[en] The analysis performed with RELAP5/MOD1/019 code and related to six problems proposed by NEA, should be considered a preliminary step in acquiring sensitivity in the code use; the work also allowed to find some errors present in the code version available at Pisa University and running on an IBM computer. As a general remark the analysis performed by varying the time step limits, evidenced the lack of stability of the code and then the noticeable dependence of output variables trends from the imposed time steps. An intervention of the code developers is then required, also with the aim to implement more appropriate criteria for the choice of time steps; in this sense the possibility to introduce some other variable (apart from volumes average density, already included) to control time step limits should be evaluated. Finally the complexity and the number of information which should be included in the nodalization of an assigned problem suggest the need to establish criteria in order to make easier the comparison among different nodalizations and, in general, their understanding