UNIVERSAL
FLOOR TESTER: AN
Sean
Gerolimatos
Schluter
Systems, L.P.
194
Pleasant Ridge Road
Plattsburgh,
NY 12901, USA
Dale Kempster
Terrazzo, Tile and
Marble Association of
30 Capstan Gate, Unit 5
Peter
Nielsen
Schluter
Systems, L.P.
194
Pleasant Ridge Road
Frank
Woeste, Ph.D., P.E.
Wood Construction & Engineering
Consultant
Keywords:
ceramic tile assembly performance testing, Robinson-Type Tester, ASTM
C627, Universal Floor Tester, System Crack Resistance Test, ANSI A118.12
Abstract
The performance of a ceramic tile assembly
depends upon how effectively the installation system integrates the tile
covering with the existing flooring structure.
Thus, it is important that ceramic tile assembly performance testing
incorporates those aspects of the existing structure that are most likely to have
a detrimental effect on the tile assembly in service conditions. With this precept in mind, the authors
compare existing ceramic tile assembly performance test methods, including the
Robinson-Type Test Method per ASTM C627 and System Crack Resistance Test Method
per the soon to be published ANSI A118.12 standard, with new test methods made
possible through the development of the Universal Floor Tester.
The installation of ceramic tile in
Thus, the authors demonstrate how the Universal
Floor Tester will allow for improved evaluation of ceramic tile assemblies and
highlight various test programs that have the potential to increase the
industry’s understanding of ceramic tile assembly performance in service. The goal is to provide consumers with more
durable and reliable ceramic tile installations, which will serve to improve
customer satisfaction and increase ceramic tile consumption.
Development
of Selected Ceramic Tile Assembly Evaluation Methods
Robinson-Type
Test Method
The advent of dry-set mortar and subsequent
introduction of the thin-set method for tile installation led the Tile Council
of America (TCA)
Figure 1 - Robinson-Type
Floor Tester
A standardized test method based upon the
Robinson tester, ASTM C627, was written in 1970 under the auspices of ASTM
Committee C-21 on Ceramic Whitewares and Related Products after four (4) years
of comparative interlaboratory testing.
While the ASTM C627 standard has been revised over the years, only small
changes have been made, leaving the test method for the most part
unchanged. The latest revision came in
1999, with the next expected in late 2005 or early 2006.
System
Crack Resistance Test Method
The Material Methods and Standards Association
(MMSA) created a sub-committee in 1995 to create a standard for crack isolation
membranes for thin-set ceramic tile and dimension stone installation. The resulting standard, ANSI A118.12,
contains various performance criteria, including a test method to evaluate the ability
of a membrane to protect the tile assembly from minor in-plane substrate
cracking. The System Crack Resistance
Test Method uses a specially designed jig, which consists of a steel apparatus
to which two (2) concrete blocks are attached: one to a fixed plate, the other to a movable
plate. The ceramic tile assembly is
installed over the concrete blocks.
Figure 2 - System Crack
Resistance Tester
The apparatus is manually operated to move the
concrete blocks apart at a designated rate until a specified “crack width” is
reached. The assembly is inspected for
cracks in the tile and is discontinued if any are found. If no cracks are found in the tile, the test
is continued. An assembly exhibiting no
cracks in the tile after the opening width reaches 1/16" receives a
“standard performance” rating; an assembly exhibiting no cracks in the tile
after the opening width reaches 1/8" receives a “high performance”
rating.
Universal
Floor Tester
Development of the Universal Floor Tester began
in the year 2000. The primary objective
was to design a fully-automated testing machine that could be used to evaluate
ceramic tile assembly performance using a variety of methods. The primary feature is a weighted
three-wheeled carriage, similar to the Robinson-Type Tester, which is mounted
on a vertical slide instead of a center axle, eliminating the need to cut a
hole in the test specimen. The vertical
slide mount also allows for increased clearance below the carriage, up to
approximately 40" (1 m), making it possible to evaluate ceramic tile
assembly performance under loading over full-scale wood-framed flooring systems
(see Figure 3a).
Figure 3a - Universal
Floor Tester Full-Scale Assembly
The tester also features a sliding table
mechanism that is used to split concrete pads and simulate crack formation,
opening, and closing below ceramic tile assemblies (see Figure 3b). The mechanism can be opened to a maximum
width of 3/8" (9 mm), with the rate of opening and closing programmed into
the tester by the operator.
Figure 3b - Universal
Floor Tester Sliding Table Mechanism
Current plans are for comparative testing to be
performed through three (3) laboratories toward development of standardized
test methods using the Universal Floor Tester.
Comparison
of Evaluation Methods for Ceramic Tile Assemblies
When evaluating the results of any given test,
it is important to understand how the test method incorporates aspects of the
tile assembly in service. To do so, one
must first understand the various aspects of the existing structure that
supports the tile assembly and how they influence the performance of the tile
assembly. The performance of a tile
assembly depends upon how effectively the installation system integrates the
tile covering with the existing flooring structure.
Ceramic
Tile Assemblies for Installation over Wood Substrates
Residential construction, the largest source of
ceramic tile consumption in
Wood frame construction presents many challenges
to the tile setter, including deflection and curvature under loading (including
creep), stress concentrations at sheathing seams, and movement and dimensional
changes of the wood products due to changes in the moisture content of the wood
materials. The discussion below will
focus on deflection, curvature, and stress concentrations at sheathing seams.
Historically, deflection of the flooring system
has been viewed as the most detrimental to the performance of the tile assembly
over wood-frame construction, with the potential to cause cracking and
delaminating of the tile and grout. In
fact, the limitation of deflection has become a central design criterion in the
North American tile industry. The 2005
Handbook for Ceramic Tile Installation, published by the Tile Council of North
America, states in various details for tile installation over wood-frame
construction: “design floor areas over
which tile is to be applied to have a deflection not greater than 1/360 of the
span when measured under 300 lb. concentrated load (see ASTM C627).” However, the influence of deflection on the
tile assembly has traditionally been left in vague terms. Bretzfield and Woeste (2002) performed
engineering analyses to investigate this issue and provide a better
understanding of the dynamics of the flooring system (joists, subfloor, and
underlayment) with regard to the tile assembly.
Their work was based on the premise that it is curvature of the
subfloor, rather than simply deflection, which stresses the tile assembly and
can lead to damage. Consider that if the
flooring system were perfectly rigid, but allowed to deflect vertically as a
unit, no stress would be placed on the tile assembly. However, in reality, deflection of the floor
is accompanied by bending that produces stress on the rigid ceramic tile
covering, which does not conform to the resulting curved surface (see Figure
4).
Figure 4 - Exaggerated
View of Curvature of Wood Subfloor beneath Tile Covering
Through basic calculations, it was shown that
the predicted curvature of sheathing panels between
joists far exceeds that along the
joists. The implication of their
study is that sheathing stiffness is far more likely to influence tile assembly
performance than joist stiffness. This
finding was supported by joint testing administered by Osterberger, Mayhew, and
Fine (2004) of Trus Joist, A Weyerhaeuser
Business (engineered wood systems manufacturer) and Laticrete International, Inc. (ceramic tile installation systems
manufacturer) on full-scale floor assemblies (15-ft and 17-ft spans).
It follows that the most straightforward,
effective manner in which to improve the capability of a wood-framed floor to
serve as a base for ceramic tile, in terms of limiting curvature[1],
is to install a second layer of plywood or OSB.
However, since the floor sheathing is comprised of panels, rather than a
monolithic layer, and thin-set tile assemblies are sensitive to differential
vertical movement, stress concentrations at plywood/OSB seams must also be
considered to ensure a successful installation.
As such, Woeste and Nielsen (2004) put forth the argument that when
installing a plywood or OSB underlayment to limit curvature of the sheathing,
the underlayment joints should be placed at a location of relatively low
bending stress under a wide range of loading scenarios. For example, under a
uniform load, the bending stress in the sheathing between joists is near zero
at approximately the ¼-point between joists.
The integration of wood substrates with the
Robinson-Type Floor Tester, per ASTM C627, has been limited. Typically, the wood frame assemblies tested
consist of nominal 2 x 2′s (38
mm x 38 mm) at the desired joist spacing, which are fully supported by a
concrete slab along the length of the joist span. These “simulated joists” provide a surface
for nailing a single 4-ft x 4-ft (1.22 m x 1.22 m) panel of wood sheathing per
building code requirements.
As a result of the test configuration, deflection
and curvature of the flooring system during loading is confined to the
sheathing between the fully supported joists.
As discussed previously, there is evidence to show that curvature of the
sheathing is of greater importance than the deflection and curvature along the
joists. Thus, while the Robinson-Type
Floor Tester does not fully simulate an existing flooring structure, it does
incorporate one of the most significant aspects of the flooring structure with
respect to tile assembly performance. This is an important point to consider,
because it serves to validate the extensive use of and reliance on the
Robinson-Type Floor Tester over the last thirty-five (35) years in testing tile
assemblies over wood frame construction.
However, there are application shortcomings that must be addressed. In the authors’ view, the primary drawback is
that the limited size of the test specimens prohibits the inclusion of seams in
the sheathing in a manner consistent with building code requirements. Seams in the subfloor sheathing are located
at areas of maximum negative bending stress (over joists). Rotations of the ends of these sheathing
panels caused by loading in adjacent joist bays have the potential to inflict
damage on the tile covering, and should be evaluated when testing the viability
of a given tile assembly.
Unlike the Robinson-Type Floor Tester, the
design of the Universal Floor Tester is such that it has the capability to load
larger test specimens; specifically, flooring assemblies that can be constructed
using full size joists, along with subfloor and underlayment sheathing that
includes seams constructed per building code requirements (i.e., two-span
minimum for sheathing panels). The ability to test such specimens represents an
opportunity for important research and improved evaluation of ceramic tile
installation systems.
One such opportunity for research is further
investigation into the relative effects of joist stiffness versus sheathing
stiffness with regard to ceramic tile assembly performance. For example, it is possible to test flooring
systems with similar joist spacing and sheathing thickness, but with different
joist stiffness or span. Such a test
program, based on Bretzfield and Woeste (2002), could confirm that maximizing
sheathing system stiffness is the optimum method to limit curvature, which can
increase the reliability of ceramic tile installations over wood-frame
construction.
Another opportunity for research is to evaluate
the effects of variation in placement of wood underlayment joints on the
performance of the tile assembly under load.
Such a test program that evaluates the results from Woeste and Nielsen
(2004), which were based on engineering analyses, may validate the hypothesis
that placing wood underlayment joints at one joist span plus a ¼ joist span
from the subfloor seam is optimum for preventing localized bending stresses
from causing cracks in the tile covering.
The use of large formats (e.g., 18" x
18", 24" x 24", etc.) has become the norm in the tile industry. The latest trends include sizes up to
24" x 48" and larger. Given
that the Robinson-Type Floor Tester pad is only 4 ft x 4 ft in size,
limitations with regard to such “giant” format tile must be considered before
evaluating the use of these tiles over wood-frame construction. It is reasonable to assume that as tile
format increases, sensitivity to curvature will increase as well (see Figure
5).
Figure 5 - Sensitivity
to Curvature Increases with Tile Format
Compared to smaller format tiles, a higher
proportion of these “giant” format tiles will span joints in the
subfloor/underlayment and transitions from bearing walls to adjacent
joists. As a result, it will be
important to evaluate sensitivity of the tile to such discontinuities in the
sheathing and framing. As tile size
continues to increase, the Universal Floor Tester will allow for the use of
larger specimens, ensuring that representative samples are tested.
Thus, the Universal Floor Tester has the
potential to improve the industry’s understanding of the challenges associated
with setting ceramic tile over wood-frame construction. Further, the results obtained from the
testing discussed above will allow manufacturers and industry organizations to
provide practical recommendations to tile setters to help improve the
durability and reliability of their tile installations.
Ceramic
Tile Assemblies for Installation over Concrete Substrates
Concrete substrates may consist of
cast-in-place, precast, prestressed, and post-tensioned slabs. These slabs may be located below, on, or
above grade. The common thread that
binds all these substrates is the material characteristics of the concrete
itself.
While concrete slabs above grade are subject to
bending and deflection under loading, the most widely held concern regarding
the installation of ceramic tile over concrete substrates is movement
associated with cracks and joints. For
example, shrinkage cracks in concrete slabs are locations of reduced stiffness. Thus, when the concrete expands and contracts
with changes in temperature and moisture, the associated movement can be
concentrated at the weak plane of the crack.
This behavior is similar to control joints, which are actually provided
to help absorb movement and prevent randomly located shrinkage cracks. As a crack or joint opens, it can cause
tensile stresses to develop in the tile covering in direct bond
applications. The result can be cracked
tile, delaminated tile, and/or cracked grout joints.
The System Crack Resistance Tester is manually
operated to simulate the opening of a crack below a tile assembly. The assembly is evaluated based upon the
“crack” width reached before the tile cracks or debonds from the substrate. Thus, the System Crack Resistance Test Method
provides a basic simulation of the action of a crack in a concrete slab and its
potential effects on the tile assembly, which has been absent from the industry
until now. However, the Universal Floor
Tester has the potential to allow for more advanced simulations.
For example, the Universal Floor Tester is a
fully automated machine that enables the operator to control crack opening
width in a more precise fashion than manual controls allow. Thus, any variations due to operator are completely
eliminated. Another advantage of the
automated control is that it allows for cycled opening and closing of the crack
over time without continuous operator involvement. This feature would make investigation of the
effects of cycling or accelerated aging more reasonable. Finally, future research into the combined
effects of crack propagation and loading is possible.
Summary
When evaluating the results of any given test,
it is important to understand how the test method incorporates aspects of the
tile assembly in service. To do so, one
must first understand the various aspects of the existing structure that
supports the tile assembly and how it influences the performance of the tile
assembly. The performance of a tile
assembly depends upon how effectively the installation system integrates the
tile covering with the existing flooring structure.
The Universal Floor Tester can provide the tile
industry with the means to evaluate tile assemblies for installation over wood
substrates more completely than the Robinson-Type Tester by incorporating full
span joists, sheathing seams, and larger test specimens. With regard to tile assemblies for
installation over concrete, the Universal Floor Tester expands upon the
capabilities of the System Crack Resistance Tester by allowing for more precise
control and the ability to simulate accelerated aging by means of cycled crack
opening and closing.
In essence, the Universal Floor Tester has the
potential to focus future research on the most important areas to provide
practical information regarding tile installation, so that consumers are
provided with the most durable, reliable installations possible. The end result is an improved perception of
ceramic tile performance and overall growth of the ceramic tile industry.
References
ANSI A118.12, “American national standard
specifications for crack isolation membranes for thin-set ceramic tile and
dimension stone installations.” ANSI A118.12-2005, American National Standard Specifications for the Installation of
Ceramic Tile, 2005, Tile Council of North America (TCNA), 100 Clemson
Research Boulevard, Anderson, SC, USA.
Note: At
the time of submission of this article, the ANSI A118.12 standard was approved
by the Accredited Standards Committee on Ceramic Tile A108, but not yet
reviewed by ANSI for final publication.
As such, release date may vary from above.
ASTM C627, “Standard test method for evaluating
ceramic floor tile installation systems using the Robinson-Type Floor Tester.”
ASTM C627-93 (1999), Annual Book of
Standards, Volume 15.2, 2001, American Society for Testing and Materials,
Bretzfield, K. T. and Woeste, F. E., “Joist
curvature verses sheathing curvature and the probable role of each on ceramic
tile performance.” TTMAC Hardsurfaces
Magazine, 2003, Terrazzo, Tile and Marble Association of
Osterberger, T., Mayhew, S., and Fine, S.,
“Ceramic Tile and Engineered Wood Floors: Dispelling the Myths.” Coverings 2004
tradeshow presentation,
TCNA, 2005
Handbook for Ceramic Tile Installation.
Tile Council of
Woeste, F. E. and Nielsen, P., “Position of
Underlayment to Prevent Cracked Tile and Grout.” TileLetter, June 2004, National Tile Contractors Association,