[en] In many applied situations, the process of quantifying the dynamic human error probability (DHEP) involves two steps. The first step is determining the parameters of two probability distributions: g(t), the critical time for accomplishing a corrective action, and f(t), the time required by operating personal to accomplish corrective action. After these distributions have been determined (or estimated), the second step is to evaluate their convolution. In this paper, an improved method is presented for carrying out the first step. Its efficacy compared to another commonly used method is demonstrated by applying it to the steam generator tube rupture (SGTR). Both methods are based on maximizing the entropy of constrained Weibull distributions. Convincing arguments have been previously presented for using this approach to rationalize the determination of probability distributions for probabilistic risk assessment (PRA) applications