In this paper, we present DPN-SOG, a software tool written in C++ for fault diagnosis of discrete event systems modeled by bounded labeled Petri nets. DPN-SOG (for Diagnosability analysis of Petri Nets using Symbolic Observation Graphs) implements the semi-symbolic diagnoser approach developed in [1, 2] for fault diagnosis of bounded labeled Petri nets. The implemented approach aims to cope with some limitations of the classic diagnoser-based approaches, namely the state-space explosion problem, the intermediate models and the double-checking procedure for diagnosability analysis. The key features of DPN-SOG are: (i) the on-the-fly building of the diagnoser and analysis of diagnosability, (ii) the generation of only the necessary part of the diagnoser to perform the diagnosability analysis and online diagnosis, and (iii) the evaluation of the time/memory consumption for the construction of the diagnoser and the analysis of diagnosability.