xMEMS Creates Ultra-Tiny Speakers to Power the OTC Hearing Aid Revolution

One of the driving forces behind both sound quality and miniaturization of hearing aids has been the constant development of tiny microphones and speakers (also called receivers) for low-power applications. Without them, we would not have incredibly discreet high-performance hearing aids, true wireless earbuds, and a variety of other popular consumer electronics.

Lately, we’ve been hearing a lot about MEMS (micro-electromechanical systems) speakers, which are compact audio transducers that generate sound. MEMS speakers and the more traditional balanced armature (BA) speakers currently found in most hearing aids differ in several aspects.

According to xMEMS, MEMS speakers are durable, compact, consume low power, and offer a wide frequency response range. More recently, they've been employed in electronics like earphones, headphones, and audio/VR glasses. On the other hand, BA speakers prioritize sound accuracy and are often found in more professional audio applications like in-ear monitors, earphones and hearing aids.

HearingTracker thought it would be interesting to get an insider’s perspective on MEMS speakers, so we turned to Mike Housholder, vice president of marketing and business development at xMEMS Labs Inc. A 25-year veteran in the tech industry, Houshoulder has held positions in several tech companies involved in semiconductors and audio-related applications. He spoke to us from xMEMS headquarters in Santa Clara, Calif.

What specific advantages do your MEMS speakers offer over the balanced armature speakers commonly used by hearing aids today?

I think xMEMS speakers—specifically our small Cowell device—bring several advantages to hearing aids.

First, and maybe most obvious, is the size. xMEMS Cowell is even smaller in total area/size versus balanced armatures (BA), enabling the same deep insertion into the ear canal. Length and width are similar to BA, but our height is 1.15 mm. This makes it appropriate for any hearing aid form factor, including Receiver in Canal (RIC) devices.

Second, Cowell is a full-bandwidth speaker capable of producing the full human hearing frequency range from 20 Hz to 20 kHz. So, not only can Cowell amplify voice, but it can also be used for amplifying rich, complex media—including music, podcasts, movies, etc. This is increasingly important as most advanced hearing aids provide Bluetooth connectivity to smartphones, TVs, and other media devices. Consumers expect their hearing aids to amplify full-bandwidth content, not just voice. Also, independent studies suggest extended high-frequencies (EHF, above 8 kHz) can improve speech intelligibility in the presence of background noise. xMEMS’ speakers have excellent response beyond 8 kHz in contrast to full-range BAs, which typically roll off after 8 kHz.

Third, these microphones have a much smoother response. If you look at their frequency response [see graph below], our MEMS speakers offer a linear response without the resonant peaks and valleys common in many BA devices.

And fourth, these speakers are very reliable, lighter in weight, and don't present problems related to electromagnetic interference (EMI) with Bluetooth^®/wireless radio frequencies. In particular, solid-state MEMS speakers are inherently more robust than BAs because they're made from one monolithic silicon structure and are more resistant to drops or mechanical shock. They're IP58 rated for moisture/dust/particulate resistance, and xMEMS even survive washer/dryer cycles. This leads to fewer failures at the consumer level and fewer returns/repairs for manufacturers.

Speaking of which, from a manufacturing standpoint, MEMs speakers are also really well-suited for automated, high-volume production. With the new OTC initiative, consumer demand for hearing aids and hearable devices is expected to increase dramatically. xMEMS high volume manufacturing capacity will be able to support this increased demand.

How does the MEMS speaker / amplifier combination compare to balanced armature power consumption?

The MEMS speaker is a capacitive device [i.e., it can store energy], it consumes very little current, and is more efficient (lower powered) than a comparable balanced armature.

Where there is area for improvement is in the amplifier efficiency, because the MEMS amplifier needs to boost to high voltage to drive the piezo actuator layer in the speaker.

Today, our existing amplifiers—including the speaker—are equivalent in efficiency to a electrodynamic speaker plus a Class-D amp; this means we are appropriate for consumer devices that expect 6-8 hours of battery life between charges. So, while this may not fit the 18+ hour requirements for medical-grade hearing aids, it can certainly fit market requirements for more casual-use hearing assistance devices including some OTC-class products.

However, future-generation xMEMS amplifiers should allow us to achieve the 18+ hour use target for traditional hearing aids. As I mentioned earlier, our speakers are essentially large capacitors. Our future amplifiers will be capable of recycling the excess energy from the speaker to feed it back to the battery (energy recycling), which should enable our solutions to meet or exceed the hearing aid market requirements for power consumption.

To what level of hearing loss is the output of xMEMS speakers able to support?

Today, at 20 Hz to 1 kHz, our speakers can support between 110 dB to 120 dB SPL. Above 1 kHz the linear gain increase of our speakers can achieve SPL above 130-140 dB.

Future generation speakers will be capable of achieving much higher SPL.

What hearing aid device types are MEMS speakers practical for in terms of form factors (RIC, etc) and category (OTC or prescription)?

From a form-factor perspective, I believe our speakers are appropriate for use in any type of hearing aid. As for categories, this may be dictated in the short-term by power consumption; in the long-term, however, I believe xMEMS will fit into all categories: Over-the-counter, prescription, etc.

When considering the total cost of the MEMS speaker and the amplifier needed to drive it, how does the cost of this solution compare to that of balanced armatures used in hearing aids today?

With the small size and automated supply chain, we have achieved cost parity with BA today with our first-generation devices. As we scale to higher volumes—in both hearing healthcare and consumer audio—our costs will be able to scale lower much faster than manually-assembled BA supply chains.

xMEMS' recent press release describes their speakers as having "no moving parts." How can a speaker have no moving parts?

What this was meant to convey was that our MEMS speakers are a monolithic device, not a series of separate mechanical components like a BA or electrodynamic speaker. The entire speaker function (actuation, frame, diaphragm) is implemented in a single silicon die. The actuator (thin-film piezo) and speaker diaphragm are layers in the semiconductor process. The monolithic implementation, combined with the consistency and uniformity of the semiconductor process, leads to a speaker that is more robust and reliable while also having improved part-to-part frequency response and phase consistency compared to manually-assembled incumbent speaker technologies.

What can you tell me about the xMEMS active vent? What products might it be useful for, and what are the limitations? How would you compare to the ActiveVent from Sonova?

The Skyline DynamicVent provides hearing aid and TWS earbuds the ability of active ambient control to help reduce the occlusion effect. It also provides more natural passthrough or transparency modes, and enables dynamic switching between passive isolation and open/spatially aware listening modes.

From a performance perspective, Skyline was designed to specifications provided by leading hearing aid OEMs to remove the occlusion effect when the vent is fully opened. The addition of a comfort mode (lowest power state, but still slightly open), provides a good balance between power consumption and openness/occlusion effect relief, while still maintaining a good low-frequency response for music and media listening.

Skyline eliminates the audible click noise common with existing active vent implementations. Similar to our speakers, its monolithic, solid-state design provides an IP58 rating for moisture/dust/particulate resistance which can even survive washer/dryer cycles. The aim, of course, is fewer failures at the consumer level and returns/repairs for manufacturers.