Gain transient response compensation

Abstract

A method, apparatus, and computer program are described, comprising: determining an absolute gain of a power amplifier over time, wherein the absolute gain is formed by dividing a feedback baseband signal, which is obtained from an output of the amplifier (e.g., by demodulating an RF signal), by a forward baseband signal, the forward baseband signal being used to form an input to the power amplifier; determining a relative gain transient response (GTR) of the power amplifier by normalizing the absolute gain to generate a relative gain of the power amplifier over time; and determining a gain transient response compensation coefficient, the gain transient response compensation coefficient having an inverse characteristic of the relative gain transient response.

Description

Technical Field

[0001] The present description relates to transient response compensation of a power amplifier, for example to a power amplifier forming part of a communication system. Background Art

[0002] In some applications, a power amplifier (eg, a power amplifier in a transmission system) outputs data in framed burst signals. If the gain of the amplifier is not stable for the duration of each burst, transient changes may occur. Summary of the invention

[0003] In a first aspect, the present specification describes an apparatus comprising: a component for determining an absolute gain of a power amplifier over time, wherein the absolute gain is formed by dividing a feedback baseband signal from an output of the amplifier (e.g., by demodulating an RF signal), the forward baseband signal being used to form an input of the power amplifier (e.g., after conversion to RF after modulation); a component for determining a relative gain transient response (GTR) of the power amplifier by normalizing the absolute gain to generate a relative gain of the power amplifier over time; and a component (e.g., a transient response calculation unit) for determining a gain transient response compensation coefficient having a characteristic that is opposite to the relative gain transient response.

[0004] Means for determining an absolute gain of the power amplifier over time may include means for averaging the feedback baseband signal divided by the forward baseband signal over successive sampling intervals (eg, an average absolute gain may be generated for each sample).

[0005] The means for determining the relative gain transient response of the power amplifier may include means for normalizing the absolute gain relative to a reference gain. The reference gain may be determined at a time when the absolute gain is considered to have stabilized. The reference gain may be determined at a time when the absolute gain is determined to have stabilized. In addition, means for determining the time when the absolute gain has stabilized is provided, the means including a reference detection unit, the reference detection unit being configured to determine a sampling period during which the power level of the forward baseband signal has maximum stability.

[0006] Some embodiments may include means (e.g., a transient pre-compensation unit) for modifying a baseband forward signal using the determined gain transient response compensation coefficient to generate a compensated baseband forward signal (thereby compensating for the gain transient response of the power amplifier). For example, the baseband forward signal may be multiplied by the determined gain transient response compensation coefficient to generate the compensated baseband forward signal.

[0007] Some embodiments may include means for storing and updating gain transient response compensation coefficients (eg, a gain transient response compensation value update and retention unit). The means for storing and updating gain transient response compensation coefficients may update a gain transient response compensation coefficient according to the length of another compensation coefficient.

[0008] The common update cycle index (s_com) may be determined according to the following relationship between the reference index of the previous coefficient (s_ref_pre) and the reference index of the current residual gain (s_ref):

[0009]

[0010] Wherein: the gain transient compensation coefficient is updated within a common update cycle, calculated by using a previous compensation coefficient in the common update cycle and a current residual gain, wherein an error in the current residual gain is calculated by the difference in the residual gain from an index (s_com), and is controlled by multiplying the update coefficient by the error; and a portion of the coefficient is combined with an updated coefficient in the common cycle, a portion of the coefficient is outside the common update cycle, and a portion of the coefficient has a length longer than a portion of a previous compensation coefficient or a portion of the current residual gain coefficient.

[0011] In some embodiments, the feedback baseband signal is a composite signal having in-phase and quadrature components and / or the forward baseband signal is a composite signal having in-phase and quadrature components and / or the baseband forward signal is a composite signal having in-phase and quadrature components.

[0012] The power amplifier may be an RF signal amplifier unit (eg, a wireless transmission device for sending a burst signal).

[0013] The input signal to the apparatus may be a burst mode signal (such that the power amplifier is a burst mode power amplifier).

[0014] The apparatus may include at least one processor; and at least one memory including computer program code, wherein the at least one memory and the computer program code are configured to, together with the at least one processor, cause execution of the apparatus.

[0015] In a second aspect, the present specification describes a method comprising: determining an absolute gain of a power amplifier over time, wherein the absolute gain is formed by dividing a feedback baseband signal by a forward baseband signal, the feedback baseband signal being obtained from the output of the power amplifier, and the forward baseband signal being used to form the input of the power amplifier; determining a relative gain transient response of the power amplifier by normalizing the absolute gain to generate a relative gain of the power amplifier over time; and determining a gain transient response compensation coefficient, the gain transient response compensation coefficient having a characteristic opposite to the relative gain transient response.

[0016] Determining the absolute gain of the power amplifier over time may include averaging the feedback baseband signal divided by the forward baseband signal over successive sampling intervals (eg, an average absolute gain may be generated for each sample).

[0017] Determining the relative gain transient response of the power amplifier may include normalizing the absolute gain relative to a reference gain. The reference gain may be determined at a time when the absolute gain is considered to have stabilized. The reference gain may be determined at a time when the absolute gain is determined to have stabilized. In addition, a component for determining the time when the absolute gain has stabilized is provided, the component including a reference detection unit, the reference detection unit being configured to determine a sampling period during which the power level of the forward baseband signal has maximum stability.

[0018] The method may further include modifying the baseband forward signal using the determined gain transient response compensation coefficient to generate a compensated baseband forward signal (thereby compensating for the gain transient response of the power amplifier).

[0019] In a third aspect, the specification describes an apparatus configured to perform any method described with reference to the second aspect.

[0020] In a fourth aspect, the present specification describes computer readable instructions which, when executed by a computing device, cause the device to perform the method described with reference to the second aspect.

[0021] In a fifth aspect, the specification describes a computer program including instructions stored thereon for performing at least the following operations: determining an absolute gain of a power amplifier over time, wherein the absolute gain is formed by dividing a feedback baseband signal by a forward baseband signal, the feedback baseband signal being obtained from the output of the power amplifier, and the forward baseband signal being used to form the input of the power amplifier; determining a relative gain transient response of the power amplifier by normalizing the absolute gain to generate a relative gain of the power amplifier over time; and determining a gain transient response compensation coefficient, the gain transient response compensation coefficient having a characteristic opposite to the relative gain transient response. The computer program may further include instructions stored thereon for modifying a baseband forward signal using a gain transient response compensation value to generate a compensated baseband forward signal (thereby compensating for the gain transient response of the power amplifier). BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Example embodiments will now be described, by way of example only, with reference to the following schematic diagrams, in which:

[0023] Figure 1a is a block diagram of a system according to an example embodiment;

[0024] Figure 1b is a flow chart illustrating an algorithm according to an example embodiment;

[0025] Figure 2 is a block diagram of a system according to an example embodiment;

[0026] Figure 3 As shown in Figure 1 and Figure 2 Output generated from example uses of the system;

[0027] Figure 4 Figure 1 and Figure 2 Output generated from example uses of the system;

[0028] Figure 5 As shown in Figure 1 and Figure 2 Data points generated from example uses of the system;

[0029] Figure 6 is a block diagram of a system according to an example embodiment;

[0030] Figure 7 , Figure 8 , Figure 9a , Figure 9b , Fig.10a and Fig.10b The data generated in the example embodiment is shown.

[0031] Fig.11 is a block diagram of a system according to an example embodiment;

[0032] Fig.12 is a block diagram of a system according to an example embodiment;

[0033] Fig.13 is a block diagram of a system according to an example embodiment;

[0034] Fig.14 shows signals generated according to an example embodiment;

[0035] Fig.15 is a block diagram of a system according to an example embodiment;

[0036] Fig.16 shows symbols generated according to an example embodiment;

[0037] Fig.17 shows signals generated according to an example embodiment;

[0038] Fig.18 is a block diagram of a system according to an example embodiment;

[0039] Fig.19 shows symbols generated according to an example embodiment;

[0040] Fig. 20 shows data generated according to an example embodiment;

[0041] Figures 21a to 21c shows data generated according to an example embodiment;

[0042] Figure 22 to Figure 26 is a block diagram of a system according to an example embodiment;

[0043] Fig. 27 is a block diagram of a system according to an example embodiment; and

[0044] Fig.28a and Fig.28b Tangible media are shown, respectively, a removable memory unit and a compact disk (CD) storing computer readable codes, which when executed by the computer, perform operations according to the embodiments. DETAILED DESCRIPTION

[0045] Figure 1a is a block diagram of a system generally indicated by reference numeral A10 according to an example embodiment. System A10 includes RF modulation circuitry 302, RF demodulation circuitry 304, power amplifier 3, and signal acquisition circuitry 11.

[0046] System A10 receives a forward IQ signal modulated by RF modulation circuit 302 to provide a radio frequency (RF) signal provided to power amplifier 3. Power amplifier 3 amplifies the RF signal to provide an RF output. The forward IQ signal is provided as a first input to signal acquisition circuit 11. The RF output is demodulated by RF demodulation circuit 304, and the resulting feedback IQ signal is provided as a second input to signal acquisition circuit 11.

[0047] As described in detail below, the forward IQ signal and the feedback IQ signal may be compared to determine various aspects of the characteristics of the power amplifier 3 .

[0048] Figure 1b is a flow chart illustrating an algorithm generally indicated by reference numeral A15 used according to an example of system A10. Algorithm A15 begins at operation A16, in which the feedback IQ signal is divided by the forward IQ signal to provide an absolute gain of the power amplifier 3. Next, at operation A17, the absolute gain is normalized to generate a relative gain transient response of the power amplifier 3 over time. At operation A18, a gain transient response compensation value for the power amplifier 3 is generated. Finally, at operation A19, the generated gain transient response compensation value can be used to modify the forward IQ signal to compensate for the gain transient response of the power amplifier 3. More details of the functionality and implementation of operations A16 to A19 are provided below.

[0049] Figure 2is a block diagram of a system generally indicated by reference numeral A20 according to an example embodiment. System A20 may be, for example, a wireless transmission device for studying the transient response of a power amplifier. As described in detail below, device A20 includes the RF modulation circuit 302, RF demodulation circuit 304, power amplifier 3, and signal acquisition circuit 11 of the above-mentioned system A10.

[0050] System A20 shows the generation of transient response characteristics associated with power amplifier 3. System A20 includes a channel filter unit 4 that converts an input waveform (such as Figure 2 The invention relates to an IQ baseband signal 310 as shown in the figure, which is filtered to a transmission channel frequency band; a crest factor reduction (CFR) circuit 5, which suppresses the peak in the IQ baseband signal; a digital pre-distortion (DPD) circuit 6, which can be used to compensate for the distortion of the power amplifier; a power amplifier 3, which amplifies the transmission signal; a directional coupler 7, which obtains an RF feedback signal of the RF output signal; a signal capture circuit 11; and a radio frequency integrated circuit (RFIC) 2, including a digital-to-analog (D / A) converter 301 and an RF modulation circuit 302, which are used to convert the composite baseband IQ signal into an RF signal of a desired carrier frequency for transmission, and an RF demodulation circuit 304 and a digital-to-analog converter 303, which are used to convert the RF feedback signal into a composite baseband signal.

[0051] The signal capture unit 11 captures the forward IQ signal and the feedback IQ signal to calculate the gain transient response (GTR) characteristic (thereby implementing the above-mentioned operation A16). Here, the feedback IQ signal converted from the RF feedback signal to the composite baseband signal by the RF demodulator 304 and the A / D converter 303 contains the same GTR characteristic as the RF output, which is used to calculate the GTR.

[0052] Figure 3 An output generally indicated by reference numeral A30 generated in an exemplary use of the above described system A10 and system A20 is shown. The output A30 shows a time series of the amplitudes of the forward IQ signal and the feedback IQ signal.

[0053] The amplitude of the forward IQ signal (at Figure 3 The amplitude of the feedback IQ signal (as shown in the upper figure of Figure 3 ) has a peak level with a variable amplitude variation at the beginning of each burst.

[0054] In the following example embodiments, a method for accurately calculating the GTR using a forward IQ signal and a feedback IQ signal is provided.

[0055] The amplitude of the forward IQ signal varies sample by sample because the forward IQ signal is a modulated signal with a bandwidth, and the LTE 20MHz bandwidth is shown here. It may be necessary to make a sampling timing adjustment between two signals (e.g., a first signal and a second signal) (such as a feedback IQ signal and a forward IQ signal) in advance to calculate the inherent gain between the two signals. M upsampling and convolution processing are performed, and the sampling timing is adjusted at a ratio of M times from the original sampling ratio with the estimated timing difference. After upsampling, a timing adjustment is performed between the first signal and the second signal. The timing adjustment process described in this article is for illustrative purposes only. The summary of the timing adjustment is shown in the following equation E1.

[0056] Let t represent the sampling index before M-fold upsampling, and let m represent the sampling index after M-fold upsampling. Thus, the first signal before M-fold upsampling is represented as Sb(t), and the first signal after M-fold upsampling is represented as Sb(m). Thus, the second signal before M-fold upsampling is represented as Sf(t), and the second signal after M-fold upsampling is represented as Sf(m). "Nsp" represents the nth sample, i.e., a sufficient number of samples to compensate for GTR from the beginning of each burst.

[0057] E1:

[0058]

[0059] Before upsampling: t = 1, ..., Nsp

[0060] After upsampling: m = 1, ..., Nsp x M (1)

[0061] After the timing adjustment between the first signal and the second signal is performed at high resolution, the upsampled first signal Sb(m) and the upsampled second signal Sf(m) are performed in equation E2. In the upsampled time series, the timing deviation between the first signal and the second signal is represented by ⊿m. The convolution process function is represented by "*". The convolution process function is used to detect the timing deviation (represented by U(⊿t)) through the peak value of the function output.

[0062] E2:

[0063] U(⊿t)=|Sf(m)*Sb(m-⊿t)|

[0064] ⊿m=arg max{U(⊿t)}⊿t=1,…,Nsp x M(2)

[0065] Downsampling is then performed at a ratio of "1 / M" times to restore the original sampling rate, as shown in Equation E3. The downsampled signals after the timing adjustment of the first signal and the second signal are represented by Sf(i), Sb(i), respectively, where the sampling index is represented by "i". The sampling index ranges from 1 to Nsp (i=1, ..., Nsp).

[0066] E3:

[0067]

[0068] The down-sampled signal Sf(i) is a forward IQ signal obtained after down-sampling. The down-sampled signal Sb(i) is a feedback IQ signal obtained after down-sampling. The sampling timing can be adjusted by the sampling index represented by "i".

[0069] In this exemplary embodiment, the ratio M is set to 8. However, the ratio M is not limited to 8.

[0070] Equation E4 shows the division process. The division process is performed in order to calculate the absolute GTR. The absolute GTR represented by "aG(i)" is the gain of the feedback IQ signal relative to the forward IQ signal as a reference (see the above-mentioned operation A16).

[0071] E4:

[0072] aG(i)=Sb(i) / Sf(i)i=1,…,Nsp(4)

[0073] Figure 4 The outputs, generally indicated by reference numeral A40, generated in an exemplary use of system A10 and system A20 are shown. The outputs show the absolute gain transient response (GTR) for each sample calculated from the forward IQ signal and the feedback IQ signal. The upper portion of output A40 shows the amplitude component, while the lower portion of output A40 shows the phase component. Sb(i) divided by Sf(i) is a division of vectors, and therefore any process errors present in the outputs may depend on the amplitude range of the samples. As shown in FIG. Figure 4 As shown, the error causes the data points to be spread relatively widely.

[0074] In one variation, the average absolute GTR ("aG_ave") is calculated using the absolute gain of the sampling reference and a predetermined sampling interval ("Nd") as shown in Equation E4a below. In Equation E4a, the index of the sampling interval, i.e., the interval index, is represented by "u", and the average absolute GTR of the sampling interval is represented by aG_ave(u). i_itv(u) represents the target sampling index for averaging at interval index "u", where i_itv(u) ranges from i_start(u) to i_end(u) (i_itv(u) = i_start(u), ..., i_end(u)). The number of samples averaged at interval index "u" is represented by Nave(u) (Nave(u) = i_end(u) - i_start(u) + 1).

[0075] E4a:

[0076] Average absolute GTR:

[0077] Figure 5 Data points generated in an example use of system A10 and system A20 are shown, generally indicated by reference numerals A50 and A52. Data points A50 and A52 illustrate the GTR averaged for each interval.

[0078] Data point A50 shows the segmented average absolute GTR for the sampled basis and the predetermined sampling interval length ("Nd") calculated in equation E4a. Figure 5 , Nd is 2560 (as an example). In the illustration of data point A50, the scaling ratio of the vertical axis (representing the amplitude component (upper part) or the phase component (lower part)) is the same as that of the example output A40. Data point A52 is the same as data point A50. In the illustration of data point A52, the scaling ratio of the vertical axis is larger than that of A50 so that A52 is magnified to focus on the changes of each data point along the vertical axis. According to data point A52, both the amplitude and the phase vary within the range of one subframe. A subframe consists of 14 orthogonal frequency division multiplexing (OFDM) symbols.

[0079] In an exemplary embodiment, when compensation of the GTR of the RF output is performed in advance using the GTR characteristics obtained in the above embodiments (operation A19 of the above-mentioned algorithm A15), the GTR characteristics may be used as a compensation coefficient series for GTR compensation (GTC) instead, and the transmission signal may be pre-compensated using the compensation coefficient series before inputting to the digital-to-analog (D / A) converter 301. The compensation coefficient series is as follows. For simplicity, the term "symbol" is used to represent an OFDM symbol, the term "symbol gain" is used to represent the average gain of an OFDM symbol, and the term "symbol power" is used to represent the average power of an OFDM symbol.

[0080] In an example embodiment, a series of compensation coefficients is determined using the number of coefficients, the time interval compensated by the coefficients, and the compensation range from the start of a burst represented by a predetermined time period. There are three parameters, namely a first parameter, a second parameter, and a third parameter, used to calculate the series of compensation coefficients. The first parameter is the time interval from the start of the burst to the time when compensation is performed. The second parameter is the number of compensation coefficients. The third parameter is the length of the time interval compensated by the compensation coefficients.

[0081] In one variant, with respect to the first parameter, in order to reduce the computational load associated with the compensation, compensation may not be performed for all signals over the entire time interval from the start of the burst. The time interval used may be the time interval in which the transient response occurs, which may reduce the number of compensation coefficients to be used in the series of compensation coefficients. With respect to the second and third parameters, the interval average of the symbols of the transient response of the sampled reference is calculated, thereby providing one compensation coefficient for each symbol.

[0082] For example, as shown by data points A50 and A52, the GTR converges within a subframe length, which is approximately equal to the range from the starting symbol to the fourteenth symbol. Therefore, the absolute gain of the fifteenth symbol is taken as a reference gain (and can be used for normalization, as discussed further below). With respect to the absolute gain of the fifteenth symbol as a reference, a series of relative GTRs are obtained.

[0083] For the purpose of providing an example embodiment, the gain characteristic is considered as a relative gain transient characteristic. For the absolute gain, the forward portion is normalized by a reference gain. The reference gain is taken from the absolute gain in the transient response convergence region.

[0084] The inverse characteristic of the relative GTR for pre-compensating the GTR is described below. As described above, the inverse characteristic of the relative GTR is calculated as a compensation coefficient series. The baseband IQ signal to be transmitted is pre-compensated by using the compensation coefficient.

[0085] In an exemplary embodiment, as described below, the average value in the interval of the symbol duration of the sample-based absolute gain (aG(i)) is calculated according to Equation E5. For a predetermined number of symbols (Nsym), the symbol-based absolute GTR is represented by aG_ave(s). The symbol index is represented by "s". For illustrative purposes, the predetermined number of symbols Nsym is equal to fifteen. The absolute gain is calculated for the predetermined number of symbols from the beginning. The target sample index for averaging at symbol index "s" is represented by i_sym(s), where i_sym(s) ranges from i_start(s) to i_end(s) (i_itv(u) = i_start(u), ..., i_end(u)). The number of samples averaged at symbol index "s" is represented by Nave (Nave = i_end(s) - i_start(s) + 1).

[0086] E5:

[0087] Symbol Absolute GTR: Symbol index: s = 1, ..., Nsym

[0088] After calculating the symbol-based absolute GTR, the symbol absolute gain transient response from the start symbol to the symbol before the reference symbol is normalized based on the division shown in Equation E6. The absolute gain aG_ave(s_ref) is used as the reference symbol (in this example, s_ref is equal to 15).

[0089] E6:

[0090] Symbol reference relative GTR: rG_ave(s) = aG_ave(s) / aG_ave(s_ref)

[0091] s = 1, ..., s_ref (6)

[0092] Figure 6 is a block diagram of a system, generally indicated by reference numeral A60, according to an example embodiment. The system A60 provides a process for calculating a relative gain. Thus, the symbol reference relative GTR rG_ave(s) is calculated using a gain relative to a detected reference gain.

[0093] like Figure 6 As shown, the subframe control signal indicates the start of a burst. The fifteenth symbol of the burst (in this example embodiment) is used as a reference gain and is used for conversion of the average absolute gain to provide a relative gain (operation A17 as described above).

[0094] Figure 7Data generated according to an example embodiment is shown, generally indicated by reference numerals A70 and A72. Data A70 shows the absolute GTR (operation A16) for each symbol. The absolute GTR uses the symbol base gain of the feedback IQ signal as a reference for the forward IQ signal. The upper portion of A70 and A72 shows the amplitude component of the composite gain, while the lower portion of A70 and A72 shows the phase component of the composite gain. Data A72 shows the relative GTR (operation A17) for each symbol. The relative GTR shows the gain in the absolute GTR normalized by the fifteenth symbol of the absolute gain as the reference gain. As shown in data A72, for the fifteenth symbol, the amplitude component (upper portion) is 1, while for the fifteenth symbol, the phase component is zero. Both the amplitude component and the phase component are shown as changing to about the fourteenth symbol. The amplitude component decreases at the fifth symbol. The amplitude component of the fifth symbol is 0.975, which is the minimum value of a given symbol. After the fifth symbol, the amplitude component continues to increase and approaches a value. The phase has a maximum error at the fifth symbol and then increases toward zero.

[0095] Figure 8 Generated data according to an example embodiment is shown, generally indicated by reference numeral A80. As described above, the characteristics of the relative GTR change after compensation is performed. Thus, the relative GTR becomes closer to a gain value where the amplitude component is 1 and the phase component is zero.

[0096] In the above example embodiment, one of the benefits of detecting the position of the reference gain and normalizing the forward portion of the absolute gain is that the response can be compensated without a discontinuity in the RF output signal between the forward compensated portion of the reference symbol and the backward non-compensated portion of the reference symbol. This is because the magnitude and phase of the gain in the reference symbol are "1" and "0", respectively.

[0097] Figure 9a , Figure 9b , Fig.10a and Fig.10b Data generated according to an example embodiment are shown, generally indicated by reference numerals A90, A92, A100, and A102, respectively. Data A90 and A100 show the measurement results of the constellation diagram of the RF output signal, and data A92 and A102 show the measurement results of the symbol error vector magnitude (EVM) of the RF output signal. Data A90, A92, A100, and A102 collectively show the effect of the gain transient response compensation described herein.

[0098] Figure 9a and Figure 9bOutputs A90 and A92 without compensation are shown. As shown, the transient response has an impact on the start of the burst generation, where a decrease in the symbol error vector magnitude (EVM) can be seen. As a result, both the amplitude component and the phase component of the signal point in the constellation gradually deviate from the ideal signal point in the time transient. It should be noted that if the deviation of the signal point occurs on the transmitter side and cannot be corrected in the demodulation process on the receiver side, the bit detection of the signal point will include errors and the throughput performance will be reduced.

[0099] Fig.10a and 10b Outputs A100 and A102 are shown, where compensation has been pre-performed by multiplying the compensation coefficient by the IQ baseband signal 310 to be transmitted. As shown, the symbol error vector magnitude (EVM) and the deviation of the signal point in the constellation diagram are improved. In this way, by using the compensation coefficient based on the inverse characteristic of the relative GTR, the improvement of the symbol EVM is obtained by compensating the GTR in advance.

[0100] Therefore, by calculating the relative GTR referenced to a specific portion of the complex absolute gain in the time series where the transient response converges, Figure 3 The transient response characteristics of the RF output signal shown are achieved. With respect to the calculation of the relative gain, the process of timing adjustment at a high ("M" times) sampling rate brings high accuracy in the gain calculation in both amplitude and phase. In addition, the GTR of the RF output signal can be compensated by multiplying the inverse characteristic of the relative GTR by the previously transmitted IQ baseband signal.

[0101] The compensation in the above example embodiments may be performed using digital signal processing that may be implemented in an integrated circuit (IC) device, such as a field programmable gate array (FPGA), a digital signal processor (DSP) device, and an SoC (system on a chip). It should also be noted that the provision of the RFIC unit 2 represents one example implementation. By way of example only, alternative implementations may implement the features of the RFIC unit in other ways, such as as part of an IC device that implements other features of the system A20.

[0102] In order to overcome problems in mass production and actual operation, for example, it may be necessary to adaptively update the compensation coefficient. According to an example embodiment, a GTR compensation circuit for adaptively updating the compensation coefficient is provided below.

[0103] Unless otherwise specified, the test model (bandwidth: 20 MHz) of LTE E-TM3.1a is also used here for the test signal.

[0104] The transient response characteristics of the RF output may be affected by temperature and device-specific changes. The change may be caused by changes in the load transient response characteristics related to the voltage and current supply of various devices such as the power supply unit of the device, the power amplifier unit, the capacitors constituting the peripheral circuit, etc. Since each device has temperature characteristics, they may be affected by the surrounding temperature. The transient response characteristics may also change based on changes in the input signal level of the power amplifier unit. The change in the input signal level may be caused by changes in the transmission service and differences in the number of subcarriers used for transmission. Therefore, for mass production and operation, it is desirable to automatically and adaptively update and optimize the compensation coefficient during the operation of the device.

[0105] The difference in transient response characteristics of the RF output may be caused based on one or more of the first factor, the second factor, and the third factor. The first factor is that the transient response varies from device to device. The second factor is the effect of ambient temperature change on the transient response; the third factor is the effect of input signal level difference on the transient response. The differences in transient response characteristics based on the first, second, and third factors are described below.

[0106] Fig.11 1 is a block diagram of a system generally indicated by reference numeral A110 according to an example embodiment. System A110 shows a configuration of a gain transient response compensation (GTC) circuit 1 according to an example embodiment. (The gain transient response compensation circuit 1 is in Fig.11 The terms "circuit" and "device" are used interchangeably in this document.)

[0107] The GTC circuit 1 includes a gain transient response (GTR) calculation unit 100, a compensation coefficient update and retention unit 201, and a compensation unit 202. The GTR calculation unit 100 calculates a compensation coefficient series, which is the opposite characteristic of the relative GTR from the forward IQ signal 111 and the feedback IQ signal 112. The baseband forward signal 210 (derived from the forward IQ signal 111) is the baseband IQ data to be transmitted, and the feedback IQ signal 112 is the output of the power amplifier unit, as described in detail above. The coefficient update and retention unit 201 updates the compensation coefficient series and stores the updated compensation coefficient series. The compensation unit 202 uses the retained compensation coefficient series to compensate the baseband forward signal 210 to be transmitted. Note that the baseband forward signal 210 is compensated by the compensation coefficient.

[0108] Fig.12 is a block diagram of a system, generally indicated by reference numeral A120, according to an example embodiment. System A120 shows a GTC circuit 1 operating with other devices involved in signal transmission.

[0109] System A120 includes the above-mentioned channel filter unit 4, CFR unit 5 and DPD unit 6, which are provided before the GTC circuit. System A120 also includes the above-mentioned RFIC unit 2, power amplifier 3 and directional coupler 7, which are provided after the GTC circuit. As described above, the RFIC unit 2 converts the baseband IQ signal into an RF signal of the desired carrier frequency, and converts the RF feedback signal from the power amplifier 3 into a composite baseband signal; the power amplifier 3 amplifies the RF signal; and the directional coupler 7 obtains the RF feedback signal from the RF output.

[0110] Next, the detailed operation of the GTC circuit 1 will be described.

[0111] Fig.13 is a block diagram of a system generally indicated by reference numeral A130 according to an example embodiment. System A130 illustrates a GTR calculation unit 100 , which is included within a GTC circuit 1 .

[0112] The input of the GTR calculation unit 100 is a forward IQ signal 112, a feedback IQ signal 111 of a burst signal, and a frame control signal 113. The frame control signal 113 indicates the time region of a time division duplex (TDD) signal. The GTR calculation unit 100 includes a control signal generation unit 101, an absolute gain calculation unit 102, a reference gain detection unit 103, and a relative gain calculation unit 104. The control signal generation unit 101 generates a symbol control signal from a frame control signal. The absolute gain calculation unit 102 calculates the absolute GTR using the forward IQ signal and the feedback IQ signal. The reference gain detection unit 103 detects a reference gain index. The relative gain calculation unit 104 calculates the relative GTR relative to the absolute GTR with reference to the absolute gain at the time position of the reference gain index. The output of the GTR calculation unit 100 is a relative GTR 117 in the start region of the burst signal and the burst information 118. The process for generating the relative GTR will be described in detail below.

[0113] Fig.14 The signal generated by the control signal generation unit 101 according to an example embodiment is shown, which is generally represented by the reference numeral A140. The signal A140 includes a frame control signal, a symbol control signal generated by the control signal generation unit 101, and corresponding forward and feedback IQ signals. The symbol control signal generated by the control signal generation unit 101 indicates the symbol timing of the modulated signal based on the frame control signal and the signal information. The signal information contains signal information about the signal format, such as the modulation parameters of the transmitted signal, the symbol duration, and the burst signal duration. The boundary of the symbol can be known from the signal information.

[0114] Fig.151 is a block diagram of a system generally indicated by reference numeral A150 according to an example embodiment. System A150 shows an absolute gain calculation unit 102, which is included in the GTC circuit 1. The absolute gain calculation unit 102 calculates a gain series of the feedback IQ signal relative to the forward IQ signal in each sample. This gain is referred to herein as "absolute gain".

[0115] The absolute gain calculation unit 102 includes a timing adjustment unit 1021, a normalization unit 1022 and an averaging unit 1023. The timing adjustment unit 1021 adjusts the sampling timing between the forward IQ signal and the feedback IQ signal. The input of the absolute gain calculation unit includes the feedback IQ signal 111, the forward IQ signal 112 and the average control signal 114. The output of the absolute gain calculation unit 102 is GTR115. For the purpose of illustration, M times upsampling and related processing can be performed. As shown in the following equation, an example for timing adjustment processing. However, any other technology can be used to perform the timing adjustment processing.

[0116] In the following equations E11 and E12, the representation of the samples before up-sampling is represented as “t”, and the representation of the samples after up-sampling is represented as “m.” Nsp is the total number of samples used to compensate for GTR from the start of the burst signal.

[0117] Before upsampling: t = 1, ..., Nsp

[0118] After upsampling: m = 1, ..., Nsp x M

[0119] E11:

[0120]

[0121] E12:

[0122]

[0123] After upsampling, the timing between the two upsampled signals Sb(m) and Sf(m) is adjusted with high resolution. Then (1 / M) times downsampling is performed to restore the original sampling rate again, as shown in the following equations E13 and E14. The timing difference is represented by ⊿m, and the downsampled signals after timing adjustment are represented by Sb(i) and Sf(i) (i=1, ..., Nsp).

[0124] E13:

[0125]

[0126] E14:

[0127]

[0128] Sb(i) and Sf(i) may be obtained such that the timing of Sb(i) and Sf(i) match at the same sampling position "i". For the purpose of illustration, timing adjustment is performed taking M=8 as an example.

[0129] As shown in Equation E15, a division process is performed to calculate an absolute GTR "aG(i)" showing a gain of the feedback IQ signal with respect to the forward IQ signal as a reference.

[0130] E15:

[0131] aG(i)=Sb(i) / Sf(i)i=1,…,Nsp(15)

[0132] The division shown in equation E15 is processed by a complex vector, and the processing error generated in the output depends on the amplitude range of the sample. Therefore, the change of each sample becomes larger for each sample. Therefore, as shown in equation E16 below, the average value of the interval of the symbol duration from the sample-based absolute gain aG(i) is calculated. The symbol-based absolute GTR aG_ave(s) of a predetermined number of symbols Nsym (=15) starting from the beginning of the symbol is obtained. The target sample index for averaging at symbol index "s" is represented by i_sym(s), where i_sym(s) ranges from i_start(s) to i_end(s) (i_itv(u)=i_start(u),...,i_end(u)). The number of samples averaged at symbol index "s" is represented by (Nave) (Nave=i_end(s)-i_start(s)+1).

[0133] E16:

[0134] Symbol Absolute GTR: Symbol index: s = 1, ..., Nsym

[0135] Fig.16 The generated symbols generally indicated by reference numeral A160 according to an example embodiment are shown. Symbol A160 includes symbols before and after the absolute gain of each symbol is averaged by the averaging unit. In the symbol averaging process, the symbol period can be determined by the symbol control signal generated in the control signal generating unit 101.

[0136] At the reference gain detection unit 103, a control signal 114 and a forward IQ signal 112 are received as inputs. In the region where the gain is determined to be stable, a portion considered to be a reference gain is detected from the symbol absolute gain. The output of the reference gain detection unit 103 is a specified reference gain index. Since the target is the absolute gain of each symbol, the reference gain index is the symbol position from the beginning of the burst, and the reference gain index is shown as an integer starting from 1.

[0137] A method for detecting a symbol position of a reference gain is described in detail below according to an exemplary embodiment.

[0138] In earlier example embodiments, the power of each symbol in the transmitted signal was considered equal, and the symbol positions where convergence was assumed to be approximately specified (e.g., Figure 7 In this way, the same symbol position is set to the reference gain. However, in the LTE (Long Term Evolution) system, the number of subcarriers activated in a symbol (the number of subcarriers whose amplitude is non-zero) varies according to the traffic volume. As a result, the symbol power fluctuates in the operating signal according to the number of control channels to be transmitted and the traffic volume of the user. Thus, the fourth factor that may cause a difference in transient response characteristics is the influence of symbol power fluctuations on GTR. Therefore, in the following example embodiment, in addition to the method for determining the reference gain, the influence of symbol power on GTR is also shown.

[0139] Fig.17 The signals generated when the symbol power fluctuates according to an example embodiment are shown, generally indicated by reference numerals A170 and A172. Signal A170 shows the symbol level with the input signal root mean square (RMS). Signal A172 represents the absolute GTR when a signal is received as input.

[0140] For the purpose of illustration, the test signal here is an LTE signal with LTE transmission mode 9. The symbol power of the first symbol and the fifteenth symbol is less than the symbol power of other symbols. In this example, the LTE signal contains a physical downlink control channel (PDCCH) in the first symbol in each subframe. Even if the amount of data in the physical downlink shared channel (PDSCH) is large, there are any unused subcarriers in the PDCCH. In the GTR shown in signal A172, the symbol level (RMS) in the fifteenth symbol is relatively small (about half the size) compared to the previous and next symbols (i.e., the fourteenth symbol and the sixteenth symbol). Therefore, the gain amplitude in the fifteenth symbol is about 0.02 larger. As described above, when the symbol power of the signal input to the power amplifier 3 fluctuates greatly, the GTR also changes.

[0141] When determining the absolute gain of a symbol, if the fluctuation of the symbol power increases before determining the reference gain, the change in the absolute gain in the symbol is superimposed on the relative gain as an offset. As a result, an error is caused in the compensation coefficient value to be calculated, and the GTR of the amplifier cannot be accurately compensated. Therefore, it is advantageous to detect the absolute gain in a stable part with a relatively same power level, and the fluctuation part can be avoided even in the convergence area of ​​the transient response. This problem is solved in the above exemplary embodiment, where the backward part is used to calculate the relative GTR if there is a significant change in the symbol power.

[0142] Fig.18 is a block diagram of a system generally indicated by reference numeral A180 according to an example embodiment. System A180 illustrates the reference gain detection unit 103 included in the GTR calculation unit 100.

[0143] The reference gain detection unit 103 includes a power calculation unit 1031, an averaging unit 1032 and a reference detection unit 1033. The power calculation unit 1031 calculates the sampled power value of the forward IQ signal; the averaging unit 1032 calculates the symbol power by averaging the sampled power in the symbol duration; and the reference detection unit 1033 detects the symbol part of the reference gain from the symbol power.

[0144] Fig.19 The generated symbols according to an example embodiment are shown, generally indicated by reference numeral A190. Symbol A190 shows the calculation of the symbol power in the averaging unit 1032. In the averaging unit 1032, the symbol power (denoted by Pow_ave) is calculated from the sample power (denoted by Pow(i)) based on the symbol control unit.

[0145] The reference gain detection unit 103 measures the maximum power level in the target detection duration, and calculates the power lower limit based on the level fluctuation range (indicated by ΔP) based on the maximum power level as the maximum limit.

[0146] Fig. 20 Data generated according to an example embodiment is shown, generally indicated by reference numerals A200 and A202. Data A202 shows parameters for detecting symbol positions of reference gains based on a time series of symbol powers. Data A200 shows symbol powers in time series, and data A202 shows parameters for detecting symbol positions of reference gains.

[0147] The operation of the reference gain detection unit 103 is described in detail below. The start of a detection interval for detecting a reference symbol is denoted by "Sdet_start", and the end of the detection interval is denoted by "Sdet_end".

[0148] The maximum power level (denoted by "Pmax") in the detection interval (Sdet_start≤s≤Sdet_end) is measured from the time series of symbol power. The minimum power level (denoted by "Pmin") of the level fluctuation amplitude is calculated from the maximum level Pmax. The level fluctuation amplitude ΔP may be a predetermined value.

[0149] The second half of the stable level region can be specified by first counting the number of symbols (denoted by "Nstb") that are continuous within the range of the level fluctuation width ΔP for a predetermined number of symbols Nstb, and then specifying the symbol position where the count value continuously exceeds the predetermined number Ndet. The symbol position of the reference gain is denoted by s_ref.

[0150] The following algorithm illustrates the process for detecting reference symbols. For illustration purposes, the following MATLAB M-file code is provided:

[0151] In the following code, ⊿P is represented by deltaP

[0152] % Parameter configuration

[0153] Nstb = 4;

[0154] PowCount_ref = 6;

[0155] Sdet_start = 1;

[0156] Sdet_end=16;

[0157] deltaP = 1.0;

[0158] %initialization

[0159] s_ref_temp=[];% empty

[0160] s_ref = []; % empty

[0161] % Detection of reference symbol index x Pmax = max(Pow(Sdet_start:Sdet_end));

[0162] for s=Sdet_start:Sdet_end

[0163] if s>=Sdet_start+Nstb-1

[0164] PowDiff(s)=min(Pow(s-Nstb+1:s))-Pmax;

[0165] PowDet(s)=PowDiff(s)>=-deltaP;

[0166] if PowDet(s) == 1

[0167] PowCount(s)=PowCount(s-1)+1;

[0168] elseif PowDet(s)==0

[0169] PowCount(s) = 0;

[0170] end if PowCount(s)>=PowCount_ref s_ref_temp=[s_ref_temp s];

[0171] end

[0172] else

[0173] PowDiff(s)=0.0;

[0174] PowDet(s)=0;

[0175] PowCount(s) = 0;

[0176] end

[0177] end s_ref=max(s_ref_temp);

[0178] if isempty(s_ref)

[0179] Flag_Detected = 0;

[0180] else

[0181] Flag_Detected = 1;

[0182] end

[0183] In the code above:

[0184] The value of PowDiff is indicated by the mark "o" in FIG. 21(b).

[0185] The value of PowDet is shown with the mark "o" in FIG. 21(c).

[0186] The value of PowCount(s) is shown with a mark "x" in FIG21(c).

[0187] We can find verification of the detection process through “Flag_Detected”.

[0188] If the process is invalid, ie, "Flag_Detected = 0", the current update process may jump to the next signal capture time.

[0189] At operation [B1], the fluctuation interval width of the symbol power within the number of level-stable symbols Nstb is calculated (represented by "PowDiff(s)"). The symbol at "s" where "PwdDiff" is within the predetermined range ⊿P is considered valid, and PowDet is set to 1. "PowDet(s)" is counted cumulatively. If the "PwdDiff" of the symbol is not within the range ⊿P, the symbol "s" is considered invalid, and PowDet is set to zero. The accumulated count value is called the symbol power continuous value (represented by "PowCont(s)").

[0190] Then, symbol positions having symbol power continuous values ​​equal to or greater than a predetermined value "PowCont_ref" are regarded as reference symbol candidates, and the last reference symbol among the plurality of reference symbol candidates in the detection interval is set as the reference symbol.

[0191] Interval power fluctuation amplitude: PowDiff(s)=max{Pow(n)}–min{Pow(n)}n=s-Nstb,…,s

[0192] Sign detection value: PowDet=0, if PowDet>⊿P,

[0193] PowDet=1,if PowDiff(s)≦⊿P

[0194] Symbol power continuous value: PowCont(s) = PowCont(s-1) + PowDet(s), if PowDet(s) = 1

[0195] PowCont(s)==0, if PowDiff(s)=0

[0196] Reference symbol index: s_ref = max{s_ref_temp}

[0197] s_ref_temp∈argment(PowCont(s)>=PowCont_ref)

[0198] Figures 21a to 21cData generated according to an example embodiment is shown, generally represented by the reference numeral A210. Data A210 shows reference gain detection in a numerical example. The specific values ​​shown in data A210 are for illustrative purposes only. ΔP is set to 1 (deltaP=1.0). Within the detection interval, Sdet_start is set to 2 (Sdet_start=1) and Sdet_end is set to 15 (Sdet_end=16). Nstb is set to 4 (Nstb=4) and PowCont_ref is set to 6 (PowCont_ref=6). In this example, the position of the reference symbol "s_ref" is calculated to be fourteen.

[0199] Fig. 22 21 , the fourteenth symbol (s_ref=14) can be considered as a reference symbol. The symbol-based relative GTR (represented by "rG_ave(s)") is calculated using the gain relative to the detected reference gain.

[0200] E110:

[0201] Symbol reference relative GTR: rG_ave(s) = aG_ave(s) / aG_ave(s_ref)

[0202] s=1,...,s_ref(110)

[0203] Therefore, the calculated relative GTR 117 (rG_ave) is sent as an output from the GTR calculation unit 100 and is received at the compensation coefficient update and retention unit 201 included in the GTR compensation device 1. This is the next stage of the complex gain series in the "(P)th" update process. A new compensation coefficient series ((p)th coefficient) can be calculated using the calculated relative gain and the retained compensation coefficient series ((p-1)th coefficient).

[0204] Fig.23 is a block diagram of a system, generally indicated by reference numeral A230 , according to an example embodiment. The system A230 illustrates the compensation coefficient update and retention unit 201 included in the GTR calculation unit 100 .

[0205] The compensation coefficient updating and retaining unit 201 includes an inverse gain calculation unit 1021, a residual gain error unit 1022, a coefficient updating unit 1023, a coefficient library unit 1025, and a coefficient retaining unit 1026. The inverse gain calculation unit 1021 calculates the inverse characteristic of the relative GTR; the compensation coefficient library unit 1025 retains the compensation coefficient related to the input level and temperature; the residual gain error unit 1022 calculates the residual gain error "dG". The coefficient updating unit 1023 synthesizes the retained compensation coefficient updated last time with the output of the product residual gain error "dG" multiplied by the step size "α".

[0206] In the inverse gain calculation unit, the inverse characteristic "iG_ave" of the relative GTR "rG_ave" is calculated according to the following equation E111.

[0207] E111:

[0208] iG(s)=1 / rG_ave(s)s=1,...,s_ref(111)

[0209] Since "iG(s)" is the opposite characteristic of GTR, in the update process, we call "iG(s)" the residual gain coefficient, because "iG(s)" represents the compensation by using the compensation coefficient. Figure 8 The relative GTR shown shows the output where the IQ baseband signal is compensated by the GTC value. Therefore, the relative GTR is close to (but not exactly) [1,0], where the error in the relative gain is still in the GTR. We can call this the residual relative gain (or "residual gain") during the update process. We can distinguish whether the GTC value used is for a compensated gain state by using the terms "residual gain" and "relative / absolute gain".

[0210] In the update process of the compensation coefficient, the reference symbol position of each burst is different according to the traffic volume. Therefore, since the coefficient length may differ significantly each time the compensation coefficient is updated, it needs to be updated.

[0211] More specifically, the compensation coefficients are updated using the hold compensation coefficient "Coef(s, p-1)" and the gain error. The hold compensation coefficient is calculated in the previous update of the (P-1)th coefficient. The gain error (here equal to the inverse characteristic rG_inv(s)) contains the remaining gain compensation coefficients. The series of compensation coefficients should be updated regularly within a predetermined period of time.

[0212] In case the shorter of the length of the previous "P-1" compensation coefficients or the length of the remaining gain coefficient length "s_ref" in the current "P" updates is the update period length, then the coefficients not in the common update period are combined therein. Therefore, the compensation coefficient series should be updated within the "common update period length", which is determined by the shorter of the length of the previous "P-1" compensation coefficients or the length of the remaining gain coefficient length "s_ref" in the current "P" updates, then the coefficients exceeding the common update period are combined thereto.

[0213] The update of compensation factors based on differences in compensation factor series is further explained in detail in the following example.

[0214] The reference symbol indices for the previous update (P-1) and the current update (P) are denoted by "s_ref_pre" and "s_ref", respectively. The last symbol index of the common update cycle is "s_com". The angle of the complex value "x" is denoted by "angle(x)".

[0215] In a first example, the length between previous compensation coefficients is the same as the length between remaining gain coefficients. Fig.24 is a block diagram of a system according to an example embodiment, generally indicated by reference numeral A240. System A240 shows the relationship between the coefficient length before updating the compensation coefficient, the coefficient length after updating the compensation coefficient, and the symbol index between reference symbols. In this example, the common update period is the same length before and after the update, as shown in equation E112 below. Therefore, the compensation coefficients are updated as follows, regardless of the common update period.

[0216] E112:

[0217] s_com=s_ref=s_ref_pre(112)

[0218] First, the residual gain error (expressed as "dG(s)") is calculated. In order to compensate for the amplitude component and phase component of the residual gain, the error of the residual gain is extracted as an error vector, as shown in the following equation E113, using the vector method expressed by equation (13). However, the error in this error vector does not result in proportional compensation of the "update coefficient α" (i.e., the "step size") in both the amplitude component and the phase component.

[0219] E113:

[0220] Error vector: dG(s) = iG(s) – 1(113)

[0221] Therefore, the amplitude component and the phase component are calculated separately, as shown in the following equation E114. Then, each of the amplitude component and the phase component is multiplied by the update coefficient.

[0222] E114:

[0223] The amplitude factor of the residual gain error is: dG(s)_amp=|iG(s)|-1

[0224] Phase factor of the residual gain error: dG(s)_pha = angle(iG(s))

[0225] s=1,…,s_ref(114)

[0226] can be the amplitude component (given by α amp ) and the phase component (denoted by α amp Each of the 100 s (denoted by "uG_amp(s)") provides an update coefficient and is multiplied by each of the residual gain errors shown in equation E115 below. A value of 1 is added to the amplitude component to update the gain (denoted by "uG_amp(s)").

[0227] E115:

[0228] Update the gain amplitude factor: uG_amp(s)=1+dG_amp(s)*

[0229] Update the phase coefficient of the gain: uG_pha(s)=dG_pha(s)*

[0230] s=1,…,s_ref(115)

[0231] Therefore, an update gain "uG(s, p)" indicating the gain difference to be updated is provided so that the compensation coefficients in the amplitude component and the phase component are updated simultaneously by performing complex multiplication. amp , α pha The updated gain uG(s,p) is shown in equation E116 below.

[0232] E116:

[0233] uG(s, p)=uG_amp(s)*exp(j*uG_pha(s))s=1,...,s_ref(116)

[0234] As mentioned above, the update coefficient α can be amp , α phaThe compensation amount is controlled by multiplying both the amplitude component and the phase component respectively. The introduction of the update coefficient can provide stable compensation coefficient control under the slight fluctuation of GTR of each burst caused by the change of traffic volume. In addition, the gain amount of both the amplitude component and the phase component of the compensation coefficient series can be updated by multiplying the update gain "uG(s)" by the previous compensation coefficient series "Coef(s, p-1)". Therefore, the compensation coefficient series in the (P)th update is obtained using the equation E117 shown below.

[0235] E117:

[0236] Coef(s,p)=Coef(s,p-1)*uG(s)s=1,...,s_ref(117)

[0237] In a second example, the length of the remaining gain coefficients in the current update is "shorter" than the length of the remaining gain coefficients of the previous compensation coefficients. Fig.25 1 is a block diagram of a system according to an example embodiment, generally indicated by reference numeral A250. System A250 shows the relationship between the coefficient length before updating the compensation coefficient, the coefficient length after updating the compensation coefficient, and the reference symbol. In this example, the following equation E118 shows the relationship between the reference symbol of the previous compensation coefficient, the remaining gain coefficient, and the symbol index between the common update period. Therefore, s_ref is lower than s_ref_pre, and s_com is equal to s_ref.

[0238] E118:

[0239] s_ref <s_ref_pre,s_com=s_ref(118)

[0240] Because s_com is equal to s_ref, the residual gain error (represented by dG) for the common update period is calculated using the following equation E119. The residual gain error is referenced to the complex value [1, 0], i.e., the reference symbol value of s_ref, and its compensation coefficient is calculated by equation 120 with reference to the same complex value [1, 0]. The previous compensation coefficient of the external index of the common update period is then cascaded with the update coefficient shown in equation 121.

[0241] E119:

[0242] The amplitude factor of the residual gain error is: dG(s)_amp=|iG(s)|-1

[0243] Phase factor of the residual gain error: dG(s)_pha = angle(iG(s))

[0244] s=1,…,s_ref(119)

[0245] E120:

[0246] Update the gain amplitude factor: uG_amp(s)=1+dG_amp(s)*

[0247] Update the phase coefficient of the gain: uG_pha(s)=dG_pha(s)*

[0248] s=1,...,s_ref(120)

[0249] E121:

[0250]

[0251] In the third example, the length between previous compensation coefficients is shorter than the length between remaining gain coefficients in the current update. Fig.26 is a block diagram of a system according to an example embodiment. Generally indicated by reference numeral A260. System A260 shows the relationship between the coefficient length before updating the compensation coefficient, the coefficient length after updating the compensation coefficient, and the reference symbol. In this example, the following equation E122 shows the relationship between the reference symbol of the previous compensation coefficient, the remaining gain coefficient, and the symbol index between the common update period. Therefore, s_ref is higher than s_ref_pre, and s_com is equal to s_ref.

[0252] E122:

[0253] s_ref>s_ref_pre, s_com=s_ref_pre(122)

[0254] Since s_com is equal to s_ref, the residual gain error (denoted by dG) for the common update period is calculated using the following equation E123. The residual gain error is referenced by the residual gain coefficient value "iG(s_ref_pre)", i.e., the reference symbol position "s_ref_pre" of the previous coefficient. This is different from the second example, in which the residual gain error is referenced by the complex value [1,0] at the reference symbol position of the residual gain.

[0255] E123:

[0256] Amplitude factor of the residual gain error: dG(s)_amp=|iG(s)|-|iG(s_ref_pre)|

[0257] Phase factor of the residual gain error: dG(s)_pha = angle(iG(s)) – angle(iG(s_ref_pre))

[0258] s=1,…,s_ref_pre(123)

[0259] In addition, the compensation coefficient is calculated using equation E124 shown below: The compensation coefficient is referenced by the same residual gain coefficient value "iG(s_ref_pre)" located in the reference symbol position "s_ref_pre" of the previous coefficient.

[0260] E124:

[0261] Update the gain amplitude factor: uG_amp(s)=abs(iG(s_ref_pre))*(1+dG_amp(s)*α)

[0262] Update the phase factor of the gain: uG_pha(s)=angle(iG(s_ref_pre))+dG_pha(s)*α)

[0263] s∈1,…,s_ref_pre(124)

[0264] As shown in equation E125 below, the remaining gain coefficients of the external index used for the common update period in the (p)th update are concatenated into an update coefficient period.

[0265] E125:

[0266]

[0267] In the above examples, i.e., the first example, the second example, and the third example, the equations can be compared. Equations E115 and E119 can be expressed as the following equation E122, because in the first and second examples the residual gain at "s_ref" is a complex value with an amplitude component equal to 1 and a phase component equal to zero, i.e., a vector complex value [1, 0].

[0268] E126:

[0269] Amplitude factor of the residual gain error: dG(s)_amp = |iG(s)| - |iG(s_ref)| Phase factor of the residual gain error: dG(s)_pha = angle(iG(s)) - angle(iG(s_ref))

[0270] s=1,…,s_ref(126)

[0271] By analyzing the common update cycle, it is determined that the coefficient update and cascading in the update process can be performed and summarized in the following steps.

[0272] The first step is to analyze the common update cycle.

[0273] For the first example,

[0274] For the second example,

[0275] For the third example,

[0276] In the second step, the residual gain error is calculated according to Equation E127 as shown below.

[0277] E127:

[0278] Amplitude factor of the residual gain error: dG(s)_amp=|iG(s)|-|iG(s_com)|

[0279] Phase factor of the residual gain error: dG(s)_pha = angle(iG(s))-angle(iG(s_com))

[0280] s=1,...,s_com(127)

[0281] In the third step, the update gain is calculated according to Equation E128 as follows:

[0282] E128:

[0283] Update the gain amplitude factor: uG_amp(s)=abs(iG(s_ref_pre))*(1+dG_amp(s)*α)

[0284] Update the phase coefficient of the gain: uG_pha(s)=angle(iG(s_ref_pre))+dG_pha(s)*α)

[0285] Complex update gain: uG(s) = uG_amp(s)*exp(j*uG_pha)

[0286] s=1,…,s_com(128)

[0287] In the fourth step, the coefficients are updated and the concatenation is performed.

[0288] For the first example, coefficient updating is performed by equation E129, and the output is used as the compensation coefficient.

[0289] E129:

[0290] Coef(s,p)=Coef(s,p-1)*uG(s)s=1,...,s_ref(129)

[0291] For the second example, coefficient updating is performed by equation E130, and the output is used as the compensation coefficient.

[0292] E130:

[0293]

[0294] For the third example, coefficient updating is performed by equation E131, and the output is used as the compensation coefficient.

[0295] E131:

[0296]

[0297] As described above, one of the advantages of the example embodiments is that even if the coefficient length is different between the previous compensation coefficient and the current residual gain coefficient, the update process can update the compensation coefficient without inconsistency by updating the compensation coefficient based on the coefficient length.

[0298] More specifically, the amplitude component and the phase component of the residual gain error are calculated by reference to the residual gain coefficient at the reference symbol position, respectively. Then, the compensation coefficient of the common period is determined by multiplying the composite update gain. The composite update gain is determined by multiplying the update coefficient (step size) by the amplitude component and the phase component of the residual gain error. The portion of coefficients outside the common update period that is longer than the portion of the previous compensation coefficient or the current residual gain coefficient can be combined into the update coefficients in the common period.

[0299] Therefore, by changing the position of the reference gain in each burst due to traffic variations, a continuous update can be obtained.

[0300] exist Fig.12 In the compensation unit 202, the GTR is compensated before transmission. The GTR is compensated by multiplying the compensation coefficient with the baseband forward signal 210. The compensation coefficient of the GTR is calculated using the compensation coefficient series Coef(s, p) for each symbol.

[0301] As previously described, the compensation coefficient is the inverse characteristic of the relative gain calculated from the start symbol to the detected reference symbol. Therefore, continuous transmission of the waveform is provided during the compensation period. The compensation period may include the forward portion of the reference symbol and the backward symbol after the compensation period. Therefore, a limited number of coefficients are provided for compensation without increasing the number according to the length of the transient response. With respect to the retained compensation coefficients, the last coefficients always converge to [1, 0] so that an update can be performed, thereby eliminating the problem of coefficient length differences.

[0302] It is not necessary to perform the above-mentioned process of continuously updating the compensation coefficient in each burst signal. The process can be performed periodically or intermediately according to the capability of the transmission system, which depends on the number of transmission antennas, the processing performance of the equipment equipped in the device, and more specifically, the number of signal branches of the RFIC, the processing capability of digital signal processing devices such as DPS, FPGA and SoC.

[0303] For example, in the above exemplary embodiment, the baseband forward signal 210 is pre-compensated by using the compensation coefficients. In the absence of DPD, the signal to be pre-compensated may be output to the CFR 312.

[0304] The gain transient response compensation circuit described in this article is particularly suitable for signals of high-order multi-level modulation schemes such as 256QAM and 1024QAM because of the accurate calculation and compensation of GTR as shown in Figures 9 and 10 above, and the adaptability of updating the compensation coefficient as described above.

[0305] For completeness, Fig. 27 2 is a schematic diagram of components for implementing one or more modules of the above algorithm, which are collectively referred to as processing system 270. Processing system 270 may have a processor 276, a memory 271 coupled to the processor and including RAM 272 and ROM 273, and optionally, a user input 277 and a display 278. Processing system 270 may include one or more network interfaces 279 for connecting to a network, such as a wired or wireless modem.

[0306] Processor 276 is connected to each of the other components to control their operation.

[0307] The memory 271 may include non-volatile memory, a hard disk drive (HDD) or a solid state drive (SSD). The ROM 312 of the memory 273 stores, among other things, an operating system 274 and may store software applications 275. The processor 276 uses the RAM 272 of the memory 271 for temporary storage of data. The operating system 274 may contain code that implements aspects of the algorithm A15 when executed by the processor.

[0308] Processor 276 can take any suitable form. For example, it can be a microcontroller, multiple microcontrollers, a processor or multiple processors. Processor 276 can include processor circuits.

[0309] Processing system 270 may be a stand-alone computer, a server, a console, or a network thereof.

[0310] In some embodiments, processing system 270 may also be associated with external software applications. These may be applications stored on remote server devices and may run partially or exclusively on remote server devices. These applications may be referred to as cloud-hosted applications. Processing system 270 may communicate with remote server devices in order to utilize software applications stored therein.

[0311] Reference Fig.28a and Fig.28bTangible media, respectively a removable non-volatile storage unit 281 and a compact disk (CD) 284, are shown, which store computer readable code, respectively, which can perform the method according to the above-mentioned embodiment when run by the computer. The removable non-volatile storage unit 281 can be a memory stick, such as a USB memory stick, which has an internal memory 283 storing computer readable code. The memory 283 can be accessed by the computer system via a connector 282. The CD 284 can be a CD-ROM or a DVD or the like. Other forms of tangible storage media can be used.

[0312] Embodiments of the present invention may be implemented in software, hardware, application logic, or a combination of software, hardware, and application logic. The software, application logic, and / or hardware may reside in memory or on any computer medium. In an example embodiment, the application logic, software, or instruction set is maintained on any of a variety of conventional computer-readable media. In the context of this document, "memory" or "computer-readable medium" may be any non-transitory medium or device that can contain, store, communicate, propagate, or transmit instructions for instruction execution or use in conjunction with a system, device, or apparatus (e.g., a computer).

[0313] Where relevant, references to "computer-readable storage medium", "computer program product", "tangibly embodied computer program", etc. or "processor" or "processing circuitry", etc. shall be understood to include not only computers having different architectures, such as single / multi-processor architectures and sequencer / parallel architectures, but also special purpose circuits, such as field programmable gate arrays FPGAs, application specific circuits ASICs, signal processing devices and other devices. References to computer programs, instructions, code, etc. shall be understood to mean software for programmable processor firmware, such as the programmable content of a hardware device, as instructions for a processor or for a fixed function device, gate array, programmable configuration or configuration setting logic device, etc.

[0314] As used in this application, the term "circuitry" refers to all of the following: (a) hardware-only circuit implementations (e.g., implementations solely in analog and / or digital circuitry), and (b) combinations of circuitry and software (and / or firmware), such as, where applicable: (i) to a combination of a processor or (ii) to a portion of a processor / software (including a digital signal processor), software, and memory that work together to enable a device such as a server to perform various functions, and (c) to circuitry such as a microprocessor or portion of a microprocessor that requires software or firmware, even if the software or firmware is not actually present.

[0315] If desired, the different functions discussed herein may be performed in a different order and / or simultaneously with each other. Furthermore, if desired, one or more of the above functions may be optional or may be combined. Similarly, it will also be appreciated that Figure 3 A flow chart can be described. Figure 1b These are merely examples, and the various operations depicted therein may be omitted, reordered, and / or combined.

[0316] It should be understood that the above exemplary embodiments are illustrative only and do not limit the scope of the present invention.Other changes and modifications will be apparent to those skilled in the art after reading this specification.

[0317] Furthermore, the disclosure of the present application should be understood to include any novel feature or novel combination of features disclosed herein, whether explicitly or implicitly, or any generalization thereof, and new claims may be raised during the prosecution of the present application or any application derived therefrom. The term "covering" may cover any such feature and / or combination of these features.

[0318] Although various aspects of the invention are set out in the independent claims, further aspects of the invention comprise other combinations of features from the described embodiments and / or the dependent claims with features of the independent claims, and not just the combinations explicitly listed in the claims.

[0319] It should also be noted here that although various examples are described above, these descriptions should not be construed in a limiting sense. Rather, many variations and modifications may be made without departing from the scope of the present invention as defined in the appended claims.

Claims

1. A device for communication, comprising: means for determining an absolute gain of a power amplifier over time, wherein the absolute gain is formed by dividing a feedback baseband signal derived from an output of the power amplifier by a forward baseband signal used to form an input to the power amplifier; means for determining a relative gain transient response of the power amplifier by normalizing the absolute gain to generate a relative gain of the power amplifier over time; as well as Means for determining a gain transient response compensation coefficient, the gain transient response compensation coefficient having an inverse characteristic of the relative gain transient response.

2. The apparatus of claim 1 , wherein the means for determining the absolute gain of the power amplifier over time comprises: Means for averaging the feedback baseband signal divided by the forward baseband signal over successive sampling intervals.

3. The apparatus of claim 1 or claim 2, wherein the means for determining the relative gain transient response of the power amplifier comprises: Means for normalizing the absolute gain relative to a reference gain. 4 . The apparatus of claim 3 , wherein the reference gain is determined at a time when the absolute gain is determined to have stabilized.

5. The apparatus according to claim 4, further comprising: Means for determining a time at which the absolute gain is determined to have stabilized, the means comprising a reference detection unit configured to determine a sampling period during which a power level of the forward baseband signal has maximum stability.

6. The device according to claim 1 or claim 2, further comprising: Means for modifying the forward baseband signal using the determined gain transient response compensation coefficient to generate a compensated forward baseband signal. 7 . The apparatus of claim 6 , wherein the forward baseband signal is multiplied by the determined gain transient response compensation coefficient to generate the compensated forward baseband signal.

8. The device according to claim 1 or claim 2, further comprising: A component for storing and updating the gain transient response compensation coefficient.

9. The apparatus according to claim 8, wherein the means for storing and updating the gain transient response compensation coefficient updates the gain transient response compensation coefficient depending on the length of a previous compensation coefficient or a current remaining gain coefficient.

10. The apparatus of claim 9, wherein the common update cycle index (s_com) is determined according to the following relationship between a reference index (s_ref_pre) of a previous compensation coefficient and a reference index (s_ref) of a current residual gain: in: the gain transient compensation coefficient, updated for the common update period, calculated by using the previous compensation coefficient in the common update period and the current residual gain, controlled by multiplying the update coefficient by an error of the current residual gain, wherein the error is calculated by a difference from the residual gain of the index (s_com); and The coefficient portion of the gain transient compensation coefficient that is outside the common update period is combined with the updated coefficient portion in the common update period, and the length of the coefficient portion of the gain transient compensation coefficient that is outside the common update period is greater than the length of the portion of the previous compensation coefficient or the current residual gain that is outside the common update period.

11. The apparatus according to claim 1 or claim 2, wherein the feedback baseband signal is a complex signal having an in-phase component and a quadrature component, and / or the forward baseband signal is a complex signal having an in-phase component and a quadrature component.

12. The apparatus of claim 1 or claim 2, wherein the power amplifier is an RF signal amplifier unit.

13. An apparatus as claimed in claim 1 or claim 2, wherein the input to the apparatus is a burst mode signal.

14. The apparatus of claim 1 or claim 2, wherein the component comprises: at least one processor; as well as At least one memory including computer program code, the at least one memory and the computer program code being configured to, together with the at least one processor, cause execution of the apparatus.

15. A method for communication, comprising: determining an absolute gain of a power amplifier over time, wherein the absolute gain is formed by dividing a feedback baseband signal derived from an output of the power amplifier by a forward baseband signal used to form an input to the power amplifier; determining a relative gain transient response of the power amplifier by normalizing the absolute gain to generate a relative gain of the power amplifier over time; as well as A gain transient response compensation coefficient is determined, the gain transient response compensation coefficient having an inverse characteristic of the relative gain transient response.

16. The method of claim 15, wherein determining the absolute gain of the power amplifier over time comprises: The feedback baseband signal divided by the forward baseband signal is averaged and / or normalized over consecutive sampling intervals.

17. The method according to claim 15 or claim 16, further comprising: The forward baseband signal is modified using the determined gain transient response compensation coefficient to generate a compensated forward baseband signal.

18. The method according to claim 15 or claim 16, further comprising: The gain transient response compensation coefficient is stored and updated.

19. A computer program product comprising instructions stored thereon for performing at least the method of any one of claims 15 to 18.

Images