In this paper

the adaptive noise cancelling technique is used to improve the signal-to-noise ratio of vibrational data induced from a bearing.Besides the LMS(Widrow) algorithm

the VS(Variable Step) algorithm with more rapid convergent speed and with only a modest increase of computational amount is discussed in comparison with the LMS algorithm.The analysis shows that

when the adaptive noise cancelling technique is adopted to monitor and diagnose a mechanical fault

the VS algorithm is quite feasible.The adaptive noise cancelling techniques at the same state and non-same state are also conducive to identify the bearing fault

but the technique at the non-same state is more applicable than that at the same state.