Page 167 - MEMS and Microstructures in Aerospace Applications

Osiander / MEMS and microstructures in Aerospace applications DK3181_c008 Final Proof page 157 1.9.2005 12:05pm

Microelectromechanical Systems for Spacecraft Communications 157

insertion loss of a switch should be small. Insertion loss can be due both to
impedance mismatch, which results in reflection, and resistive losses. The return
loss, which is not always quoted, is the attenuation of the signal reflected back to the
input.
Transition time is the time required for the signal voltage to go from 10 to 90%
(for on-time) or from 90 to 10% (for off-time) of its full value. The switching time
13
includes the transition time as well as delays in the control system. The control (or
actuation) voltage is the voltage required to open or close the switch. In some
switches, a control current might also be specified. The control power characterizes
the power required to operate the switch. This should not be confused with the
power handling capability of the switch, which is how much signal power the
switch is capable of routing.
Linearity with respect to power is of great importance in microwave switches,
particularly for solid-state switches, which can be highly nonlinear. As power levels
increase, energy will generally shift from the first-order harmonic of the signal to
higher order harmonics. The IP3 point is the power at which the third-order
harmonic intersects with the first-order signal. Ideally the transmission should be
independent of signal level resulting in a very large IP3 point magnitude.
The cut-off frequency is often specified at a figure of merit. The cut-off
frequency is the frequency at which the ratio of the off-impedance to on-impedance
degrades to unity. The cut-off frequency theoretically sets the upper limit for switch
function, although it neglects the effects of inductance that can become significant
at high frequencies. 17
Lifetime is usually measured in switching cycles. The switch lifetime depends
on the microwave signal, and so measurement conditions must be specified if one
wishes to compare different devices. ‘‘Cold-switching’’ refers to a measurement
without any microwave signal present, and measures just the mechanical lifetime of
the device. Since the predominant failure mode is degradation of the electrical
contacts or dielectric charging, the cold-switching lifetime will usually be much
greater than the operation lifetime. The lifetime for a signal-carrying switch is
referred to as the ‘‘hot’’ lifetime. A long lifetime is desirable.

8.2.2.2 Example Performance
As an example, the RF switch performance goals given by the Air Force Research
Laboratory (AFRL) in Rome are as follows: insertion loss < 0.1 dB from 0 to
4 GHz, isolation > 50 dB at 2 GHz, switching time < 10 ms, CMOS-compatible
control voltage levels (generally 0 to 5 V), power handling capability > 30 dBm,
9
IP3 > 55 dBm, and hot lifetime > 10 cycles. 30 It would be useful to extend these
performance levels up to 40 GHz.
8.2.2.3 Failure Modes

For capacitive switches, the two dominant failure mechanisms that limit power
handling are RF latching and RF self-actuation. RF self-actuation occurs when the
root-mean-square (rms) signal voltage becomes large enough to close the switch

© 2006 by Taylor & Francis Group, LLC