There are many problems in practical diagnostics of high-power, squirrel-cage induction machines, operated in industrial conditions. These problems make the diagnostics a difficult matter. The problems appear already at the stage of gathering diagnostic signals. The only practically accessible signals are the currents on secondary sides of current transformers monitoring supply currents of separate machines. Due to the proximity of other conductors, carrying high currents, these signals are sometimes accompanied by heavy background noise. The next stage of the diagnostic procedure is off-line analysis of registered signals referring to investigated objects. The main problem at this stage is proper interpretation of gotten results, especially in the cases where determination of the cage state, of investigated machines, is difficult due to ambiguity of diagnostic signals. The ambiguities and doubts by taking diagnostic decision result from that the diagnostic-like signal could be generated not only by cage damage but also by slightly asymmetrically manufactured rotor cage. In such difficult cases, the many-year diagnosing experience is the indispensable precondition for proper diagnostic verdict. The present paper includes some reports on case studies encountered during practical diagnoses of high power, 6kV induction machines, operated in real industrial conditions.