<p>Accurate fault diagnosis in industrial machinery requires robust feature extraction and a comprehensive understanding of normal operating conditions. Advanced signal processing techniques, such as the Fast Wavelet Transform and Empirical Mode Decomposition, are commonly employed to enable early fault detection. However, these approaches often lack the systematic integration of baseline data that defines healthy machine behaviour. Furthermore, existing methods predominantly rely on interdependent signals such as vibration, current, and voltage, which are heavily influenced by vibration-based fault signatures. This dependence limits their capacity to differentiate between fault types and to capture the complex dynamics of machine operation, often resulting in incomplete diagnostics. This study proposes a novel methodology to address these limitations by enhancing fault detection in induction motors through detailed analysis of the line-to-neutral voltage signal. A unique fault frequency indicator, extracted via an advanced Fast Fourier Transform technique, is introduced for the diagnosis of outer race bearing faults. Unlike conventional vibration-based approaches, the proposed method mitigates the influence of shaft eccentricity and enables the precise detection of incipient faults, even under complex operating conditions. The effectiveness of the proposed approach is rigorously validated through both numerical simulations and experimental investigations. Results demonstrate a significant improvement in fault detection accuracy and establish a reliable foundation for predictive maintenance. This innovation offers early, actionable insights into machine health and represents a cost-effective, reliable solution for industrial fault diagnostics.</p>