Looking at the huge opportunities offered by the consumer market, the microphone industry is more important than ever. Yole Développement (Yole) has had the chance to interview Matt Crowley (CEO) and Bobby Littrell (CTO) of Vesper, the first company offering MEMS microphones that withstand real-world use in smartphones and other connected devices. As Yole has highlighted in its latest report “Sensors for Cellphones and Tablets 2016”, this technology is poised to grow in the coming years, especially in the smartphone business.
Indeed, Vesper’s are the only MEMS microphones rugged enough to withstand water, dust and particle contaminants. They’re high-performance and enable acoustically rich experiences for any application – from smartphones and wearables to Internet of Things (IoT) devices and connected automobiles. As analysed by Knowmade and System Plus consulting in their report “MEMS Microphone Technology and Patent Infringement Risk Analysis“, the existing microphone products are sharing a lot of similar technologies (which means significant patent discussions among the players…) and Vesper is proposing both new performances and also a new microphone design.
We asked Matt Crowley and Bobby Littrell to give an analysis of their technology and what this can change in the future…
Yole Développement: Introduce us to Vesper and the activities that your company is currently undertaking.
Matt Crowley: Vesper is a Boston-based startup company, spun out of the University of Michigan in 2010, uniquely specializing in piezoMEMS microphone technology. Our strategy is to get to market very quickly. Our first product, the VM1000 went to market in about two years, which is extremely fast for a new MEMS technology. We are currently in low-volume production; last month we shipped our first 100,000 microphones to customers, mostly in China and Taiwan. We currently provide our technology to Asian companies who are designing it into ear-buds, hands-free Bluetooth systems, and waterproof cameras, among other applications. This year, we are planning to release the VM2000, a differential output device which has better RF interference rejection and a superior acoustic overload point. We’re in the process of transitioning from an R&D company to a production company and we’re rapidly growing.
YD: What are the advantages of piezo microphones compared to capacitive technology?
Bobby Littrell: The primary advantage of piezo technology is environmental robustness. Compared to capacitive technology, degradation over time is less of a problem, because piezo microphones lack a small capacitive gap that is sensitive to dust. In capacitive microphones, if you get a particle between the diaphragm and the back plate, that will shift the performance of the capacitive microphone. In fact, we have microphones that have been exposed to the environment since 2009 and still have the same sensitivities that they did when we built them. The other advantage of piezo technology is the ability to achieve a better signal-to-noise ratio (SNR). Currently, our components are on par with what is available commercially for capacitive microphones, but because we don’t have the primary design constraints that the capacitive microphones have, we believe that we’ll be able to get much better SNR as we go forward. We’ve already identified factors we need to improve on and built prototypes that can achieve up to 68dB SNR, and we believe we can achieve more than 75dB SNR with known approaches and have theoretical concepts that could exceed 80dB.
YD: How important is SNR for advanced voice-control systems?
MC: Customers have told us that microphone SNR is still the limiting bottleneck in their systems. Algorithms currently in use are very sophisticated, but if we can provide better data, they would be even more effective. Our customers have told us that 90dB is the amount of SNR that a microphone would have to achieve before there was no more benefit, so there is still a way to go. To put that in perspective, InvenSense has a capacitive microphone that has an advertised 70dB SNR, the highest on the market. However, this is in a very large package, and it uses four MEMS diaphragms. When we talk about SNR in our piezo microphones, we’re talking about a single-diaphragm device. The fact that they employ four diaphragms to hit 70dB shows that this may be the limit of capacitive technology.
YD: What kind of new applications will the piezo microphone enable?
BL: We believe that in ten years’ time, voice interface is going to become the primary interface with the internet and devices in consumers’ houses. For this, you need to use an array of microphones, which can boost accuracy in noisy environments from around 70% to >95%. Piezo microphones can be used indoors, outdoors, in smoky kitchens, in any kind of location, which is critical in these large voice-controlled and monitoring arrays where reliability issues would be a big problem. In addition, whereas capacitive systems have to be constantly on, listening for keywords such as ‘Alexa’ or ‘Siri’, piezo microphones do not have a charge pump, and so have a very fast startup time. They can therefore do duty cycles very rapidly when in ‘always listening’ mode, allowing up to 90% reduction in power usage. We also have a new microphone we’re announcing called the VM1010, which actually uses the power of the piezo element itself to turn on the microphone. This is going to be the world’s first quiescent-sensing MEMS device. In the future, people are going to want to have many low power MEMS and sensors deployed, necessitating very low power consumption in those devices. With our new quiescent-sensing microphone, we can actually wake and listen to sound, while only consuming 3µA of power. This is much less than even the normal drainage of an unconnected battery. Once you say “I can put sensing and voice sensing into a device without any power penalty”, it opens up a lot of applications. I can have a voice-controlled remote control, a battery-powered voice-controlled smart speaker, voice-controlled garage door openers, sprinklers, smoke detectors, thermostats, any kind of thing in your house without any power penalty.
YD: What is your strategy for the smartphone and Internet of Things (IoT) markets?
MC: Our strategy for both smartphones and the IoT is to partner with some of the large semiconductor and packaging companies like AAC who are incumbent suppliers to the smartphone companies. These companies have large support needs, and so it makes sense to work with the large packaging companies. We want to leverage AAC and other packaging partners to do all the sales and distribution, and we’ll focus more on PR and marketing, and really just try and generate interest and demand for the product. My belief is that the smartphone market will demand the largest volumes in the short term, especially if smartphones start going from four microphones per device to six microphones and eventually to ten microphones. We’ve also heard smartphone companies are thinking about adding noise-cancelling headphones, which would require five microphones per headphone set which would have to be high quality. With video recordings on smartphones, the visual quality is very good, but the audio quality is still lacking, so we think there’s a clear end-user demand for higher quality audio recording, and then of course for voice-recognition accuracy, both of which are facilitated by using arrays. However, I also think the IoT market for microphones is potentially huge volume, because all the IoT devices are going to want to use arrays of microphones. If you look at something like the Amazon Echo and the Amazon Dot, the Amazon vision is to have a microphone array in every room of your house. That’s potentially a huge market.
YD: So, you partner with AAC Technologies, could you describe your relationship with them?
MC: AAC Technologies – one of the world’s largest suppliers of MEMS microphones – invested in our company back in 2014, and has been a great partner. We sell MEMS and ASIC wafers, and AAC works with us to deliver a differentiated product in the marketplace. AAC and Vesper complement each other almost perfectly. Vesper’s expertise is in innovative design and marketing, while AAC has great expertise in volume manufacturing, packaging and testing. They also have fantastic resources in quality, customer sales and customer support. We don’t have an exclusive relationship with AAC – although I would say they are our largest partner – we also have a few other partners, typically smaller, Asian packaging companies.
YD: How do you intend to avoid the commoditization paradox, where products are sold in high volumes but at low prices, that has proved an issue for other MEMS sensors?
MC: Something becomes a commodity if three people can make it, or if the customer doesn’t value it. If we look at the statistics, of all the MEMS sensors, the microphone seems to be fighting off this commoditization paradox better than others.
The MEMS microphone industry has always been able to improve performance, and the new improved product can maintain reasonable pricing, while the old products diminish in market value. Additionally, the microphone content per phone keeps going up. In inertial sensing, oscillators and pressure sensing, the number of sensors per phone actually goes down, or is limited to one sensor per device. Because the smartphone market requires multiple devices per phone, we think the business model for microphones is much more like that of an FBAR RF filter. Where the filter content has gone up for phones, the FBAR makers have been able to maintain very good pricing and differentiation, and there really aren’t that many suppliers of the technology. Likewise, due to the unique features of the piezo microphone such as having this wake-on sound capability, as well as our strong patent portfolio, we are really the only vendor that can supply it. We also plan to increase the SNR very rapidly, and get into an SNR space where other suppliers simply can’t compete. In the future, we could even make smart microphones that leverage some unique aspects of piezo technology. For example, with the quiescent-sensing capability, we could put our microphone into a door and due to the low power consumption you would only have to change the battery every few years. If I combine that with an algorithm that does biometric voice authentication, I can have a very secure password-based system for door entry that leverages a lot of the cool software, and it’s a great solution, but also a solution that could only be delivered if you used the piezoelectric MEMS element. As we’re a small company, we aren’t working on software right now. We have to focus, but these are options for avoiding the commoditization paradox in the future.
YD: What are the challenges you face in moving to larger scale production with piezo MEMS?
BL: Fortunately for us there are not a whole lot of issues with depositing high-quality films, especially at the thicknesses we want. As we move forward, we want to be doing things like doping to get higher coupling coefficients in the piezoelectric films to improve performance, and also reducing the thickness of the films while retaining high quality.
MC: An important point that most people don’t quite understand yet is that piezo MEMS has a Moore’s Law type of scaling; if you make the piezo film half as thick, you can make the whole piezo device half the size. We think in the very long term, once piezo technology gets up and running, and the industry gets even more behind it, you’re going to see an increase in demand and a very rapid advancement in piezo sensor technology similar to what you’ve seen with electronics over the past several decades. We’ll get this virtuous cycle where tool vendors invest more money into manufacturing, make money, and then can start making the tools better.
YD: You are collaborating with GlobalFoundries to manufacture your piezo microphones. Why did you choose this company?
MC: I worked with GlobalFoundries in my previous company, so we had a good relationship with them, which was part of it, but one of the big reasons we selected GlobalFoundries was because they already have a lot of expertise in piezo MEMS processing. They’re also the second largest foundry in the world, and so have the financial and operational capabilities to expand into very high volume sockets and applications. If we wanted to target a very high volume smart phone, for example, having the second largest pure-play foundry as your supplier is a huge help in convincing customers that we, as a small company, can reliably supply the part.
YD: Do you think the leadership in MEMS innovation has recently shifted from the US to China?
MC: I wouldn’t say that; I still think a lot of the core MEMS research goes on in the US. What I will say is that China is probably a bit faster at adopting new technologies and commercializing them. They’re a little more willing to take risks on new technology because China is a very competitive and diverse market. Just having a market that large means that there are lots of different sockets and we can go for lots of people, certainly we have had very good successes and experiences working with Chinese companies. I think that in the future, you will see more and more innovation coming out of China, especially in the MEMS and semiconductor space. We’ve seen a focused effort in the Chinese government to acquire US or European MEMS and semiconductor expertise, but I’m not sure if I’d say innovation has completely shifted to China. I think it’s more that the US, China, Europe and Japan and others are almost becoming just one technology market.
YD: What can we expect from Vesper over the coming months and in the longer term?
MC: We’re in production, trying to get a million microphones shipped as quickly as possible and signing up more customers. We’re extremely busy in a good way. We want to keep pushing the technology frontier, keep heavily investing in technology, keep announcing new products, and continue coming out with follow-on products as quickly, or even more quickly than we were able to with the VM1000. I think in the longer-term we’re going to look at other MEMS devices, potentially even open cavity sensor nodes, or a sensor node which is exposed to the environment for smartphones and other applications. We believe that with piezo technology you could combine a lot of these more environmentally-focused sensors, such as ultrasonic devices, microphones, pressure sensors, chemical sensors on a single chip.