In the past decades, MEMS sensors have come a long way. In the late 1990s, mass production of MEMS accelerometers for airbag impact sensors and MEMS gyroscopes for rollover detection and anti lock braking system (ABS) began. In the early 2000s, MEMS sensors made a leap from automobile to mobile and consumer electronics, first using MEMS microphone in Motorola’s highly successful Razr mobile phone, and then using MEMS accelerometer in the first Nintendo Wii Remote control.
In the early stage of MEMS commercialization, the schedule of mass popularization of MEMS and non MEMS sensors is followed. The first iPhone released in 2007 has a MEMS accelerometer and a proximity sensor. Ten years later, the iPhone x includes four MEMS microphones, a barometer, a three-axis gyroscope, a MEMS accelerometer, as well as proximity sensors, ambient light sensors and infrared (IR) sensors, magnetometers and multiple image sensors.
In order to let you know more about the development of MEMS in the past few years, we can see that since 2007, more than 2 billion iPhones have been sold worldwide, which makes iPhones the main growth engine of MEMS. According to yole D é In 2020 alone, MEMS products will generate US $10.9 billion in revenue (non MEMS sensor revenue will be higher), covering automotive, consumer and mobile, Internet of things (IOT), healthcare, aerospace, industry and other markets, according to the report of velocity.
With so much growth, what should we do next? Jens fabrowsky, executive vice president of automotive electronics at Robert Bosch GmbH, recently shared their views.
Semi: what are the major challenges facing the MEMS industry?
Fabrowsky: the cost of developing a new generation of MEMS sensors is increasing, which has led to several major changes. In order to compensate for the rising development costs and reduce risks, MEMS sensor suppliers are looking for a broader and more diversified market, not just for mass applications. At the same time, terminal equipment manufacturers demand greater product differentiation, but they don’t want to pay a high price for it, and they don’t want to wait for new hardware iterations. To remain competitive, sensor vendors offer software solutions that support new features. This method is more cost-effective and speeds up the design to production cycle.
Semi: what factors are increasing the development cost of new MEMS sensors, and what measures can the company take to reduce its R & D risk?
Fabrowsky: like most electronic components, the cost of MEMS is determined by development and capital expenditure. The increasing complexity of the content, especially in the interface ASIC and software, makes MEMS development a multi-disciplinary feat, which requires multiple capabilities across multiple design centers to meet the growing schedule.
Manufacturing also plays a role. We often see dedicated production lines built for new MEMS products, which emphasizes both investment and capacity planning. As an industry, we can reduce risk and cost by applying the same manufacturing process to multiple generations of products, which will accelerate product time to market, increase production and improve return on investment.
To what extent will the semi: covid-19 pandemic continue to affect sensor suppliers?
Fabrowsky: MEMS manufacturing process has been affected by supply chain disruption. While the benefits of multiple sourcing and more direct ownership of the process itself help us, no one in the industry can assert that they are already at risk, especially in the context of a new wave of infectious diseases. Our industry relies heavily on just in time manufacturing and logistics, and we are all concerned about the impact of possible process changes. The pandemic has made all of us realize that the important competitive advantage is predictable and safe supply, and the final customer must cherish this cost.
Semi: how can traditional hardware companies like Robert Bosch bring differentiated platforms to terminal equipment manufacturers?
Fabrowsky: we always believe and still invest in our own manufacturing plants. These include the 12 inch ASIC Fab in Dresden, Germany, where we hope to produce the next generation of power and control electronics to meet the growing demand for silicon for automotive electrification.
In our opinion, one of our biggest differences is that our product portfolio does not only include components: close cooperation with internal partner departments provides us with comprehensive system expertise throughout the automotive supply chain. In consumer electronics, we have established extensive partnerships with manufacturers of application processors, wireless systems and sensor processing software. With our expertise, we can offer flexible system integration options to end customers who also benefit from a mature supply chain that supports mass production and field tested quality.
Semi: what does customer demand for software solutions mean to sensor suppliers, and how will suppliers develop to meet this demand?
Fabrowsky: in some silicon product business departments, the work of developing software is higher than that of designing hardware! Software is not only needed in the application layer. It also runs processor interfaces, drivers. In addition, increasingly complex test software ensures high production and minimizes defects.
At the application level, we are increasingly using and promoting open source platforms to encourage better collaboration across the ecosystem. On the contrary, companies that charge for access to their proprietary software environment lose the opportunity to remain competitive for a long time.
Semi: Why are terminal device manufacturers looking for plug and play solutions instead of stand-alone devices?
Fabrowsky: consumers of electronic devices have always wanted products with more features and lower prices. Their demands have had a trickle effect, which has spread to parts suppliers like us. This requires us to manage healthy innovation channels and select products and technologies that can achieve growth and high output. However, it’s not always that simple, and most of the time the component itself is not enough.
Think about our CD-ROM projector for Bosch smart glasses. The only way we hope to win design in this market is to achieve a fully integrated module by using our own scanning mirror and driver chip as well as the integration of laser module and display system. This enables us to provide the final product that is ready for assembly and has been individually tested and calibrated.
Editor in charge: Tzh