Page 17 - EE Times Europe Magazine | June2020
P. 17

EE|Times EUROPE   17

                                                      The Growing Importance of Innovative MEMS Microphones


                                                                            AOP is key for undistorted concert music recordings
                                                                            and echo cancellation when a microphone is placed
                                                                            close to a speaker. Good phase matching is important
                                                                            for microphone arrays and improving the performance
                                                                            of beamforming algorithms. In other words, if you are
                                                                            building a system that has to perform beamforming,
                                                                            pick up audio from a distance, and cancel echoes,
                                                                            you need an array of microphones with all of the
                                                                            above-mentioned features.
                                                                              In addition to providing all of the audio parame-
                                                                            ters for the use cases mentioned above, microphones
                                                                            should be robust against water and dust. They must
                                                                            also work within their specifications under variations
                                                                            in temperature, humidity, pressure, and other environ-
                                                                            mental factors. In an ideal world, microphones would
                                                                            not generate any self-noise at all but would emulate the
                                                                            human ear, which has an extremely low noise floor. In
                                                                            reality, however, there are technical limitations defined
                                                                            by physics. There are typically four noise contributors
                                                                            in MEMS microphone systems: the MEMS itself, the
           Figure 3: FluSense is an innovative concept for monitoring influenza trends. It   ASIC, the package, and its sound port. The highest noise
           combines a microphone array with a camera to analyze coughing and patient   contributor for state-of-the-art microphones used in
           counts. (Image: University of Massachusetts)                     smartphones is the MEMS component. Infineon has
                                                                            improved the MEMS component, pushing the boundar-
                                                                            ies of MEMS microphone performance beyond even the
           microphones contains more information and less self-noise, which   current best-in-class dual-backplate technology. The company’s efforts
           makes the subsequent processing on the edge processor and in the   here culminated in the development of Infineon’s innovative, patented
           cloud more efficient. Better raw audio data input results in a higher   sealed dual-membrane (SDM) MEMS technology.
           hit rate, lower false acceptance instances, and, subsequently, a lower
           error rate for system wakeup. In short, having better microphones in   A LOOK INSIDE A MEMS MICROPHONE
           your smart speaker means that the next time you ask it a question, you   Before explaining how the new SDM works, it makes sense to take a
           won’t have to repeat yourself.                        quick look at the inside of a MEMS microphone and see what challenges
             In general, each acoustic parameter has its own importance and   the different components have to overcome to deliver the best perfor-
           enables certain specific use cases (Figure 4). High SNR, for example, is   mance. A typical design combines a MEMS sensor and an ASIC. MEMS
           important for picking up distant and whispered voices in smart speak-  microphones extract audio pressure changes as electrical signals. The
           ers, video conferencing devices, and similar far-field applications. A   ASIC then processes these signals either in differential analog or in
           study conducted by Infineon showed that higher-SNR microphones can   digital format at the output. The first challenge is handling the pressure
           capture whisper audio 40% better than standard microphones. High   levels of loud sounds, as they produce large mechanical movements in


































           Figure 4: High-performance microphones are characterized by a number of different parameters.

                                                                                           www.eetimes.eu | JUNE 2020
   12   13   14   15   16   17   18   19   20   21   22