Monolithic Microwave Integrated Circuits
A. General Description
Monolithic Microwave Integrated Circuits (MMICs) are used in satellite systems
that require smaller, less expensive circuits or when the parasitic reactance inherent in
hybrid integrated circuits degrades the circuit performance, typically in the upper
microwave and the millimeter-wave spectrum. Examples of systems that use MMICs are
receivers and transmitters for communications, phased-array antennas where small size
and uniform circuit performance are required, and sensors and radars that operate at high
frequencies. The types of circuits required for each of these systems are illustrated by
examining the simple receiver and transmitter systems shown in Figures 3-38 and 3-39,
respectively. In both schematics, a phase shifter—which may be placed in either the
local oscillator (LO), the RF, or the IF portion of the system—has been added to make
the system perform as if each circuit were coupled to a single radiating element of a
phased-array antenna. For non-phased-array applications, the schematic is unchanged
except for the removal of the phase shifter. A photograph of a completely monolithic 30-
GHz receiver is shown in Figure 3-40. Although the high level of circuit integration
illustrated in Figure 3-40 decreases the packaging and interconnect costs, this integration
is not necessary or common. Instead, each function of the system is typically fabricated
on an individual die to permit the optimization of the material system and device type for
each application. Regardless of the level of circuit interconnection, the reliability of the
system is dependent on the continuous operation of each circuit.
Research Center.)
This is understood by examining the receiver circuit shown in Figure 3-38. The
input (RF) signal typically has a very low power level that may be close to the noise
floor. The low-noise amplifier (LNA) amplifies the received signal while at the same
time introduces very little new noise. If the gain of the LNA is sufficiently large, the
noise contributions of the rest of the system will be small since the noise created by later
circuits is divided by the gain of the LNA. Thus, the LNA gain and noise figure, the
measure of noise added by the LNA, determine the receiver noise characteristics. If the
receiver has poor noise characteristics, it will not be able to receive weak signals. The
signal may then pass through a narrow-band filter and into the mixer. The LO generates
a signal that is also fed into the mixer. The mixer combines the two signals through a
nonlinear device, such as a MESFET or diode, and generates a signal at the intermediate
frequency (IF) of fRF – fLO or fLO – fRF and harmonics of the IF, RF, and LO frequencies.
All but the desired IF components must be filtered out. The conversion efficiency of the
mixer is usually dependent on the LO drive power. In addition, a variation in the LO
frequency will cause a shift in the IF that may cause the signal to be attenuated in the
narrow-band filters that are part of the mixer. If the system is to be associated with a
phased-array antenna, the direction and shape of the main beam radiated or received by
the antenna is dependent on the relative phase shift and power level of each transmitter
(and receiver). The relative phase of each radiating element is set by the phase shifter.
Thus, if the phase shift through the circuit varies because of unexpected conditions, the
efficiency of the entire antenna will degrade. It is thus seen that a parametric shift by any
of the components may cause the entire system to fail.
The phase shifter, local oscillator, and mixer circuits are common to the receiver
and transmitter with the exception of a shift in the design frequency. The real difference
Amplifiers
Both low-noise and power amplifiers are used to increase the power of the RF
signal. In almost all systems, this is accomplished by using the transconductance of
MESFETs and HEMTs or the current gain of HBTs. The amount of signal increase is
called "gain" and is usually given in dB, where gain in dB = 10 log (gain). For example,
if the output power is twice the input power, the amplifier has 3 dB of gain. Typically,
the input power and the output power are also specified in dB, permitting the output
power to equal the sum of the input power and the gain. This ideal operation of an
amplifier is accurate for low power levels. Unfortunately, as power levels increase, the
amplifier becomes nonlinear. In the nonlinear region of operation, the output power is
less than the sum of the input power and the amplifier gain in the linear region, or it can
be stated that the amplifier gain is lower in the nonlinear region. Figure 3-41 shows a
typical amplifier characteristic. The point at which the output power drops by 1 dB from
the linearly extrapolated value is called the 1-dB compression point [1]. This value
separates small-signal or linear amplifiers from large-signal or power amplifiers. Note
that this is also the criteria used to differentiate small-signal and large-signal transistors,
since a transistor can be viewed as a simple, unmatched amplifier. This differentiation is
important in determining the failure mechanisms that need to be addressed and the type
of reliability tests that should be performed.
Power Amplifiers
Power amplifiers, by their very nature, must handle high input and output powers.
The maximum voltage swing of the input signal is limited by the breakdown voltage of
the transistor, and thus transistors with high breakdown voltages are required. The
current through each transistor is limited by the resistance in the gate or emitter of FETs
and HBTs, respectively, since ohmic losses are converted to heat, which decreases the
device's reliability. To increase the current handling capability of the device, power
transistors combine many gates or emitters in parallel. This parallel combination
increases the total gate width or emitter area and decreases the resistance, while at the
same time increases the difficulty in matching the input impedance of the transistor to the
output impedance of the prior stage. In addition, the spacing required between the
transistor elements to permit sufficient thermal dissipation creates large devices that are
more difficult to maintain with a uniform voltage [3]. To dissipate the heat from the
transistors, power amplifiers are fabricated on thin wafers, less than 100 mm thick and
typically between 25 and 50 mm, to reduce the thermal path between the transistor's
active region and a good heat sink, such as a metal or diamond carrier. Generally,
thermal constraints limit the design and performance of power amplifiers more than
frequency constraints. Thus, the efficiency of power amplifiers is one of the most critical
specifications, especially in space applications where satellite power is limited, where
dissipation of the thermal load requires heat sinks that increase the system weight, and
where circuit heating can decrease reliability.
Low-Noise Amplifiers\
Since low-noise amplifiers are used on the front end of receivers, they are
designed to handle very low power levels. Thus, the thermal problems and high bias
currents and voltages that affect power amplifier reliability are generally not a concern
for LNA designers. The most important criterion in specifying or measuring an LNA's
performance is the noise figure, and since HEMTs and PHEMTs have the lowest noise
figure, they are used in almost all LNAs. To minimize the noise figure, small gate
lengths and low parasitic gate and source resistances are required [4]. Thus, state-of-theart
LNAs are usually comprised of 0.1 to 0.25 mm gate-length HEMTs or PHEMTs, and
the reliability concerns—such as gate metal sinking and ohmic contact diffusion (see
Chapter 4)—arising from small gate lengths and corresponding small channel thicknesses
are the most important.
To decrease the noise figure of the system, it is important to reduce the circuit
losses, especially before the first stage of the LNA. This includes the package feed losses
and transmission line losses from the antenna since they introduce noise into the system
before the LNA. Besides reducing the circuit losses, noise can be reduced by operating
the amplifier at lower temperatures and lower bias currents and voltages. Lastly, the
noise figure of the LNA is dependent on the matching circuits, which are designed with
an input matching network that minimizes the noise figure and an output matching
network to maximize the gain. The optimum input matching network can be found
through noise parameter measurements of the HEMT. From these measurements, an
equivalent circuit model of the HEMT that includes noise sources can be generated.
Mixers
Mixers convert an input signal at one frequency to an output signal at another
frequency to permit filtering, phase shifting, or some other data processing operation at a
frequency more easily implemented by the circuits. For example, a system may require
the data to be received at W-band, 75 to 110 GHz, but W-band filters have a low Q or a
high loss, which degrades the receiver noise characteristics. Therefore, it may be
advantageous to shift the received signal's frequency to a lower value where low-loss
filters are possible. Ideally, this operation is accomplished without degrading the input
signal's amplitude or introducing additional noise.
Oscillators
Oscillators generate microwave energy for communications, radars, and
navigation systems. For example, modulators, superheterodyne receivers, and phasedlocked
loops depend on a good microwave source to function. In principal, any amplifier
could be made into an oscillator by providing positive feedback to the input terminals so
that the reflection coefficient of the amplifier is greater than one. More often than not,
this is accidentally done by amplifier designers. Therefore, an oscillator is basically an
LNA with a feedback loop that introduces delay-of-integer multiples of 2p. The choice
of the load and terminating impedance to achieve this condition should also guarantee the
proper oscillation frequency and maximize the efficiency or RF power delivered to a
load. In general, there are two types of oscillators: fixed-frequency oscillators designed
to operate at a single frequency and variable-frequency oscillators or voltage-controlled
oscillators (VCOs) with tuning circuits that change the oscillation frequency. The
schematic of a simple oscillator is shown in Figure 3-45. It consists of a transistor with
feedback between the gate and drain, an output matching circuit, and a resonant structure
on the input. Oscillator performance specifications or figures of merit that affect the
system reliability include phase noise and thermal stability
Phase Shifters
Phase shifters are used to impart a repeatable and controllable change of phase to
a microwave signal with no effect on the signal's amplitude. Although they are usually
associated with phased-array antennas, where they are used to control the beam shape and
direction, they are also used in communication systems, radar systems, and microwave
instrumentation. Two methods are commonly used to change the phase in MMICs. The
first method switches the signal between a short and a long length of transmission line to
develop a phase shift of b where b is the propagation constant of the transmission line
and is the differential transmission line length. This type of phase shifter is called a
switched-line phase shifter and is a true time-delay phase shift. The second method
changes the reactance of a transmission line, which changes the propagation constant
along the line. The implementation of MMIC phase shifters is broadly characterized as
either reflection type or transmission type.
By Edgar Alberto Servita 18.856.338
CAF
No hay comentarios:
Publicar un comentario