Ensuring performance in mulled window systems

by arslan_ahmed | July 7, 2023 4:00 pm

Photos courtesy Simpson Gumpertz & Heger Inc.[1]
Photos courtesy Simpson Gumpertz & Heger Inc.

By John C. Stuart and Christopher N. Grey

When specifying fenestration, design professionals must select appropriate products for the project’s various opening sizes. However, today’s construction landscape is rapidly changing with higher material costs, tight tolerances, and expectations of floor-to-ceiling glazing in multifamily projects (Figure 1).

Many project teams are opting to specify delegated design mulled window systems that combine multiple windows into a single unit in a larger glazed opening, rather than a storefront or curtain wall system. While these mulled systems have become increasingly popular, their mulling components can create potential performance issues, including air infiltration, water penetration, structural failure, and reduced thermal performance.

Highlighting examples from past projects, this article explores the rationale and risks of using mulled window systems; reviews methods for addressing these risks; and demystifies standards, such as NAFS-17 and AAMA 450-20, that are crucial for design professionals specifying window performance.

Large mulled vinyl window installation.[2]
Figure 1 Large mulled vinyl window installation.

Fenestration: A performance perspective

AAMA/WDMA/CSA 101/I.S. 2/A440-17, commonly known as the North American Fenestration Standard or NAFS-17, defines window ratings based on air infiltration, water penetration, and structural (AWS) performance during laboratory testing.1 Rated products are assigned a performance class (R, LC, CW, and AW), with R (residential) windows having the least rigorous testing requirements, and AW (architectural window) having the highest and most diverse requirements (Figure 2, below). Products also carry a performance grade, indicating the maximum wind pressure that product has been tested to, without unacceptable deflection or damage. Water penetration resistance is also tested at a fixed fraction of the structural pressure, typically 15 or 20 percent, depending on the performance class. This system is widely used by window manufacturers throughout North America so designers can compare and specify window and door products.

NAFS-17 performance grade comparison.[3]
Figure 2 NAFS-17 performance grade comparison.4Photo courtesy

A window product line is typically tested only up to a size determined by its manufacturer that minimizes the frame depth and weight while optimizing for the most common residential opening sizes. Unfortunately, this optimization for smaller openings means the most widely available and economical products often do not have the stiffness to achieve the nearly floor-to-ceiling spans expected in many recent multifamily projects. In this scenario, many designers turn to storefront and curtain wall systems specifically designed for such spans. However, if the project’s design intent or budget do not allow for these systems, window manufacturers frequently propose a mulled window system as a solution.

Comparison of AAMA ratings by product type.[4]
Figure 3 Comparison of AAMA ratings by product type.Photo courtesy

 Mulled window systems

Mulled systems consist of multiple prefabricated window units, each with its own rating and often with relatively lightweight flexible frame extrusions, which are then combined in the field by fastening and sealing them together or attaching them to less flexible sub-framing members. These members include vertical mulls, horizontal transoms, subsills, head receptors, full perimeter receptor frames, or any combination of the above. This contrasts with storefront and curtain wall, which can also be used to fill a punched opening, but typically consist of deeper, heavier extrusions, and usually require a greater degree of field assembly and installer expertise to construct.

Each mulled system is assembled from multiple units, each with their own rating.2 This contrasts with composite windows (e.g. a casement over fixed configuration), which are similar in appearance, but are factory-fabricated with a shared mullion between each lite, so each sub-unit could not exist on its own if disassembled. The key difference for specifiers is that composite units carry one overall AAMA rating for the entire assembly while mulled systems often do not.

Comparison of AAMA ratings by product type.[5]
Figure 3 Comparison of AAMA ratings by product type.Photo courtesy

Since manufacturers often need to introduce a new component which is not common to the windows being mulled, this introduces new joinery and modes of deflection to the assembly, which were not present when the constituent windows were tested individually. However, since sub-framing is not directly addressed by NAFS-17, the connection between the window and its sub-framing may have a lower air, water, or structural performance rating than its constituent windows or no rating at all.3 The authors have encountered numerous projects where accepting an unrated mulled system without fully understanding these limitations resulted in air and water infiltration issues that did not meet performance expectations. Figure 3 shows examples of how this can impact performance depending on the configuration.

Designers navigating the complexities of mulled windows on their projects can easily find themselves in uncharted waters with few tested product options and unforeseen air, water, or structural issues. Several methods exist for charting a path past these risks such as avoidance strategies early in design, understanding and specifying a more appropriate performance standard such as AAMA 450-20, and as a last resort, accommodating unrated components via a program of careful shop drawing review, laboratory testing, and field testing to address any “weak link” sub-framing components that could compromise an installation’s performance.

Avoidance

Planning around performance requirements should begin early in design. An important rule of thumb to remember is product options and performance tend to decrease as opening size increases. A good first step to a successful project is to carefully consider opening sizes early in design and size each based on the limits of a preferred window product line with a known track record of success to ensure the initial design is realistic.

For projects with typical opening sizes of traditional residential construction, many options for standalone rated windows exist. However, for the larger window sizes seen in today’s multifamily construction, few manufacturers offer a rated window product that can fill oversized openings without sub-framing. The cutoff for what constitutes as an “oversized opening” varies by product type and the manufacturer, but designers can identify these issues by reviewing the basis-of-design document to understand product constraints and avoid costly redesigns, testing fees, or change orders for additional framing members.

However, even in smaller openings, the question of whether to use sub-framing may still come up. Contractors sometimes propose the use of receptor frames in inconsistent existing openings to provide tolerance during window installation. Similarly, designers may specify them to accommodate movements such as slab deflection. Receptor frames capture the full perimeter of the window, leaving space for adjustment and movement. However, this creates yet another system of joints where water and air leakage can occur, unless adequately vetted beforehand and properly detailed.5,6 Identifying where additional installation and movement tolerances are needed in advance can help a design team identify these risks early and adjust the design to mitigate them.

Alternately, when openings become large or irregular enough, it is often simpler to use a storefront or curtain wall which can perform over larger opening sizes, rather than using a mulled window. Storefront and curtain wall frequently utilize starter sills and receptor frames to accommodate adjustment and movement, but these have a greater track record and higher availability of rated components. For this reason, designers targeting more reliable performance in large, punched openings may start by developing elevations and specifications that are more conducive to storefront and curtainwall systems.

AAMA 450

For mulled fenestration, the most crucial addition to a specifier’s toolbox is a performance specification that references AAMA-450, Performance Rating Method for Mulled Combination Assemblies, Composite Units, and Other Mulled Fenestration Systems. Although this industry standard for rating systems with sub-framing components has been in place since 2000, in the author’s experience, it was not widely embraced until a few years ago—since NAFS-17 lists it as a voluntary standard—and it did not address crucial detailing considerations that impact air and watertightness until its most recent update in 2020.

AAMA 450 is now quickly gaining traction in the industry, possibly due to the streamlined testing and submittal process it allows for. Previously, applicable mulled window test reports were often unavailable since testing all possible mulling configurations was often infeasible; however, AAMA 450 provides a standardized method for combining testing and structural analysis to certify product lines in a variety of configurations. This simplifies the process for manufacturers and gives specifiers both design flexibility and better performance documentation for the systems they select.

Typical AAMA 450 performance chart format.[6]
Figure 4Typical AAMA 450 performance chart format.Photo courtesy

Incorporating AAMA 450 can be as simple as specifying a sub-framing that carry an equal or greater rating then its fenestration when tested in accordance with AAMA 450-20, and specifying a suitable AAMA 450-20 tested basis of design product. As an example, Figure 4 shows a generic mulled system performance grade chart. The chart can be interpreted by finding the mullion height and width of a mulled assembly to find the expected performance. The curved regions with a design pressure (DP) label indicate performance estimates established by calculation only, whereas the rectangular regions that include performance grade (PG) labels represent the largest size of each product physically tested. This chart also shows that even when using mulling, there are limits to how large a combined assembly can be. For example, a 1,625 mm (64 in.) wide by 2,159 mm (85 in.) tall window could be rated PG-80, but could only achieve a DP-20 performance when mulled together to span a much larger opening.

While AAMA 450 performance charts provide a user-friendly way of reviewing simple mulled systems, it has limitations. Unusual components designed for a specific project are often not tested. Similarly, end conditions where mullion components interface directly with perimeter flashings are often overlooked as well and can compromise the mull’s drainage strategy. Thus, even with a rated assembly, it is still important for the design team to carefully review the bid qualifications, test reports, and shop drawings of manufacturers to remain vigilant for these discrepancies between the tested system and the proposed system that could otherwise be overlooked.

Accommodation

The authors have observed the decision to use a mulled system is frequently driven by a contractor or manufacturer, either in a design-assist role early in the project, or as a substitution request due to perceived cost advantage. In cases where projects change to a mulled system instead of the original design intent, it can require significant effort from the designer to carefully review the system, identify its risks, and work with the project team to accommodate the changes if accepted.

The first step would be to request all available cut sheets and test data for a proposed system, including test reports documenting the ASTM E283, E331, and E330 laboratory tests necessary to validate air leakage, water penetration resistance, and structural performance respectively. If the proposed sub-framing is not included in these test reports, additional AAMA 450 certifications should be requested. If such reports are unavailable, some manufacturers may be able to perform additional analysis at the request of the design team on a project-by-project basis.

Proposed manufacturers should be vetted as early as possible to identify their limitations and account for them in design choices. Further, if the selection of the window manufacturer is driven by the contractor on a cost basis, designers must carefully research any substitution requests to confirm they are in fact equal to the specified system and flag any differences the owner or contractor may be unaware of.

 Mull and receptor frame configuration on educational institution project.[7]
Figure 5 Mull and receptor frame configuration on educational institution project.

The design team should also review shop drawings and identify areas for further development such as corner conditions, drainage strategies, and other internal seals, such as heel beads the manufacturer or installer may offer as a means of improving a system’s integration with its sub-framing. This, coupled with stringent requirements for mockup testing, can provide better assurance that the system will perform as intended. However, accommodating a change to use a mulled system after the design phase typically requires significantly more work from the designer, especially when attempting to develop mulling systems for the already lower-performing R and light commercial (LC) windows.

Case study: Water penetration

The risks associated with mulled windows can be most easily seen in practice. For example, one of the author’s projects involved an educational institution targeting an AW-40 rating for its punched windows in existing 3 x 3 m (10 x 10 ft) concrete openings (Figure 5). On this project, the author provided building envelope commissioning services to
the owner.

The contract documents specified a curtain wall system for the large, punched openings on the project. However, the contractor opted for an unrated receptor frame and vertical mull to combine two narrower rated windows and accommodate variations in the existing concrete openings. This mulled assembly had no performance rating and during field quality control (QC) testing, it exhibited severely compromised performance compared to a window tested in isolation. This consistently resulted in leakage at significantly less than the specified performance rating (Figure 6).

Leaks at the intersection of the window, receptor frame, and center mull.[8]
Figure 6 (top and bottom) Leaks at the intersection of the window, receptor frame, and center mull. Photo courtesy

In particular, the receptor frame presented a major weak point in the system. However, the vertical mulls also played a role by further complicating the assembly as installers attempted to integrate all components without published installation instructions or support from the manufacturer tailored to this mulled configuration.

Repairs were eventually implemented to bring the windows back into compliance with the contract documents. However, this issue brought with it significant costs, as well as additional seals that would need to be maintained at the owner’s expense, and aesthetic changes had to be made to the windows as the drainage path needed to be modified by adding field-drilled weeps.

Inset image: Building elevation with a four-way mulled window assembly. Right: Typical elevation of a four-way mulled window assembly. [9]
Figure 7 Inset image: Building elevation with a four-way mulled window assembly. Right: Typical elevation of a four-way mulled window assembly.

This project shows the importance of weighing the pros and cons of fenestration decisions. Including an earlier mockup testing phase, investing in improving the rough opening tolerance issues, or investing in the additional up-front cost of installing a curtain wall may have helped avoid this situation by eliminating the problematic sub-framing. However, budgetary and schedule constraints often preclude these approaches. In such cases, communicating these risks and putting additional effort into the shop drawing phase are key to avoiding installation issues.

Case study: Structural failure

One of the author’s projects involved a mixed-use residential wood-framed building targeting a PG-50 rating for its punched vinyl windows in 2.4 x 2.6 m (8 x 8.5 ft) wood framed openings (Figure 7). On this project, the author provided peer review building envelope consulting services to the architect, as well as field testing services to the owner.

The contract documents only required that each window carry a PG-50 rating and AAMA certification, but did not include performance requirements for the sub-framing. The contract documents required field testing though and included details for a flange window system with perimeter flashings, including a membrane-wrapped rated sill pan with upturned legs. A window manufacturer claimed during bidding they could meet all specification and detailing requirements.

Observed deflections during structural testing.[10]
Figure 8 (Top): Observed deflections during structural testing.

During the submittal phase, the window manufacturer submitted test reports for a double hung window with a fixed transom configuration, which did not exist on the project. Instead, most of the configurations in the design required four-way mulled fixed lites and casements (Figure 6, page 5). The window manufacturer could not produce any performance test reports for a four-way mulled configuration of the size necessary on the project, so the project team pursued lab performance testing for the product as per the specified project air, water, and structural loading.

During testing, excessive deflections of approximately 25 mm (1 in.) elastic and 13 mm (0.5 in.) inelastic permanent set occurred during the uniform structural load tests (Figure 8, left). During structural testing, the excessive deflections caused the casement sash to disengage from the perimeter frame and damage operating hardware (Figure 8).

Following testing, the manufacturer disassembled the windows and found the reinforcement welds were fractured at the four-way mull joint. They also observed excess grinding of the welds occurred to fit the reinforcing within the mull joint. After further testing, the window achieved the minimum requirements; however, excessive deflections continued to allow gaskets and window joints to open up under loading, resulting in water leakage.

 Casement operable sash disengagement.[11]
Figure 8 (Bottom) Casement operable sash disengagement.

The above issues presented a major weak point in the system. Luckily, the observed issues occurred in the lab well ahead of window fabrication for
the project and the manufacturer was able to implement additional QC measures during fabrication of the mull joints, limiting schedule delays and added cost.

Conclusion

Understanding the effect mulled systems and their sub-framing can have on air, water, and structural performance is an important skill for a successful project. The complexities of fenestration design are plentiful, however, having a standardized method of identifying, specifying, and reviewing these products from design through construction gives architects a powerful tool for tailoring their design to meet the project’s needs and avert costly delays and changes during construction.

Throughout a project, it is critical to understand the pros and cons of mulled fenestration and know when to implement strategies such as avoidance, AAMA 450 performance specifications, or accommodation as a means of addressing these complex fenestration systems.

Notes

1 See the North American Fenestration Standard, AAMA/WDMA/CSA 101/I.S. 2/A440-17 (FGIA, WDMA, and CSA Group, 2017).

2 Learn more at Performance Rating Method for Mulled Combination Assemblies, Composite Units, and Other Mulled Fenestration Systems, AAMA 450-20 (Schaumburg, IL: Fenestration and Glazing Industry Alliance, 2020).

3 Read the article, “Window Walls: Blurring the Lines Between Glazing Products.[12]” The Construction Specifier, May 2017.

4 See note 1.

5 See the article, “Window Receptor Frames: What You Need to Know.” RCI (IIBEC), 2007.

6 Read more about window receptor frames, “Window Receptor Frames in 2018: More than Just Accessories.” RCI (IIBEC), 2018.

Endnotes:
  1. [Image]: https://www.constructionspecifier.com/wp-content/uploads/2023/06/TITLE-PHOTO-1.jpg
  2. [Image]: https://www.constructionspecifier.com/wp-content/uploads/2023/06/FIGURE-1.jpg
  3. [Image]: https://www.constructionspecifier.com/wp-content/uploads/2023/06/Figure-2.jpg
  4. [Image]: https://www.constructionspecifier.com/wp-content/uploads/2023/06/Figure-3-TOP.jpg
  5. [Image]: https://www.constructionspecifier.com/wp-content/uploads/2023/06/Figure-3-BOTTOM.jpg
  6. [Image]: https://www.constructionspecifier.com/wp-content/uploads/2023/06/Figure-4.jpg
  7. [Image]: https://www.constructionspecifier.com/wp-content/uploads/2023/06/Figure-5.jpg
  8. [Image]: https://www.constructionspecifier.com/wp-content/uploads/2023/06/Figure-6-RIGHT.jpg
  9. [Image]: https://www.constructionspecifier.com/wp-content/uploads/2023/06/Figure-7.jpg
  10. [Image]: https://www.constructionspecifier.com/wp-content/uploads/2023/06/Figure-8-LEFT.jpg
  11. [Image]: https://www.constructionspecifier.com/wp-content/uploads/2023/06/Figure-8-RIGHT.jpg
  12. Window Walls: Blurring the Lines Between Glazing Products.: https://www.constructionspecifier.com/window-walls-blurring-line-glazing-products/

Source URL: https://www.constructionspecifier.com/ensuring-performance-in-mulled-window-systems/