USound GmbH is a fast growing audio start-up, founded with the mission of developing and producing advanced audio systems for mobile applications based on MEMS technology. The technological platform developed by USound is enabling the production of a whole new generation of MEMS micro speakers. Their primary market, estimated at $8.7B in Yole Développement recent analysis Acoustic MEMS and Audio Solutions, will focus on small systems requiring low size and low power consumption, a perfect match for MEMS technology.
Following our last exchange nine months ago, we had the pleasure to get back to them and review their progress. Read below the synthesis of our recent interview with Thomas Gmeiner, VP Global Program Management and Business Development at USound and Nick Renaud-Bezot is Business Development Manager at USound.
Yole Développement (YD): What have been your latest improvements since our last interview (here) in June 2016?
USound: We are focusing on two product categories this year: InSound for in-ear applications, and NearSound for on/over-the-ear ones. InSound is presented in the same form factor as balanced-armature receivers, but solve the issues commonly associated with that product category (high cost, product variations and drop-test reliability). NearSound targets earpieces and headphones, offering better sound quality for those applications. Since our last talk, we have improved the MEMS and package of our Moon (InSound) and Ganymede (NearSound) speakers and are currently ramping up production. For both, the sound output has been improved and distortion has been reduced.
YD: What defines a good loudspeaker?
USound: A good loudspeaker depends on the application, as each product developer is looking for a specific sound signature. For Moon, we have decided to target as wide a potential market as possible by offering a flat output over the full bandwidth (currently 10 Hz to 16 kHz). This enables us to shape the sound profile through filtering (passive- or DSP-based), to compose whichever target is required. In particular, we have managed to recreate what Sean Olive of Harman presented at last year’s AES conference as being the ideal earphone signature.
YD: What is typically the power consumption of the MEMS speakers compared to a classic one?
USound: Piezoelectric devices are pure capacitors in the operating range (in fact up to 2 MHz). This means that the power consumption is purely reactive, thus not comparable with standard speakers (resistive loads). We have identified a few off-the-shelf amplifiers compatible with our MEMS speakers. With them, the power consumption of the audio chain (amplifier and MEMS speaker) is similar to that of electrodynamic speakers. We are also prototyping our own ASIC, which will use energy recovery to dramatically slash the power consumption of the audio chain (expected savings of 50 %).
YD: How does the PZT layer behave over time? Some people reported that the manufacturing of piezoelectric material could be difficult, according to you, what is the most critical step of the manufacturing?
USound: Reliability of the PZT layer has always been one of the most critical development aspects. Already at the beginnings of our technology, the MEMS achieved five billion actuations at full excursion without major performance degradation. This test was the equivalent of 60 days of operation at full power, more than enough for consumer applications.
PZT manufacturing is of course impacting reliability of that layer, and thus of the utmost importance. We have developed recipes with our foundry partners to achieve our goals, and will benefit from their experience in ramping up production to achieve commercial success with our technology.
YD: How do you measure performances of your speakers? Who will be your preferential partnership for such tests?
USound: Typical testing starts with SPL (Sound Pressure Level) and THD (Total Harmonic Distortion) over the product bandwidth. We are finalising ATE development with our equipment partner, to enable at-scale testing common in the MEMS industry. This platform will be implemented by our OSAT vendor, to efficiently ramp that production step.
YD: How do your speakers behave regarding the dynamic range and the distortion?
USound: Due to the low weight of the actuator, our speakers exhibit low distortion from the lowest frequencies and wide bandwidth. For example, we are close to reaching our target of less than 2 % of THD from 10 Hz to 16 kHz for Moon.
YD: Last time, you said that through disassociation of the vibrating structure and the membrane which is amplifying the soundwave, the packaging structure is much more simplified, and so too the integration of the driver within the package, do you have any picture to share with us to describe your principle?
USound: This simplification of the architecture is particularly visible in the case of our InSound products (see below). Whereas a standard balanced-armature receiver (see Jerry Harvey’s blog for example) has to pack a complex combination of coil and magnets within its back volume (thus wasting space as well), ours is completely empty (thus fully dedicated to improving the acoustic performance), and simply built by a simple sequence of SMT process steps.
Cross-section of Moon (MEMS-based in-ear speaker) – Courtesy of USound
YD: How much would you save on the BOM if we consider a classic speaker compared to a MEMS speaker?
USound: We are working on improving the total cost of ownership for the acoustic solution by combining several functions in one device. Our target is to be cost neutral for our customers, with the potential to actually slash costs by at least half in earphone applications.
YD: What would be the final cost of such a solution?
USound: Depending on the speaker size and purchasing volume, unit price should range from less than $1 to a few dollars.
YD: We see a massive adoption of more and more speakers in handheld devices (Smartphones, Voice personal assistant, smartwatches, HMD), you claim wanting to address the consumer (difficult) business in priority, why such a choice?
USound: MEMS businesses thrives on high volumes, so this applies to MEMS speakers as well. We are starting with consumer accessories, as OEMs push this product category to compensate for falling margins of smartphones. The generated revenues will allow us to fund R&D to address other product categories.
YD: With your speakers, you are able to select which directions are selected, can you tell us what type of applications it could open?
USound: One major field of interest is the large bandwidth. Our product can operate in the ultrasound band (we have currently characterised them up to 80 kHz). We are currently investigating how to operate them as both a speaker and a proximity sensor, which should help the industrial design of future smartphones. Other functions are also in development, and will be revealed over time.
YD: What is the investment needed to create such business in the MEMS field? What have been the critical steps of USound’s birth?
USound: MEMS development with its tape outs is capital-intensive business. With their expertise in the field, the founders had identified the right application, and secured rights for the technology. This was followed by proof of concept that a MEMS could indeed generate sounds. The third step was the creation of an acoustic team, so that the speakers would sound good. We are now in the fourth one, which presenting to the world and generating revenue!
YD: We’re approaching the 2017 Q2, are you in line with your objectives, can you remind us your milestones?
USound: Having slipped by a few months, as is common when bringing a new technology to the market, we are roughly in line with our objectives for this year. We are ramping up production of Moon, which has demonstrated high SPL and low distortion over its wide bandwidth.
YD: What are your objectives in term of volume for this year and the next 5 years?
USound: We are targeting a volume of one million Moons this year, and are forecasting a cumulative 500 million units shipped by the end of the next five-year period.
YD: You recently raised €12M saying that part of this money will be engaged to accelerate organizational growth, what will be the others primary needs?
USound: The target of the last investment round is revenue generation. The funds enable the following steps: – Development of a local sales team to accelerate customer adoption, – Industrialisation of the various products, – Development of the next product generations and categories.
YD: You recently announced your partnership with STMicroelectronics. Can you tell us more behind the agreement? What convinced you to select STM?
USound: STMicroelectronics is one of the most successful piezo-MEMS foundries in the world, with a proven track record. With two of the founders having worked at that company in the past, there is a deep understanding of the vendor’s culture, which will help us ramp production in the shortest possible time.
YD: Do you seek other partners to diversify your manufacturing/sourcing?
Our roadmap covers various frequency bands, application profiles and architectures, so we are in talks with several foundry, OSAT and integration vendors to help us bring the MEMS revolution to the audio world.
Thomas Gmeiner is VP Global Program Management and Business Development at USound, having previously held the position of Vice President R&D at Knowles Sound Solutions in Austria. Prior to that, he was the Director R&D at AKG Vienna (Austria). He has more than 20 years of experience in the audio industry, and a holds master degrees from Graz University of Technology and from Pennsylvania State University.
Nick Renaud-Bezot is Business Development Manager at USound, bringing experience in sales, applications and product management from various industries (FA, epoxies, PCB/packaging). He studied MEMS technologies at the ENSMM in Besançon (France), automotive engineering at the HS Karlsruhe (Germany) and technology management at Grenoble School of Management (France).
Key features of the report:
– Microphone, microspeaker and audio IC market forecast in Munits through 2022
– Microphone, microspeaker and audio IC market forecast in $M through 2022
– Microphone, microspeaker, audio IC ecosystem mapping
– Application trends: mobile and consumer, automotive, medical
– Technology trends: MEMS, capacitive, piezoelectric, optical
More information here