In the present time, the inverter-driven induction motor drives are being widely employed in the industries for variable speed applications. These drives are replacing DC motors and thyrister bridges day-by-day in the industries. Therefore, health monitoring of these variable speed drives are receiving large attention from the researchers. In this research paper, the rotor broken bar fault has been diagnosed in the steady state as well as transient conditions by advanced signal processing techniques. For fault detection in the steady state condition, the Fast Fourier Transform (FFT) algorithm is used and for transient fault detection, time domain as well as wavelet transform techniques have been used. The time domain and frequency domain techniques show some kind of lacks for diagnosis of this fault. Therefore, a wavelet transform based new technique for early broken rotor bar, fault diagnosis purpose has been proposed and the rotor fault is efficiently diagnosed in the transient condition. These approximation signals based new technique has diagnosed the fault in the transient condition and consequently, fault may be averted before they become more catastrophic. As a result, industries may save large revenues and unexpected failure conditions.