JP2012244553A - Power amplifier, lut control circuit, and predistortion compensation method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a power amplifier for accurately performing DPD on a burst signal and amplifying the burst signal, an LUT, and a predistortion compensation method.SOLUTION: Level data and loopback data obtained on a sample unit 11 are written into a RAM 15. The number of times that the obtained level data exceeds a predetermined peak power threshold is counted on a level comparison unit 12, and an average power value of the level data is calculated on a power amplification unit 13. If the average power value is equal to or greater than a predetermined average power and if the number of times is equal to or greater than the predetermined count number, a determination processing unit 14 reads the level data and the loopback data from the RAM 15 for an analysis unit 31, calculates a predistortion control signal Ds by using a predetermined program, and overwrites a LUT 32, thereby performing control of updating data stored on the LUT.

Description

  Embodiments described herein relate generally to a power amplifier LUT control circuit and a predistortion compensation method for a mobile communication radio transmitter.

  In recent years, high-speed mobile communication systems such as LTE (Long Term Evolution) and WiMAX systems are being introduced. In such a mobile communication system, signals modulated with OFDM at burst timing are transmitted and received between a base station and a mobile terminal. Such a broadband modulation signal has a large PAPR (Peak to Average Power Ratio), and the transmitter power amplifier is required to have high efficiency and low distortion. Particularly in the vicinity of the peak output, since countermeasures for distortion such as output saturation are required, the input signal is preliminarily subjected to DPD (Digital Pre-Distortion) processing for applying predistortion according to the signal level. At this time, the input signal is monitored and distortion having an inverse characteristic according to the input signal level is given, but the transmission timing of the burst signal is not constant, and the level of the burst signal also varies during transmission.

  A method of detecting the start of transmission of a burst signal, measuring the level, and predicting an inverse characteristic for giving predistortion is also being studied (for example, Patent Document 1).

  However, since the burst signal is switched between time-divided transmission signals, the reverse characteristics cannot be accurately estimated if the signal is acquired and DPD is executed when the input / output level during transmission of the channel is low before and after switching. .

FIG. 5 is a functional block diagram for explaining the operation of a power amplifier using a conventional DPD.
In FIG. 5, the power amplifier PA is input to the DPD unit 2 that performs digital predistortion compensation (DPD), for example, a high-speed digital modulation signal that is OFDM-modulated. The compensated output signal is converted into an analog signal by the D / A 4 and then input to the final stage amplifying unit 5 and the output is supplied to the antenna. The input digital modulation signal is preliminarily applied to the modulation signal (input signal) corresponding to the input signal level by referring to the LUT (Look Up Table) 320 as the level of the signal input by the preamplifier 20 of the DPD unit 20. Read distortion control data Ds. Then, DPD such as gain and phase adjustment is applied to the input signal in accordance with the predistortion control data Ds.

FIG. 6 is a conceptual diagram of signal level measurement of a conventional burst signal, and FIG. 7 is an input / output characteristic diagram showing input signal versus output signal amplitude.
In FIG. 6, Q1, Q2,... Qn are burst signals transmitted in a time-sharing manner. m1, m2, mn1, and mn2 are sample pulses indicating the timing for measuring the level of each burst signal.

  FIG. 7 shows the relationship between the amplitudes of the input and output signals. When distortion correction is not performed, the output curve Op becomes saturated as the input level increases. Therefore, in order to compensate for this saturation, the gain of the power amplifier is increased so as to obtain a preset reverse characteristic curve Rp. The compensation data having the reverse characteristic is written and stored in advance in the LUT 320 in correspondence with the input signal level.

  When the input signal level corresponding to m2 and mn1 in FIG. 6 is k1, which is relatively low, the output curve Op is in a straight line state, and distortion compensation can be reduced. On the other hand, when the input signal level corresponding to mn2 in FIG. 6 is k2, which is relatively high, the output curve Op is in a saturated state, and considerable distortion compensation is required. The power amplifier estimates the inverse characteristic with reference to the LUT 320 that sets the gain corresponding to k1 and k2.

  Therefore, as in the case where the sampling timing in the burst signal Qn is mn1, originally, distortion compensation should be performed on the assumption of a high level input signal, but only a small amount of compensation is performed by determining that the input level is low. Therefore, there is a problem that distortion at a high input level may not be canceled.

JP 2004-112252 A

  In the conventional DPD processing for a burst signal, distortion compensation should be performed assuming a high level input signal, but only a small amount of compensation is performed because it is determined that the input level is low. There was a problem that it could not be countered.

  The problem to be solved by the present invention is to provide a power amplifier, an LUT control circuit, and a predistortion compensation method capable of accurately applying DPD to a burst signal transmitted at random timing and amplifying the burst signal.

  In order to achieve the above object, the power amplifier according to the present embodiment monitors an input signal transmitted in a burst modulated in a wide band and refers to an LUT that stores predistortion control data applied to the input signal. In a power amplifier that outputs a signal amplified by DPD means for applying DPD to the input signal, the input signal and the amplified signal are received, a signal level is sampled and measured at a predetermined sampling period and sampling rate, Sample means for outputting as level data and loopback data with identification information, storage means for storing and holding a predetermined number of the output level data and loopback data, level data and loopback data being input, Calculate the predistortion data to be subjected to DPD, and write the calculated predistortion data to the LUT Analyzing means for storing, power calculating means for calculating and outputting an average power value obtained by cumulative addition and division by the predetermined number, and level data to which the level data is inputted and inputted And a signal level comparison means for counting and outputting the number of times the threshold is exceeded during the cumulative addition, and controlling the start and end timing of the sample measurement, and When the average power value and the number of counts are input, the average power value is equal to or greater than a predetermined average power threshold value, and the count number exceeds a predetermined number of times, the analysis unit is transferred from the storage unit to the analysis unit. And determination processing means for performing control for instructing to read out the stored level data and loopback data.

  In order to achieve the above object, the LUT control circuit of the present embodiment monitors an input signal to be transmitted in a burst modulated in a wide band, and stores a predistortion control data to be applied to the input signal. In the LUT control circuit that writes the predistortion control data to the LUT of the power amplifier that amplifies the input signal by applying DPD to the output and outputs the received signal, the input signal and the amplified signal are received, and a predetermined sample period Sample means for measuring a signal level at a sampling rate and outputting it as level data and loopback data with identification information for each sample, and storage means for storing and holding a predetermined number of the output level data and loopback data And calculating the predistortion data to which the DPD is applied by inputting level data and loopback data, Analysis means for writing and storing the calculated predistortion data in the LUT; power calculation means for calculating and outputting an average power value obtained by accumulating and dividing the level data by the predetermined number; and the level Signal level comparison means for inputting data, comparing the inputted level data with a predetermined peak power threshold value, and counting and outputting the number of times the threshold value is exceeded during the cumulative addition, and the sample measurement The average power value and the count number are input, the average power value is equal to or greater than a predetermined average power threshold, and the count number exceeds a predetermined number of times. In this case, control is performed to instruct the analysis means to read out the stored level data and loopback data from the storage means. Characterized by comprising a determination processing unit.

  Furthermore, a predistortion compensation method for a power amplifier according to an embodiment of the present invention includes a sampling unit, a storage unit, a power calculation unit, a signal level comparison unit, a determination processing unit, an analysis unit, an LUT, and a predistortion. And amplifying means for amplifying by applying DPD according to the control data, the amplifying means monitors the input signal transmitted in a burst modulated in a wide band, and stores predistortion control data corresponding to the input signal In a predistortion compensation method for a power amplifier that outputs an amplified signal obtained by applying the DPD to an input signal with reference to an LUT, the sampling means receives the input signal and the amplified signal, and receives a predetermined sample period. The signal level is sampled at a sampling rate and output as level data and loopback data to which identification information for each sample is added. A predetermined number of stored level data and loopback data is stored, and the power calculation means receives the level data, calculates and outputs an average power value cumulatively added and divided by the predetermined number, and outputs a signal The level comparison means receives the level data, compares the input level data with a predetermined peak power threshold, counts and outputs the number of times the threshold is exceeded during the cumulative addition, The determination processing means controls the start and end timing of the sample measurement, receives the average power value and the count number, the average power value is equal to or greater than a predetermined average power threshold, and the count number is When the predetermined number of times is exceeded, the stored level data and loopback data are stored in the analysis unit from the storage unit. The analysis means that receives the instruction reads out the stored level data and loopback data from the storage means, calculates the predistortion data to be subjected to the DPD, and calculates the calculated predistortion data. Distortion data is written and stored in the LUT.

The functional block diagram explaining operation | movement of the power amplifier of this embodiment. The wave form diagram explaining the sampling process of the input signal of the power amplifier of this embodiment. The wave form diagram explaining the sampling process of the input signal of the power amplifier of this embodiment. The flowchart which shows the sampling process sequence in this embodiment. The functional block diagram of the conventional power amplifier. The wave form diagram explaining the input signal sampling process of the conventional power amplifier. The input-output characteristic figure which shows the input signal versus output signal amplitude of a power amplifier.

  A power amplifier according to an embodiment will be described below with reference to the drawings.

FIG. 1 is a functional block diagram for explaining the operation of the power amplifier of the present embodiment.
In FIG. 1, in the power amplifier PA of the present embodiment, for example, an OFDM-modulated digital modulation signal that is burst-transmitted in the LTE system is input to the preamplifier 20 of the DPD unit 2. In addition, the digital modulation signal is branched and input to the sample unit 11 of the selection processing unit 1. Hereinafter, the same parts as those in FIG.

  The preamplifier 20 acquires the predistortion control data Ds from the LUT 32 of the LUT generation unit 3 and performs DPD processing such as gain adjustment. The LUT 32 analyzes the input signal by the analysis unit 31 so that accurate DPD is possible. The predistortion control data Ds generated in this way is rewritten and updated.

  The selection processing unit 1 receives the digital modulation signal and the loopback signal from the final stage amplification unit 5 digitally converted by the A / D 6 and outputs high-speed sampling data, and is output from the sample unit 11 Input power calculation unit 16 that calculates power for each sample and outputs it as level data, RAM 15 that stores sampled data (including loopback data) as a buffer, level comparison unit 12, average power calculation unit 13, and determination processing Part 14.

  The sampling data of the modulation signal is input to the input power calculation unit 16 and the average power calculation unit 13. The power level data calculated by the input power calculation unit 16 is also output to the level comparison unit 12, and the result of comparison according to predetermined peak power thresholds Vpt1 to Vptn (comparison criteria) is output to the determination processing unit 14. Further, the level data is input to the power calculation unit 13, and the average power value Ao is calculated by cumulative calculation, and the result is input to the determination processing unit 14. The average power value Ao is an average of a predetermined number of times, for example, 4500 times.

  The threshold values Vpt1 to Vptn set by the level comparison unit 12 are set corresponding to the average power value Ao in order from the lowest in a plurality of, for example, 3 dB steps so as to classify the distribution of peak levels in a stepwise manner. . The number of times exceeding the threshold values Vpt1 to Vptn is counted as N1, N2,... Nn and is output to the determination processing unit 14.

  Although the peak threshold value Vpt may be one, it is possible to accurately determine that a signal having an appropriate peak level corresponding to the assumed PAPR average power value Ao is included by using a plurality of peak threshold values Vpt.

  Then, the determination processing unit 14 determines whether to accept the level data from the input average power value Ao and the distribution range, and outputs a RAM read flag F to the analysis unit 31 when it is determined to adopt the level data.

  That is, the determination processing unit 14 compares the average power value Ao with a predetermined average power threshold Vath. If the average power value Ao is not exceeded, the data input state is insufficient or the burst signal sampling timing is inappropriate. As a result, the accumulated 4500 sample data are discarded as a group.

  This average power threshold Vath is a value set as an allowable lower limit input value, for example, a value 8 dB lower than the rated input level Sp assuming standard operation.

  On the other hand, when the average power value Ao exceeds the average power threshold value Vath, the level comparison unit 12 examines the distribution of peak levels exceeding the preset threshold values Vpt1 to Vtn. That is, since the PAPR of the power of the transmission signal is adjusted in advance to approximately 9 dB, for example, when the average power value Ao is low, the number N1 of times exceeding the low peak level threshold Vpt1 is checked, and the number is 50 times. If the specified number of times is exceeded, it is determined that the assumed OFDM signal has been accurately sampled. Then, it decides to adopt the group of 4500 level data, and outputs a RAM read flag F to the analysis unit 31.

  If the specified number is not satisfied, it is determined that the sampling timing of the burst signal is not accurate, and the 4,500 pieces of data are discarded.

  The analysis unit 31 reads the adopted sample data (level data and loopback data) from the RAM 15 at the timing when the flag F is set, and clears and updates the content of the RAM 15. The input level data and loopback data are compared and collated, and predistortion compensation data Ds for performing DPD processing is predicted and generated by a predetermined program, and written and updated in the LUT 32.

  This program causes the LUT 32 to treat the LUT 32 as a low level timing signal with a large amplitude output or to reduce the peak of a small amplitude signal, even if the input signal level is the same, due to the difference between the level and the loopback data. Predistortion control data is generated that accurately reflects whether distortion correction is performed as to whether the signal is handled as a close timing signal.

  In other words, the selection processing unit 1 performs control to output from the RAM 15 after eliminating unnecessary sample data in the burst signal that causes noise from the input signal so that the program for performing the DPD processing operates properly. ing. Hereinafter, the operation will be described in detail.

  The determination processing unit 14 determines the setting of the sampling period (timing) for monitoring the burst signal and the adoption of the sampled data. In the present embodiment, sampling data is repeatedly acquired continuously for a continuous period close to the transmission time of the channel for burst transmission, and sampling data that is statistically used as a reference for DPD is selected.

FIG. 2 is a waveform diagram for explaining the sampling process of the power amplifier PA of this embodiment, and shows a combination of the relationship between the sampling period and the burst signal.
In the present embodiment, a sampling period (gate window) GW is provided so that level data can be acquired in a continuous manner to some extent during burst transmission.

However, there is no relationship in which the start timing of GW, the sampling length, and the transmission time (timing) TS of the burst signal can be synchronized, and some combinations occur.

  2 (a1), (b1), and (c1) have a shorter GW than the burst transmission time TS, and FIG. 2 (a2), (b2), and (c2) have a longer GW. In FIG. 2 (a1), there are cases where the GW falls within the transmission time TS, or vice versa, as shown in FIG. 2 (a2), the transmission time TS falls within the GW. Fig. 2 (b1) and (b2) show the case where only data at the beginning of the transmission time can be obtained from before the beginning of transmission. 2 (c1) and 2 (c2) both include empty data so that only data acquisition from the middle of transmission to the end of transmission can be performed. Therefore, it is determined whether or not the data acquired while the GW is open is at a required level, and this empty data and data that causes an error are excluded.

  Depending on the combination of burst transmission time, GW period, and sampling rate, the form of sample data that can be acquired varies, but the sample data processing procedure will be described below as a representative example when the GW open period is shorter than the burst transmission time TS. Will be explained.

FIG. 3 is a waveform diagram illustrating the sampling process of the input signal of the power amplifier according to this embodiment.
For example, the burst signal is transmitted to the power amplifier in a series of 1 msec subframes such as Q1 to Qn in FIG. Subframes may be continuous or one of them may be paused (missed). For example, a maximum of 10 subframes may have a total radio frame structure of 10 ms.

  Here, for simplicity of explanation, as shown in FIG. 3 (a), there is no burst signal before and after, and a single burst signal having a length of 1 msec is targeted, and the GW is as shown in FIGS. 3 (b) to 3 (d). The case of sampling at 100 MHz (sampling pulse sp repeated 10 nsec) with a length of 200 μsec is targeted. Note that the timing at which the GW opens and closes, the sampling length, and the sampling speed are not limited to these values, and are set in the determination processing unit 14 in advance by a program or an information terminal connected from outside (not shown). Yes.

FIG. 4 is a flowchart showing a sampling processing procedure in the present embodiment.
Hereinafter, the sampling processing of this embodiment will be described with reference to FIGS. 1, 3, and 4.
The determination processing unit 14 transmits a sampling start command to the sample unit 11, the level comparison unit 12, the average power calculation unit 13, and the input power calculation unit 16 at a predetermined timing (step s1 in FIG. 4). Here, although detailed explanation is omitted, since the burst signal itself is random, the start timing is triggered by receiving information indicating that the transmission mode has been entered from the control means of the wireless transmitter (not shown), or It may be set based on periodic repetition or random number generation.

  As soon as the GW is opened, the sample unit 11 acquires the input signal, that is, sample data of the digital modulation signal (hereinafter referred to as level data (amplitude information)) until it exceeds 200 μsec (step sx is No). .) Identification code For example, a number from the head is assigned and output to the RAM 15 and the input power calculation unit 16 (step s2).

  The input power calculation unit 16 calculates the power for each sample and outputs it as level data to the level comparison unit 12 and the power calculation unit 13. The power calculator 13 outputs the average power value Ao and the statistical distribution (for example, level data corresponding to σ1 to σ3 of the Gaussian distribution) to the determination processor 14 when 4500 level data are accumulated (step s3). .)

  The level comparison unit 12 compares the input level data with the peak threshold values Vpt1 to Vptn, counts the number of times the threshold values Vpt1 to Vptn are exceeded as N1, N2,... Nn, and outputs them to the determination processing unit 14. (Step s4).

  This count number is a number during the cumulative addition of the peak levels of 4500 data by referring to a timer (not shown). In addition, the count may be notified by the determination processing unit 14 instead of being performed by the level comparison unit 12, and the determination processing unit 14 may perform the count.

  The determination processing unit 14 compares the average power value Ao with a predetermined average power threshold Vath (if step s5 is No), the data input state is insufficient or the sampling timing of the burst signal Are discarded as a group (step s11).

  On the other hand, if it exceeds (when step s5 is Yes), the distribution of data exceeding the threshold value Vpt corresponding to the average power value Ao (for example, Vpt1 if the average power is low) is examined (step s6). ).

  Here, if the number of times exceeding the threshold value Vpt exceeds the specified number of times such as 50 (when step s7 is Yes), it is determined that the group of 4500 level data is adopted, and the RAM read flag F is output to the analysis unit 31 (step s8).

  If step s7 is No if the specified number of times is not satisfied. ), It is determined that the sampling timing of the burst signal is not accurate, and the 4500 pieces of data are discarded (step s11).

  The specified number of times may be the same number for all peak power thresholds Vpt1 to Vptn, or may be set independently for each peak threshold value.

  At the timing when this flag is present, the determination processing unit 14 outputs 4500 pieces of sample data from the RAM 15 to the analysis unit 31 to update the RAM 15 (step s9). That is, the analysis unit 31 reads out the level data and the loopback data from the RAM 15 and enters a predetermined procedure of a program for reediting and updating the LUT.

In the following, modifications related to level data processing will be described.
In FIG. 3B, the GW is opened at the timing t10 during continuous burst transmission, and the process of calculating and outputting the average power value Ao by accumulating the input signal level is performed when 4500 samplings are collected. .

  As another method, a process of calculating the provisional average power value a0 and the like is started from the beginning of the level signals input before 4500 input level signals are collected and continuously performed. Then, the determination processing unit 14 immediately checks whether the average power value Ao exceeds the average power threshold Vath when accumulation of 4500 samples is completed, and determines whether or not to output the data held in the RAM 15. May be performed.

  When the stable input data is continuously acquired, the timing t12 after 45 μsec has elapsed, and when it reaches a predetermined cumulative number, for example, 4500, it is output to the determination processing unit 14 as the average power value Ao. After that, the next 4500 items are counted. The output update of the RAM read flag F is performed in synchronism with the timing of counting 4500 pieces.

  As another method, after reaching 4500, the average power value Ao for every 4500 samples is compared with the average power threshold value Vath each time, and even if the average power value Ao exceeds a predetermined level, it is not immediately compared with the count number exceeding the peak threshold value Vpt. Hold as is. If it is the longest, the latest level data is always written to the RAM 15 one after another until the timing t11 when the GW is closed, and the top (oldest) level data exceeding 4500 is repeatedly discarded. In this case, the latest count value at 45 μsec is updated and input from the level comparison unit 16 by a timer (not shown). The determination processing unit 14 monitors the count number of the peak level, and outputs the RAM read flag F at the timing when the peak level reaches a specified number of times corresponding to the current average power value Ao range.

  In this method, even when burst transmission starts after opening the GW, the LUT 32 can be updated at a fast timing, and accurate DPD can be realized more quickly.

  As yet another method, the determination processing unit 14 preliminarily sets and stores data conditions that can be compared with the LUT 32 on the statistical distribution, for example, within 1σ to 3σ. The determination processing unit 14 monitors the spread of 1σ of the statistical distribution obtained when calculating the average power value Ao, and when the input level data spreads, for example, 9 dB or more, even if it is greater than a predetermined count number The output of the RAM read flag F may be stopped because the data is too discrete. In this method, when there is an unstable transient response at the burst transmission start timing signal, it is possible to remove data that is likely to cause an error.

  The analysis unit 31 that has received the level data and the loopback data from the RAM 15 by the above procedure compares and collates with the predistortion control data of the LUT 32 corresponding to the input level. Then, the predistortion control data Ds is calculated from the level data and the loopback data so that the output signal (loopback signal) falls within the required distortion by a predetermined inverse characteristic calculation program, and the LUT 32 is overwritten (step s10). .

  When the level data and the loopback data are correctly corrected, that is, when a predetermined correspondence (amplitude, phase) is obtained so as to be on the dotted input / output straight line in FIG. Therefore, it is determined that the DPD process is properly performed, and the data in the LUT 32 is maintained as it is. On the other hand, if the error is out of the predetermined error range, the LUT 32 is overwritten and updated to further compensate whether the compensation is insufficient or excessive.

  This inverse characteristic calculation program compares the input signal level with the loopback data, which is the output signal level of the final amplifier, and preliminarily adjusts the gain characteristic of the preamplifier so that the input level and the output level are kept linear. The distortion control data is calculated. At this time, the analysis unit 31 reads the predistortion control data written in the LUT 32 and overwrites it with new predistortion control data in which the gain adjustment amount is corrected.

  The concept to be corrected will be described. For example, the gain is 10 in the input level range of the rated operation. If the power of the level data read from the RAM by the analysis unit 31 is “8”, the power of the loopback data is “80” if the linearity operation is performed. However, if the read loopback data is “72”, it is determined that the gain needs to be increased by about 11% from the current level, and the LUT 32 is further referred to. Assuming that “1.1” has already been written in the LUT 32 as a coefficient for adjusting the gain, that is, the predistortion control data Ds, the analyzing unit 31 performs predistortion control on the level data whose power is “8”, that is, the input signal. The data Ds is corrected to 1.1X1.1 = 1.21 and written to the LUT 32 and updated. If the relationship between the level data and the loopback data is within the required range, either the predistortion control data is not recalculated and the LUT 32 is maintained without being overwritten, or is recalculated and overwritten. But it goes without saying.

  In addition, the gain characteristic adjustment / compensation method has been described above, but the predistortion control data may be adjusted similarly for the phase.

  Thus, in the ideal example shown in FIG. 3B, the accumulated number reaches 4500 when 45 μsec or more elapses (timing t12) from the timing when the GW that starts sampling during burst signal transmission is opened. A power value Ao is obtained.

  After that, the average power is also sequentially updated until the end of sampling (timing t11), and the level data input during that time is compared with the loopback data by the analysis unit 31, and the predistortion control data Ds is continuously updated correctly.

  Next, in the example shown in FIG. 3C, burst transmission has not yet started at the start timing t20 when the GW is opened, and burst transmission is finally started at timing t00, but is lower than the threshold value. Since only the level data is acquired, timing t22 is reached and the threshold is gradually exceeded. Then, the provisional average value ao is calculated, but the GW may end at timing t21 before 45 μsec elapses.

  In principle, including the example shown in FIG. 3B, the level data acquired 45 seconds before the start of the sample is small in number of acquisitions and insufficiently compared with the distribution state. Data is not compared with the above described procedure.

  In this case, although the number of samples is less than the required 4500, the provisional average power value a0 may be obtained from a plurality of samples and compared with a predetermined threshold value Vath. In this case, the count number exceeding the peak level threshold value Vpt is also reduced at the same ratio as the sample number reduction ratio. It is possible to avoid the influence of the noise-like input signal level by determining whether or not to accept data based on the determination criteria based on two kinds of thresholds and, although small, by the average and smoothing effect of the data.

  On the other hand, when the required average number of samples is insufficient as shown at timing t21 in FIG. 3B, when it is found that the GW is closed after 200 μsec, the GW is kept open until 45 μsec has elapsed, or the resampling is immediately performed. You may start. In this case, since the determination processing unit 14 monitors the cumulative number notified from the power calculation unit 13, the determination processing unit 14 determines that the cumulative number is insufficient, invalidates the sampling in this GW, and again executes the GW. Perform the opening process.

  In the GW that is reopened, the sample data is cleared (discarded) and the accumulation is performed again. Alternatively, when level data exceeding the average power threshold Vath is continuously input, the average power value average Ao is calculated by further extending the GW, for example, by 45 μsec or 200 μsec and accumulating the level data. Anyway.

  In the case of FIG. 3D, burst transmission ends while the GW is open, but the determination processing unit 14 monitors the accumulated number and the level data that is permitted to be calculated before reaching the prescribed number is a predetermined level. If the input is stopped for more than the number of times, either the GW is closed and the update process of the LUT 32 is finished, or the GW is opened again and either of the commands for starting the sampling is output. Since the processing procedure can be easily estimated from the above description, a detailed description thereof will be omitted.

  According to the power amplifier, the DPD circuit, and the predistortion compensation method of at least one embodiment described above, the LUT can be accurately determined by sampling the level data of the input modulation signal continuously for a predetermined period and performing statistical processing. Since the pre-distortion compensation data is updated, it is possible to amplify the burst signal transmitted at random timing by applying accurate DPD.

  Although several embodiments of the present invention have been described, these embodiments are presented by way of example and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the spirit of the invention. These embodiments and their modifications are included in the scope and gist of the invention, and are also included in the invention described in the claims and the equivalents thereof.

DESCRIPTION OF SYMBOLS 1 Selection processing part 11 Sample part 12 Level comparison part 13 Power calculation part 14 Determination processing part 15 RAM
16 Input power calculation unit 2 DPD (Digital Pre Distortion) unit 20 Preamplifier 3 LUT generation unit 31 Analysis unit 32 LUT (Look Up Table)
4 D / A (digital / analog converter)
5 Final stage amplifier 6 A / D (analog / digital converter)

Claims (5)

A wideband modulated input signal to be transmitted in a burst is monitored, and a signal amplified by DPD means for applying DPD to the input signal with reference to an LUT storing predistortion control data applied to the input signal is output. In the power amplifier,
A sample that receives the input signal and the amplified signal, samples the signal level at a predetermined sampling period and a predetermined sampling rate, and outputs level data and identification information for each sample and loopback data Means,
Storage means for storing and holding a predetermined number of the output level data and loopback data;
Analysis means for receiving the level data and the loopback data, calculating the predistortion data to be subjected to the DPD, and writing and storing the calculated predistortion data in the LUT;
Power calculation means for inputting the level data, calculating and outputting an average power value obtained by cumulative addition and division by the predetermined number;
Signal level comparison means for inputting the level data, comparing the inputted level data with a predetermined peak power threshold, and counting and outputting the number of times the threshold is exceeded during the cumulative addition;
Controlling the start and end timing of the sample measurement;
When the average power value and the number of counts are input, the average power value is equal to or greater than a predetermined average power threshold value, and the count number exceeds a predetermined number of times, the analysis unit stores the storage unit. A power amplifier comprising: determination processing means for performing control to instruct to read out the stored level data and loopback data. A power amplifier that monitors an input signal to be transmitted in a burst modulated in a wide band and refers to an LUT that stores predistortion control data to be applied to the input signal, and amplifies the input signal by applying DPD and outputs the amplified signal In the LUT control circuit for writing the predistortion control data to the LUT,
The input signal and the amplified signal are received, the signal level is sampled at a predetermined sampling period and a predetermined sampling speed, and the level data and identification information for each sample are output. Sample means;
Storage means for storing and holding a predetermined number of the output level data and loopback data;
Analysis means for receiving the level data and the loopback data, calculating the predistortion data to be subjected to the DPD, and writing and storing the calculated predistortion data in the LUT;
Power calculation means for inputting the level data, calculating and outputting an average power value obtained by cumulative addition and division by the predetermined number;
Signal level comparison means for inputting the level data, comparing the inputted level data with a predetermined peak power threshold, and counting and outputting the number of times the threshold is exceeded during the cumulative addition;
Controlling the start and end timing of the sample measurement;
When the average power value and the number of counts are input, the average power value is equal to or greater than a predetermined average power threshold value, and the count number exceeds a predetermined number of times, the analysis unit stores the storage unit. An LUT control circuit for a power amplifier, comprising: determination processing means for performing control for instructing to read the stored level data and loopback data. Sample means, storage means, power calculation means, signal level comparison means, determination processing means, analysis means, LUT, and amplification means for applying DPD according to predistortion control data and amplifying,
The amplification means monitors the input signal to be transmitted in a burst modulated in a wide band, refers to the LUT storing predistortion control data corresponding to the input signal, and amplifies the DPD on the input signal In a predistortion compensation method for a power amplifier that outputs a signal,
The sample means comprises:
The input signal and the amplified signal are received, the signal level is sampled and measured at a predetermined sample period and a predetermined sampling rate, and level data and loopback data with identification information for each sample are output,
The storage means stores and holds a predetermined number of the output level data and loopback data,
The power calculation means includes
The level data is input, an average power value obtained by cumulative addition and division by the predetermined number is calculated and output,
The signal level comparison means
The level data is input, the input level data is compared with a predetermined peak power threshold, and the number of times the threshold is exceeded during the cumulative addition is output,
The determination processing means includes:
Controlling the start and end timing of the sample measurement;
When the average power value and the number of counts are input, the average power value is equal to or greater than a predetermined average power threshold value, and the count number exceeds a predetermined number of times, the analysis unit stores the storage unit. Control to instruct to read the stored level data and loopback data,
The analyzing means that has received the instruction includes:
The stored level data and loopback data are read from the storage means, the predistortion data to be subjected to the DPD is calculated, and the calculated predistortion data is written and stored in the LUT. Predistortion compensation method. The peak power threshold is set to a plurality of levels corresponding to the average power value,
4. The predistortion compensation for a power amplifier according to claim 3, wherein the determination processing unit gives the read instruction when the number of times exceeds the specified number of times exceeding the peak power threshold corresponding to the level of the average power value. Method. The power calculation means includes
Outputting statistical distribution information of the level data to the determination processing means;
The determination processing means discards the level data from which the average power is calculated from the storage means when the spread of the statistical distribution exceeds a range of predetermined level data. 5. A predistortion compensation method for a power amplifier according to item 4.

Images