CN1249912C - High-efficiency modulating RF amplifier - Google Patents

Abstract

The present invention provides for high-efficiency power control of a high-efficiency (e.g. hard-limiting or switch-mode) power amplifier. The spread between a maximum frequency of the desired modulation and the operating frequency of a switch-mode DC-DC converter is reduced by following the switch-mode converter with an active linear regulator. The linear regulator controls the operating voltage of the power amplifier with sufficient bandwidth to reproduce the desired amplitude modulation waveform. The linear regulator rejects variations on its input voltage even while the output voltage is changed in response to an applied control signal. Amplitude modulation is achieved by varying the operating voltage on the power amplifier. High efficiency is enhanced by allowing the switch-mode DC-to-DC converter to also vary its output voltage such that the voltage drop across the linear regulator is kept at a low and relatively constant level.

Description

High Efficiency Modulation RF Amplifier

technical field

This invention relates to RF amplifiers and signal modulation.

Background technique

In wireless communication devices such as cellular telephones, pagers, wireless modems, and the like, battery life is of great concern. In particular, radio frequency transmission consumes a large amount of power. Contributing to this power dissipation is inefficient power amplification operation. Typical RF power amplifiers for wireless communications operate at about 10% efficiency. Undoubtedly, low-cost techniques to significantly increase the efficiency of amplifiers will satisfy the above pressing needs.

Additionally, most modern digital wireless communication devices operate on packets. That is, the transmitted information is sent in short bursts of one or more, with the transmitter active only during the burst time and inactive at all other times. Therefore, control of pulse firing and cessation in an energy-efficient manner is required in order to further extend battery life.

Divide power amplifiers into different groups: Class A, Class B, Class A and B, etc. Different kinds of power amplifiers usually represent different bias conditions. When designing RF power amplifiers, there is usually a trade-off between linearity and efficiency. Different kinds of amplifier operation offer designers different ways to balance the above two parameters.

In general, power amplifiers are divided into two different categories: linear and nonlinear. Linear amplifiers (eg, Class A and Class B push-pull amplifiers) maintain a high degree of linearity, resulting in a faithful reproduction of the input signal at their output because the output signal is linearly proportional to the input signal. In nonlinear amplifiers (eg, single-ended Class B and Class C amplifiers), the output signal is not directly proportional to the input signal. The resulting amplitude distortion on the output signal makes these amplifiers ideal for signals without amplitude modulation (also known as constant envelope signals).

Amplifier output efficiency is defined as the ratio between RF output power and input (DC) power. The main reason for the inefficiency of power amplifiers is the power dissipated in the transistors. Class A amplifiers are inefficient because current is continuously passed through the device. Conventionally, linearity has been traded off for efficiency to increase efficiency. For example, in a Class B amplifier, the bias conditions are chosen such that the output signal is cut off for half the cycle unless the second transistor provides the other half (push-pull). Therefore, the waveform is less linear. By filtering the higher and lower frequency portions with a tank or other filter, the output waveform can still be sinusoidal.

A Class C amplifier conducts less than 50% of the period to further improve efficiency; ie, if the conduction angle of the output current is less than 180 degrees, the amplifier is called a Class C amplifier. The efficiency of the above modes of operation is greater than that of a Class A or B amplifier, but generally its distortion is greater than that of a Class A or B amplifier. In the case of a Class C amplifier, when changing the input amplitude, there is still a change in the output amplitude. The reason for this is that the Class C amplifier acts as a constant current source—even a brief current source—rather than a switch.

Other kinds of amplifiers advantageously address power dissipation within transistors by using only transistors as switches. The rationale for this type of amplifier is that in theory the switches consume no power because either zero voltage or zero current passes through. Because the V-I product of the switch is always zero, there is no dissipation in the device. Class E power amplifiers use one transistor, whereas Class D power amplifiers use two transistors.

In reality, however, switches aren't perfect. (Switches have on/off times and on-resistance.) The associated dissipation reduces efficiency. Accordingly, the prior art has sought to modify so-called "switching" amplifiers (in which transistors are driven as switches at the frequency of operation so that the power dissipated when the transistors conduct current is minimized) so that non-transitory switching occurs instantaneously. The switching voltage is zero during the zero time interval, thereby reducing power dissipation. Class E amplifiers use a reactive output network that provides enough freedom to shape the switching voltage to have zero value and zero slope when the switch is on, reducing switching losses. Class A amplifiers are a further class of switching amplifiers. Class A amplifiers generate a more square output waveform than a normal sine wave. "Squaring" of the output waveform is achieved by encouraging the generation of homogeneous harmonics (ie, x3, x5, x7, etc.) and suppressing the generation of even-order harmonics (ie, x2, x4, etc.) in the output network.

Figure 1 shows an example of a known power amplifier used in a cellular telephone. For example, a GSM cellular phone must be able to program output power over a range of 30dBm. In addition, the switching on and off of the transmitter must be accurately controlled to prevent spurious launches. Power is directly controlled by the cell phone's DSP (Digital Signal Processor), via a DAC (Digital to Analog Converter). In the circuit of Figure 1, signal GCTL drives the gate of the external AGC amplifier, which controls the RF level to the power amplifier. Feedback section output via directional coupler for closed loop operation. The amplifier of Figure 1 is not a switching amplifier. On the contrary, the amplifier is at best a class AB amplifier going into saturation, thus proving relatively low efficiency.

Figure 2 shows an example of a known Class E power amplifier as described in US Patent 3,919,656. The RF input signal is connected via conductor 1 to an exciter 2 which controls an active device 5 via a signal connected via conductor 3 . In essence, the active device 5 acts as a switch when the actuator 2 is properly activated. Therefore, the output port of the active device is represented as a SPST switch 6 . The DC power source 7 and the input port of the load network 9 are combined in series via a switch 6 . Connect the output port of load network 9 to load 11 . When the switch 6 is cycled according to the desired AC output frequency, at the switching frequency (and its harmonics), the DC energy of the source 7 is converted to AC energy.

US Patent 3,900,823 to Sokal et al. describes feedback control of a Class E power amplifier. The need for feedback control implies an inability to fully characterize the device, which in turn implies a practical departure from the operation of the device as a true switch. Sokal also demonstrated a solution to the feedthrough power control problem at low power levels: by applying negative feedback techniques to control the RF input drive level to control one or more pre-stage DC supplies. Feedback control is required to impose feedback loop dynamic constraints on the system.

Although theoretically the Class E amplifier structure of Figure 2 is capable of high conversion efficiency, it has the disadvantage of large voltage swings at the output of active devices due to damped oscillations. The large voltage swing, typically more than 3 times the supply voltage, prevents the use of Class E circuits with certain active devices that have low breakdown voltages.

In order to operate a switched RF power amplifier, the output transistors must be driven rapidly between off and fully on and then back off in a repetitive fashion. The means to achieve the above fast switching depends on the type of transistor selected as the switch: for field effect transistors (FETs), the control parameter is the gate source voltage, while for bipolar transistors (BJT, HBT), the control parameter is the base emission pole current.

However, the drive circuit in the RF amplifier of Figure 2 typically includes a matching network consisting of a tuned (resonant) circuit. Referring to Figure 3, in this type of configuration, the RF input signal is connected to the driver amplifier, typically Class A operation. The output signal of the driver amplifier is connected to the control terminal of the switching transistor (represented as FET in Figure 3) through a matching network. As with the design of the load network shown in Figure 2, properly designing the matching network is not an easy task.

Various designs attempt to improve different aspects of the basic Class E amplifier. One design is illustrated in Physically Analytical Model-Based FET Class E Power Amplifiers—Maximum PAE Design by Choi et al., IEEE Transactions on Microwave Theory and Technology, Vol. 47, No. 9, September 1999. This literature simulates various non-ideal FET switches, and according to the model, draws conclusions that are beneficial to the design of Class E amplifiers. For selected topologies, a maximum power added efficiency (PAE) of approximately 55% occurs at power levels below 0.5W. At higher powers, the PAE is automatically reduced, eg, less than 30% at 2W.

Set the PAE of the power amplifier based on the amount of DC power required to achieve the final 26dB of gain required to achieve the final output power. (At these gain levels, the power supplied to the amplifier by the excitation signal—not easily measured—becomes insignificant.) Currently, there are no known amplifying devices capable of generating more than 1W of output power at RF and providing at least 26dB of power gain. Therefore, one or more amplifiers must be provided before the final stage, and the DC power consumed by such amplifiers must be included in determining the overall PAE.

Conventional design practice requires the amplifier designer to impedance match the driver output impedance to the input impedance of the final switching transistor. Thus, the actual output power required by the driver stage is defined in terms of the voltage (or current) required into the (usually lower) effective deep impedance of the switching components. The specific impedance of the input to a switching transistor is not definable, since the impedance concept requires linear operation, whereas switches are non-linear.

Fig. 4 shows an example of an RF amplifier circuit according to the method described above. An interstage "T-section" consisting of inductor L1, bypass capacitor C, and inductor L2 is used to match the drive stage to an assumed 50 ohm load (ie, the final stage).

The above conventional practice regards the interstage between the excitation stage and the final stage as a linear network, while the interstage is not a linear network. Additionally, conventional practice maximizes power transfer between the driver stage and the final stage (a corollary of impedance matching). So, for example, to generate the drive voltage required for a FET acting as a switching transistor, the driver must generate an in-phase current in order to provide impedance-matched power.

Fig. 5 shows another example of a conventional RF power amplifier circuit. The circuit uses "resonant interstage matching" where the excitation stage and final stage are connected using a coupling capacitor Ccpl.

As noted above, conventional practice fails to achieve high PAE at high output powers (eg, 2W, a power level typically encountered when operating cellular telephones). Therefore, there is a need for RF power amplifiers that exhibit high PAE at higher output powers.

Control of the output power of an amplifier has been consistently shown to require a feedback structure, as exemplified by Sokal et al. and in the following US Patents: 4,392,245, 4,992,753, 5,095,542, 5,193,223, 5,369,789, 5,697,072, and 5,697,074. Other references, such as US Patent 5,276,912, teach controlling amplifier output power by varying the amplifier load circuit.

A related problem is generating modulated signals, such as amplitude modulated (AM) signals, quadrature amplitude modulated signals (QAM) and the like. Figure 6 shows a known IQ modulation structure. The data signal is applied to a quadrature phase modulation encoder that generates I and Q signals. The I and Q signals are applied to a quadrature phase modulator along with a carrier signal. The carrier signal is generated by a carrier generation block, where the tuning signal is applied to the carrier generation block.

Typically, the output signal of the quadrature phase modulator is applied to a variable attenuator controlled according to a power control signal. In other examples, power control is achieved by changing the gain of the amplifier. This is achieved by adjusting the bias voltage of the transistors within the linear amplifier, taking advantage of the effect that the transconductance of the transistors changes with the applied bias conditions. Since amplifier gain has a strong relationship to the transconductance of a transistor, changing the transconductance will effectively change the amplifier gain. The resulting signal is amplified with a linear power amplifier and then applied to the antenna.

In an AM signal, the amplitude of the signal is made roughly proportional to the amplitude of an information signal such as sound. In essence, information signals such as sound are not constant, so the resulting AM signal is constantly changing in output power.

As early as 70 years ago, textbooks such as Terman's Radio Engineer's Handbook (McGraw-Hill, 1943) had explained the use of nonlinear Class C amplifiers to generate accurately amplitude modulated signals, called "plate modulation". In a typical plate modulation technique, the output current of the modulating amplifier is linearly added to the supply current of the amplifying element (tube or transistor), so that the supply current increases and decreases from the average value according to the amplitude modulation. The variable current causes the apparent supply voltage across the amplifying element to vary according to the resistive (or conductive) characteristic of the amplifying element.

By directly controlling the output power, AM can be achieved as long as the bandwidth of the variable operating voltage is sufficient. That is, the above-described nonlinear amplifier actually acts as a linear amplifier with respect to the amplifier operating voltage. The output signal can be linearly modulated in the sense that the operating voltage can be changed at any time when driving the nonlinear power amplifier.

Other methods of achieving amplitude modulation include combining many constant amplitude signals, as described in US Pat. US Patents 4,896,372, 3,506,920, 3,588,744 and 3,413,570 describe methods of amplitude modulation using pulse width modulation to vary the power supply of a power amplifier. However, the above-mentioned patent considers that the operating frequency of the switching DC-DC converter must be much higher than the highest modulation frequency.

Applicant's US Patent 5,126,688 to Nakanishi et al. addresses the control of nonlinear amplifiers by using feedback control to set the actual amplifier output power in combination with periodic adjustments of the power amplifier's operating voltage to improve the operating efficiency of the power amplifier. The main disadvantage of this technique is that additional control circuitry is required to read the desired output power, to determine if (or not) the power amplifier operating voltage needs to be changed to improve efficiency, and to make the change if necessary. Additional control circuitry adds complexity to the amplifier and draws additional power beyond that of the amplifier itself, which directly reduces overall efficiency.

Another challenge is generating high power RF signals with the desired modulation characteristics. This object is achieved according to the teaching of the applicant's US Patent 4,580,111 to Swanson by using a number of high efficiency amplifiers providing fixed output powers, wherein the amplifiers are activated sequentially so that the total combined output power required is the output power of the individual fixed power amplifiers multiples of . In this approach, the minimal change in overall output power is substantially equal to the power of each amplifier of the plurality of high efficiency amplifiers. If finely graded output power resolution is required, a large number of highly efficient amplifiers may be required. This undoubtedly increases the overall complexity of the amplifier.

US Patent 5,321,799 performs polar modulation, but is limited to full response to data signals and cannot be used with high power, high efficiency amplifiers. The patent considers the application of amplitude variations to the modulated signal through a digital amplifier and signal generation stage following phase modulation. The final analog signal is then generated by using a digital-to-analog converter. As explained in the state of the art, signals with information realized in amplitude variations are incompatible with high-efficiency, non-linear power amplifiers since signal amplitude variations can be severely distorted.

Despite the technical descriptions of the above references, there are still many problems to be solved, including: Achieving high efficiency of RF signals by changing the operating voltage using switching converters without requiring high-efficiency switching operation (compared to frequency modulation) Amplitude modulation; unified power level and burst control with modulation control; enabling efficient modulation of any desired characteristic (amplitude and/or phase); and enabling high-power operation without sacrificing power efficiency (e.g., for operation of base stations) .

Contents of the invention

In general, the present invention provides efficient power control of efficient (eg, hard-limited or switched) power amplifiers in a manner that achieves the desired control or modulation. Unlike the prior art, no feedback control is required. That is, the amplifier can be controlled without continuous or frequent feedback adjustments. In one embodiment, by emulating a switching converter with an active linear regulator, the difference between the highest frequency of modulation required and the operating frequency of the switching DC-DC converter is reduced. The goal of designing a linear regulator is to control the operating voltage of the power amplifier with enough bandwidth to faithfully reproduce the desired amplitude-modulated waveform. Another purpose of designing a linear regulator is to reject changes in the input voltage even though the output voltage is changed in response to an applied control signal. This suppression will occur even if the input voltage changes at a frequency comparable to or even lower than the controlled output's change. Amplitude modulation is achieved by directly or effectively changing the operating voltage of the power amplifier, while efficiently converting the original DC supply to an amplitude modulated output signal. Efficiency is improved by allowing a switching DC-DC converter to vary its output voltage so that the voltage drop across the linear regulator is maintained at a relatively constant low level. It is possible to combine time division multiple access (TDMA) burst capability with effective amplitude modulation, provided that the combined functions described above are controlled. Additionally, changes in the average output power level consistent with commands from a certain communication system can be combined within the same structure.

The efficient AM structure can be extended to arbitrary modulations. Modulation is performed in a polar manner, ie in a manner that does not require quadrature modulation.

Many individual high-efficiency stages can be combined to form high-power, high-efficiency modulation structures.

According to one aspect of the present invention, there is provided a variable output radio frequency power amplifier, which is characterized in that it includes: a voltage regulating device, according to a control signal for performing level control, burst control and modulation for one of them, generating a specific voltage within a voltage range, said voltage regulator comprising: a switching converter stage and a linear regulator stage; and a power amplifier comprising: a final amplifier stage, the final amplifier stage Having said specific voltage as the supply voltage, and also having an excitation signal which enables said final amplifier stage to be excited repeatedly between two states - a hard on state and a hard off state - without the need for Operating the amplifier in a linear region of operation within an appreciable amount of time; wherein said amplifier is controlled without the need for continuous or frequent feedback adjustments.

According to another aspect of the present invention, there is provided a variable output radio frequency power amplifier, which is characterized in that it includes: a voltage regulating device, which is used to generate A specific voltage within a voltage range, the voltage regulating device comprising: a switching converter stage and a linear regulator stage; and a power amplifier comprising: a final amplifier stage having A specific voltage is used as the supply voltage and also has an excitation signal that repeatedly excites the final amplifier stage between two states - hard on state and hard off state - without being online for an appreciable amount of time operating an amplifier in a non-volatile region of operation; wherein said amplifier is controlled without the need for continuous or frequent feedback adjustments.

According to another aspect of the present invention, there is provided a method for controlling a power amplifier, which is characterized in that it includes: generating a specific voltage according to a control signal that performs at least one of level control, pulse train control and modulation; as a power amplifier the supply voltage for the final amplification stage, applying a specific voltage to the power amplifier; and the ability to switch between two states—hard on and hard off—without operating the amplifier in the linear region of operation for an appreciable amount of time repeatedly excites the final amplifier stage; wherein the amplifier is controlled without the need for continuous or frequent feedback adjustments.

Description of drawings

The invention will be better understood from the following description read in conjunction with the accompanying drawings. The accompanying drawings are:

Figure 1 is a block diagram of a known power amplifier that controls output power by varying the supply voltage;

Figure 2 is a simplified block diagram of a known single-ended switch-mode RF amplifier;

Figure 3 is a schematic diagram of a portion of a known RF amplifier;

4 is a schematic diagram of a conventional RF power amplifier circuit;

5 is a schematic diagram of another conventional RF power amplifier circuit;

Figure 6 is a block diagram of a known IQ modulation structure;

FIG. 7 is a block diagram of a power amplifier according to certain exemplary embodiments;

Figure 8 compares the output power of a saturated Class A and B PA with the mathematical model as a function of operating voltage $V=\sqrt{\mathrm{PR}}.$

Figure 9 is a waveform illustrating the operation of an embodiment;

Figure 10 is a waveform showing the operation of another embodiment;

Figure 11 is a waveform representing burst AM operation;

Figure 12 is a waveform showing burst AM operation with power level control;

Figure 13 is a block diagram of a polar modulation architecture using a high-efficiency amplifier;

Figure 14 is a block diagram of a first high power, high efficiency amplitude modulated RF amplifier;

Figure 15 is a waveform showing the operation of the amplifier of Figure 14;

Figure 16 is a block diagram of a second high power, high efficiency AM RF amplifier;

Figure 17 is a waveform showing the operation of the amplifier of Figure 16;

FIG. 18 is a block diagram of an RF switching amplifier according to an embodiment;

Figure 19 is a schematic diagram of a portion of an RF switching amplifier according to an embodiment of the present invention;

FIG. 20 is a schematic diagram of a suitable load network for use in the RF switch-mode amplifier of FIG. 19;

Fig. 21 is a waveform showing the input voltage and related waveforms of the RF switching amplifier of Fig. 19;

Figure 22 is a waveform representing the base and collector current waveforms of the switching transistor of Figure 19;

Figure 23 is a waveform representing the output voltage of the RF switching amplifier of Figure 19;

24 is a schematic diagram of a portion of an RF switched-mode amplifier according to another embodiment;

Figure 25 is a waveform showing the input voltage and related waveforms of the RF switching amplifier of Figure 24;

Figure 26 is a waveform showing the collector current waveform of the drive transistor of Figure 24;

Figure 27 is a waveform representing the gate voltage waveform of the switching transistor of Figure 24;

28 is a schematic diagram of an RF power amplifier circuit according to another embodiment; and

FIG. 29 is a waveform showing waveforms occurring at selected nodes of the amplifier circuit of FIG. 28. FIG.

Detailed ways

Reference is now made to FIG. 7, which is a schematic diagram of a power amplifier that overcomes many of the above-mentioned disadvantages. A switched (or saturated) non-linear amplifier has applied to it the voltage generated by the power control stage. In a typical implementation, essentially according to the formula

$\mathrm{<span\; class="notranslate">\; V</span>}<span\; class="notranslate">\; =</span>\sqrt{\mathrm{<span\; class="notranslate">\; PR</span>}}$

Controls the voltage V applied to the nonlinear amplifier, where P is the desired power output level of the amplifier and R is the resistance of the amplifier. In the case of a switching or saturated amplifier, the resistance R can be considered constant. The power control stage receives a DC input voltage (for example, from a battery), receives a power level control signal, and outputs a voltage according to the above equation.

The efficiency of directly controlling the output power of a nonlinear amplifier over a wide dynamic range by varying the operating voltage alone is illustrated using Figure 8, which compares the output power of a saturated Class AB PA as a function of operating voltage with a mathematical model $V=\sqrt{\mathrm{PR}}.$

Referring again to FIG. 7, this figure shows a power control circuit according to an exemplary embodiment. The power control circuit consists of a switching converter stage and a linear regulator stage connected in series. The switching converter can be a Class D device, for example, or a switched mode power supply (SMPS). The switching converter effectively steps down the DC voltage to a voltage slightly above but close to the desired power amplifier operating voltage level. That is, the switching converter performs effective overall power level control. Switching converters may or may not provide sufficiently fine control to define the ramp portion of the desired power envelope.

A linear regulator performs a filtering function on the output of a switching converter. That is, the linear regulator controls precise power envelope modulation during TDMA bursts. Like switching converters, linear regulators may or may not provide level control capability.

Note that depending on the speed of the switching converter and linear regulator, power control and/or amplitude modulation can be performed using the power control circuit. The control signal PL/BURST/MOD is input into the control block which outputs suitable analog or digital control signals for the switching converter and linear regulator. This control block can be implemented in the form of a ROM (Read Only Memory) and/or a DAC (Digital to Analog Converter).

Referring to FIG. 9, FIG. 9 is a waveform showing operation according to one embodiment of the present invention. Waveforms A and B represent the analog control signals applied to the switching converter and linear regulator, respectively. Waveforms _V1 and _V2 represent the output voltages of the switching converter and linear regulator, respectively. It is assumed that the switching converter has a relatively large time constant, ie a slow ramp. When the control signal A is set to a first non-zero power level, the voltage _V1 will ramp towards the commensurate voltage. Due to the switching nature of the converter, the voltage V ₁ may have large fluctuations. The amount of time required to reach the desired voltage defines the wake-up period. When this voltage is reached, control signal B is raised or lowered to define a series of transmit bursts. When the control signal B is increased, the voltage _V2 ramps up quickly towards the corresponding voltage direction, and when the control signal B is decreased, the voltage _V2 ramps down rapidly. After a series of bursts (in this example), the control signal A is raised to increase the RF power level of the subsequent burst. During the waiting period, control signal B remains low. When voltage _V1 reaches a certain level, control signal B is raised or lowered to define other transmit bursts.

Voltage V ₂ is represented by a dotted line superimposed on voltage V ₁ . Note that voltage V ₂ is slightly smaller than voltage V ₁ and larger than the negative peak fluctuation on voltage V ₁ . The small difference between the input voltage _V1 of the linear regulator and the output voltage _V2 of the linear regulator makes all efficient operation possible.

Referring to FIG. 10 , according to various embodiments, it is assumed that the switching converter has a shorter time constant; ie, the ramp is faster. Therefore, when the control signal A is increased, the voltage _V1 ramps rapidly towards the commensurate voltage. When the control signal B is raised, the voltage _V2 is ramped. The time difference between raising the control signal B and raising the control signal A defines the wake-up time, which can be short to maximize sleep time and save power. When the transmit burst is finished, control signal B is lowered, after which control signal A is lowered. Following the example of Fig. 9, in Fig. 10, when the control signal A is raised next time, it defines a higher power level. Furthermore, the voltage V ₂ is superimposed on the voltage V ₁ in a dotted line.

In addition to power and burst control, amplitude modulation can be performed using the same structure. Referring to FIG. 11, this figure is a waveform showing burst AM operation. The output signal of the switching transition is shown in solid lines. At the beginning of the burst, the output signal of the switching converter ramps up. Alternatively, as shown by the dotted line, the switching converter ramps up towards a fixed level while the linear regulator achieves all the amplitude modulation on the output signal. From an efficiency standpoint, it is preferable that the switching converter implements amplitude modulation, thereby generating an output signal that is a small fixed offset ΔV above the desired output signal when noise is ignored. A linear regulator removes noise from the switching converter's output signal, effectively reducing the signal by ΔV. The dashed line in Figure 11 represents the output signal of the linear regulator. When the burst ends, the signal ramps down.

Maintain complete control of the output signal power level (average power of the signal). For example, as shown in Figure 12, subsequent bursts may occur at higher power levels. Compared to Figure 11, in Figure 12 all signals are properly scaled to achieve a higher average power output.

Although adding amplitude modulation to a phase-modulated signal complicates signal generation methods, it is often just what is needed, since the signal typically occupies less bandwidth than a purely phase-modulated signal. Reference is made to Figure 13, which is a block diagram of a polar modulation architecture using a high efficiency amplifier of the type described so far. This polar modulation structure enables any desired modulation. The data signal is applied to a modulation encoder, which generates an amplitude and phase signal. The phase signal is applied to a carrier generation block that supports phase modulation, while the tuning signal is applied to the generation block. The resulting signal is then amplified with a non-linear power amplifier of the type described above. Simultaneously, the amplitude signal is applied to the amplitude exciter. The amplitude exciter also receives a power control signal. In response, the amplitude driver generates an operating voltage that is applied to the nonlinear amplifier. As shown by the dotted line in Fig. 13, the amplitude driver and the nonlinear amplifier are realized in the same manner as shown in Fig. 7 explained above.

Among other applications, the modulation structures described so far are suitable for use in cellular telephone handsets. Efficient RF signal generation is also required in cellular telephone base stations. However, base stations operate at higher power than cell phones. The following structures are used to achieve high power, efficient RF signal generation.

Referring to FIG. 14, a first high power, high efficiency AM RF amplifier includes a number of switched mode power amplifier (SMPA) blocks, each implemented in the manner shown in FIG. 7, for example. Input the RF signal to be amplified into all common SMPA blocks. The amplitude exciter generates independent control signals for each SMPA block in response to the amplitude input signal. Sums the output signals of all SMPA blocks to produce a single composite output signal.

Specifically, the AM radio frequency amplifier shown in Figure 14 includes: a plurality of amplifier modules (five SMPA blocks as shown in Figure 14); a radio frequency signal (RF input (phase)) for the plurality of amplifier modules All of the RF power amplifiers are common and applied to these amplifier modules; and an amplitude driver. In addition, as shown in Figure 7, each amplifier module includes: a switching converter (SM converter shown in Figure 7); a linear regulator; an amplitude driver (not labeled in Figure 7); and a RF Power Amplifier (SMPA). The switching converter has: a power input (V _DC ); a power output (V ₁ ); and a control input (A). The linear regulator has: a power input terminal (V ₁ ); a power output terminal (V ₂ ) and a control input terminal (B). The power input (V ₁ ) of the linear regulator is coupled to the power output of the switching converter. Amplitude exciter (not numbered in Fig. 7) responsive to a modulation signal (PL/BURST/MOD) for generating a first control signal (A) coupled to the control input of the switching converter , is also used to generate a second control signal (B), which is coupled to the control input of the linear regulator. Each radio frequency power amplifier (SMPA) has a non-linear mode of operation. The power output terminal (V ₂ ) of the linear regulator provides the working voltage of the RF power amplifier. In this embodiment shown in FIG. 14, separate amplitude excitation signals are generated for each RF power amplifier, as shown by the respective vertical lines from the amplitude exciter to the five SMPA boxes. In another alternative embodiment shown in FIG. 16 and described in detail below, a single amplitude excitation signal is common to and applied to all RF power amplifiers, as shown in Fig. The amplitude exciter and its branches to each of the five SMPA boxes are shown as a single vertical line.

By referring to FIG. 15, the manner in which the amplifier of FIG. 14 operates can be understood. The left shows the full amplitude signal applied to the amplitude exciter. The right side shows the SMPA excitation signal applied to each SMPA output by the amplitude exciter. Note that the sum of the individual excitation signals produces the overall amplitude signal.

Fig. 16 shows another embodiment of a high power amplifier. In this embodiment, a common excitation signal is generated and then applied to each SMPA, instead of separately generating excitation signals for each SMPA. At a given moment, make the value of the common excitation signal one-Nth of all amplitude signals applied to the amplitude exciters, where N is the number of SMPAs. Fig. 17 shows the results. Also, note that the sum of the individual excitation signals produces the overall amplitude signal.

Referring now to FIG. 18, this figure shows a block diagram of an RF switching amplifier according to another embodiment. Apply an RF input signal to a non-reactive excitation circuit. Connect the excitation circuit to the active device to actuate the active device to switch. Connect an active device switch to a load network that generates an RF output signal that is applied to a load such as an antenna. A fast time-varying power supply, preferably implemented by combining a switching mode power supply and a linear regulator in series, applies power to an active device switch in order to vary the operating voltage of the active device switch. By varying the operating voltage in a controllable manner, power control, burst control and modulation can be achieved in the manner described above.

Active device switches can be bipolar transistors or FET transistors. Referring to Figure 19, this figure is a block diagram of a portion of an RF switching amplifier in which the active device switch is a bipolar transistor with collector, emitter and base terminals. Through an RF choke L, the collector of bipolar transistor N1 is connected to the operating voltage V _PA and to the output matching network. Connect the emitter of bipolar transistor N1 to circuit (AC) ground.

The base of bipolar transistor N1 is connected to the emitter of another bipolar transistor N2 (driver transistor) in a Darlington manner. The collector of drive transistor N2 is connected to operating voltage V _DRIVER and to a bypass capacitor. Associated with drive transistor N2 is a bias network which, in the embodiment shown, includes three resistors, R1, R2 and R3. Connect resistor R1 from the emitter of the drive transistor to circuit ground. Connect resistor R2 from the base of the drive transistor to ground. Connect resistor R3 from the base of drive transistor N2 to V _DRIVER . The RF input signal is applied to the base of the drive transistor through the DC blocking capacitor C _in .

Referring to Figure 20, the output network may take the form of an impedance matched transmission line TL and capacitor _Cout .

As shown in waveform 1 of Figure 21, the RF input voltage signal is a sine wave. As shown in waveform 2, the input voltage is shifted horizontally upwards to generate a voltage that drives the base of transistor N2. As shown in waveform 3, the emitter voltage of drive transistor N2 drops by _Vbe and is applied to the base of switch transistor N1. At the beginning of the positive half cycle, the driving transistor N2 acts as an emitter output amplifier, and its output (emitter) voltage is much lower than the turn-on voltage of the switching transistor N1, so the switching transistor N1 is turned off. As shown in Figure 22, when the signal increases, drive transistor N2 turns on switching transistor N1 and saturates it. As shown in Figure 23, the current passes through the RF choke L and the switching transistor N1, and when the capacitor C _out is discharged, the output voltage decreases. As the end of the positive half-cycle approaches, the output voltage of driver transistor N2 drops below the turn-on voltage of switching transistor N1, allowing it to turn off. The value of resistor R1 is chosen such that switching transistor N1 turns off quickly. The current continues to flow through the RF choke L, charging the capacitor C _out and increasing the output voltage.

Referring to Figure 24, which is a schematic diagram of a portion of an RF switching amplifier in which the active device switches are FET transistors (MESFET, JFET, PHEMT, etc.) with drain, source, and gate pinouts. Through the RF choke L1, the drain of the FET transistor M1 is connected to the operating voltage V _PA and at the same time to the output network. Connect the source of the FET transistor to circuit (AC) ground.

The gate of the FET transistor is biased from the supply -V _B through a large value resistor R1 and connected to a pair of bipolar transistors (driver transistors) connected in a push-pull configuration through a DC blocking capacitor C1. The driving transistors include an NPN transistor N1 and a PNP transistor P1. Connect the collector of the NPN driving transistor N1 to the operating voltage V _CC and to the bypass capacitor at the same time. Connect the collector of the PNP driver transistor to the negative reference voltage -V _B and also to the bypass capacitor. Connect the bases of the drive transistors in common. Large value resistors R2 and R3 are connected to the common node of each power supply rail.

Connect NPN bipolar transistor N2 in a common base configuration. The emitter of the bipolar transistor is connected to _-VB through resistor R4 and to the RF input signal through capacitor C3. Connect the collector of the bipolar transistor to V _CC through inductor L2 and also to the bypass capacitor.

Referring to FIG. 25 , there is shown input voltage waveforms 1-4 for the circuit of FIG. 24 . The input voltage 1 is horizontally shifted down by one _Vbe (generating voltage 2) and then applied to the emitter of bipolar transistor N2. Using the influence of inductor L2, a large voltage swing 3 is generated at the collector of bipolar transistor N2. The above voltage swing is shifted horizontally downward to generate a voltage 4, which is applied to the base of the drive transistor at node N. In operation, during the positive half cycle, bipolar transistor N2 is initially turned off. Current flows through inductor L2 into capacitor C2 connected to the bases of the transistor pair, thereby turning on NPN transistor N1 and turning off PNP transistor P1 (FIG. 26). DC blocking capacitor C1 is charged from the V _CC supply, which increases the gate potential of FET M1, turning it on (Figure 27). During the negative half cycle, bipolar transistor N2 is turned on. Current flows through inductor L2, through transistor N2, and to the -V _B rail. Current flows out of the base of PNP transistor P1, turning this transistor on. The DC blocking capacitor C1 discharges, thereby reducing the gate potential of FET M1, turning it off. The output network operates in the same way as above.

Referring now to FIG. 28, there is shown a schematic diagram of a multi-stage RF power amplifier circuit that may be used with the drive circuit described above. Use an input matching circuit consisting of capacitors C ₁ , capacitor C ₂ , and inductor L ₁ to set the input impedance of the circuit. The driver stage _M1 and the final stage _M2 are shown as FETs, although in other embodiments bipolar transistors could be used. The drain electrode of FET _M1 is connected to supply voltage _Vd1 through a drain bias network comprising RF choke _L3 and capacitor _C5 . Likewise, the drain electrode of FET _M2 is connected to supply voltage _Vd2 through a drain bias network comprising RF choke _L7 and capacitor _C10 .

A gate bias network is provided for stages _M1 and _M2 respectively. In the case of stage _M1 , the gate bias network consists of inductor _L2 , capacitor _C3 and capacitor _C4 connected at a common node to voltage _Vg1 . In the case of stage _M2 , the gate bias network consists of inductor _L6 , capacitor _C8 and capacitor _C9 connected at a common node to voltage _Vg2 .

The drive stage and the final stage are connected using an interstage network represented by a series LC combination of inductance _L4 and capacitance _C6 , the inductance and capacitance values being chosen so as to provide resonance with the input capacitance of the final stage _M2 . Connect the final stage _M2 to the conventional load network, in this example represented as a CLC Pi network composed of capacitor _C11 , inductor _L8 and capacitor _C12 , wherein the capacitance and inductance values are determined according to the characteristics of the final stage _M2 .

In a typical implementation, the component values are as follows, where capacitance is measured in picofarads and inductance is measured in nanohenries:

Table 1 capacitance pf inductance no Voltage V C ₁ 27 L ₁ 8.2 V _d1 3.3 C ₂ 10 L ₂ 33 V _d2 3.2 C ₃ 0.01 L ₃ 33 V _g1 -1.53 C ₄ 27 L ₄ 4.7 V _g2 -1.27 C ₅ 27 L ₅ NA C ₆ 27 L ₆ 39 C ₇ NA L ₇ 15 C ₈ 27 L ₈ 2.7 C ₉ 0.01 C ₁₀ 27 C ₁₁ 1.5 C ₁₂ 5.6

In the example of FIG. 28 , the excitation stage (stage M ₁ ) is operated in switching mode. Referring to FIG. 29, this figure shows the input voltage of stage _M2 at node A, the drain voltage of stage _M1 at node B, the drain voltage of stage _M2 at node C, the drain current of stage _M1 at node D, and The waveform of the drain current of stage _M2 at node E. Note that the peak value of the gate voltage of the final stage—stage M ₂ (waveform A)—is much higher than in conventional designs. In the above configuration, the input drive of the switches is high enough to reduce the operating voltage of the driver stage. In this way, the DC power of the exciter can be further reduced, thereby improving PAE.

By using a circuit of the type shown, the measured PAE is 72% when the output power is 2W.

Thus, a power amplifier circuit structure including an excitation circuit and a multi-stage amplification circuit that can accurately generate a desired RF waveform with high power increasing efficiency without feedback is described.

Claims (21)

1. a variable output radio-frequency power amplifier is characterized in that, comprising:

Voltage regulating device, be used for the control signal of one of them according to being used to carry out level control, pulse train control and modulation, be created in a specific voltage in the voltage range, described voltage regulating device comprises: a switch mode converters level and a linear regulator level; With

A power amplifier, comprise: a final stage amplifying stage, this final stage amplifying stage has described specific voltage as supply voltage, also has a pumping signal, this pumping signal can two states-hard on-state and hard off-state-between make described final stage amplifying stage repeat to be energized, and need not in the appreciable time this amplifier of operation in the linear operation zone;

Wherein said amplifier is controlled under the situation that does not need continuous or frequent feedback adjustment.

2. a variable output radio-frequency power amplifier is characterized in that, comprising:

Voltage regulating device is used for being created in a specific voltage in the voltage range according to the control signal of carrying out one of level control, pulse train control and modulation at least, and this voltage regulating device comprises:

A switch mode converters level and a linear regulator level; With

A power amplifier, this amplifier comprises: a final stage amplifying stage, this final stage amplifying stage has specific voltage as supply voltage, also has a pumping signal, this pumping signal can two states-hard on-state and hard off-state-between this final stage amplifying stage of repeat actuation, and need not in the appreciable time operational amplifier in the linear operation zone;

Wherein said amplifier is controlled under the situation that does not need continuous or frequent feedback adjustment.

3. device according to claim 2 is characterized in that, also comprises:

A switch mode converters level, it is set for provides thick level control and comes work as described switch mode converters level; With

A linear regulator level, being set for provides the control of meticulous slope and comes work as described linear regulator level.

4. variable output radio-frequency power amplifier according to claim 3 is characterized in that, also comprises: the power amplifier of a hard restriction, come work as described power amplifier.

5. variable output radio-frequency power amplifier according to claim 4 is characterized in that, described power amplifier is a kind of regulex of selecting from Class A, Class B and class C amplifier group.

6. variable output radio-frequency power amplifier according to claim 3 is characterized in that, described power amplifier is a switching regulator amplifier.

7. variable output radio-frequency power amplifier according to claim 3 is characterized in that, described power amplifier is a class C amplifier.

8. variable output radio-frequency power amplifier according to claim 2 is characterized in that, also comprises: a switch mode converters level, it is set for provides level control and slope control, in order to come work as described switch mode converters level.

9. variable output radio-frequency power amplifier according to claim 2 is characterized in that, also comprises: a linear regulator level, it is set for provides slope control and level control, in order to come work as described linear regulator level.

10. variable output radio-frequency power amplifier according to claim 2, it is characterized in that, also comprise: be used to receive and respond the device of described control signal, be used to described one first control signal of switch mode converters level generation and be one second control signal of described linear regulator level generation.

11. variable output radio-frequency power amplifier according to claim 2, it is characterized in that, also comprise: an amplitude excitations device, respond a modulation signal, be used to described one first control signal of switch mode converters level generation and be one second control signal of described linear regulator level generation.

12. variable output radio-frequency power amplifier according to claim 2, it is characterized in that, also comprise: be used to respond a phase control signal and the device that produces a carrier signal, this carrier signal has the phase modulated characteristic, and this carrier signal is applied to described radio-frequency power amplifier.

13. variable output radio-frequency power amplifier according to claim 12 is characterized in that, also comprises: an amplitude control signal, in order to come work as described modulation signal, described radiofrequency signal is amplitude modulation.

14. variable output radio-frequency power amplifier according to claim 13 is characterized in that, also comprises: a modulating coder, respond a data-signal, be used to produce described amplitude control signal and described phase control signal.

15. variable output radio-frequency power amplifier according to claim 14 is characterized in that, also comprises: a modulating coder, it is set in polar coordinate system as described modulating coder work.

16. variable output radio-frequency power amplifier according to claim 2 is characterized in that, also comprises:

A plurality of amplifier modules, each amplifier module comprises:

A switch mode converters has: a power input, a power take-off and a control input end;

An adjuster has: a power input, a power take-off and a control input end, and the described power input of this adjuster is coupled to the described power take-off of described switch mode converters;

An amplitude excitations device, respond a modulation signal, be used to produce one first control signal, this first control signal is coupled the described control input end with described switch mode converters, also produce one second control signal, this second control signal is coupled to the described control input end of described adjuster; With

A radio-frequency power amplifier has a kind of nonlinear operation pattern, and the power take-off of described adjuster provides the operating voltage of described radio-frequency power amplifier;

A radiofrequency signal is shared for all described radio-frequency power amplifiers; With

An amplitude excitations device responds all amplitude signals, is used to produce one or more amplitude excitations signals, and this amplitude excitations signal application is in each radio-frequency power amplifier.

17. variable output radio-frequency power amplifier according to claim 16, it is characterized in that, also comprise: an amplitude excitations device, respond all amplitude signals, be used for respectively for each radio-frequency power amplifier produces an independent amplitude excitations signal, to work as described amplitude excitations device.

18. variable output radio-frequency power amplifier according to claim 16, it is characterized in that, also comprise: an amplitude excitations device, respond all amplitude signals, be used to produce a single amplitude excitations signal, this amplitude excitations signal is shared for all radio-frequency power amplifiers, to work as described amplitude excitations device.

19. the method for a control power amplifiers is characterized in that, comprising:

Control signal according to carrying out one of level control, pulse train control and modulation at least produces specific voltage;

Supply voltage as the final stage amplifying stage of power amplifier is applied to this power amplifier with specific voltage; And

Need not in the appreciable time this amplifier of operation in the linear operation zone, just can two states-hard on-state and firmly off-state-between this final stage amplifying stage of repeat actuation;

Wherein do not need continuously or situation that frequent feedback is adjusted under control amplifier.

20. method according to claim 19 is characterized in that, also comprises: radio-frequency input signals is applied to radio frequency amplifier, and wherein radio-frequency input signals is a phase modulated signal.

21. method according to claim 19 is characterized in that, also comprises:

In polar coordinates, data are encoded, so that produce an amplitude signal and a phase signal; With

According to phase signal, produce radio-frequency input signals;

Wherein said modulation signal is led according to described amplitude signal.

Images