Optimization of Electrostatic MEMS Speakers and Array Weighting Methods for Digital Sound Reconstruction

To address the issues of low transient sound pressure output and insufficient beamforming capability under binary drive constraints in digital sound reconstruction (DSR) technology, this paper optimizes the structural parameters of electrostatic MEMS speaker units and investigates array directivity and weighting methods. First, by optimizing the diaphragm thickness, side length, and cavity height, under 12 V DC bias and 12 V AC drive, the sound pressure level of the unit at a distance of 1 mm is increased to 72.6 dB at the resonant frequency of 76 kHz, achieving high transient sound pressure output capability, with a response time of less than 23 μs. The acoustic output satisfies the linear superposition requirement, conforming to DSR requirements. Second, a uniform planar array is constructed to analyze the influence of the number of elements and element spacing on directivity. The optimal array configuration is determined as 32 × 32 with an element spacing of λ/2, achieving a main lobe width of 3.14°. Finally, in view of the binary drive characteristic of DSR arrays, a unit density mapping weighting method based on error diffusion is proposed, which converts continuous weights into binary switching states. Simulation results show that with Taylor weighting, the 32 × 32 array achieves a peak sidelobe level of −27.49 dB, which is significantly improved compared with −13.29 dB for the unweighted case, indicating that the method can approximate ideal continuous weighting and effectively suppress sidelobe levels to a certain extent. This work provides an effective technical solution for sound field control in DSR loudspeaker arrays.