People focus too much on materials or individual members in structure, and not enough on systems. You can’t understand modern structures – by which I mean ones that follow the cutting edge of thought since the 1880s – without thinking about them in terms of structural systems.
The picture above shows the underside of a typical floor at our hospital conversion project in Queens. Pretty much everything you’re seeing in that picture is terra cotta, but it’s not a tile-arch floor. It’s a concrete two-way rib system…what’s usually called a concrete waffle. The underside of the floor is terra cotta, as you see here, but the interesting question is why.
Every “fireproof” floor used in the US from the beginnings of wrought-iron beams in the 1850s, through the development of steel- and concrete-frame buildings between 1880 and 1910, through the use of almost-modern versions of those frames in the 1940s and early 50s required wood formwork of some kind as temporary support during construction. The different floors – tile arch, concrete draped-mesh, reinforced concrete, and the many oddball types – all needed support until they were complete and could carry load on their own. The rapid spread of Q-deck (cellular, non-composite light-gage steel deck) and then composite steel deck in the mid and late 1950s was sped by the cost savings in eliminating all the carpentry of temporary supports. So when we look at any pre-metal-deck floor now, we have to see it from the viewpoint of how it was built.
The minimum structural requirements of the floor were that it (a) span between the steel beams of the building frame, and (b) meet the code requirement of a 3-hour fire rating. There were other requirements: (c) provide attachment for plaster ceiling finishes (more on this below), (d) minimize temporary work other than the wood formwork, (e) minimize weight to reduce the load on the frame and foundations, and (f) provide easy access during construction for all the piping required in a hospital. The system that was used is a concrete waffle built in the following manner: terra cotta blocks, 16 inches square) were laid on top of the wood formwork in a grid pattern, with 4 inches between any two blocks. Thin tiles of terra cotta were laid between the blocks, in the strips that would become the ribs. Rebar was placed between the blocks and above the tiles as reinforcing for the ribs, then concrete placed for the ribs. After that, a very thin, wire-reinforced slab was cast over the top of the blocks and ribs, and then the formwork was removed. This system is structurally as deep as the ribs but weighs about a third as much as a solid slab that depth.
At that point, the floor was structurally complete, and plaster should have been applied to the terra cotta underside. The corrugations on the terra cotta surface are there specifically to provide anchorage for plaster keys, to hold the plaster stable. (The square undersides of the blocks and the rectangular undersides of the tiles are clearly visible above.) But it wasn’t, and a hung ceiling was installed instead. Two possibilities come to mind: either (1) at some time between the original design and the beginning of interior fit-out work, the decision was made to use a hung plaster ceiling rather than direct-applied plaster or (2) the tiles were put in the forms even though there was never any intention to put plaster directly on the floor underside. Option 2 is not as silly as it sounds: the fire-rating was based on the entire floor system, so it may have been drilled into people’s heads that the system couldn’t be modified. Option 1 is the kind of mid-course change made all the time.
So to understand what I’m seeing here, I have to step back from concrete analysis and look at how the floor was built and the issue of plaster application. There are similar non-design issues when we examine any piece of an existing building, which is why I insist we can only really understand them by thinking in terms of systems. This is a floor system, not just a concrete waffle.