Full scale dynamic testing of seismic isolation devices in combined compression and shear is nowadays required by all international standards on seismically isolated structures and/or on antiseismic devices, for prototype tests, and sometimes for production control (or acceptance) tests as well. Full scale dynamic testing is essential to demonstrate the ability of the isolators to withstand severe earthquakes - as well as non seismic actions - thus guaranteeing safety to the structure in which they are installed. Furthermore, type tests are aimed at understanding the effects on the isolator's hysteretic behaviour of different actions, such as vertical loads, amplitude and velocity of the horizontal displacements, and at introducing such effects in proper modelling of seismically isolated structures. Thus, the dynamic tests required by the standards are usually cyclic sinusoidal tests carried out under different vertical loads, with different amplitude, frequency, number of cycles. Therefore, the complete testing protocol on each isolator results very demanding, in terms of global energy dissipation and cumulated displacement applied to it, and quite long and costly. Testing protocols for production control usually are simpler than prototype tests, but in some standard they are still long and complex; and of course the cost incidence of production control tests on the global cost of seismic isolation is higher due to the large number of tests required.

As a contribution to future revision of the testing protocols in the standards, the paper presents a comparison between cyclic loading tests and Time History (TH) tests carried out under the design vertical load on a pendulum isolator characterized by maximum displacement capacity of ± 450 mm. The input of TH tests are the displacement TH resulting as output of non linear dynamic analyses on the structure in which the isolator are installed. The seismic input for such analyses were 3 couples of records of Turkish earthquakes, scaled to the MCE spectrum, and 3 generated earthquakes compatible with the same spectrum. Thus, a total of 9 time history tests were carried out on the isolator. The results are compared with that of cyclic sinusoidal tests similar to those required by standards, in terms of energy dissipation and cumulated displacement.