GB2527816A - Improvements to digital pre-distortion processing - Google Patents
Improvements to digital pre-distortion processing Download PDFInfo
- Publication number
- GB2527816A GB2527816A GB1411871.5A GB201411871A GB2527816A GB 2527816 A GB2527816 A GB 2527816A GB 201411871 A GB201411871 A GB 201411871A GB 2527816 A GB2527816 A GB 2527816A
- Authority
- GB
- United Kingdom
- Prior art keywords
- distortion processor
- power
- data
- coefficients
- input
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000012545 processing Methods 0.000 title description 6
- 238000000034 method Methods 0.000 claims description 33
- 238000005259 measurement Methods 0.000 claims description 23
- 230000003321 amplification Effects 0.000 claims description 17
- 238000003199 nucleic acid amplification method Methods 0.000 claims description 17
- 230000001960 triggered effect Effects 0.000 abstract description 3
- 230000003044 adaptive effect Effects 0.000 abstract 1
- 230000006870 function Effects 0.000 description 13
- 230000005540 biological transmission Effects 0.000 description 7
- 238000010586 diagram Methods 0.000 description 6
- 230000004044 response Effects 0.000 description 6
- 238000012546 transfer Methods 0.000 description 6
- 230000008859 change Effects 0.000 description 5
- 230000001360 synchronised effect Effects 0.000 description 5
- 230000009286 beneficial effect Effects 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 238000012360 testing method Methods 0.000 description 4
- 238000013459 approach Methods 0.000 description 3
- 230000001419 dependent effect Effects 0.000 description 3
- 230000001413 cellular effect Effects 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 102000000429 Factor XII Human genes 0.000 description 1
- 108010080865 Factor XII Proteins 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 230000003111 delayed effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000009877 rendering Methods 0.000 description 1
- 230000002459 sustained effect Effects 0.000 description 1
- 230000036962 time dependent Effects 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03F—AMPLIFIERS
- H03F1/00—Details of amplifiers with only discharge tubes, only semiconductor devices or only unspecified devices as amplifying elements
- H03F1/32—Modifications of amplifiers to reduce non-linear distortion
- H03F1/3241—Modifications of amplifiers to reduce non-linear distortion using predistortion circuits
- H03F1/3247—Modifications of amplifiers to reduce non-linear distortion using predistortion circuits using feedback acting on predistortion circuits
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03F—AMPLIFIERS
- H03F1/00—Details of amplifiers with only discharge tubes, only semiconductor devices or only unspecified devices as amplifying elements
- H03F1/32—Modifications of amplifiers to reduce non-linear distortion
- H03F1/3241—Modifications of amplifiers to reduce non-linear distortion using predistortion circuits
- H03F1/3258—Modifications of amplifiers to reduce non-linear distortion using predistortion circuits based on polynomial terms
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03F—AMPLIFIERS
- H03F3/00—Amplifiers with only discharge tubes or only semiconductor devices as amplifying elements
- H03F3/20—Power amplifiers, e.g. Class B amplifiers, Class C amplifiers
- H03F3/24—Power amplifiers, e.g. Class B amplifiers, Class C amplifiers of transmitter output stages
- H03F3/245—Power amplifiers, e.g. Class B amplifiers, Class C amplifiers of transmitter output stages with semiconductor devices only
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03F—AMPLIFIERS
- H03F2200/00—Indexing scheme relating to amplifiers
- H03F2200/321—Use of a microprocessor in an amplifier circuit or its control circuit
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03F—AMPLIFIERS
- H03F2200/00—Indexing scheme relating to amplifiers
- H03F2200/435—A peak detection being used in a signal measuring circuit in a controlling circuit of an amplifier
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Nonlinear Science (AREA)
- Algebra (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Amplifiers (AREA)
Abstract
New predistortion coefficients calculated by the processor 13 are used to update the existing coefficients in the predistorter with a weight that depends on the average amplitude of the input signal during the signal capture interval. The weight may be one or zero, or a number between one and zero. Greater weight is applied when the input signal has a large amplitude, because a large amplitude signal will experience greater distortion and so yield more effective predistortion coefficients. A smaller weight is applied when the signal amplitude is low, so that at least some updating of the coefficients is performed even when there has been an absence of high amplitude signals. The capture interval may be triggered when the input signal exceeds 7 an adaptive threshold 6 based on the peaks 5 of the input signal.
Description
Improvements to Digital Pre-di stortion Processing
Technical Field
The present invention relates generally to digital pre-distortion for linearisation of a power amplifier for use as a transmitter in a wireless network, and more specifically, but not exclusively, to updating operating coefficients of a digital pre-distortion processor.
Background
Modern wireless networks place increasing demands on the design of power amplifier transmission chains, in particular for base stations of the wireless networks. Power amplifier transmission chains are required to have a highly linear transfer function, but are also required to be power efficient, to both reduce power consumption and to reduce the amount of heat that needs to be dissipated t5 by the power amplifier, and so reduce the associated cost of the measures needed to dissipate the heat. Efficient power amplifiers tend to be inherently non-linear, and so techniques have been developed to linearise a power amplifier by means of pre-distortion of input signals, to correct for the non-linear response of the power amplifier. Typically, the non-linear response of the power amplifier has a characteristic that is temperature dependent, and that is also dependent on the input and output powers, and various other factors such as power supply voltage.
The temperature of the power amplifier may be dependent on the signal power, and also on the environment. As a result, the non-linear characteristic is typically time dependent, as the various factors affecting the linearity of the amplifier change with time. The pre-distortion characteristic is typically applied digitally to a digital input signal, typically at baseband, and is typically implemented by a digital transversal filter approach, in which the digital filter has a number of operating coefficients, also known as weights, which can be set to appropriate values, for example in a polynomial model, to counteract the non-linear characteristic of the power amplifier, Typically, the operating coefficients may be set on the basis of a comparison ofthe input signal supplied to the pre-distortion processor and the output signal generated by the power amplifier. Granted patent US8049560 describes a pre-distortion processor and a method of updating pre-distorter operating coefficients.
It is at the peaks of the input waveform that the need for accurate correction of the power amplifier transmission characteristic is at its greatest, since the peaks of the waveform will suffer the most distortion in the power amplifier because they exercise a greater range of the power amplifier response.
Generally, operating coefficients that have been derived from a high power sial will also be acceptable for use with a low power signal, but the converse is generally not true; operating coefficients that have been derived for use with a low power signal are unlikely to be acceptable for use with a high power signal.
So, a capture threshold is generally employed, so that only input signals that exceed a certain level are used to train a pre-distortion processor to derive operating coefficients.
However, it may be problematic to set the capture threshold, since it is undesirable that update of the coefficients be inhibited for long periods of time, because the required operating coefficients are likely to change with time. During periods of light use of the wireless network, for example at night, a high threshold may not necessarily be crossed, so that update of coefficients may be inhibited for long periods, so that the coefficients do not track changes in the amplifier response caused, for example, by changes in temperature. However, if the threshold were to be set at a lower level, then the coefficients may be trained on lower input power signals, so that the updated coefficients would not necessarily be suitable for use with higher power input signals, should they be presentcd to the pre-distortion processor.
It is an object of the invention to address at least some of the limitations
of the prior art systems. n
Summary
In accordance with a first aspect of the present invention, there is provided a method of updating operating coefficients of a digital pre-distortion processor for a power amplifier of a transmitter of a wireless network, the method comprising: updating a capture threshold on a basis comprising a measure of a peak power of input signals to the pre-distortion processor during a first measurement period; generating a trigger signal in dependence on a comparison of magnitudes of samples of the input signals to the pre-distortion processor with the capture threshold; in dependence on the trigger signal, capturing a block of data, the captured block of data representing signals before amplification in the power amplifier and corresponding signals output from the power amplifier; calculating provisional updated coefficients for the digital pre-distortion processor using the captured block of data; determining a weighting factor for the provisional updated coefficients in dependence at least on a measure of an average power of the input to the pre-distortion processor in the captured block of data; and updating the operating coefficients of the digital pre-distortion processor using the provisional updated coefficients on the basis of the weighting factor.
This allows the capture threshold to fall during periods of reduced peak power, typically during periods of reduced traffic in the wireless network, so that some updates of operating coefficients may still take place to correct for changes in the power amplifier characteristic, typically due to temperature changes. The use of the weighting factor limits updating of the operating coefficients if the block of captured data is found to represent signals having insufficient power to provide a reliable update.
In an embodiment of the invention, the weighting factor has a value selected from 0 or 1, the value of 0 causing the operating coefficients of the digital pre-distortion processor not to be updated using the provisional updated coefficients, and the value of I causing the operating coefficients of the digital pre-distortion processor to be updated using the provisional updated coefficients.
This reduces processing load and provides a simpler implementation as compared to the use of weights that may be set to intermediate values, while still providing beneficial performance.
In an embodiment of the invention, the method comprises determining the weighting ftctor at least by comparison of a root mean square (RMS) power of the input to the pre-distortion processor in the captured data with a first threshold.
It has been found that the comparison of RMS power with a threshold provides a good indication of the acceptability of captured data to be used as the basis of operating coefficients.
In an embodiment of the invention the method comprises determining the first threshold from a root mean square power at the input to the pre-distortion processor during a second measurement period.
This allows a threshold to be set that is related to a RMS power determined dwing a reference period, which may be synchronised with a signal frame structure, to give a measure of expected RMS power that is expected to provide a reliable update.
In an embodiment of the invention, the second measurement period is the same as the first measurement period.
The first and second measurement periods need not necessarily be the same, but if they are processing may be simplified, since RMS and peak power may be measured for the same data.
In an embodiment of the invention, the first measurement period is greater than lOms.
This allows average power to be measured over at least an LTE frame.
In an embodiment of the invention, the captured block of data comprises between one thousand and four thousand time samples.
This provides a convenient sample size encompassing a peak of the waveform. A captured block of data comprising 2048 time samples has been found to be particulaily advantageous.
In an embodiment of the invention, the weighting factor is determined in further dependence on a measure of a peak power of the input to the pre-distortion processor in the captured data.
This allows further limiting of updates to take into account both average and peak power of the captured data, which may improve the reliability of updates.
In an embodiment of the invention, the weighting factor is determined at least by comparison of the measure of peak power of the input to the pre-distortion processor in the captured data with a second threshold.
This allows a limit to be set on acceptable peak power for use in an update, to provide a further test to limit updates in situations in which the RMS power may appear acceptable, but peak power is inadequate to exercise sufficient of the amplifier non-linear characteristic to provide reliable updates.
In an embodiment of the invention, the method comprises detentning the second threshold from a peak power at the input to the pre-distortion processor dwing a third measurement period.
This allows a threshold to be set that is related to a peak power determined during a reference period, which may be synchronised with a signal frame siructure, to give a measure of expected peak power that is expected to provide a reliable update.
In an embodiment of the invention, the wireless network comprises equipment having a first radio access technology and equipment having a second radio access technology, and the input signals to the pre-distortion processor comprise signals for the first radio access technology and signals for the second radio access technology.
The combination of signals of different radio access technology, e.g. 3G and LTE, to be passed through a power amplifier provides the potential for a wide variety of peak to mean signal power characteristics so that limiting of updates according to signal characteristics may be beneficial.
The signals in the captured block of data that are captured before amplification in the power amplifier may be signals input to the pre-distortion processor or signals output from the pre-distortion processor, depending on the architecture used to train the pre-distortion processor.
In accordance with a second aspect of the invention there is provided a radio head for a wireless network, the radio head comprising at least a digital pre-distortion processor and a power amplifier, the radio head being configured to perform the method of the claimed invention.
lii accordance with a third aspect of the invention, there is provided program code for configuring a radio head of a wireless network comprising at least a digital pre-distortion processor and a power amplifier to operate according to the method the claimed invention.
Further features and advantages of the invention will be apparent from the following description of preferred embodiments of the invention, which are given by way of example only.
Brief Description of the Drawinizs
Figure I is a schematic diagram illustrating a linearising arrangement for a power amplifier; Figure 2 is a schematic diagram illustrating operation of apparatus according to a first embodiment of the invention; Figure 3 is a schematic diagram illustrating operation of apparatus according to a second embodiment of the invention; Figure 4 is a graph illustrating updating of a capture threshold and application of weights to updates in a period of relatively high mean signal power in an embodiment of the invention; Figure 5 is a graph illustrating updating of a capture threshold and application of weights to updates in a period of relatively low mean signal power in an embodiment of the invention; and Figure 6 is a flow diagram illustrating an embodiment of the invention.
Detailed Description
By way of example, embodiments of the invention will now be described in the context of detection of pre-distortion processors for power amplifiers in cellular wireless networks such as USM, 30 (UMIS) and LIE networks comprising GERAN, UTRAN and/or E-UIR.AIN radio access networks, but it will be understood that embodiments of the invention may relate to other types of wireless network, for example IEEE 80216 WiMax systems.
Figure 1 shows a linearisation arrangement for a power amplifier 19. The power amplifier may, for example, be a Doherty type amplifier, which achieves high efficiency at the cost of a non-linear transfer characteristic. Input signals X, typically in digital form, are passed to pre-distortion processor 16, which processes the input signals to counteract the effects of non-linearity in the power amplifier 19. The pre-distortion processor typically applies a polynomial function to the input signals. The polynomial function typically has variable operating coefficients, which apply weighting values to terms of the polynomial ifinction, and the terms of the polynomial function may include delayed polynomial terms.
The operating coefficients are set by an estimator 24. The output Z of the pre-distortion processor 16 is passed, typically at baseband as inphase (I) and quadrature (Q) components to Digital to Analogue Converters (DAC) 17, and then to a radio frequency (RE) up-converter 18, which typically modulates a carrier with the I and Q components and may translate the signal frequencies to radio frequency by mixing with appropriate local oscillator signals. Typically, the pre-distortion processor processes signals at baseband. A Quadrature Modulation Correction (QMC) function may be included between the pre-distortion processor 16 and the DAC 17, to correct for imperfections in the process of modulating inphase (I) and quadrature (Q) components onto a carrier in the analogue domain.
Signals are output from the RF upconverter 18 at radio frequency and amplified by the power amplifier 19 for transmission by one or more antennas. The output of the power amplifier passes through a coupler 20 before transmission.
The coupler 20 couples a relatively small proportion of the signal, as compared to the proportion of the transmitted signal, to an observation receiver comprising an RF downconverter 21, typically resulting in inphase and quadrature outputs and analogue to digital converters (ADC) 22 which convert the inphase and quadrature outputs to the digital domain.
A capture block 22 captures samples of signals before amplification in the power amplifier 19, and also captures samples of signals after amplification in the power amplifier 19, at the output of the observation receiver. So, in the example of the architecture of Figure 1, signals Z are captured at the output of the pre-distortion processor before amplification in the power amplifier 19, and signals Yi are captured at the output of the ADC 22 of the observation receiver after amplification by the power amplifier 19. The signals Y1 and I are typically represented as inphase and quadrature components. The capture block 23, operating under control of a system control function 25, captures a block of N samples Z1N YiN of signals before and after amplification in the power amplifier, and the block of samples is passed to an estimator 24. The estimator typically compares the samples of signals before and after amplification by the power amplifier, normalised to a similar power level, to determine a function by which the signal has been distorted by the amplifier. The estimator then typically determines an inverse of the distortion function, for application to the input signal by the pre-distorter in order to counteract the effects of the distortion in the amplifier, rendering the path from the input of the pre-distortion processor to the output of the amplifier substantially linear and so substantially distortion free of distortion and spurious signal components. The inverse of the distortion function is typically applied by applying operating coefficients calculated in the estimator 24 to a polynomial frmnction in the pre-distortion processor.
Conventionally, in the system of Figure 1, the capturing of data by the capture block 23 is triggered by the comparison of the magnitude of input samples with a capture threshold. The capture threshold may be set at a level that would be expected to be triggered by peak powers with sufficient power to exercise the range of the power amplifier characteristic sufficiently to excite a significant non-linear response. However, it may be found that in periods when the traffic in the wireless network is low, the aggregate power transmitted by the power amplifier may not be high enough to cross the threshold, and in this case a change in temperature in the power amplifier may cause the transfer characteristic of the non-linear response to change, so that the previously calculated operating coefficients may no longer be correct to counteract the effects of the non-linearity in the power amplifier. While the signal power is low, this may not cause a problem, since the excursions of the signal waveform may not extend into the now badly-corrected non-linear region of the amplifier characteristic, However, if a burst of high power signals should suddenly appear, for example due to a sustained data download, it may be found that the overall non-linearity of the combination of the pre-distortion processor, also known as a pre-distorter, is not adequate and spurious signals may be generated as a result.
In an embodiment of the invention, the capture threshold is lowered during periods of low signal power, so that some data is captured to potentially allow operating coefficients to be updated before they become out of date. However, lowering the capture threshold increases the probability of capturing data that does not adequately exercise the non-linear characteristic of the power amplifier, so that updates to operating coefficients generated on the basis of this data may degrade the linearization performance. To guard against this, in an embodiment of the invention, the updates to the operating coefficients are weighted according to a figure of merit of the captured data, typically based on the root mean square (RMS) and/or peak power of the data that has been captured as the result of the trigger. The weighting of the updates may be all or nothing, i.e. the update may be used or discarded, with the weight having a or 0 value, or else the weighting may include intermediate values, for example set in proportion to how high a peak and/or RIvIS power is above a threshold Figure 2 is a block diagram illustrating a method of updating operating coefficients of a digital pre-distortion processor for a power amplifier of a transmitter of a wireless network in a first embodiment of the invention.
Input data, typically at baseband in inphase and quadrature form, is applied to an amplifier linearisation arrangement, such as that illustrated in Figure 1. A peak detector detects 5 a peak of the incoming data, and this peak level is used to set a capture threshold, by updating 6 the capture threshold on a basis comprising the measure of a peak power of input signals to the pre-distortion processor during a first measurement period. The first measurement period is typically greater than 10 ms, allowing average power to be measured over at least an LTE frame, and as shown in Figure 2 a first measurement period of 100 ms may be advantageous.
A trigger signal is generated in dependence on a comparison 7 of magnitudes of samples of the input signals to the pre-distortion processor with the capture threshold, and in dependence on the trigger signal a block of data is captured. "Input" signals are captured 8 before amplification in the power amplifier, and "output" signals are captured 14 after amplification in the power amplifier. So, the captured block of data represents signals before amplification in the power amplifier and corresponding signals output from the power amplifier.
The captured block of data may comprise between one thousand and four thousand time samples, which provides a convenient sample size to at least encompass a peak of the waveform, A captured block of data comprising 2048 time samples has been found to be particularly advantageous.
Provisional updated coefficients are calculated 13 for the digital pre-distortion processor using the captured block of data. The provisional updated coefficients may be new values of the operating coefficients calculated based on a comparison of captured data before and after amplification in the power amplifier, or the provisional updated coefficients may be incremental values to be applied to the currently applied operating coefficients in the pre-distortion processor. An example of an algorithm for generating updated operating coefficients is disclosed in granted patent US8049560.
In the embodiment of the invention shown in Figure 2, a weighting factor is determined 9 for the provisional updated coefficients in dependence at least on a measure of an average power of the input to the pre-distortion processor in the captured block of data. The operating coefficients of the digital pre-distortion processor are then updated using the provisional updated coefficients on the basis of the weighting factor 12, The weighting factor may be determined at least by ii comparison of a root mean square (RMS) power of the input to the pre-distortion processor in the captured data with a first threshold, which may provide a good indication of the acceptability of captured data to be used as the basis of operating coefficients. The root mean square power at the input to the pre-distortion processor may be determined during a second measurement period, which may be a reference period synchronised with a signal frame structure, to give a measure of expected ifiviS power that is expected to provide a reliable update. The second measurement period, used to measure the RMS power for determining the first threshold, used for determining the weighting factor, may be the same period as the first measurement period in which peak power is measured to determine the capture threshold, The first and second measurement periods need not necessarily be the same, but if they are processing may be simplified, since RMS and peak power may be measured for the same data.
The weighting factor may, in addition to depending on a measure of RNIS power the input to the pre-distortion processor in the captured data, also depend on a measure of a peak power of the input to the pre-distortion processor in the captured data, allowing further limiting of updates to take into account both average and peak power of the captured data, which may improve the reliability of updates. This may be done by comparison of the measure of peak power of the input to the pre-distortion processor in the captured data with a second threshold, allowing a limit to be set on acceptable peak power for use in an update, to provide a further test to limit updates in situations in which the RMS power may appear acceptable, but peak power is inadequate to exercise sufficient of the amplifier non-linear characteristic to provide reliable updates. The second threshold may be determined from a peak power at the input to the pre-distortion processor during a third measurement period, so that a threshold to be set that is related to a peak power determined during a reference period, which may be synchronised with a signal frame stmcture, to give a measure of expected peak power that is expected to provide a reliable update.
The updated pre-distortion coefficients are applied 10 to the input data, the input data representing input signals, and the pre-distorted signals are then amplified II in the power amplifier and output, typically to a diplexer and antenna.
Signals of different radio access technology, e.g. 3G and LIE, may be passed together, at different frequencies, through the pre-distortion processor and power amplifier, if the wireless network comprises equipment having more than one radio access technology. As a result, a wide variety of peak to mean signal power characteristics may be generated, based on the combination. So, it may be beneficial to measure peak and mean signal characteristics at the input as a characterisation of expected values, which may be compared with the measured characteristic of captured data to generate weighting values and/or go/no go decisions for an update based on the captured data.
Figure 3 shows an alternative embodiment to that of Figure 2. In the architecture of Figure 2, the "input" signals 8 in the captured block of data, that is to say the signals that are captured before amplification in the power amplifier, are signals at the output of the pre-distortion processor, that is to say after application 10 of the pre-distortion coefficients. In the alternative arrangement shown in Figure 3, the "input" signals 8 in the captured block of data are signals at the input to the pre-distortion processor, before application 10 of the pre-distortion coefficients. Either the approach of Figure 2 or Figure 3 can be used to update, i.e. train, the pre-distortion coefficients, In the case of Figure 3, the coefficients may be trained to reduce differences between the signal represented by the input data and the output signal from the power amplifier, suitably frequency translated and normalised for power. In the case of Figure 2, as already mentioned, the coefficients may be trained to apply an inverse function to the non-linear transfer frmnction of the power amplifier, so that application of the two transfer functions in succession results in an overall linear transfer function, In the case of Figure 3, similarly to Figure 2, the weighting factor may have a value selected from 0 or 1, the value of 0 causing the operating coefficients of the digital pre-distortion processor not to be updated using the provisional updated coefficients, and the value of 1 causing the operating coefficients of the digital pre-distortion processor to be updated using the provisional updated coefficients. This approach reduces processing load and provides a simpler implementation as compared to the use of weights that may be set to intermediate values, while still providing beneficial performance.
The linearisation and amplifier circuitry, including the digital pre-distortion processor, power amplifier and observation receiver may be implemented as part of a radio head for a wireless network. The term "radio head" is used in the field of cellular wireless to refer to electronic circuitry for transmission and/or reception of radio frequency signals that is installed at a cell site and connected to antennas at the cell site. Typically a radio head is mounted on a transmission tower, and connected, for example by a fibre, to further digital electronic parts in a cabinet on the ground for connection to a wider telecommunications network, A radio head typically comprises up converter and down converter circuitry for converting digital signals, typically at baseband, to and from radio frequencies (RF).
Program code may be provided to configure the radio head to operate according to the method of embodiments of the invention. For example, the process flow of Figure 2 or Figure 3 may be performed under control of a microcontroller or other digital processor or computer according to the program code. The program code may comprise executable software instructions and/or firmware for configuring a gate array circuit or circuits, For example, in an embodiment of the invention as shown in Figure 2, applying pre-distortion coefficients 10 to the input data may be performed by a Field Programmable Gate Array (FPGA) configured by program code, and the functions 5, 6, 7, 8, 9, 13, N for controlling the updates of the pre-distortion coefficients may be performed by a microcontroller performing instructions according to program code held in program memory, The pre-distorted data is converted to the analogue domain and amplified 11 by the power amplifier.
Figure 4 is a graph illustrating updating of a capture threshold and application of weights to updates in a period of relatively high mean signal power in an embodiment of the invention, Curve 1 shows an example of the input sial level as a ifinction of time, In this example, there are alternating periods of 00 ms at a relatively high signal level, followed by period of 500 ms at a lower signal level. This example may be used as a test signal for the system; signals measured on a network may be less regular. It can be seen that a capture threshold 2 varies with time, ramping up or down according to the power of the input signal, in this example applying a time constant to historical measurements ofpealc input power.
So, it may be seen that during a high power period ti the capture threshold may rise and during a lower power period t2 the capture threshold may fall. When the signal power level of the input data crosses the capture threshold, this generates a trigger signal that triggers the capture of a block of data sampled from signals coupled from both before and after the power amplifier. In the example of Figure 4, it can be seen that capture triggers occur at times 3a... 3h, causing capture of respective data blocks 4a.... 4h. In this example, a captured block of data comprises 2048 time samples, so that captured data typically represents a small proportion of the total input data. Captured data is intended to represent the peaks of the signal characteristic, that may be expected to exercise the extremes of the amplifier characteristic, i.e., the extremes of the output voltage range, in which non-linear behaviour is expected. In the example of Figure 4, the captured data corresponds well with peaks of the input signal that are taken when the RIvIS signal level is relatively high, So, updates to the operating coefficients generated using this data may be expected to be reliable, and high weights are applied to these updates. In one embodiment, the weights may have values selected between 1 and 0, so in this case a high weight would be a weight of 1, i.e. the update would be enabled.
Figure 5 shows the effect on capture threshold when there are longer periods of low signal power. It can be seen that the capture threshold falls during the longer period of low signal power. Again, similarly to Figure 4, Figure 5 is an example of a test signal for the system, and so it is somewhat more regular than a typical signal received in a network, but it illustrates the principle of operation.
In this example, there are periods of 100 ms at a relatively high signal level, but some periods t5 at lower signal level are longer than those in the case of Figure 4, in this example 3 seconds, It can be seen that a capture threshold 2 ramps down, with a time constant, during the period of lower signal level t3, on the basis of peak input power received in this period. It can be seen that capture triggers occur at times 3i. 3s, causing capture of respective data blocks 4i 4s. In this example, the capture triggers 3i and 3j result in captured blocks of data 4i and 4j, S which have a relatively high RMS power, so that updates of operating coefficients based on this data would be expected to be reliable, and high weights are applied.
By contrast, the capture triggers 3k results in captured block of data 4k, which in this example has a relatively low RIvIS power, so that updates of operating coefficients based on this data would be expected to be less reliable, and a low weight is applied for the update of operating coefficients. The low weight may be a 0, so that no update is applied. In some circumstances, it may be advantageous to update coefficients during the periods of low signal power, because if the coefficients are not updated for a long period they may become out of date, and so the capture threshold requirement may be relaxed by reducing the tS capture threshold. But not all blocks of data captured with the relaxed capture requirement may be suitable, so in an embodiment of the invention weights may be applied to the updates based on characteristics of the captured data, comparing the characteristics of the captured data with characteristics, e.g. RIvIS power, that would be expected for reliable data. It can be seen in Figure 5 that capture triggers 31 and 3m result in capturing data blocks 41 and 4m which in this example has a relatively high RMS power compared with that at 3k, so that updates of operating coefficients based on this data would be expected to be more reliable, and a higher weight is applied for the update of operating coefficients than for the case of 3k.
If the weights are restricted to 0 or 1 values, a I value may be used for the updates calculated on the basis of triggers 31 and 3m, i.e. the update is enabled. If variable weights are used, the update at 3k may be weighted with a lower weight than that at 3i or 3j, due to the lower power in the signal level I at 3k than at 3i or 3j, but at least some update is possible to the coefficients in the period between 3j and 31, which may not have been the case had a fixed threshold been used for the capture threshold, For example, a lower limit may be set to RMS power may of a captured data block, which is set to be below an expected power by a margin.
The weight value may be varied between 0 and I over the range equivalent to the margin, for example in proportion to the difference between the RMS and/or peak power of the input and the margin. Alternatively, a non-linear relationship between input power and weight value may be used, for example held in a look
up table.
In an embodiment of the invention, the capture threshold is derived from an estimate of the peak power of the signal being transmitted. This may be done with a peak detector which is read at too ms intervals, and which holds the peak magnitude of the signal that was transmitted during a 100 ms interval, The sampled value may then be filtered with a suitable time constant and used to set the capture threshold.
To determine the weighting values to be applied to an update, an RMS power estimate may be sampled, which may be synchronised with the frame structure of input signals, for example an average power may be measured over a t5 complete lOms LTE frame. The RIVIS power estimate may be sampled at lOOms intervals and filtered with an appropriate time constant filter, This may be used as a measure of expected signal metrics, together with a filtered version of the peak detector output. The expected signal metrics may be compared with corresponding metrics of the captured signals in the captured block of data to determine the weights, The filtering of the peak detector output used for the setting of the weights may be different of the peak detector output used to set the capture threshold, The time constant of the change in the capture threshold may be set to match an estimate of a time constant for temperature changes in the power amplifier. Typically the range of time constant may be between 10 ms and 10 seconds, but values outside this range may also be used, Figure 6 is a flow diagram illustrating a method an embodiment of the invention, according to steps S 6.1 to S 6.5.
The above embodiments are to be understood as illustrative examples of the invention, It is to be understood that any feature described in relation to any one embodiment may be used alone, or in combination with other features described, and may also be used in combination with one or more features of any other of the embodiments, or any combination of any other of the embodiments.
Furthermore, equivalents and modifications not described above may also be employed without departing from the scope of the invention, which is defined in the accompanying claims.
Claims (17)
- Claims 1. A method of updating operating coefficients of a digital pre-distortion processor for a power amplifier of a transmitter of a wireless network, the method comprising: updating a capture threshold on a basis comprising a measure of a peak power of input signals to the pre-distortion processor during a first measurement period; generating a trigger signal in dependence on a comparison of magnitudes of samples of the input signals to the pre-distortion processor with the capture threshold; in dependence on the trigger signal, capturing a block of data, the captured block of data representing signals before amplification in the power amplifier and corresponding signals output from the power amplifier; calculating provisional updated coefficients for the digital pre-di stortion processor using the captured block of data; determining a weighting factor for the provisional updated coefficients in dependence at least on a measure of an average power of the input to the pre-distortion processor in the captured block of data; and updating the operating coefficients of the digital pre-distortion processor using the provisional updated coefficients on the basis of the weighting factor.
- 2. A method according to any preceding claim, wherein the weighting factor has a value selected from 0 or 1, the value of 0 causing the operating coefficients of the digital pre-distortion processor not to be updated using the provisional updated coefficients, and the value of I causing the operating coefficients of the digital pre-distortion processor to be updated using the provisional updated coefficients.
- 3. A method according to claim 2, comprising determining the weighting factor at least by comparison of a root mean square power of the input to the pre-distortion processor in the captured data with a first threshold.
- 4. A method according to claim 3, comprising determining the first threshold from a root mean square power at the input to the pre-distortion processor during a second measurement period.
- 5. A method according to claim 4, wherein the second measurement period is the same as the first measurement period.
- 6. A method according to any preceding claim, wherein the first measurement period is greater than 10 ms.
- 7. A method according to any preceding claim, wherein the captured block of data comprises between one thousand and four thousand time samples.
- 8. A method according to claim 7, wherein the captured block of data comprises 2048 time samples.
- 9. A method according to any preceding claim, comprising determining the weighting factor in further dependence on a measure of a peak power of the input to the pre-distortion processor in the captured data.
- 10. A method according to claim 9, comprising determining the weighting factor at least by comparison of the measure of peak power of the input to the pre-distortion processor in the captured data with a second threshold.
- 11. A method according to claim 10, comprising determining the second threshold from a peak power at the input to the pre-distortion processor during a third measurement period.
- 12. A method according to any preceding claim, wherein the wireless network comprises equipment having a first radio access technology and equipment having a second radio access technology, and the input signals to the pre-distortion processor comprise signals for the first radio access technology and signals for the second radio access technology.
- 13. A method according to any preceding claim, wherein said signals before amplification in the power amplifier are signals input to the pre-distortion processor.
- 14. A method according to any one of claims 1 to 12, wherein said signals before amplification in the power amplifier are signals output from the pre-distortion processor.
- 15. A radio head for a wireless network, the radio head comprising: a digital pre-distortion processor; a power amplifier; means for updating a capture threshold on a basis comprising a measure of a peak power of input signals to the pre-distortion processor during a first measurement period; means for generating a trigger signal in dependence on a comparison of magnitudes of samples of the input signals to the pre-distortion processor with the capture threshold; means for capturing, in dependence on the trigger signal, a block of data, the captured block of data representing signals before amplification in the power amplifier and corresponding signals output from the power amplifier; means for calculating provisional updated coefficients for the digital pre-distortion processor using the captured block of data; means for determining a weighting factor for the provisional updated coefficients in dependence at least on a measure of an average power of the input to the pre-distortion processor in the captured block of data; and means for updating the operating coefficients of the digital pre-distortion processor using the provisional updated coefficients on the basis of the weighting factor.
- 16. A radio head according to claim 15, comprising means for performing a method according to any one of claims 2 to 14.
- 17. Program code for configuring a radio head of a wireless network comprising at least a digital pre-distortion processor and a power amplifier to operate according to the method of any one of claims I to 14.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| GB1411871.5A GB2527816B (en) | 2014-07-03 | 2014-07-03 | Improvements to digital pre-distortion processing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| GB1411871.5A GB2527816B (en) | 2014-07-03 | 2014-07-03 | Improvements to digital pre-distortion processing |
Publications (3)
| Publication Number | Publication Date |
|---|---|
| GB201411871D0 GB201411871D0 (en) | 2014-08-20 |
| GB2527816A true GB2527816A (en) | 2016-01-06 |
| GB2527816B GB2527816B (en) | 2018-10-24 |
Family
ID=51410579
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| GB1411871.5A Expired - Fee Related GB2527816B (en) | 2014-07-03 | 2014-07-03 | Improvements to digital pre-distortion processing |
Country Status (1)
| Country | Link |
|---|---|
| GB (1) | GB2527816B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112953869A (en) * | 2021-01-27 | 2021-06-11 | 武汉正维电子技术有限公司 | Adaptive data screening method and system for base station digital intermediate frequency predistortion system |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111131104B (en) * | 2019-12-12 | 2022-12-06 | 京信网络***股份有限公司 | Pre-distortion processing method, device, storage medium and equipment |
| CN113468814B (en) * | 2021-07-09 | 2024-02-27 | 成都德芯数字科技股份有限公司 | Neural network-based digital predistortion training data screening method and device |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1983659A2 (en) * | 2007-04-20 | 2008-10-22 | TelASIC Communications, Inc. | Method and apparatus for dynamic digital pre-distortion in radio transmitters |
| US7801245B2 (en) * | 2006-12-28 | 2010-09-21 | Alcatel-Lucent Usa Inc. | Increased reliability of data captures for signal predistortion |
| WO2011131122A1 (en) * | 2010-04-21 | 2011-10-27 | Zte Wistron Telecom Ab | Improved capture buffer of digital predistortion systems |
| US8049560B2 (en) * | 2008-11-11 | 2011-11-01 | Philip Brown | Increasing the density of larger magnitude amplifier output samples for estimating a model of digital pre-distortion to compensate for amplifier non-linearity |
| US8229025B1 (en) * | 2008-08-29 | 2012-07-24 | Xilinx, Inc. | Method of and circuit for accepting a sample of an input signal to be used to calculate parameters for a predistortion circuit in an integrated circut |
-
2014
- 2014-07-03 GB GB1411871.5A patent/GB2527816B/en not_active Expired - Fee Related
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7801245B2 (en) * | 2006-12-28 | 2010-09-21 | Alcatel-Lucent Usa Inc. | Increased reliability of data captures for signal predistortion |
| EP1983659A2 (en) * | 2007-04-20 | 2008-10-22 | TelASIC Communications, Inc. | Method and apparatus for dynamic digital pre-distortion in radio transmitters |
| US8229025B1 (en) * | 2008-08-29 | 2012-07-24 | Xilinx, Inc. | Method of and circuit for accepting a sample of an input signal to be used to calculate parameters for a predistortion circuit in an integrated circut |
| US8049560B2 (en) * | 2008-11-11 | 2011-11-01 | Philip Brown | Increasing the density of larger magnitude amplifier output samples for estimating a model of digital pre-distortion to compensate for amplifier non-linearity |
| WO2011131122A1 (en) * | 2010-04-21 | 2011-10-27 | Zte Wistron Telecom Ab | Improved capture buffer of digital predistortion systems |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112953869A (en) * | 2021-01-27 | 2021-06-11 | 武汉正维电子技术有限公司 | Adaptive data screening method and system for base station digital intermediate frequency predistortion system |
Also Published As
| Publication number | Publication date |
|---|---|
| GB201411871D0 (en) | 2014-08-20 |
| GB2527816B (en) | 2018-10-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9998241B2 (en) | Envelope tracking (ET) closed-loop on-the-fly calibration | |
| KR101715422B1 (en) | Pre-distortion for a radio frequency power amplifier | |
| US8737526B2 (en) | Predistortion of complex modulated waveform | |
| US8620233B2 (en) | Method of power amplifier predistortion adaptation using compression detection | |
| US6175270B1 (en) | Method and apparatus for tailored distortion of a signal prior to amplification to reduce clipping | |
| US20040142667A1 (en) | Method of correcting distortion in a power amplifier | |
| KR101711175B1 (en) | Under-sampling digital pre-distortion architecture | |
| US8705595B2 (en) | Digital output power measurement in radio communication systems | |
| US20100033246A1 (en) | Adaptive digital predistortion of complex modulated waveform using peak and rms voltage feedback from the output of a power amplifier | |
| JP2012509614A (en) | Efficient adaptive digital precompensation system for resources | |
| EP2795792B1 (en) | Adaptive predistortion for a non-linear subsystem based on a model as a concatenation of a non-linear model followed by a linear model | |
| US9680422B2 (en) | Power amplifier signal compensation | |
| JP6554265B2 (en) | Baseband digital predistortion architecture | |
| US20120264380A1 (en) | Polar Amplification Transmitter Distortion Reduction | |
| EP2237512A1 (en) | Method and apparatus, transmitter and transmitting method of determining open loop pre-distortion parameter | |
| CN113054917A (en) | Method and system for correcting distortion of non-linear power amplifier, and predistortion component | |
| US9363130B2 (en) | Dynamic digital predistortion for a microwave radio system | |
| CN103281274B (en) | A kind of digital pre-distortion measures system and performance number calibration measuring method thereof | |
| GB2527816A (en) | Improvements to digital pre-distortion processing | |
| US7120561B2 (en) | Phase-error based signal alignment | |
| CN203800956U (en) | Digital pre-distortion system of off-line estimation pre-distortion coefficient | |
| US8538349B2 (en) | Method and device for pre-distorting an exciter and predistortion exciter | |
| Tripathi et al. | A Low-cost test bench for the characterization of microwave devices using modulated envelope signal | |
| KR101415541B1 (en) | Apparatus for measuring rf distortion signal, radio power amplifying apparatus with the same and method for measuring rf distortion signal | |
| EP2056496A1 (en) | Wideband feedback path frequency response calibration method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PCNP | Patent ceased through non-payment of renewal fee |
Effective date: 20230703 |