MXPA97005617A - Distortion compensation control for a poten amplifier - Google Patents

Abstract

The present invention relates to a control facility and method for controlling a gain and phase adjuster used to control the gain and phase of a compensation signal to compensate for the distortion produced by a power amplifier. The control installation makes use of a reference signal, of known frequency, which is amplified together with the desired carrier signals. The reference signal component of the amplified signal is then isolated and a comparison is made, either with the actual reference signal, for example, by means of a QAM demodulator, or alternatively with the known frequency of the reference signal, for example , by means of an FM discriminator, in order to determine the gain and phase differences of the reference signal component of the amplified signal compared to the reference signal. For example, a QAM demodulator is used to produce a first difference signal indicative of the phase switching difference, and a second difference signal indicative of the gain switching difference. These signals are then correlated with the signal envelope of the main signal in order to produce signals indicative of the sign and magnitude of the control signals used to adjust the gain and phase adjuster in a feedback loop.

Description

DISTORTION COMPENSATION CONTROL FOR A POWER AMPLIFIER This invention relates to the control of a distortion compensation facility for a power amplifier, such as a multi-channel linear power amplifier for a cellular radio communications system. BACKGROUND OF THE INVENTION As is well known, an RF amplifier that operates at high power levels has non-linear amplifier characteristics that produces unwanted intermodulation distortion due to the interaction between the signals that are being amplified. In order to compensate for the non-linear characteristics, the prior art systems provide distortion compensation facilities in which a part of the incoming signal is subtracted, with an appropriate delay, and phase and gain adjustments, from a part of the amplified signal that is subject to the characteristics of the non-linear amplifier to isolate a distortion signal component. This distortion signal is a form of compensation signal used to predistort the main signal in order to compensate for the distortion before amplification. Other known systems utilize a "feed advance" installation where the compensation signal comprises an adjusted copy of the original signal that is used to compensate for the distortion after amplification. Prior art systems, for example, those conceived in U.S. Patent No. 4,879,519 issued November 7, 1989, which names Robert E. Myer as inventor, entitled "Linear Amplifier Compensated by Predistortion", and the U.S. Patent No. 4,885,551 issued December 5, 1989, naming Robert E. Myer as inventor, entitled "Linear Feed Advance Expander", recognizes the difficulty of controlling compensation because it requires modifying the gain and phase of the component. of compensation signal (e.g., predistortion) to compare the gain and phase shift of the amplified signal on a continuous basis in order to compensate for changes in operating conditions. For this purpose, these patents describe the use of frequency scanning to detect the presence of intermodulation signals. The system controls are then adjusted in order to reduce such intermodulation signals below a predefined threshold, using a trial-and-error, iterative approach. In such systems, an initial adjustment is made to the compensation signal. The output is then scanned frequently in order to evaluate the results of the initial adjustment. If such adjustment does not reduce the intermodulation signals below the threshold, another adjustment is made. The resulting intermodulation signal detected by the frequent scanning process is then compared to the previous intermodulation signal in order to determine the next adjustment. This iterative approach is continued until the intermodulation signals are below a predefined threshold or a predetermined number of adjustments have been made. Such an approach can take a period of time while iterative adjustments are made before the intermodulation products are properly reduced. During this period of time the unwanted distortion is potentially produced by such a system. There is a need for a system that quickly determines the gain and phase adjustments required for the compensation signal to minimize the distortion signals produced. Summary of the Invention An object of this invention is to provide a compensation installation for a linear power amplifier that adaptively controls the gain and phase of the compensation signal component in order to reduce the total distortion produced by the system. Another object of the invention is to provide a compensation facility that quickly determines the gain and phase adjustments required for the compensation signal to minimize the distortion signals produced. The signal uses a reference signal, of known frequency, which is amplified together with the desired carrier signals. Such a system evaluates the errors introduced in the known reference signal, and determines the errors, (for example, the intermodulation distortion) caused by the amplification of the reference signal by the power amplifier. These error signals are then used to determine the current adjustments to the compensation signal that reduces the resulting error signals. In this way, a control cycle is introduced which, by reducing the distortion introduced to the reference signal, reduces the total distortion introduced to the carriers. Such a reference signal may be added to the carriers before the amplification and subsequently canceled subsequently. Alternatively, if the carriers to be amplified include a suitable signal that occurs continuously, such a signal can be used as the reference signal. In the preferred embodiment, the distortion (e.g., the intermodulation products surrounding the reference signal frequency) input to said reference signal is evaluated by comparing an unamplified copy of the reference signal with the reference signal component. of the combined signal amplified by the amplifier. For this purpose, a quadrature amplitude modulation (QAM) demodulator is used. In another embodiment of the invention, the distortion introduced to the reference signal is evaluated by using an FM discriminator centered on the known frequency of the reference signal and an envelope detector. According to one aspect of the invention, a circuit distortion compensation is provided for a power amplifier comprising an input path to forward a combined signal to the amplifier to form an amplified signal, the combined signal comprising at least one carrier signal and a reference; a compensation path to produce a compensation signal to compensate for the distortion produced by the amplifier; a control unit; a coupler for coupling the amplified signal to the control unit; the control unit comprising error signal means to produce a first difference signal indicative of the gain difference between the reference signal and an amplified reference signal component of the amplified signal and to produce a second difference signal indicative of the phase difference between the reference signal and the amplified reference signal component; a gain and phase adjuster, in response to the control unit, to adjust the gain and phase of the compensation signal in order to minimize the error signals. According to another aspect of the invention, there is provided a method for minimizing the intermodulation distortion produced by an RF power amplifier comprising the steps of: (a) introducing an RF signal to be amplified, including the RF signal a reference signal; (b) producing a compensation signal that broadly cancels the distortion produced by the RF power amplifier; (c) producing a pair of error signals indicative of the distortion introduced by the RF power amplifier to the reference signal; (d) adaptively adjust the gain and phase of the compensation signal in order to minimize the error signals. BRIEF DESCRIPTION OF THE DRAWINGS The present invention, together with the objects and additional advantages thereof, will be understood more from the following description of the exemplary embodiments with reference to the drawings in which: Figure 1 schematically illustrates a modality of the invention incorporating a power amplifier with a pre-distortion compensation facility; Figure 2 schematically illustrates an amplifier and a predistortion compensation control device according to an embodiment of the invention; Figures 3 and 4 show respectively a phase plane diagram and a frequency spectrum diagram with reference to which the operation of the installation of Figure 2 is explained. Figure 5 illustrates a feed advancement facility that is used in together with the embodiment of figures 1 and 2. Figure 6 illustrates an alternative embodiment of the control installation. Detailed Description of the Preferred Modes The preferred embodiment of the invention will be described with respect to its implementation for controlling a predistortion facility. It should be appreciated that the control mechanism can also be adapted appropriately for use in controlling the compensation signal used in a feed advancement system. Referring to Figure 1, there is illustrated an input 48 for an RF signal to be amplified, a linear power amplifier (PA) 12, couplers 46 and 26, an output 14 for an amplified RF signal, a predistortion compensation facility. which is constituted by the interior box of components 42, a control facility 40 and a polar rotation unit 78. The RF signal is, for example, a multi-channel signal including multiple channels with frequencies within a portion of the 25 MHz band from 869 to 894 MHz for use in a cellular radio communications system. The distortion compensation facility 42 serves to compensate for the non-linear characteristics of the amplifier 12, especially at higher powers, thereby reducing the intermodulation distortion between the multiple channels. The predistortion compensation facility comprises a drive amplifier (DA) 16 which has characteristics that exactly match those of the power amplifier 12; an additional amplifier 18, couplers (C) 20, 22 and 24; delay units 28 and 30; a subtractor 32; gain and phase adjustment units 34 and 36, and control unit 38 for the control unit 34. The predistortion and power amplifier installation (PD & PA) are known inside the case 42. As shown in FIG. Figure 1, the control signals for the gain and phase adjustment unit 36, which originate from the control unit 40, pass through a polar rotation box 78. Referring to the case 42 of FIG. Figure 1, the incoming signal from the input 48 is coupled to a reference signal 100 that originates from the control unit 40 through the coupler 46. This combined signal 10 is then coupled to the amplifier input of the amplifier. power 12 through the coupler 20, the gain and phase adjustment unit 34, the drive amplifier 16, the coupler 22, the delay unit 30, and the coupler 24 in sequence. A portion of the signal amplified by the drive amplifier 16 is coupled by the coupler 22 to an additive (+) input of the subtractor 32, a subtractive input (-) which is supplied with a portion of the input signal 10 from the coupler 20, after a compensation delay provided by the delay unit 28. An output 33 of the subtractor 32 is supplied as a feedback signal to the control unit 38, which controls the gain and phase adjustment unit 34. to substantially eliminate the components of the incoming signal at the output of the subtractor 32. Consequently, the output 33 of the subtractor 32 is a distortion signal component due to the non-linearity or distortion in the drive amplifier 16, corresponding (due to the exact comparison of the amplifier 16 with the amplifier 12) to the non-linearity or distortion in the power amplifier 12. It should be appreciated that the Gain and phase adjustment 34 could alternatively be provided at other positions in any of the input paths to the subtractor 32. The displayed facility has the advantage of the positioning unit 34 in a feedback loop capable of compensating for gain variations. of the amplifier 16 resulting in changes in operating conditions, such as variations in temperature, supply voltage, aging components, etc., which do not appear in the path from 20 to 28 to 32. In this way, the elements 20, 34, 16, 22, 28, 32, and 38 serve the basic function of producing distortion signal 33 as the output of the subtractor 32 which closely approximates the distortion added by the amplifier 12, as is known in the art. . This distortion signal 33 is then used to predistort the main signal by amplifying as follows. This distortion signal 33 is then supplied, via the gain and phase adjustment unit 36, to the amplifier 18 in which the signal is amplified. The signal also undergoes a phase shift of 180 degrees, effectively reversing the signal, either in the amplifier 18 or in the adjuster 36. The output of the amplifier 18 is supplied to the coupler 24, in which the amplified distortion signal (due to its inversion) it is subtracted from the amplified combined signal (ie, the output of DA16), the combined signal being provided with a compensation delay in the delay unit 30. After subtraction in the coupler 24, the signal The predistorted power is supplied to the power amplifier 12. The gain and phase adjustment unit 36 has two adjustable input settings (or parameters). A parameter, which in FIG. 1 is controlled by the signal 110, adjusts the gain of the distortion signal passing through the unit. The second parameter, which in this case is controlled by the signal 120, adjusts the phase of the distortion signal passing through the unit. The graduations of the gain and phase settings in the unit 36 are initially pre-set in order to produce a distortion signal that passes to the amplifier 12., which will broadly cancel the distortion signal produced by the amplifier 12, thus minimizing the intermodulation distortion at the output 14, as is known in the art. A portion of the output signal 14 from the power amplifier 12 is coupled by the coupler 26 back to the control unit 40 to form a feedback loop, which controls the gain and phase adjustment unit 36. This cycle of feedback adjusts the graduations of the unit 36 in order to adaptively compensate for changes in operating conditions, such as variations in temperature, humidity, power supply voltage, aging components, etc. The two types of distortion produced by PA12 are AM / AM (amplitude modulation modulation to amplitude modulation) and AM / PM (amplitude modulation distortion to phase modulation). For small errors in the graduations in the unit 36, the effect on the composite non-linearity of the whole system is as follows: the errors in the gain control signal 110 result in a non-linear distortion composed of the same type (AM / AM or AM / PM) to that produced by amplifiers 16 and 12, provided that errors in the phase control signal 120 result in a composite non-linear distortion of the type opposite to that produced by amplifiers 16 and 12. For example, suppose that amplifiers 16 and 12 produced only AM / PM (phase rotation dependent on the envelope of the input signal). For such an amplifier, an error in the gain control signal 110 results in a composite AM / PM (distortion of the same type) whose sign and magnitude are proportional to the gain graduation error, as long as an error in the signal of phase control 120 results in a composite AM / AM (RF gain dependent on the envelope of the input signal; that is, distortion of the opposite type) whose sign and magnitude are proportional to the phase graduation error. Similarly, if the amplifiers 16 and 12 only produce AM / AM distortion, then the gain control signal 110 adjusts the composite AM / AM distortion and the phase control signal 120 adjusts the composite AM / PM distortion. Typically, an amplifier exhibits both AM / AM and AM / PM in some weighted combination. Accordingly, an error in the gain control graduation 110 produces a similarly weighted weighting of an AM / AM and AM / PM distortion and an error in the phase control graduation 120 produces a differently weighted combination of distortion AM / AM and AM / PM. For the purposes of automatic control according to the embodiment of the invention shown in Figures 1 and 2, the control unit 140 produces a pair of control signals (a, p), where "a" is a control signal of AM / PM distortion compensation, and "p" is an AM / PM distortion compensation control signal, as will be explained below. A polar rotation is applied to the pair of control signals (a, p) produced by the control unit 140, which is a complex number, in the polar rotation unit 78 in order to produce the gain control signal 110 and the phase control signal 120. This is done in such a way that the change in the control signal "a" results in a huge distortion of AM / AM, and the changes in the control signal "p" result in a huge distortion of AM / PM. The polar rotation unit 78 produces a gain control signal 110 = a * eos (theta) + p * sin (tetha) and produces a phase control signal 120 = a * sin (theta) - p * eos (theta ), where theta is a fixed angle selected according to the mix of AM / AM and AM / PM produced by the amplifiers. For example, for amps dominated by AM / AM theta approaches zero degrees while for amps mostly with AM / PM, theta approaches 90 degrees. For amplifiers that produce similar amounts of AM / AM and AM / PM theta approaches 45 degrees. This actual graduation for theta in the polar rotation unit 78 is initially established depending on the characteristics of PA 12. This invention particularly relates to the manner in which the signals? A 'and p' are produced, which are the inputs to the polar rotation unit 78, as illustrated in FIG. 2. In FIG. 2, all the components of the installation of FIG. 1 inside the housing 42 are referred to as a power amplifier unit and a predistortion unit for advancing the power (PD & PA). Figure 2 also shows the polar rotation unit 78, the initial coupler 46 for producing the input 10 from an initial input signal 48 to the unit 42, the coupler 26, and the output 14 of the installation of figure 1 The remaining components illustrated in Figure 2 serve as the control unit 40, as shown in dotted lines. Referring to Figure 2, the control unit 40 is shown including a pilot generator 44 which generates a pilot signal 100 at a frequency of, for example, 910 MHz, which serves as the reference signal. This signal is chosen for convenience because it is a little outside the signal band of the input signal 48 (which, for this example, is between 869, and 894 MHz), so that the intermodulation products, such As discussed below, they can be more easily distinguished from the desired signals. The coupler 46 serves to couple the pilot signal from the generator 44, and the RF input signal from the path 48, to form the input signal 10 for the unit 42. As can be seen from the previous description, the characteristics The non-linear amplification of the power amplifier results in the interaction between the RF signal to be amplified and the pilot signal. This interaction produces intermodulation products at frequencies centered at the pilot signal frequency, hereinafter referred to as Reference Intermodulation Products (RIPS), and the intermodulation products are centered around the actual signals to be amplified (hereinafter referred to as Signal Intermodulation Products (SIPs)), as shown in Figure 4. The invention makes use of a specific reference signal, for example, the pilot signal to isolate these RIPs in order to produce the feedback control signals a and p which minimize the RIPs produced. As a result of controlling the gain and phase adjustment unit 36 to minimize the RIPs produced, the SIPs produced by the predistortion unit 42 are minimized. To do this, the reference signal is divided into two trajectories. The first path is coupled to the incoming RF signal and amplified together with it, forming the amplified output signal 14 described above. This amplified signal is then demodulated by QAM (Quadrature Amplitude Modulation), with respect to a copy of the reference signal 44 (2nd path) offset by phase in an appropriate manner. As discussed below, this produces appropriate precursors to the "a" and "p" control signals; that is, an "I" signal representing the change of gain in the reference signal after amplification in PA 12, and a "Q" signal representing the phase change in the reference signal. More particularly, the "I" signal, which in this mode is the first difference signal, is a signal having a voltage that indicates the difference in gain, at a given moment in time, between the reference signal and the reference signal component of the amplified signal. Similarly, the "Q" signal, which in this case is the second difference signal, is a signal having a voltage that indicates the phase difference, at any given moment in time, between the reference signal and the reference signal component of the amplified signal. The second path is used to control a feedback loop to control the QAM demodulator. In the embodiment shown in Fig. 2, the portion of the output signal 14 which is decoupled by the coupler 26 is supplied to a bandpass filter 50. The bandpass filter 50 has a bandpass centered on the a pilot signal frequency and a band amplitude corresponding to two times the bandwidth of the RF signal being amplified, and substantially reducing the amplitude of the amplified RF signal outside this bandpass. In other words, the bandpass filter 50 serves to isolate the RIPS. The output of the bandpass filter 50 is supplied to a signal input of a QAM demodulator (Quadrature Amplitude Modulation) 52. The QAM demodulator 52 also has a local oscillator input which is supplied with the second path of the pilot signal from the generator 44 through a phase adjuster 56, which is controlled to adjust the phase of the pilot signal as described below. The QAM demodulator 52 produces demodulated output signals in phase (I) and quadrature phase (Q) which are supplied to optional low pass filters 58 and 60 respectively. The demodulated signals I and Q produced at the outputs of the QAM demodulator 52 track the phase and phase quadrature components of the pilot tone and its intermodulation products as illustrated in Figures 3 and 4. Figure 4 is a diagram of frequency spectrum, illustrating, by way of example, the relative amplitudes of different frequencies of a multi-channel RF signal, the pilot signal, and resulting, intermodulation products of the amplified RF signal through the unit 42. In this example, the multi-channel signal comprises a plurality of RF carriers that constitute the desired signal, at frequencies in a frequency band from 869 to 880 MHz. The bandwidth of the RF signal in this case is 11 MHz. As is well known in cellular radio communication systems, other RF carrier frequencies and other RF signal band amplitudes can also be used. Adjacent to this desired RF signal band, at lower and higher frequencies, Figure 4 shows that there are signal intermodulation products (SIPS) at lower amplitudes, due to the distortion in the power amplifier (non-linearity) 12. The Figure 4 also shows the pilot tone at a frequency of 910 MHz as described above, having an amplitude that is chosen for example at -30 dBc with respect to the total signal power. Reference intermodulation products (RIPS) are also displayed at lower amplitudes, at frequencies above and below the pilot tone frequency, due to the intermodulation between the pilot signal and the RF carriers of the multi-channel signal. Figure 4 also indicates a bandwidth of 22 MHz of the predominant RIPS centered around the pilot signal. In this example, this bandwidth of 22 MHz is equal to twice the RF signal band amplitude of 11 MHz for a power amplifier having non-linearity characteristics of 3er. order, conventional. In this example, a suitable bandpass of the filter 50 is 22 MHz in order to isolate the RIPS and the low pass band amplitude of the filters 60 and 58 is 11 MHz. As a result of proper filtering, the demodulator of QAM 52 is provided with substantially only the pilot tone and the intermodulation products to which the pilot tone contributes; in other words, band pass and low pass filters reject high amplitude RF carriers. Filters 58 and 60 may be omitted in some circumstances, depending on. the quality of the QAM demodulator 52, especially where the bandpass filter 50 is presented. These filters can also be complemented with, or replaced by, limiting, quadrature, or other non-linear circuitry for the purpose of compensating for distortion effects of higher order in the power amplifier 12.

Referring again to Figure 2, the Q signal from the LPF 60 is preferably supplied through the integrator 64 as a control signal for the phase adjuster 56, thereby forming a control cycle 65 for adjusting the signal phase of the signal. unamplified reference so that this signal is in phase with the amplified reference signal components (e.g., RIPS) of the output 14 coming from the amplifier 12. This control cycle can be understood with reference to Figure 3. Figure 3 is a phase plane diagram, with axes representing in a conventional manner the amplitudes of the demodulated signal in phase I and the quadrature phase demodulated signal Q produced in the outputs of the QAM demodulator 52. A point P represents an arbitrary position corresponding to the signal magnitudes I and Q at an arbitrary moment. This control cycle serves to move the point P in the direction of an arrow A along a circular path (ie, only the phase is changed) to the point Pl on the axis I, that is, the value d.c. of Q is zero. When the phase shifter 56 is thus adjusted, the Q signal responds to the phase changes through the unit 42 and the I signal responds to the gain changes through the unit 42. A direct current component (dc) of the resulting signal I is removed by a dc remover circuit 62. As illustrated, the d.c. 62 comprises a subtractor 66, a series resistor 68 and a bypass capacitor 70. The signal I is supplied to the additive input (+) of the subtractor 66, and an attenuated or integrated version of the incoming signal, produced by an integrator that it comprises the series resistor 68 and the bypass capacitor 70, it is supplied to the subtractive input (-) of the subtractor. The output of subtractor 66 constitutes the output of the d.c. 62. This output represents the fluctuations of time variation in the gain of the unit 42 caused by the non-linearity of AM / AM. A d.c. similar 72 is supplied with the output of a envelope detector 54, which serves to provide a signal representative of the amplitude of the envelope of the RF signal. As shown, the input to the envelope detector 54 is from the coupler 26. It should be appreciated that the input to the envelope detector 54 could be derived alternatively from any point in the main RF signal path. The output signals from the d.c. 62 and the low pass filter 60 each correlate with the RF signal envelope amplitude by multiplying these signals by the output of the d.c. 72 in multipliers 74 and 76, respectively. The demodulated signal of Q, derived from the output of the QAM demodulator 52 through the low pass filter 60, consists of a variable signal indicating AM / PM distortion (amplitude modulation to phase modulation) in the reference signal amplitude produced by the power amplifier 12. This AM / PM distortion is represented in figure 3 by fluctuations of the point Pl in the directions shown by a double-headed arrow B. In this way, the point Pl moves as indicated by arrow B along part of the circular path. This movement corresponds to the phase modulation of the amplified reference signal resulting from the amplitude modulation of the main RF signal and the non-linear distortion of the power amplifier 12. This amplitude-dependent phase variation correlates with the amplitude of the amplitude. the RF signal envelope by the multiplier 76, which multiplies the output of the low pass filter 60 by the output of the envelope detector 54 after removal of its dc component by the d.c. 72. The output of the multiplier 76 constitutes a signal of estimation or error of AM / PM whose average value is indicative of the sign and magnitude of the error in the AM / PM distortion compensation control? p '. This AM / PM error signal determines both the magnitude and direction of the phase adjustment to be made. The demodulated signal I derived from the output of the QAM demodulator 52 through the low pass filter 58 has its d.c. removed by the d.c. 62. The resulting signal indicates the AM / AM distortion produced by the power amplifier 12 and is represented in FIG. 3 by the movement of the point Pl in the directions shown by a double-headed arrow C. In this way, the point Pl moves as indicated by the arrow C radially with respect to the intersection of the I and Q axes, corresponding to the amplitude modulation as a result of the amplitude modulation of the main RF signal and the non-linear distortion of the power amplifier 12. This amplitude variation correlates with the amplitude of the RF signal envelope by multiplying the outputs of the dc remover circuits 62 and 72 in the multiplier 74. The output of the multiplier 74 constitutes an AM / AM estimation or error signal whose average value is indicative of the sign and magnitude of the error in the AM / AM distortion compensation control a '. This AM / AM error signal determines both the magnitude and direction of the gain adjustment to be made. The resulting control signals are stored in the buffer in the integration and storage unit 80 to control the gain and phase adjuster 36 through the polar rotation unit 78. It should be noted that the sensitivity of the AM / PM detector and AM / AM approaches zero in the absence of either an RF carrier or sufficient envelope fluctuations. In this way, integration and storage 80 are only updated preferentially if the main RF signal is of sufficiently large average signal strength and exhibits sufficient envelope fluctuation. For example, a single modulated FM carrier may have insufficient envelope fluctuation, while the presence of two such FM carriers or a single broadband modulated amplitude carrier such as a signal from a Multiple Access base station may occur. Division of Code IS-95 of multiple users is likely to provide a sufficient fluctuation of envelope. If the integrator is updated when there is no carrier or insufficient envelope fluctuation, the displacement voltages can cause the integrated signals to increase or decrease to the control limits. This can potentially result in transient intermodulation when additional carriers are added. In this situation, this transient intermodulation is avoided by storing the integrator outputs when there is insufficient RF power or envelope fluctuations. For this purpose, for example, the updating of the store 80 is enabled by an output of a compactor (not shown) that compares the output of the envelope detector 54 with a fixed threshold level. The installation described above operates on the principle that a reference signal, for example the pilot signal, added to the RF carriers, is subject to a phase shift and a gain change of • agreement with the carrier characteristics of carriers Instantaneous RF and non-linearity of the unit 42. In this way, the reference signal is subjected to a phase shift and a gain change dependent on amplitude (compression or expansion). The demodulation of the reference signal at the output of the unit 42 allows the observation of the relationship between the phase and the gain of the reference signal and the instantaneous RF envelope as detected at the output of the unit 42. The multiplication of Q and the RF envelope together provide an AM / PM error signal. The high-pass filtering of the I and RF envelope terms and the multiplication of these together provide an AM / AM error signal. The AM / PM and AM / AM error signals are used to adjust the gain and phase of the compensation pre-distortion signal in a feedback loop in order to minimize the error signals, thereby providing an optimal compensation for the error signal. distortion of the power amplifier. The pre-distortion installation above is preferably used in conjunction with a feed advancement installation, as illustrated in Figure 5. This feed advance circuitry (comprising the coupler 8, the compensation delay 82, the subtracter 84, the phase inversion amplifier 92 and the coupler 96, together with the compensation delay 94 to produce the output 98, as is known in the art) acts to cancel the distortion that is not canceled by the predistortion facility. In addition, after the pilot has served his purpose in aiding predistortion control, the feed advance circuitry serves the purpose of largely canceling the pilot at output 98 via coupler 96; this cancellation is optimized by adjusting the gain and phase of the feed advance signal in the gain and phase adjuster 88, as is known in the art. It should be noted that, as shown in Fig. 4, a pilot signal frequency was chosen so that there would not be a considerable overlap of the frequency band of the pilot signal and its intermodulation products with the frequency band of the carriers. of RF and its intermodulation products. Any such overlap would reduce the effectiveness of the bandpass filter (s), resulting in increased variability or noise at the outputs of the PD and PA unit 42. In order to attenuate the RF carriers in In this case and therefore to reduce the output variability, it is desirable to derive the input from the QAM demodulator 52 after a cancellation point (eg, after the subtractor 84) of the feed path of the system as shown in FIG. Figure 5. In this manner, Figure 5 also shows a coupler 86 for using the subtracted signal produced by the subtractor 84 as an alternative input (not shown) to the QAM demodulator 52 in place of the input as shown in the figure 2. When using this signal, the RF carriers are substantially attenuated but with a relatively minimal attenuation of the reference signal from which the AM / AM and AM / PM estimates are derived. As a result, the variability in the outputs of the PD and PA unit 42 is reduced. In this particular case, the need for the bandpass filter 50 is reduced and can thus be omitted.

It should be noted that although in Figure 4 the pilot signal is illustrated as a signal frequency tone, it should be appreciated by a person skilled in the art that the reference signal can be modulated to spread its energy over a wide frequency band, for example , using FM modulation or PSK (phase shift key) diffusion spectrum. Although a particular implementation of the control facility has been described above, it should be appreciated that several alternatives may be used. For example, the control cycle 65 comprising the integrator 64 can be eliminated. Instead, the phase adjuster 56 is preset based on the factory look-up tables determined by the actual frequency chosen as the pilot signal. In this case, a DC stirring circuit, similar to that shown at 62, will be used to move the DC component from the output of the low pass filter 60 before mixing in the mixer 76. In addition, instead of inserting a signal pilot as a reference signal, the reference signal can be extracted from the bandwidth of the input signal. For example, the reference signal may be selected to be a carrier normally present in the RF signal, such as a guide or control channel. If this approach is used, then the bandpass filter 50 will not be able to isolate the amplified reference signal since it is part of the carrier. Accordingly, the feed advance cycle approach shown in Figure 5 is required, which uses the output from the coupler 86 as the input to the QAM 52 demodulator. The factory look-up table approach will also be required for establish the phase adjuster 56. It should be appreciated that one or both of the AM / PM and AM / AM error signals can be produced in other ways without requiring a QAM demodulator, and in some control facilities only signals of error. For example, the reader addresses McNicol et al., U.S. Patent No. 5,193,224, issued March 9, 1993, the disclosure of which is incorporated herein by reference. For another example, as shown in Figure 6, in which the pilot signal is at a frequency separated from the carriers, as shown in Figure 4, the AM / PM error signal may occur as an alternative to integrate the output of an FM discriminator centered on the pilot signal. Such a discriminator has its input supplied with a signal derived from the output of the bandpass filter 50. Similarly, the AM / AM error signal is produced by removing the DC from the output of a shell detector. whose input is supplied with the output of the bandpass filter 50. The above description has shown various embodiments of the invention directed to a predistortion facility and how to control the gain and phase adjuster of such a predistortion facility. It should be appreciated by one skilled in the art that such a control facility can also control the gain and phase of a compensation signal which, instead of being subtracted from the signal to be amplified before the amplification as in the predistortion approach described above, as an alternative, the compensation signal can be used to cancel the distortion introduced after the amplification in a feed advance cycle. Numerous other modifications, variations and adaptations may be made to the embodiments of the invention described above without departing from the scope of the invention as defined in the claims.

Claims (18)

1. NOVELTY OF THE INVENTION Having described the present invention, it is considered as a novelty and therefore the property described in the following claims is claimed as property. A distortion compensation circuit for a power amplifier comprising: an input path for advancing a combined signal to said amplifier to form an amplified signal, said combined signal comprising at least one carrier and a reference signal; a compensation path to produce a compensation signal for compensation of the distortion produced by said amplifier; a control unit; a coupler for coupling said amplified signal to said control unit; said control unit comprising error signal means for producing a first difference signal indicative of the gain difference between said reference signal and an amplified reference signal component of said amplified signal and for producing a second indicative difference signal of the phase difference between said reference signal and said amplified reference signal component; and a gain and phase adjuster, responsive to said control unit, for adjusting the gain and phase of said compensation signal in order to minimize said error signals.
2. 2. A distortion compensation circuit according to claim 1, characterized in that said control unit further comprises: an envelope detector for detecting the instantaneous RF signal envelope of said amplified signal; first mixing means for correlating said first difference signal with said signal envelope in order to form a signal indicative of the AM / PM distortion within said amplified reference signal component; and second mixing means for correlating said second difference signal with said signal envelope in order to form a signal indicative of the AM / AM distortion within said amplified reference signal component.
3. 3. A distortion compensation circuit according to claim 2, characterized in that said error signal means comprises an FM integrator and discriminator to produce said first difference signal and an envelope discriminator or AM discriminator to produce said second signal. Of diference. A distortion compensation circuit according to claim 2, characterized in that said error signal means comprises a QAM demodulator and wherein said control unit further comprises an oscillating coupling means for coupling said reference signal to said input of local QAM demodulator oscillator. 5. A distortion compensation circuit according to claim 2, characterized in that said input means comprises: means for receiving an RF signal to be amplified; means for producing a reference signal; and means for coupling said reference signal to said RF signal to form a combined signal to be amplified. 6. A distortion compensation circuit according to claim 4, characterized in that said input means comprises: means for receiving an RF signal to be amplified, said RF signal comprising at least one carrier signal and said reference signal; and means for coupling said reference signal to said oscillator coupling means. 7. A distortion compensation circuit according to claim 4, characterized in that said oscillator coupling means further comprises a phase adjuster for adjusting the phase of said reference signal. A distortion compensation circuit according to claim 7, characterized in that said coupling means further comprises a feedback loop coupled to a quadrature phase demodulated output signal (Q) from said QAM demodulator to control said phase. 9. A distortion compensation circuit according to claim 8, characterized in that said power amplifier has a non-linear characteristic and further comprises a polar rotation unit having an angle selected according to the non-linear characteristic. A distortion compensation circuit according to claim 9, characterized in that said error signal means further comprises an integration and storage unit. 11. A method for minimizing the intermodulation distortion produced by an RF power amplifier comprising the steps of: (a) inputting an RF signal to be amplified, said RF signal including a reference signal; (b) producing a compensation signal that will broadly cancel out the distortion produced by the RF power amplifier; (c) producing a pair of error signals indicative of the distortion introduced by said RF power amplifier to the reference signal; and (d) adaptively adjusting the gain and phase of said compensation signal in order to minimize said error signals. A method according to claim 11, characterized in that step (c) comprises isolating the amplified reference signal component from the output of said RF power amplifier; comparing said amplified reference signal component with said reference signal in order to produce: a first difference signal indicative of the gain difference between said reference signal and said amplified reference signal component; and a second difference signal indicative of the phase difference between said reference signal and said amplified reference signal component. A method according to claim 11, characterized in that it further comprises the step of determining the instantaneous RF signal envelope of said amplified signal and wherein step (c) further comprises the steps of: correlating said first error signal with said signal envelope in order to form a signal indicative of the AM / PM distortion within said amplified reference signal component; and correlating said second error signal with said signal envelope in order to form a signal indicative of the AM / AM distortion within said amplified reference signal component. A method according to claim 12, characterized in that said comparison step comprises the step of using a QAM demodulator to demodulate said amplified reference signal component by QAM, using an unamplified copy of said reference signal as the input of oscillator to said QAM demodulator. A method according to claim 13, characterized in that step (a) comprises the step of generating and coupling a reference signal to the RF signal to be amplified, 16. A method according to claim 14, characterized in that said compensation signal it is a predistortion signal. A method according to claim 14, characterized in that said compensation signal is used to cancel said distortion after amplification by said RF power amplifier in a feed advance cycle. A method according to claim 11, characterized in that step (c) comprises the steps of: isolating the amplified reference signal component from the output of said RF power amplifier; using an FM discriminator centered on said reference signal frequency to produce a first difference signal indicative of the gain difference between said reference signal and said amplified reference signal component; and using an envelope detector or AM discriminator to produce a second difference signal indicative of the phase difference between said reference signal and said amplified reference signal component. SUMMARY A control installation and a method for controlling a gain and phase adjuster used to control the gain and phase of a compensation signal to compensate for the distortion produced by a power amplifier are set forth. The control installation makes use of a reference signal, of known frequency, which is amplified together with the desired carrier signals. The reference signal component of the amplified signal is then isolated and a comparison made, either with the actual reference signal, for example, by means of a QAM demodulator, or alternatively with the known frequency of the reference signal, by example, by means of an FM discriminator, in order to determine the gain and phase differences of the reference signal component of the amplified signal compared to the reference signal. For example, a QAM demodulator is used to produce a first difference signal indicative of the phase switching difference, and a second difference signal indicative of the gain switching difference. These signals are then correlated with the signal envelope of the main signal in order to produce signals indicative of the sign and magnitude of the control signals used to adjust the gain and phase adjuster in a feedback loop.