BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a power amplifier and
more particularly to a microwave power amplifier topology that functions as a low-noise
amplifier when the input signal level is low and automatically switches to high-power
amplification for relatively high input signal levels.
2. Description of the Prior Art
Radio frequency and microwave communication systems are
known to place ever-increasing demands on the linearity and efficiency of power
amplifiers. Unfortunately, conventional power amplifiers operate at maximum efficiency
at or near saturation. Thus, in order to accommodate communication signals having
varying amplitudes, systems utilizing conventional power amplifiers normally operate
at less than peak efficiency for a substantial portion of the time.
In order to solve this problem, so-called Doherty amplifiers
have been developed. Doherty amplifiers were first introduced by an inventor having
the same name and described in; "Radio Engineering Handbook" 5th edition,
McGraw Hill Book Company, 1959, pp. 18-39, as well as U.S. Patent No. 2,210,028,
hereby incorporated by reference. The standard topology for a Doherty amplifier
includes a carrier amplifier, operated in a Class AB mode, and peak amplifier, operated
in a Class C mode. A quadrature Lange coupler is used at the input so that the carrier
amplifier and peak amplifier signals will combine in phase. A quarter wave amplifier
is provided at the outputs of the amplifier. In such amplifiers, as the input RF
drive signal to the carrier amplifier increases, the carrier amplifier is driven
to the point of saturation for maximum linear efficiency. The peak amplifier is
used to maintain the linearity of the output signal when the carrier amplifier begins
Such Doherty amplifiers have been known to be used in various
microwave and RF applications. Examples of such applications are disclosed in U.S.
Patent Nos. 5,420,541; 5,880,633; 5,886,575, 6,097,252 and 6,133,788. Examples of
such Doherty amplifiers are also disclosed in "A Fully Integrated Ku-Band Doherty
Amplifier MMIC," by C. F. Campbell, IEEE Microwave and Guided Wave Letters,
Vol. 9, No. 3, March 1999, pp. 114-116; "An 18-21 GHz InP DHBT Linear Microwave
Doherty Amplifier", by Kobayashi, et al, 2000 IEEE Radio Frequency Integrated
Circuits Symposium Digest of Papers
, pages 179-182, "A CW 4 Ka-Band Power Amplifier Utilizing MMIC Multichip
Technology," Matsunaga, et al., 1999, GaAs IC Symposium Digest, Monterey,
California, pp. 153-156, all hereby incorporated by reference.
The systems mentioned above are known to provide relatively
good phase linearity and high efficiency since the power amplifier only needs to
respond to constant or near constant RF amplitude envelopes. Unfortunately, the
RF amplitude envelopes of multi-carrier signals (multi-frequency signals), used
for example in satellite communications systems, change with time as a function
of the bandwidth thus exhibiting noise-like characteristics. Power amplifiers utilized
in multi-carrier systems must be able to operate over a relatively large instantaneous
bandwidth while providing relatively good phase linearity for RF signals having
In addition, such power amplifiers, used as low-noise amplifiers
(LNA), for example. as a first amplification stage in an RF or microwave receiver,
must be able to provide linear amplification at a relatively high efficiency. Unfortunately,
in applications when the RF drive signal has a non-constant RF envelope, for example,
as in a multi-carrier satellite communication system, the Doherty amplifier is operated
such that the carrier amplifier is operated at about one half of its maximum power
capability during low power operation to provide relatively low noise performance.
Such operation results in relatively low linearity and low efficiency.
In order to increase the linearity and efficiency of Doherty
amplifiers used in applications of non-constant RF envelopes, various techniques
have been used. For example, U.S. Patent No. 5,739,723 discloses an active bias
circuit which biases the peaking amplifier to improve the efficiency. Unfortunately,
the active bias circuit includes a number of resistors which increases the overall
power consumption of the circuit thus providing a less than optimum efficiency.
In order to minimize the bias power consumption of such
a Doherty amplifier, U.S. Patent No. 5,568,086 discloses a combining network for
combining the output signals of the carrier amplifier and the peak amplifier to
provide improved impedance matching. The combining network includes a pair of quarter
waves transformers and a number of quarter wave phase shifting elements. Unfortunately,
efficiencies of only 40% - 50% were realized in the combining network disclosed
in the '086 patent. Power efficiency in many applications such as satellite communication
systems is a critical factor. Thus there is an ever-increasing need to further improve
the efficiency of power amplifiers used in such applications.
SUMMARY OF THE INVENTION
The present invention relates to a microwave amplifier
and more particularly to a microwave amplifier configured as a Doherty amplifier.
The amplifier includes a carrier amplifier, a peak amplifier, a Lange coupler at
the input of the amplifiers and quarter wave amplifier at the output of the amplifiers.
In order to further increase the efficiency, the Doherty amplifier is formed from
HEMT/HBT technology to take advantage of the low-noise performance of HEMTs and
the high-linearity of HBTs to form a relatively efficient amplifier that functions
as a low-noise amplifier at low power levels and automatically switches to high-power
amplification for relatively high-impact RF power levels.
DESCRIPTION OF THE DRAWINGS
These and other advantages of the present invention will
be readily understood with reference to the following specification and attached
DETAILED DESCRIPTION OF THE INVENTION
- FIG. 1 is a schematic diagram of the microwave power amplifier in accordance
with the present invention.
- FIG. 2 is a graphical representation of the output power as a function of the
gain for various biasing points of the carrier and peak amplifiers of an HBT Doherty
- FIGs. 3A-3C illustrate matching networks for use with the present invention.
- FIGs. 4A-4B illustrate biasing networks for use with an HBT and a HEMT, respectively,
for use with the present invention.
The present invention relates to a microwave power amplifier
configured as a Doherty amplifier, generally identified with the reference numeral
20. The microwave power amplifier 20 includes a carrier amplifier 22 and a peak
amplifier 24. In known Doherty amplifiers, both the carrier amplifier and the peak
amplifier are formed from heterojunction bipolar transistors (HBT) and, for example,
as a prematched 1.5 x 30 µm2 x four finger DHBT device with a total
emitter area of 180 µm2. An example of such a device is disclosed
in "An 18-21 GHz InP DHBT Linear Microwave Doherty Amplifier", by Kobayashi, et
al, 2000 IEEE Radio Frequency Integrated Circuits Symposium Digest of Papers,
pages 179-182, hereby incorporated by reference. In order to improve the efficiency
and linearity of the Doherty amplifier 20, the Doherty amplifier 20 is formed from
HEMT/HBT technology to take advantage of the low-noise performance and low intermodulation
distortion of HEMTS and the high-linearity of HBTs.
An example of a HEMT integrated on the same substrate as
an HBT is disclosed in European Publication No. EP0710984A1, published on May 8,
1996, corresponding to European Patent Application No. 95115137.2, commonly owned
U.S. Patent Nos. 5,838,031 and 5,920,230, as well as: "Monolithic HEMT-HBT Integration
by Selective MBE," by D. C. Streit, D. K. Umemoto, K. W. Kobayashi and A. K. Oki,
IEEE Transactions on Electron Devices, Vol. 42, No. 4, April 1995, pp. 618-623;
"A Monolithic HBT-Regulated HEMT LNA by Selective MBE," by D. C. Streit, K. W. Kobayashi,
A. K. Oki and D. K. Umemoto. Microwave and Guided Wave Letts., Vol. 5, No. 4, April
1995, pp. 124-126.
In one embodiment, the carrier amplifier 22 is formed from
an HEMT while the peak amplifier 24 is formed from an HBT. In this case, a low noise
HEMT device acts as the carrier amplifier 22 operating in class A under small input
signal conditions, providing low noise and high linearity. As the input signal strength
increases and the class A amplifier begins to compress and clip the input signal,
the HBT peak amplifier 24 turns on to extend the linear amplification. Since the
HBT has an abrupt base-emitter turn-on resembling a diode characteristic, it optimally
acts as a peak amplifier whose turn-on threshold can be set at the base of the HBT
transistor. The more abrupt the turn-on, the more efficient amplification of the
increasing input power.
In an alternative embodiment of the invention, the carrier
amplifier 22 is formed as an HBT and the peak amplifier 24 from an HEMT. This embodiment
is for applications in which high linearity and efficiency are required under low
input power conditions where operating a more efficient class A HBT amplifier as
the carrier amplifier is more attractive. Using a higher speed low noise HEMT device
as a peak amplifier helps prevent increased noise transmission and additional distortion,
which would be present if an HBT amplifier is used as a peak amplifier due to the
abrupt diode/mixer like characteristics of the HBT which is noisier than a HEMT.
The HEMT would turn-on slower with increased input power where it would operate
closer to its low noise bias operating regime. In addition, the wider band HEMT
device could also preserve the data signal integrity as it turns on from class B/C.
The result is less AM-AM and AM-PM distortion for the peak amplifier when it begins
to turn on.
In order for the output signals from the carrier amplifier
22 and the peak amplifier 24 to be in phase at the output, a Lange coupler 32 is
provided. One input terminal of the Lange coupler 32 is used as an RF input port
34. The other input terminal is terminated to an input resistor 36. One output terminal
of the Lange coupler 32 is coupled to the input of the carrier amplifier 22 while
the other output terminal is coupled to the input to the peak amplifier 24. A 8/4
impedance transformer having a characteristic impedance Zo = 2RL
+ Ropt is provided at the output of the amplifiers 22 and 24. An output
terminal of the power amplifier 20 is terminated to load impedance RL.
Both the carrier amplifier 22 and the peak amplifier 24 are configured to deliver
maximum power when the load impedance RL is Ropt.
The carrier amplifier 22 is operated as a Class A amplifier
while the peak amplifier 24 is operated as a Class B/C amplifier. In order to improve
the isolation between the carrier amplifier 22 and the peak amplifier 24, for example,
when the carrier amplifier 22 is biased as a Class A amplifier and the peak amplifier
24 is biased between Class B and C, matching networks 26 and 28 are coupled to the
output of the carrier amplifier 22 and the peak amplifier 24. As such, the impedance
of each amplifier stage will not contribute to the inter-modulation (IM) performance
of the other stage.
In order to fully understand the invention, a discussion
of known Doherty type amplifiers is presented first. More particularly, as set forth
in: "A Fully Integrated Ku-Band Doherty Amplifier MMIC," supra, the loading impedance
presented to the carrier and peak amplifiers of known Doherty amplifiers is a function
of the output power delivered by the peak amplifier. During low input drive levels
(i.e. levels in which the RF input amplitude is low), the peak amplifier is turned
off resulting in a configuration in which the carrier amplifier saturates at a relatively
low input drive level. As such, the carrier amplifier 22 will result in a higher
power added efficiency (PAE) at lower input power levels. As the input power level
increases, the peak amplifier will begin to turn on as the power delivered by the
peak amplifier increases. The load presented to the carrier amplifier decreases
allowing the carrier amplifier 24 to increase to provide power to the load.
The matching networks 26 and 28 are serially coupled to
the outputs of the carrier and peak amplifiers 22 and 24, respectively. These matching
networks 26 and 28 may be provided as low pass networks, for example, as illustrated
in FIGs. 3A-3C. As shown in FIGs. 3A-3C, the matching networks 26, 28 may be implemented
as a series inductance 40 or transmission line 42 and a shunt capacitance 44 or
open stub 46. In operation, when the carrier amplifier 22 is on and the peak amplifier
24 is off, the matching networks 26, 28 provide a relatively high impedance (mainly
due to the high impedance transmission line 42 or inductance 40) such that the peak
amplifier 24 does not load down the carrier amplifier 22, operating in class A,
to achieve optimum linearity and efficiency under low input power conditions.
The theory of operation of the matching networks 26, 28
is contrary to the operation of matching networks used for conventional power amplifiers.
More particularly, typically in a power amplifier application a low impedance series
transmission line or low impedance shunt capacitance or open stub is provided at
the output of the power transistor in order to efficiently transform the low impedance
of the power transistor to a higher manageable impedance as well as provide isolation
between the amplifying transistors.
In accordance with another aspect of the invention, the
carrier amplifier 22 and peak amplifier 24 may be DC biased tuned so that the optimum
IM performance can be achieved for the specific operating frequency of the amplifier.
For example, for a 21 GHz carrier frequency, a microwave amplifier 20 can be DC
biased tuned to minimize the IM performance at 20 GHz.
Various biasing networks can be used for tuning the carrier
and peak amplifiers 22 and 24. Exemplary biasing networks 48 and 50, are illustrated
in Figs. 4A and 4B. In particular, FIG. 4A is a schematic diagram of an exemplary
biasing circuit 48 for an HBT while FIG. 4B is a schematic diagram of a biasing
circuit 50 for a HEMT.
Referring first to FIG. 4A, the HBT biasing network 48
includes a biasing resistor Rbbc, coupled to an external source of DC,
Vbc. A low pass capacitor Cclp is coupled to the biasing resistor
Rbbc, the external source DC voltage Vbc, and ground to filter
out noise. A coupling capacitor Ccc, may be used to couple the carrier
and peak amplifiers 22 and 24 to the Lange coupler 32.
FIG. 4B is a schematic diagram of an exemplary biasing
network 50 for a HEMT. The HEMT H1 may be either the carrier amplifier
22 or the peak amplifier 24. The gate of the HEMT H1 is coupled to the
gate of another HEMT (or FET) H2 by way of an RF choke L1.
A bypass capacitor C1, coupled between the gate of the HEMT H2
and ground, isolates the bias network from the RF network. The HEMT H2
functions as a current mirror and is selected to have an area 10-20 smaller than
the HEMT H1.
Both HEMTS H1 and H2 are connected
in a common source configuration. The drain of the HEMT H1 is connected
to a DC voltage supply VDD. The drain of the HEMT H2 is also
coupled to the DC voltage supply VDD by way of a variable resistance
R1, which may alternatively be implemented as a FET (or HEMT not shown)
configured as a voltage variable resistance. Either the variable resistance R1
or voltage variable resistance FET may be used to adjust the bias and linearity
of either the carrier amplifier 22 or peak amplifier 24, as discussed above. In
a configuration which utilizes a voltage variable resistance FET, the drain and
source terminals are connected to the drain terminal of the HEMT H2 and
the DC voltage source VDD, in place of the variable resistance R1,
shown in FIG. 4B. A variable voltage supply of DC (not shown) may be connected to
the gate of the voltage variable resistance FET to vary its resistance and, in turn,
the biasing level of the HEMT H1.
A bipolar transistor or HBT may be used in the biasing
network 50 to provide a low impedance voltage source to the HEMT H1.
The transistor Q1 also provides for ramp up of the current in the HEMT
H1 under class B/C operation. The transistor Q1 is configured
such that its collector and emitter terminals are coupled to the DC voltage source
VDD and the gate of the HEMT H2. The base of the bipolar transistor
is connected to the node between the variable resistance R1 and the drain
of the HEMT H2.
A bypass capacitor Cbypass may be used to isolate
the DC voltage supply. The bypass capacitor Cbypass is coupled between
the DC voltage supply VDD and ground.
The biasing circuits, for example, the biasing circuits
48 and 50, enable one or the other or both the carrier amplifier 22 and peak amplifier
to be electronically turned. In the case of the exemplary biasing circuits 48 and
50, illustrated in FIGs. 4A and 4B, respectively, the biasing and thus the linearity
of the carrier and peak amplifiers 22 and 24 may be varied by varying the amplitude
of the external DC voltage coupled to the biasing networks 48 and 50.
The tuning of the carrier and peak amplifiers 22 and 24,
as provided by the biasing circuits 48 and 50, provides many important advantages
in accordance with the present invention. First, the tuning allows the carrier and
peak amplifiers 22 and 24 to be tuned for optimal linearity. Secondly, tuning allows
for improved intermodulation distortion over a relatively wide input power range.
As such, the amplifier 20 can be tuned such that the operating range (i.e. carrier
amplifier frequency) has the maximum IM rejection possible. Moreover, as discussed
above, the relatively high impedance of the matching networks 26 and 28 results
in the virtual isolation of the IM products of the carrier amplifier 22 and peak
amplifier 24, therefore, providing less IM products. Lastly, the tuning can also
be used to provide gain expansion and phase compression for use in predistortion
FIG. 2 illustrates the measured gain and IM3 (third order
modulation products) as a function of output power at 21 GHz for various biasing
conditions of the amplifier 20 in which both the carrier amplifier 22 and peak amplifier
24 are formed from HBTs for illustration purposes. An amplifier 20 formed from composite
HEMT/HBT as in the present invention will be somewhat different. FIG. 2 is presented
to illustrate the electronic tuning capability of the device. In particular, the
IM3 and gain is illustrated for Class A bias operation (Icl = 64mA; Ic2 = 64 mA)
as well as asymmetric bias conditions. In particular, the asymmetrically biased
conditions are illustrated when the peak amplifier 24 is off and the carrier amplifier
22 is biased in a Class A mode (IC1 = 60-64 mA) and the peak amplifier is bias in
Class B (IC2 = 0.3-10 mA). As illustrated in FIG. 2, adjustment of the peak amplifier
biased current (IC2) allows the shape and performance of the IM3 linearity ratio
to be significantly improved across a relatively wide output power range. During
a biasing condition (i.e. Ic1 = 60 mA; Ic2 = 0.3 mA), when the peak amplifier is
nearly shut off, the microwave power amplifier 20 in accordance with the present
invention achieves a relatively dramatic improvement of the IM3 ratio resulting
in a deep IM3 cancellation of about -43 dBc. For multi-carrier communication systems,
an IM3 ratio of about 30 dBc is a typical requirement for linearity. With such linearity,
the microwave power amplifier 20 is able to achieve about 20% power added efficiency
(PAE) and an output power of about 20.1 dBm which is a significant improvement compared
to conventional linear Class A bias mode which achieves about 13% PAE and 18.8 dBm
output power for the same linearity.
Obviously, many modification and variations of the present
invention are possible in light of the above teachings. For example, thus, it is
to be understood that, within the scope of the appended claims, the invention may
be practiced otherwise than as specifically described above.