Nightmares on the Edge: Lawsuits About Concrete Slab Tolerance and Deflection Related to Windows and Curtainwalls Mark Meshulam, REWC Façade Consultants 4244 Kayla Lane, Northbrook, IL 60062 (847) 878-8922 | mark@chicagowindowexpert.com No v e m b e r 9 , 1 0 , 1 2 , 1 6 , 1 7 , 1 9 , 2 0 2 0 | 2 0 2 0 I I BE C B u i l d in g E nc l o s u r e S y m p o s i u m Me s h u l a m | 7 1 ABSTRACT SPEAKER Mark Meshulam, REWC Façade Consultants | Northbrook, IL Mark Meshulam has over 40 years of construction experience. Starting as a sales engineer, he eventually became an owner of one of the top commercial window companies in Chicago. He oversaw engineering, shop drawings, laboratory and field testing, and project management for hundreds of buildings. In the past decade, he has offered clients the benefit of his experience as a consultant and expert witness. He works nationally and internationally— sometimes in high-profile legal cases, including crimes where glass is important evidence. His websites attract thousands of viewers who educate themselves by reading his 80 original articles. The author has been involved as an expert witness in multiple litigations where confusion, poor language, and competing demands have led to million-dollar lawsuits. The negative domino effect starts with specification language, may be affected by value engineering, continues through contract language, develops further in shop drawings, and culminates in a train wreck in the field. The train wreck later falls off a cliff when the slabs deflect over the long term in unanticipated ways. The blame is often shared by multiple parties, including owners, architects, structural engineers, general contractors, façade manufacturers, and façade contractors. Presenting such a complex case to a judge or jury is challenging. After all, if the construction professionals can’t figure it out, how can non-construction judges and juries get it right? In this presentation, the author will focus on what went wrong in the construction process, and he will offer solutions that can bring about positive shared expectations of all parties and ultimate project success. 7 2 | Me s h u l a m 2 0 2 0 I I BE C B u i l d in g E nc l o s u r e S y m p o s i u m | No v e m b e r 9 , 1 0 , 1 2 , 1 6 , 1 7 , 1 9 , 2 0 2 0 No v e m b e r 9 , 1 0 , 1 2 , 1 6 , 1 7 , 1 9 , 2 0 2 0 | 2 0 2 0 I I BE C B u i l d in g E nc l o s u r e S y m p o s i u m M e s h u l a m | 7 3 This article is intended as a cautionary tale to parties who design building façades or the structures that support them. Such parties include architects, structural engineers (SEs), general contractors (GCs), concrete formwork contractors (CFCs), and façade contractors (FCs). When completed, the buildings produced by these teams look simple and beautiful. Because of this, they belie the hard work, and sometimes grief, that takes place on the long road to façade completion. One of the root causes of the pain and suffering endured by such construction groups is, in this author’s opinion, a lack of common understanding and language that sometimes never gets resolved. Or, when misunderstandings are finally resolved, the resolution happens in front of a judge or arbitration panel, with high legal costs incurred by most participants. The author has been involved in three litigations as an expert witness, wherein poor understanding and communication about concrete slab deflection and tolerance resulted in excessive costs and delays to the point where litigation was the ultimate result. The buildings are (see Figure 1): • Building A: 21-story residential high rise in Oak Park, IL • Building B: 48-story residential high rise in Chicago, IL • Building C: 80-story residential high rise in Chicago, IL BUILDING A The author started out as a consultant to the GC just after an FC was selected. In the very first meeting, the question of slab tolerance and deflection was discussed, since it was not fully addressed in the contract documents. The FC was cautioned to carefully craft a request for information (RFI) with specific language. However, instead they issued a sloppy RFI that returned an irrelevant answer from the SE. This started a months-long string of miscommunications that saw delays, the FC demanding more money for new extrusion dies that were not needed, the FC generally failing to perform, and their subcontract being terminated. The GC engaged a new FC at much higher cost, then sued the original FC for delays and increased costs. The original FC in turn countersued the GC. The case went to trial in a “bench trial,” wherein there is a judge but no jury. The judge misunderstood certain key premises and ruled generally in favor of the original FC; however, no monetary damages were assessed. Thus, both sides expended much in time and legal costs, and both basically went nowhere. So far, an appeal has not been initiated and appears unlikely. BUILDING B The author was engaged as an expert witness on behalf of the FC after the building was completed. Although there were many issues of disagreement, the issue of slab tolerance and deflection took an inordinate amount of attention in the researching and structuring of arguments which were eventually presented before an arbitration panel of four construction attorneys. The panel found generally in favor of the FC with significant financial benefit awarded to the FC. BUILDING C The author became involved after the original FC’s subcontract was terminated following disagreements about anticipated concrete movement. In this case, the concrete deflection criteria were made more Nightmares on the Edge: Lawsuits About Concrete Slab Tolerance and Deflection Related to Windows and Curtainwalls Figure 1 – Elevations of buildings A, B, and C. stringent by the owner’s consultant, as the project was already in progress, but the GC resisted the FC’s attempts to negotiate with them or assess additional charges for costs to meet the new criteria. The increased stringency was a late addition of seismic requirements that are inappropriate for Chicago, which is not in a seismic zone. The FC had fabricated much product for the project in anticipation of schedule pressures once the submittals were approved and was therefore left with substantial cost when the GC terminated his contract. The FC is currently in litigation with the GC. It is believed that the second FC also had problems meeting the seismic requirement, and that for the second FC, the seismic requirements were loosened. WHY ALL THE TROUBLE? In the author’s experience, the problems came from two main sources. First, the issues are complex. They involve understanding three-dimensional movements and dimensions that might change over a period of time. The second problem is in the communication of such issues in a way that professionals from different disciplines can commonly understand. CONCRETE TOLERANCES A CFC places formwork to support concrete slabs that will eventually support a window wall or curtainwall. But the concrete won’t be perfect. It may be too high, too low, projecting out too far or too little (in/out). The amount of inaccuracy that is allowed is called the tolerance. Concrete tolerances are defined in ACI 117; however, the tolerances defined therein can be too generous for the capabilities of the façade system and therefore may need to be negotiated between the GC and the CFC. An experienced GC will know this and will incorporate tighter tolerances in the CFC’s subcontract. Other GCs learn this too late and are stuck with somehow resolving competing needs of different subcontractors. Complicating this and muddying the outcome are situations in which the GC self-performs the concrete. In this case, the FC may be battling the GC at a distinct disadvantage. See Figures 2 and 3. CONCRETE MOVEMENTS Altough concrete appears to be an immoveable object, it can move over time. When concrete does move, it does so with great force, so whatever is attached to it must either accommodate the movement, suffer failure, or both. Concrete slabs can sag soon after the reshoring is removed (“instantaneous deflection”), it can sag over the long haul under its own weight (“dead-load deflection”), it can sag under the weight of objects and people (“live-load deflection”), it can sag over the long term (“long-term creep”), and it can simply travel lower as the columns shrink (“column foreshortening”). The key to making façades work well with concrete is to have the SE clearly predict the concrete movements, and to have these movements designed into the façade system. But the SE can’t answer clearly if the question is not posed to him or her clearly. TIMING! The most persistent miscommunication between parties seems to be: when will these concrete movements happen? The lucky part about all the sagging that the concrete is doing is that much of it will have already occurred when the façade is installed. Therefore, the key to asking the SE about slab deflections is to first define the point at which the façade will be installed, then ask the question in two parts: How much deflection is expected before the façade is installed? How much after the façade is installed? This simple point was a major misunderstanding of the original FC in Building A, and despite much effort on the part of the author, it remained poorly misunderstood even at the time of the bench trial. In order to give the SE the proper timeframe, one must first ask the GC: how many weeks after the reshoring is removed will the façade be installed? SEISMIC MOVEMENT Earthquakes can shake a concrete structure, and the façade that is attached to it, in ways that are different from all the downward deflections previously discussed. When an earthquake shakes a structure, it racks the grid created by structure and façade. It momentarily and violently turns rectangles into parallelograms, then does so rapidly in the opposite direction, while twisting flat objects out of plane. Obviously, this imposes more stress on the façade, and it requires that the façade is more flexible and forgiving of such abuse. That being said, Chicago is not in a seismic region, yet the “deal breaker” at Building 74 | Meshulam 2020 IIBEbeC BuildINg ENClosure Symposium | November 9, 10, 12, 16, 17, 19, 2020 Figure 2 – Excerpt from ACI 117-10 regarding slab tolerances. Figure 3 – Allowable tolerances in poured slab location. C was the introduction of seismic requirements mid-stream in the project. The FC’s system wasn’t designed for seismic movements. CAMBERS Sometimes an SE will design a concrete slab with a slight upward arch (camber) in order to counteract the effects of gravity and therefore reduce sagging. The FC will not know, without asking whether a camber is present at the slab edge, how high it might start out being and, most importantly, where it will be at the time of façade installation. POST-TENSIONED (PT) CONCRETE Concrete slabs are commonly fitted with embedded steel cables (“posttensioned cables”) that are stretched and tightened after the concrete is formed. This places the concrete in a permanent state of longitudinal compression and makes the concrete stronger with regard to resisting downward deflections. That’s the good news. The bad news is that if a PT cable is accidentally cut or drilled into, it can snap with explosive force and do significant damage to itself and whomever is nearby. In order to avoid this catastrophe, much effort goes into laying out façade attachments so that they miss the PT cables, or special anchors are embedded into the concrete during forming. In the experience of the author, the parties have been doing a good job of coordinating façade attachments with PT cables without excessive conflict other than occasional embedment placement errors. WINDOW WALLS Window walls are runs of windows that typically go from floor to ceiling. The bottom of a window wall is mounted to the top of the slab; and the top of the window wall is mounted to the bottom of the slab above it. Often a slab cover—a C-shaped length of extruded or formed aluminum—is used to conceal the slab edge and create a quasicurtainwall look. This façade assembly would then be called a “window wall with slab covers.” See Figure 4. Window walls employ a continuous track at the top and bottom of the opening. When this track is set, the glazed panels are placed into the tracks and locked together at their jambs. The top track is called the “head receptor” and the bottom track is called the “subsill.” Together, they are considered to comprise a “subframe.” The head receptor typically allows the top of the glazed panel to slide up and down inside the head receptor. This is where all slab movements that occur after the window wall is installed are absorbed. The subsill bears the weight of the panels, so no movement will occur between subsill and panel. Any and all concrete tolerances and concrete movements that occur before the window wall is installed are accommodated in the adjustable space between the subframes and the concrete, sometimes thought of as the “shim space” or the “caulk-joint space.” The space between the subframes is created by the installers who are tasked with installing straight and level subframes despite the fact that the concrete is not straight and level. Installers No v e m b e r 9 , 1 0 , 1 2 , 1 6 , 1 7 , 1 9 , 2 0 2 0 | 2 0 2 0 I I BE C B u i l d in g E nc l o s u r e S y m p o s i u m M e s h u l a m | 7 5 Any and all concrete tolerances and concrete movements that occur before the window wall is installed are accommodated in the adjustable space between the subframes and the concrete, sometimes thought of as the “shim space” or the “caulk-joint space.” Figure 4 – Window wall with slab cover. use hard plastic shims to level out the subframes at their proper height before fastening them in place, then fill the resultant gap with a flexible caulk—usually silicone based. See Figures 5, 6, and 7. CURTAINWALLS A curtainwall, on the other hand, runs outboard of the slab in a continuous plane. Curtainwalls attach to the edge of the slab at one connection per each vertical mullion. In recent years, the curtainwall industry has gone to “unitized” design, in which fully assembled and glazed panels are hung on specialized anchors and stacked onto one another like a checkerboard. This type of curtainwall will be considered in this paper. Generally, the curtainwall anchoring systems are tolerant and adjustable. Current curtainwall anchor designs include serrated washers that lock a sliding anchor plate in the proper in/out position, while jacking screws on the sides of the mullions allow for precise up/down adjustment. When concrete movements occur after curtainwall installation, the movement is taken at the horizontal joint between panels. This joint allows up-and-down sliding. Typically, the top of the lower panel has a gasketed up-leg called a “chicken head” that engages into a slot in the bottom of the upper panel. The chicken head defines the front-to-back plane of water management. See Figures 8, 9, and 10. WHICH SYSTEM GETS INTO MORE TROUBLE? All three of the buildings discussed in this paper utilized window walls, not curtainwalls. This may be because there are more window-wall jobs built, or it could be that curtainwall jobs might have higher skill levels in the team. Unitized curtainwalls typically cost more, but it could be that their legal costs are lower. CALCULATING CAULK JOINT SIZES It may seem strange to think of calculating caulk joint sizes, and stranger still that such joints must be “engineered” for long-term performance, but that is exactly what is needed. Most well-designed caulk 76 | Meshulam 2020 IIBEbeC BuildINg ENClosure Symposium | November 9, 10, 12, 16, 17, 19, 2020 Figure 5 – Window wall systems accommodate up/down slab tolerances by varying the shim height and caulk-joint size. This is done at the time of window installation. Figure 6 – Window wall systems accommodate slab movement by varying the engagement of the window head into the head receptor. This occurs after installation, during the life of the façade. Figure 7 – Summary of how and when slab tolerances and movements are accommodated in window walls. Concrete Tolerances and Movement at Window Walls Where accommodated? When accommodated? Slab tolerance Shim space and caulk joint At time of façade installation Slab movement Engagement at head receptor After installation, during life of façade joints should be hourglass shaped in cross section for maximum adhesive surface and flexibility. The calculations come into play when the maximum expansion, contraction, and shear capabilities of the caulk become known. Let’s say a particular sealant (caulk) has a +/- 50% movement capability. It can be stretched up to 50% of its freely moving dimension, compressed up to 50% of that dimension, and also placed in shear by that same amount. Beyond that amount, that caulk will rip apart (which is known as experiencing a cohesive failure) and costly leaks will ensue. See Figures 11 and 12. DISASTROUS COMMUNICATIONS With the background provided above, it is hoped that the reader will be able to discern problems and shortcomings in actual project communications that led to litigation. November 9, 10, 12, 16, 17, 19, 2020 | 2020 IIBEbeC BuildINg ENClosure Symposium M Meshulam | 77 Figure 8 – Unitized curtainwalls use specialized anchors to manage concrete tolerances and stacking horizontal joints using a “chicken head” to manage up/down movement. Concrete Tolerances and Movement at Curtainwalls Where accommodated? When accommodated? Slab tolerance Adjustable anchoring system At time of façade installation Slab movement At horizontal stack joint After installation, during life of façade utilizing a chicken head detail Example: Calculating caulk joint size for concrete tolerances (simplified) A Concrete tolerance (provided by GC) +/- 1/4” B Estimated worst-case (typically at middle of span) deflection prior -1/4” to window installation (provided by SE) C Total worst-case downward position of slab at time façade is installed (A + B) -1/2” D Minimum caulk joint size allowed (per sealant manufacturer) 1/4” E Minimum designed caulk joint size (-C + D) 3/4” Figure 11 – Simplified example of caulk joint size calculation for window wall. Figure 9 – Adjustable anchoring system of unitized curtainwall. Figure 10 – Summary of how and when slab tolerances and movements are accommodated in window walls. Example: Calculating thermal movement When thermal movements are calculated, thermal movement of the concrete is generally ignored, and the focus is on the aluminum framing. A Coefficient of linear thermal expansion of aluminum 13 x 10^-6 inch per inch per degree F = .000013 in/in/°F B The longest aluminum part is the vertical mullion; say it is 10’ long 120” C Thermal operating range 0°F to 150°F = 150°F D Total thermal expansion (A x B x C) .234” E Assuming the mullion is cut and installed at the middle +/-.117” of the thermal range, the expansion is: (D / 2) F This will generally be rounded to: +/-1/8” Figure 12 – An example of the calculation of thermal movement in aluminum mullions. Disastrous Communication – Building A The FC’s RFI may be viewed in Figure 13. The SE’s response is shown in Figure 14. This led to months of back and forth wherein the FC, because he failed to clearly ask the question, concluded that +/-7/8 in. of movement was needed at the window head, and then somehow determined that he needed to redesign 28 new extrusion dies in order to accomplish that. When the new, more experienced FC came on board, they immediately issued the RFI shown in Figure 15, which went to the heart of the issue of deflection over time, and also clarified the location of the area of interest. Now that the RFI was in language the SE could understand, the SE returned the interesting and critical response shown in Figure 16. This information, the result of asking the right question, clearly told the parties that: Of the total slab deflection of .75 in., .30 in. of it would already have occurred when the façade was installed. Therefore, the façade would only experience a slab deflection of .45 in. (.75-.3). This deflection could have been easily accommodated by the first FC’s system without modification. Disastrous Communication – Building B This became an issue on the project due to the window wall not fitting at the balconies. The window wall met the contractual requirement for construction tolerance; however the concrete was poured out of tolerance, which caused the windows not to fit. See the subcontractor’s language in Figure 17. The movement requirement of +1/2 in. and -1 in. was met by the FC, and this was undisputed. However, the windows did not fit at the balconies. Interestingly, there were no conflicts at the adjacent curtainwalls. The balcony openings were tight. The GC, FC, and installer then engaged in four to five months of back and forth wherein the FC was blamed for making the windows too tall, and the SE was involved to see if some movement capability could be reduced in order to compress the window system and fit the windows into the opening. This would be “robbing movement to pay for tolerance.” See Figure 18. Ultimately, the installer surveyed the slabs before window installation and reported to the GC that numerous areas of concrete grinding were needed. The pattern of grinding was consistent from floor to floor, but as the upper floors were poured, the concrete height error was corrected so that by the 29th floor and above, no grinding was required. Grinding costs money and causes delays, so the GC, who self- performed the concrete, saw fit to push 78 | Meshulam 2020 IIBEbeC BuildINg ENClosure Symposium | November 9, 10, 12, 16, 17, 19, 2020 Re: Slab Edge Deflection – Please provide the engineer’s anticipated floor slab deflection at 60 days after pouring and fully loaded. At the worst-case slab-edge conditions, where windows are above and below the slab edge, please supply the following anticipated dimensions: Total deflection from designated elevation after 2 months: Total deflection from designated elevation after 3 months: Total deflection from designated elevation after 4 months: Total deflection from designated elevation after 10 years: 2.16. Window-wall system head receptor designed to allow for a plus 1/2 in. or minus 1 in. of deflection in addition to a 1/2 in. plus or minus constrution tolerance. Additionally, the concrete slab edge will have a +/- 1 in. in/out tolerance. The following slab deflections are based on code-required loading and theoretical slab behavior; they do not take into account construction tolerances which are allowed by the ACI code. There is no discernible difference between the slab deflection at 2 months vs. at time 4 months. The worst-case slab deflection at the exterior of the building at the typical level slab are as follows: Deflection at time 2 months = ˜0.30 inches Deflection at time 3 months = ˜0.30 inches Deflection at time 4 months = ˜0.30 inches Deflection at time 10 years = 0.75 inches ADDITIONAL INFORMATION IS REQUIRED IN ORDER TO ACCOMODATE THE REQUEST AND PROVIDE THE THEORETICAL DEFLECTIONS. WHAT FLOOR? WHAT LOCATION ON THE FLOOR? WORST CASE OR TYPICAL? EXTERIOR OF THE FLOOR AT THE PERIMETER OR INTERIOR OF THE FLOOR IN THE CENTER OF THE BAY? HAS THE FACADE LOAD BEEN PLACED ON THE SLAB AT 60 DAYS? IF SO, WHEN? PLEASE CLARIFY. Figure 14 – SE’s answer to first RFI from original FC of Building A asked for specificity regarding location and timeframe. Figure 13 – First RFI from original FC of Building A was lacking in specificity regarding location and timeframe. Figure 15 – First RFI from second FC of Building A was clear and specific regarding location and timeframe. Figure 17 – Building B subcontractor’s language addressing slab tolerance and deflection. Figure 16 – SE’s answer to second FC’s RFI is clear and has actionable information. back and charge the FC for the cost of the delays, including supervision and overtime in excess of $1 million. The charge was not allowed by the arbitrators based upon the GC’s failure to produce a critical path analysis of the delay. It was also telling that many areas were “OK” and did not need concrete grinding, so obviously the windows were not too tall. Disastrous Communication – Building C This story starts with particularly onerous subcontract language (see Figure 19). Then the fun begins. The original concrete movement requirements started off as stringent, but the FC agreed to the requirements. However, as the job progressed, various parties in the design team added new, even more stringent requirements mid-stream that the FC’s system was not designed for. Under pressure from the GC, who unreasonably wielded the contract language in the “Changes” section, the FC resisted but eventually buckled under and allowed the performance mockup (PMU) of the system to be tested at the higher performance levels; the system failed, as the FC had predicted, at the areas of greater movement. The GC and design team refused to allow the FC to test the system at the originally specified performance levels, nor according to the test protocols defined in the specification. The situation became heated, the GC terminated the FC’s subcontract, and made a claim on the FC’s bond. Against the FC’s wishes, the surety paid a large sum to the GC, which the FC now owes to the surety under an indemnity agreement. The replacement FC’s PMU failed numerous tests for similar reasons, but the GC and design team gave the replacement FC much leniency and forgiveness in the acceptance of PMU results and even allowed the second FC to install windows before the PMU was successful. See Figure 20. SUMMARY AND CONCLUSION Conflicts and disputes can easily occur when the parties don’t understand the needs of one another and can’t find a common language to express it. Such conflicts can be eased and mitigated when a common understanding and language for expressing slab tolerance and movement is developed, and particularly when it is coordinated with applicable timeframes. November 9, 10, 12, 16, 17, 19, 2020 | 2020 IIBEbeC BuildINg ENClosure Symposium M Meshulam | 79 Figure 18 – Installer’s survey of floor slab heights at Building B. Figure 19 – Onerous language regarding changes that the GC used to force the FC to allow movement changes to his subcontract. Figure 20 – A history of movement-related scope changes in Building C.
