WITHDRAWAL PROPERTIES OF HARDWOOD DOWELS IN END JOINTS I - EVALUATIONOF SHEAR PERFORMANCE OF ADHESIVES

Withdrawal strengths of hardwood-dowel joints do not increase in propo rtion to embedded-dowel lengths, because shear stress concentrates at ends of bond lines due to elongations of dowels and wood members. Appl ications of elastic adhesives with lease dowel-hole clearances are dis cussed to moderate the stress concentrations. A series of tests was co nducted to determine withdrawal strengths of dowels (Acer mono) 8 mm i n diameter, glued parallel to the grain of jointed wood members (Crypt omeria japonica). A considered theory(3) based on the Volkersen model for lap joints contains two bond-line parameters : Shear strength fv a nd shear stiffness Gamma. These parameters were determined by pull-out tests of dowel joints. curve-fitting Eq.(8) on the observed strength as a function of embedded-dowel length. f(v)s determined for polyureth ane and epoxy adhesives were in good agreement with he shear strengths of each adhesive measured directly by push-out tests of dowels, 5 mm in embedded length, in which shear stress in bond line was assumed to be distributed uniformly (Table 3). Furthermore, withdrawal displaceme nts of dowels calculated from Gamma agreed well with measurements in p ull-out tests at ultimate load levels (Figs. 8 and 10). These results suggest good applicability of the theoretical expressions considered. Among three adhesives tested, polyurethane adhesive, which had a small er Gamma, achieved larger pull-out strength than epoxy and resorcinol- formaldehyde adhesives in the range of embedded lengths between 6 and 10 in length/diameter ratios (Fig. 11). No statistically significant d ifference in joint strengths was observed between 1.0 mm and 0.5 mm do wel-hole clearances for polyurethane adhesive. However, average streng th observed for 1.0 mm clearance was larger than that for 0.5 mm; Gamm a determined from curve-fitting for 0.5 mm clearance was found to be t w;ce as large as that for 1.0 mm, while f(v)s for the two conditions w ere almost the same (Fig. 9).