A new method for loudspeaker small noise vibration detection based on wavelet package

Currently, many digital analysis methods are widely used for loudspeaker fault detection. These methods are mainly focused on characteristics like frequency response, impedance and Thiele-Small parameters etc. However, for small noise vibration detection, especially for online detecting, no suitable method is available. Loudspeaker small noise vibration detection is still relied on human hearing, which may harm the tester. By researching characteristics of loudspeaker small noise vibration, we propose a new detecting method using wavelet package. The key algorithm for this new method is how to separate the response signal of the loudspeaker and choose suitable wavelet package for each subsection of response signal to quickly generate the noise vibration curve for each loudspeaker online. By comparing the noise vibration curve of loudspeaker with those of normal ones, we can easily pick out loudspeakers with small noise vibration faults. The simulation result shows that the new method is effective and stable.