INTRODUCTION Metal flashings are commonly used in buildings as parapet caps, at the windowsills and heads, and at horizontal junctions of building components. Water flow over vertical cladding is greatly affected by small changes in the geometry of horizontal projections. Figure 1 shows the water flow pattern at two different horizontal projections on the cladding of a building in downtown Toronto. Water staining can be seen below the parapet coping. At windowsill locations, a metal flashing with different drip geometry prevents this phenomenon by more effectively shedding water. Although the presence of water can be noted at two ends of the windowsill, there is no water below the drip, indicating that it is effectively shedding water. In two case studies reported by Doshi (2005), the rainwater flowing over cladding was greater than could be predicted by the effect of wind-driven rain deposition only. This was attributed to water flow over buildings due to ineffective shedding at horizontal projections. Maurenbrecher (1998) reports on water-shedding details that improve masonry and emphasizes the key role played by metal flashing. This paper details a test method that was developed to quantify the water-shedding effectiveness of metal flashing used in buildings. The test apparatus and the testing procedure developed for this purpose, as well as the parameters which impact the effectiveness of flashing, are described. BACKGROUND Literature Review Despite the importance of metal flashing in building construction, a literature search revealed that there are no consistent guidelines for the design of a metal flashing. Practice guides and building codes include some prescriptive information on metal flashing. This provides details about the material and thickness of flashing, but very little or no information on aspects that affect its water-shedding ability. The included information varies in content and detail. The Sheet Metal and Air Conditioning Contractors National Association (SMACNA, 1993), the National Roofing Contractors Association (NRCA, 2001), and the Canadian Roofing Contractors’ Association (CRCA, 1997) are examples of trade association whose manuals are frequently referenced by designers for components and installation specification details. These manuals include general information about the shape and size of metal flashing used for parapets, but they do not provide any particular information on the design of the water-shedding aspect of the flashing. The Aluminum Association (2000) manual also provides general information on metal flashing. Again, it does not provide Figure 1 – the effect of small changes in geometry of horizontal projections on water flow over cladding. 24 • I N T E R FA C E S E P T E M B E R 2006 any detailed information about the requirement pertaining to the geometry of the drip edge. Building codes (IRC, 1995) used in Canada contain information on sheet metal flashing. This information pertains predominantly to the type of materials used for flashings and their applicable thicknesses. Very little information is available that relates to the drip geometry. Recently, the Canadian Mortgage Housing Corporation (CMHC, 1997a, b) produced practice manuals that provide some prescriptive information pertaining to the design of drip. This information is not consistent among different cases. The literature survey was limited to North American sources. It has been brought to the attention of the authors that there might be some emerging ideas from Europe. The authors were unable to source the documents from Europe through many of the publicly available databases in Canada. It is expected that a review of these European practices will be conducted in future versions of this work. Review of existing practices In addition to the published information, the authors reviewed the practices of design offices and of contractors as it relates to the geometrical considerations of the drip edge of metal flashing. Both new construction and retrofit construction projects were examined (Construction Control Inc., 2004a and b; Quadrangle Architects Limited, 2004; and Young and Wright Architects Inc., 2004). For each of these cases, the drawings, the specifications, and the field practices were examined. It was noted that the drawing and specifications referenced some of the abovementioned trade publications for the flashing requirements. Discussion with the contractors revealed that the girth of the flashings, the angle of the drip, and the length of the drip were considered in forming the metal flashings. The actual value of these parameters was determined by the shops forming the metal, rather than the designers. Summary of Findings from Literature Review and Review of Existing Practices Test method to determine watershedding effectiveness of flashing The literature survey revealed that there is a lack of information on testing the water-shedding effectiveness of different configurations of metal flashing. The authors did not find any test method applicable to the measure of the performance of the drip edge of flashing. Variables affecting water-shedding performance of metal flashing Three parameters that impact the water-shedding effecttiveness of metal flashing that can be controlled by designers and fabricators were identified in reviewing the existing information. They are shown in Figure 2: 1. The angle of the drip-edge. 2. The length of the drip-edge or its horizontal projection from the wall. 3. The gap between the face of flashing and the vertical surface of the wall. Figure 2 – Flashing design parameters. PARAMETER SPECIFIED RANGE (OUT OF 3 SOURCES) Angle of drip edge 0 – Horizontal projection 3 10mm – 25mm Gap 2 25mm Table 3. Reported window sill flashing requirements. S E P T E M B E R 2006 I N T E R FA C E • 2 5 PARAMETER SPECIFIED RANGE (OUT OF 3 SOURCES) Angle of drip edge 0 – Horizontal projection 3 5mm – 10mm Gap 0 – Table 2. Reported through-wall flashing requirements. PARAMETER SPECIFIED RANGE (OUT OF 17 SOURCES) Angle of drip edge 4 45° Horizontal projection 9 13.5mm – 35.8mm Gap 0 – Table 1. Reported parapet flashing requirements. Other factors will impact the performance of the flashing that cannot be controlled by designers and fabricators. These include parameters related to the nature of winddriven rain such as the quantity of rain and the spatial and temporal nature of wind. The focus of this work is on studying the impact of controllable factors. Values for variables affecting water-shedding performance Based on a review of published information and a review of practices, values for the parameters affecting water-shedding performance were found to be as shown in Tables 1 to 3. The column titled “Parameter” lists the variables of interest as explained above. The next column, titled “Specified,” lists the number of sources reporting a value for the respective variable. The total number of sources reporting the particular flashing configuration is shown in the column heading. The last column provides a summary of the range of values. As can be seen from Tables 1 through 3, there is a wide variation among the existing requirements and practices of the three variables that impact the water-shedding effectiveness of metal flashing. In the following section, a test method is presented which attempts to study this effectiveness. PROPOSED TEST METHOD A test method was developed to study the impact of the three mentioned variables on the water-shedding effectiveness of a metal flashing. In order to do this, the following issues were considered. Test Considerations Efficiency measure Water flowing over a flashing should ideally be directed away from the building face. Some of it will flow back toward the building and run down the wall cladding below the flashing. The efficiency of a flashing profile is directly proportional to the amount of water that drips away from the drip-edge without coming in contact with the vertical wall surface below it. The greater this amount, the better the performance of the flashing. Equation 1 expresses the water-shedding efficiency of a flashing profile as a percentage: where E = the efficiency of a profile; O = the original amount of water flowing on the flashing; and Dw = the amount of water dripped back on the wall surface. The method by which the original amount of water (O) is introduced over the flashing, and manner in which the water that flows back on the wall surface (Dw) is collected and measured, is explained in subsequent sections. Figure 4 – Water being sprayed at the test specimen (flashing flatness in the test assembly was ensured during installation by fastening). Equation 1 Figure 3 – Tested flashing profiles. 26 • I N T E R FA C E S E P T E M B E R 2006 Flow rate The rate of water flow will affect the water-shedding performance of flashing. A higher rate of flow will direct more water away from the flashing than a lower rate of flow. This should be a control variable in the test. Test results reported in this paper were done using only one rate of flow of water. However, the test method can be easily adapted to include a flow rate control mechanism. Wind pressure Wind will have a considerable impact on the water deposition on the wall and the subsequent flow over the cladding. This paper focuses on the study of water flow over different flashing profiles and evaluating the profiles based on their watershedding effectiveness. It does not take into account the impact of wind on the water flow. In relative comparisons between different flashing profiles, wind will likely have the same effect on tested specimens and will change the test results equally. It is for this reason that ignoring the effect of wind is not expected to change the relative effectiveness of different flashing profiles. Surface flow considerations Test specimens are 500-mm-wide galvanized sheet steel profiles. This width is considered to be reasonable in order to achieve the desirable test results. Profiles are designed with a 300-mm-tall vertical flange at the top end. Water is sprayed at this surface first to create a uniform water flow at the top of the flashing. Two sides of the flashing specimens are treated with a strip of foam tape to prevent water run-off at the ends (Figures 3 and 4). Height of fall of water from drip edge Profiles to be tested are placed in the test assembly at a height of 1 meter from the water collection assembly. This is a common distance from a windowsill flashing to the ground or a parapet flashing to a window head. Field observations showed that water dripping from a flashing hits the wall surface at a distance of approximately 0.5 meters from the flashing; therefore, a Figure 5 – Test frame of assembly (approximate size of assembly = 300mm deep x 500mm wide x 1600mm high). with Next generation …the Lighter, Stronger “P”Curb www.chemlinkinc.com CHEMCURBTM No Mixing 10 Minute Installation Durable 10 Year Warranty Profitable Maintenance Free Environmentally Friendly Non-Shrinking, No V.O.C.’s CONTRACTOR HOT LINE 800-826-1681 Call for FREE CD! S E P T E M B E R 2006 I N T E R FA C E • 2 7 height of 1 meter will provide enough distance for water to either drip away or hit the wall surface. Other consideration The metal flashing test specimen is held in place by using two magnets. These magnets are powerful enough to keep the flashing in place while being tested and enable the researcher to change the position of the profile in order to create different gap sizes between the flashing and wall face. Design of Test Apparatus Test frame The test assembly is made of a 300mm x 500mm x 1600mm steel frame that contains the flashing profile to be tested, as well as water spray and water collection assemblies. A piece of plywood connected to the front side of the frame performs as the surface of the building. This is referred to as the “back-wall” in the test assembly. The plywood is protected with a waterproof layer that provides a non-absorbent surface over which water can flow without loss due to surface absorption (Figures 5 and 6). Water spray assembly The flashing profile is sprayed with water dispersed from the water spray apparatus. A 400 x 200mm bucket filled with 14 liters of water is placed on top of the test assembly at a height of 700mm from the test specimens. Water is sprayed to the vertical flange of the flashing through the nozzles placed on a copper pipe. The copper pipe is connected to the bucket of water. A valve is placed at the end of the loop to control the water flow. Once this valve is open, water flows until all the water in the bucket has drained. The results in this paper are for a flow that was timed to last for ten minutes. Figure 6 Figure 7 – Water spray assembly. Figure 8 – Water collection assembly. 28 • I N T E R FA C E S E P T E M B E R 2006 Sprayed water flows over the flashing and either drips away from the drip-edge, or drips back and runs down the back wall (Figure 7). Water collection assembly To collect the water runoff, a plastic tube is placed at the bottom end of the back wall. This is placed in such a way so as to minimize any extraneous water except for the amount to be collected. The tube is inclined in order to direct the runoff water to a container placed at the end of the tube. The amount of collected water is measured (Figure 8). The efficiency of a flashing profile is calculated by substituting this amount in Equation 1. Description of test specimen Flashing profiles are constructed of galvanized sheet steel. Tested flashing profiles are constructed based on the design parameters explained previously. Combinations of three sizes of angle (30˚, 45˚, and 60˚), three sizes of horizontal projection (10mm, 15mm, and 25mm), and two sizes of gap between the face of the flashing and the vertical surface of the wall (0 and 25mm), were used. This made 18 different flashing profiles that were tested. These sizes were chosen because they were seen more frequently in the reviewed design guidelines. Figure 9 – Effect of horizontal projection on efficiency for angles shown. S E P T E M B E R 2006 I N T E R FA C E • 2 9 TEST RESULTS A series of tests was conducted on the fabricated flashing profiles. The results reveal that a change in the design parameters of a flashing has a definite effect on its performance in water shedding. Using the test assembly described in this paper, 18 different flashing profiles were tested. Each test was repeated three times and the average was calculated for the results reported in this paper. The results from each repetition did not show any significant variation. The efficiency of each profile was calculated as described above. The following observations could be made from the results: As the horizontal projection of a dripedge increases, the size of the angle (and consequently the size of the drip-edge itself) should be larger for that profile to have a better performance. This relationship is shown in Figure 9. When testing profiles with 10mm horizontal projection, a profile with a 30˚ angle has the best efficiency. By increasing the size of the horizontal projection to 15mm, the profile with a 45˚ angle has the best efficiency. Finally, with a 25mm horizontal projection, the 60˚ angle profile performs the best. Increasing the size of the horizontal projection (drip-edge) of a profile will improve its performance for a given angle of drip edge. Also seen in Figure 9 (as the lines representing performance of each angle move higher in the graph, showing a better efficiency), the size of the horizontal projection increases toward the right side of the graph. Increasing the size of the gap between the flashing and the vertical face of the wall will increase the efficiency of the flashing. This relationship can be noted for the majority of tested profiles, except for those with a 60˚ drip-edge angle, in Figure 10. In this figure, each two joined bars represents a specific profile and the numbers below it explain the specification of that profile. For example, profile 60-10 is a flashing with 60˚ angle and 10mm horizontal projection. The bar with lighter color represents the efficiency of the profile with no gap between the flashing and the face of the wall. The bar with darker color represents its efficien- 30 • I N T E R FA C E S E P T E M B E R 2006 Figure 10 – Effect of increasing gap size on efficiency. cy with a 25mm gap. The results indicate that at larger drip-edge angles, the impact of gap is reduced. POSSIBLE FUTURE IMPROVEMENTS Further improvements can be applied to the proposed test assembly to more fully discern the impacts of the variables tested and to test the effect of additional variables. Keeping the same duration of water flow over the specimen was a challenge in these tests. This problem was due to an inconsistent water spray pattern. Small holes in the copper pipe get plugged with particles in the water, which results in increasing the duration of water flow over the specimen. Using a more accurate water spray apparatus could eliminate this problem. Due to unavailability of an accurate flow meter at this stage, tests were conducted using no particular flow, although the amount of water and duration of spray were kept consistent for each test. We attempted to use a flow meter at the early stages of the design; however, a very high flow was required for the apparatus to show the flow rate. Use of a sensitive flow meter to accurately show the rate of flow will be a significant improvement to this test method. Also, the effect of various rates of flow on watershedding effectiveness of metal flashing could be studied in the future. Metal flashing specimens used for testing were fabricated from unpainted galvanized sheet steel. This material was chosen because it was readily available at the time and could be easily bent to the required profiles. The unpainted nature of the surfaces seemed to impact the flow. Water did not seem to flow uniformly over the unpainted metal surface and it tended to form random streams of water that splashed away from the edge. Further testing can be conducted on painted metal flashing specimens. This might give a more representative result. CONCLUSION This paper has described the need to study the water-shedding performance of metal flashing and one test method to serve this purpose. Conducting a literature survey on current design guidelines revealed that there are no consistent design requirements for the drip-edge portion of a metal flashing. Also, studying some construction practices revealed that the drip portion of the metal flashing varies among projects. A test method was developed to study the water-shedding performance of metal flashing. The test assembly consisted of a frame, which allows the testing of various flashing profiles one at a time, as well as water spray and water collection assemblies. Different flashing profiles were constructed based on the variables that were identified to impact drip-edge performance. These flashing profiles were placed in the test assembly and were sprayed with water. The efficiency of each tested profile was calculated based on the collected water run-off and the results were compared. Test results revealed that small variations in the geometry of the drip can have a significant impact on the water-shedding characteristics of the flashing. Improving this test method by using a more accurate water spray apparatus and a sensitive flow meter has the potential for use as a standard for comparing the performance of flashings used in modern cladding systems. ACKNOWLEDGEMENTS The authors wish to acknowledge assistance from Ability Fabricators Inc. for help in fabricating the test specimens; Halcrow Yolles and Construction Control Inc. for providing resources and information; and staff from the workshop of the Department of Architectural Science at Ryerson University for assistance in constructing the test assembly. References Canadian Mortgage Housing Corporation (CMHC). Best Practice Guide: Brick Veneer Concrete Masonry Unit Backing. Ottawa: CMHC. 1997a. Canadian Mortgage Housing Corporation (CMHC). Best Practice Guide: Flashing. Ottawa: CMHC. 1997b. Canadian Roofing Contractors’ Association (CRCA). Roofing Specifications. Ottawa: CRCA. 1997. Construction Control Inc. Specifications for the Replacement of the Roof Membrane and Associated Work at 1615 Clark Boulevard, Brampton. Woodbridge: Construction Control Inc. 2004a. Construction Control Inc. Specifications for the Replacement of the Roof Membrane and Associated Work at 2780-2800 Skymark Ave., Mississauga. Woodbridge: Construction Control Inc. 2004b. Doshi, H. “Rainwater on Building Test your knowledge of roofing with the following questions, developed by Donald E. Bush Sr., RRC, FRCI, PE, chairman of the RRC Examination Development Subcommittee. 1. Which are the five basic types of construction identified under the International Building Code that define the degree of fire resistance for buildings? 2. Where would a designer look to determine the IBC-required fire rating for the roof assembly? 3. FM Global’s classification of resistance to interior fire divides roof assemblies into two basic categories: sprinklered and unsprinklered. Which roof assemblies do not require sprinklers? 4. Which four tests for external fire resistance does UL perform to determine roof covering fire classification? 5. Flame spread depends chiefly on which three parameters? Answers on page 32 S E P T E M B E R 2006 I N T E R FA C E • 3 1 Cladding: Two Case Studies.” 10th Canadian Conference on Building Science and Technology, May 12-13, 2005. Ottawa: IRC/NRC. 2005. Institute for Research in Construction (IRC). National Building Code of Canada. Toronto: IRC. 1995. Maurenbrecher, A. H. P. “Water-shedding Details Improve Masonry Performance.” Construction Technology Update No. 23, National Research Council of Canada. Ottawa: NRC. 1998. National Roofing Contractors Association (NRCA). Roofing & Waterproofing Manual, 5th Ed. Rosemont: NRCA. 2001. Quadrangle Architects Limited. City- Gate2 Architectural Drawings and Specifications. Toronto: Quadrangle Architects Limited. 2004. Sheet Metal and Air Conditioning Contractors National Association (SMACNA). Architectural Sheet Metal Manual, 5th ed. Chantily: SMACNA. 1993. The Aluminum Association. Specifications for Aluminum Sheet Metal Work in Building Construction, 4th Ed. Washington, D.C: The Aluminum Association. 2000. Young and Wright Architects Inc. The Uptown Architectural Drawings and Specifications. Toronto: Young and Wright Architects Inc. 2004. Answers to questions on page 31: 1. Type I — Non-combustible, protected. Type II — Non-combustible, protected and unprotected. Type III — Non-combustible exterior walls, interior building elements of any permitted material. Type IV — Non-combustible exterior walls, interior building elements of heavy timber. Type V — Combustible; i.e., any materials permitted by code. 2. Each of the five types of construction identified by the IBC has subclasses and each subclass carries its own required fire rating for the roof assembly. 3. a. Class 1 steel deck assemblies. b. Non-combustible decks – concrete, gypsum, asbestos cement, and preformed structural mineralized wood fiber. c. Wood decks treated with fireretardant in organic salts limiting flame spread to 25 feet or less. 4. a. Flame spread. b. Flame exposure. c. Burning brand. d. Flying brand. 5. a. Roof slope. b. Nature and quantity of surfacing. c. Membrane composition. Reference: Manual of Low-Slope Roof Systems – Fourth Edition 32 • I N T E R FA C E S E P T E M B E R 2006 Saba Saneinejad has a bachelor’s degree in architectural science from Ryerson University. She is currently employed as a building science specialist at Halcrow Yolles in Toronto. This is her undergraduate thesis paper, conducted under the supervision of Professor Hitesh Doshi. Saba Saneinejad Hitesh Doshi is a professional engineer and a faculty member in the Department of Architectural Science at Ryerson University in Toronto. He is involved in practice, research, and teaching related to a variety of building science issues. The information presented in this paper is a result of ongoing work conducted by students at Ryerson under Doshi’s supervision. Hitesh Doshi The Building Enclosure Technology and Environment Council (BETEC) of The National Institute of Building Sciences (NIBS) announces its new relationship with Matrix Group Publishing in the production of the Journal of Building Enclosure Design. As the need for better-performing building wall, below-grade, roof and fenestration systems and related indoor environments grows, the journal will provide a forum for new ideas and initiatives. Focus will be on the latest developments in building energy conservation, security, and improved building envelope design and performance. The journal will be produced biannually, with the first edition expected in August 2006. For more information, visit www.nibs.org. Journal of Building Enclosure Design
