An Investigation of the Earphone Design with Regression Analysis

Abstract

The current research work provides an investigation into the connection between the design parameters of an earphone and the sound quality performance in order to generate recommendations for future designs. As the earphone design was highly dependent on the empirical experience of the engineers, along with extensive trial and error, it was found to be difficult for research and development. Therefore, the current study sought a systematic approach to deal with the design. Along with the simplification of the earphone design problem, the sound quality performance outcomes were characterized by six factors, including total harmonic distortion, output power, frequency response, signal-tonoise ratio, speaker impedance, and headroom with seven levels with eight design parameters for the selection of types of divers (dynamic driver/moving coil, balanced armature driver, and planar magnetic driver), magnet (N35, N40, N45, and N50 Grade Neodymium Magnet), voice coils (copper wire, copper-covered aluminium wire, and silver wire), and diaphragm (polyethene terephthalate, polyethene naphtholate, polyetheretherketone, and polyetheretherketone + polyurethane) for the two drivers with zeros for the earphone products with a single driver. After the consolidation of the data from the manufacturer, regression analysis was conducted, while the six equations were formulated, and the models highlighted the importance of the selection of the type of primary driver and the irrelevance of the type of secondary driver. The result not only provides a direct and evidence recommendation for the research and development of earphones, but the research work also proposes an innovative way for knowledge generation from the information as well as existing data.