Squirrel cage induction motors (SCIMs) account for approximately 87% of all AC motors in the industry. However, during their operational lifetime, SCIMs are subjected to various stresses, including thermal, mechanical, electrical, and environmental factors, which can lead to faults despite robust design and manufacturing standards. Faults in SCIMs are classified into electrical (stator and rotor faults) and mechanical (bearing and rotor faults). This study investigates the detection of broken rotor bar (BRB) faults in SCIMs using stray magnetic flux measurements with randomly positioned sensor. The research aims to validate the measurement method and assess its efficacy in detecting BRB faults in a steady state using random sensor positioning. The experimental setup involved two groups of SCIMs (Siemens and Končar), each comprising two identical motors—one maintained in a healthy state and the other with an induced BRB. Three analytical approaches were employed: statistical analysis of raw data, time-domain feature analysis, and Fast Fourier Transformation (FFT) analysis. The statistical analysis aimed to validate the consistency of the measurement method. Results showed that the measurements were time-independent for each motor condition, indicating the method’s reliability. However, statistical analysis alone was inconclusive for BRB detection due to significant differences between healthy motors. The feature analysis focused on 19 time-domain features, revealing that peak-to-rms, impulse factor, and clearance factor were reliable indicators for detecting BRB in Siemens motors, especially when a large number of measurements were used. For Končar motors, however, no features consistently indicated BRB faults at high reliability levels. The FFT analysis proved to be the most effective approach for BRB detection. By averaging the frequency spectra over the interval of 0-100 Hz and analysing specific frequency indicators, the FFT method reliably detected BRB faults. The reliability of detection increased with the number of measurements, achieving high accuracy with as few as 10 random measurements. This research demonstrates that BRB faults in SCIMs can be reliably detected using stray magnetic flux measurements with random sensor positioning, particularly when employing FFT analysis. Future research should explore shorter measurement durations, a wider variety of motor types, and the detection of multiple broken rotor bars. The limitations of this study include the controlled laboratory environment, the specific motor types tested, and the constant measurement duration and sampling frequency.