Roofing Boards: The Next Generation

Roofing Boards:
The Next Generation

By Adriana Galli, Engineer
INTRODUCTION The poor performance of one roofing compo¬
nent often leads to failure of the entire
waterproofing system. Thus, roof consultants
and building owners are realizing that the
insistence on durable components throughout the
assembly plays an important role in maximizing the
structure’s life cycle.
Finex Inc., a company in the business of develop¬
ing proprietary non-combustible roofing systems,
believes that the use of fiber cement underlayment/
overlayment panels in roofing assemblies pro¬
vides potential lifecycle advantages over traditional
wood, gypsum, and even mesh-reinforced concrete
boards. The company hired Advanced Building
Materials to conduct practical comparison testing that
would closely represent the jobsite and in situ reality.
In order to confirm the performance characteristics
of fiber cement roofing boards, important field perfor¬
mance criteria, such as flute spanability and mem¬
brane adhesion, were comparison tested with other
roof support boards. These field performance criteria
were chosen to measure the resistance of roofing
boards to cracking and damage during traffic, and the
membrane compatibility under various climatic condi¬
tions.
The results of both the flute spanability and peel
adhesion strength testing clearly demonstrated the
superiority of fiber cement over gypsum and mesh
reinforced concrete boards.
The test results indicate that fiber cement boards
should enhance the roof assembly’s life cycle when
used in the following applications:
Underlayment (deck sheathing): Over steel deck to
resist fire within the building and limit the spread of
flame along the underside of the roof assembly (in
either a protected membrane or conventional roof
assembly)
Overlay board: To protect the insulation from
being crushed by normal foot or equipment traffic.
The fiber cement board can be adhered to the insula¬
tion or mechanically fastened over low compressive
strength insulation.
Separation board: To keep the new roof separated
from the old roof when chemical incompatibility
between the systems is encountered, e.g., covering a
standard BUR with PVC membrane.
Recover board (overiayment): Over an existing
roof to provide an acceptable surface to receive the
new membrane (commonly used when a new mem¬
brane is installed over an existing BUR system).
FLUTE SPANABILITY TESTING
PROCEDURE AND APPARATUS
The objective of this test was to
verify the flute spanability perfor¬
mance of fiber cement boards—
48″x96″xl/4″ in comparison to treated
and reinforced gypsum board—
48″x96″xl/4″, treated and reinforced
gypsum board—48″x96″xl/2″, regular
gypsum roof board—48″x96″xl/2″,
and mesh-reinforced concrete—
36″x72″x3/8″. Testing was conducted
for a metal roof deck with a rib open¬
ing of 2-5/8″ and a flute spacing of
6″, and metal roof deck with a rib
opening of 6-7/16″ and a flute spac¬
ing of 12″.
A roof assembly of 22 ga. metal
deck and a mechanically-attached
underlayment board was subjected to
the weight of a wheelbarrow filled
with a given load. The load (i.e. a
concentrated dynamic load) was
applied to the center of a rib opening
until ultimate value was achieved.
(Refer to Figures 1 and 2. This was
accomplished by increasing the load
in increments of 10 lb. until failure
occurred. The following controls
were used:
▼ The contact surface area was
maintained within ±10%; i.e.,
width of the wheelbarrows
wheel was equal to 4″ ±0.4″.
▼ The same area (at rib opening)
was subjected to the applied
load only once.
▼ The relative humidity and tem¬
perature were monitored.
The fastening pattern, the board
dimensions and the staggering of the
joints were equivalent for all speci¬
mens that were tested. Refer to
Figure 3.
Once the ultimate load was
achieved, the test was repeated a
minimum of six times to ensure the
accuracy of the results. The failure
load was within ±10% repeatability.
September 1997 Interface • 15
Figure 1—Detail of location of load on the roof support board.
The failure (ultimate) load as well as
the type of failure that occurred were
recorded. Failure was characterized
as cracking, warping and/or collapse
of the support board at the rib open¬
ing when subjected to the ultimate
concentrated dynamic load.
The testing was conducted under
the following two conditions:
▼ Equilibrium: Interior Ambient
Relative Humidity and Temperature
(for both metal decks with a
flute spacing of 6″ and 12″).
Refer to Table 1 for the test
parameters.
▼ 1 Hour Exposure of Simulated
Rainfall (for the metal deck
with a flute spacing of 6″ only).
Refer to Table 1 for details about
the test parameters.
The apparatus utilized for the test
was a roof assembly that consisted of
a structural framing of 8’xl6’ onto
which a 22-ga. metal deck (with a 6″
and 12″ flute spacing) was installed.
The underlayment boards were
mechanically attached to the metal
deck with roofing fasteners and 3″x3″
square galvanized metal plates. A
wooden platform was placed next to
Figure 2—Photo of wheelbarrow.
the roof assembly (parallel to the
direction of the flutes) and level with
the height of the roof assembly. This
would allow for a smooth transfer of
load onto the deck. The dynamic
concentrated load was applied with a
conventional wheelbarrow with a 15″
diameter and 4″ wide wheel. (Refer
to Figure 4).
FLUTE SPANABILITY
TEST RESULTS
Table 1
Test Parameters for Flute Spanability Tests
Test
Condition Products Tested
Product
Dimension
Dimension of
Metal Deck’s
Flute Spacing
Equilibrium • Fiber cement boards (1/4″)
• Treated & reinforced gypsum board (1/4″)
• 4’x8′
• 4’x8′
6″*
Equilibrium • Fiber cement boards (1/4″)
• Treated & reinforced gypsum board (1/2″)
• Regular gypsum roof board (1/2″)
• Mesh-reinforced concrete board (3/8″)
• 4’x8′
• 4’x8′
• 4’x8′
• 3’x6′
12″
1 Hour of
Simulated
Rainfall
• Fiber cement boards (1/4″)
• Reinforced gypsum board (1/4″)
• Reinforced gypsum board (1/2″)
• Regular gypsum roof board (1/2”)
• Mesh-reinforced board (3/8″)
• 4’x8′
• 4’x8′
• 4’x8′
• 4’x8′
• 3’x6′
6″*
Test results for underlay¬
ment boards at the equilib¬
rium condition are shown in
Figure 5 for the 2-5/8″ rib
opening deck. Figure 6 sum¬
marizes the results obtained
with the 6-7/16″ rib opening
deck. The results obtained
when roof support boards
were subjected to 1 hour of
simulated rainfall are
recorded in Table 2. Figure
7 shows the results obtained
with the 2-5/8″ rib opening
deck.
ADHESION TESTING
PROCEDURE AND
APPARATUS
* Treated and Reinforced Gypsum Board /1/4″) was not tested with the deck that has a flute
spacing of 12″ because, as specified in the literature, it is not recommended to install the prod¬
uct on a metal deck with a 12″ flute spacing. In addition, the treated and reinforced gypsum
board (112″) and the mesh-reinforced concrete board (3/8″) were not tested on the metal deck
with a 6″ flute spacing because the objective of this study was to test the worst case scenario.
The objective of this test
was to determine how vari¬
ous climatic conditions that
involve moisture and tem¬
perature changes would
affect the adhesion between
16 • Interface September 1997
Figure 3—Fastening patterns of roof support boards.
Figure 4 – “Wheelbarrow used for flute spanability tests. Figure 5—Bar chart of results from Table 2 with 2-5/8″ rib
opening.**
exposed to cold winter conditions.
Condition C— 24 hour saturation in
water before application of membrane.
This condition assumed a situa¬
tion where moisture would be intro¬
duced into the system during instal¬
lation before the primer and/or mem¬
brane was applied to the wet boards.
An example of such an occurrence
can be if the roof support boards
were installed and then left exposed
to rainfall or snow.
primed and unprimed roof support
boards and bituminous membranes.
The following four test conditions
were designed to simulate some of
the most common occurrences:
Condition A—Ambient tempera¬
ture and relative humidity.
This condition assumed an ideal
situation for the membrane to be
applied to the roof support boards
where moisture or extreme tempera¬
ture gradients were not present.
All boards tested on 6″ flute span.
8 ft
Condition B— 30 minute satura¬
tion in water at 5°C before applica¬
tion of membrane; followed by 65
hours of exposure to -10°C after
application of membrane.
This condition assumed a situa¬
tion where the waterproofing mem¬
brane was applied to roof support
boards that were exposed to a rainfall
or snow. After the application of the
primer and/or the membrane to the
wet boards, the system was left
All boards tested on 12″ flute span.
8 ft
Mesh-reinforced board tested on 6″ flute span.
6 ft
Mesh-reinforced board tested on 12″ flute span.
6 ft
September 1997 Interface • 17
Table 2
Flute Spanability Test Results
(Underlayment boards subjected to 1 hour of simulated rainfall)
** Since a pneumatic tire was used on the wheelbarrow, an exact relationship to
failure load does not exist. Therefore, the load is not directly comparable between
charts and the load is not indicative of the true values on the boards.
2-5/8″ rib opening
Product
Load at which failure
occurred (Ib./sq. in.) Description of Failure
Fiber cement board (1/4″) 54 (666 lbs.) Board cracked along point
of load application
Treated & reinforced gypsum
board (1/4″)
24 (300 lbs.) Wheelbarrow fell through
the substrate
Treated & reinforced gypsum
board (1/2″)
37.3 (466 lbs.) Gypsum core crushed
Regular gypsum roof board (1/2″) 12.2 (153 lbs.) Gypsum core crushed
Mesh-reinforced board (3/8″) 37.3 (466 lbs.) Concrete core spalled
UNDERLAYMENT BOARD AT EQUILIBRIUM CONDITION
WITH A 6-7/16″ RIB OPENING DECK
– 50
O 30
§ 10
g °
Figure 6—Bar chart of results with 6-7/16″ rib opening.**
Reinforced
Gypsum
Board
Reinforced
Concrete
Board
Cement
Board
(1/4-)
Gypsum
Roof Board
(1/2″)
Reinforced
Gypsum
Board
PRODUCT
Figure 7 – Bar chart of results from Table 2 with 2-5/8″ rib opening.**
Condition D— 24 hour saturation
in water after application of mem¬
brane.
This condition assumes a situation
where moisture would be introduced
into the roof system due to failure of
the roof membrane system; conden¬
sation within the insulation resulting
from the lack of a vapor retar¬
der; or inadequate insulation R-value
where high interior moisture levels
occur.
The roof support boards tested
were fiber cement boards—
48″x96″xl/4″; treated and reinforced
gypsum board—48″x96″xl/4″; treated
and reinforced gypsum board—
48″x96″xl/2″; regular gypsum roof
board—48″x96″xl/2″; and a meshreinforced
concrete board—
36″x72″x3/8”. To increase the statisti¬
cal significance of the test, a series of
eight samples was tested for each
roof support board (both primed and
unprimed) and for each test condi¬
tion specified below.
Part I of the test procedure con¬
sisted of sample preparation. An SBS
modified bitumen membrane base
sheet with a non-woven polyester
reinforcement was torch applied to
4’x8’ boards of fiber cement boards,
treated and reinforced gypsum board
(1/4″ and 1/2″), regular gypsum roof
board (1/2″), and 3’x6’ boards of the
mesh-reinforced concrete board
(3/8″) with and without a penetrating
asphalt primer under the four test
conditions described above.
Part II consisted of peel tests con¬
ducted by separating the membrane
from the underlayment board until
the separation angle between the
membrane and the underlayment
board was 180°. This was executed
by cutting samples (from the full
scale samples prepared) of 2″x6″ with
a 6″ overlap of membrane at one end
(refer to Figure 8) and placing them
in the apparatus described in Figure
9 and Figure 10. In this tensile tester,
the substrate was clamped at one
end and the membrane was clamped
at the other end. However, before
placing the specimens in the appara¬
tus, they were undercut with a sharp
blade along the shorter edge of the
substrate where the membrane over¬
laps the support board to produce a
proper separation at the interface of
the membrane and the roof substrate
and to prevent premature failure of
the roof substrates during the test.
The rate-of-jaw separation to remove
the membrane from the substrate
was 3 in./min and readings were tab¬
ulated every 5 seconds in pounds per
minute until failure occurred. The
final results were recorded in Ib./in.
of sample width.
The tests were conducted in
accordance to ASTM D903,
“Standard Test Method for Peel or
Stripping Strength of Adhesive
Bonds” and ASTM D1970-78,
“Standard Specification for Self¬
Adhering Polymer Modified
18 • Interface September 1997
Bituminous Sheet Materials Used as
Steep Roofing Underlayment for Ice
Dam Protection.” These test meth¬
ods were chosen in order to quantita¬
tively evaluate the results for peel
strength.
The mode of failure is character¬
ized according to ASTM as follows:
▼ Cohesive Failure—when the
membrane breaks apart as it is
being pulled from the substrate.
▼ Adhesive Failure—when there
is a lack of adherence of the
membrane to the substrate it is
being bonded with.
▼ Substrate Failure—when there
is a weak link in the substrate
which causes it to come apart
when the membrane is being
pulled away from it.
The most favorable modes of fail¬
ure for roof support boards, in order
of preference, are cohesive failure,
adhesive failure, and the substrate
failure. The reason is mainly because
cohesive failure is an indicator that
there is a strong bond between the
membrane and the roof substrate.
This is a property that is of great sig¬
nificance to membrane manufactur¬
ers. Adhesive failure is still possible
Figure 9—Schematic of power-driven
machine per ASTM D 903-93 for
adhesion (stripping strength).
along this edge TOP VIEW
Membrane
at this location Roof support board –
SIDE VIEW
Figure 8—Dimensions of specimen for adhesion tests.
due to poor adhesion during torching
of the membrane. The most unfavor¬
able mode of failure is substrate fail¬
ure, since it is an indication that
there is a weak link in the roof sup¬
port board.
The apparatus utilized for this
test was a power-driven machine
(refer to Figure 9) that maintained a
constant crosshead speed of 3
in./min. throughout the tests. In this
pendulum-type machine, the weight
lever swung as a free pendulum
without engagement of pawls. The
applied tension as measured and
recorded was accurate within ±1%.
The specimens were held in the
testing machine by grips which
clamped firmly and prevented slip¬
ping at all times. The machine
capacity was such that the maximum
applied tension during the test proce¬
dure did not exceed 85% nor was it
less than 15% of the rated capacity.
ADHESION TEST RESULTS
All products tested were abbrevi¬
ated as follows:
FCB(l/4″) = Fiber cement board,
1/4″ thick
RG (1/4″) = Treated and
reinforced gypsum
board, (1/4″) thick
RG (1/2″) = Treated and
reinforced gypsum
board, (1/2″) thick
G (1/2″) = Regular gypsum board,
(1/2″) thick
Table 3
Percent Advantage in Peel Strength (Ib/in)
of Fiber Cement Boards to other Products
Fiber cement
board vs:
Test A Test B Test C Test D
(P) (UP) (P) (UP) (P) (UP) (P) (UP)
RG (1/4″) 39.89 73.92 77.77 88.73 73.55 72.56 67.38 58.16
RG (1/2″) 42.70 71.90 76.14 60.23 60.17 81.67 70.29 67.91
G (1/2″) 43.00 78.35 54.52 63.93 82.19 61.92 100 100
MRC (3/8″) 43.12 54.48 25.79 33.33 40.15 23.59 32.82 82.27 |
September 1997 Interface • 19
Figure 10—Condition 71: Product (P) and (UP) Versus Peel Strength (Ib.Hn.).
Figure 11—Condition B: Product (P) and (UP) Versus Peel Strength (Ib.Hn.).
MRC (3/8″) = Mesh-reinforced
concrete board,
3/8″ thick
where (P) = Primed roof support
board, and (UP) =
Unprimed roof
support board
The peel strength data are sum¬
marized in the bar charts of Figure 10
to Figure 13. The data from Figure 10
to Figure 13 were used to calculate
the percentage increase in perfor¬
mance of fiber cement boards com¬
pared to the other roof support
boards (for both primed and un¬
primed boards) as shown in Table 3.
The mode of failure for each product
is shown in the referenced figures as
follows:
▼ Cohesive failure for fiber
cement boards as shown in
Figure 14.
▼ Adhesive and substrate failure
for treated and reinforced gyp¬
sum board (1/4″) as shown in
Figure 15 and Figure 16.
▼ Adhesive and substrate failure
for treated and reinforced gyp¬
sum board (1/2″) as shown in
Figure 17 and Figure 18.
▼ Substrate failure for regular gyp¬
sum roof board (1/2″) as shown
in Figure 19.
▼ Cohesive and adhesive failure
for the mesh-reinforced concrete
boards as shown in Figure 20.
Conclusions
FLUTE SPANABILITY
In every test condition, fiber
cement outperformed the other types
of materials. Common elements of
moisture and cold working condi¬
tions were considered. The fiber
cement boards widened the perfor¬
mance gap when compared to the
other materials.
In the most common deck module
of 6″, the fiber cement boards
demonstrated the most resistance to
failure when subjected to a concen¬
trated dynamic load from a wheelbar¬
row in the flute spanability tests.
The order of performance from best
to worst for the flute spanability
tests under an equilibrium condition
was the fiber cement board (1/4″),
followed by the treated and rein¬
forced gypsum board (1/4″).
The order of performance from
best to worst for a flute spanability
of 12″ (6-7/16″ rib opening) under an
equilibrium condition was the fiber
cement board (1/4″), treated and re¬
inforced gypsum board (1/2″), meshreinforced
concrete board (3/8″), and
regular gypsum roof board (1/2″).
The order of performance from
best to worst when tested on deck
with a flute spanability of 6″ (2-5/8″
rib opening) under a simulated rain¬
fall condition was the fiber cement
board (1/4″), treated and reinforced
gypsum board (1/2″) and the meshreinforced
concrete board (3/8″) at
par, treated and reinforced gypsum
board (1/4″) and regular gypsum roof
board (1/2″).
The test results confirm that in
applications where roof traffic and
loading from heavy equipment must
be considered, fiber cement panels
perform well.
MEMBRANE ADHESION
Based on the adhesion test
results, fiber cement boards unprimed
followed by fiber cement board
primed performed the best for Test
A and D, and Fiber Cement Board
primed followed by Fiber Cement
Board unprimed performed the best
for Test B and C. The most evident
mode of failure was cohesive failure
20 • Interface September 1997
Figure 12—Condition C: Product (P) and (UP) Versus Peel Strength (lb. jin-).
Figure 13—Condition D: Product (P) and (UP) Versus Peel Strength (lb. jin.).
and in some instances, adhesive fail¬
ure. The mesh-reinforced concrete
board produced the second highest
set of results for Tests A to D. Both
cohesive and adhesive failure were
noted, and substrate failure was
recorded for Test A only. The most
evident mode of failure for the treat¬
ed and reinforced gypsum board
(1/4″) and (1/2″) was substrate failure.
The second most evident mode of
failure was adhesive failure. For all
the test conditions (A to D) the regu¬
lar gypsum roof board samples
cracked at midspan either before or
during testing. Therefore, a smaller
population (approximately half com¬
pared to the other products) was
tested due to the large number of
samples that were not capable of
remaining intact. Due to the severe
nature of Test D, the core of the
gypsum samples was completely sat¬
urated and therefore, most of the
specimens cracked along the width
and then crumbled to pieces. The
remainder of the samples were so
saturated that the paper facing
delaminated from the gypsum core
and thus it was not possible to test
any of the specimens for Test D.
The most evident mode of failure for
all the specimens was substrate fail¬
ure. For those that were tested (from
Test A to Test G), the mode of fail¬
ure was also substrate failure where
the paper facing would peel away
from the gypsum core, taking with it
some of the gypsum core.
Consequently, the fiber cement
boards had the highest peel strength
with and without primer in compari¬
son to the other products. Fiber
cement board also provided the best
surface preparation in order for a
example) the substrate failure that
occurred with the treated and rein¬
forced gypsum board and the regular
gypsum roof board.
The test results confirm that in
applications where moisture may
eventually be present, or where roof
membranes are installed in incle¬
ment conditions, fiber cement panels
perform better than gypsum and
mesh-reinforced concrete board.
membrane to be
applied to it with
or without primer,
regardless of whe¬
ther fiber cement
boards are in a sat¬
urated or equilibri¬
um state. This fur¬
ther confirms the
recommendation
that priming is a
good practice, but
not necessarily
required before
applying a water¬
proofing membrane
onto its surface. In
addition, fiber ce¬
ment boards have
the most favorable
mode of failure
(cohesive failure)
compared to (for
About The Author
Adriana Galli graduated from
the Center for Building Studies at
Concordia University in Building
Engineering, and is a registered
engineer with the Order of Engineers
of Quebec. At a research and development engineer with
Materiaux de Batiment d’Avant Garde (Advanced
Building Materials), she did testing on fiber cement prod¬
ucts, including flat boards for roofing applications, and
corrugated profiles for roof deck applications. Prior to
joining AIBA, Galli worked as a technical and marketing
services manager with a fiber cement manufacturer. Galli
is currently the marketing services manager of Sika
Canada Inc., a manufacturer of construction chemicals
and cementitious products for concrete repair and restora¬
tion, located in Montreal.
September 1997 Interface • 21