The mechanical equipments driven by induction motor are widely used in manufacturing. Aiming at this type of equipment, some holes are advisedly drilled on the bars to simulate the broken bar faults of the motor, and with on-off loads changes of output circuit of the load generator, forced torsional vibration of rotor was generated. Using the above simulation test ways, the typical faults of motor and its driven equipment are tested and analyzed. In addition, an signal analysis method using Hilbert conversion envelope spectrum and real modulation zoom envelope spectrum is proposed, this method can effectively extract the faults information of stator current, reject the useless power frequency. The experimental results indicate that: this method can identify not only the faults of the motor itself, but also municipal fault types of the motor’s driven equipment. Especially, through the contrastive experiments on the unbalance and the torsional vibration of the rotor, the conclusion is made that the method is more sensitive to torsional vibration detection. Also, it develops a new direction for the application and the research of sensorless detection and diagnosis method.