JP5048593B2 - Transmitter - Google Patents

Description

  The present invention relates to a transmitter having a predistortion type distortion compensation circuit that compensates for distortion components generated by an amplifier.

  In general, when a large main signal is input to a power amplifier of a transmitter, a distortion component may be added to the output signal. In particular, a transmitter for a mobile phone base station requires high power efficiency and low nonlinear distortion characteristics. Therefore, nonlinear distortion generated according to the level of the input signal of the amplifier becomes a problem, and distortion compensation is necessary. . In order to satisfy such a requirement, distortion compensation using a feedforward method or a predistortion method is performed.

  In distortion compensation by the feedforward method, amplification is performed by a main amplifier, an input signal having an opposite phase is synthesized with a signal including distortion, and only a distortion component is extracted. The extracted distortion component is amplified by the distortion amplifier and then brought into the opposite phase and added to the main amplifier output to perform distortion compensation. However, in this system, power consumption increases due to the distortion amplifier, and it is difficult to improve power efficiency.

  In distortion compensation by the predistortion method, distortion is added in advance by modulating the amplitude and phase so as to be the inverse characteristics of AM (Amplitude) -AM (Amplitude) characteristics and AM (Amplitude) -PM (Phase) characteristics, respectively. The predistortion type distortion compensation is realized by digital signal processing without inputting the processed signal into the amplifier and using the distortion amplifier.

  In recent years, various distortion compensation techniques for power amplifiers have been developed in order to reduce the size, power consumption, and price of power amplifiers. In general, as the signal becomes wider, the memory effect in the amplifier (being affected by past signals) increases and the distortion compensation characteristics deteriorate. In order to enhance the distortion compensation effect, there is a problem that the distortion compensation circuit scale becomes large. Therefore, Non-Patent Document 1 proposes a method of reducing the increased circuit scale in order to perform distortion compensation due to the memory effect of the amplifier.

  Non-Patent Document 2 discloses several methods for obtaining a good distortion compensation amount, using a PH (Parallel Hammerstein) model as a representative. Furthermore, in Patent Document 1, in order to reduce the influence of the memory effect of the amplifier, the distortion component generated due to the memory effect of the amplifier is compensated by using the temporal difference of the result of raising the input signal to an even power. Technology to do is shown.

Japanese Patent Laying-Open No. 2005-101908 Naoki Hone, Tetsuhiko Miyatani, Yoichi Okubo, Yoshihiko Akaiwa, “Digital Predistorter for Memory Power Amplifiers”, IEICE Transactions Vol. J88-B No. 10 pp. 2062-2071, 2005/10/01 Isaksson, M.M. Wisell, D.W. Ronnow, D.D. “A comparative analysis of behavioral models or RF power amplifiers”, Microwave Theory and Techniques, IEEE Transactions on publicJ. 2006 Volume: 54, Issue: 1 On page (s): 348-359

  Usually, the distortion compensation amount and the circuit scale / computation amount are in a trade-off relationship, and it is difficult to satisfy both at the same time. Moreover, the method of the patent document mentioned above and the nonpatent literature has many parameters which should be converged, and tends to become unstable. In addition, in order to improve stability, it is necessary to collect a large number of signals for distortion compensation, and it is difficult to shorten the collection time and processing time, and it is not possible to execute sequential distortion compensation during transmission.

  For the reasons described above, there has been a demand for further development of a distortion compensation circuit that can obtain a good distortion compensation effect even with a simple circuit and can obtain a stable convergence characteristic in a short time. Accordingly, an object of the present invention is to provide a transmitter having a distortion compensation circuit that can reduce the influence of the memory effect in a short time.

  In order to achieve the above object, a transmitter according to the present invention includes a signal generator that generates a training signal and a training signal to compensate for distortion components generated by an amplifier that amplifies an input signal. In a transmitter having a predistortion type distortion compensation circuit that compensates for a component, the signal generator constellates the increase / decrease of the amplitude signal level, the time differential value of the positive / negative amplitude, and the time differential value of the positive / negative phase. An 8-character training signal that draws an 8-character above is generated, and the distortion compensation circuit detects the distortion generated by the amplifier to detect the distortion generated by the amplifier. A training signal is input to the amplifier, and distortion components generated due to nonlinearity of the AM-AM characteristic and AM-PM characteristic of the amplifier are detected at least for each time differential value of the amplitude. Based on the calculated compensation values of the plurality of sections, the distortion detecting means for calculating the compensation values of each section by dividing the distortion component detected by the strain detection means into a plurality of preset sections And compensation means for superimposing the compensation signal on the input signal.

  In the transmitter according to the present invention, the distortion detection means detects a distortion component using at least one of the amplitude of the input figure-8 training signal, the time differential value of the amplitude, and the time differential value of the phase. It is characterized by.

  In the transmitter according to the present invention, the computing means computes the compensation value by dividing it into at least two categories of an upper limit value and a lower limit value using the input 8-character training signal.

  In the transmitter according to the present invention, the compensation means includes an interpolation means for interpolating discrete distortion components detected by the plurality of distortion detection means.

  In the transmitter according to the present invention, the threshold used for the division is a threshold based on a linear division including the peak of the input signal modulated by a predetermined modulation method, or the peak of the input signal and the input signal. The threshold value is determined by a polyline approximation section including at least an average value.

  Furthermore, the transmitter according to the present invention is characterized in that the compensation means includes residual distortion removal means for further removing residual distortion remaining in distortion compensation by a predistortion method.

  By using a transmitter having a distortion compensation circuit according to the present invention, even with a simple circuit, a good distortion compensation effect can be obtained in a short time, and a stable convergence characteristic can be obtained. .

  Hereinafter, the best mode for carrying out the present invention (hereinafter referred to as an embodiment) will be described with reference to the drawings.

  FIG. 1 shows the configuration of the transmitter 1. The transmitter 1 receives an output signal Y that is fed back and a reference signal X that is a signal to be amplified, an inverse distortion signal generation circuit 10 that functions as a distortion compensation circuit according to the first embodiment, and an inverse distortion signal. The mixer 14 that upconverts Z, the power amplifier 11 that amplifies the signal, the directional coupler 12 that extracts the output signal Y, the antenna 62, and the output signal Y distributed by the directional coupler 12 are downconverted. It has a mixer 15 and a local oscillator 13 for frequency conversion.

  The inverse distortion signal generation circuit 10 which is a distortion compensation circuit is configured by a digital circuit, and includes a signal generator 63 for measuring distortion characteristics, a switch 61 for switching signals, an amplitude detector 19, and a power detector 18. The amplitude differential value analyzer 25 and the phase differential value analyzer 26 which switch each memory table and instruct weight interpolation, and the weight which interpolates by weighting the amplitude differential value and the phase differential value based on an instruction from these analyzers. An interpolator 27 and a multiplier 28 that multiplies the reference signal X by a compensation value. Further, the inverse distortion signal generation circuit 10 generates an amplitude differential value table creation unit 16 that creates a compensation table based on the time differential value of the amplitude, and a phase differential value table creation that creates a compensation table based on the time differential value of the phase. And a memoryless table 21 of amplitude differential values and a memoryless table 22 of phase differential values.

  Note that “memoryless” in the memoryless tables 21 and 22 refers to the effect of outputting a predetermined value one-to-one with respect to the input without being affected by past signals. A characteristic feature of the present invention is that two types of tables as shown in FIG. 1 are created based on the fact that the distortion characteristics of a power amplifier, which will be described later, differ depending on the amplitude differential value and the phase differential value even when the same instantaneous power is used. And a plurality of memoryless tables that are not affected by past signals are divided into predetermined sections, and each memoryless table is switched and used to compensate for the memory effect.

  An amplitude differential value and a phase differential value of an output modulation signal with respect to an input modulation signal such as a W-CDMA signal will be described with reference to FIG. In general, an amplifier having a large output is realized by connecting a plurality of amplifiers. FIG. 2B shows a power amplifier that inputs and amplifies the input modulation signal Pi from the modulation signal generator 31 to the power amplifiers 11a and 11b and outputs the output signal Po. For example, the latter-stage amplifier 11b in FIG. 2B operates as a memoryless amplifier. However, noise or the like transmitted through the power supply line wraps around the previous-stage amplifier 11a, so that the amplifier 11a becomes a past signal. May be affected and have a “memory” effect. FIG. 2A shows a constellation diagram of the input modulation signal Pi, and the length in the radial direction represents the amplitude. Further, dA / dt which is a time differential value of dA which is a change in the amplitude direction represents an amplitude differential value, and dθ / dt which is a time differential value of Δθ indicating a phase by a change in the circumferential direction represents a phase differential value. .

  FIG. 12 shows a modulation pattern generated by the signal generator 63 of FIG. 1 in order to measure the distortion characteristics of the amplifier in this embodiment. A signal indicated by a solid line in FIG. 12 represents an input signal of the amplifier, and a signal indicated by a broken line represents an output signal of the amplifier.

  This modulation pattern is an 8-shaped modulation signal (Honeybee's Dance) that draws an 8 figure on the constellation of an increase / decrease in amplitude signal level, a time differential value of positive / negative amplitude, and a time differential value of positive / negative phase. Modulation). The characteristic of the figure-8 modulation signal is to realize desired conditions with positive and negative amplitude differential values and phase differential values with a small number of samples, and by using them as a training pattern for the distortion compensation circuit, stable distortion compensation in a short time. It became possible to obtain a value.

  Table 1 shows six characteristic modes from “Up +” to “Down−” of the figure 8 modulation signal. The symbol “Up +” in the table is a signal that has started to move from the origin of the constellation. Since the amplitude increases and the phase also increases, the sign of the amplitude differential value is “positive” and the sign of the phase differential value is “positive”. " If such an increase phenomenon continues, “Peak +”, in which the amplitude is constant and the phase is increased, becomes “Peak +”, the amplitude differential value remains “zero”, and the amplitude differential value remains “positive”. Next, the amplitude decreases, but the phase continues to increase to “Down +”, the amplitude differential value becomes “negative”, and the phase differential value remains “positive”. In addition, “+” shown after Up, Peak, and Down in FIG. 12 means the sign of the phase differential value. In the same manner, the process returns to the beginning through “Up−” to “Down−”.

  FIG. 3 shows an example of statistical processing showing the dispersion of the amplitude differential value (A) of the input modulation signal Pi with respect to the instantaneous power of the input modulation signal Pi and the phase differential value (B) with respect to the instantaneous power. In the amplitude differential value of FIG. 3A, the amplitude differential value is dispersed from −0.5 to +0.5 when the instantaneous power is 0 dB, and the instantaneous power is 0 dB in the phase differential value of FIG. At this time, it is dispersed from -10 to +10.

  Therefore, in order to examine the dependency of the amplitude differentiation on the instantaneous power, in the amplitude differential value of FIG. 3A, this variance is −0.5 or more and less than −0.3 and −0.3 or more and −0 from the bottom. Threshold values were set for each of the following five categories: less than 1, less than -0.1 and less than +0.1, more than +0.1 and less than +0.3, and more than +0.3 and less than +0.5. .

  Further, in order to investigate the dependence of the phase differential value on the instantaneous power, in the phase differential value of FIG. 3B, this variance is -5 or more and less than -3, -3 or more and less than -1, and -1 The threshold values are set for each of the five categories of less than +1, less than +1 and less than +3, and more than +3 and less than +5.

  FIG. 4 shows the dependence of the amplitude differential value on the instantaneous power. Data is extracted for each threshold value category shown in FIG. 3 (A), and AM-AM characteristics: FIG. 4 (A) and AM -PM characteristics: FIG. 4B is plotted. In the AM-AM characteristic of FIG. 4 (A), the data of five sections are shown in shades, and the AM characteristic (dB) decreases and decreases in each of the areas where the instantaneous power is low and high. It can be seen that the conditions are different.

  Also, in the AM-PM characteristic of FIG. 4B, the data of five sections are similarly shown in shades, and the PM characteristic (deg) increases for each section as the instantaneous power increases. It can be seen that the degree of increase differs for each category and instantaneous power.

  FIG. 5 shows the dependence of the phase differential value on the instantaneous power. Data is extracted for each threshold value category shown in FIG. 3B, and AM-AM characteristics: FIG. 5A and AM -PM characteristics: FIG. 5B is plotted. In the AM-AM characteristic of FIG. 5 (A), the data of five sections are shown in shades, and the AM characteristic (dB) decreases and decreases in each of the sections where the instantaneous power is low and high. It can be seen that the conditions are different from each other and are different for each category and instantaneous power as in the characteristics of FIG.

  Similarly, in the AM-PM characteristic of FIG. 5B, the data of the five categories are shown in shades, and the PM characteristic (deg) increases for each category as the instantaneous power increases. It can be seen that the degree of increase is different, and is different for each category and instantaneous power as in the characteristics of FIG.

  In FIG. 6, a table of AM-AM characteristics and AM-PM characteristics (Look Up Table) divided by phase differential values in the horizontal direction is arranged, and AM-AM characteristics and AM divided by amplitude differential values in the vertical direction. A compensation table in which LUTs having -PM characteristics are arranged is shown. These horizontal tables correspond to the rotational direction of the constellation, and the vertical tables correspond to the radial rise and fall of the constellation. A mechanism for compensating for distortion due to the dependency of the amplitude differential value and the phase differential value of the modulation signal will be described with reference to FIG.

  Since the AM-AM characteristic and AM-PM characteristic of the power amplifier depend on the amplitude differential value and the phase differential value of the modulation signal, a four-dimensional table is required for each of AM and PM in order to compensate all of them. However, in order to realize a four-dimensional table, the circuit scale becomes too large, which is not realistic. Therefore, as described above, the characteristics of the amplitude differential value and the phase differential value can be simplified by dividing the characteristics into predetermined sections to form a plurality of memoryless tables and a compensation table as shown in FIG. Further, it is possible to prepare at least two memoryless tables divided into predetermined sections to compensate for the dependency due to the amplitude differential value and the phase differential value.

  FIG. 7 is an example of a distortion compensation effect actual measurement result by the distortion compensation circuit according to the present invention. An input modulation signal when a 4-carrier signal of W-CDMA (about 2 GHz band) is input to the power amplifier, and no distortion compensation are shown. An output signal and an output signal with distortion compensation are shown. Moreover, the vertical axis | shaft of FIG. 7 shows electric power, and the horizontal axis has shown the offset frequency. As shown in FIG. 7, in the case of no distortion compensation, the power 10 MHz away from the center frequency was −55 dB, but in the case of compensation, it has stable convergence characteristics and can be improved to −80 dB. It was.

  FIG. 8 shows a configuration of the inverse distortion signal generation circuit 20 according to the second embodiment. This embodiment has a configuration that emphasizes the convergence and stability of distortion compensation. In the inverse distortion signal generation circuit 20 of FIG. 8, a multiplier 33 superimposes a memoryless inverse distortion signal on the reference signal X, and a multiplier 28 further superimposes an inverse distortion signal including a memory effect. With such a configuration, after starting the distortion compensation circuit, it is possible to compensate for memoryless distortion at high speed, and then slowly compensate for memory distortion to improve distortion compensation characteristics.

  Further, the switch 41 is connected in advance to the one having the better distortion compensation characteristic according to the distortion characteristic of the power amplifier. For example, in the distortion model power amplifier 11a shown in FIG. 2B, it is considered that the memory effect is superimposed on the input signal on which the inverse distortion signal is superimposed. Therefore, the switch 41 is connected to the output side of the multiplier 33. As a result, memory distortion due to the input signal on which the inverse distortion signal is superimposed can be generated.

  FIG. 9 shows a configuration of an inverse distortion signal generation circuit 30 according to the third embodiment. The inverse distortion signal generation circuit 30 includes an amplitude table creation unit 36, a power detector 34, a memoryless table 35, a multiplier 33, a switch 41 for selecting distortion compensation, a phase differential value analyzer 26, a level An adjuster 42 and a multiplier 28 for multiplying the reference signal X by the imaginary component of the level adjuster 42 are provided. When the AM-PM characteristic shows a characteristic that depends solely on the phase differential value due to the characteristic of the power amplifier, it is possible to obtain a distortion compensation effect with a simplified circuit configuration without using a memoryless table.

  FIG. 10 shows a configuration of the table creation unit 50 of the inverse distortion signal generation circuit according to the present embodiment. The table creation unit 50 includes a Y / X divider 51 that generates a signal obtained by dividing the output signal Y signal of the power amplifier by the reference signal X, an amplitude differential value analyzer 52, a phase differential value analyzer 53, and an amplitude. A switch 55 for selecting a signal from the Y / X divider 51 according to the differential value and the phase differential value, and a histogram calculator 56 for performing histogram formation and averaging based on the selected data are obtained. A polynomial coefficient calculation circuit 57 that calculates a polynomial from the discrete values, and an LUT value calculation circuit 58 that performs inverse characteristics.

  The amplitude differential value analyzer 52 in FIG. 10 obtains the amplitude differential value of the reference signal X and gives it to the threshold value determination circuit 54. Similarly, the phase differential value analyzer 53 calculates the phase differential value of the reference signal X. It is obtained and given to the threshold judgment circuit 54. The threshold determination circuit 54 determines whether or not the reference signal X is within a preset threshold range, and operates the switch 55 only when it is within a predetermined range. Note that the threshold value can be set to only the amplitude differential value, only the phase differential value, or a combination of both. The switch 55 validates the data from the Y / X divider 51 only when permission is given from the threshold determination circuit 54.

  The histogram calculator 56 in FIG. 10 classifies the effective data by the input power value, accumulates it for each section element, and measures the frequency. This frequency is accumulated for a certain period of time, and then averaged by dividing by frequency for each section element to create a histogram. The generated histogram is approximated by a polynomial by a polynomial coefficient calculation circuit 57, and a polynomial coefficient is output. For the approximation, for example, a least square method is used. The LUT value calculation circuit calculates the LUT value using the approximated polynomial coefficient, and reverse-characterizes it when the Y / X divider is used. With the above configuration, a plurality of LUTs (AM-AM characteristics, AM-PM characteristics) divided by threshold values can be created. As a method for analyzing the phase differential value, a method using an arc tangent table, a method using Hilbert transform, or the like can be suitably used.

  FIG. 11 is an example for threshold setting according to the present embodiment. The linear threshold setting as shown in FIG. 3 simplifies the process, but there is a region where there is no instantaneous power peak (for example, −0.5 or more and less than −0.3). There is a disadvantage that the approximation accuracy is degraded. Therefore, the peak value and the average value (near 0 dB) of the instantaneous power are set by setting the area of Table 1 and the area of Table 2 with threshold values based on multiple lines according to the distribution of the modulation signal as shown in FIG. 6 can be deleted, the polynomial approximation accuracy can be improved and the distortion compensation performance can be improved.

  As described above, by using the transmitter having the signal generator and the distortion compensation circuit of the present embodiment, a good distortion compensation effect can be obtained in a short time even with a simple circuit, and stable. Convergence characteristics can be obtained. In this embodiment, the W-CDMA distortion compensation circuit with envelope variation has been described. However, the present invention is not limited to this, and the present invention is also suitable for use in the OFDM, multilevel QAM, and QPSK systems. .

1 is a configuration diagram illustrating a configuration of a transmitter including an inverse distortion signal generation circuit that functions as a distortion compensation circuit according to a first embodiment of the present invention. It is explanatory drawing explaining the amplitude differential value and phase differential value of an output modulation signal with respect to an input modulation signal. It is an example of the measurement data figure which shows dispersion | distribution of the amplitude differential value and phase differential value of an output modulation signal with respect to the instantaneous electric power of an input modulation signal. It is explanatory drawing explaining the dependence of an amplitude differential value. It is explanatory drawing explaining the dependence of a phase differential value. It is explanatory drawing explaining the mechanism for compensating the distortion by the dependence of the amplitude differential value and phase differential value of a modulation signal. It is an example of the distortion compensation effect measurement result by the distortion compensation circuit which concerns on this invention. It is a block diagram which shows the structure of the reverse distortion signal generation circuit used as the distortion compensation circuit which concerns on the 2nd Embodiment of this invention. It is a block diagram which shows the structure of the reverse distortion signal generation circuit used as the distortion compensation circuit which concerns on the 3rd Embodiment of this invention. It is a block diagram which shows the structure of the table preparation part which concerns on embodiment of this invention. It is an Example which shows an example for the threshold value setting which concerns on embodiment of this invention. It is an example of the modulation pattern for measuring the distortion characteristic of the power amplifier in this embodiment.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 Transmitter 10, 20, 30 Inverse distortion signal generation circuit, 11 Power amplifier, 12 Directional coupler, 13 Local oscillator, 14, 15 Mixer, 16 Amplitude differential value table creation part, 17 Phase differential value table creation part, 18, 34 Power detector, 19 Amplitude detector, 21, 22, 35 Memoryless table, 25, 52 Amplitude differential value analyzer, 26, 53 Phase differential value analyzer, 27 Weighted interpolator, 28, 33 Multiplier, 31 modulation signal generator, 36 amplitude table creation unit, 37, 38 divider, 41, 55, 61 switch, 42 level adjuster, 50 table creation unit, 51 Y / X divider, 56 histogram calculator, 57 polynomial coefficient Calculation circuit, 58 LUT value calculation circuit, 62 antenna, 63 signal generator.

Claims (6)

In order to compensate for distortion components generated by an amplifier that amplifies an input signal, in a transmitter having a signal generator that generates a training signal and a distortion compensation circuit of a predistortion method that compensates for distortion components using the training signal,
The signal generator generates an 8-character training signal that draws an 8-character on the constellation by increasing / decreasing the amplitude signal level, a time differential value of positive / negative amplitude, and a time differential value of positive / negative phase,
The distortion compensation circuit
In order to detect distortion generated by the amplifier, an input signal modulated by a predetermined modulation method or a figure 8 training signal of the signal generator is input to the amplifier, and the AM-AM characteristics and AM-PM characteristics of the amplifier are nonlinear. Distortion detection means for detecting distortion components generated due to the property at least for each time differential value of amplitude;
An arithmetic means for dividing the distortion component detected by the distortion detecting means into a plurality of preset sections and calculating a compensation value for each section;
Compensation means for superimposing the compensation signal on the input signal based on the calculated compensation values of the plurality of sections;
A transmitter characterized by comprising: The transmitter of claim 1, wherein
The distortion detecting means detects a distortion component by using at least one of the amplitude of the input figure-8 training signal, the time differential value of the amplitude, and the time differential value of the phase. The transmitter according to claim 1 or 2,
The calculating means calculates a compensation value by dividing into at least an upper limit value and a lower limit value using the input figure 8 training signal. The transmitter according to claim 1 or 2,
Compensation means
A transmitter comprising: interpolation means for interpolating discrete distortion components detected by a plurality of distortion detection means. The transmitter according to claim 3, wherein
The threshold used for the division is a threshold based on a linear division including the peak of the input signal modulated by a predetermined modulation method, or a multiple linear approximation division including at least the peak of the input signal and the average value of the input signal. Transmitter characterized by a threshold value according to. The transmitter according to claim 4, wherein
The transmitter according to claim 1, wherein the compensation means includes residual distortion removal means for further removing residual distortion remaining due to distortion compensation by a predistortion method.