Research On TV Loudspeaker Fault Detection Based On Time-frequency Analysis

Fault detection of TV speakers is an important part of the enterprise in the process of assembling and producing TV sets,and its detection results are directly related to the sound quality and product quality of TV sets.At present,the domestic TV manufacturing enterprises use the detection method of direct listening by the listener is being phased out due to the personal state of the listener is difficult to ensure the accuracy and stability of the test results.And in the use of automated detection methods,enterprises will give priority to equipment and labor cost issues,which makes the requirement of algorithm speed as fast as possible,making people give priority to the traditional time-frequency method for analysis and detection.In signal processing analysis,the time-frequency analysis method can express the change pattern between signal frequency and time,and can well extract the time and frequency information from the signal,and the time-frequency method is faster compared with other methods.Therefore,combined with the time-frequency analysis processing method,we have done the following work for TV speaker fault detection:(1)There are three main types of TV speaker fault sound signals: response side fault signal,impurity sound fault signal and polarity inverse fault signal.The causes of the three fault signals are analyzed,and the fault simulation and signal acquisition are performed according to the causes.The data collected from different types of TV sets are pre-processed,and the endpoint detection of the collected signals is performed using a single-parameter double-threshold method,and software is produced for processing,which enables more convenient pre-processing operations.(2)According to the requirement of faster detection speed in the production line,this thesis compares the processing speed of different methods and then selects the fast and simple calculation of the zero crossing rate for analysis,and proposes a fault detection method on the response side of the adjacent zero crossing rate.Since the frequency of the sweep response signal increases with time,the corresponding zero crossing rate also increases with time,and it is possible to characterize the sweep response signal.And for the special environmental sound characteristics in the production line,the grouping difference strategy is used to filter out the special environmental sound,and it is able to analyze and extract the characteristics of the response side fault signal.Through different experimental comparisons,the results show that the adjacent zero crossing rate method for the detection of response side faults can run faster with guaranteed accuracy,which illustrates the accuracy,rapidity and adaptability of the method.(3)Comparing the effect of different time frequency methods for the analysis of normal response signal and impurity sound fault signal,the article chooses the short-time Fourier transform method,combines with the energy segmentation method,and proposes the shorttime Fourier energy segmentation method.The method is mainly to short-time Fourier transform of the time-frequency matrix according to the energy size of the method to carry out segmentation,segmentation of the energy region can correspond to the various components of the signal,the high energy segment of the sweep signal and the low energy segment of the impurity sound signal to eliminate,can highlight the characteristics of the impurity sound fault signal,so as to extract the energy percentage of its features.Through different experimental comparisons,the results show that the accuracy of the short-time Fourier transform in the impurity sound fault detection is better than the comparative time-frequency analysis method,and can be well adapted to different types of TV sets,which is of great significance for the detection of TV speaker faults on the production line,and can reduce the cost of labor in the factory and complete the work of automated detection.