Cross-Laminated Timber (CLT) combines layers of dimension lumber in alternating grain direction to form a mass timber panel that can be used to create entire wall, floor and roof elements. The viability of CLT as an element to resist lateral forces from racking has been of great interest (Dujic et al. 2004, Blass and Fellmoser 2004, and Moosbrugger et al. 2006). However, most research to date has been conducted on full-scale wall panels connected with proprietary fasteners according to European Test Methods. Little research has focused on non-proprietary connections, including nails, bolts and lag screws. The behavior of CLT full-scale wall panels is dependent upon the individual connection properties including the panel-panel connections between adjoining CLT panels within the wall.

The purpose of this research is to evaluate the behavior of three small-scale CLT connection configurations using non-proprietary fasteners. Three different connections -LVL surface spline with lag screws, half-lap joint with lag screws, and butt joint with a steel plate fastened with nails - were tested in both monotonic and cyclic tests. In all, 30 connection tests were conducted, with 15 monotonic test and 15 cyclic tests. Connection strength, stiffness, and ductility were recorded for each connection. Experimental values were compared to National Design Specification for Wood Construction, or NDS (AWC 2012) predictions for connection strength.

Nailed steel plate connections yielded much greater loads and behaved in a more ductile manner than did the lag screwed connections. The surface spline and half-lap connections often failed in a catastrophic manner usually due to splitting of the spline and fastener failure. Experimental results were generally lower than predicted by the yield models for the surface spline and steel plate connections. The half-lap connection resulted in higher experimental results than predicted. A discussion of the connection strength for materials with a non-homogeneous grain direction is also included.