Thomas E. Remmele1
Abstract
Water leakage into wall construction is a problem that has been occurring for many years and seems to be increasing. The result is damage to components of wall construction and costly repairs. One of the principal sources of leakage is residential windows in wood frame construction. Although building codes call for windows to be flashed at the sill which would help to control window leakage and prevent damage, the installation of flashing beneath residential windows is not practiced by builders. In many cases windows are replaced or the cladding is removed to repair underlying water damage within the first ten years of service. This paper explains the need for window sill flashing in residential wood frame construction and shows a material, engineered PVC flashing, and method for installing the flashing beneath windows without having to remove or replace them. Laboratory and field test results, and limitations of the flashing are presented.
Introduction
Flashing is a component of wall construction that is sometimes overlooked, yet it is vital for the long term successful performance of all exterior wall assemblies, including vinyl or wood siding, brick veneer, conventional stucco, and EIFS (exterior insulation and finish systems), also known as synthetic stucco. By definition, flashing is “a thin impervious material placed in construction......to
prevent water penetration and/or provide water drainage, especially between a roof and wall, and over exterior door openings and windows.1” A standard rule of thumb for the installation of flashing is that it should terminate in daylight, meaning that it should be bent, folded or configured so that when placed in wall construction, it functions to direct water to the wall exterior, not into the wall cavity.
__________________
1 Manager, Technical Services, Sto Corp., 6175 Riverside Drive, S.W., Atlanta, GA 30331
The need for flashing beneath residential windows
Recent history: In August of 1995 a local building inpsector in Wilmington, North Carolina discovered a pattern of water damage to wood frame homes clad with EIFS. The damage varied from minor water stains on sheathing to structural rot of wood sheathing and, in the most severe cases, rot of wood framing members. The damage was the result of water intrusion into the wall cavity and an inability of the wood materials in the structure to dry readily. The inability to dry was a function of the excess amounts of water getting into the wall, the wet, humid climate, and the makeup of the wall assembly: EIFS on the outside of 50 x 100 mm (2” x 4”) wood frame/wood sheathed construction, batt insulation in the stud cavity, a vapor retarder and 13 mm (1/2”) drywall on the inside (figure 1).
Figure 1. Typical wall construction with EIFS wall covering in Wilmington, NC
One of the main sources of water intrusion into the wall cavity was window leakage. In fact, the local AIA (American Institute of Architects) chapter led a task force that gathered statistics on the sources of water intrusion and found that roughly 2/3 of the windows surveyed leaked into the wall cavity. Sto Corp., a manufacturer of the EIFS used on many of the homes, conducted its own survey in its efforts to identify the sources of damage, and similarly found windows to be a major source of leaks into wall construction.
To gain more insight into the window leakage phenomenon Sto Corp. conducted tests through an independent testing agency to learn whether or not windows typically leaked into wall cavites and if the window leakage phenomenon in Wilmington was unique. While not having the resources to test a broad range of windows, Sto Corp. tested an expensive brand name vinyl clad twin casement style wood window and a much less expensive brand name double hung
wood brick mold window. The windows were installed in accordance with the written instructions provided and then tested in accordance with ASTM E 547, Standard Test Method for Water Penetration of Exterior Windows, Curtain Walls
and Doors by Cyclic Static Air Pressure Differential. This test subjects the window to static water spray at a prescribed flow rate from a spray rack while a prescribed pressure is applied. The test duration is 5 minutes and then the test is repeated two times for a total of 15 minutes of exposure to the test conditions. The required performance level is that “there is no water leakage through the window into the wall cavity or the room.2” The results of the tests are shown in table 1.
Test Pressures in Pa (psf)Leakage in Liters (Gal): Wood brick mold windowLeakage in Liters (Gal): Vinyl clad casement window137 (2.86)0.49 (0.13)0.0 (0.0) 212 (4.43)1.47 (0.39)0.0 (0.0)287 (6.0)2.38 (0.63)0.026 (0.007)
Table 1. Measurement of water leakage through windows
Both windows leaked into the wall cavity. This information coupled with the field investigations in Wilmington led us to the conclusion that water penetration resistance of residential windows was unpredictable at best.
Current window water penetration resistance standards: Window performance standards today acknowledge the fact that window leakage can or should be expected into wall cavities. The American Architectural Manufacturers Association and the National Wood Window and Door Association (AAMA/NWWDA) co-publish performance standards for windows entitled,Voluntary Specification for Aluminum, Vinyl (PVC) and Wood Windows and Glass Doors. This document, designated AAMA/NWWDA 101/I.S. 2 - 97, explains that “the perimeter plane of water and air penetration resistance4” for windows coincides with the face of the mounting flange or back surface of the brick mold (figure 2).
Figure 2. Typical mounting flange window with dotted line showing plane of water penetration resistance
The standard requires no leakage “into the plane of the innermost face of the test specimen....nor water penetration through the frame of the test specimen5.” The current window industry interpretation of this standard is that leakage outbound of
the perimeter plane of water penetration resistance (or outbound of the mounting flange) is not considered to be a window leak. The amount of outbound leakage that can be expected is typically not indicated on window manufacturer literature, nor is there any criteria in performance standards for minimizing the amount of leakage. Although the window industry considers this leakage “normal” it can and does lead to material degradation of components within the wall cavity--corrosion of metal ties in brick veneer walls, corrosion of metal lath in conventional stucco assemblies, and rot of wood sheathing in vinyl, wood or EIFS assemblies. Even with the installation of a moisture barrier over sheathing, degradation of the moisture barrier (typically one layer of Grade D building paper) in any of these assemblies will occur over time depending on the amount of window leakage, climate, and the overall makeup of the wall assembly.
Building codes: The Application and Commentary Manual for The CABO One and Two Family Dwelling Code, which governs most residential construction in the United States, depicts sill flashing beneath windows and states that it is “required to minimize leakage in the interface between wall coverings and other materials.3”
This statement and the accompanying illustration from the Manual leave no doubt that window sills must be flashed, yet flashing is rarely installed in residential construction.
Performance requirements and flashing design
Given the dilemna posed by leaking windows in residential wall construction, a product and method of repair had to be identified to control window leakage. Although flashing is not usually thought of as an engineered component of a wall assembly, the idea of engineering the flashing to perform up to a minimum specification that exceeded the water penetration resistance performance of the window was one of the principal design parameters in its development. Residential windows are graded for structural design pressure. They are required to have a minimum rating of 15 which means the structural design pressure is 720 Pa (15 psf). Water penetration resistance test pressure is taken as 15 percent of structural design pressure and not less than 137 Pa (2.86 psf). Thus, for a window with a structural design pressure rating of 1920 Pa (40 psf), the test pressure for water penetration resistance is 290 Pa (6.0 psf). Most residential windows do not exceed a 1920 Pa (40 psf) design pressure and therefore would not be required to exceed a water penetration resistance test pressure of more than 290 Pa (6.0 psf). Therefore the minimum performance threshold chosen for the flashing to successfully demonstrate control of window water leakage was 290 Pa (6.0 psf). The standard value used for water penetration resistance of EIFS is 299 Pa (6.24 psf), so this value was used as the final target performance level for the engineered flashing . At this minimum test pressure level it was determined that the flashing would have to demonstrate control of window leakage.
The final configuration of the flashing also had to be such that it could be installed beneath windows from the exterior without removing the windows, as well as capture and discharge window water leakage to the exterior. The configuration that was developed was a block and a wedge built in one which incorporated a slope for positive drainage and internal fins that served as end dams (figure 3). The flashing material, PVC (polyvinyl chloride), was chosen because a combination of molded and extruded parts could be manufactured to assemble the finished flashing into a single unit near identical to the configuration originally envisioned by the inventor. In addition PVC was relatively easy to cut in the field, non-corrosive, paintable, and could be made to be UV (ultra-violet) resistant.
Figure 3. Rigid PVC Engineered Flashing with internal slope and built-in end dams.
Testing
Testing was performed at an independent lab to evaluate the performance of the flashing. Results are reported in table 2.
Test Pressure in psf (Pa)Leakage in L (Gal) Loc. #1Leakage in L (Gal) Loc. # 2Leakage in L (Gal) Loc. # 3Leakage in L (Gal): total00000137 (2.86)0immeasurable00299 (6.24)01.60 (0.423)3.35 (0.885)4.95 (1.31)575 (12.00)03.85 (1.02)8.10 (2.14)11.95 (3.16)862 (18.00)0.255 (0.067)9.39 (2.48)16.5 (4.36)26.1 (6.91)
Table 2. Results of Engineered PVC Flashing water penetration resistance testing.
The test followed ASTM E 331, Standard Test Method for Water Penetration of Exterior Windows, Curtain Walls, and Doors by Uniform Static Air Pressure Difference. The test is similar to method E 547 described above except the test pressure is uninterrupted for the 15 minute test interval. The flashing successfully controlled window leakage up to 12.00 psf (575 Pa) without allowing water entry into the wall cavity, where leakage was defined as window water leakage that infiltrated the wall cavity outbound of the window’s plane of water penetration resistance, i.e., “normal” window leakage (location no. 1 in table no. 2).
Installation
The most challenging aspect of installation was the development of a safe and reliable technique to remove wood framing material to create a slot beneath the window to receive the flashing. The method uses a drill guide and bit to make successive holes in the sill framing . The final configuration of the slot with installed flashing is shown in figure 4. Figure 5 shows the flashing being installed.
Figure 4. Plan view of slot for installation of pvc engineered flashing.
Figure 5. Installation of PVC engineered flashing into slot beneath window.
Case study
A homeowner who had a leaking window problem in an EIFS clad residence in Wilmington, NC called upon Sto Corp. and a local contractor to repair the windows. The residence had double hung vinyl clad wood windows. Initial moisture surveys of the home indicated windows with elevated moisture readings (above 20%) in the wood sheathing beneath them. Within three weeks of the installation of the flashing moisture levels receded to acceptable levels. The results indicated that wood sheathing with elevated moisture levels dried to
acceptable levels once the source of leakage was stopped and controlled, even in a wet, humid climate, on homes with low peremeability exterior cladding (EIFS) and
an interior vapor retarder (figure 6). The drying took place via diffusion through the EIFS wall covering and, we believe, some degree of drying to the interior.
Figure 6. Moisture readings in wood sheathing after installation of engineered PVC flashing. Moisture content below 20% is considered acceptable.
Limitations
The engineered flashing will not control all forms of window leakage. It was designed for repair situations with the goal of repairing the predominant form of window leakage into wood frame wall cavities, i.e., leakage outbound of the face of sheathing (or the window plane of water penetration resistance), without removing the window. Windows may leak in other ways too. Some leaks occur within the rabbets of windows and accumulate between the panes of glass, while other leaks may be driven by pressure differentials to the interior of the window’s plane of water penetration resistance before they ever find their way to the flashing. Our experience has shown that these types of leaks occur with less frequency than the “normal” type of window leak. Although the PVC engineered flashing can be installed in traditional stucco, vinyl and wood siding walls, it usually cannot be installed beneath windows in brick construction because the brick sill will interfere with the nosing of the flashing. Windows with integral mounting flanges require special consideration because the installation of the flashing requires removal of the sill mounting flange. In most cases single unit mounting flange windows do not need the attachment at the sill to resist design wind pressures. Double unit windows usually do require the attachment. In such cases an alternate method of attachment (such as screw attachment through the window sill behind the sash into framing) can usually be accomplished, however the window manufacturer should be consulted in these cases.
Summary
In summary, engineered PVC flashing for repair of leaking windows in wood frame construction is an effective way to control window leakage and to stop or avoid water damage to wall components. Both laboratory and field tests have proven to be successful in controlling leaks from windows. The installed cost of the flashing is about 1/3 the cost of window replacement. The installation of the flashing requires special skills and tools and should not be performed by the do-it-yourselfer. Remodeling contractors or other specialty repair contractors are the tradespeople who would be the most qualified to install the flashing. Installation involves the removal of wood framing material beneath the window to create a slot to receive the flashing. The placement of a seal within the slot before the flashing is inserted and an outer perimeter seal after the flashing is inserted in the slot are also part of the installation sequence. The flashing is paintable and generally does not detract from the appearance of the window unit. It will control the predominant form of window leakage although it may not control all forms of window leaks. It is generally suitable for use with wood, vinyl or stucco siding but not with brick veneer. Mounting flange windows may require special consideration and the involvement of the window manufacturer before making a decision to install the flashing.
Footnotes
1 Cyril M. Harris, ed., Dictionary of Architecture and Construction (New York, 1993), p. 340.
2 NWWDA Industry Standard I.S. 2-93, Wood Windows, (Des Plaines, 1993), p.
3 Application and Commentary CABO One and Two Family Dwelling Code 1995 Edition, (Country Club Hills, Whittier, Birmingham, 1996), p. 134.
4 Voluntary Specifications for Aluminum, Vinyl (PVC) and Wood Windows and Glass Doors (Schaumburg, Des Plaines, 1997), p. 6.
5 Ibid.
References
Application and Commentary CABO One and Two Family Dwelling Code 1995 Edition. Country Club Hills, Whittier, Birmingham: International Code Council, Inc., Building Officials and Code Administrators, Inc. International Conference of Building Officials and Southern Building Code Congress International, Inc., 1996.
Harris, Cyril M., ed. Dictionary of Architecture and Construction. New York: McGraw-Hill, 1993.
NWWDA Industry Standard I.S. 2-93, Wood Windows. Des Plaines: National Wood Window and Door Association, 1993.
Voluntary Specifications for Aluminum, Vinyl (PVC) and Wood Windows and Glass Doors. Schaumburg, Des Plaines: American Architecutral Manufacturers Association and National Wood Window and Door Association, 1997.