"Experimental" technologies are often slow to gain ground among high-frequency engineers. For example, high-electron-mobility-transistor (HEMT) devices, which are commonplace today, were used by only the most courageous of engineers during the early years of the technology. At present, microelectromechanical systems (MEMS) devices are viewed as a technology with high associated risk, although evidence from the recent Microwave Theory & Techniques Symposium (MTT-S) in Fort Worth, TX hints that greater acceptance of MEMS may be forthcoming.
MEMS devices can be thought of as silicon integrated circuits (ICs) with moving parts. The technology has existed since the 1970s in the form of sensors, but RF devices have been rare. The most common RF device is the MEMS switch (which sacrifices the speed of a PIN diode for greatly enhanced isolation). MEMS microwave switch technology has existed since the 1980s, developed by Dr. Larry Larson of Hughes Research Labs (Malibu, CA) with support from DARPA.
Any new technology must provide advantages over an existing technology before it can replace the older approach. While MEMS switches, for example, offer those advantages over other technologies in terms of size, power handling, power consumption, and their ability for high levels of integration on silicon, they pose a mystery for many RF/microwave engineers in terms of reliability. According to Dan Hyman, President of XCom Wireless (www.xcomwireless.com, Signal Hill, CA), who spoke recently at a MEMS-inspired panel session at the MTT-S, his company's work on MEMS switches, switch matrices, filters, phase shifters, and antenna assemblies has shown the technology to be extremely reliable. He and his designers are such firm believers in the technology, in fact, that he noted a recent development project for a front-end design in which the majority of the passive signal-routing components were MEMS devices.
On the MTT-S exhibit floor, Dow-Key Microwave (www.dowkey.com, Ventura, CA) showed results for accelerated lifetime testing of MEMS switches, with almost imperceptible degradation in electrical performance even after millions of switching operations. The company also introduced their model M1C06-CDK2 single-pole, double-throw (SPDT) DC-to-6-GHz switch rated for an amazing 100 million cycles.
While questions may remain among engineers considering MEMS devices for their designs, reliability does not appear to be one of them. The technology is not the answer for all applications (the lack of speed being an essential limiting factor), but it is an approach worth considering.
