Systems Innovation Dominate MEMS EC 2011
November 3, 2011 - Monterey, CA - The MEMS Executive Congress continued the second day with presentation from further members including end users who were creating finished goods with the devices. This brought an overview of the development ecosystem for MEMS to start with process technology and devices and continue through final product development with system and applicaiton software.
As a discussion from the top - “Software is the lifeblood of today’s innovation and is changing the design paradigm within many markets”. IBM’s Rational Division discussed a “system of systems” approach to bringing products to market based on software flexibility and systems objectives rather than component based design. This holistic top down approach was used by GM to bring the new drive system for the Chevy Volt from concept to production in just 29 months.
As a result, the MEMS world has to shift from the traditional stocking distibutor model with standard product components. The cost of selecting a vendor for an end product is not only the unit price at time of assembly, but also includes a pro-rated loaded cost factor for the development and test of the system that incorporates the new components. As the cost of manufacturing is similar for like products (accelerometers, gyros, etc) there is not a lot of variability in the high volume unit. prices. The differences are in the SDKs that are created by the MEMS supplier to allow for the cost and schedule for the ind product development to be as short as possible. reference designs, interface approved vendor lists (AVLs), and multi-OS development platforms are now a key differentiating factor on the parts.
The MEMS sensors are now moving from a human interface and feedback role to automated sensor to sensor / machine to machine communication. In this new role, the sensor network plays an equal role to the sensor itself for being the key role of the data integrity and data transfer that is picked up from the sensor. These include the power control and the wireless communications for the sensors. A key factor in these environments is the control of the EM environment around the sensors to make sure that the wireless traffic in the area does not impact or get falsely reported by the sensors.
The other aspect of the sensor network is the actual interpretation of the collected data. The data mining and database schema are integral points in the sensors providing a beneficial vs a detrimental function. Historically, data analysis and trend information was deduced from minimum data sets, and the gap avoidance and fill were the main portion of the effort. With high density, high performance monitor networks, upwards of 1 million sensors can provide 24/7 reporting to the data collection system. For the first time the task has shifted from gap filling to data filtering and data reduction. The problem is amplified by the high precision of the MEMS devices which require threshold setting for interpretation of the data. The high resolution of these sensors, in some cases have a system noise floor that is significantly measurable with low order bits & bytes of the sensor readouts. All of this network interface is now pushing the ubiquitous security issues down to the component and SDK level. A sample of the current end products that show the full ecosystem is the goggle based information system from Recon Instruments.
As MEMS become mainstream, not only in mobile devices for human interface and automotive systems but move into the connected world of consumer devices, their ecosystem and necessity is now an indispensable part of the $1T+ electronics industry and has moved up from being an “emerging technology”.