The data shown in Tables of XII to XVII is obtained assigning the signal source of the circuit shown in Fig. 7 by a fast front surge obtained from lab tests for each varistor. The frequency-dependent models present good results for residual voltage of the three varistors. Whereas, the Pinceti model presented numerical oscillation and the IEEE model had more consistent results. In the case of energy absorption, the conventional model had the best results. It occurs because the fitting parameters of the frequency-dependent model takes into account only the final value of the residual voltage, i.e., the voltage waveform is neglected. This procedure may result in a good fit of the waveform, but there are cases in which it is not occur. The conventional model does not represent the dynamic characteristics of the MOSA, even so, there are situations in which it has a best fit of the voltage waveform. Consequently, the energy obtained from this model provides best results.