Sonic Edge is a company that specializes in MEMS miniature speaker technology. MEMS (for micro-electromechanical systems) speakers employ drivers that are fabricated on silicon chips, which can be made to be very small, and can offer advantages in audio quality, and manufacturing. Sonic Edge, based in Israel and Denmark, offers a unique modulated-ultrasound MEMS driver, a second-generation MEMS speaker that is a “one-speaker-solution,” unlike first-generation MEMS speakers that can only work as tweeters, operating in parallel with a traditional dynamic-driver speaker. Sonic Edge’s driver is based on ultrasound, where ultrasound is generated from multiple speakers, each the width of a human hair, and then converted to audible sound via a patented acoustic frequency transformer or modulator.
I spoke with Moti Margalit, CEO and co-founder of Sonic Edge, about their approach to MEMS speaker design, which has applications in earphones, smart glasses, headphones, hearing aids and other products.
Moti Margalit of Sonic Edge.
Frank Doris: Let’s talk about who you are, what Sonic Edge does, and your AirDriver technology.
Moti Margalit: Most of the MEMS speakers on the market today are only good for tweeters. They're not a full-range solution and always need a standard driver to provide the low-frequency sounds. This is an intrinsic aspect of trying to make sound from a MEMS speaker, because the MEMS material is not compliant enough. We started thinking about MEMS speakers back in 2012, when we got our first patent, and formed Sonic Edge in 2019 to take the technology to market. I looked at the general situation and the challenges, and said that to make a MEMS speaker, we needed to go back to first principles, keeping in mind that size is the critical parameter. Making a smaller speaker has huge benefits for customers, but also helps drive costs down. At the end of the day, MEMS also needs to be about making things smaller and cheaper.
We started with a speaker that's the width of a hair that is good for making ultrasound. And we could put about 200 of our speakers on a 10-square-millimeter chip and had a great ultrasound speaker, but it couldn’t be heard as music. We somehow needed to take that ultrasound and make it into audible sound. That is how we invented the active modulation principle. In the past, people have done what's called a parametric speaker. [These generate ultrasonic waves which are modulated with audio signals as they travel through the air. The waves are demodulated and heard as sound. – Ed.]
The sound was not good, and lots of power was required to drive the speakers, required, the size of the speakers as large; in short, everything was against that technology. It's good for making very directional sound, but that's really the only benefit. That's why we thought of using active modulation. Basically, this means that the ultrasound is going through an acoustic channel, and we can open and close that acoustic channel. If we modulate it fast enough, we can generate sound from ultrasound. This is the basis of our technology, a totally new way of generating sound. It took a couple of years to develop and refine.
But why does this enable a smaller and better speaker? Another way of looking at our device it is that functions as a very high-speed air pump. When the speaker membrane is pushing air by moving in one direction, the channel is open and air can flow, and when it is pulling air by moving back in the opposite direction the channel is closed and air is not pulled back, so our speaker is actually a pump working at ultrasonic rates, 400,000 times a second.
Compare this to a standard speaker pushing air at let's say one kilohertz or a thousand times a second. So we have 400 times the advantage. We can make the speaker area 10 times smaller. The movement of the speaker can be 10 times less, so the displacement can be much smaller at ultrasound compared to audio frequencies. Yet we are generating more sound pressure level than a standard speaker. Going to modulated ultrasound is like using a very high-speed pump to replace a slow-moving membrane. That's the reason why we can make a much smaller speaker than existing speakers.
Diagram of the Sonic Edge AirDriver principle.
FD: Just to be clear, this is a mechanical way of converting ultrasound to audible sound. The signal is not going through any kind of electronic or D/A conversion. It’s strictly mechanical.
MM: Exactly.
FD: Is it patent-pending or proprietary?
MM: We have 25 patents, of which 12 are already granted worldwide protection. To make the AirDriver, we are using the same manufacturing companies that today make MEMS microphones. These fabs [fabrication facilities] are already making billions of MEMS microphones, and are adapted to what we need, so we can scale our manufacturing in a rapid manner. Our MEMS speaker is made of pure silicon and we use electrostatic drive to move the ultrasound speakers and modulator, which provides very low power and high fidelity.
Some of our competitors are using PZT as an actuator, and PZT is a problematic material. First of all, it has lead in it, limiting its use in some applications, and second, it does not have the manufacturing infrastructure and reliability history that pure silicon has. Our design is highly reliable, scalable, and low-cost. It’s similar in design to what's called a double-backplate microphone, the standard MEMS microphone most people use today. So, we just use that platform, but in a different design.
Sonic Edge AirDriver modules.
FD: I'm trying to wrap my head around this. How does the modulator get you from say 400,000 movements a second to the audio band?
MM: We’re doing amplitude modulation of the ultrasound.
FD: Ah. Like radio broadcasting!
MM: Exactly. Like radio waves. This is exactly the principle. The active modulation creates a sideband that we can actually hear, and in a very similar way to electromagnetic waves. If instead of modulating it in a single frequency we modulate it in multiple frequencies, we will get the whole audio spectrum.
FD: The obvious applications would be headphones and smartphones. Are there any others? And let me talk about the sound of the MEMS headphones I’ve heard so far. There’s a sort of electrostatic clarity and detail.
MM: Our driver delivers increasing sound pressure level as it goes down to low frequencies. It actually goes down to DC. No other speaker can do something like that. Regarding other applications, it’s scalable. So, you can increase the active area and get more air flow. On the other hand, you can use just a single ultrasound speaker and channel and incorporate it into something like gas sensors, or medical applications for pumping air or gases. Another far-reaching application is for very-low-frequency applications like hydrophones for sonar, where you can make infrasound sonar with a very small and energy-efficient source.
So it really is a new way of pumping air, with sound being a large, but not the only, application. However, for speakers we believe that our technology provides a scalable way to generate sound and will be the speaker of choice in the future.
We think this can first of all replace micro speakers, wherever they are in earphones, headphones, glasses, cell phones. But as we mature the technology, we have a roadmap where every two years we will increase the sound pressure level by a factor of two for the same driver area. Combining the increase sound pressure level with a scalable architecture, we can also create larger and larger speakers with more volume. As an example, in cars, we can take the speakers out from the door and put them in the roof, or in the headrest.
FD: As an audiophile it drives me a little crazy when I’m in my car and the speakers are at the sides of the doors, aiming at my legs. I mean, it's just silly.
How loud are the AirDriver speakers capable of playing? Obviously good enough for headphones and other applications.
MM: The best way to think of it is like a pump, and one unit, let's say 10 square millimeters, generates a certain airflow. Now if we want to replace a speaker, we need to look at the maximum airflow that speaker is generating, and then see what the area is that we need [for an AirDriver in order to] to replace that speaker. If we're looking at a 2-inch speaker, we’d replace it with something with the same area, but with factor of ten times smaller thickness, so instead of the 1-1/2-inch width of a 2-inch driver, we would need only 1/8-inch, and we also only need a much smaller back cavity. Even more importantly, the configuration of MEMS drivers doesn't need to be square or round, it can be any shape. And because they are truly identical drivers, we can do very good beamforming, so the listener can get personalized sound while the other people in the car hear their own personalized music. [Beamforming is the process of changing the output of multiple speakers using delay, volume differences or cancellation effect to control the behavior of an acoustic wave. – Ed.]
FD: Can you tune the frequency response of the drivers, or is that more a question of the acoustic space that the drivers would be placed in?
MM: Our device is basically a pump, so in lumped-element speaker design “speak” it's a “current flow,” and you can do whatever you are used to doing in acoustic design in terms of introducing cavities, pipes, or tubes to tweak that. We don't need the same size back cavity as a standard speaker. When you have a standard speaker and the back cavity is too small, it will affect the compliance of the driver and as a result, push up the resonant frequency, and you won’t get really a nice low frequency performance. But a pump can handle pressures up to a thousand pascal without effecting performance so we can have smaller cavities.
You can think of our drivers as a way to build up very loud speakers by using a series of isolated emitters. This is similar to using multiple LEDs for anything from flashlights to large displays. In the future we can build arrays of speakers and do meaningful things with those arrays.
Diagram of a Sonic Edge AirDriver unit.
FD: Are you going to be selling these under your own brand or are you going to be an OEM to other manufacturers?
MM: We are an OEM. We sell our speakers and the ASICs (application-specific integrated circuits) that drive them, and a module which connects to a standard audio interface. We will not have an earphone brand and we will not compete with our customers.
FD: So there might be instances where a company will say they’re using Sonic Edge technology, or is it going to be more invisible to the general public?
MM: If we do a good job in branding and marketing, then our customers will be proud to say that their products are powered by Sonic Edge. We need to work to make that happen. But that is the plan.
FD: Tell us more about Sonic Edge, the company.
MM: The original concept was developed in 2012. My business partner Ari uses hearing aids, and we were discussing between us, why can't hearing aids really provide good sound? And we started studying why.
FD: I confess I use them myself at times.
MM: It's better than nothing, but it's not good.
FD: They've gotten better. But I tell people I don't want to listen to life through bad D-to-A converters. For speech intelligibility when sitting in a crowded restaurant, they’re great, but I can’t listen to an audio system with hearing aids. It's too sad for me.
MM: So, this is really the issue. The electronics are really good. There are really a lot of very smart things going on there. but the bandwidth of the transducers is limited. That's why if you go to a high-end IEM (in-ear monitor), you have five, 10, even 20 balanced armatures to really get the full frequency spectrum. And we tried to understand the limitations of why speakers can't be small and good. That's how we invented this modulated ultrasound technology.
Life takes you in many [directions and] we did not actively pursue that at [first]. We started Sonic Edge back in 2019, and then the coronavirus stuck. The first two years were basically working out of my attic and developing the technology. In 2022 we had working prototypes, working with some consumer electronics companies to demonstrate earphones and glasses. By the end of this year, we should have the manufacturing up and running, with the goal of having products in the market in 2025.
I think this is a very big step, and our modulated ultrasound approach is the right way to make MEMS speakers. We've spoken with many companies and they think that MEMS speakers will be the speaker of the future. Some are saying next year sounds reasonable. Some are saying it'll take another five years. But it's not a question of if it'll happen, only when.
