W. SAMARASINGHE, S.J. LAWRENCE CSIRO Building, Construction & Engineering, Sydney
and A.W. PAGE
The University of Newcastle, NSW

The bond wrench test, first developed in Australia, has become the preferred method of evaluating flexural tensile strength of masonry for on-site quality control and for obtaining parameters required to design walls against wind and earthquake loads. The paper describes investigations into the variability of bond strength results obtained by this method and the adverse influence of this variability on characteristic design strengths. Numerical and experimental evaluations of the performance of the bond wrench are presented. Comparisons are made between the bond wrench method and the testing of simply supported stack-bonded beams, and a calibration block has been used to study the stress distributions imposed by the bond wrench in various circumstances. It is shown that the conditions imposed by the bond wrench on the test specimen can vary significantly from the assumed linear stress distribution used in the calculation of the joint strength. It is also shown that the local stresses caused by the bottom clamps are substantial. As a consequence, there is an urgent need for standardisation of the bond wrench geometry and testing procedures.