Marshall Begel
Published: 2004
Total Pages: 20
Get eBook
This paper discusses results of tests conducted to expand the data base on the performance of timber rivet connections in U.S. domestic species to verify existing and proposed design procedures. Eight-, 80-, and 200-rivet connections were tested. The 8-rivet joint tests illustrate the effects of material type, rivet length, and load direction on the behavior of rivet?connections when gross wood failures are avoided. The 80-rivet connections, loaded perpendicular to grain, show the effect of rivet spacing as well as test support conditions on failure mode and strength. The 200-rivet joints, with load applied parallel to grain, provide data for previously reported failure modes, including the rivet yield/wood crush modes modeled by European yield model (EYM) and gross wood failures. The 200-rivet tests also show the effect of rivet spacing on joint strength and failure mode. The tests were conducted in two phases. In Phase I, small?specimens were made from Southern Pine and ponderosa pine solid sawn lumber, Hem-Fir and Southern Pine glulam, and parallel strand lumber. The small specimens failed by a combination of rivet yielding and wood crushing as described by EYM, rather than by a gross failure of the wood block. Joint yield load was determined by fitting test data to a curve and finding the load at 5% offset displacement. Basic trends can be seen in the data, but yield load is a poor predictor of ultimate load. In Phase II, large specimens were made from Southern Pine glulam. Larger rivet spacing, wider specimens, and longer rivets generally increased joint capacity; end spacing had no significant effect. Short continuously supported beams had a higher capacity load than did longer simply supported beams. The results show that when wood failure modes occur, thicker glulam has higher capacity, contradicting the size effect of the design code.