Behind the failure of many engineered structural elements lies the very instrument designed to hold them together. Engineered wood products are for the most part composite structural elements that rely on two or more separate materials to achieve strength and ultimately carry a design load. Many fire service professionals recognize gusset plates in the field as indicative of such an engineered or, lightweight structural element. Correctly identified as metal plate connectors, “gusset plates” enable the joining of wood products in a fast, cost-effective way. They have been adopted as one of the main avenues to design and connect engineered systems such as parallel chord wood trusses and roof trusses. The failure of the metal plate connector or “gusset plate” means the compromise of that individual structural element. A recent fire in a private dwelling was revisited and the assessment was truly astonishing. It opened up some of our firefighters’ eyes and provided great insight into what was really going on during the fire that was unseen at the time. The photos provided will show you the dangers of these connections in a wood truss system, and the lessons learned will provide a glimpse into experience-based observations instead of relying on test methods within a controlled environment. Let’s take a look…

A metal plate connector or “gusset plate” hails originally from Fort Lauderdale, Florida, in 1958. Originally dubbed “grip plate for truss,” they are metal plates made of light gauge steel that are hydraulically pressed on one side, forming small spikes or teeth on the other side (figures 1 and 2). The teeth of the metal plate connector are about 1/4-3/8 inch. The plates are secured by hand at first to truss members, and are then passed on to be secured by a hydraulic press. There is an allowance of a 1/8-inch max space between web and chord members.

Experience has shown us that the weakest part of a wood truss is the connection or panel point. This is where web members and chords meet by, you guessed it, the connection via a metal plate connector. As previously stated, they penetrate the wood fibers 1/4-3/8 inch (figure 3). This connection is limited to the premise that there is no mishandling of the truss during transport and installation. Should a truss undergo a significant impact to the panel point and the penetrations are dislodged, they cannot be appropriately set back in. It has been observed on one occasion that the metal plate connector that was dislodged was re-secured with a ½-inch steel staple. Metal plate connectors are made from galvanized light gauge cold formed steel. This type of steel experiences compromise at 800°. 

Following a recent fire, we looked at the conditions of a parallel chord wood truss floor system in the daylight. It was during this evaluation of structural integrity that we could draw certain conclusions based on real events. The metal plate connectors were the weakest link in the truss system (figure 4). Littered throughout the debris were hundreds of metal plate connectors, almost if it had been raining gusset plates. Fire damage to the trusses were significant. Charred trusses released the bottom chord once the metal plate connector had dislodged and fell. The remaining metal plate connectors did not stay buried into the wood; they pulled away during the shifting of the floor system after multiple trusses had collapsed. Contrary to may opinions (including this author’s own, in the past), the metal plate connectors did not peel back, warp, or bend. Once the wood surrounding and embedded into the plate connector burned, the plate fell off in one straight piece, releasing its members and precipitating other plate failures. The rust on the galvanized plate indicates to us that it was superheated, ultimately deglazing the zinc coating. This is evidence to us of its conductive heat transmission into the wood. This supports the narrative of induced pyrolytic decomposition of the wood fibers of the truss from the metal plate connector’s teeth. Because the teeth of the metal plate connector were superheated and therefore conductively transmitted heat into the penetration, the teeth loosened and ultimately allowed the plate to fall. This disputes the theory that heat is reflected off of the plate, therefore benefiting the panel point.

Structural Collapse Under Fire Conditions

As you can see in the photos provided, the top chord of the truss remained secured to the bottom of the second floor decking due it its nailing from above. You can see a significant 10-12-inch deflection of the floor as well, with no structural integrity underneath (figure 5). This also proves that, due to massive tensile forces upon the bottom chord of the truss via fire damage, the metal plate connectors could not support the compression from above, therefore releasing the plate, as evidenced by teeth bent all in one direction.

The trusses were found only10 minutes into the fire. When this was communicated, members evacuated the second floor and opened a wall between the first and second floors and extinguished fire in the truss loft on one side of the house. What’s astonishing is that the postfire assessment showed this allowed the truss members to retain integrity, with little damage to the system, however once the heated metal plate connectors were quenched, there were a number that pulled from their connection.

Luckily, members were not on the second floor for long. Should the loading of multiple firefighters have continued, including the live loading of water, it is obvious the floor would have collapsed completely. The trusses were free burning throughout the truss loft, and the first floor at times only had a haze. This is very dangerous. The gypsum board prevented firefighters from seeing that structural compromise was significant. It is truly a lesson learned.

There have been lab tests surrounding metal plate connectors, including the NISTIR 7393. In opinion of this author, this test is not conducive to your health. The test uses two individual one-foot sections of lumber connected via metal plate connectors. The test does not induce heating by fire, rather a radiant heat panel—and oh yea, only on one side (trusses don’t only burn on one side). Also, the disclaimer reads: “Load carrying ability of the metal plate truss connections was not measured during these tests.” That’s a big deal. Metal plate connectors experience loading during a fire, and the tensile forces are pulling the members apart. This is why the plates remained connected to the wood in this test, and not in actual fires. The test did not replicate compartmental heat, movement due to fire damage, impact loads, concentrated loads, or tensile forces.

There are many factors that ultimately lead to the collapse of structural elements. Connections within engineered wood materials are the precipitating factor to the majority of such instances. Will a truss always collapse in five-10 minutes? No. But that’s not always what we are worried about. We should be concerned with the weakening of materials and connections ,such as with the metal plate connector, that can mask a compromised structural system such as a truss floor, where the only precipitating factor needed to complete the collapse process is….us. (Figure 6)

SALVATORE ANCONA is a deputy chief fire instructor at the Nassau County (NY) Fire Service Academy; a member of the Seaford (NY) Fire Department; a former captain and training officer for the Bellmore (NY) Fire Department; and a paramedic supervisor in Queens, New York. He has an A.S. degree in fire science from Nassau Community College and has a bachelor’s degree in fire and emergency services administration from John Jay College of Criminal justice. Ancona is the author of the building construction page “The Sons of Brannigan” on Facebook and was a recipient of the 2019 FDIC International Honeywell DuPont Scholarship.

Combustible Construction: Podium Wood-Frame Private Dwellings

Veneer vs. Façade: What’s the Difference?

Attic vs. Cockloft: What’s the Difference?

Firefighters and Construction: Beams and Girders: How They Fail