This study investigates how concealed glulam beam-end connectors influence shear capacity and tension-perpendicular-to-grain fracture relative to traditional notched beam ends. Current U.S. design provisions assume that concealed connectors behave as reduced-depth sections, yet this assumption lacks experimental validation and may not reflect the true fracture mechanics governing connector pockets. To address this gap, a full-scale experimental program was conducted on Douglas-fir glulam beams at remaining-depth ratios of 0.75d and 0.90d, using both notched and concealed-connector configurations. Twenty-two shear tests were performed following ASTM
D7147 procedures, complemented by three-dimensional finite-element modeling. Experimental results showed that both detail types produced brittle tension-perpendicular-to-grain failures, but concealed connectors exhibited distinct cracking patterns influenced by fastener rows and localized stress interactions. Measured shear capacities were compared against NDS empirical reduction equations and Eurocode 5 fracture-based predictions, revealing that the NDS provisions were consistently conservative for both notched and connector configurations, while Eurocode 5 more closely reflected observed behavior. Analytical stress fields from finite-element models aligned with experimentally observed crack initiation zones, confirming the role of remaining-depth ratio as the dominant parameter governing fracture. These findings provide the first systematic experimental evidence evaluating concealed-connector cross-grain tensile behavior and highlight the need for refined design guidance for mid-depth connector installations in mass-timber construction.
