As houses get tighter and energy codes stricter, it has become increasingly common for builders to apply one or more layers of rigid foam to the outside of wood-framed shells, then install horizontal siding over vertical strapping. The method adds R-value to the walls, cuts thermal conduction through the framing, and when done carefully, creates a flashed drainage space for any water that makes it through the cladding layer. The technique's been covered in several JLC articles over the past few years; after reading "Building a High-Performance Shell" (5/10), my colleagues and I decided to test the strength of the connection between the strapping and the framing - especially given the fact that the fastener is being driven through several inches of foam, which acts essentially the same as an air gap.
Test Procedure
Using our hydraulic testing equipment, we set up an experiment to measure the real performance of the type of assembly seen in the JLC article. Our test specimens consisted of a No. 2 Douglas fir 2x6 stud, 7„16-inch OSB attached to the stud with three 10d nails (.148-inch diameter), two layers of 2-inch foil-faced polyisocyanurate foam insulation glued to each other and to the OSB with construction adhesive, and 1x3 No. 2 hem-fir strapping secured through the foam with two 7-inch-long FastenMaster HeadLok screws spaced 16 inches on-center. We predrilled the screw holes to avoid splitting the strapping, and aligned the foam and strapping layers so that we could observe the displacement as the hydraulic ram pushed down on top of the strapping.
We loaded each sample assembly to determine its maximum load-carrying capacity, and used a dial indicator at the base of the strapping to measure displacement while capturing the progress with a digital video camera. Matching the data from the computer-controlled hydraulic ram and the real-time video log, we were able to accurately determine the "yield" point of the fasteners - the load at which the steel develops a permanent bend. As the table below shows, the yield loads are much smaller than the maximum loads at failure.
Displacement of the strapping against the dial indicator was typically greater than 2 inches at the maximum load capacity. The displacement at the yield point, however, was quite small - a few hundredths of an inch.
Analyzing the Results
Next, we compared our test results against "predicted," or calculated, results based on formulas found in Technical Report 12 (TR12) from the American Forest & Paper Association, which is used to design wood structural connections that rely on steel fasteners. Our average tested yield value per fastener was 46 pounds. Calculations using TR12 methods, which allowed us to account for the 4-inch "air" gap, gave a predicted yield of 53 pounds. (We also ran the numbers, per TR12, for a connection with no insulation gap at all, and the yield rose by a factor of six, to 321 pounds.)
Since the calculated value was reasonably close to our tested values, we used TR12 to create the HeadLok schedule above. We applied the TR12 safety factor of 2.2 for design values, the load duration adjustments for dead load, and an additional safety factor of 2. We added the extra safety factor for several reasons. One, we wanted to account for possible installation variables, given that carpenters are having to blindly screw through several inches of material in order to hit the studs. Not every screw will be dead center. Second, we wanted to account for creep - the tendency for installed materials to deform slowly over time. And finally we wanted to err on the side of caution, especially given that FastenMaster, the maker of HeadLok screws, has published a much more conservative schedule (see Technical Evaluation Report 1009-01 at fastenmaster.com).
When using the schedule, keep in mind that it does not address seismic or wind loads - only the gravity loads from the weight of the assembly itself.
Zeno Martin, P.E., S.E., is a senior associate with Wiss, Janney, Elstner Associates in Seattle, Wash.
