Wood Floor Behavior: Experimental Study

A series of 13 wood floors was tested under uniformly distributed load. The floors were manufactured with 12 (nominal) $28\mathrm{hemlock}$ joists, “No. 2 and better” grade, 5/8 in. (15.8 mm) Douglas‐fir plywood, and 8d common nails at 6 in. (152 mm) o.c. In two floors the nail spacing was increased to 12 in. (305 mm), and in one floor, 5/8 in. (15.8 mm) waferboard replaced plywood as the sheathing. Experimental results are presented for floor deflections, first joist failure load, and subsequent failure loads. Test results were compared with theoretical predictions using the computer program FAP. It is concluded that, although first joist failure should be used in reliability analysis, the load carrying capacity of the floor assembly is normally higher and several more joist failures usually occur before total collapse. The variability in floor strength is substantially lower than the variability in bending strength of individual joists. The tests showed that floor behavior was nearly linear, and the results compared well with FAP predictions. In particular, the required nail stiffness may be obtained from a simple nailed joint test in compression. It was also concluded that the weakest joists are not generally the ones that fail first in a floor, and that reliability predictions of floor strength can only be obtained when the bending strength is known for the weaker as well as the stronger joists in the population.