A M E R I C A N
A R C H I T E C T U R A L
AAMA CW-DG-1-96 Editorial Revision: 5/2005
Curtain Wall Design Guide Manual
M A N U F A C T U R E R S
A S S O C I A T I O N
TABLE OF CONTENTS Foreword ..........................................................1 Aluminum Curtain Wall Types And Systems ............1 The Function Of A Wall .......................................1 Basic Terms Defined ...........................................2 Wall Types ........................................................3 Classification Of Wall Systems..............................3 Summary ..........................................................9 Primary Concerns In Aluminum Curtain Wall Design ....................................10 Function Of The Wall ........................................10 Natural Forces And Their Effects .........................11 Basic Design Considerations...............................12 Other Design Considerations ..............................15 Summary.........................................................16 Guidelines For The Architect In Detailing The Wall ..................................17 Advice During Early Design Stages ......................17 The Architect's Details ......................................18 Manufacturers' Suggestions Often Helpful.............22 Importance Of Tolerances And Clearances ............26 Building Frame Tolerances .................................27
Installation Clearances And Tolerances ................27 Table A – Standard Tolerances For Steel Building Frames ............................28 Table B – Standard Tolerances For Poured Concrete Building Frames.............29 Table C – Standard Tolerances For Precast And Prestressed Concrete ............30 Testing Of Aluminum Curtain Walls .....................36 Reasons For, And Value Of Laboratory Testing.......36 Performance Characteristics Subject To Pre-testing............................................37 The Test Specimen............................................38 Order Of Testing...............................................39 Test For Air Leakage .........................................40 Tests For Water Penetration ...............................40 Test For Structural Performance ..........................42 Thermal Tests ..................................................43 Sound Transmission Test....................................45 Evaluation Of Test Results .................................45 Field Checks During Installation .........................46 Summary Recommendations ...............................47
AAMA. The Source of Performance Standards, Product Certification and Educational Programs for the Fenestration Industry. This voluntary specification was developed by representative members of AAMA as advisory information and published as a public service. AAMA disclaims all liability for the use, application or adaptation of materials published herein. 2005 © American Architectural Manufacturers Association 1827 Walden Office Square, Suite 550, Schaumburg, IL 60173 PHONE 847/303-5664 FAX 847/303-5774 EMAIL
[email protected] WEBSITE www.aamanet.org All AAMA documents may be ordered at our web site in the “Publications Store”. Original Publication: CW-DG-1-79 Preceding Document: CW-DG-1-96 (Released: 1996) Editorially Revised: 5/2005 Published: 5/2005
FOREWORD AAMA in 1970 announced its intent to provide Architects with up-to-date information and technical data on aluminum curtain wall construction with the organization of its curtain wall division. To achieve this objective, the, "Aluminum Curtain Wall Series" publications were initiated. The reception by architects, as well as others concerned with curtain wall construction, has been excellent, In view of this reception, AAMA has undertaken the continuation and revision of the series as an on-going project. This ALUMINUM CURTAIN WALL DESIGN GUIDE MANUAL is comprised of the latest revisions to the earliest volumes. AAMA plans to continue these publications with the release of individual volumes as they are completed. In addition to this manual, volumes are available on Joint Sealants, Glass and Glazing, Fire Safety, Installation of Curtain Walls, Anodic and Painted Finishes, Structural Silicone Glazing, Properties of Glass, Care and Handling of Architectural Aluminum and Windloads. Not a part of
the Curtain Wall Series but addressing associated topics are the Metal Curtain Wall Manual containing guide specifications for storefront and curtain wall and the Aluminum Storefront and Entrance Manual which addresses entrance requirements. This manual was developed by representative members of the American Architectural Manufacturers Association. It is published primarily as a service to architects and secondarily as a service to manufacturers and installers of these products. AAMA disclaims any responsibility of liability of any kind in connection with material contained in the manual and makes no warranties expressed or implied, of any kind whatsoever regarding the information contained herein. Furthermore, none of the contents of the manual shall be construed as a recommendation of any patented or proprietary application that may be included in such contents.
ALUMINUM CURTAIN WALL TYPES AND SYSTEMS In general, the metal curtain walls of today, even the simpler types, are far more sophisticated products than their early counterparts, though many of the earliest walls are still performing admirably. Fifty years of experience and development have eliminated the major difficulties of the pioneering designs, resulting in better products. Beginning with the relatively simple, but innovative concept of the early 1950's - a series of window units and panels joined and supported by simple framing members metal curtain wall technology has developed, over the years, into a proliferation of highly engineered designs. Throughout this development, however, the basic principles of good curtain wall design have not changed. Recognition of these principles has grown with experience, and the criteria of good design have now become well defined. And, as with any vital and developing product, the industry continues to find ways of improving performance. Although the aluminum curtain walls of today appear to present endless variety of design, the majority of these designs can be identified as representing one of several elementary types. Because it is important that architects recognize these principal types, their characteristics and their relative merits, it is the purpose of this article to clarify these distinctions and the extent to which they may be applicable. First, however, one should have a clear understanding of the primary function of any such wall, regardless of what type it may be.
AAMA CW-DG-1-96
THE FUNCTION OF A WALL Prior to the advent of metal curtain wall, it seems that little attention had been paid to analyzing and identifying the rather complex function of an exterior building wall. Generally it was thought of as performing either one or both of two functions: 1) providing structural support for floors and roof, if a bearing wall, and 2) forming a protective enclosure excluding the elements, but with openings for vision and ventilation as required. One of the early studies of metal curtain wall potentials,* however, pointed out the fact, now generally recognized but still sometimes overlooked, that the exterior wall of a building actually serves, in effect, as a two-way filter, controlling the through flow, both inward and outward, not only of heat, light and air, but also of a number of other penetrants such as moisture, dirt, sound, vermin, animals and, of course, people. A properly designed aluminum curtain wall has the capability of being able to provide any degree of such control as is desired. * “Curtain Walls of Stainless Steel,” a report by the School of Architecture, Princeton University, 1955
Page 1
BASIC TERMS DEFINED As a further prerequisite to discussing wall types, and to facilitate communication and avoid misunderstandings, certain key terms should be defined. Three of the most commonly used terms, "curtain wall," "metal curtain wall" and "window wall," still mean different things to different people. Often they are used interchangeably, with no clear distinction being made between them. As their meanings are interrelated and overlapping, this will likely continue to be the case, but for the purposes of this article these terms are defined as follows:
1 2
Curtain Wall: Any building wall, of any material, which carries no superimposed vertical loads, i.e., any "nonbearing" wall. Metal Curtain Wall: An exterior curtain wall which may consist entirely or principally of metal, or may be a combination of metal, glass and other surfacing materials supported by or within a metal framework. Window Wall: A type of metal curtain wall installed between floors or between floor and roof and typically composed of vertical and horizontal framing members, containing operable sash or ventilators, fixed lights or opaque panels or any combination thereof. It follows, then, that a metal curtain wall is a particular type of curtain wall, and a window wall, in turn, is a certain form of metal curtain wall. In today's common usage, metal curtain wall is often referred to, and recognized, as simply "curtain wall," a simplification of terms that is widely accepted but not precisely accurate.
AAMA CW-DG-1-96
1: Framing Unit. 2: Building Spandrel. Other variations: Framing units may be individual units or stick assemblies. Glazing infills may be installed in the framing before frame installation or glazed after frame erection. FIGURE 1: WINDOW WALL – SCHEMATIC OF TYPICAL VERSION
Page 2
WALL TYPES Because of the wide and increasing variety of aluminum curtain wall designs, it is difficult, if not impossible, to precisely identify every design as representing one or another of a few basic types. Certain broad distinctions can usually be made, but in some cases, accurate classification under one of a limited number of subcategories becomes subjective and therefore debatable. Nevertheless, because there are so many variations, some generally accepted system of identifying the most common design forms becomes essential. One method of identification which is useful and allinclusive is based on the extent to which the wall design is unique. Any aluminum wall is either 1) a custom type, 2) a standard type, or, in some cases 3) a combination of the two. The two basic types are described as follows: Custom walls are those which are designed specifically for one project (either a single building or a group of related buildings). Such walls usually, though not necessarily, have substantial areas of glass, and may be used on buildings of any type or size. Typically they are chosen for the more glamorous high-rise structures and for commercial, institutional and monumental buildings of high quality. Standard walls are those which employ components and details which are designed and standardized by their manufacturer. They may be assembled in stock units, but more often their arrangement is dictated by the architect's design. Standard walls may be either of two types: Architectural: Walls having formed framing members (usually extrusions) and sizeable areas of glass, often with opaque panel areas also. Industrial: Walls composed either of preformed metal sheets made in stock patterns and sizes, used in combination with standard windows, or of large metalfaced insulated panels, used either with or without fenestration. Typical usage of such walls is on industrial type structures.
Standard industrial walls are generally the most economical type. No definite statement can be made, however, regarding the relative costs of standard architectural walls and custom walls. As a rule, the standard type, because of the economies inherent in quantity production, tend to be less expensive than custom walls, but this is not always the case. Some of the more complex standard designs may cost as much or more than the simplest custom designs, and on large projects where quantities are sufficient to warrant large scale production methods, custom work may impose little, if any cost premium. Also, when standard designs are modified to incorporate special features or details - when they become "custom standard" - the resulting cost may equal or exceed that of a custom wall designed specifically to provide such features. The character of the job, and its aesthetic and functional requirements, usually determine which type of wall is the best choice. The architect, before choosing between custom and standard, should contact the potential suppliers and fully investigate the complete overall costs of each option. CLASSIFICATION OF WALL SYSTEMS Both custom and standard walls may be further classified according to their "system" or method of installation. The majority of aluminum curtain walls built to date may be identified as representing one of five different systems, but some of the newer designs do not fall neatly into any of these categories. With more new design expressions constantly appearing, it's quite likely that additional "systems" may become common. The five generally recognized systems referred to are these: 1) the stick system, 2) the unit system, 3) the unitand-mullion system, 4) the panel system, and 5) the column-cover-and-spandrel system. Each of these systems is illustrated schematically and is briefly described in the following:
The term "standard" should not be interpreted as in any way implying rigidly fixed or static design concepts. Standard designs vary from one manufacturer to another, and are adaptable to a broad range of aesthetic expression. It has been suggested, in fact, that three subcategories of standard architectural walls should be recognized: "stock standard," "modified standard" and "custom standard," to indicate the degree of design flexibility offered in the socalled "standard" types, and the fact that distinction between standard and custom walls may be nebulous.
AAMA CW-DG-1-96
Page 3
STICK SYSTEM In the stick system the wall is installed piece by piece. Usually the mullion members are installed first, followed in turn by the horizontal rail members, the panels (if any), and finally the glazing or window units. However, in designs accenting the horizontal lines the process may be altered to first install the larger horizontal members. In either case, the horizontal and vertical framing members are often long sections designed to either be interrupted or extend through at their intersections.
1 5 2
The stick system was used extensively in the early years of metal curtain wall development, and is still in wide use in greatly improved versions. Some manufacturers consider it to be superior to other systems. The advantages of this system are its relatively low shipping and handling costs, because of minimal bulk, and the fact that it offers some degree of dimensional adjustment to site conditions. Among its disadvantages are the necessity of assembly in the field, rather than under controlled factory conditions, and the fact that preglazing is obviously impossible. Current designs offer considerable variation in the location of the glass with respect to the mullion or rail depth. It may be near the indoor face, near the outdoor face, providing a virtually flush exterior surface, or anywhere between the two.
3 1 4
6 7 1: Anchors. 2: Mullion. 3: Horizontal rail (gutter section at window head). 4: Spandrel panel (may be installed from inside building). 5: Horizontal rail (window sill section). 6: Vision glass (installed from inside building). 7: Interior mullion trim. Other variations: Mullion and rail sections may be longer or shorter than shown. Vision glass may be set directly in recesses in framing members, may be set with applied stops, may be set in sub-frame, or may include operable sash. FIGURE 2: STICK SYSTEM – SCHEMATIC OF TYPICAL VERSION
AAMA CW-DG-1-96
Page 4
UNIT SYSTEM In the unit system the wall is composed entirely of large framed units pre-assembled at the factory, complete with spandrel panels (if any) and sometimes also pre-glazed. The vertical edges of the units join to form mullion members, top and bottom members join to form horizontal rails, and the units may be one, two or sometimes three stories in height. This system offers obvious advantages, but has certain disadvantages also. It provides assembly under controlled shop conditions, where the work can be carefully inspected, and facilitates rapid enclosure of the building with a minimum of field labor and relatively few field joints. On the other hand, the units are bulky and require more space for shop assembly, shipping and on-site storage. They also necessitate more elaborate protective measures, both in transit and in storage at the site prior to installation. Another problem sometimes encountered is the detailing and installation of "leave-out" units. Typical wall units are designed for sequential interlocking installation, and when openings have to be left in the wall to facilitate the handling of construction materials the units installed later to close these openings usually require special joint details and installation procedures.
1
2
1: Anchors. 2: Pre-assembled framed unit. Other variations: Mullion sections may be interlocking “split” type or may be channel shapes with applied inside and outside joint covers. Units may be unglazed when installed or may be pre-glazed. Spandrel panel may be either at top or bottom of unit. FIGURE 3: UNIT SYSTEM – SCHEMATIC OF TYPICAL VERSION
AAMA CW-DG-1-96
Page 5
UNIT AND MULLION SYSTEM The unit-and-mullion system, as the term implies, is a compromise between the two previously described systems. In this system the mullion members are separately installed first, then pre-assembled framed units are placed between them. These units may be full story height, or they may be divided into a spandrel unit and a vision glass unit. The system is often employed when the mullion sections are unusually deep or large in cross section, making it impractical to incorporate them as part of a pre-assembled unit. The advantages and disadvantages of this system are generally comparable to those of the unit system. The shipping bulk of the units themselves is somewhat less, because of the omission of the mullion section, but the amount of field labor, field jointing and erection time is likely to be somewhat greater.
3
1 4 2
1: Anchors. 2: Mullion (either one- or two-story lengths). 3: Pre-assembled unit-lowered into place behind mullion from floor above. 4: Interior mullion trim. Other variations: Framed units may be full-story height (as shown), either unglazed or pre-glazed, or may be separate spandrel cover units and vision glass units. Horizontal rail sections are sometimes used between units. FIGURE 4: UNIT-AND-MULLION SYSTEM – SCHEMATIC OF TYPICAL VERSION
AAMA CW-DG-1-96
Page 6
PANEL SYSTEM The panel system is similar in concept to the unit system, the chief difference being that the panels are not preassembled framed units but homogeneous units formed from sheet metal or as castings, with few if any internal joints except at the glass periphery. The panels may be full story height, with or without openings for glazing, or they may be smaller units. Unlike the other three systems, the panel system usually provides an overall pattern for the wall, rather than a grid pattern or a design having strong vertical or horizontal accents.
1
Two kinds of panel systems should be recognized: those of an "architectural" character and those used in industrial walls. The architectural type (the panel system) is always custom made, and consequently is relatively expensive, whereas the industrial type panels are produced in large quantity, as standard products, by roll- or brake-forming, and are relatively inexpensive.
2
The architectural panel system offers all of the advantages of the unit system, with probably a wider range of design flexibility, and has the added advantage of a minimal amount of shop labor. However, with present techniques, the cost of dies, if panels are formed by stamping, or of molds, if formed by casting, makes this system economically attractive only when a large number of identical panels are to be used.
1
1: Anchors. 2: Panel. Other variations: Panels may be formed sheet or castings, may be full story height (as shown) or smaller units, and may be either pre-glazed or glazed after installation. FIGURE 5: PANEL SYSTEM – SCHEMATIC OF TYPICAL VERSION
AAMA CW-DG-1-96
Page 7
COLUMN-COVER-AND-SPANDREL SYSTEM The column-cover-and-spandrel system is a relatively recent development, compared with the other four systems, all of which have been in use for at least twentyfive years. It may, in fact, be questioned whether the use of this design concept has been sufficient to warrant its classification as a "system," but its popularity appears to be increasing. As its name implies, the elements of this system consist of column cover sections, long spandrel units which span between column covers, and infill glazing units which may be either pre-assemblies or separate framing members and glass.
1
2
This type of system, unlike the other four described, provides a facade design which clearly expresses the structural frame of the building, rather than a superimposed grid or overall pattern. It permits a wide latitude of aesthetic expression in all three of its basic components, and because column spacings and spandrel depths vary with almost every building, its use is necessarily limited to custom type walls.
3
1: Column cover section. 2: Spandrel panel. 3: Glazing infill. Other variations: Column covers may be one piece or an assembly, may be of any cross-sectional profile, and either one or two stories in height. Spandrel panel may be plain, textured or patterned. Glazing infill may be a preassembly, either glazed or unglazed, or be assembled in place. FIGURE 6: COLUMN COVER AND SPANDREL SYSTEM – SCHEMATIC OF TYPICAL VERSION
AAMA CW-DG-1-96
Page 8
These are the five systems, based on method of installation, which have been most commonly used to date. It should be obvious that aluminum curtain wall design, contrary to some views, is by no means limited to grid patterns, or to patterns accenting either vertical or horizontal lines. More and more, other forms of aesthetic expression are appearing; virtually flush walls, walls with little or no exposed framing, walls in which exposed metal serves as a permanent form for concrete framing or fireproofing, and other fresh new concepts. Perhaps in the future, some of these innovations will become common systems deserving identification, but at present no attempt is being made to "tag" them. They are referred to simply as "other systems," and as the design potentials of aluminum curtain wall are further explored there will certainly be more systems in this catch-all category.
AAMA CW-DG-1-96
SUMMARY While design variations in aluminum curtain wall are innumerable, and it is impractical to attempt to categorize all of them, the great majority of designs to date can be identified as to type and system. The generally accepted method of classifying them has been described, and the use of this method is recommended, insofar as it is applicable, to facilitate communication. Accordingly, it is common practice to identify aluminum curtain walls as being either a custom or standard type and representing one of the five common systems of installation. For example, we may refer to a custom wall of the unit-and-mullion system, a standard wall of the stick system, a custom column-cover-and-spandrel system, a standard industrial wall, and so forth. Custom walls may employ any of the five (or other) systems. Standard architectural walls may be either stick system, unit system or unit-and-mullion system; very few, if any, have been panel or column-cover-and-spandrel systems. Standard industrial walls, on the other hand, are always some type of panel system.
Page 9
PRIMARY CONCERNS IN ALUMINUM CURTAIN WALL DESIGN It has often been said before, but it bears repeating: metal curtain wall technology is different. The design of aluminum-and-glass walls requires careful attention to matters which normally receive little consideration when designing with the more traditional wall construction materials. It isn't because the laws of nature are any different for metal curtain walls; they aren't. But the materials used in its construction react quite differently to some of these laws than do other wall materials. It seems there are still many architects who either don’t fully understand the principles involved in good curtain wall design or fail to give these matters the attention they deserve. One reason for this seems clear. In school, architects are taught how to build with wood, brick, stone and concrete, but they are told little if anything about the architectural metals, glass and sealants. Small wonder, then, that they sometimes make mistakes - errors in judgment which, with a better understanding of these newer materials, could probably have been avoided. Although aluminum curtain wall design has been steadily improving, it appears that there is still an information gap. If so, it is probably due, in part at least, to inadequate educational efforts on the part of the metal curtain wall industry itself. To help bridge this gap, and provide guidance for the architect, this article highlights briefly the basic principles and essential requirements of good curtain wall design. In later issues of the Aluminum Curtain Wall series these matters will be explored in greater depth.
FUNCTION OF THE WALL Until the invention of the structural steel building frame, over a hundred years ago, most exterior building walls were substantially solid elements, pierced by windows to provide light and ventilation. They served chiefly two purposes: to form an enclosing barrier and to provide support for upper floors and roof. But with the advent of skeleton frame construction, and with the imaginative daring of such leaders as Sullivan and Gropius, glazed areas became much larger, sometimes dominating the facade, with only enough masonry to cover the bones of the structure. Today we often see glass used as the predominant wall material, with no masonry at all. “Whether the all-glazed air-conditioned building with fixed glass has no windows or no walls becomes merely a nice point in theory and terminology, but it does emphasize the need to examine a building enclosure as a single complex system, rather than an assembly of separate elements with clearly defined functions."1 1
Rostron, R. Michael, "Light Cladding of Buildings," The Architectural Press, London, 1964 The aluminum curtain wall, regardless of what proportion may be glazed, should be thought of as an enclosure system. And the functions of this system are far more complex than the twofold purpose of the early masonry bearing wall. It serves, in effect, as a filter, selectively impeding or controlling the flow inward, outward, or in both directions, not only of people and property, but of all that affects the internal environment of the building. This concept of the exterior wall as a filter is not new; it was suggested as long as 50 years ago, and has often been expressed since, but still deserves emphasis. The aluminum curtain wall, relieved of the inhibitions of mass necessary to support vertically imposed loads, has become in truth a filtering envelope for the building. Properly designed, it serves the multiple functions of 1) withstanding the action of the elements, 2) controlling the passage inwards and outwards of heat, light, air and sound, and 3) preventing not only access by intruders but the entrance also of deteriorating influences affecting its own integrity. It is this concept of the wall as a selective filter system which should dictate its physical design.
AAMA CW-DG-1-96
Page 10
OUTDOORS
NATURAL FORCES AND THEIR EFFECTS Obviously all exterior walls, of whatever material, are subject to, and must withstand the ravaging effects of nature. Chief among these natural forces are sunlight, temperature, water, wind and gravity. Except for gravity, the intensity and relative significance of these forces vary somewhat from one region to another, but all of them must be considered, and their effects provided for, in all locations. They may act upon the wall either individually or more often in concert, but to understand their impact on design requirements the effects of each should be separately examined.
INDOORS
RAIN DIRT FIRE NOISE
VAPOR
VANDALS THIEVES INSECTS
LIGHT AIR HEAT COLD
VISION AIR HEAT COLD
AUTHORIZED
PERSONS
Sunlight is great; we couldn’t live without it. It provides warmth, color, visual definition and life itself. But it also creates certain problems in curtain wall design. One of these problems is its deteriorating effect on organic materials such as color pigments, plastics and sealants. The actinic rays, particularly those found in the ultraviolet range of the spectrum, produce chemical changes which cause fading or more serious degradation of materials. It’s essential, therefore, that materials and finishes vulnerable to such action be thoroughly investigated before being used, and that sealants be tested for resistance to ozone attack and ultra-violet radiation. Another problem resulting when uncontrolled sunlight passes through the wall is the discomfort of glare and brightness and degradation of interior furnishings. Conventionally, such effects are combated by use of some type of shading device, either inside or outside of the vision glass. A newer approach, gaining in favor, is the use of glare-reducing or reflective types of glass which provide relief without restricting vision. Temperature also creates two kinds of problems in curtain wall design: 1) the expansion and contraction of materials, and 2) the necessity to control the passage of heat through the wall. It is the effect of solar heat on the wall which creates one of the major concerns in aluminum curtain wall designthermal movement. Although minimum outdoor temperatures vary about 56°C (100°F) throughout the United States, maximum surface temperatures of metal on buildings, in most locations, are in the neighborhood of 77°C (170°F). Of course it’s the temperature fluctuations, both diurnally and seasonally, that critically affect wall details. All building materials expand and contract to some extent with temperature changes, but the amount of movement is greater in aluminum than in most other building materials.
FIGURE 7: THE WALL AS A FILTER
AAMA CW-DG-1-96
Page 11
The control of heat passage through the wall affects both heat loss in cold weather and heat gain in hot weather, the relative importance of the two varying with geographic location. Thermal insulation of opaque wall areas becomes an important consideration when such areas constitute a substantial part of the total wall area, but when vision glass areas predominate, the use of insulating glass, and the minimizing of through metal or “cold bridges” are more effective in lowering the overall Uvalue of the wall. Water, in the form of rain, snow, vapor or condensate, is probably the most persistent cause of potential trouble. As wind-driven rain, it can enter very small openings and may move within the wall and appear on the indoor face far from its of point of entrance. In the form of vapor it can penetrate microscopic pores, will condense upon cooling and, if trapped within the wall, can cause serious damage that may long remain undetected. Leakage may be a problem in a wall built of any material. Most masonry walls, being porous, absorb a good deal of water over their entire wetted surface, and under certain conditions some of this water may penetrate the wall, appearing as leaks on the indoor side. But the materials used in metal curtain wall are impervious to water, and potential leakage is limited to joints and openings. Though this greatly limits the area of vulnerability, it greatly increases the importance of properly designing the joints and seals. Wind acting upon the wall produces the forces which largely dictate its structural design. On the taller structures in particular, the structural properties of framing members and panels, as well as the thickness of glass, are determined by maximum wind loads. Winds also contribute to the movement of the wall, affecting joint seals and wall anchorage. The pressures and vacuums alternately created by high winds not only subject framing members and glass to stress reversal, but cause rain to defy gravity, flowing in all directions over the wall face. Thus wind must be recognized also as a major factor contributing to potential water leakage. Gravity, unlike the other natural forces, is static and constant, rather than dynamic and variable. Because of the relatively light weight of materials used in curtain walls, it is a force of secondary significance, rarely imposing any serous design problems. It causes deflections in horizontal load-carrying members, especially under the weight of large sheets of heavy glass, but because the weight of the wall is transferred at frequent intervals to the building frame, gravity forces affecting structural design are generally small in comparison with those imposed by wind action. But far greater gravity forces, in the form of floor and roof loads, are acting on the building frame to which the wall is attached. As these loads may cause deflections and
AAMA CW-DG-1-96
displacements in the frame, the connections of the wall to this frame must be designed to provide for sufficient relative movement to insure that displacements do not impose vertical loads on the wall itself. BASIC DESIGN CONSIDERATIONS An analysis of the effects of these natural forces reveals the major problem areas to be anticipated. Experience verifies that in the design of aluminum curtain wall there are generally three matters of chief concern: 1) structural integrity, 2) provision for movement, and 3) weathertightness. Of course there are a number of other considerations, most of which are of less critical importance and some of which vary in importance with the location and type of building. But let us review briefly these major considerations applying in all cases. a) Structural Integrity Because structural failure may jeopardize human life, the structural integrity of the wall may be said to be the primary concern in its design. But the structural design of curtain wall involves the same procedures as used in any other wall, and deficiencies in this respect are less likely to occur than deficiencies in providing for movement and weathertightness-requirements which, by comparison, present unique problems in metal construction. Structurally, the requirements of stiffness rather than strength usually govern, and though excessive deformations may, in some cases, lead to damage, such rare instances of actual failure as have occurred have, for the most part, been limited to faulty anchorage details. As vertical loads in the wall system are relatively light, structural design is chiefly a matter of providing proper resistance to lateral wind forces. This is a routine procedure, provided that the nature and magnitude of the wind loads are known. But herein may lie the problem. We have a fairly good knowledge of wind loads on low and medium height buildings, but still have much to learn about the nature and intensity of such loads on tall structures. This a subject which is more fully addressed in the AAMA Curtain Wall 11 on, “Design Windloads for Buildings and Boundary Layer Wind Tunnel Testing.” It is well known that maximum wind velocities, and consequently design wind loads, vary not only with geographic location but also with height above the ground. It is less generally recognized that the nature of the building’s surroundings-whether open country, suburban character or dense urban building-are even more important influences on wind action. Another fact not generally recognized, even in some of the major building codes, is that the wind loads acting on the skin of the building are of a different character and magnitude than those which govern the design of the building frame. As compared with the overall design loads, those acting on the wall are more severe in intensity, have a specific rather than cumulative effect, and change more drastically and more rapidly.
Page 12
Often too little significance is attached to the negative wind loading, or suction forces, acting on the wall, and the fact that internal building pressures due to air conditioning may augment such forces. Many designers tend to think only in terms of wind pressure, whereas in fact, even with moderate winds, more of the total perimeter wall surface of a rectangular building is likely to be subjected to a vacuum than to pressure. On highrise buildings these negative pressures are usually maximum near the building corners, where they may be more than twice as great an any positive load on the wall. When wind damage does occur, it is more often in the form of a blow-out than a blow-in. This explains why the most common deficiency in structural design is the failure to provide adequate resistance, particularly in anchorage details, to the suction action of the wind. SAME WALL UNDER DIFFERENT WIND DIRECTIONS
-24
-20 -23
25
-22
24
-24
-24 -25
22 -26 20 18 -22
16
14
-20 -18 -16 -14
12 12
GROUND LEVEL
MAXIMUM POSITIVE PRESSURES
-12
MAXIMUM NEGATIVE PRESSURES
(UNITS ON ISOBARS ARE POUNDS PER SQUARE FOOT)
Distribution of design wind loads on walls of 64-story building, triangular in plan, as determined by model tests in boundary layer wind tunnel. FIGURE 8
AAMA CW-DG-1-96
b) Provision for Movement A most important consideration, in designing any aluminum curtain wall, is ample provision for movement. Failure to do this lies at the base of many instances of malperformance. No building is a static thing, and this goes double for metal curtain wall. Movement is constantly taking place-movement within the wall components themselves, relative movement between the components, and relative movement between the wall and building frame to which it’s attached. These movements are caused not only by temperature changes, but by wind action, by gravity forces and by deformations or displacements in the building frame. To disregard such movements in designing the wall is an urgent invitation to trouble. The effect of temperature changes is of course uniquely significant, because of the relatively high coefficient of expansion of aluminum, but the amount of such movement is predictable. In most parts of the country the probable seasonal range of metal surface temperature is at least 66°C (150°F), and in some parts it may be as much as 94°C (200°F). This translates into a movement of from 6 mm to 8 mm (1/4 in to 5/16 in) in a 3 m (10 ft) length of aluminum. In a sheet of glass used alongside the aluminum the amount of movement will be less than half as much. Movement due to the other causes mentioned are generally not accurately predictable, but may be equally significant. Whatever the cause, however, the problem of providing for movement reduces to the problem of joint design, because it’s at the joints that movement must be accommodated. It becomes axiomatic, therefore, that the secret of a functionally successful curtain wall lies in the design of its joints. Consequently, the detailing of the joints is the most critical, and often the most difficult aspect of any curtain wall design. It doesn’t necessarily follow, however, that by using larger wall units and thus fewer joints the problem will be simplified. This is seldom the case. The larger the units, or the longer the members, the greater will be the amount of movement to be accommodated at each joint, and this tends to complicate, rather than simplify the joint design. Provision must be made, of course, for both vertical and horizontal movement in the plane of the wall, either by some kind of slip joints or bellows action. Detailed information as to various means of accomplishing this are presented in AAMA’s entitled “Installation of Aluminum Curtain Walls.” It should be recognized, though, that expansion and contraction are not necessarily translated entirely into displacement. In some situations they can be absorbed, to some degree at least, by increased stress within the member, resulting in a calculated deformation, and often they are accommodated by a combination of stress and deformation. Except in a few cases, however, reliance upon stress increase alone to accommodate expansion is not advisable, as excessive bending or buckling may result.
Page 13
MOVEMENT ACCOMODATED BY SLIP JOINTS
MOVEMENT ACCOMODATED BY SLIP JOINTS
ANCHOR BRACKET
SPLIT MULLION (MATING TYPE)
BELLOWS MULLION
SPLIT MULLION (BATTEN TYPE)
In wall designs featuring mullion members it is common practice to design these members in such a way as to accommodate lateral movement in the plane of the wall. These drawings illustrate three ways this may be done. FIGURE 9 c) Weathertightness Weathertightness means protection against both water leakage and excessive air infiltration. It depends in large measure on adequate provision for movement, and is closely related to proper joint design. Undoubtedly, a major share of difficulties experienced with metal curtain wall over the years has been due to the lack of weathertightness. Water leakage was an all-too-common occurrence in the earlier walls, due to faulty design, materials or workmanship, or a combination of these. But with improved materials and design techniques its prevention has now become the rule rather than the exception. By comparison, excessive air leakage is less critical and more easily prevented. High winds cause rain water to flow in all directions over the windward surface of a wall, and on surfaces of impervious materials much of it tends to collect at the joints-the major points of vulnerability. Early in the history of metal curtain wall experience it became apparent that to provide all joints at their outer surface with a permanently waterproof seal was essentially impossible, because of their continual movement, and this approach to weathertightness was soon abandoned. Instead, two other methods have been developed for preventing leakage through the wall, and either of these, when intelligently applied, is highly dependable. One is referred to as the “internal drainage” or “secondary defense” system, and has long been used by competent designers. The other is the "pressure equalization" method, a more recent development in metal curtain wall technology. Both of these methods are applicable to the design of windows as well as complete wall systems.
AAMA CW-DG-1-96
The internal drainage method is based on the philosophy that it is impractical if not virtually impossible to totally eliminate, for any length of time, all leakage at all points in the outer skin of the wall, but that such minor leakage can be prevented from penetrating to the indoor face of the wall or even remaining within the wall. This is accomplished by providing within the wall itself a system of flashing and collection devices, with ample drainage outlets to the outdoor face of the wall. The method of pressure equalization, based on the “rain screen principle,” is generally a more sophisticated and complex solution, but is claimed by its proponents to be completely infallible when properly applied. This method is discussed in detail in the AAMA publication entitled, “The Rain Screen Principle and Pressure Equalized Wall Design,” and need not be expounded here. Briefly, it requires the provision of a ventilated outer wall surface, backed by drained air spaces in which pressures are maintained equal to those outside the wall, with the indoor face of the wall being sealed against the passage of air. The successful use of these methods depends on a clear understanding of the action of wind driven rain, careful detailing and, of course, proper installation. And in both cases ample weepholes or drainage slots, strategically located and properly baffled, play a critical role.
Page 14
INDOORS
ALUM. MULLION
THERMAL BREAKS
OUTDOORS A
B
C
When a minimum portion of the mullion surface is exposed to cold outdoor temperatures, as in A, its indoor surface temperature is higher than when the ratio of exposed areas is reversed, as at B. Drawing C shows a still more effective method of eliminating condensation on the interior faces of framing members, the provision of a thermal break within the member itself. FIGURE 10 OTHER DESIGN CONSIDERATIONS Often there are other design criteria, perhaps of major significance also, which influence the design of the wall. Generally these relate to the control of the environment within the building-other important aspects of the filtering function of the wall. Among the more important of these, the following should be noted:
1. 2. 3. 4.
Moisture Control Because metal and glass are not only impermeable to moisture, and thus highly efficient vapor barriers, but also have low heat retention capacity, the control of condensation is essential in any metal curtain wall design. Unless proper controls are provided, moisture, or even frost, may occur on the indoor face of the wall, and condensation may collect within the wall, causing damage which can become serious before it’s detected. Fortunately, the control of moisture is a comparatively simple matter, provided that the problem is anticipated and preventive measures are incorporated in the wall when it’s built. An understanding of the causes of condensation, where it will likely occur, and how to minimize its potential damage is essential, if trouble is to be avoided. But to explain these matters is beyond the scope of this summary review, intended only to flag out the importance of the matter. In capsule form, the important precautions to be remembered are these:
AAMA CW-DG-1-96
A vapor barrier should be provided on or near the indoor side of the wall; Impervious internal surfaces should be sufficiently insulated to keep them warmer than the dew point of the air contacting them; Provision should be made for the escape of vapor to the outdoors, and The wall should be so detailed that any condensation occurring within it will be collected and drained away.
Thermal Insulation In some cases the insulating value of the wall may be one of the major design considerations. Whether to reduce heat loss and prevent condensation in cold weather, or to minimize heat gain and air conditioning cost in hot weather, reduction of the overall U-value of the wall is usually a good long-term investment. Metal and glass are materials which inherently have low resistance to heat flow, but with proper attention to details aluminum curtain walls can be designed to provide good thermal performance. Generally this is accomplished by minimizing the proportion of metal framing members exposed to the outdoors, eliminating thermal short circuits by means of “thermal breaks,” using double rather than single glazing, and providing good insulation in the large opaque areas of the wall.
Page 15
Sound Transmission Under normal conditions, even in densely built urban areas, metal curtain walls compare favorably with any other wall construction having equivalent fenestration, as a barrier to airborne sound. However, the increasing concern with noise pollution and the mushrooming of building near airports has focused attention on the need for “soundproofing” exterior walls. According to the law of mass, the transmission of sound through any barrier is inversely proportional to the mass of the barrier, and any lightweight construction such as metal curtain wall can claim no natural advantage as a sound barrier. But with careful detailing, based on an understanding of the principles of sound transmission, aluminum curtain walls have been designed to provide quiet enclosures near many airports. It must be remembered that the efficiency of a barrier to airborne sound depends, in large degree, upon its weakest link, and the weak links in most walls are glazed areas and openings, however small the latter may be. Where a high degree of sound insulation is required, air leakage through the wall must be minimized, and double glazing, well separated and sealed, is usually essential.
SUMMARY Aluminum curtain wall is uniquely amenable to the concept of the wall as a filtering system. But because of the nature of its component materials the technology of its design involves some special considerations of major significance. Chief among these are structural integrity, provision for movement and weathertightness, and the latter two in particular require careful attention to details. Care must be taken also to prevent condensation within the wall itself. In the structural design of aluminum curtain wall, stiffness rather than strength usually governs, and because the wall does not rely on its weight to provide stability, its anchorage to the building frame must be designed with due regard to both the suction and pressure effects of wind loading. Because of thermal effects and other causes, substantial movements occur at many of the joints between its parts, and provision must be made to accommodate these movements without jeopardizing the integrity of the wall. And because the wall materials are impervious to water, the major impact of heavy rains is concentrated at the joints also. With the dual requirement of accommodating movement while still preventing leakage, joint design becomes the key to wall performance, and the most critical aspect of design. Weathertightness, often regarded as the chief measure of acceptable performance, is achieved, not by attempting to maintain a completely unbroken impervious membrane at the other wall surface, but by other more dependable means. Just as with any other type of exterior building wall, thermal insulating value may be an important consideration in many locations and the necessity of low sound transmission is also sometimes a requirement affecting wall design. Such requirements are satisfied in aluminum curtain walls by applying in their design the accepted principles of heat flow and acoustics. The performance of any metal curtain wall depends in large measure upon how well its designer understands not only the principles of natural laws but how they affect the detailing of the wall. In the belief that more information on those matters will be helpful, some of the subjects discussed only briefly here will be examined in much greater depth, with supporting data, in future publications.
AAMA CW-DG-1-96
Page 16
GUIDELINES FOR THE ARCHITECT IN DETAILING THE WALL Because aluminum curtain walls are factory-made products, the role of the architect in their design and detailing is somewhat different than his role in the design of a wall built on the site, using traditional materials. Some architects who have been using metal curtain wall for years have come to recognize this, and have learned by experience how to minimize their costs in detailing the wall, yet insure the achievement of their design concepts. But those less familiar with metal curtain wall construction, or perhaps using it for the first time, are likely to spend a great deal of time in minute detailing, only to discover later that some of their labor was wasted. Often the details and joinery of the wall as shown on the shop drawings for their approval, and as later built, differ from those they provided, yet are completely acceptable. This seeming inconsistency can readily be explained. The architect cannot be an expert in all fields; he can hardly be expected to know as much about metal-working techniques as those who have made it their life's work. But he is the expert in matters pertaining to aesthetics, and he is responsible for establishing performance requirements. Insofar as his details reflect these concerns, the architect's requirements, providing they are reasonable and logical, should be faithfully met.
ADVICE DURING EARLY DESIGN STAGES Whether the curtain wall being considered is for a small one-story building or a monumental high-rise structure, it's generally advisable to call on the advice of one or more wall manufacturers before finalizing the design. Manufacturers are usually pleased to provide such advice and counsel at no obligation, and prefer to have the opportunity to offer suggestions while the concept is still fluid.
Reliable manufacturers of aluminum curtain wall make every effort to satisfy the architect's stipulations in respect to the constituent materials, the appearance and the performance of the wall, though they may sometimes question the wisdom of some of his decisions. They quite naturally prefer, however, to meet these criteria by methods which they have found by long experience to be the most dependable and efficient. The manufacturer sincerely desires not only to assist the architect in the development of his design, but to help him present its concept and details in such a manner that will minimize waste effort on his part, yet insure a clear understanding of his intent. In the hope of clarifying this interest, it is the purpose of this article to recommend efficient detailing procedures and to call attention also to certain considerations often overlooked in architects' designs.
With their extensive experience and their knowledge of production processes and installation methods, the wall fabricators are able to analyze the general concept, recognize potential problems and offer suggestions which will aid the designer, facilitate production and usually result in cost savings. The failure to seek such advice, and profit by it, often leads to difficulties later. It's not unusual to discover, for example, after the job is out for bids, that some parts of the wall, as detailed, are very difficult or practically impossible to produce, or that they are structurally inadequate; that the size of the wall units is such as to create transportation problems; that the tolerances allowed in the design are unrealistic, or that there isn't enough clearance provided to permit easy installation.
The need for, and value of, the manufacturer's advice varies, of course, with the size, importance and complexity of the job in question. Even when some type of standard* wall is to be used, as is common practice on many of the smaller buildings, there are often certain details requiring clarification, or practical limitations to be explained, especially when modifications of the design are contemplated. If the wall is to be custom designed, the importance of competent advice by qualified manufacturers is usually much greater, regardless of the size of the job. *Aluminum curtain wall is generally classified as either standard or custom. For clarification of the distinctions, see Aluminum Curtain Wall, Types and Systems.
Due to his limited experience, the architect, working alone, may not recognize the importance of such matters, which are critical considerations in metal curtain wall design. But with the aid of competent advice by manufacturers during the early stages of design, he can avoid many of the difficulties that might otherwise be encountered.
AAMA CW-DG-1-96
Page 17
THE ARCHITECT'S DETAILS The amount of detailing which should be done by the architect depends on several factors: whether it's a standard or custom wall, the size of the job, and whether the contract is to be negotiated or competitively bid. Standard Walls For obvious reasons, less detailing is usually required for a standard type of wall than for custom wall, regardless of the size of the building. It is essential, of course, that the architect establish the wall pattern and locate all principal members, designate the infill materials to be used and the finishes required, the type, size and location of operable window units, if any, and provide details of perimeter conditions where the wall adjoins other materials. The cross sectional dimensions of framing members should be shown, as in most standard systems a range of sizes is offered. In selecting the sizes of principal members strict attention should be paid to the manufacturer's data regarding their structural capacities, and these capacities should not be exceeded. The type and thickness of glass, and the material to be used for opaque panels must be shown, but the specific glazing methods to be used will likely vary with different wall manufacturers. It should be made clear, however, that only those methods meeting the specified performance standards will be accepted. Figure 11 illustrates the extent of detailing normally required for a small two-story standard wall which is to be bid competitively. There is no need to detail standard members, but in cases where such members are to be modified, sufficiently large scale details [at least 75 mm (3 in) scale; preferably larger] should be provided to clearly explain the nature and extent of modification. In many installations the surround conditions are unique; there may be special sill conditions, special corner
AAMA CW-DG-1-96
treatments or a specially designed coping. It is essential that all such conditions be clarified by large scale details showing, not the fabrication and joinery, but the profiles desired, along with all critical dimensions and the clearances to be provided. Most manufacturers of standard wall systems provide large scale or full size section drawings, and often engineering data as well, to facilitate such detailing. If the work is to be done under a negotiated contract, with the wall manufacturer selected in advance, the use of that manufacturer's details in preparing the architectural drawings will, of course, present no problem. But usually competitive bids, by fabricators of several comparable systems, will be required and this may raise questions as to how the details should be prepared. Unlike structural steel sections, the sections used in standard curtain wall systems offered by different manufacturers are not identical. They are "standard" as far as their own producer is concerned, but not a standard of the industry. But there is a good deal of similarity in general appearance between the sections offered by one manufacturer and those offered by several of his competitors. If the performance requirements and the aesthetic effect desired are clearly defined, the architect will generally experience little difficulty in obtaining a sufficient number of bids to ensure a good competitive price. This may be done by listing several acceptable bidders, although the details shown represent one specific system. To ensure that the bids presented by competitors represent work of acceptable appearance and quality the architect should specify that they submit for his approval, well in advance of the bid date, sufficient details to show how they propose to construct the wall.
Page 18
3785 mm (12 ft 5 in) 25 mm ± 6 mm (1 in ± 1/4 in)
406 mm ± 12 mm (1ft 4 in ± 1/2 in) 2
6
3734 mm (12 ft 3 in)
6 mm (1/4 in) PLATE GLASS HOPPER VENT
WOOD HANDRAIL BEHIND GLASS 3
STEEL ROOF DECK 64 mm (2 1/2 in) 686 mm (2 ft 3 in)
914 mm (3 ft)
FLUSH ALUMINUM SPANDREL PANELS
5
FASCIA NOT IN C.W. CONTR
MASONRY OPNG. 7.7 m (25 ft 3 in) ± 6 mm (1/4 in) CURTAIN WALL 7.7 m (25 ft 2 in) 6 EQUAL BAYS 1
FLUSH ALUMINUM PANEL
64 mm (2 1/2 in)
1
METAL SOFFIT
4
292 mm (1 ft 1/2 in)
NOTE : GLAZING METHOD TO BE SPECIFIED BY WALL MANUF.
WOOD HANDRAIL
3100 mm (10 ft 2 in)
ELEVATION
PLASTER FACE TO CEILING HEIGHT
GRADE
CONCRETE
6.35 mm (1/4 in) PL. GLASS ELEC. HEATER
6 mm (1/4 in)
METAL COVE
CALKING
140 mm (5 1/2 in) MULLION
SHIM & GROUT SPACE
4 VERTICAL SECTIONS
6 mm (1/4 in) PL. GLASS
5
SOUND INSULATION
FIN. 2ND FL.
STEEL FLOOR DECK
FLUSH ALUMINUM PANEL
44 mm (1 3/4 in)
100 mm (4 in) 140 mm (5 1/2 in)
WD HANDRAIL
3
2
FIRE STOP
±152 mm (6 in) PAINTED METAL COVER NOT IN CURT WALL CONTR.
6
7.7 m (25 ft 2 in) CURTAIN WALL 7.7 m (25 ft 3 in) ± 6 mm (1/4 in) M.O.
3100 mm (10 ft 2in) TO BOTTOM OF C. WALL
25 mm ±6 mm (1in ±1/4 in)
ALUMINUM WATER
152 mm (6 in)
140 mm (5 1/2 in)
140 mm (5 1/2 in) MULLION
635 mm (2 ft 1 in)
3100 mm (10 ft 2 in)
914.4mm (3 ft 0 in) TO FIN FLOOR
ELEC. HTR.
PLATE GLASS
12 mm (1/2 in) ± 3 mm (1/8 in)
VERTICAL SECTIONS
PLAN SECTIONS
FIGURE 11
AAMA CW-DG-1-96
Page 19
Custom Walls As each custom wall is substantially unique in design, though it may employ certain standard sections, its detailing presents a somewhat different problem. It must, of course, be completely detailed, but the bulk of this work may be done by the manufacturer rather than the architect, as will be explained. Generally the architect has two options as to how he may handle this aspect of custom wall work. He may specify that selected bidders submit with their bids certain drawings explaining in detail their proposals, or he may choose to provide his own complete details on which all bids are to be based. In either case he controls the aesthetics of the design and it is his responsibility to establish the performance requirements of the wall and see that they are met. The choice between these two procedures depends on several factors: the time allowed for the preparation of bids (especially on jobs to be bid competitively); the size of the job; the expertise of the architect's staff, and above all, the prior determination as to whether the manufacturers to be asked to bid will be willing to provide proposed details for the job in question. Each of the two procedures has its advantages and disadvantages, as will be explained. Under the submission procedure each bidder submits for the architect's approval "submission" or "proposition" drawings showing how he proposes to construct the wall. All such drawings are, of course, developed from the design provided by the architect in his contract documents. The architect's details must therefore clearly show the aesthetic requirements, e.g. the dimensioned locations of all principal members, large scale details showing their exterior profiles, the preferred glazing methods, general methods of anchorage and so forth, and he must state in his specifications the performance requirements to be met. He should not undertake the complete detailing of all parts of the wall and their methods of assembly. This becomes the obligation of the bidders, each of whom will likely propose details differing not only from each other but from those which the architect would have prepared. The drawings submitted by the bidders usually include only representative typical details sufficient to show the proposed character of construction; they are by no means complete shop drawings. At a later stage, when the successful bidder submits his complete shop drawings, all aspects of the design will, of course, be subject to the architect's approval. An example of drawings used in this procedure is shown in Figure 12. Drawing A is the architect's detail at the edge of a masonry opening, and submission drawings by four different bidders are shown
AAMA CW-DG-1-96
in Drawings B, C, D and E. Note that in at least one of the proposed designs the location of the finished plaster in respect to the glazing bead is questionable. This method of developing custom wall details is commonly used with negotiated contracts, but is applicable also to jobs bid competitively, provided that sufficient time is allowed in the bidding period for bidders to prepare their proposed details. It must be recognized, however, that the proposals submitted by the various bidders may very likely use differing methods of glazing, anchorage or other details affecting the work of collateral trades. It becomes essential, therefore, that for each submission approved as being acceptable, information as to how that design will affect the work of other trades be provided to subcontractors bidding that work, so as to avoid either the duplication of bids or the omission of essential items. While this "submission procedure" does not require complete detailing by the architect, it does impose on him the necessity of being able to intelligently evaluate the merits of each proposal and judge whether it will provide an attractive and trouble-free wall. In cases where the General Contractor has already been chosen, he should also be available to advise as to the relative merits of the proposals taking into account their overall cost effects. It is essential, in any case, that the architect have a knowledge of the fundamental requirements of good curtain wall construction and that he understands how the materials going into it will react under the conditions to be imposed upon it. Under the alternative procedure, when submission drawings are not required of the bidders, the architect's details must be much more complete, as they establish the basis on which all bids are tendered. It is the architect's responsibility in this case not only to provide details showing the materials and finishes to be used, the methods of glazing, provisions for accommodating movements, methods of anchorage, all critical dimensions and profiles, and clearances to be provided, but also to verify the structural adequacy of his design. Even under this procedure, however, he should not attempt to provide details so complete that they may be used as shop drawings. He should leave to the discretion of the fabricator the choice of certain fabrication and connection details, the internal configuration of extruded sections and the location of inconspicuous non-working joints. He should anticipate, too, that some of the bidders may suggest certain revisions in his details, and should be receptive to such suggestions. If they are offered and accepted during the bidding period, all other bidders should, of course, be so advised.
Page 20
PLASTER FIN. 19 mm (3/4 in)
SNAP-ON 33 mm (1 5/16 in) 16 mm (5/8 in) 6 mm (1/4 in) PL. GLASS
12 mm (1/2 in)
150 mm x 100 mm x 6 mm (6 in x 4 in x 1/4 in) 50 mm (2 in)
37 mm (1 1/2 in)
A
STONE 70 mm to 8 0 mm (2 3/4 in to 3 3/8 in )
56 mm (2 1/4 in)
37 mm (1 1/2 in) 19 mm (3/4 in)
SHIMS
EXTRUDED CLIPS 50 mm (2 in) LONG
CLIPS
B
C
CLIP ANGLES
SHIMS
D
E
FIGURE 12
AAMA CW-DG-1-96
Page 21
MANUFACTURERS' SUGGESTIONS OFTEN HELPFUL Unquestionably it is the architect's business to determine the aesthetic character of the wall, its overall pattern and texture, the materials to be used, the proportions of its elements and the profiles of its members. Most architects, though, can claim no expertise in metal fabrication techniques. For them to insist upon the manufacturer following details which he considers impractical not only places the burden of responsibility upon themselves but is likely to result in dissatisfaction. Most successful curtain walls are the result of a team effort, with the architect/engineer, the contractor, and the fabricator pooling their knowledge and talents to produce an attractive and efficient design. Frequently the wall manufacturers bidding the work suggest minor changes in the design or detailing of a curtain wall which result in improvements, and usually a savings of cost. The following case histories of actual experience are examples of such instances.
150 mm (6 in)
The mullion section shown on the architect's details was a 90 mm x 150 mm (3 ½ in x 6 in) tubular section as shown at A. When costs had to be reduced to meet the budget, the manufacturer, by engineering analysis, found that this section was far stronger than necessary and wasteful of material. He advised the architect that the 50 mm x 90 mm (2 in x 3 ½ in) section shown at B would have ample strength and stiffness, and this much smaller section was approved and used.
90 mm (312 in)
EXAMPLE NO. 1
50 mm (2 in)
90 mm (312 in)
B
A
FIGURE 13: EXAMPLE NO. 1 EXAMPLE NO. 2 This is one of those cases where the architect, contrary to the advice of the manufacturer, insisted upon retaining a detail which, because it was impractical, not only increased costs but failed to achieve its intended purpose. The detail, as shown at A, required a 6 mm x 3 mm (1/4 in x 1/8 in) "accent groove" in the face of the jamb frame adjoining a precast facing. The dimensions to the outer edge of the groove and the width of the masonry opening were shown to be identical. The manufacturer pointed out that, since no allowance was made for working tolerances, the strict alignment intended could not be achieved and the aesthetic effect of the groove would be lost. But the architect could not agree, so special extrusions were made, and with special tooling the design was provided. Tolerance in the metal frames were held within ±0.8 mm (1/32 in), but the precast concrete, when installed, varied by as much as 12 mm (1/2 in). The result, as exaggerated at B, was that in some places the groove was entirely covered, while in others the jamb face beyond the groove was exposed.
1000 mm (42 in) OPENING HEIGHT
GROOVE
MASONRY
A 6 mm x 3 mm (1/4 in x 1/8 in) ACCENT GROOVE
PLAN SECTION AT 'A'
B
FIGURE 14: EXAMPLE NO. 2
AAMA CW-DG-1-96
Page 22
EXAMPLE NO. 3 In this case the architect's design called for mullion spacing shown in Drawing A, requiring two widths of glass and glazing frame in each typical bay. The manufacturer suggested that consideration be given to relocating the mullions as shown in Drawing B, making all of the glazed units the same width. This suggestion was accepted, resulting in a substantial saving.
CL COL.
C L COL. 6600 mm (21 ft 8 in)
DOUBLE MULLION
MULLIONS
GLASS 1320 mm (4 ft 4 in)
1320 mm (4 ft 4 in)
660 mm (2 ft 2 in)
1320 mm (4 ft 4 in)
1320 mm (4 ft 4 in) 660 mm (2 ft 2 in)
5200 mm (17 ft 4 in)
A - ARCHITECT'S DESIGN
60 mm (2 1/2 in)
1270 mm (4 ft 2 in)
1270 mm (4 ft 2 in)
700 mm (2 ft 3 1/2 in)
1270 mm (4 ft 2 in)
1270 mm (4 ft 2 in)
5200 mm (17 ft 1 in)
60 mm (2 1/2 in) 700 mm (2 ft 3 1/2 in)
B - MODIFIED DESIGN
FIGURE 15: EXAMPLE NO. 3
NOTE: It should be recognized that the advisability of changing from a solid to a hollow extrusion should not be considered a general rule, but depends on the quantities involved. In this case the value of the metal saved far exceeded the increase in die and extrusion costs.
AAMA CW-DG-1-96
12 mm (12 in)
72 mm (2 7 8 in)
100 mm (4 in)
For this wall the mullion section detailed by the architect, as shown at A, called for a heavy extrusion. The manufacturer suggested changing to a lighter extrusion having the same exterior profile, and adequate stiffness, as shown at B. Although this is a more complex configuration, incorporating a hollow element, its use resulted in a saving of 45,360 kg (100,000 lbs) of aluminum on this one job. Also, by substituting a snap-on cover on the indoor face, the modification eliminated the field labor required for the screw assembly of mullion parts.
16 mm (5 8 in)
EXAMPLE NO. 4
87 mm (312 in)
A. ARCHITECT'S DESIGN
B. MODIFIED DESIGN
FIGURE 16: EXAMPLE NO. 4
Page 23
EXAMPLE NO. 5
230 mm (9 in)
As in this case, architects sometimes unknowingly detail members in ways that require special control measures resulting in higher costs. Here the architect required a color anodized mullion section of such depth that it required two extrusions, as shown at A. The manufacturer, foreseeing possible problems in color matching, suggested the modified design shown at B, pointing out that the recess in the side face of the mullion would serve not only to conceal the connecting screws but also to make less noticeable any slight variation in color between the two parts. The architect agreed, and the modified design was adopted.
70 mm (23 4 in)
A
B
FIGURE 17: EXAMPLE NO. 5
AAMA CW-DG-1-96
Page 24
EXAMPLE NO. 6 This is a case where the architect's details, shown at A, were followed when the wall was built, but altered for a later building. As will be seen, the original design included an applied rectangular member used to frame and accent each 1295 mm x 3430 mm (4 ft-3 in x 11 ft-3 in) unit. Later the architect was asked to design another building using a very similar wall design, and the same fabricator was given the job. In detailing the second wall, however, the fabricator suggested simplifying the fabrication to reduce costs by providing the frame integrally, rather than as an applied element. This was done, as shown at B, and reduced the number of principal extrusions required for the mullion assembly from five to two.
VISION GLASS
OPAQUE GLASS
KEY ELEVATION
3
2 4
5 1
A. ORIGINAL DESIGN
35 mm (13 8 in)
SPRING CLIP
38 mm (112 in)
2
COMPRESSED SEALING TAPE 1
290 mm (1112 in)
B. MODIFIED DESIGN
90 mm (3 12 in)
FIGURE 18: EXAMPLE NO. 6
AAMA CW-DG-1-96
Page 25
EXAMPLE NO. 7 This is an illustration of how a minor detail, more significant than it seems, may be improved, not to reduce costs but to insure against objectionable appearance. The original detail of a welded lap joint is shown at A. With this detail there is likely to be discoloration of the exposed face opposite the weld, due to welding heat. At the manufacturer's suggestion the detail at B was substituted, providing a cover to hide discoloration. WELD
A
WELD
B FIGURE 19: EXAMPLE NO. 7
IMPORTANCE OF TOLERANCES AND CLEARANCES A common deficiency in architects' detailing of aluminum curtain wall construction is the failure to recognize the full significance of standard tolerances and to provide adequate clearances for installing the wall. Lack of attention to these matters often necessitates changes and adjustments in the field, not only delaying the work but usually resulting in unnecessary extra costs, and sometimes impairing the appearance of the wall. Because tolerances and clearances may be closely related, the two terms are often confused. The have distinctly different meanings, however, and this distinction should be clearly understood. A tolerance is a permissible amount of deviation from a specified or nominal characteristic, whether it be a dimension, color, shape, composition or other quality. In this discussion the concern is with dimensional tolerances. A clearance is a space or distance purposely provided between adjacent parts, either to allow for movements or for anticipated size variations, to provide working space, or for other reasons. The recognition of normal dimensional tolerances and the provision of proper clearances are of critical importance in several aspects of aluminum curtain wall design. One such area is the detailing of glazing frames, where ample edge clearance and sufficient "bite" are prime factors affecting glass performance. These considerations will be explored in detail in later articles dealing with glazing methods in general. Another area of even greater concern is the matter of tolerances in the building frame and the clearances provided between this structure and the curtain wall. It's appropriate that this latter subject be discussed here, as it is a matter of fundamental importance in the detailing of any metal curtain wall.
AAMA CW-DG-1-96
Metal curtain wall construction involves the covering of a field-constructed skeleton with a factory-made skin. Thus, like other modern systems of building, it requires the combination of two different levels of discipline in respect to dimensional control. In contrast to traditional masonry wall construction, where adjustments can be made by cutting, fitting and patching, significant deviations from true alignment in the building frame are incompatible, if not intolerable, with metal curtain wall construction. With the growing use of metal walls, this fact has become widely recognized, and dimensional tolerance standards for building frame are being upgraded. There are many architects, though, who fail to recognize the critical importance of strictly controlling the alignment of the building frame. It is not uncommon to find specifications in which no mention is made of the tolerance to be held in the construction to which the wall is being attached, and clearance dimensions shown on the drawings which allow for no such tolerance at all, yet the wall is expected to be installed "plumb and true". Such disregard of the "facts of life" have often led to delays and arguments during installation of the wall when, because the structure is not located where shown on the drawings, it's been found impossible to properly align the wall. The problem of allocating responsibilities for errors in such situations is always frustrating, and sometimes impossible to solve. And it is impracticable to attempt to avoid such problems by specifying that the wall manufacturer take field measurements before proceeding with fabrication. This would result in intolerable delays and excessive costs.
Page 26
BUILDING FRAME TOLERANCES Obviously, it's the architect's responsibility to control, by his details, his specifications and alert supervision in the field, the tolerances permitted in the basic building structure. In the case of a reinforced concrete building frame the maximum tolerances he should permit, except where otherwise specifically stated, should be those listed in Section 2.4.1 of the American Concrete Institute Standard Recommended Practices for Concrete Formwork (ACI 347). The tolerances permitted under this standard are shown in Table B. Tolerances for structural steel building frames should be specified to conform with the American Institute of Steel Construction (AISC) Code of Standard Practice, Section 7, paragraph 7.11, the provisions of which are shown in Table A. It must be recognized, though, that the ACI Standard applies only to reinforced concrete buildings, and the AISC Code only to steel building frames. Neither of these standards applies to buildings of composite construction (e.g. concrete floor slabs carried by steel columns) or to concrete encasing structural steel members (e.g. fireproofing). Obviously, though, the location of the face of the fireproofing on the steel, as well as that of the steel member itself, are both critical. As the alignment of composite constructions, fireproofing and masonry work are not controlled by referencing these standards, the architect should require that the location of all such materials contiguous to the curtain wall be controlled within tolerances which are, at most, no more than those specified in ACI 347. Should there be some doubt as to what these tolerances should be, the curtain wall manufacturer should be consulted for advice. In all cases the tolerances specified must be reasonable and realistic. It may be possible to limit them to less than those given in the standards referred to, but often this tends to increase costs, and on the larger jobs it can be expensive. When considering the necessity of abnormally tight tolerances it is always advisable to re-study the wall design in an effort to minimize or eliminate such requirements.
AAMA CW-DG-1-96
INSTALLATION CLEARANCES AND TOLERANCES With reasonable tolerances for the building frame established, it is equally important that the designer provide proper clearances, based on the full range of these tolerances, in detailing the wall design and its relationship to the building structure. It is common practice among architects to specify that the curtain wall be installed plumb and true within relatively small tolerances, and the wall contractor does his best to achieve this within the limitations imposed by the wall design on the one hand and the job conditions on the other. Although metal curtain walls, if their anchorage systems are properly designed, are inherently capable of being built nearer to true planes than most other types of wall, realistic installation tolerances depend, to some extent, on the area of wall involved. Particularly on the larger jobs, it may be advisable for the architect to establish practical and acceptable tolerances in consultation with the general contractor, the wall contractor and other sub-contractors responsible. It must be understood, however, that the wall can be installed within the tolerances established only if: 1) the building frame (or contiguous construction) is built within the tolerances specified for that work, and 2) the clearance dimensions shown on the architect's drawings provide adequate working space, taking into account in all details both limits of the tolerances allowed. Otherwise alignment of the wall as specified will likely necessitate delays and extra costs, or may even be impossible. The clearance necessary for installation of the wall will depend on the wall design and the limits of adjustment permitted by its anchorage details. If anchorage is required to the face of spandrel beams or columns or their fireproofing, more clearance will be needed to install fastenings than when the anchors are located on the top and/or bottom faces of beams and the sides of columns, as is usually done. In no case, however, should the actual clearance provided be less than 50 mm (2 in). The nominal clearance dimension shown on the drawings, then, should be equal to the actual clearance required plus the outward tolerance permitted for the adjacent construction, and should be determined on the assumption that this construction will be as far out of position in the wrong direction as is allowed.
Page 27
TABLE A - STANDARD TOLERANCES FOR STEEL BUILDING FRAMES (Excerpted from American Institute of Steel Construction, Code of Standard Practice-September 1, 1986, Section 7, Erection, Paragraph 7.11, Frame Tolerance) Erection tolerances are defined relative to member working points and working lines as follows: a) For members other than horizontal members, the member work point is the actual center of the member at each end of the shipping piece. b) For horizontal members, the working point is the actual center line of the top flange or top surface at each end. c) Other working points may be substituted for ease of reference, providing they are based upon these definitions. d) The member working line is a straight line connecting the member working points. The tolerances on position and alignment of member working points and working lines are as follows: Individual column shipping pieces are considered plumb if the deviation of the working line from a plumb line does not exceed 1:500, subject to the following limitations: The member working points of exterior column shipping pieces may be displaced from the established column line no more than 25 mm (1 in) toward nor 50 mm (2 in) away from the building line in the first 20 stories; above the 20th story, the displacement may be increased 1.5 mm (1/16 in) for each additional story, but may not exceed a total displacement of 50 mm (2 in) toward nor 75 mm (3 in) away from the building line. The member working points of exterior column shipping pieces at any splice level for multi-tier buildings and at the tops of columns for single tier buildings may not fall outside a horizontal envelope, parallel to the building line, 38 mm (1 ½ in) wide for buildings up to 90 m (300 ft) in length. The width of the envelope may be increased by 12 mm (1/2 in) for each additional 30 m (100 ft) in length, but may not exceed 75 mm (3 in). The member working points of exterior column shipping pieces may be displaced from the established column line, in a direction parallel to the building line, no more than 50 mm (2 in) in the first 20 stories; above the 20th story, the displacement may be increased 1.5 mm (1/16 in) for each additional story, but may not exceed a total displacement of 75 mm (3 in) parallel to the building line. Members Other Than Columns Alignment of members which consist of a single straight shipping piece containing no field splices, except cantilever members, is considered acceptable if the variation in alignment is caused solely by the variation of column alignment and/or primary supporting member alignment within the permissible limits for fabrication and erection of such members.
AAMA CW-DG-1-96
The elevation of members connecting to columns is considered acceptable if the distance from the member working point to the upper milled splice line of the column does not deviate more than plus 4.5 mm (3/16 in) or minus 7.5 mm (5/16 in) from the distance specified on the drawings. The elevation of members which consist of a single shipping piece, other than members connected to columns, is considered acceptable if the variation in actual elevation is caused solely by the variation in elevation of the supporting members which are within permissible limits for fabrication and erection of such members. Individual shipping pieces which are segments of field assembled units containing field splices between points of support are considered plumb, level and aligned if the angular variation of the working line of each shipping piece relative to the plan alignment does not exceed 1:500. The elevation and alignment of cantilever members shall be considered plumb, level and aligned if the angular variation of the working line from a straight line extended in the plan direction from the working point at its supported end does not exceed 1:500. The elevation and alignment of members which are of irregular shape shall be considered plumb, level and aligned if the fabricated member is within its tolerance and its supporting member or members are within the tolerances specified in this code. Adjustable Items The alignment of lintels, wall supports, curb angles, mullions and similar supporting members for the use of other trades, requiring limits closer than the foregoing tolerances, cannot be assured unless the owner's plans call for adjustable connections of these members to the supporting structural frame. When adjustable connections are specified, the owner's plans must provide for the total adjustment required to accommodate the tolerances on the steel frame for the proper alignment of these supports for other trades. The tolerances on position and alignment of such adjustable items are as follows: a) Adjustable items are considered to be properly located in their vertical position when their location is within 10 mm (3/8 in) of the location established from the upper milled splice line of the nearest column to the support location as specified on the drawings. b) Adjustable items are considered to be properly located in their horizontal position when their location is within 10 mm (3/8 in) of the proper location relative to the established finish line at any particular floor.
Page 28
TABLE B - STANDARD TOLERANCES FOR POURED CONCRETE BUILDING FRAMES (Excerpted from American Concrete Institute Recommended Practice for Concrete Formwork. ACI 347-78 Reaffirmed 1984, Section 3.3.1) Tolerances for reinforced concrete buildings† 1.
Variations from the plumb. a) In the lines and surfaces of columns, piers, walls, and in arrises In any 3 m (10 ft) of length ....................................................................................................................... 6 mm (1/4 in) Maximum for entire length......................................................................................................................... 25 mm (1 in) b) For exposed corner columns, control-joint grooves, and other conspicuous lines In any 6 m (20 ft) of length ....................................................................................................................... 6 mm (1/4 in) Maximum for entire length...................................................................................................................... 12 mm (1/2 in)
2.
Variation from the level or from the grades indicated on the drawings. a) In slab soffits*, ceilings, beam soffits, and in arrises In any 3 m (10 ft) of length ....................................................................................................................... 6 mm (1/4 in) In any bay or in any 6 m (20 ft) of length................................................................................................ 10 mm (3/8 in) Maximum for entire length...................................................................................................................... 19 mm (3/4 in) b) In exposed lintels, sills, parapets, horizontal grooves, and other conspicuous lines In any bay or in any 6 m (20 ft) of length.................................................................................................. 6 mm (1/4 in) Maximum for entire length...................................................................................................................... 12 mm (1/2 in)
3.
Variations of distance between walls, columns, partitions, and beams. 6 mm per 3 m (1/4 in per 10 ft) of distance, but not more than 12 mm (1/2 in) in any one bay, and not more than 25 mm (1 in) total variation
4.
Variation of linear building lines from established position in plan.................................................................. 25 mm (1 in)
5.
Variation in the sizes and locations of sleeves, floor openings, and wall openings. Minus......................................................................................................................................................... 6 mm (1/4 in) Plus .......................................................................................................................................................... 12 mm (1/2 in)
6.
Variation in cross-sectional dimensions of columns and beams and in the thickness of slabs and walls. Minus......................................................................................................................................................... 6 mm (1/4 in) Plus .......................................................................................................................................................... 12 mm (1/2 in)
† Variations from plumb and linear building lines on upper stories of high-rise structures [above 30 m (100 ft) high] are special cases which may require special tolerances. * Variations in slab soffits are to be measured before removal of supporting shores: the contractor is not responsible for variations due to deflection, except when the latter are corroboratory evidence of inferior concrete quality or curing, in which case only the net variation due to deflection can be considered.
AAMA CW-DG-1-96
Page 29
TABLE C - STANDARD TOLERANCES FOR PRECAST AND PRESTRESSED CONCRETE BEAMS AND SPANDRELS (SEE FIGURE 20) The following erection tolerances apply to beams and spandrels and particularly, precast element to precast element to castin-place concrete and masonry, and precast element to steel frame. a
=
Plan location from building grid datum.................................................................................................... ±25 mm (1 in)
a1
=
Plan from center line of steel*.................................................................................................................. ±25 mm (1 in)
b
=
Bearing elevation † from nominal elevation at support Maximum low ......................................................................................................................................... 12 mm (1/2 in) Maximum high .......................................................................................................................................... 6 mm (1/4 in)
c
=
Maximum plumb variation over height of element Per 300 mm (12 in) height......................................................................................................................... 6 mm (1/4 in) Maximum ................................................................................................................................................ 12 mm (1/2 in)
d
=
Maximum jog in alignment of matching edges Architectural exposed edges...................................................................................................................... 6 mm (1/4 in) Visually noncritical edges ....................................................................................................................... 12 mm (1/2 in)
e
=
Joint width Architectural exposed joints ....................................................................................................................±6 mm (1/4 in) Hidden joints ......................................................................................................................................... ±19 mm (3/4 in) Exposed structural joint not visually critical ......................................................................................... ±12 mm (1/2 in)
f
=
Bearing length ‡ (span direction) .......................................................................................................... ±19 mm (3/4 in)
g
=
Bearing width ‡ ..................................................................................................................................... ±12 mm (1/2 in)
*For precast elements erected on a steel frame, this tolerance takes precedence over tolerance dimension "a" † Or member top elevation where member is part of a frame without bearings. ‡ This is a setting tolerance and should not be confused with structural performance requirements set by the architect/engineer.
AAMA CW-DG-1-96
Page 30
a
d
BEARING AREA
e
f g COLUMN GRID LINE
PLAN VIEW
PRECAST CONCRETE BEAM c c
a1
d
b
DESIGN ELEVATION PRECAST CONCRETE BEAM
a
PRECAST OR CAST-IN-PLACE COLUMN COLUMN GRID LINE
C L
STEEL STEEL SUPPORT STRUCTURE
ELEVATION FIGURE 20: ERECTION TOLERANCES FOR BEAMS AND SPANDRELS
AAMA CW-DG-1-96
Page 31
TABLE C - STANDARD TOLERANCES FOR PRECAST AND PRESTRESSED CONCRETE (Con’d) FLOOR AND ROOF MEMBERS (SEE FIGURES 21 & 22) The following erection tolerances apply to floor and roof members and particularly, precast element to precast element, precast element to cast-in-place concrete and masonry, and precast element to steel frame. a
=
Plan location from building grid datum.................................................................................................... ±25 mm (1 in)
a1
=
Plan location from centerline of steel* ..................................................................................................... ±25 mm (1 in)
b
=
Top elevation from nominal top elevation at member ends Covered with topping ............................................................................................................................ ±19 mm (3/4 in) Untopped floor ........................................................................................................................................±6 mm (1/4 in) Untopped roof........................................................................................................................................ ±19 mm (3/4 in)
c
=
Maximum jog in alignment of matching edges (both topped and untopped construction) ................................................................................................... 25 mm (1 in)
d
=
Joint width 0 to 12 m (0 to 40 ft) member length..................................................................................................... ±12 mm (1/2 in) 12.1 to 18.2 m (41 to 60 ft) member length........................................................................................... ±19 mm (3/4 in) 18.3 m (61 ft) plus .................................................................................................................................... ±25 mm (1 in)
e
=
Differential top elevation as erected Covered with topping .............................................................................................................................. 19 mm (3/4 in) Untopped floor .......................................................................................................................................... 6 mm (1/4 in) Untopped roof.......................................................................................................................................... 19 mm (3/4 in)
f
=
Bearing length † (span direction) .......................................................................................................... ±19 mm (3/4 in)
g
=
Bearing width † ..................................................................................................................................... ±12 mm (1/2 in)
h
=
Differential bottom elevation of exposed hollow-core slabs ‡ .................................................................. 6 mm (1/4 in)
*For precast elements erected on a steel frame, this tolerance takes precedence over tolerance dimension "a" † This is a setting tolerance and should not be confused with structural performance requirements set by the architect/engineer. ‡ Untopped Installations will require a larger tolerance.
AAMA CW-DG-1-96
Page 32
a BUILDING Y GRID DATUM
BEARING BLDG X GRID DATUM
a
g
d
c
PLAN
CLEARANCE
PRECAST CONCRETE FLOOR OR ROOF MEMBERS
PRECAST CONCRETE FLOOR OR ROOF MEMBERS
e
b f
PRECAST OR CAST-IN-PLACE CONCRETE SUPPORT MEMBER BUILDING ELEV DATUM
ELEVATION
FIGURE 21: ERECTION TOLERANCES FOR FLOOR AND ROOF MEMBERS
AAMA CW-DG-1-96
Page 33
BLDG. Y GRID DATUM
a
f g a
BLDG. X GRID DATUM
g f d c
PRECAST CONCRETE FLOOR OR ROOF MEMBERS STEEL STRUCTURE
PLAN
a1
e C L OF STEEL STRUCTURE f
b
STEEL SUPPORT STRUCTURE
BUILDING ELEVATION DATUM
ELEVATION FIGURE 22: ERECTION TOLERANCES FOR FLOOR AND ROOF MEMBERS
AAMA CW-DG-1-96
Page 34
TABLE C - STANDARD TOLERANCES FOR PRECAST AND PRESTRESSED CONCRETE (Con’d) COLUMNS (SEE FIGURE 23) The following erection tolerances apply to columns and particularly, precast element to precast element. a
=
Plan location from building grid datum Structural applications ........................................................................................................................... ±12 mm (1/2 in) Architectural applications...................................................................................................................... ±10 mm (3/8 in)
b
=
Top elevation from nominal top elevation Maximum low ......................................................................................................................................... 12 mm (1/2 in) Maximum high .......................................................................................................................................... 6 mm (1/4 in)
c
=
Bearing haunch elevation from nominal elevation Maximum low ......................................................................................................................................... 12 mm (1/2 in) Maximum high .......................................................................................................................................... 6 mm (1/4 in)
d
=
Maximum plumb variation height of element [element in structure of maximum height of 30 m (100 ft)] 25 mm (1 in)
e
=
Plumb in any 3 m (10 ft) of element height............................................................................................... 6 mm (1/4 in)
f
=
Maximum jog in alignment of matching edges Architectural exposed edges...................................................................................................................... 6 mm (1/4 in) Visually noncritical edges ....................................................................................................................... 12 mm (1/2 in)
a d
BLDG. Y GRID
d BLDG. X GRID
a PRECAST CONCRETE COLUMN
f
SPLICE AREA
e
b
f
c
3 m (10 ft) ELEVATION DATUM
ELEVATION DATUM
a BLDG. X GRID DATUM OR Y GRID DATUM
FIGURE 23: ERECTION TOLERANCES FOR COLUMNS
AAMA CW-DG-1-96
Page 35
TESTING OF ALUMINUM CURTAIN WALLS Although the testing of curtain wall systems for their effectiveness in resisting wind and rain has now become rather common practice, such testing is by no means a universal requirement of all curtain wall designs. Often tests are essential, but frequently they are not. When they are needed, they should be the proper kind of tests, governed by appropriate criteria, but unnecessary testing is obviously a waste of time and money. It is essential, therefore, that the architect, before specifying any tests for aluminum curtain walls, fully understands the reasons for testing, what tests are advisable, and what performance criteria should be established as reasonable requirements. Present concepts of wall testing originated with the advent of metal curtain wall construction, though generally similar tests had been conducted on metal windows for some time prior to that. Two reasons in particular led to the introduction of wall testing at this time: first, it was not until factory-built wall systems came into use that representative pre-assembled units were available for evaluation by testing, and second, the problems of sealing against water leakage are far greater in metal-and-glass construction than in masonry construction. This is due both to the inherent nature of the materials themselves, and the fact that much larger wall units are used. Less water enters metal walls during a rainstorm than enters masonry walls, but whereas the relatively absorptive masonry construction absorb within themselves much of this water, later evaporating it, much of the water striking an impervious metal wall accumulates at the joints. These joints must be capable of accommodating movement, yet any water entering them must be drained back to the exterior or it may cause corrosion and may also appear conspicuously as a leak on the interior of the building. The need for pre-testing of curtain walls may therefore be said to have been dictated by the very nature of the materials being used. At some stage during its design development, any metal curtain wall should be tested for leakage of both air and water. But most standard types of wall, the walls which supply a large share of the market, have already been extensively tested, both by their manufacturers and by impartial testing agencies, during their design development, before they were placed on the market. And, in addition, such walls have been extensively tested in actual use. When using an established wall system of this type, further testing for the specific job at hand is usually unnecessary, provided that no special features or design changes are incorporated and the installation is to be identical in all respects with the system tested.
AAMA CW-DG-1-96
With custom type walls, it's a different matter. Most walls in this category contain new and previously untried features, and therefore should be tested. It follows then, that by far the bulk of performance testing specified by the architect and performed by commercial testing agencies is concerned with custom designed walls. In short, the need for testing depends upon both the type of wall being used and the circumstances of its use. If the manufacturer will certify that the wall, as it is to be installed on the building, has already been tested and qualified by an impartial authority and meets the specified criteria, further testing should be unnecessary. But when a previously unproven wall design is being used, thorough pre-testing is usually not only advisable but necessary. The architect should be sufficiently informed regarding the nature and value of testing to determine what testing procedures, if any, are appropriate. REASONS FOR, AND VALUE OF LABORATORY TESTING In general, laboratory pre-testing of metal curtain walls is aimed at evaluating performance of the wall under exposure to simulated environmental conditions before full scale production of the wall system is begun. A secondary benefit of such testing is that, in constructing the test specimen, or mock-up, an opportunity is provided to check installation procedures, and in some cases this experience in itself leads to design improvements. Laboratory tests may be conducted for either of two purposes: to provide the wall manufacturer himself with information about the performance of his design, or to provide official evidence and certification that the performance of the wall meets specified standards. Design check, or "exploratory" tests are made during the development of the wall design, and are conducted by the wall manufacturer, usually with his own facilities and staff. Such test may be unrealistically severe, even to destruction, in order to disclose design weaknesses and suggest potential improvements. Acceptance tests are those which are conducted for the purpose of verifying that the wall conforms with the architect's performance specifications, or to prove its acceptability to the architect or owner. These tests are conducted, or witnessed and certified, by an impartial test agency designated or approved by the architect. Either the facilities of the agency or those of the manufacturer may be used, but in either case the results must be reported and certified by the agency.
Page 36
It must be recognized, however, that even the most conscientious laboratory testing cannot reliably predict with accuracy the performance of the wall in actual use. To a large degree, field performance depends upon the care used in installing the wall, on proper anchorage, the fit of mating parts and the effectiveness of field seals. These, in turn, depend upon the alignment of the building frame, working conditions at the building site, quality of workmanship and proper supervision. Proper allowance for all of those unknowns cannot be made in laboratory testing, nor can the detrimental effects of time and aging be simulated. Nevertheless, standard laboratory performance tests do have substantial value. Although they provide no positive proof that the wall when installed will function properly, they often do reveal design weaknesses or fabrication faults requiring correction, and the discovery of such deficiencies in advance of production may well save many times the cost of conducting the tests. PERFORMANCE CHARACTERISTICS SUBJECT TO PRE-TESTING Almost any type of performance can be pre-tested by using a proper full-size specimen of the wall and the proper testing facilities. There are three performance characteristics in particular that are commonly investigated ⎯ resistance to air infiltration, resistance to water penetration and structural adequacy ⎯ and standard methods have been developed for conducting such tests. Other characteristics such as heat and sound transmission are also critical concerns in some cases and may require testing. All of these tests will be discussed, with the more common "standard" tests being examined in greater detail. The performance characteristics which are usually of greatest concern are structural performance under wind loading and the ability of the wall to prevent water penetration during heavy rain storms. These represent two levels of concern, however. Structural failure, of course, may endanger human life, so structural adequacy is a basic essential. The occurrence of water leakage will not likely be dangerous, but may cause discomfort and substantial property damage. It does not follow, however, that structural testing is more essential than testing for water penetration. In fact, the order of importance, as far as the need for testing is concerned, is usually the reverse. The reason for this is that structural requirements are well recognized, can be calculated with reasonable accuracy, and are found to be amply satisfied in most cases. Resistance to water penetration, on the other hand, cannot be accurately calculated or predicted, but requires testing for verification, and is often found to be deficient.
AAMA CW-DG-1-96
A third common reason for testing is to determine resistance of the wall to air infiltration, and this is a matter of particular concern when the design includes a number of operating window units. Generally, the concern about air leakage is of secondary importance, though the amount of air passing through the wall must always be limited to a small amount, usually a specified maximum, in order to minimize heat loss and condensation. Contrary to some beliefs, there is no direct and constant relationship between the amount of air infiltration and the amount of water penetration occurring in a wall. If the amount of air infiltration is high, the wall will likely be susceptible to water penetration also, but walls which are relatively airtight may also have serious water leakage problems. The resistance to both air infiltration and water leakage depend entirely upon the design details, and are essentially indeterminate, except by testing. Varying degrees of importance are attached to the tests for these three characteristics which are tested by standard methods, and the architect may, of course, be selective in specifying them. Each type of performance is measured individually, by its own test, and only those which are considered to be in doubt need be tested. In the experience of most commercial testing laboratories, the water penetration test is always specified, the structural test not quite as frequently, and the air infiltration test far less often than either of the others. It should be recognized, however, that major expense of testing is the cost of preparing and instrumenting the test specimen. After this is done, the difference in cost of running three or four different types of tests on the same specimen, rather than only one or two, is relatively small, provided of course that the laboratory is equipped to conduct all of the standard tests. Thermal tests, as applied to aluminum curtain walls and windows, are of several types and are essential to the determination of energy-conserving capabilities. Efforts to legislate energy conserving measures into building codes may make testing of thermal performance as common a requirement in the future as testing for air leakage, water penetration and structural strength is at present. Sound transmission tests are also being required more often as designers strive for wall construction which effectively reduces transmission of air-borne noise.
Page 37
THE TEST SPECIMEN It is essential that the wall test specimen be, as nearly as possible, a faithful representation of the intended design. It should be constructed just as the wall is to be installed on the building, using the same methods of support and attachment, similar conditions of continuity in all structural elements, the same type of glass, same sealants and so forth. As far as practicable, the building frame which supports the wall should also be simulated in the test set-up. In some cases, full-size steel or concrete framing has been constructed as part of the test structure, but in normal testing practice this expense is avoided by using heavy wood or steel members which provide equivalent stiffness and the same kind of support and anchorage as will be furnished by the actual building frame. In any case, all details of the intended anchorage system ⎯ the steel angles, clips, shims, brackets, bolts and welds ⎯ should be used on the test specimen just as detailed for the ultimate installation.
spaced not more than 7.5 m (25 ft) on centers, the width of the test specimen should be one full bay between such piers, plus the metal-to-masonry joints and representative width of masonry at both sides. The selection of the area of building facade to be represented by the specimen is, of course, the architect's decision. The choice deserves careful consideration, and should be made in consultation with the wall manufacturer and the testing laboratory.
Whenever possible, the same parties who will later be installing the wall on the building should also construct the test specimen. This is desirable for several reasons: 1) it provides an installation more representative of field workmanship than would likely be provided by laboratory technicians, 2) it acquaints the parties involved with the details of construction and the critical aspects of the installation procedure, and 3) it often leads to valuable suggestions by the workmen themselves regarding minor revisions in details that will facilitate installation. The size of the test specimen, as well as the selection of the wall area which it represents, are important considerations also. Usually the same test specimen is used for all three of the "standard" tests (air, water and structural), and the standard methods for these tests, which will be identified later, stipulate the general requirements as to size. Some types of thermal tests, if conducted by the same laboratory performing the standard tests, may also be performed on this same specimen. Certain types of thermal testing, and tests for sound transmission, are done by different laboratories and require different kinds of test specimens which may not be quite as large or elaborate, though the tests themselves are more complex. Some supplementary advice and recommendations regarding the nature of the test specimen, not provided in the standard test methods, should be noted. The area of wall represented by the specimen should include the most critical and vulnerable conditions, as illustrated in Figure 24. Horizontal joints between units, designed to accommodate movement, should be near the lower edge of the specimen, so as to collect rundown water from a sizeable area and, in some cases, when the extra cost is justified, especially with walls designed for use on tall buildings, it may be advisable to represent a corner condition, which is subject to the greatest pressures. If the wall design includes masonry piers or column facings
AAMA CW-DG-1-96
A curtain wall specimen is erected in a chamber for conducting tests under static pressure.
Page 38
ROOF COPING
WALL UNIT
B
X
EXTENDS ONE BAY AROUND CORNER
A
BUILDING CORNER
HORIZONTAL JOINTS BETWEEN UNITS
C
ORDER OF TESTING The three "standard," and by far most common, tests are generally conducted, quite logically, in order of severity of loading. First is the air infiltration test, which usually employs loads of from 75 to 300 Pa (1.57 to 6.24 psf). If the wall contains operable window units, for example, it is essential that this test be conducted on a dry wall, because the wetting of some types of weatherstripping may improve their sealing ability by as much as 100%, and only a test of the dry material can give a true measure of its capability of sealing against air infiltration. The second test is normally the test for water penetration. This may require uniform loading of 390 to 580 Pa (8 to 12 psf), depending on the criteria specified. The last test in the series, then, is the structural test, in which no less than the full design (wind) loading is applied. The amount of this loading varies widely, depending on design requirements, but may be as high as five to ten times the loading used for testing for water penetration. Tests for thermal and acoustical properties, if such are required, may generally be run at any time. Some of these tests, as previously noted, require special laboratory facilities and a different type of test specimen. If only critical interior surface temperatures and the effectiveness of thermal breaks are to be determined, the same specimen as is used for the standard tests may be employed, and it is customary to conduct such tests before subjecting the specimen to the water penetration test.
Areas of a curtain wall which can be used for test specimens are shown on an elevation of a typical curtain wall. Partial Building Façade Illustrating Possible Selection of Areas to be Represented in Test Specimen. Area A – Normal Choice Area B – Better Choice, but more elaborate and more expensive Area C – Most complex and expensive, but may be advisable in some cases Area X – Inadequate and unacceptable FIGURE 24
AAMA CW-DG-1-96
Page 39
TEST FOR AIR LEAKAGE This test is always conducted by the so-called "static" method, using an air chamber. Briefly, the procedure consists of constructing a relatively airtight assembly in the form of a large box, with the wall test specimen constituting or being contained in one of the two large sides of this box. Air is then supplied into, or exhausted from this assembly by means of a blower system, producing a pressure differential across the specimen, and the amount of air passing through the specimen itself is carefully measured. A schematic drawing of such a test assembly is shown in Figure 25. The procedure used for this test is described in ASTM E 283, "Test Method for Determining the Rate of Air Leakage Through Exterior Windows, Curtain Walls and Doors Under Specified Pressure Differences Across The Specimen." This standard, which originally was applicable only to the testing of windows, was later revised to include the testing of walls and doors. It supersedes the old NAAMM Standard TM-1-68T which was specifically intended for testing air leakage as well as water penetration and structural performance of curtain walls. The ASTM standard method of test for air leakage prescribes the testing procedure but, unless otherwise specified, calls for air leakage tests to be conducted at 75 Pa (1.57 psf), representing the velocity pressure of a 11 m/s (25 mph) wind. The 75 Pa (1.57 psf) is normally used for testing residential and commercial windows. Air leakage should not exceed 0.2 L/s•m (0.37 cfm/ft) of crack under this pressure. A pressure of 300 Pa (6.24 psf) representing a 22 m/s (50 mph) wind is used for testing monumental windows. Here too, air leakage should not exceed 0.2 L/s•m (0.37 cfm/ft) of crack. For windows having high thermal performance requirements air leakage should not exceed 0.2 L/s•m (0.37 cfm/ft), and for certain architectural applications it may be desirable to limit the leakage to considerably less than this. Refer to AAMA 101, "Voluntary Specifications for Aluminum and Poly (Vinyl Chloride) (PVC) Prime Windows and Glass Doors," for more specific values. Pressures higher than 75 Pa (1.57 psf) may also be used to test curtain walls, but for walls tested at 75 Pa (1.57 psf) good performance through the fixed glass and panel area would require that the air leakage not exceed 0.08 L/s•m (0.06 cfm/ft) of gross wall area. Since the permissible air infiltration of a curtain wall is usually specified in Liters per second per square meter (cfm per square foot) of projected wall area, a reasonable criterion for good performance ⎯ though subject, of course, to modification as dictated by circumstances ⎯ would be 0.08 L/s•m (0.06 cfm/ft) of wall area plus the air leakage in L/s (cfm) per linear meter (foot) of crack for the operable window units (if any) contained in the wall.
AAMA CW-DG-1-96
TESTS FOR WATER PENETRATION Two different methods are used for testing the resistance of walls to water penetration. One of these is the static method, using an air chamber, as described for the air infiltration test. The procedure is similar to that of the air test, except that higher pressures are used, and the outdoor (high pressure) side of the wall is subjected, while under pressure, to a uniform application of water at a specified rate. The other method, referred to as the "dynamic" method, employs a wind generator ⎯ usually an aircraft motor and large propeller ⎯ to simulate wind (and provide the test pressure), while water is fed into the air stream and onto the wall at the same rate as is used in the static method. In both cases the standard rate of water application is 3.4 L/m•m2 (5 gal/h•ft2) of wall surface. (Refer to AAMA 501, "Methods of Test for Exterior Walls.") Since the inception of curtain wall testing there have been, and undoubtedly will continue to be, valid differences of opinion as to the relative merits of these two methods of testing. In the first wall tests, in 1951, the dynamic method was used. It was soon found, however, that the static method is capable of producing much higher, and readily measurable pressures, provides results which are reasonably reproducible, and often reveals leakage failures not found by dynamic testing. Consequently, many in the industry have favored the static test, considering it to be more severe and therefore more reliable. In recent years, however, with the advent of pressure-equalized wall designs, there has appeared to be growing evidence that the dynamic test does have important significance. Some wall and window designs which have supposedly employed the rain screen principle to prevent leakage, and which have been tested by both methods, have been found to leak under dynamic testing, but under static tests, run at higher pressures, have shown no evidence of leakage. It is generally agreed that the dynamic method more closely simulates the action of wind, producing similar gusting, buffeting and vibrational effects and driving the water in all directions over the surface of the test specimen. But with present equipment the maximum pressure produced on the wall surface by this method is about 1200 Pa (25 psf). (Actual pressures on the wall surface are highly variable and practically impossible to measure; the maximum referred to is based on comparing deflections with those produced by measured static pressure.) The static method, on the other hand, easily provides much higher pressure which can be accurately controlled and measured. Some walls have "passed" the dynamic test with no sign of leakage, but have failed under the static test, while with other walls, as noted above, the reverse has been true. It might be concluded, from this experience, that certain types of design should be tested by one method, other types by the other, but to identify the distinction would require a rare understanding of the mechanics of water leakage. Consequently, some
Page 40
authorities recommend in cases of doubt, the use of both methods, and this has frequently been done, especially for some of the more important high-rise buildings. Many testing laboratories, as well as manufacturers, have good facilities for static testing, whereas the more expensive facilities for dynamic testing are comparatively rare. For the large majority of wall designs ⎯ standard type walls and low-rise custom walls ⎯ static testing is generally considered to be quite adequate. As long as the performance of the wall is being guaranteed by the manufacturer, as it usually is, the method of testing used to ensure this is chiefly his concern. When dynamic testing is required, however, to verify the effectiveness of pressure equalized designs or to insure the weathertightness of walls to be subjected to critical exposures on high-rise buildings, the additional cost of such testing will be minimal if it is done at a laboratory having facilities for both types of testing. As mentioned before, the major expense of testing lies in the preparation of the test specimen; after that is done, the difference in cost of running two, three or four tests on the same specimen is relatively small.
Nozzles spray water on a curtain wall test specimen in a static test for water penetration.
The standard method used for static testing is described in ASTM E 331, "Test Method for Water Penetration of Exterior Windows, Curtain Walls and Doors by Uniform Air Pressure." Since this standard was promulgated by ASTM it has been accepted as one method of evaluating the ability of a wall to resist water penetration. There is as yet no ASTM Standard Method for water penetration testing of either windows or walls by the dynamic method. The only standard for such tests on curtain walls is AAMA 501.1 "Standard Test Method for Exterior Windows, Curtain Walls and Doors for Water Penetration Using Dynamic Pressure." Regardless of which method is used, the amount of pressure and water flow to which the specimen is subjected is usually the same. Water is applied uniformly over the entire area of the test specimen by means of a nozzle spray system, at the rate of 3.4 L/m•m2 (5 gal/h•ft2) of projected wall area, measured at the face of the wall. This simulates the amount of "run-down" water at the base of the medium tall building in a heavy rain storm. The maximum test pressure applied, while the wall is being subjected to this flow of water, is from 10 to 20% of the positive design (wind) load, and it is applied for a period of 15 minutes. Generally the requirement is that there shall be no evidence of any "uncontrolled" leakage at any time during this period, i.e. no leakage which is not contained and/or drained away in such a manner as to cause no damage to the wall or to adjacent construction or finishes.
AAMA CW-DG-1-96
An airplane engine sprays water on a curtain wall test specimen in a dynamic test for water penetration.
Page 41
TEST FOR STRUCTURAL PERFORMANCE As previously mentioned, although the structural adequacy of any curtain wall is a prime essential, there is frequently no need for verifying structural performance by physical testing. If the wall is of conventional and relatively simple, straight forward design, the adequacy of structural members can often be determined by standard engineering analysis, and the thickness of glass required for various loadings may be found in standard reference charts. Many building codes will accept structural calculations by a qualified engineer in lieu of physical testing. The limiting factor in the structural design of framing members and panels is usually stiffness, rather than strength, but in the design of fastenings and anchors actual strength in shear or bending is usually the chief concern. If proper safety factors are used, calculations may often suffice. The typical curtain wall, however, is a composite assembly of many interdependent elements. It contains a number of non-grid joints and consequently there are reactions which are, to some extent, indeterminate. Its action under wind loading is influenced by the unpredictable performance of seals and gaskets, the tendency of members to twist, and other interactions which are difficult to anticipate. Physical testing may often, therefore, be the only reliable means of verifying overall performance, and it is well worth the small additional cost if the specimen is to be subjected, in any case, to static testing for air infiltration and water penetration. In exploratory testing, when the structural test is the final test in the standard series, useful information may be obtained by loading the specimens to failure, in order to determine the ultimate capacity of the wall and identify its weakest parts.
On high-rise buildings the negative pressure may often, in fact, be much higher than the positive pressure. The subject of design wind loads for walls, and their effect on wall design are discussed in other AAMA publications. The method recommended for structural testing of aluminum curtain walls is described in ASTM E 330, "Test Method for Structural Performance of Exterior Windows, Curtain Walls and Doors by Uniform Static Air Pressure Difference." The methods prescribed in this standard are widely used to test wall structures under uniform load conditions.
Curtain wall specimen before negative pressure is applied in a structural performance test.
Structural testing is conducted by the static method, using an air chamber as described for the air infiltration test. The pressure used, however, is much higher. It should be at least as great as the specified design wind load, and may be as much as 50% greater. As wind pressures generally increase with height above the ground, it is quite common practice to specify higher design loads near the top of tall buildings than are specified at lower levels on the same buildings. It must be remembered, too, that structural performance under outward, or negative, pressure is equally as critical as performance under inward, or positive pressures. The test specimen should always be subjected to pressure acting first in one direction, then the other, and the negative test (and design) pressure should be at least as great as the positive pressure, and it is critical, of course, that wall anchorage be designed to withstand these forces. Curtain wall specimen during application of negative pressure in structural performance test. Deformation of the wall is shown in the reflected images.
AAMA CW-DG-1-96
Page 42
WALL TEST SPECIMEN
SEAL
AIR CHAMBER
1.2 to 1.5 m ± (4 to 5 ft ±)
FLOW METER
ACCESS DOOR
CONTROL VALVE
SEAL
BLOWER SUPPLY OR EXHAUST PRESSURE MEASURING DEVICE
Section through a typical static pressure test assembly. FIGURE 25 THERMAL TESTS Some aluminum curtain walls ⎯ chiefly those to be used on projects of major significance in the colder climates ⎯ are subjected to tests designed to evaluate their performance under controlled temperature conditions. Such tests, which are relatively expensive, may have a variety of objectives, but most of them are used: •
• • •
to determine the effectiveness of thermal breaks incorporated in the wall design, measure indoor surface temperatures, and find where condensation will likely occur; to test the effectiveness of thermal insulation of steel building frame members; to investigate the effect of temperature fluctuations on exterior wall elements, or to determine the overall thermal conductivity, or Uvalue, of the composite wall assembly.
Except for the last type of test mentioned, these tests are conducted in a "double-chamber" apparatus, one variation of which is illustrated schematically in Figure 26. The wall test specimen is constructed with its indoor side sealed against one chamber, in which normal room temperature is maintained by heating and air conditioning and humidity can be controlled, while against the outdoor side of the specimen is sealed a second chamber, heavily
AAMA CW-DG-1-96
insulated, in which temperature can be varied over a wide range to simulate outdoor conditions. Air is circulated over refrigerated coils in this "outdoor" chamber to provide below-zero temperatures, and some suitable heating method is used to alternately provide wall surface temperatures in range of 71°C (160°F). Temperatures in both chambers are, of course, controlled, those in the indoor chamber being held constant while those in the outdoor chamber are cycled to simulate diurnal or seasonal changes. The design of proper apparatus for conducting this type of test is still under study, however, in some tests heat lamps have been used as a heat source in the outdoor chamber, as shown in the drawing, but some authorities maintain that this method provides erroneous results, due to unrealistic radiation effects. AAMA has developed a method of testing the thermal performance of standard size windows, using a test chamber with a warm room and cold room. Test procedures prescribe that the warm room temperature be held at 20°C (68°F) and the cold room at -8°C (18°F). The window sample to be tested is mounted between the rooms. Thermocouples mounted on the metal and glass surfaces exposed to the warm room measure the temperatures of these surfaces. From this data a Condensation Resistance Factor, CRF, can be calculated. With a known CRF the conditions of outside
Page 43
temperature and inside relative humidity under which a window will perform without objectionable condensation can be determined. But means of calorimeter or calibrated hot box techniques the thermal transmittance of windows can be determined in the same thermal test chamber. Since air leakage is important in total thermal performance this is also measured by applying a pressure differential across the test sample. The thermal performance of operable windows and curtain wall may be determined by testing in accordance with AAMA 1503, "Voluntary Test Method for Thermal Transmittance and Condensation Resistance of Windows, Doors and Glazed Wall Sections." Condensation resistance and thermal transmittance determined by this method can be very useful in determining the effectiveness of thermal breaks and wall insulation provisions. This type of thermal testing for windows has been increasing rapidly in response to the need for energyconserving construction to offset the high cost of fuels and help reduce future fuel shortages. The need for actual test data on the thermal performance of wall systems is equally as important as that for windows. Thermal testing of walls, therefore, will also increase rapidly. Variations of the double-chamber apparatus have been used not only to locate indoor "cold spots" and thermal short circuits, but to determine optimum methods of supplying building heat, to measure the range of temperature on building frame columns, to check the flatness of large spandrel panels under temperature fluctuations, and even to check for noise in the wall system caused by temperatureinduced movements at the joints.
Thermal chambers are often monitored and controlled by computer.
AAMA CW-DG-1-96
Completed test specimen built against indoor chamber. Chamber has not yet been insulated.
Closed test chamber with liquid nitrogen supply used for cold cycle and humidity control.
Page 44
INSULATION
WALL TEST SPECIMEN
A
INDOOR CHAMBER
OUTDOOR CHAMBER B A
ACCESS DOORS
A = FAN COIL UNITS
*
HEATED. AIR CONDITIONED CONTROLLED HUMIDITY
B = BATTERY OF
FOR COOLING (AIR FLOW PARALLEL TO WALL)
HEAT LAMPS IN GRID PATTERN
*
MAY BE STANDARD STATIC TEST AIR CHAMBER
Section through a thermal test assembly. FIGURE 26
SOUND TRANSMISSION TEST On occasion, when curtain walls are to be used on buildings at airports, or in other locations exposed to high noise levels, special sound barrier features have been incorporated in their design, and it has been found advisable to subject such specially designed walls to sound transmission tests. When required, such tests are conducted at one of several well-qualified acoustical laboratories, some of which are equipped to accommodate specimens up to 3 m to 6 m (10 ft x 20 ft) in size. The method used is that described in ASTM E 90, "Test Method for Laboratory Measurement of Airborne-Sound Transmission Loss of Building Partitions," and ASTM E 413, "Classification for Rating Sound Insulation." The design of curtain walls to reduce sound transmission, and the methods used in measuring transmission, are addressed in AAMA TIR-A1, "Sound Control for Aluminum Curtain Walls and Windows." EVALUATION OF TEST RESULTS As mentioned, the ASTM Test Methods referenced for the "standard" wall tests specify only the method to be used for testing; they do not stipulate the required standards of performance to be met. These must be spelled out by the architect in his specifications for the aluminum curtain wall. In the foregoing descriptions of the tests for air infiltration, water penetration and structural performance, mention was made of the criteria of performance generally recommended by AAMA. It will be recalled, however, that specific test pressures were stated for only the air infiltration test; the pressures to be used in the other two tests are functions of the design wind load, which must be determined by the architect. This load will vary not only with the location and height of the building, but sometimes also with the location of the wall area in question on the building facade. To comply with the requirements of the ASTM test methods, the testing agency is required to include in its report such information as: • the date of the test and the report, • a complete description and detailed drawing of the test specimen, showing all pertinent features of the construction, • a tabulation of the pressures used in testing, and • a detailed statement of performance and full description of any deficiencies of performance observed.
AAMA CW-DG-1-96
Page 45
On the basis of this report, the architect must then determine whether or not the wall has performed acceptably. In the event that it has failed to do so ⎯ as may happen ⎯ the necessary corrections in design should be made and whatever tests are necessary to confirm that this has been done should be repeated. Sometimes two or more re-runs may be required before all deficiencies are eliminated. Obviously, therefore, it's essential that the architect clearly states in his specifications that acceptability of the wall is contingent upon performance as specified. And, as previously pointed out, it must be remembered that even excellent performance under the most rigorous laboratory testing does not guarantee good performance on the building. Proper installation of the wall is equally essential. FIELD CHECKS DURING INSTALLATION Some of the leading wall installation contractors have found that a systematic "testing" of the wall for watertightness during the course of its installation on the building is highly advantageous and well worth the relatively small cost involved. Such field checks serve either to verify that the design and installation are satisfactory or to reveal deficiencies in the field work or possibly the need for minor adjustments, points of vulnerability not disclosed in previous laboratory testing. Consequently, they should be made early in the installation process, so that any faults discovered may be remedied before much of the wall is in place. On tall buildings the check is usually repeated once or twice at higher levels as an added insurance. Obviously it's much better to discover and remedy deficiencies as the work proceeds than to find them after the installation is completed, necessitating expensive callback remedial work in the whole wall. Several procedures are used for this field check for water leakage. (AAMA 501.3, "Field Check of Water and Air Leakage Through Installed Exterior Windows, Curtain Walls and Doors by Uniform Air Pressure Difference," AAMA 502, "Voluntary Specifications for Field Testing of Windows and Sliding Glass Doors," and AAMA 503, "Voluntary Specification for Field Testing of Metal Storefronts, Curtain Walls and Sloped Glazing Systems.) In some cases a portable air chamber has been employed, and static pressure tests have been conducted on completed sections of the wall in much the same manner as described for the standard laboratory static test for water penetration. More often, though, a simpler, less sophisticated and much less expensive "hose test" is used.
AAMA CW-DG-1-96
One such hose test is that prescribed in AAMA standard 501.2, "Field Check of Metal Storefronts, Curtain Walls, and Sloped Glazing Systems for Water Leakage." This method requires that after all framing members or wall units are installed in the lower two stories of the building, and at least 23 lineal meters (75 lineal feet) of this twostory height of wall is completely glazed, with the indoor wall finish omitted, the architect shall select a wall area two bays in width, to be checked. This area should include all typical horizontal and vertical expansion joints or other conditions where leakage may occur. Working from an exterior scaffold, while observing from the inside, the wall is tested in accordance with this standard. Water from a 19 mm (3/4 in) garden hose with a specified nozzle at a controlled flow rate is directed at the joints working from the lowest horizontal joint upward. If leakage occurs and the source cannot be identified, all joints must be covered with a waterproof adhesive masking tape and the test must be repeated by progressively removing the masking tape at the lowest joint and working across and up until sources of leaks have been identified. Wherever leakage is discovered, joints are made watertight in a manner acceptable to the architect, and the check is repeated as often as may be necessary. Any remedial measures found necessary are of course applied to the remainder of the wall as it is installed, even if this necessitates some change in the fabrication of the wall members or units themselves. The potential value of such a check as this, especially on large installations, is quite clear, even though the method used may be rather crude. However conducted, though, a field check should be considered only as a supplementary insurance measure. In no case should it be considered as a substitute for the more comprehensive, more precise and more rigorous laboratory test previously described. Laboratory testing is usually done in advance of full scale production of the wall, and deficiencies which it may reveal can thus be corrected before the wall design is completely finalized. Field checks, on the other hand, are made after the wall is manufactured and is in the process of being delivered to the site ⎯ too late to make any substantial changes in the design of the wall system.
Page 46
SUMMARY RECOMMENDATIONS The foregoing description of the various types of curtain wall tests and the methods used in performing them has been presented in an effort to clarify this rather complex subject. Hopefully it may help the architects to better understand their value and significance, to determine what tests, if any, are needed, and how they should be specified. In summary, the more important guidelines to be kept in mind, when considering this aspect of the specifications for aluminum curtain wall, are the following. 1. Whether or not tests are to be required, be sure that all performance requirements of the wall are clearly specified, and that these requirements are realistic, reasonable and consistent with the capabilities of the design. 2. Specify only such tests as are needed to confirm that the wall meets that performance criteria specified. If the wall manufacturer will certify that the wall, as it is to be installed, has already been properly qualified by a responsible impartial testing authority, further testing may be unnecessary. 3. Do not rely on testing as a substitute for thorough study of wall details in the design process. Such a course generally proves to be inefficient and wasteful.
5. The selection of the wall area to be represented by the test specimen should be carefully considered, in conference with the wall manufacturer, keeping in mind that building corner and coping details are often more vulnerable to leakage than typical wall areas. 6. Provide the testing agency with detailed drawings of the test specimen. This will facilitate both the location of any design defects revealed by testing and recording of any changes found to be necessary. 7. If possible, have the test specimen constructed by those parties who will be installing the wall on the building. 8. Make adequate allowances in the construction schedule for the conduct of tests and the effective use of test results. Usually this will require at least six to eight weeks. 9. See that a responsible member of your staff is present to witness installation of the specimen and all tests. 10. Specify that, in the event that the tests disclose deficiencies in either workmanship or design, such deficiencies shall be corrected and tests shall be repeated until all specified performance requirements are satisfied. 11. Consider the advisability of requiring a field check for water leakage.
4. If performance tests are required, they should be performed and certified by a responsible impartial testing agency, either at the agency's laboratory or at the manufacturer's plant.
AAMA CW-DG-1-96
Page 47
American Architectural Manufacturers Association 1827 Walden Office Square, Suite 550 Schaumburg, IL 60173 PHONE
(847) 303-5664 WEBSITE EMAIL
FAX
(847) 303-5774
www.aamanet.org
[email protected]
