For this purpose, one or more clocks synchronized with the
increases. The number of thermocouple connectors should be data acquisition time system should be clearly in view of all
kept to a minimum. cameras or integrated in the photographic or video system.
Digital clocks are generally easier to read in photographs and
6.5.4.2 Thermocouples should be placed at vertical intervals
movies. Also, a length-scale grid on one or more walls and
of 76 to 152 mm (3 to 6 in.) at the following locations: (1)
doorway opening should be used. Resolution may be better
within the fire plume of the ignition source, (2) near the center
of the room, and (3) in any door or window opening. Individual with appropriate still photography.
thermocouples should also be located adjacent to all air 6.5.6.2 Photographic records of the compartment fire can be
velocity probes, optical density light paths, and heat flux gages. supplemented by observation with audio recorders running
6.5.4.3 When measuring a material surface temperature, the continuously. A written listing of observations with time should
thermocouples should be placed with junctions in contact with be provided in the report of the experiment.
the surfaces. The wires on either side of the junction should lie 6.5.7 Smoke:
in contact with the surface for a length of at least 5 diameters 6.5.7.1 The generation of rate of smoke during the fire is
in order to reduce heat conduction losses along the wires. In very important. The time history provides information on the
some cases, it may be appropriate to use a bonding material, in rate of smoke buildup and therefore allows the average smoke
which case the bonding material should be specified as well as concentration to be calculated for the compartments commu-
its method of application. However, the bonding material will nicating with the fire area. The measurements are made by
affect the surface temperature to some extent. If possible, it is determining the attenuation of a light beam as it travels from
better to drill two tiny holes 10-mm apart normal to the the source to the target. A collimated light source and a
surfaces. The junction would be located midway between the photometer placed directly opposite to it is a typical arrange-
holes. The lead wires would be threaded through the holes and ment. In the present state-of-the-art, the concentration of
pulled tight. smoke is best measured by the extinction coefficient (45) using
6.5.5 Air Velocity—Air velocity is normally determined by a laser source. A design of such a system, now commercially
means of the bidirectional flow probe. This probe is less available, is shown in Fig. 4. Careful alignment between
affected by high temperature sensitivity problems and fouling transmitting and detecting optics is necessary for optimal
from soot deposition than older techniques, such as pilot tubes. operation. The smoke generation rate can be obtained by
The probe basically consists of two large pitot tubes facing measuring the extinction coefficient and flow of smoke. Pos-
opposite directions. An electronic (variable capacitance) ma- sible locations of smoke measurements are the exhaust duct,
nometer connected to the probe then allows an accurate vertical or horizontal within the room, and any door or window
determination of flow velocity magnitude and of whether there opening. At compartment openings, smoke measurements
is an inflow or outflow. While the probe itself is inexpensive, should be conducted at multiple intervals so that the measure-
each pressure measurement channel is rather expensive. The ment errors due to stratification effects can be minimized. In
cost can be reduced by the use of a fluid switch to connect the exhaust duct, the measurements should be made at a
several probes successively to a manometer. However, this location at which the flow is well mixed. Another possible
slows down the data acquisition process, since time must be optical smoke measurement system design is based on an
allowed for the pressure to equilibrate at each station. Texts by incandescent light source (46). In selecting a system, it is

10
 E 603 – 01

FIG. 4 Laser Extinction Beam

important to understand the nature of the measurement errors rate of fuel production and degree of combustion. Levels of
associated with it (47). carbon monoxide (CO), hydrogen chloride (HCl), hydrogen
6.5.7.2 Changes in the intensity of a well-collimated light cyanide (HCN), and nitrogen oxides (NOx) account for the
beam that may extend from floor to ceiling, pass horizontally common toxic gas concentration components. The presence of
across the room or doorway, or be limited to a few inches in a complex gas or gases that may be highly toxic in trace
length are used to measure the smoke. In choosing one or more amounts requires the collection of grab samples that can be
path lengths for measurement, consideration should be given to analyzed by mass spectrometry or gas chromatography, or
the normal tendency of smoke to stratify, particularly in the both. Toxicity assessment may require bioassays and behav-
early stages of a fire. ioral studies of exposed laboratory animals (see 6.5.9).
6.5.7.3 Deposition of soot within the optics is one of the 6.5.8.4 Oxygen concentrations are measured continuously
large sources of error in smoke measurements. This can be with meters that respond to the paramagnetic properties of
overcome by providing a gentle flow of forced air to purge the oxygen or with electrochemical cells, by a polarographic
system. method.
6.5.8 Gas Concentrations: 6.5.8.5 Nondispersive infrared instruments are the most
6.5.8.1 Sampling probes should be located with the holes common technique used for recording the levels of carbon
facing downstream, away from the inlet, to avoid clogging the monoxide, carbon dioxide, and water continuously.
holes with soot. The recommended designs of gas sampling 6.5.8.6 Measurements of water concentrations require
probes are a ring sampler, such as that used in Test Method heated filters, heated sampling lines, heated pumps, and heated
E 1537 or Test Method E 1590 (which also contains overall cells in the analyzers, to avoid condensation. Moreover, it is
construction details), or a straight tube crossing the duct more difficult to calibrate water analyzers than other analyzers.
section, such as the recommended design in ISO 9705. This can be solved with special calibration gases, but different
6.5.8.2 The measurement of oxygen is used to determine the calibration gases may be necessary for different analyzers. The
rate of heat release in conjunction with the flow rate of smoke measurement of water concentrations should be restricted to
(see 6.5.2). Measurements of carbon dioxide, carbon monox- special circumstances, because of the experimental difficulties
ide, and water concentrations in the outgoing smoke can aid in involved.
the precision of calculations of rate of heat production in the 6.5.8.7 Gas sampling tubes that change color when samples
room, for analysis of smoke toxicity and as indicators of of specific gases are drawn through them can be used to
combustion completeness. determine the approximate concentrations of a large number of
6.5.8.3 For condensable combustion gases, heated sampling combustion products, including carbon monoxide, carbon di-
lines (150 to 175°C) should be used, to minimize losses oxide, HCl, HCN, and NOx. These gas sampling tubes gener-
through deposition. This includes many gases, except for ally have very low precision and should only be used as a
carbon monoxide, carbon dioxide, and oxygen, notably hydro- rough guide. Comprehensive methods for measurement are
gen chloride (HCl), hydrogen cyanide (HCN), acrolein, am- provided in Guide E 800.
monia, unburnt hydrocarbons, and water. The length of sam- 6.5.9 Toxicity:
pling lines should be kept to a minimum and made from 6.5.9.1 The toxic potency of smoke is a quantitative expres-
non-reactive materials (for example, polytetrafluoroethylene). sion that relates smoke concentration and exposure time (that
This inhibits the reactivity of the combustion products with the is, the exposure dose) to a certain adverse effect, on exposure
tubing material and concentration losses, typically occurring of a test animal; the effect is usually lethality. The toxicity of
with hydrogen chloride (48 and 49). This information, along smoke is a function of its composition, which, in turn, is
with that from a combustible products gas meter, can yield the dependent both on the material being burnt and the way in

11
 E 603 – 01
which it is burnt. Thus, as the composition of the smoke 6.5.9.8 A number of toxic gases decrease their concentration
generated from the same material in different tests can vary in fires by virtue of reactions with surfaces (for example,
broadly, so will its toxic potency. It must be stressed that the walls), with the most notable example being hydrogen chloride
toxic potency of smoke is also heavily dependent on the (55, 56).
conditions under which the smoke has been generated, since 6.5.9.9 The propensity for smoke (or combustion products)
the mode of generation will affect both the amount and type of from any material or product to have the same effects on
combustion products being generated by the burning of any humans in fire situations as the effect it has on rats is only
material. inferable to the extent that the rat is correlated with humans as
6.5.9.2 Chemical analysis of smoke thus does not yield a a biological system.
comprehensive evaluation of all of the components in smoke. 6.5.10 Selection and Location of Instrumentation:
However, the toxicity of the smoke from the majority of 6.5.10.1 The large array of instrumentation described in this
common materials is usually described adequately by measure- guide is expensive. Thus, for research and development or
ment of the concentrations of a very small number of indi- demonstration experiments, some subset of instrumentation is
vidual toxicants. The most important of those are the carbon likely to be used. The selection of such a subset should be made
oxides (carbon monoxide and carbon dioxide) and low oxygen so as to satisfy the objectives of the experiment. However, a
(which are present in all fires) and other gases such as minimum amount of instrumentation is essential, even in a
hydrogen cyanide, hydrogen chloride, hydrogen bromide, and research experiment, so as to obtain data for the energy balance
acrolein (50-54). calculations necessary to adequately assess fire growth. These
6.5.9.3 For some materials (very few), it has been shown same measurements are necessary to establish exposure con-
that smoke toxicity is not well represented by the simple ditions for fire-test-response standards for igntion, flame spread
method described in 6.5.9.2. In that case, an adequate evalua- or heat release. In a single room test, only one vent is necessary
tion of smoke toxicity requires the use of laboratory animals. from which to measure the release rate. Appropriate locations
6.5.9.4 If laboratory test animals are used in room fire for vent include the doorway and a window. In a multi-room
experiments, place the animals within or near the compartment. facilty, one appropriate site for measurements is the end of an
Extract the smoke at a measuring point in the fire room and adjoining corridor, where the effluents are directed. It is
cool the smoke stream before exposing the animals, to avoid necessary to have data acquisition equipment, which can obtain
subjecting them to heat stress. If the smoke is cooled, the data at intervals no longer than 6 s (at least in the periods of
airborne concentration of those combustion products that decay high activity), including an analog/digital converter and a
in fire atmospheres (for example water soluble gases and solid computer that scans test activity. Details of instrumentation
particles) will decrease (7, 8) (see also 6.2.2.3 and 6.5.9.8). required for particular materials or products are given in the
Thus, cooling the smoke is likely to result in a mixture of specific test methods.
combustion products which is more representative of smoke at 6.5.10.2 A recommended set of measurements for a research
a location remote from the fire. It is also advisable to return the experiment includes the following:
smoke to the fire room, after animal exposure, to retain the (1) Oxygen concentration in the exhaust duct, to determine
original fire room environment. heat release (rates and amounts).
6.5.9.5 It is recommended that certain physiological char- (2) Burning rate of the item (or items) being tested, for
acteristics of the animals be monitored, as well as determining example, by measuring their mass loss rate.
the end point, which is usually lethality. Measurements that (3) Burning rate of the ignition source.
have been conducted include blood concentrations of carboxy- (4) Flame spread rates on objects or wall surfaces.
hemoglobin, oxyhemoglobin and total hemoglobin, blood pres- (5) Smoke obscuration measurements in the exhaust duct.
sure, respiration rate, heartbeat rate, body temperature, and (6) Carbon dioxide and carbon monoxide concentrations of
brain activity. gas in the exhaust duct.
6.5.9.6 Gas analyses for carbon monoxide, carbon dioxide, (7) Continuous measurements of mass of any items being
oxygen, and halogen acids (typically HCl) are usually con- tested.
ducted at various sampling points, animal chambers, and the 6.5.10.3 Other important measurements include the follow-
test room. The gas analysis data can be used to calculate the ing:
fractional effective dose (FED) of toxicity in the smoke. This (1) Vertical air temperature profiles in the doorway and
method evaluates the combined effect of selected primary center of the room.
gases on humans and animals. (2) Surface temperature at the center of the ceiling and
6.5.9.7 A complement to the chemical analysis of smoke is upper half of the walls.
the measurement of toxic potency. The concentration of smoke (3) Total heat flux measurements at the center of the floor,
(kg/m 3) required to kill 50 % of the animals is the toxic on the wall surface above the ignition source, and near the
potency measure, or LC50. Animals are preselected for mortal- center of the ceiling.
ity studies; hereby, mortality is not only observed during the (4) Vertical air flow rates in the doorway (45).
test but for 14 days after exposure as well. Several authors (5) Upper layer height.
(50-54) discuss the significance of these measurements. In a (6) Surface temperatures of target items.
similar fashion, incapacitation measures (for example, an IC50) (7) Heat flux to target objects.
can also be made. (8) Ignition times of target objects.

12
 E 603 – 01
(9) Gas concentrations. the fire area, especially materials that could add fuel to the fire
6.5.10.4 Some measurements in the test room may also or cause tripping accidents during subsequent operations. A
yield useful information. Examples are as follows: safety checklist should be prepared and used to ensure that all
(1) Carbon dioxide and carbon monoxide concentrations of safety procedures are followed properly.
the gas 0.025 m (1 in.) below the top of the doorway. 7.1.1.3 Careful consideration should be given to the electri-
(2) Smoke measurements in the center of the room and cal aspects of the experiment instrumentation and wiring.
doorway. Wiring of the appropriate temperature rating should be used.
(3) Temperature measurements at various locations. Electrical circuitry is necessarily extensive in such an experi-
6.5.10.5 The foregoing measurements should be provided mental arrangement, and care must be taken to avoid patch-
with time-tracked, photographic, or video coverage and with work systems. In large projects, there is a tendency by some to
audio-recorded, visual observation of the fire as seen through use recognized unsafe practices with the excuse that they are
the doorway.
temporary and everybody knows about them. Separate electri-
6.5.10.6 Certification experiments and research experiments
cal circuits should be dedicated to each major phase of the
aimed at the development of materials might be limited to
experiment. Electrical service to the test room should be
temperature measurements with photographic and visual ob-
evaluated with special consideration for needs during extin-
servation.
guishment of fire.
6.6 Location of Instrumentation:
6.6.1 The instrumentation in the exhaust duct should be 7.1.1.4 The use of flammable liquids for ignition fuels or for
located in such a way that the combustion products are well aiding the ignition of other fuels should be watched closely.
mixed and that there is a nearly uniform flow velocity across Where a large volume of gaseous fuel is used for the burner, a
the duct cross section. safety cutout device should be used to detect and cut off the
6.6.2 A minimum of six duct diameters is the recommended supply of fuel to the burner. For example, a combination of a
distance of straight duct, downstream from the last turn, before flame detector (flame rod) and a solenoid valve may be used for
placing any measuring instrument. A minimum of four duct this purpose. While it is not reasonable to prohibit the use of
diameters is recommended after the last measuring instrument certain materials in a research effort, a list might be prepared of
before placing any other fans or turns. materials, including indications of potential hazards and cau-
6.6.3 When flow velocity, smoke obscuration, and gas tions for their use. For example, one of the flammable liquids
concentrations are measured, the recommended order in which that might be included in the list is heptane, a highly volatile
the instrumentation is to be placed is as follows: velocity substance that must be used with extreme caution. Heptane
measurements followed by gas concentrations (including oxy- evaporates rapidly and propagates fumes over floor surfaces,
gen, for heat release) and smoke obscuration. A suitable which can be ignited easily, causing minor explosions. As such,
separation between the different instruments is 0.3 m. all liquids must be stored in approved safety containers. Light
fuel oils can be used more safely since they volatilize slowly at
7. Procedure
room temperatures; however, they are smokey. Absolute ethyl
7.1 Safety Precautions: alcohol volatilizes moderately, burns cleanly, and can be used
7.1.1 Prior to Experiment: with relative safety. When such fuels are used for ignition, the
7.1.1.1 Prior to performing the experiment, safety planning containers should be removed from the experiment site prior to
should occur and safety preparations should be made. The local ignition.
fire department may be notified in advance. Technicians should
7.1.1.5 Prior to and during an experiment, it is important
be assigned in pairs and should have responsibility for one
that one individual have full authority with respect to safety
another during the period of the experiment. The size of the fire
considerations and experiment termination. This individual
and the fuel loading should be considered and the following
should have the responsibility of reviewing the experiment
determinations made: (1) the primary and the backup method
arrangement with respect to safety, as well as deploying
of extinguishment; (2) the expected toxic pyrolysis products
and fire gases; (3) the potential for an explosion caused by a personnel and reviewing the understanding of individual re-
buildup of flammable gases such as carbon monoxide; (4) the sponsibilities. Prior to the conduct of the experiment, consid-
most appropriate place for visiting observers and experiment eration should be given to means of controlling particulates and
personnel to be located; and (5) the ventilation necessary to toxic gases in the outside enclosure. Experimental situations
remove toxic gases and limit the obscuration problem. often generate considerable smoke, making visibility difficult,
7.1.1.2 Extinguishment equipment, gas analyzers, protec- and accidents can occur easily.
tive clothing, breathing apparatus, medical first aid, and ven- 7.1.2 During the Experiment—All experiment personnel
tilation equipment should be provided based on these determi- should be in fire-retardant clothing with hard hats and safety
nations. All observers and experiment personnel should be glasses. Observers should be in protective clothing commen-
provided with appropriate protective clothing and safety eye- surate with their proximity to the fire. The extinguishing team
wear and should be briefed about the evacuation route. should be ready, clothed in turn-out-coats, self-contained
Properly lighted EXIT signs should be posted at appropriate breathing apparatus, and fire helmets. All backup equipment
places. Once all of these preparations have been made and just should be readily accessible. Toxic gases and potentially
prior to ignition, the experiment site should be safety checked explosive gases should be monitored throughout the experi-
to verify that all extraneous materials have been removed from ment. A continuous assessment of the fire damage should be

13
 E 603 – 01
made, with special attention given to any weakening of the content of the materials used should be recorded just prior to
structure that could permit falling debris. the test (follow the guidance in Test Methods D 4442 and
7.1.3 After the Experiment: D 4444).
7.1.3.1 Extinguishment should begin as soon as possible 7.2.3 Data Recording:
after the necessary data are recorded. Electric power to items 7.2.3.1 It is not necessary to scan the data continuously, but
within the compartment must be cut off prior to entry of the it should be done at least every 15 s. If the system is variable,
extinguishing team. The primary extinguishing team should a more frequent scan may be required when rapid changes
not engage the fire, however, until the backup equipment is occur. Computer data acquisition equipment is highly recom-
charged and manned. All loose material should be pulled from mended in the conduct of room fire tests. The analog output of
the upper walls and overhead as the extinguishing crew all transducers should be read, processed, stored converted to
approaches the fire. Forced ventilation should begin once the digital output, and printed to screen every 6 s for up to 50
channels of data and every 10 s for up to 100 channels. The
fire is extinguished completely. Known toxic gases should be
screen output serves as an on-the-spot reference of critical data.
monitored continuously from the very start of the experiment
The 60-Hz noise should be filtered simultaneously. Recogniz-
and up until their concentrations have returned to acceptable
ing that sharp peaks may be missed with this frequency of
limits. All enclosed experiment sites should also be monitored
channel scanning, sophisticated equipment is available that is
for oxygen levels to preclude any oxygen starvation situations.
capable of scanning 100 channels per second. Alternatively, a
7.1.3.2 No one should enter the experiment area after the few (up to 10) selected channels may be read every second by
fire has been extinguished without a self-contained breathing one economical data acquisition system while another system
apparatus and protective clothing until the room is thoroughly is scanning all other channels at a 6 s or greater interval.
ventilated. Melts of synthetic materials should be handled 7.2.3.2 If a computer data acquisition system as described in
carefully because of the possibility of retained heat and 7.2.3.1 is used, strip chart recorders become optional backup
continuous generation of toxic products. The possible contami- equipment for data acquisition. Although it is far better to have
nation with toxic gases and skin irritants should be considered a second computerized system for backup, in a practical sense,
for all materials taken from the experiment area. The experi- it may not be feasible to have backup for every channel of data
ment site should be watched for at least 1 h, to ensure complete being collected by the computer system.
extinction of the fire. Unexpected fires may occur from piles of 7.2.4 The mass loss of objects and materials in the room
smoldering, hot, and charred debris. Post-experiment cleanup should be collected using load cells sized appropriately for the
may be facilitated by the use of a suitable disposal material laid mass being measured. Care should be taken to protect the load
on the floor if such would not be expected to alter the outcome cell and electrical wiring from excessive heat and radiation.
of the experiment (see 6.2.3). These cells are highly temperature sensitive. Two other impor-
7.2 Observations and Data Gathering: tant factors that should be addressed are buoyancy effects and
7.2.1 Observations: signal noise. Signal noise can be reduced through electronic
7.2.1.1 The following critical times should be noted by an damping or left and dealt with by best curve fitting or
assigned observer during the test: (1) ignition of each separate averaging the noise. While it is possible to read the load cell
combustible item in the room, including wall, ceiling, floor, output manually, computerized data collection is recommended
furnishings, and combustible indicator panels, if used; ( 2) highly.
onset of flame passage through the doorway (flameover); and 8. Analysis and Use of Results
(3) sudden deformation or change in noncombustible items in
8.1 Criteria:
the room.
8.1.1 The criteria for evaluating the performance of the
7.2.1.2 The flame spread rate on the walls and ceiling, finishing and furnishing materials of a room can be based both
maximum flame travel distance from the ignition source, on what occurs within the room and what occurs outside of the
height of the lower boundary of the smoke layer passing out room. One key factor affecting occupants outside the room of
through the doorway, and degree of involvement of the various fire origin is the total energy released. This energy makes the
combustibles should also be noted periodically during the room a heat pump that distributes smoke and hot gases
experiment. A grid premarked on the walls and ceiling will throughout the building. When the room reaches flashover, the
assist in estimating the flame spread rate. An audio recording in possibility of escape for the occupants in an adjacent area is
real time is one of the most efficient methods of documenting reduced substantially. Considerations within the room may
rapidly changing events. These observations should be backed include vertical temperature profiles, smoke and combustion
up with video tape or photographic coverage whenever pos- gas species profiles, surface temperatures, heat flux levels
sible. developed at the center of the floor, total rate of energy release,
7.2.2 Ambient Conditions—Since the relative humidity of and time to flashover.
the air and moisture in the experiment materials may affect the 8.1.2 Consequently, an important criterion is the elapsed
experiment results markedly, measurements of ambient humid- time from ignition to flashover in the room. The ignition time
ity and air temperature should be made for at least 12 h could be defined as ignition of either the heat source or the
preceding the experiment. These ambient conditions should be specimen.
measured both within the compartment and outside of the 8.1.2.1 Flashover has been found to correlate reasonably
compartment just prior to the test. In addition, the moisture well with an average air temperature of 500 to 600°C measured

14
 E 603 – 01
TABLE 1 Typical Thermal Property Values of Some Common Materials (To be Used for Guidence Only)

NOTE 1—The data provided in this table (Peacock, et al., NIST, 1994) defines one set of properties for common materials which are not well defined,
and are provided for approximate guidance only. The numbers listed within this table cannot be assumed to fully reflect the properties of all materials
within the generic class described. Data for common brick and clay brick were provided by the Brick Institute of America.
Thermal Conductivity, Specific Heat, Density,
Materials, in. Thickness, m Emissivity, (-)
W(m K) J/(kg K) kg/m3
Gypsum board, 1⁄2 in. 0.16 900 790 0.013 0.9
Gypsum board, 5⁄8 in. 0.16 900 790 0.016 0.9
Gypsum board, 3⁄4 in. 0.16 900 790 0.019 0.9
Gypsum wallboard, ranges 0.16–0.22 900–1047 790–400 0.024–0.050 0.90–0.97
Gypsum board, type X, 5⁄8 in. 0.14 900 770 0.016 0.9
Gypsum board, type X, 3 in. 0.22 1085 1680 0.076 0.9
Gypsum substrate, w. glass matte 0.16–0.04 900–720 790–10 0.024–0.050 0.9
Brick, common, 3 in. 0.72 921 1920 0.076 0.9
Clay brick, 3 in. 1.3 1004 2082 0.076 0.9
Fire brick 0.36 750 1040 0.113 0.8
Fire brick composite, range 0.17–0.36 1040 128–750 0.005–0.113 0.95
Concrete, normal weight, 6 in. 1.75 1000 2200 0.15 0.94
Cement mortar, 1 in. 0.72 780 1860 0.025 0.9
Glass plate, 1⁄4 in. 1.4 750 2500 0.006 0.1
Aluminum, pure, 1⁄8 in. 231 1033 2702 0.003 0.9
Aluminum alloy 2064-T6, 1⁄8 in. 186 1042 2770 0.003 0.9
Carbon steel, plain, 1⁄8 in. 48 559 7854 0.003 0.9
Carbon steel, plain, sheet, 1⁄16 in. 48 559 7854 0.0015 0.9
Stainless steel 304, 1⁄8 in. 19.8 557 7900 0.003 0.9
Plywood building board, 1⁄2 in. 0.12 1215 545 0.013 0.9
Hardwood siding, 1⁄2 in. 0.094 1170 640 0.013 0.9
Hardboard, high density, 1⁄2 in. 0.15 1380 1010 0.013 0.9
Particle board, low density, 1⁄2 in. 0.078 1300 590 0.013 0.9
Particle board, high density, 1⁄2 in. 0.17 1300 1000 0.013 0.9
Hardwoods (oak, maple), 3⁄4 in. 0.16 1255 720 0.019 0.9
Softwoods (fir, pine), 3⁄4 in. 0.12 1380 510 0.019 0.9
Wood board, shredded, cemented, 1⁄2 in. 0.087 1590 350 0.013 0.9
Sheathing, regular density, 1⁄2 in. 0.055 1300 290 0.013 0.9
Ceremic (kaolin) fiber insulation 0.22 1047 128 0.116 0.97
Glass fiber insulation, 3-1⁄2 in. 0.04 720 105 0.088 0.9
Glass fiber, organic bonded, 1⁄2 in. 0.036 795 105 0.013 0.9
Glass fiber, poured or blown, 1⁄2 in. 0.043 835 16 0.013 0.9
Glass fiber, coated, duct liner, 1⁄2 in. 0.038 835 32 0.013 0.9
Acoustic tile, 1⁄2 in. 0.058 1340 290 0.013 0.9
Vermiculite flakes, 1⁄2 in. 0.068 835 80 0.006 0.9
Urethane insulation, rigid foam, 1⁄2 in. 0.026 1045 70 0.013 0.9

100 mm below the ceiling, or above 600°C measured at the top engineering approximation to determine the rate of heat release
of the doorway (31). A more direct indication of flashover is that will result in room flashover (22).
provided by the radiant energy flux incident on the floor. A
RHR 5 E 3 F 3 V̇ (2)
level of 20 kW/m2 at the center of the floor is indicative of
incipient flashover in the room. For some classes of products
and certain fire scenarios, flame emergence from the doorway where:
is well correlated to flashover (22). The most convincing RHR = rate of heat release (MW),
indication of flashover can come from the ignition of raw E = energy release per kg of combustion air (E = 3.0
cotton fiber indicators or crumpled newsprint. Consequently, a MJ/kg),
criterion that may be used in a room fire experiment could be F = fraction of maximum air flow into the room at the
the time to flashover. Flashover can be inferred from measure- onset of flashover (0.3 # F # 0.5),
ments made by one or a combination of the following: (1) the V = maximum air flow (kg/s) into the room following
thermocouples in the upper part of the room, (2) a total heat flashover,
flux gage at the center of the floor, (3) the emergence of flames V = 0.5 A=H, with A = area of ventilation opening
from the doorway (for some products), or (4) by the ignition of [m2], and
cotton indicators or a ball of crumpled newsprint. H = height of ventilation opening (m).
8.1.3 If the energy release rate produced during a room fire This expression permits the calculation of a range of values
test is measured using oxygen consumption calorimetry (Ref of rate of heat release sufficient to cause flashover in a room
(38)), criteria can be based on the energy release rates produced whose floor area does not exceed 500 m2.
by the lining or furnishing materials, or both. Criteria based on 8.1.4 The potential growth and spread of fire within the
energy release rates can be tailored to ensure that only a given room of fire origin is another criterion that is useful for
fraction of the energy required to produce room flashover is determining the ease with which a major fire may develop. This
permitted. The following relationship can be used as an criterion may be applied by evaluating the size of the igniting

15
 E 603 – 01
source (whether a wood crib, waste container, or gas burner) (presented by high temperatures, poor visibility, or the pres-
necessary to cause flashover or flame emergence from the ence of toxic gases) within the building.
doorway for a specific test specimen. This requires multiple 8.1.10 Another criterion that could be useful is based on the
experiments to determine the size of the ignition source potential of post-flashover fire spread from one compartment to
required. another in multi-room scenarios. By measuring the temperature
8.1.5 Another criterion that may be applied is based on an rise on the unexposed sides of the walls of the room of fire
evaluation of the fire-damaged area for a fixed size of ignition origin or measuring heat flux emitted by the unexposed sides of
source (for example, equivalent to a 5-lb (2.3-kg) wood crib) these walls, or both, one can gain a measure of the likelihood
for the experiment specimen. The rate of fire growth can also of intercompartmental fire spread through the room walls.
be documented photographically and used to develop criteria. Should the adjacent rooms be furnished, the same methodology
8.1.6 Another criterion that might be used is based on the as that outlined in 8.1.2.1 could also be used to determine
total amount or rate of smoke generated. The rate of smoke whether flashover has occurred in rooms adjacent to the room
generation can be estimated from the optical density per meter of fire origin.
(at the top of the doorway) times the volume flow of air from 8.1.11 In summary, the time to flashover of the room, rate of
the room. The dilution, coagulation, deposition, and stratifica- heat release by the room linings and contents, size of the
tion of the smoke make analytical predictions of smoke ignition source required to produce flashover of the room, rate
concentration at remote locations in the building difficult even and total volume of smoke and combustion gases produced by
for a known smoke generation rate. the fire, and rate of room to room spread of fire, smoke, and
8.1.7 Optical density correlates well with visibility, and other toxic combustion products can all be used to develop
hence it is a good performance criterion. The attenuation level criteria that can be useful for evaluating the fire performance in
relative to some arbitrary product is certainly a good perfor- the room experiment arrangement.
mance measure. On the other hand, attenuation levels mea- 8.2 Analysis—The analysis of data on the room fire experi-
sured in full-scale multi-compartment simulations could be ments can serve several purposes:
used to establish criteria at locations remote from the room of 8.2.1 To relate the severity of the room fire experiment to
fire origin. the laboratory-measured fire properties of the materials. This is
8.1.8 Other criteria that might be used could be based on the an attempt to establish empirical relationships that would
rate of production or the concentration of certain species of validate the strengths and weaknesses of the various fire test
gases produced by the burning materials. This information methods currently in use or proposed for controlling the
could be used in experiments of smoldering fires to ensure that products to be used in the rooms.
the concentrations of specific gas species do not exceed 8.2.2 To compare data obtained from room fire experiments
tenability limits at critical locations within the room, such as to results from room fire models. This comparison indicates the
the elevation at which a person sleeping in a chair could be strengths and weaknesses of the models. Studies by Benjamin
exposed. For non-smoldering fires, the rate of production and and Peacock (57, 58) demonstrate the validation of fire models
concentration of gases could be used to assess the effects on and fire hazard assessment using full-scale tests. Guide E 1355
building occupants located outside the room of fire origin. is a good validation guide for fire models.
Criteria could be developed by comparison with the production 8.2.3 To provide input data for room fire models. Paragraph
rates of some arbitrary experimental materials. Another crite- 8.5.10 lists some of the properties that are helpful to modelers.
rion could be to ensure that the concentrations of specific gases Full-scale data are very important for developing empirical
or total smoke dose at selected locations outside the room of models of fire behavior where a theoretical basis is lacking.
fire origin do not exceed tenability limits. Tenability limits These empirical estimates serve further to improve the models.
could also be observed for radiant heat (burns) and convected 8.2.4 To evaluate materials and products when estimating
heat stress. their fire performance, particularly those for which no adequate
8.1.9 A more challenging approach would be to develop material property tests exist.
criteria based on the likelihood of the safe evacuation of
occupants from the building. The multi-room fire scenario 9. Keywords
could be used to determine whether the time for detection of 9.1 burning characteristics; experiment design; flashover
the fire and the time for escape to an area of safety are adequate guide; full-scale test; guide; ignition; instrumentation; room
in view of the time for the development of untenable conditions fire experiment; ventilation

16
 E 603 – 01
REFERENCES

(1) Fire Research 1968, Report of the Fire Research Steering Committee (19) Babrauskas, V., Lawson, J. R., Walton, W. D., and Twilley, W. H.,
with the Report of the Director Fire Research, Ministry of Technology, “Upholstered Furniture Heat Release Rates Measured with a Furni-
JFRO (Joint Fire Research Organisation), London, United Kingdom, ture Calorimeter,” NBSIR 82-2604, National Bureau of Standards,
1969. Gaithersburg, MD, December 1982.
(2) Kawagoe, K.,“ Fire Behavior in Rooms,” Report No. 27, Building (20) Babrauskas, V., “Combustion of Mattresses Exposed to Flaming
Research Institute, Ministry of Construction, Tokyo, Japan, p. 732, Ignition Sources, Part II: Bench-Scale Tests and Recommended
1958. Standard Test,” NBSIR 80-2186, National Bureau of Standards,
(3) Hägglund, B., Jansson, R., and Onnermark, B., “Fire Development in Gaithersburg, MD, February 1981.
Residential Rooms After Ignition from Nuclear Explosion,” FOA (21) Gross, D., “Data Sources for Parameters used in Predictive Modeling
Report C20016-D6 (A3), Stockholm, Sweden, 1974. of Fire Growth and Smoke Spread,” NBSIR 85-3223, National
(4) Drysdale, D.,“ An Introduction to Fire Dynamics,” Wiley-Interscience Bureau of Standards, Gaithersburg, MD, 1985.
Publication, Chichester, United Kingdom, 1985. (22) Babrauskas, V., and Krasny, J., “Fire Behavior of Upholstered
(5) Parker, W. J., Tu, K.-M., Nurbakhsh, S., and Damant, G. H., “Furniture Furniture,” NBS Monograph 173, National Bureau of Standards,
Flammability: An Investigation of the California Technical Bulletin Gaithersburg, MD, 1985.
133 Test, Part III: Full Scale Chair Burns,” NISTIR 90-4375, National (23) Holmlund, C., “Burning of Small Crib Ignition Sources,” Report No.
Institute of Standards and Technology, Gaithersburg, MD, 1990. 569, Technical Research Center of Finland, Espoo, Finland, Novem-
(6) Kokkala, M., Gǒransson, U., and Šoderbom, J., “Experiments in a ber 1988.
Large Room with Combustible Linings,” Presented at 2nd Interna- (24) Ahonen, A., Kokkala, M., and Weckman, H., “Burning Characteris-
tional Heat Release and Fire Hazard Symposium, Brussels, Belgium, tics of Potential Ignition Sources of Room Fires,” Research Report
February 1991. 285, Technical Research Centre of Finland, Espoo, Finland, June
(7) Galloway, F. M., and Hirschler, M. M., “The Use of a Model for 1984.
Hydrogen Chloride Transport and Decay to Predict Airborne Hydro- (25) Castino, G. T., Beyreis, J. R., and Metes, W. S., “Flammability
gen Chloride Concentrations in a Full-Scale Room-Corridor Sce- Studies of Cellular Plastics and Other Building Materials Used for
nario,” Fire Safety Journal, 19, 1992, pp. 73–101. Interior Finish,” Underwriters’ Laboratories, Inc., Northbrook, IL,
(8) Galloway, F. M., and Hirschler, M. M., “Transport and Decay of 1975.
Hydrogen Chloride: Use of a Model to Predict Hydrogen Chloride (26) Quintiere, J. G., and McCaffrey, B. J., “The Burning of Wood and
Concentrations in Fires Involving a Room-Corridor-Room Arrange- Plastic Cribs in an Enclosure: Volume 1,” NBSIR 80-2054, National
ment,” Fire Safety Journal, Vol 16, 1990, pp. 33–52. Bureau of Standards, Gaithersburg, MD, 1980.
(9) Gritzo, L. A., and Nicolette, V. F., “Coupled Thermal Response of (27) Babrauskas, V., “Flame Lengths Under Ceiling,” Fire and Materials,
Objects and Participating Media in Fires and Large Combustion Vol 4, 1980, pp. 119–126.
Systems,” Numerical Heat Transfer, Part A, Vol 28, pp. 531–545, (28) Thomas, P., and Karlsson, B., “On The Length of Flames Under
1995. Ceiling,” Report LUTVDCT/TVBB-3059, Lund Institute of Technol-
(10) Magnusson, S. F., and Thelanderson, S., “Temperature-Time Curves ogy, Dept. of Fire Safety Engineering, Lund, Sweden, 1990.
of Complete Process of Fire Development—Theoretical Study of (29) Lie, T. T., Fire and Buildings, Applied Science Publishers, Ltd.,
Wood Fuel Fires in Enclosed Spaces,” Acta Polytechnica Scandina- London, United Kingdom, 1972.
via, Civil Engineering and Building Construction Series No. 65, (30) Thomas, P. H., “The Size of Flames from Natural Fires,” Ninth
Stockholm, Sweden, 1970. Symposium on Combustion, Academic Press, New York and London,
(11) Kawagoe, K., and Sekine, T., “Estimation of Fire Temperature-Time United Kindom, 1963, pp. 844–859.
Curve in Rooms,” Occasional Report No. 11, Building Research (31) Tarifa, C. S., Del Notario, P. P., and Valdes, E. C. F., “Open Fires,”
Institute, Tokyo, Japan, 1963. Instituto Nacional de Tecnica Aeroespacial8 Esteban Terrada’,
(12) Odeen, K., “Standard Fire Endurance Tests—Discussion, Criticism, Madrid, Spain, 1967, p. 86.
and Alternatives: Fire Test Performance,” ASTM STP 464, ASTM, (32) Yokoi, S.,“ Study on the Prevention of Fire Spread Caused by Hot
Philadelphia, PA, 1979, pp. 30–56. Upward Current,” Report No. 34, Building Research Institute,
(13) Silbulkin, M., “Effects of Material Properties on Burning and Ministry of Construction, Tokyo, Japan, 1969, pp. 34–36.
Extinction Fires on Vertical Surfaces,” Final Report, Brown Univer- (33) Yokoi, S., “On the Heights of Flames from Burning Cribs,” Occa-
sity, Providence, RI, February 1982. sional Report No. 12, Building Research Institute, Ministry of
(14) Lee, B. T., “Effects of Wall and Room Surfaces on the Rates of Heat, Construction, Tokyo, Japan, 1963, p. 12.
Smoke, and Carbon Monoxide Production in a Park Lodging Bed- (34) Alpert, R. L., “Fire Induced Turbulent Ceiling Jet,” Technical Report
room Fire,” NBSIR 85-2998, National Bureau of Standards, Gaith- 19722-2, Factory Mutual Research Corporation, May 1971.
ersburg, MD, February 1985. (35) Alpert, R. L., “Calculation of Response Time of Ceiling-Mounted
(15) Tran, H. C., and Janssens, M. L.,“ Room Fire Test for Fire Growth Fire Detectors,” Fire Technology, Vol 8, 1972.
Modeling—A Sensitivity Study,” Journal of Fire Sciences, Vol 7, (36) Steckler, K., Tu, K. M., and Twilley, W., “Calibration Technique for
July/August 1989, pp. 217–236. Heat Flux Sensors Used in Fire Experiments and Standard Fire
(16) Peacock, R. D., Jones, W. W., Forney, G. P., Portier, R. W., Reneke, Tests,” Report to Federal Aviation Administration, National Institute
P. A., Bukowski, R. W., and Klote, J. H., “An Update Guide for of Standards and Technology, Gaithersburg, MD, 1989.
HAZARD I Version 1.2,” NISTIR 94-5410, National Institute of (37) Olsson, S., “Calibration of Radiant Heat Flux Meters: The Develop-
Standards and Technology, Gaithersburg, MD, 1994. ment of a Water Cooled Aperture for Use with Black Body Cavities,”
(17) Klitgaard, P. S., and Williamson, R. B., “The Impact of Contents on Nordtest Project 873-90, SP Report 1991:58, Swedish National
Building Fires: Journal of Fire and Flammability Consumer Products Testing and Research Institute, Boras, Sweden.
Flammability Supplement,” Vol 2, 1975, pp. 84–113. (38) Janssens, M., “Measuring Rate of Heat Release In Full-Scale Fire
(18) Gross, D., and Fang, J. B., “The Definition of a Low Intensity Fire: Tests By Oxygen Consumption,” Fire Technology, Vol 27, August
Performance Concept in Buildings,” NBS Special Publication 361, 1991, pp. 237–249.
Vol 1, Proceedings of the Joint RILEM-ASTM-CIB Symposium, (39) Babrauskas, V., and Grayson, S. J., “Heat Release in Fires,” Elsevier,
Philadelphia, PA, May 2–5, 1972. London, UK, 1992.

17
 E 603 – 01
(40) Gardon, R., “An Instrument for the Direct Measurement of Intense Toxicology—Principles and Test Methods,” Technomic, Lancaster,
Thermal Radiation,” Review of Scientific Instruments, Vol 24, No. 5, PA, 1983.
May 1953, pp. 366–370. (51) Purser, D., “Toxicity Assessment of Combustion Products,” SFPE
(41) Borell, G. J., and Diller, T. E., “Convection Calibration Method for Handbook of Fire Protection Engineering, National Fire Protection
Local Heat Flux Gages,” Journal of Heat Transfer, Vol 109, February Association, Quincy, MA, Section 1, Chapter 14, 1988.
1987, pp. 83–89. (52) Hirschler, M. M., “Smoke Toxicity Measurements Made so that the
(42) Keltner, N. R., and Wildin, M. W., “Transient Response of Circular Results can be Used for Improved Fire Safety,” Journal of Fire
Foil Heat-Flux Gages to Radiative Fluxes,” The Review of Scientific Sciences, Vol 9, 1991, pp. 330–347.
Instruments, Vol 46, No. 9, September 1975. (53) Babrauskas, V., Harris, R. H., Braun, E., Levin, B. C., Paabo, M., and
(43) Benedict, R. P., Fundamentals of Temperature, Pressure and Air Flow Gann, R. G., “The Role of Bench-Scale Data in Assessing Real-Scale
Measurement, 3rd ed., Omega Press. Fire Toxicity,” NIST Technical Note No. 1284, National Institute of
(44) Ower, E., and Pankhurst, R. C., The Measurement of Air Flow, Standards and Technology, Gaithersburg, MD, 1991.
Pergamon Press, 5th ed., 1977, pp. 112–147.
(54) Babrauskas, V., Levin, B. C., Gann, R. G., Paabo, M., Harris, R. H.,
(45) Peacock, R. D., and Babrauskas, V., “Analysis of Large-Scale Fire
Peacock, R. D., and Yusa, S., “Toxic Potency Measurement for Fire
Test Data,” Fire Safety Journal, Vol 17, 1991, pp. 387–414.
Hazard Analysis,” NIST Special Publication No. 827, National
(46) Bukowski, R. W., “Smoke Measurements in Large- and Small-Scale
Institute of Standards and Technology, Gaithersburg, MD, 1991.
Fire Testing,” Fire Technology, Part I, Vol 15, No. 3, 1979, pp.
173–179; Part II, Vol 15, No. 4, 1979, pp. 271–281. (55) Beitel, J. J., Bertelo, C. A., Carroll, W. A., Grand, A. F., Hirschler, M.
(47) Mulholland, G., “How Well are we Measuring Smoke?” Fire and M., and Smith, G. F., “Hydrogen Chloride Transport and Decay in a
Materials, Vol 6, 1982, pp. 65–67. Large Apparatus, II: Variables Affecting Hydrogen Chloride Decay,”
(48) Beitel, J. J., Bertelo, C. A., Carroll, W. F., Gardner, R. A., Grand, A. Journal of Fire Sciences, Vol 5, 1987, pp. 105–145.
F., Hirschler, M. M., and Smith, G. F., “Hydrogen Chloride Transport (56) Galloway, F. M., and Hirschler, M. M., “A Model for the Spontaneous
and Decay in a Large Apparatus. I. Decomposition of Poly(Vinyl Removal of Airborne Hydrogen Chloride by Common Surfaces,”
Chloride) Wire Insulation in a Plenum by Current Overload,” J. Fire Fire Safety Journal, Vol 14, 1989, pp. 251–268.
Sciences, 4, 1986, pp. 15–41. (57) Benjamin, I. A.,“ Toxic Hazard Analysis: Electrical Non-Metallic
(49) Galloway, F. M., Hirschler, M. M., and Smith, G. F., “Small Scale Tubing,” Journal of Fire Sciences, Vol 5, January/February 1987, pp.
Experiments for Measuring HCl Decay in Fire Atmospheres,” in 25–48.
Proc. Interflam ’90, Ed. C. A. Franks, Sept. 3–6, 1990, Canterbury, (58) Peacock, R. D., Davis, S., and Lee, B. T., “An Experimental Data Set
UK, Interscience, London, UK, 1990, pp. 127–143. for the Accuracy Assessment of Room Fire Models,” NBSIR 88-3752,
(50) Kaplan, H. L., Grand, A. F., and Hartzell, G. E.,“ Combustion National Bureau of Standards, Gaithersburg, MD, 1988.

ASTM International takes no position respecting the validity of any patent rights asserted in connection with any item mentioned
in this standard. Users of this standard are expressly advised that determination of the validity of any such patent rights, and the risk
of infringement of such rights, are entirely their own responsibility.

This standard is subject to revision at any time by the responsible technical committee and must be reviewed every five years and
if not revised, either reapproved or withdrawn. Your comments are invited either for revision of this standard or for additional standards
and should be addressed to ASTM International Headquarters. Your comments will receive careful consideration at a meeting of the
responsible technical committee, which you may attend. If you feel that your comments have not received a fair hearing you should
make your views known to the ASTM Committee on Standards, at the address shown below.

This standard is copyrighted by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959,
United States. Individual reprints (single or multiple copies) of this standard may be obtained by contacting ASTM at the above
address or at 610-832-9585 (phone), 610-832-9555 (fax), or [email protected] (e-mail); or through the ASTM website
(www.astm.org).

18