Part II: Efficiency of Electrostatic MEMS Transducers
Part I explained that wearables feature more and more functions. These call for increasing edge computing and thus battery power. Since the possible energy amount in a wearable device is limited, especially in a tiny device like a hearable, there is only one option: the power demand of the single components must come down. In an in-ear-device, a major part of the available energy budget is consumed by the sound generation. Finding alternative ways to create sound will therefore enable the further development of smart hearables.
Moving coils lose valuable energy
Currently, most TWS earbuds use conventional moving coil miniature speakers. As the name suggests, a coil moves in a gap due to an electric current and a permanent magnetic field. But this concept is not free from losses. Not all energy that goes in is transformed into motion. For example, the voice coil movement leads to the generation of heat. That may prove convenient in the winter to keep your ears warm. But it certainly does not contribute to the efficiency of a TWS system.
Electrostatic speakers at an advantage
Electrostatic speakers on the other hand are free from losses that are associated with magnetic hysteresis, eddy currents and ohmic resistance in a coil. Dielectric losses are also quite low because there is air between the electrodes. Therefore, electrostatic transducers have conceptually an advantage over electrodynamic transducers: The transducer itself does not dissipate any energy! It just requires electric charges to be moved towards the electrodes.
This electric current is handled by an audio amplifier. And since this driving circuit has limited efficiency, that is is where the energy is potentially burned when using electrostatic speakers. The level of power dissipation depends on the voltage and the amount of charges that needs to be moved.
The lower the voltage and the lower the electrical capacitance that needs to be charged, the lower is the complexity of the circuitry that drives the transducer. Lower complexity means higher efficiency. The target is to provide a novel class of extremely power efficient audio amplifiers for electrostatic speakers, without any external components – making system designers happy.
Nobody wants to have hundreds of volts in their ear…
Conventional electrostatic transducers are usually used in high end loudspeakers and headphones. The membrane in these transducers needs a certain excursion to enable the required sound pressure level. A rather large gap between the membrane and the electrodes is necessary to allow that excursion. But to provide the required force, high electrical voltages are inevitable – for loudspeakers in the range of several thousand volts and for headphones usually a couple of hundred volts.
As mentioned above, this is certainly not helpful for aiming at a fully integrated drive circuitry, keeping in mind that this voltage needs to be generated out of a 3.6V lithium polymer battery for mobile applications. And who really wants to carry around hundreds of volts in their ear?
Precise silicon manufacturing allows for voltages well below 48V
And this is where silicon process technologies with their very precise manufacturing opportunities come into play. Electrostatic microspeakers, based on micro-electromechanical systems (MEMS), have a big advantage: it is possible to design electrode gaps in the range of micrometers and to lever tiny movements [see video below] to enable sufficient sound pressure for in-ear applications. As a ballpark figure: It only requires 0.5 mm³ compression of air in the ear canal to reach a sound pressure level of 120 dB. The required voltages are well below 48V and can be generated by fully integrated charge pumps.
The video above shows a recording of electrostatic actuated cantilevers at resonance frequency. In the upper left part the tiny movement of the three electrodes within the cantilver can be seen. This tiny movement causes a force along the beam and results in a deflection of the entire beam. The two gaps between the three electrodes have a nominal size of 2.5 µm each. There is no electrostatic field between the cantilevers.
So what is the challenge?
Frederick Hunt, the great pioneer of moder acoustic engineering from Havard University, wrote in his famous textbook  that as soon as the power dissipation in the signal source can be solved, “… the electrostatic loudspeaker and its signal source will emerge as the most efficient member of the family of electroacoustic energy transducers.” That is exactly what we want to achieve.
As we have explained here, modern microelectronics and novel audio amplifier technologies pave the way to virtually eliminate power dissipation in the signal source – provided we can design MEMS microspeakers with very low electric capacitance.
The third part of this series explains the fundamental laws of physics determining the transducer capacitance and why these laws of nature lead to very small numbers, when using a smart electrostatic MEMS microspeaker design.
 Frederick V. Hunt, Electroacoustics – The Analysis of Transduction and Its Historical Background, Originally published in 1954; Reprinted in 1981