Our team at Washington State University has been studying the effects of different types of lateral loads on decks (see “Measuring Lateral Loads on Decks” at deckmagazine.com/structure/measuring-lateral-loads-on-decks_o.aspx). Some people have misinterpreted our research, suggesting it indicates that the tension-tie detail in the IRC is unnecessary. But while our testing shows that a deck ledger that is properly fastened to the house framing is quite strong, we also found that the deck joist-to-hanger connection is a potentially weak link. In fact, nailed joist hangers alone simply aren’t adequate for carrying tension forces resulting from lateral loads.
When decks are laterally loaded, the entire load path from the deck flooring to the house must be considered. Wind, seismic events, and even traction loads on the decking from occupants create lateral loads that are transferred to the joists, causing them to pull or push laterally against the deck ledger. In our initial testing, we were surprised by the relatively low loads that were needed to pull joists away from the ledger when typical deck-joist hangers were used.
The problem was two-fold: Smooth nails easily pull out of (or withdraw from) the deck ledger, and some joist hangers use a toenail-type of attachment to the joist that does not “grab” enough of the joist, resulting in tear-out. And when a deck ledger goes through moisture cycling in service, the withdrawal capacity of a smooth nail is reduced even further, to 25% of the tabulated value in the National Design Specification for Wood Construction (NDS Table 10.3.3).
For these reasons, we reinforced the joist-to-ledger weak link by using hanger-manufacturer-approved screws in place of nails, and by selecting a hanger model that accommodated perpendicular joist fastening (see illustration, below). It’s important to note that the hangers used in our study were non-typical for deck construction, because they didn’t have the level of corrosion protection required by the code for use with preservative-treated wood. The only deck-joist hangers (with appropriate galvanizing) available to us used the toenail-type attachment that we had found to be inadequate.
Our testing shows that using screws instead of nails creates a significantly stronger joist-to-ledger connection when joists are loaded in withdrawal; however, a number of variables remain to be investigated before broad conclusions can be made. For example, there are different types of joist hangers: those with a fastener pattern entirely perpendicular to the member face and those with toenail-type fastening. Size is also important. The 2x10 hangers we tested were fastened to the ledger with 10 fasteners and to the joist with six fasteners, while similar 2x6 or 2x8 hangers would have six fasteners into the ledger and four fasteners into the joist. Would hangers with fewer fasteners still adequately transfer withdrawal loads? And as hardware manufacturers are quick to point out, hangers are rated for vertical—not lateral—loads. This raises the question of whether or not a hanger’s vertical capacity is affected by being subjected to a lateral load.
We also have yet to analyze the effects of hidden deck-board fasteners, deck-board orientation, deck bracing, and aspect ratios on loads and load paths. With the “weak link” being the deck’s joist-to-ledger connection, clearly some sort of hardware (or detail) is needed to carry lateral loads from the deck joists to the house diaphragm. The tension-tie detail in the IRC is one way to accomplish this load path. Our research is continuing in order to find other solutions that are more practical and economical.