Nonlinear Characteristic Identification and Analysis of Micro-Electroacoustic Transducer by Using High-Precision Capacitance Micro-Displacement Sensor.

• Main Authors: Yin-Cheng Jian, 簡胤丞
• Other Authors: Yu-Ting Tsai
• Format: Others
• Language: zh-TW
• Published: 2016
• Online Access: http://ndltd.ncl.edu.tw/handle/23nfx7

碩士 === 逢甲大學 === 電聲碩士學位學程 === 104 === Miniature Loudspeakers (MS) have been developed for many years, and the manufacturing quality is getting better related to the competitive marketing. Thus, theaudio quality assessment of MS is very important. However, due to its miniaturization trend and composite materials research, the conventional laser displacement sensor may be useless due to the senseof the material, as well as the random inevitable measurement errors would begenerated from the tiny vibration of the diaphragm. Thus, the laser measurementsystem may need to be reused in many times to determine the correctness of nonlinearparameters. In view of the composite metal foil ofMS, this thesis proposed a non-contact capacitance displacement sensor techniqueto measure its tiny diaphragm displacement. Since the tiny vibration between thediaphragm and the sensor tip would reflect the small variation of magnetic fieldsensitivity corresponding to the changesof the capacitance value. In terms of solving the nonlinear parameters optimization problem, there are three novel improvements in this thesis, a) For the improvement of the iterative efficiency, Brent method is joined to get the superior search step during optimization; b) In order to avoid the indeterminate parametersolutions, the speaker impedance is usedto be the constraint equations for optimization. c) Gaussian distribution function is applied to replace the original polynomial function for reducing the numerous measurement errors. Both simulation and experimental results of this study explore the different measurement errors for identify the nonlinear parameters and validated to the Klippel measurement suite, respectively. Results show thatthe measurement error in a certain rangecan be allowed to accurately predict therobust solution of nonlinear parameters.