The mechanical component that most fails in gearboxes are the gears. These failures usually occur before the end of useful life projected by criteria of failure standards, due to teeth defects as severe wear and cracking. The condition monitoring of gearboxes evaluates parameters which can indicate the mechanism of failure in process in the gear. The most commonly used monitoring techniques of gearboxes are vibration analysis and lubricant analysis. The experimental analysis consists of the evaluation of an experimental workbench under two conditions: notched gear for crack simulation and; gear with variation of notch for simulation of a crack propagation. The workbench condition was evaluated using vibration signal treatment techniques such as Time Synchronous Averaging, Residual Signal, Demodulation, Statistical Moments, Crest Factor and Statistical Analysis using Beta Probability Density Function .The vibration techniques were adequate to identify the presence of an evolution of the notch. Statistical analysis using Beta PDF was useful to identify the degradation of a tooth as the notch size evolved. The paper's novelty lies in its experimental analysis of a gear with a notch, which simulates the presence of a fracture. It aims to determine whether the system progresses towards catastrophic failure over time (tooth breakage) and examines the system's behavior under different loads and speeds.