Eight full-scale reinforced concrete beam-column subassemblages, 4 for each non-ductile and ductile detail, were tested under quasi-static cyclic loading. All test specimens were mainly designed for gravity service loads. Four specimens named as non-ductile connections were constructed with typical reinforcement details using in Thailand. The others were assembled with details described by the seismic design building code of Thailand, known as the 1301/1302-61 code. Their performance was examined in terms of lateral load resistance, ductility index, stiffness, energy dissipation, failure modes, including joint shear strength. The test results indicated that the specimens with ductile details achieved better seismic performance in every aspect, particularly specimens with adequate joint shear strength designed according to the SST method. The ductile details have no effect to retard the stiffness degradation in the elastic range, but helped to slow the rates of stiffness degradation after the specimens reached their peak loads. The transverse reinforcement in the joints was more efficient to enhance the ductility of joints with lower joint shear strengths (J1 and J2) than the sufficient joints (J3 and J4). Based on the SST method, the connections with low shear strength capacity, J1, J1D, J2, J2D, J3, and J4 suffered with severe joint shear failure (JS), although their beam sections reached the flexural moments. On the other hand, specimens with high joint shear strengths, J3D and J4D, experienced the beam failure (BF) with mild joint shear failure (JS). Test data also demonstrated that a specimen with lesser amount of joint reinforcement exhibited satisfactory seismic behavior, as long as the joint was provided with the adequate shear strength designed by the SST method. Finally, the ACI 318 Building Code overestimated the shear strength of exterior beam-column joints. Moreover, the simplified SST model much better predicted the demand joint shear strength than the ACI 318 Code.
 HIGHLIGHTS
 
 Numerous surveys of RC buildings after earthquakes also confirmed the inadequate detailing of beam-column joints such as presence of splices, lack of hoops, deficiency of beam bar anchorage, and no steel joint stirrups in the joint
 The ductile details have no effect to retard the stiffness degradation in the elastic range, but helped to slow the rates of stiffness degradation after the specimens reached their peak loads. The transverse reinforcement in the joints was more efficient to enhance the ductility of joints with lower joint shear strengths (J1 and J2) than the sufficient joints (J3 and J4)
 The ACI 318 Building Code overestimated the shear strength of exterior beam-column joints. Moreover, the simplified SST model much better predicted the demand joint shear strength than the ACI 318 Code
 Based on the SST method, the beam-column connections with low shear strength capacity, J1, J1D, J2, J2D, J3 and J4 suffered with severe joint shear failure (JS), although their beam sections reached the flexural moments. On the other hand, specimens with high joint shear strengths, J3D and J4D, experienced the beam failure (BF) with mild joint shear failure (JS)
 The deterioration of beam-column joints under seismic could be effectively restrained by the ductile details according to 1301/1302-61 code with adequate shear strength capacity determined by using the equation of SST model
 
 GRAPHICAL ABSTRACT
Read full abstract