Vibroacoustic Characterization & Modeling of the Balanced Armature Receiver in Hearing Aids

Balanced armature receivers are powerful and power-efficient high-fidelity, miniature loudspeakers. The combination of these traits makes it the only receiver technology used in hearing aids, where it is one of the key components, alongside the microphones and amplifier. One of the main challenges in hearing aid design is feedback, which is experienced by the user as a high-pitched whistling, due to an unintended feedback loop between the receiver and microphones. When the balanced armature receiver produces sound, the vibrations of the internal components propagate to the chassis, where they propagate to the hearing aid structure and create sound in the outer air. Both sound and vibrations are picked up by the microphones that are in close proximity, and sent back to the receiver to be reproduced. A solid understanding of the vibration pattern is therefore key to design better transducers and better hearing aids. The chassis vibration pattern is governed by the coupled movement of the internal components as well as the boundary conditions posed by the hearing aid. To build a robust, predictive model it is essential to abide by the laws of physics that govern the working principles of the receiver. Therefore, we present a multi-mass lumped element model of the mechanical vibrations that distinguish the internal moving components to capture how their coupled motion propagate to the chassis, whilst we account for the boundary conditions of the receiver. The model covers forces normal to the membrane plane as well as rotation about the membrane axis of rotation and asymmetry induced rotation about the long axis of the receiver. As the working principles and general topology is preserved across different types of receivers, and different manufacturers, the models presented in this thesis are generally applicable to standard single receivers. Distinct geometries and variations in component material, shape and size are captured by the model parameters, which we determine through measurements and analytical methods. To validate the lumped model, we have performed bidirectional exterior and interior laser vibrometer measurements under fixed boundary conditions, as well as electrical impedance measurements. The vibration pattern of the chassis is difficult to characterize experimentally, also because it is difficult to create well-defined boundary conditions. For this reason, we present a 3D finite element model of mechanical system, based on the full geometry of the receiver. The finite element model is first validated against the lumped model response of the translation of the internal components, which has been verified against measurements. Finally, we can use the finite element model to validate the response of our lumped element model for the translational and rotational movement of the chassis and the internal components, when the chassis is suspended by soft springs that mimic the suspension in the hearing aid. The comparative study demonstrated good agreement between the lumped model and the finite element model, and while very asymmetric boundary conditions seem to be more challenging for the lumped model, it is still able to capture the overall frequency dependency of the translational and rotational behaviour of the receiver quite well.