Low-Slope Roofs are Rotting

Low-slope roofs are rotting,
resulting in major repair costs.
Recently investigated roofs on
three multifamily residential
buildings had failed prematurely.
All three buildings were less
than ten years old and were located in a
northern climate.
These buildings had similar nonventilated,
low-slope roof assemblies utilizing wood
trusses with a polyethylene vapor retarder
on the bottom of the trusses covered with
a gypsum ceiling. Blown-in fiberglass or
cellulose insulation filled the truss space
from the ceiling to the bottom of the roof
deck. Oriented strand board (OSB) of ½-in.
thickness was installed over the trusses as
the roof deck. Rigid board insulation was
installed over the OSB, followed by the roof
membrane, which was a gravel-surfaced,
built-up roof in one case; a ballasted EPDM
single-ply membrane on another; and a
mechanically fastened TPO single-ply membrane
on the third (Figure 1).
There was no water intrusion evident
in any of the buildings. A survey of the roof
surfaces showed that they were in good
condition. Occupants in two of the buildings
reported that they thought they had a mold
problem, which led to further investigation.
Maintenance people walked on the third
building roof and reported soft spots, which
turned out to be locations where the OSB
roof deck had lost its structural integrity
due to moisture degradation.
Invasive inspection openings made into
the roof assembly revealed the OSB decking
was severely deteriorated in many areas,
such that it could not support the roofing
materials above. The top portion of the top
cord of the wood
trusses was also
wet and rotted in
some cases. The
steel tie plates
of the trusses
were corroded.
Framing lumber
at the exterior
walls in the truss
space was also
wet.
The polyethylene
vapor
retarder was
turned down the
vertical face of
the exterior walls
and lapped with
the polyethylene
vapor retarder of
the wall, with the
overlap sealed
with plastic
tape. At partition
walls, the vapor
retarder was attached to the vertical face
of the top plate of the wall with a thin bead
of adhesive. Overlaps in the ceiling vapor
retarder were also sealed with plastic tape.
It was apparent that moisture-laden
air migrated into the truss space and condensed
in the upper reaches of the roof
1 6 • I n t e r f a c e J u l y 2 0 1 6
Figure 1 – Problem assembly.
assembly. So how did this moisture-laden air get into the
truss space and condense in quantities that caused such
extensive damage? The vapor retarder should act to minimize
the amount of moisture vapor from the interior of the
building getting into this space. In the northern climate
where these projects were located, the vapor drive in the
winter is mainly from the warm interior to the cold exterior.
The warm interior air carries moisture vapor that will
condense on the surfaces that are below the dew point, the
temperature at which condensation can occur.
The investigations found many bypasses in the vapor
retarder that would allow the warm, moist interior air to
migrate into the truss space (Figures 2-5). The party walls
between apartment units, which were a double-stud wall
J u l y 2 0 1 6 I n t e r f a c e • 1 7
Figure 2 – Rotted deck collapsed into truss space revealing
rotted top chord of truss and corroded tie plates.
Figure 4 – Vapor retarder bypass
at ceiling light electrical box.
Figure 5 – Vapor retarder bypass
through top plate of partition wall.
Figure 3 – Vapor retarder bypass
at partition and party walls.
construction, interrupted the continuity of
the vapor retarder.
Likewise, any interior partition walls
resulted in a discontinuity in the vapor
retarder at the ceiling. This was exacerbated
by the penetrations through the top
plate of the wall by plumbing stacks and
wiring. Penetrations through the ceiling,
such as sprinkler heads and electrical
boxes for light fixtures, were also found
not to be sealed to the vapor retarder.
Two of the buildings had ducts from
bathroom and dryer vents running
through the truss space (Figure 6). Some
of these ducts were not well sealed at the
joints, which introduced very moist air
directly into the space where condensation
was most likely to occur.
The roof assembly would likely perform
satisfactorily if the ceiling vapor retarder
were perfectly constructed. However, in
this type of construction, it was virtually
impossible to perfectly construct a vapor
retarder at the ceiling level due to all of
the discontinuities. Ventilation of the truss
space is not an effective option to manage the moisture.
Unlike a steep-slope attic space, there is very
little, if any, space to create airflow over relatively long
distances.
Traditional methods for moisture analysis of the
1 8 • I n t e r f a c e J u l y 2 0 1 6
Figure 6 – Leaky dryer and bathroom vent ducts in truss space.
Figure 8 – Recommendation for a roof system and parapet detail.
Figure 7 – Hygrothermal model graph of the roof deck in the
roof assembly shown in Figure 1.
Figure 9 – Hygrothermal model graph of the plywood roof
deck in the roof assembly shown in Figure 8.
exterior envelope, such as the dew point
method, the Glaser diagram, and the Kieper
diagram, are steady-state analysis tools.
These methods have significant limitations
due to the fact that wetting and drying
cycles cannot be accurately analyzed when
considering only a specific temperature at
one moment in time. These tools neglect the
moisture storage capacity in the building
materials and the transient effects of vapor
drive.
Hygrothermal modeling has become
more widely used over the past 20 years to
simulate the transient heat and moisture
conditions in roof and exterior wall assemblies.
Hygrothermal modeling, in contrast
to the traditional methods, looks at heat
and moisture conditions over time and can
take into account a number of variables.
Most importantly, it can show whether the
system has a propensity for moisture to
accumulate at levels that can result in rot,
mold, and corrosion.
Figure 7 is the hygrothermal model
graph of the roof deck in the roof assembly
shown in Figure 1, assuming the presence
of vapor retarder bypasses, over a fiveyear
period. The water content of the roof
deck exceeds 19% for about 40% of the
year, peaking in April. Calculating the dew
point using traditional methods by selecting
temperature and humidity conditions in
January may not have shown the potential
problem of excessive moisture in the
system that peaks in April. The moisture
damage to the roof deck, as observed in the
field, strongly correlates to the hygrothermal
model results.
Perhaps the motivation to design an
assembly as shown in Figure 1 would be to
minimize the insulation costs related to the
energy code requirements of recent years.
While the amount of insulation installed
exceeds the code requirement, the material
and labor to install it were less than a
code-compliant insulation installed above
the roof deck. Filling the truss space with
noncombustible insulation also provided
the opportunity to eliminate the need for
firestopping and draftstopping as noted in
the 2000 International Building Code (IBC)
in force at the time these buildings were
constructed.
A more constructible approach would be
to install a vapor retarder on the roof deck
level (Figure 8). Some insulation could be
installed in the truss space, but most of the
insulation would need to be installed above
the vapor retarder so that the dew point
occurs above the vapor retarder. Insulating
above the roof deck with rigid board insulation
would be more expensive than insulating
with blown-in insulation in the truss
space. However, eliminating condensation
is less expensive than costly repairs after
the fact.
Figure 9 is the hygrothermal model
graph of the plywood roof deck in the roof
assembly shown in Figure 8. This shows the
moisture content of the plywood roof deck
staying below 19% and actually drying out
over time from its initial peak moisture content
at the time of construction. Anecdotal
observations of roofs similar to this design
in place for 20 years or more substantiate
the hygrothermal model results.
Installing the vapor retarder at the roof
deck level affords a much better opportunity
for achieving a complete vapor retarder.
This would be a relatively easy way to provide
continuity across party walls and to
seal penetrations. The vapor retarder must
also be continuous from the roof to the
exterior walls. This might be accomplished
by using spray foam insulation within the
J u l y 2 0 1 6 I n t e r f a c e • 1 9
If you have a roof hatch that is damaged or leaking,
the VersaMount® is a simple, cost-eff ective solution.
With its low profi le frame and overlapping fl ange
and apron, the VersaMount installs easily over
existing curbs. A VersaMount roof
hatch will provide weathertight
performance, ease
of operation and years of
maintenance-free service
— all without touching the roof.
Roof Hatch Replacement
Made Easy…
With the VersaMount® Roof
Hatch, you don’t need to re-roof
VersaMount® Curb
• Standard fully enclosed curb with
built-in apron ensures a fast,
weather-tight installation
• Supplied with adjustable anchor tabs
for secure mounting
• Standard 4” curb height provides a
low profi le installation for end
user convenience
Existing Hatch Curb
Patented
For more information, call (800) 366-6530
or log-on to www.bilco.com
BLC1001 Versamount Interface Ad_.indd 1 6/14/16 2:20 PM
truss space at the exterior walls.
Recent investigations of three buildings
in a northern climate clearly demonstrated
the well-intended cost-saving measure to
insulate within the truss space resulted
in premature roof failure and expensive
repairs. Those repairs included complete
replacement of the roof covering and roof
deck, along with repairs to the structural
trusses and installation of new insulation.
This was an expensive lesson from which we
should all learn.
2 0 • I n t e r f a c e J u l y 2 0 1 6
Pamela Jergenson,
CCS, CCCA, is a
senior consultant
for exterior walls
with Inspec, a
building envelope
consulting engineering/
architectural
firm. She is
an expert in hygrothermal
analysis.
Pamela Jergenson,
CCS, CCCA
Dwight D. Benoy,
PE, has been
employed at Inspec,
a building envelope
consulting engineering/
architectural
firm, since 1974.
He is a licensed
professional engineer
in 30 states,
focusing his practice
in forensic engineering
of the building
envelope.
Dwight D. Benoy, PE
ISSUE SUBJECT SUBMISSION DEADLINE
October 2016 Product manufacturing July 15, 2016
November 2016 New mtrls. & constructions August 15, 2016
December 2016 Metal walls and roofs September 15, 2016
January 2017 Miscellaneous (bldg. envel.) October 14, 2016
February 2017 Codes and standards November 15, 2016
March 2017 Trends in wall & roof design December 15, 2016
Publish in Interface
Interface journal is seeking submissions for the following issues. Optimum article size is 2,000
to 3,000 words, containing five to ten graphics. Articles may serve commercial interests but
should not promote specific products. Articles on subjects that do not fit any given theme may be
submitted at any time.
Submit articles or questions to Executive Editor Kristen Ammerman at 800-828-1902
or kammerman@rci-online.org.
Under new rules by the U.S. Department
of Labor (DOL), the salary level under
which employees qualify for overtime pay
will increase from $455 per week ($23,360
annually) to an estimated $913 per week
($47,476 annually). The salary threshold
will update automatically every three years
to the 40th percentile of earnings of fulltime
salaried workers in the lowest-wage
census region (currently the South). The
rule amends the salary basis test to allow
employers to use nondiscretionary bonuses
and incentive payments to satisfy up to 10
percent of the new standard salary level.
The rule is effective December 1, 2016.
OVERTIME
THRESHOLDS
TO RISE
Oak Ridge National Laboratory
(ORNL) is searching for owners of commercial
buildings who will be retrofitting
their building envelope in the next
three years. Their goal is to gather data
that will allow them to estimate energy
savings and payback periods, given that
minimal information is available on this
subject. These data should be helpful to
owners who are considering retrofitting
their building envelopes.
ORNL will offer, free of charge,
blower door tests before and after the
retrofit, hygrothermal analysis of the
proposed renovation, simulations that
estimate potential energy savings, and
discounts on some air barrier and insulation
materials. For further information,
contact Diana Hun at hunde@ornl.gov.