Page 326 - System on Package_ Miniaturization of the Entire System

300 Cha pte r F i v e

to obtain a constructive effect at the output port of the combiner by adjusting the
embedded microstrip line, which connects the LPF with a vertical coupling structure.
The insertion loss of the RF signal at the output port of the combiner was measured as
1.9 dB. The isolation between ports 1 and 2 of the combiner was greater than 10 dB in
the baseband and 38 dB at 14 GHz (VNA HP8510). A 3-dB bandwidth of 7 GHz is
achieved between ports 1 and 3. Figure 5.39b shows the measured frequency spectrum
at the output of the combiner when a 7-Gbyte/s probability random bit sequence (PRBS)
with a 14-GHz sinusoidal wave is sent. The figure also shows the measured eye opening
of the 7-Gbyte/s PRBS at the output port.

5.4.10 RF MEMS Switches

History and Role of MEMS Switches
MEMS switches as we know them today were conceptualized in the late 1980s and
early 1990s. These switches were of great interest to RF engineers for their potential to
reduce the total area, power consumption, and cost of their devices. MEMS were initially
fabricated exclusively on silicon, since integrated circuit (IC) fabrication at the time was
on silicon. RF MEMS enable switching and tuning of front-end circuits for applications
of mode switching, antenna tuning, and antenna beam steering with phase shifters.
Early MEMS research was funded and performed by industry leaders looking for
applications in a variety of areas, including optics, transportation, aerospace, robotics,
chemical analysis systems, biotechnologies, and medical engineering. Devices such
as microactuators, microsensors, and microrobots were desired for a plethora of
devices, such as for automobile airbags. In time, it was expected that MEMS could be
used for flat-panel displays, optical switches, fiber optics, and integrated sensors. It
was understood that many reliability issues would have to be solved before these
advanced technologies were possible. Early MEMS switches were plagued with
electrical and mechanical issues, such as dielectric charging, substrate delamination,
creep, and fatigue.
Prior to the new millennium, accurate numerical solvers were not available to MEMS
designers. Research was primarily performed by fabricating, testing, and redesigning.
This of course is a slow and expensive process. In 1996, the Defense Advanced Research
Projects Agency (DARPA) in the United States funded research aimed at improving the
computer-aided design technology for MEMS devices. A number of useful software
programs were developed from this funding, including MEMCAD (developed by MIT
and Microcosm), IntelliCAD (developed by IntelliSense), and CAEMEMS (developed by
the University of Michigan). MEMS devices quickly surpassed the RF performance of
their solid-state equivalents. Even early MEMS switches had an insertion loss of 0.15 dB
at 20 GHz, compared to an on-state insertion loss of approximately 1 dB for a typical
GaAs-FET or PIN-diode switch at the same frequency. Today, many of the devices are
commercially available. MEMS switches in particular can be purchased with insertion
losses as low as 0.1 dB up to 50 GHz with the potential for operating more than 500 billion
cycles and handling low watt power levels.
The Holy Grail for many MEMS designers is the cell phone market. Utilizing a
bank of RF MEMS switches, capacitors, and inductors, a cell phone could offer
unparalleled reconfigurability. The cell phone could conceivably work then at any
frequency, on any channel, for any standard, and in any location. Dropped calls would
be a thing of the past.