WO2006066461A1

WO2006066461A1 - Butler doherty power amplifier

Info

Publication number: WO2006066461A1
Application number: PCT/CN2004/001517
Authority: WO
Inventors: Paul Gareth Lloyd; Mark Anthony Briffa
Original assignee: Huawei Technologies Co., Ltd.
Priority date: 2004-12-24
Filing date: 2004-12-24
Publication date: 2006-06-29
Also published as: CN101091322A; CN101091322B

Abstract

A Butler Doherty power amplifier suitable for wireless communications systems comprising of a Butler matrix network and a Doherty power amplifying stage, is disclosed. The Butler matrix network which consists of an input processing stage comprising a first hybrid or coupler, a first phase difference setting stage, a second hybrid or coupler, and a second phase difference setting stage, a plurality of power amplifier driver stages, a third hybrid or coupler and a plurality of phase setting stages is capable of accepting an RF input signal, maintaining equal power sharing in a plurality of power amplifier driver stages and supplying the Doherty power amplifying stage with independently controlled drive signals. The Doherty power amplifying stage which consists of a carrier power amplifier with an associated matching network, a peaking power amplifier with an associated matching network, a plurality of phase setting stages and Doherty combining network, is capable of accepting drive signals from the Butler splitting stage, amplifying these signals with high efficiency under Doherty operation and providing the resultant signal to an output load. The Butler matrix network facilitates the optimal distribution of power to the constituent stages of the Doherty amplifier such that optimum linearity and efficiency is maintained for high peak to average ratio RF signals.

Description

BUTLER DOHERTY POWER AMPLIFIER

Field of the Invention

The present invention relates to high efficiency power amplifiers utilizing the Doherty technique for use in wireless telecommunications applications in the radio and microwave frequency bands.

Background of the Invention

There is a well-known need for high efficiency linear RF (radio frequency) power amplifiers. This need stems from a number of applications for such amplifiers. In general these applications are found in portable transmitter applications such as mobile phones, and radio base-station applications.

In cellular base stations, satellite communications and other communications and broadcast systems, many RF (radio frequency) carriers, which can be spread over a large bandwidth, are typically amplified simultaneously in the same PA (power amplifier). For the PA, this has the effect that the instantaneous transmit power will vary widely, and potentially very quickly. This is because the sum of many independent RF carriers that is termed a multicarrier signal tends to have a large peak to average ratio (PAR). It also tends to have a similar amplitude distribution as bandlimited Gaussian noise, which has a Rayleigh distribution.

The main difficulties for a PA amplifying such signals are power efficiency, linearity and bandwidth. A conventional "class B" PA exhibits maximum DC-to-RF power conversion efficiency when it delivers its peak power to the load. Since the quasi-Rayleigh distribution of amplitudes in the summed transmit signal has a large difference between the average power and the peak power, the overall efficiency when amplifying such a signal in a conventional class B amplifier is very low. For a "class A" amplifier, the efficiency is even lower. The key issue with operation with such signals is therefore to improve efficiency when the PA is operated below its peak power i.e. under back-off.

The linearity of an RF PA is usually characterized by its AM-AM (AM=amplitude modulation) and AM-PM (PM=phase modulation) distortion characteristics. Non- linearity manifest themselves as cross-mixing of different parts of the signal, leading to a leakage of signal energy into unused channels, and the literal distortion/degradation of the signal/information being transmitted. The signal to be transmitted can be backed-off and therefore restricted to a smaller part of the total voltage swing to increase linearity. However, this reduces the efficiency of the amplifier even further. The linearity of a PA is also greatly reduced if the amplifier saturates. This means that it is not possible to increase efficiency by driving the amplifier into saturation, since the distortion would then reach unacceptable levels.

There are a number of well-known methods that can enhance the efficiency of a power amplifier when operated with a high PAR signal under back-off. One way of increasing the efficiency of a RF PA is to use a so-called "Doherty" amplifier disclosed in U.S. Pat. No. 2,210,028. With its development, numerous Doherty amplifiers have been disclosed in patents and reported in literatures since Doherty' s original disclosure. Typical improved Doherty amplifiers are disclosed in U.S. Pat. No. 6262692 to Stengel et al, which will serve to summarize the state of the majority of pertinent prior art.

Figure 1 shows one possible example block diagram schematic of a prior art conventional Doherty amplifier having a shared driver line-up, which is termed the shared driver line-up 100. Doherty amplifier with the shared driver line-up 100 includes, a driver amplifier 102 for amplifying a RF input signal at point input port 101, a splitter block 107 for splitting the driver signal from driver amplifier 102 into two paths so that the two signals are created with a 90 degree phase shift between them to have the correct amplitude and phase difference required by a subsequent Doherty stage 140, the Doherty stage 140 connecting to load 120. Wherein Doherty stage 140 includes a carrier amplifier 114 connecting at point 131, a peaking amplifier 111 connecting at point 130, a first matching network 116 used for additional matching or phase alignment to the output from the carrier amplifier 114, a second matching network 115 used for additional matching or phase alignment to the output from the peaking amplifier 111, a Doherty output combining network 119 accepting the two correctly aligned and power amplified signals and combining them.

Splitter block 107 comprises an in-phase splitter 106 and a 90-degree delay line 108. Carrier amplifier 114 comprises a power amplifying device 112 and a load matching network 113. Similarly, the peaking amplifier 111 comprises a power amplifying device 109 and a load matching network 110. Doherty output combining network 119 comprises a λ/4 impedance inverter 117 followed by an impedance matching network 118. Load 120 connects with impedance matching network 118.

The RF input signal is provided by an RF source having characteristic impedance (e.g. 50Ω). Driver amplifier 102 serves the need for sufficient overall PA gain when the amplifier stages of the Doherty stage 140 are not sufficient by themselves at the RF frequencies of operation. The two signals that result at points 130 and 131, should have the correct amplitude and phase difference required by the subsequent Doherty stage 140. Typically this means that the signal at point 130 applied to the peaking amplifier 111 is larger in amplitude, and is phase delayed by 90 degrees compared to the signal at point 131 that is applied to the carrier amplifier 114. This is because the peaking amplifier 111 is typically operated in class C mode and has therefore lower gain, than the carrier amplifier 114 which is typically operated in class AB mode. First matching network 116 is useful due to the fact that matching network 113 most likely will introduce an impedance inversion that will destroy Doherty PA operation if not corrected by the additional matching network 116. This impedance inversion issue is reported in U.S. Pat. No. 6,262,629 to Stengel et al. where an alternative solution is disclosed whereby a so-called inverted Doherty architecture is chosen. Similarly, second matching network 115 is useful due to the fact that matching network 110 most likely will introduce an impedance inversion that will destroy Doherty PA operation if not corrected by the additional matching network 115.

The key aspect of Doherty PA with the shared driver line-up 100 is that the limited RF power gain is overcome by driver amplifier 102 ahead of the Doherty stage 140, but it doesn't allow adjustable power levels to the carrier and peaking amplifiers respectively. The issue of the driver stage is important since it only can degrade efficiency because it draws DC power, but contributes zero to output power. A further benefit and Doherty's prime reason for introducing feedback is to solve the problem of poor linearity. Although feedback can be an excellent means of improving linearity, it is prone to instability when the loop gain and/or loop delay is too high for a given transfer function/bandwidth. Feedback is therefore not a suitable way of improving linearity when the amplifying stages have low gain and require substantial impedance matching in order to achieve as desired output power since delay will be relatively high. What's more, because there is no isolation between carrier amplifier 114 and peaking amplifier 111, this might cause power to be reflected back towards the in-phase splitter 106 and driver amplifier 102. Ultimately this reflected power may cause the power delivered into the carrier amplifier 114 to be a non-linear function of the drive power presented to the peaking amplifier 109 thus causing the potential for increased non-linearity.

Figure 2 shows a block diagram schematic of a prior art conventional Doherty amplifier. This is termed the independent driver line-up 200. The key difference of this line-up exists with a first driver amplifier stage 206 to carrier amplifier 212 and a second driver amplifier stage 205 to a peaking amplifier 209. The driver amplifier stage is now provided independently to each of the carrier and peaking amplifiers, instead of one shared driver, as in the previous prior art line-up 100. The rest of the Doherty amplifier with independent driver line-up circuit 200 is the same as that of shared driver line-up 100.

Since peaking amplifier 209 is typically operated in class C, its input impedance is likely to change significantly as the power is increased and the peaking amplifier device 207 is turned on. Therefore, benefit of the Doherty amplifier with independent line-up 200 is the potential for improved isolation between the carrier and peaking amplifiers. What's more, Doherty amplifier with independent line-up 200 solves the problem of achieving a suitable overall gain when the power gain of the individual transistor in the Doherty stage is limited. However both Doherty amplifier with independent line-up 200 and that of shared driver line-up 100 are failed in allowing adjustable power levels to the carrier and peaking amplifiers.

FIG. 3 shows a block diagram schematic of a prior art conventional Doherty amplifier using RF signal decomposition 300 which is disclosed in U.S. Pat. No. 6,639,464 to Hellberg. The RF input signal which is a digital input applied at point 301 has been decomposed into two parts. This signal is inputted to a D/A (digital to analog) converter 303 which converts the digital signal into an analog one. This is then upconverted to the appropriate RF frequency by up-converter 306. The RF output from 306 is inputted to the carrier amplifier 309. A portion of this RF signal is taken off and delayed by 90° in 307. The digital input signal at point 301 is also applied to an attenuation unit 302. The purpose of this attenuation unit is to modify the magnitude of the RF signal applied to the peaking amplifier 308. The function that is applied by this attenuation unit is intended to improve the linearity of the Doherty PA 300. After the attenuation unit 302, the digital signal is converted into an analog one by a D/A 304. The resulting analog signal that represents the additional magnitude function to be applied to the peaking amplifier 308 is then modulated onto to the 90° delayed RF signal by using a mixer 305. This new RF signal is then applied to the peaking amplifier 308. The remaining parts of the prior art Doherty power amplifier 300 are similar to the previous prior-art examples; the output from the carrier amplifier 309 is connected to the peaking amplifier 308 via a Doherty impedance inverter 310. The load in this case is an antenna 311. In this way, because the RF input signal is decomposed to digital baseband signal, this means that the Doherty power amplifier can no longer accept an entire RF input signal.

U.S. Pat. No. 5,604,462 to Gans et al. discloses a method by which a so called Butler matrix (referred to as a "power sharing network" by Gans et al.) can be used to share power between a number of amplifiers and then transmit this power on one or several antennas.

FIG 4 shows a prior-art example 400 of how power sharing networks such as Butler matrix networks might be used with multiple power amplifiers. In the example shown, a number of inputs are provided. Signal Tl at point 401 is a test signal. When this signal is passed through the power sharing network 405, it is divided equally and applied to power amplifiers 406 407 408 409. The outputs from these amplifiers are re-combined in another power sharing network 410, re-distributed to a load 411 and multiple antenna loads 412 413 414. The nature of the power sharing networks is such that signal Tl arrives only on load 411, but more importantly, is amplified and shared by all of the power amplifier stages. In a similar fashion, signal S 1 is applied at input 402, is then amplified by all of the power amplifier stages but appears only on the output Bl on antenna 412. Similarly, signal S2 is applied at input 403, is then amplified by all of the power amplifier stages but appears only on the output B2 on antenna 413. The concept can effectively be further extended for any number of signals, up to signal SN on input 404 that arrives at as signal BN on antenna 413. The key point is that each of the power amplifiers 406 407 408 409 will each amplify an equal amount of the input signals Tl, Sl, S2. SN.

However, Gans et al. was to use the Butler matrix as a way of detecting non- linearity and do not address the problem of splitting power for Doherty PA applications and allowing adjustable power levels to the carrier and peaking amplifiers.

Summary of the Invention

It is desirable to have a high efficiency PA that is efficient with high PAR signals and is therefore operated under back-off. In a Doherty amplifier operating at high RF and microwave frequencies it is desirable to be able to accept an RF input, divide this RF signal into two parts and then apply these signals independently to the carrier and peaking amplifiers.

In order to overcome the deficiencies of the prior art, it is an object of the present invention to have a Doherty PA, that has an RF input splitting arrangement capable of splitting and diverting RF power independently to the carrier and peaking amplifiers with allowing adjustable power levels to carrier and peaking amplifiers, so that linearity can be improved and the various requirement of supplying RF power into the constituent Doherty power amplifiers are sufficed. It is a further object of the present invention to perform this splitting arrangement with a driver amplifier stage whereby the Doherty PA comprising driver amplifier stage has equally shared power.

The scheme is as following:

A quasi-linear high efficiency power amplifier comprises a Doherty stage comprising a carrier amplifier and a peaking amplifier, and a Doherty output combining network which is connected to outputs of the carrier and peaking amplifiers; and a driver amplifier stage being capable of supplying the Doherty stage with drive signals; wherein the said power amplifier further comprises, an input processing stage for receiving an input signal, maintaining equal power sharing in and supplying power to the driver amplifier stage with controllable phase differences; a driver power combining stage for receiving and combining the output signals from the driver amplifier stage, converting phase differences to amplitude differences, and supplying requisite and independently controllable drive signals to the Doherty stage.

The input processing stage comprises, a first hybrid or coupler having one input, used for dividing the input signal into a plurality of signals with the same amplitude and a relative phase difference of 90°; a termination for proper matched operation of the first hybrid or coupler; a fine phase difference setting stage, used for accepting signals from the first hybrid or coupler and applying a phase difference between the signals to produce a nominal phase difference of 0° or 180°; a second hybrid or coupler for converting phase differences from the nominal to nominal amplitude differences; a gain difference setting stage, accepting signals from the second hybrid or coupler, wherein, a phase difference is applied to produce signals having a desired phase difference, for application to the driver amplifier stage.

The said input processing stage comprises, a first hybrid or coupler having one input, used for dividing the input signal into a plurality of signals with the same amplitude and a relative phase difference of 90°; a termination for proper matched operation of the first hybrid or coupler; a gain phase adjuster stage, used for accepting signals from the first hybrid or coupler, and applying a gain and phase difference to the signals to produce a desired gain difference and a nominal phase difference of 0° or 180°; a second hybrid or coupler, used for converting phase differences from the nominal to nominal amplitude differences and vice versa, and applying to the driver amplifier stage.

The fine phase difference setting stage comprises a first phase adjuster in one path, and a first phase delay element in the other path; the gain difference setting stage comprises a second phase adjuster in one path, and a second phase delay element in the other path. The gain phase adjuster stage comprises a gain difference setting attenuator which is connected with a fine phase difference setting phase shifter in one path and a phase delay element in the other path.

The power amplifier further comprises a coarse phase difference setting stage for facilitating a desired phase difference between the signals that are delivered to the Doherty stage, which is connected between outputs of the driver combining stage and inputs of the Doherty stage.

The said driver combining stage comprises a hybrid or coupler used for converting phase differences from the nominal to nominal amplitude differences and vice versa, and whose inputs are connected to outputs of the driver amplifier stage and outputs are connected to inputs of the coarse phase difference setting stage.

The said driver combining stage comprises isolators for absorbing power that is reflected back from the input of the peaking amplifier as it turns on and off in accordance with Doherty operation, which are respectively connected to outputs of the driver amplifier stage, and a hybrid or coupler used for converting phase differences from the nominal to nominal amplitude differences and vice versa, and whose inputs are connected to outputs of the isolators and outputs are connected to inputs of the coarse phase difference setting stage.

The said driver combining stage comprises a hybrid or coupler used for converting phase differences from the nominal to nominal amplitude differences and vice versa, and whose inputs are connected to outputs of the driver amplifier stage, and isolators for absorbing power that is reflected back from the input of the peaking amplifier as it turns on and off in accordance with Doherty operation, which are connected the outputs of the hybrid or coupler and inputs of the coarse phase difference setting stage.

The said driver combining stage comprises a converting phase differences from the nominal to nominal amplitude differences and vice versa, and whose inputs are connected to outputs of the driver amplifier stage, and an isolator for absorbing power that is reflected back from the input of the peaking amplifier as it turns on and off in accordance with Doherty operation, which is connected between one output of the hybrid or coupler and one input of the coarse phase difference setting stage, and a phase delay element compensating for the additional delay that the isolator may introduce, which is connected between the other outputs of the hybrid or coupler and the other output of the coarse phase difference setting stage.

The said coarse phase difference setting stage comprises phase delay elements for introducing the coarse phase difference required by the Doherty stage.

The power amplifier further comprises phase delay elements for adjusting that the direction of desired load impedance changes is correct for Doherty operation, which are respectively connected between outputs of the Doherty stage and inputs of the Doherty output combining network.

According to the invention, a high efficiency Doherty PA is provided whereby an RF input signal is split and a suitable characteristic impedance is supplied to the RF source. The resulting signals are then processed conveniently at low power before being applied to the driver amplifier stage arrangement incorporating the input processing stage and the driver power combining stage, which server as Butler matrix structure. The processing is provided to enable the independent control of the drive power delivered to the carrier and peaking amplifiers respectively. Although the power delivered into the carrier and peaking amplifiers can be independently controlled, the power provided by the driver amplifier stage is equally shared due to the use of the Butler matrix structure. Phase delay elements on each of the outputs of the carrier and peaking amplifiers allow the recombination of the split signal in an optimal way.

By virtue of having equally shared power, the driver amplifier stage capacity is fully utilized at all power levels by both the carrier and peaking amplifiers. At low power levels, when the peaking amplifier is switched off, all drivers in the driver amplifier stage are supplying power to the carrier amplifier. Compared with the shared driver line-up or independent driver line-up prior art concepts, the overall linearity is improved by virtue of the fact that the active driver periphery is at least twice the size of that of the prior art. Compared with the prior art, no degradation in efficiency is experienced, as all drivers are switched on in any case. At high power levels, the equally shared drivers are contributing the correct amount of power to both carrier and peaking amplifiers. Even though these two power requirements are different, the power remains equally shared in the driver amplifier stage, providing a further linearity advantage.

What's more, it is provided that an independent power splitting arrangement which does not necessarily use attenuators. By means of the phase delay elements between the carrier/peaking amplifier and Doherty output combining network, the outputs of the peaking and carrier amplifiers are optimally combined. Due to the use of the isolators in the driver combining stage, power reflectedby the constituent carrier and peaking amplifiers may be automatically "isolated".

Brief Description of the Drawings

FIG. 1 shows a block diagram schematic of a prior art conventional Doherty amplifier having a shared driver line-up.

FIG. 2 shows a block diagram schematic of a prior art conventional Doherty amplifier having an independent driver line-up.

FIG. 3 shows a block diagram schematic of a prior art conventional Doherty amplifier using RF signal decomposition.

FIG. 4 shows a block diagram schematic of a prior art arrangement having power sharing networks.

FIG. 5 shows a Doherty amplifier according to an embodiment of the invention.

FIG. 6 shows an alternative embodiment of the input processing stage.

FIG. 7 shows an alternative embodiment of the driver combining stage.

FIG. 8 shows an alternative embodiment of the driver combining stage.

FIG. 9 shows an alternative embodiment of the driver combining stage.

Embodiments of the Invention

For a complete understanding of the present invention and for further objects and advantages thereof, preferred embodiment is now made to the following description taken in conjunction with the accompanying drawings of the invention.

Butler structure consisting of a input processing stage, a driver amplifier stage, a driver combining stage and a coarse phase difference setting stage has the property to maintain equal power sharing in a plurality of power amplifiers and controllable performance. The core of component of the Butler structure is the coupler or hybrid which can work as a transform between amplitude difference and phase difference. Through utilizing Butler structure to form a so-called Butler-Doherty PA which achieves splitting arrangement having capable of splitting and diverting RF power independently to the carrier and peaking amplifiers, whereby efficiency and linearity maximization can be achieved simultaneously.

FIG 5 shows an embodiment of a Doherty PA 500 in accordance with the present invention. Doherty PA 500 includes a input processing stage 550, a driver amplifier stage consisting of a first amplifier 510 and a second amplifier 509 which are respectively in two-path; a driver combining stage 551; a coarse phase difference setting stage 532; a Doherty stage consisting of a carrier amplifier 520, a peaking amplifier 518; and a pair of phase delay elements 521 and 522; Doherty output combining network 524. Wherein two-path outputs of input processing stage 550 is connected to first driver amplifier 510 and second driver amplifier 509 respectively; first driver amplifier 510 and second driver amplifier 509 are respectively connected to two-path inputs of driver combining stage 551 of which two-path outputs are respectively connected to two-path inputs of coarse phase difference setting stage 532; two-path outputs of coarse phase difference setting stage 532 are respectively connected to input of carrier amplifier 520 and input of peaking amplifier 518. There is a fifth phase delay element 522 between output of carrier amplifier 520 and one input of Doherty output combining network 524. Similarly, there is a sixth phase delay element 521 between output of peaking amplifier 520 and another input of Doherty output combining network 524. Output of Doherty output combining network 524 is connected to load 526 as prior art.

In Doherty PA 500, input processing stage 550, driver combining stage 551, coarse phase difference setting stage 532 and a pair of phase delay elements 521 and 522 are newly added. The input processing stage 550 establishes the power sharing property like Butler matrix so that driver amplifier stage 509 510 carry the same amount of power but have a purposely introduced phase difference between them. The driver combining stage 551 converts this purposely introduced phase difference into an amplitude difference in order to meet an amplitude difference requirement for the input signals of the Doherty stage. The coarse phase difference setting stage 532 sets the bulk of the desired phase difference that the power delivered into the carrier and peaking amplifiers 518 520 such that signals passing through these amplifiers are summed constructively. The pair of phase delay elements 521 and 522 are used to ensure that the direction of desired load impedance changes is correct for proper Doherty operation and optimization of combining signals at point 541.

In succession, the input processing stage 550 will be described. An RF input signal from an RF source having characteristic impedance (e.g. 50Ω) is applied at input 501 of the input processing stage 550. This input signal is divided into two parts with the same amplitudes by a first 90° hybrid or coupler 502. A termination 503 is provided for proper matched operation of the first 90° hybrid or coupler 502. The two signals obtained from the first hybrid or coupler are ideally 90° apart from each other. A further phase difference can be applied by using a fine phase difference setting stage 530, which comprises a first phase adjuster 505 in combination with a first phase delay element 504. The objective is to achieve a phase difference on the output of the fine phase difference setting stage 530 that is 0° or 180° and to enable small or fine adjustments in this phase to be performed between the two-path separated signal. Two-path outputs of the fine phase difference setting stage 530 are connected with a second hybrid or coupler 506. It is important to maintain 0° or 180° at the input of the second hybrid or coupler 506 since it means that the signals coming out of the second hybrid or coupler 506 are equal in magnitude. This in turn, will allow the power sharing property to be exploited. The two-path output signals of second hybrid or coupler 506 have essentially the same magnitude and a phase of 0° or 180° compared to each other. An additional phase difference can be applied to these signals by using a gain difference setting stage 531, which comprises a second phase adjuster 507 in combination with a second phase delay element 508. The input processing stage 550 serves as the Butler matrix network of which nature is such that a phase adjustment just after the second hybrid or coupler 506 results in a gain adjustment after the next hybrid or coupler i.e. a third hybrid or coupler 513. This is a key and useful attribute of the present invention.

The equal amplitude signals from the second phase adjuster 507 and the second phase delay element 508 are subsequently applied to driver amplifier stage 509 510, which receive and amplify an equal amount of RF power. Apart from the 0° or 180° phase difference introduced by the previous stages, the only difference between the two-path signals being amplified is the phase difference intentionally introduced by the second phase adjuster 507 in combination with the second phase delay element 508. The driver amplifier stage 509 510 contain enough stages to reach the desired gain and are sufficient enough to supply enough RF power to both carrier and peaking amplifiers (520 518) when both of these amplifiers are producing their full peak power.

The outputs from the driver amplifier stage are next applied to the driver combining stage 551. In the preferred embodiment, this contains optional isolators 511 512 which are connected to two-path outputs of the driver amplifier stage respectively. The purpose of these isolators is to absorb power that is reflected back from the input of the peaking amplifier (518) as it turns on and off in accordance with Doherty operation. The position of these isolators 511 512 depends on the quality of said isolators. If the relative phase difference between manufactured isolators is not controlled, the position of the isolators should be as drawn in FIG 5 just before the third hybrid or coupler 513. This is because phase differences between isolators can be adjusted away by the second phase adjuster 507 in combination with the second phase delay element 508. However, if the relative amplitude difference between manufactured isolators is not controlled, the position of the isolators should be after the third hybrid or coupler 513 as drawn in FIG 8. This is because amplitude differences can be adjusted away by using the first phase adjuster 505 in combination with the first phase delay element 504.

The signals from isolators 511 512 are next applied to the third hybrid or coupler 513 of which purpose is to transform the phase difference applied by the previous stages using the second phase adjuster 507 in combination with the second phase delay element 508 into the desired gain differences. This is a key aspect of the present invention and allows the independent control of power arriving at the carrier 520 and peaking amplifiers 518.

The phase difference of the signals arriving at the output of third hybrid or coupler 513 will be the phase difference of the signals applied at the input of the second hybrid or coupler 506. In order to equally share power in the driver amplifier stage 509 510, this phase difference will be either 0° or 180° between the two-path signals. A small deviation from this 0° or 180° condition can be tolerated and exploited to fine control the phase difference at the output of the third hybrid or coupler 513 at the slight expense of power imbalance in the driver amplifier stage 509 510. In this case, the fine control of the phase difference would be performed in the fine phase setting stage 530.

The two-path output signals from the third hybrid or coupler 513 are then each applied to a coarse phase difference setting stage 532 which comprises third phase delay element 514 and forth phase delay element 515. The purpose of these phase delay elements is to introduce the coarse phase difference required by the two Doherty PA branches.

One of the outputs from the coarse phase difference setting stage 532 is applied to carrier amplifier 520. This amplifier could itself be comprised of an RF power amplifying device 516 and an output matching network 519. This carrier amplifier 520 will also typically have an RF matching network on the input but this has been omitted for the sake of clarity in FIG 5 since this is well known in the prior art. As is well-known in the prior art, this PA will be typically operated in class AB mode. The rest of the operation of the carrier amplifier 520 is in accordance with the prior art of Doherty PA operation.

Another output from the coarse phase difference setting stage 532 is applied to peaking amplifier 518. This amplifier could itself comprise of a RF power amplifying device 527 and an output matching network 517. This peaking amplifier 518 will also typically have an RF matching network on the input but this has also been omitted for the sake of clarity in FIG 5 since this is well known in the prior art. As is well-known in the prior art, this PA will be typically operated in class C mode. The rest of the operation of the peaking amplifier 518 is in accordance with the prior art of Doherty PA operation.

Doherty output combining network 524 is comprised itself of a λ/4 impedance inverting network 523 and an output load matching network 525. Finally, the output from the Doherty output combining network 524 is connected to the desired load 526.

Signal strength taken from one path is added to another one in the invention so that efficiency maximization is achieved by virtue of the fact that adjustments of signals to the different output paths may be done without losing signal power. In the prior art signal taken away is lost as useless heat, which reduces efficiency. Linearity maximization is achieved by virtue of the fact that the driver resource in the driver amplifier stage 509, 510 is pooled. The linearity of the Doherty amplifier is limited, ideally, only by the linearity of the carrier amplifier, and its driver amplifier stage 509, 510 performance. In the invention, the driver resource is shared. Therefore the signal directed to the carrier amplifier 520 is handled at a greater level of back-off in the driver amplifier stage 509, 510, and therefore is supplied to the carrier amplifier 520 with improved linearity.

These two advantageous effects may be achieved simultaneously in the Doherty amplifier of the invention. In the prior art, only one quantity is optimization.

The adjustment step of the Doherty PA 500 will be described. The said adjustment of phase delay elements 521 522 is the first adjustment of the Doherty PA 500 to be made. There are a number of additional adjustment steps that are necessary for optimum Doherty PA 500 operation.

The adjustment step is to adjust the fine phase setting stage 530 such that equal power sharing is achieved in the driver amplifier stage 509 510. The most straightforward way this can be performed by ensuring that the two-path RF power gains between RF input 501 and the outputs of driver amplifier stage 509 510 or isolators 511 512 are equal by adjusting the fine phase difference setting stage 530.

The next adjustment step is to adjust the gain difference setting stage 531 such that the desired power gain difference is achieved between the inputs of the peaking amplifier 518 and carrier amplifier 520. The most straightforward way this can be performed by measuring the RF gains between RF input 501 and the outputs of the third hybrid or coupler 513, or adjusting the gain difference setting stage 531 and phase delay elements 514 515 until the desired gain difference is achieved. Typically the power delivered into the peaking amplifier 518 will need to be higher since it is typically operated in class C but this power difference requirement also depends on the type, class and sizes of the RF power amplifying device used in peaking amplifier 518 and carrier amplifier 520. The present invention offers the benefit that different power requirements of any peaking amplifier 518 and carrier amplifier 520 can be readily accommodated. The final adjustment procedure is the coarse phase difference setting stage 532 such that signals arrive in-phase at the Doherty combining point 541 and therefore add constructively. One approach is to measure the phase shift from the RF input 501 to the point 540 and compare it with the phase at point 541 with the output Doherty combiner board removed. This might require an elevated RP input power in order to ensure that the peaking amplifier is turned on or alternatively, the bias of the peaking amplifier can be increased to turn it on instead. The phase difference between these two points needs to be 90° in order to make up the difference caused by the λ/4 Doherty impedance inverter. This said procedure should be sufficient for achieving a coarse adjustment without considering the influences of AM/PM distortion.

As a final test, the AM/ AM and AM/PM characteristics can be examined. Adjusting the phase adjusters 505 507 will allow trade-offs between linearity and efficiency to be made in a very straightforward manner compared to prior-art solutions. Furthermore, if said phase adjusters 505 507 are electrically adjustable, these settings can be adjusted by a controller or linearization system.

Fig 6 shows another embodiment of input processing stage. An RF input signal from an RF source having characteristic impedance (e.g. 50Ω) is applied at input 6501 of the input processing stage 6550. This input signal is divided into two parts by 90° hybrid or coupler 6502. A termination 6503 is provided for proper matched operation of the 90° hybrid or coupler 6502. The two-path signals obtained from the hybrid or coupler 6502 are ideally 90° apart from each other. The hybrid or coupler 6502 can be replaced with an input signal divider which can also be of Wilkinson type in which case the phase difference will be 0°. The two-path outputs from 90° hybrid or coupler 6502 are next applied to a gain phase adjuster stage 603. The purpose of gain phase adjuster stage 603 is to apply a gain/loss and phase difference between the two-path signal such that said signals have a desired gain difference and 0° or 180° phase difference. In the figure shown, gain phase adjuster stage 603 comprises, a gain difference setting attenuator 601, and a fine phase difference setting phase shifter 602 which is connected to one input of a hybrid or coupler 6506, and a phase delay element 604 in another path between 90° hybrid or coupler 6502 and the hybrid or coupler 6506. Wherein the gain difference setting attenuator 601 is connected with the fine phase difference setting phase shifter 602. Of course, the desired gain and phase differences also can be achieved by a number of alternative means e.g. I Q multipliers.

It is again important to maintain 0° or 180° at the input of hybrid or coupler 6506 since it means that the signals coming out of hybrid or coupler 6506 are equal in magnitude. This in turn, will allow the power sharing property to be exploited.

There are three embodiments in Driver combing stage, which are shown in figure 7, 8, 9. In fig 7, the driver combining stage 7551 is only comprised of a hybrid or coupler 7513. The key feature of this alternative embodiment is the absence of isolators which in turn results in so-called power recycling. Namely, power reflected back from the peaking amplifier (typically when it is biased off) will not be isolated but instead reflect back towards the driver amplifier stage and then ultimately to the carrier amplifier. This will have result of altering the power delivered to the carrier amplifier 520 as function of driver power thereby affecting linearity. The coarse phase difference setting stage can be used to control how these reflections affect the linearity of the overall Doherty PA.

In FIG 8, the driver combining stage 8551 is comprised of a hybrid or coupler 8513 which in turn is connected to isolators 801 and 802 respectively. The main advantage of such a configuration is the ability to adjust away unpredictable loss variations between the isolators.

In FIG 9, the driver combining stage 9551 is comprised of a hybrid or coupler 9513 which in turn is connected to isolator 902 and phase delay element 901 respectively. The main advantage of such a configuration is the reduction of the number of isolators to one. The phase delay element 901 is required however to compensate for the additional delay that the isolator may introduce. Of course, having two isolators shown in fig 5 is a straightforward way of equalizing out the delay.

Besides the above embodiment, there are a number of other means about the invention. For example, phase adjuster 507 and 505 is replaced with a gain adjuster, while one hybrid or coupler can be removed; Fixed Wilkinson input is spitted (i.e. no possibility of fine phase adjustment) with a phase adjuster to control gain; the PA can comprises N-path quasi Doherty amplifiers or Doherty amplifiers with M-way final outputs (e.g. Antenna diversity); phase adjusters are replace with fixed lines; different modes of PA operation, e.g. quasi- linear class F (class AB-C with harmonic terminations).

Claims

1. A quasi-linear high efficiency power amplifier comprises a Doherty stage comprising a carrier amplifier and a peaking amplifier, and a Doherty output combining network which is connected to outputs of the carrier and peaking amplifiers; and a driver amplifier stage being capable of supplying the Doherty stage with drive signals; wherein the said power amplifier further comprises, an input processing stage for receiving an input signal, maintaining equal power sharing in and supplying power to the driver amplifier stage with controllable phase differences; a driver power combining stage for receiving and combining the output signals from the driver amplifier stage, converting phase differences to amplitude differences, and supplying requisite and independently controllable drive signals to the Doherty stage.

2. The quasi-linear high efficiency power amplifier according to claim 1, the input processing stage comprises, a first hybrid or coupler having one input, used for dividing the input signal into a plurality of signals with the same amplitude and a relative phase difference of 90°; a termination for proper matched operation of the first hybrid or coupler; a fine phase difference setting stage, used for accepting signals from the first hybrid or coupler and applying a phase difference between the signals to produce a nominal phase difference of 0° or 180°; a second hybrid or coupler for converting phase differences from the nominal to nominal amplitude differences; a gain difference setting stage, accepting signals from the second hybrid or coupler, wherein, a phase difference is applied to produce signals having a desired phase difference, for application to the driver amplifier stage.

3. The quasi-linear high efficiency power amplifier according to claim 1, the said input processing stage comprises, a first hybrid or coupler having one input, used for dividing the input signal into a plurality of signals with the same amplitude and a relative phase difference of 90°; a termination for proper matched operation of the first hybrid or coupler; a gain phase adjuster stage, used for accepting signals from the first hybrid or coupler, and applying a gain and phase difference to the signals to produce a desired gain difference and a nominal phase difference of 0° or 180°; a second hybrid or coupler, used for converting phase differences from the nominal to nominal amplitude differences and vice versa, and applying to the driver amplifier stage.

4. The quasi-linear high efficiency power amplifier according to claim 2, the fine phase difference setting stage comprises a first phase adjuster in one path, and a first phase delay element in the other path; the gain difference setting stage comprises a second phase adjuster in one path, and a second phase delay element in the other path.

5. The quasi-linear high efficiency power amplifier according to claim 3, the gain phase adjuster stage comprises a gain difference setting attenuator which is connected with a fine phase difference setting phase shifter in one path and a phase delay element in the other path.

6. The quasi-linear high efficiency power amplifier according to claim 1,2 or 3, the power amplifier further comprises a coarse phase difference setting stage for facilitating a desired phase difference between the signals that are delivered to the Doherty stage, which is connected between outputs of the driver combining stage and inputs of the Doherty stage.

7. The quasi-linear high efficiency power amplifier according to claim 6, the said driver combining stage comprises a hybrid or coupler used for converting phase differences from the nominal to nominal amplitude differences and vice versa, and whose inputs are connected to outputs of the driver amplifier stage and outputs are connected to inputs of the coarse phase difference setting stage.

8. The quasi-linear high efficiency power amplifier according to claim 6, the said driver combining stage comprises isolators for absorbing power that is reflected back from the input of the peaking amplifier as it turns on and off in accordance with Doherty operation, which are respectively connected to outputs of the driver amplifier stage, and a hybrid or coupler used for converting phase differences from the nominal to nominal amplitude differences and vice versa, and whose inputs are connected to outputs of the isolators and outputs are connected to inputs of the coarse phase difference setting stage.

9. The quasi-linear high efficiency power amplifier according to claim 6, the said driver combining stage comprises a hybrid or coupler used for converting phase differences from the nominal to nominal amplitude differences and vice versa, and whose inputs are connected to outputs of the driver amplifier stage, and isolators for absorbing power that is reflected back from the input of the peaking amplifier as it turns on and off in accordance with Doherty operation, which are connected the outputs of the hybrid or coupler and inputs of the coarse phase difference setting stage.

10. The quasi-linear high efficiency power amplifier according to claim 6, the said driver combining stage comprises a converting phase differences from the nominal to nominal amplitude differences and vice versa, and whose inputs are connected to outputs of the driver amplifier stage, and an isolator for absorbing power that is reflected back from the input of the peaking amplifier as it turns on and off in accordance with Doherty operation, which is connected between one output of the hybrid or coupler and one input of the coarse phase difference setting stage, and a phase delay element compensating for the additional delay that the isolator may introduce, which is connected between the other outputs of the hybrid or coupler and the other output of the coarse phase difference setting stage.

11. The quasi-linear high efficiency power amplifier according to claim 6, the said coarse phase difference setting stage comprises phase delay elements for introducing the coarse phase difference required by the Doherty stage.

12. The quasi-linear high efficiency power amplifier according to claim 1, 2 or 3, the power amplifier further comprises phase delay elements for adjusting that the direction of desired load impedance changes is correct for Doherty operation, which are respectively connected between outputs of the Doherty stage and inputs of the Doherty output combining network.