HK1084513A

HK1084513A - Blind linearization using cross-modulation

Info

Publication number: HK1084513A
Application number: HK06102841.1A
Authority: HK
Inventors: G．J．巴兰泰勒
Original assignee: 高通股份有限公司
Priority date: 2002-06-28
Filing date: 2003-06-26
Publication date: 2006-07-28

Description

Blind linearization using cross modulation

Background

Technical Field

The present invention relates generally to a method of forcing a non-linear circuit (i.e., a circuit whose application of a given function varies with changes in the amplitude of an input signal) to apply its function to an amplitude modulated source signal without such non-linear behavior. This is done by combining the amplitude modulated source signal with one or more pseudo-signals (one for the dominant third order non-linear characteristic, two for the fifth order non-linear characteristic, etc.) to provide a combined signal that can be treated linearly, without modifying the characteristics of the circuit itself. The spurious signals, as well as other signals generated by the introduction of spurious signals, are then filtered out of the output of the circuit.

Background

Circuits are "linear" when they apply the same function to them regardless of the characteristics of the input signal. For example, if the circuit applies the same function to input signals regardless of whether they have small or large amplitudes, the circuit has no amplitude-dependent non-linear characteristics. Conversely, if the function of the circuit varies according to the amplitude of the input signal, it exhibits a nonlinear characteristic that depends on the amplitude. An example of a circuit with amplitude dependent non-linearity is an amplifier that multiplies an input signal with a small amplitude by 10, but as the amplitude of the input signal increases, it in turn multiplies them by smaller numbers, such as 9.8, 9.7, 9.6, 9.5, etc. The behavior of an amplifier is therefore dependent on the magnitude of its input signal.

Non-linear characteristics are an inherent property of many circuits and various circuit elements such as transistors, and are therefore even desirable in many different situations. However, non-linear circuit elements are generally undesirable when processing amplitude modulated communication signals. By definition, an amplitude modulated signal represents information in the form of a change in the amplitude of the signal envelope. Due to this amplitude-based variation, the non-linear circuit may process the amplitude-modulated input signal non-uniformly — not applying the same function uniformly. The effect of this is that the frequency bandwidth of the input signal is broadened. For example, an input signal that initially occupies a narrow frequency bandwidth will eventually occupy a wider frequency range. Thus, a circuit with amplitude dependent non-linear characteristics will generally increase the bandwidth of the amplitude modulated input signal.

This frequency spreading creates a number of problems. For example, the output signal of a communication device, which widens due to the non-linear effect described above, may overlap with a frequency channel used by another device of the same type. As a more specific example, the signal of one cordless telephone may overlap with the frequency channel used by another cordless telephone. This is known as "interference" and can significantly degrade the operation of other devices. Furthermore, if the master device is using a channel on the edge of the band allocated for such a device, the output signal of the device can even overlap with the frequency band of an unrelated device. Thus, a cordless telephone may interfere with a different device that is not even a cordless telephone.

Currently, engineers typically attempt to remove or compensate for non-linear characteristics within a signal transmitter by techniques such as limiting the range of input signals using non-linear circuits and filtering the output of the non-linear circuits to remove signals of undesired frequencies. Other techniques are also known, such as predistortion linearization, feedforward linearization, and modulation feedback.

However, these techniques are not sufficient in all cases. Problems still exist, for example, because predistortion requires an accurate nonlinear model, feedforward requires accurate and adaptive RF circuit matching, and modulation feedback tends to be unstable.

Disclosure of Invention

A method of linearizing a circuit having a nonlinear characteristic that depends on amplitude ("nonlinear circuit") enables the circuit to apply its function without its inherent nonlinear characteristic and without modifying the operating characteristics of the circuit. This is done by combining the amplitude modulated source signal with a dummy signal to provide a combined signal that can be linearly treated by the non-linear circuit. The spurious signals, as well as other signals resulting from the introduction of spurious signals, are then filtered out of the output of the circuit.

In accordance with a more specific aspect of the present invention, the following operations are performed. First, an amplitude modulated source signal is received, the signal having a source frequency bandwidth and a source envelope. A dummy envelope is calculated and a constant is derived if the source envelope and the dummy envelope can be combined in a predetermined manner. An amplitude modulated pseudo signal is generated which exhibits the calculated pseudo envelope and has a prescribed frequency bandwidth different from the source frequency bandwidth. The source signal and the dummy signal are added to form a combined signal that is directed to an input of the non-linear circuit. The output is filtered of spurious frequency bandwidth signals and other signals resulting from the introduction of spurious signals, thereby providing a linearized output attributable entirely to the source signal.

Drawings

FIG. 1A is a block diagram of the hardware components and interconnections of an exemplary linearizer.

FIG. 1B is a block diagram of the hardware components and interconnections of a linearizer having multiple pseudo signal generators.

Fig. 2 is an exemplary digital data processor.

Fig. 3 is an exemplary signal-bearing medium.

Fig. 4 is a flow chart illustrating an exemplary blind linearization operation sequence.

Fig. 5A-5B are signal diagrams illustrating a source signal having a source envelope and a dummy signal having a dummy envelope, respectively.

Detailed Description

The features, objects, and advantages of the present invention will become more apparent to those skilled in the art from the following detailed description when taken in conjunction with the accompanying drawings.

Hardware components and interconnects

Brief introduction to the drawings

One aspect of the present invention relates to a linearizing apparatus, which may be embodied by various hardware components and interconnections, one example of which is depicted by the informatization circuitry 101 of FIG. 1A. The linearizer is illustrated in an exemplary application environment 100 that includes a nonlinear circuit 114. In the environment 100, various inputs and outputs are depicted, such as 102, 113, 114a, 114b, 118, and so forth. Depending on the context, the hardware input/output lines ("input" and "output") and the input and output signals present on these input/output lines are denoted by these reference numerals. Moreover, although the term "circuitry" is used for ease of reference, the circuitry described herein may be implemented in discrete electronic devices, printed circuit board traces, integrated circuits, firmware, software, hardware, or any combination thereof. The composition of some exemplary sub-components is described in more detail below with reference to exemplary digital data processing apparatus, logic circuits, and signal-bearing media.

Typically, an input signal (e.g., 102) is directly input to a nonlinear circuit (e.g., 114) that merely processes the input signal and provides its output (at 114 b). Instead of this known method of providing the input signal 102 directly to the nonlinear circuit 114, an aspect of the present invention redirects the input signal 102 to a linearizer 101, the linearizer 101 generates a conditioned signal 113, and the conditioned signal 113 is input to the nonlinear circuit 114 in place of the input signal 102. Furthermore, instead of treating the output 114b of the non-linear circuit 114 as the final output, additional components of the linearizer 101 are used to further process the output 114b to provide the final linearized output 118. The linearized output 118 has no nonlinear effects of the circuit 114 that would be present at 114b if the input signal 102 were provided directly to the circuit 114.

Non-linear circuit

The nonlinear circuit 114 applies a given function to the signal at its input 114a and produces a resulting output at 114 b. However, since a given function varies with the amplitude of the signal arriving at 114a, circuit 114 is nonlinear. As a simple example, the circuit 114 may try to double the amplitude of its input signal 114 a. In this case, if the input signal is 2mV, the output of the circuit 114 is 4 mV. However, continuing with this example, the performance of the circuit 114 begins to degrade as the amplitude of the input signal increases. Instead of multiplying the input signal by 2, the circuit 114 begins multiplying the input signal by 1.95, however by 1.9 for larger amplitude input signals, then 1.85, 1.8, and so on. Thus, circuit 114 exhibits amplitude-dependent non-linear behavior because the function to which it is applied varies in accordance with changes in the amplitude of the input signal.

Preferably, the present invention can be implemented without knowledge of the degree, behavior, or other specific characteristics of the non-linear characteristics of the circuit 114. In this regard, one aspect of the present invention is "blind" linearization. Only the non-linearity class of the circuit 114 needs to be known, in particular the circuit shows a non-linear behavior that is amplitude dependent. Thus, the circuit 114 produces AM-AM and AM-PM distortion, meaning that Amplitude Modulation (AM) of the input signal causes non-linear amplitude modulation of the output signal and/or amplitude modulation of the input signal causes non-linear Phase Modulation (PM) of the output signal.

Without limitation, some examples of non-linear circuit 114 include amplifiers, filters, isolators, RF components, mixers, and the like.

Linearization circuit

Linearizer 101 is used in conjunction with nonlinear circuit 114 to force circuit 114 to provide a linear output. Rather than providing the source signal 102 directly to the nonlinear circuit 114, the nonlinear circuit 101 pre-processes the input 102 and provides a pre-processed input 113 to the nonlinear circuit 114; the linearizer 101 also post-processes the output 114b of the nonlinear circuit to ultimately provide a linearized output 118. Thus, the linearizer 101 comprises some pre-processing components 104, 112 residing between the nonlinear circuit 114 and the input 102, and some post-processing components 116 residing between the nonlinear circuit 114 and the final output 118.

As described above, there are various pre-processing components between the input 102 and the nonlinear circuit 114. This includes complement generator 104 and adder 112. The complement generator 104 includes an envelope detector 106, an envelope complement calculator 108, and a pseudo signal generator 110. The envelope detector 106 measures, quantizes, estimates, computes, or determines the envelope of the signal arriving at the input 102. This is called the source envelope. The envelope detector 106 may be implemented by various widely known envelope detectors such as one or more diodes, capacitors, resistors, etc. circuit configurations. Alternatively, in applications where the envelope information is known, the source envelope description arrives at the complement generator 104 from another source (not shown), for example in digital form.

The envelope complement calculator 108 calculates a "pseudo envelope" that is complementary to the source envelope. Broadly speaking, the dummy envelope is calculated such that if the dummy envelope were added to the source envelope, the result would be a constant. Thus, in a basic implementation, the value of the pseudo-envelope is calculated at any instant by subtracting the source envelope from a constant. The calculation of the pseudo-envelope is described in more detail below.

In one example, the complement calculator 108 can be constructed using discrete circuits such as transistors. Alternatively, the complement calculator 108 may be implemented in software, especially if the envelope detector 106 is default and the source envelope description arrives in digital form.

The pseudo signal generator 110 modulates the amplitude of the carrier signal to provide a pseudo signal characterized by a pseudo envelope. As a representative example, without limitation, the dummy signal generator 110 may include an oscillator and a multiplier, wherein the multiplier calculates the product of the carrier and the dummy envelope calculated by 108. This can be achieved with, for example, polar modulation. In a different example, the pseudo signal generator 110 may comprise a quadrature modulator including circuitry to calculate I and Q components based on the calculated pseudo envelope, and a multiplier to calculate the product of these I and Q components. Regardless of the modulation scheme, the spurious signals 105 occur at one or more frequencies (frequency "bandwidth") that are intentionally different from the source frequency bandwidth in order to assist in removing the spurious signals from the final output 118, as described in detail below.

Thus, the output of the complement generator 104 at 105 includes a pseudo signal, the envelope of which is described by a pseudo envelope. The frequency bandwidth of the signal is determined by the pseudo signal generator 110. Adder 112 combines dummy signal 105 with original source signal 102. The components 104, 112 then constitute preprocessing components that condition the source signal 102 before it reaches the nonlinear circuit 114, as described above. The output of the preprocessing component is a conditioned output 113. This signal is fed to a nonlinear circuit 114, which nonlinear circuit 114 processes its input 114a and provides an output at 114 b.

As described above, the linearizer 101 also includes various post-processing components between the nonlinear circuit 114 and the final output 118. That is, the filter 116 is used to remove any "intermodulation products," which refer to signals having spurious frequency bandwidths and signals resulting from the joint interaction of the input signal and spurious signals having non-linear characteristics. Thus, the final output 118 contains only signals attributable to the source signal 102. The filter 116 may include, for example, one or more baseband filters.

Multi-complement code generator

Optionally, different configurations of linearizer 101a are contemplated to implement the multiple pseudo signal generators 110a, 110B shown in FIG. 1B. To the extent that the components of circuit 101A differ from the components of circuit 101 (fig. 1A), they are given different reference numbers, as discussed below. The envelope detector 106 performs the same function in fig. 1A and 1B. That is, the envelope detector 106 measures, quantizes, estimates, calculates, or determines the source envelope.

Although the envelope complement generator 108a performs in a similar manner to the calculator 108 of fig. 1, the calculator 108 includes some additional functionality. That is, the envelope complement calculator 108a calculates two dummy envelopes (instead of one), where the dummy envelopes are complementarily combined with the source envelope. Exemplary methods known for generating a plurality of pseudo envelopes are discussed in more detail below.

In fig. 1B, there are a plurality of dummy signal generators 110a, 110B. Each of the pseudo signal generators 110a, 110b modulates a different carrier signal to provide a pseudo signal that exhibits a different calculated pseudo envelope (shown below as a)_de1And A_de2). Like the single dummy signal generator 110, each of the dummy signal generators 110a, 110B in the embodiment of FIG. 1B may utilize, for example, polarization or quadrature modulation.

The pseudo signal generators 110a, 110b contain pseudo signals on lines 105a, 105b, the outputs of which are directed to an adder 112. The adder 112 combines the pseudo signals 105, 106a with the source signal 102. Thus, the adder 112 provides a regulated output 113, the regulated output 113 having a constant or near constant envelope. This signal 113 is fed at 114a to a non-linear circuit 114. The circuit 114 processes the input 114a and provides an output at 114 b.

Like filter 116 of fig. 1A, filter 116a removes spurious signals (and signals resulting from the joint interaction of the input signal and spurious signals) from the output 114b of the non-linear circuit. However, since the linearizer 101a utilizes a plurality of pseudo signals 105a, 105b, the filter 116a is configured to remove the signals of each pseudo frequency bandwidth as well as any intermodulation products of these signals.

Exemplary digital data processing apparatus

As mentioned above, the data processing entity may be implemented in various forms, such as an envelope detector, an envelope complement calculator, a pseudo signal generator, an adder, a filter, or any one or more of their subcomponents. One example is a digital data processing apparatus, which is illustrated by the hardware components and interconnections of the digital data processing apparatus of FIG. 2.

The apparatus 200 includes a processor 202, such as a microprocessor, personal computer, workstation, controller, microcontroller, state machine, or other processing machine, coupled to a memory 204. In this example, the memory 204 includes fast access memory 206 and non-volatile memory 208. The fast-access memory 206 may include random-access memory ("RAM") that may be used to store programming instructions that are executed by the processor 202. The non-volatile memory 208 may comprise, for example, battery backed-up RAM, EEPROM, flash PROM, one or more magnetic data storage disks such as a "hard drive," tape drive, or any other suitable storage device. The device 200 also includes an input/output 210 such as a line, bus, cable, electromagnetic link, or other means for the processor 202 to exchange data with other hardware external to the device 200.

Although described in detail above, those skilled in the art, having the benefit of the present disclosure, will appreciate that the above-described apparatus may be implemented in machines of different configurations, without departing from the scope of the present disclosure. As a specific example, one of the components 206, 208 may be deleted; moreover, the memory 204, 206, and/or 208 can be integrated on the processor 202, or even provided external to the apparatus 200.

Logic circuit

In contrast to the digital data processing apparatus described above, a different embodiment of the present invention uses logic circuitry rather than computer-executed instructions to implement the various processing entities described above. Depending on the particular requirements of the application in the areas of speed, expense, tooling costs, etc., this logic may be implemented by an Application Specific Integrated Circuit (ASIC) constructed with thousands of tiny integrated transistors. This ASIC may be implemented in CMOS, TTL, VLSI, or another suitable configuration. Other alternatives include digital signal processing chips (DSPs), discrete circuits such as resistors, capacitors, diodes, conductors, and transistors, Field Programmable Gate Arrays (FPGAs), Programmable Logic Arrays (PLAs), Programmable Logic Devices (PLDs), and so forth.

Operation of

Having described the structural features of the present invention, various operational aspects of the present invention will now be described. As noted above, operational aspects of the present invention generally include methods that force a nonlinear circuit to apply its function in a linear fashion to an amplitude modulated source signal. This is done by combining the amplitude modulated source signal with a dummy signal to provide a combined signal that will be treated linearly, if the characteristics of the circuit itself are not modified. The spurious signals, as well as any intermodulation products, are then filtered out of the output of the circuit. These operations are discussed further below.

Signal bearing medium

As long as any of the functions of the present invention are implemented in one or more machine-executable program sequences, such sequences may be embodied in various forms of signal-bearing media. In the context of FIG. 2, such signal-bearing media may include, for example, the memory 204 or another signal-bearing media, such as a magnetic data storage diskette 300 (FIG. 3), directly or indirectly accessible by a processor 202. Whether contained in the memory 206, diskette 300, or elsewhere, the instructions may be stored on a variety of machine-readable data storage media. Some examples include direct access storage (e.g., a conventional "hard drive," a redundant array of inexpensive disks ("RAID"), or other direct access storage device ("DASD")), serial access storage such as magnetic or optical tape, electronic non-volatile storage (e.g., ROM, RPROM, flash PROM, or EEPROM), battery backed-up RAM, optical storage (e.g., CD-ROM, WORM, DVD, digital optical tape), paper "punch" cards, or other suitable signal-bearing media including analog or digital transmission media, as well as analog and communication links and wireless communications. In an illustrative embodiment of the invention, the machine-readable instructions may comprise software object code compiled from a language such as assembly language, C, or the like.

Logic circuit

In contrast to the signal-bearing media discussed above, some or all of the functionality of the present invention may be implemented in logic circuitry, rather than in instructions executed by a processor. Such logic circuitry is thus configured to perform operations to implement the method aspects of the present invention. As mentioned above, the logic circuit may be implemented with many different types of circuits.

Total sequence of operations

Fig. 4 shows a sequence 400 illustrating the operational aspects of the present invention. For ease of illustration, but without limitation, the example of FIG. 4 is described above in the context of the environment 100 of FIG. 1A.

In step 402, linearizer 101 receives a source signal on input 102. The source signal also goes to an adder 112. The source signal does not go to the non-linear circuit 114 because the linearizer 101 is designed to perform specific pre-processing tasks to help the non-linear circuit 114 process the source signal in a linear fashion.

Fig. 5A depicts a representative source signal 502. Source signal 502 comprises an amplitude modulated signal at a single frequency, although the techniques of the present invention may be used with a variety of frequency/phase modulated source signals. The source signal is said to have a source frequency bandwidth that encompasses a single frequency or multiple frequencies.

In step 403, the envelope detector 106 calculates a source envelope representative of the source signal 102. Fig. 5A depicts the envelope of the source signal 502 at 504. The envelope detector 116 operates by measuring, quantizing, estimating, calculating or otherwise determining the envelope of the signal arriving at the input 102. The output of the detector 106 is referred to as the "source envelope" and is used to describe the envelope 504 in terms of an analog waveform, digital information, or any other data, depending on the manner in which the detector 106 and/or the complement calculator 108 are implemented.

Step 403 is optional, however, as the envelope detector 106 may be omitted if the envelope information is known. For example, depending on the application, data and/or signals describing the envelope may already be available from a computer, analog circuitry, or other source separate from the linearizer 101. In this case, the input 102 need not be coupled to the complement generator 104, since the source envelope description arrives at the complement calculator 108 directly from a point at which it is separate from the source.

In step 404, the envelope detector 106 calculates a dummy envelope from the source envelope 504 calculated by the envelope detector 106 (step 403) or received from another source point. Broadly speaking, the pseudo-envelope is computed to derive a predetermined constant if the source envelope 504 and the pseudo-envelope are combined in a particular manner at any time. Fig. 5B illustrates an exemplary pseudo-envelope 508, the pseudo-envelope 508 being calculated based on the source envelope 504.

As a more specific example, the pseudo signal envelope is calculated such that when the source envelope and the pseudo envelope are processed in a predetermined formula (described below) and the processed signals are added, a predetermined constant is always derived. One embodiment of this predetermined formula is represented by the following formulas 1 and 2.

Formula 1K ═ a_se ²+2^*A_de ²

Wherein: k is a constant.

A_seThe magnitude of the source envelope.

A_deThe amplitude of the pseudo envelope.

In other words, the envelope complement calculator 108 in this embodiment calculates the magnitude of the pseudo envelope to satisfy the following equation 2.

Equation 2A_de＝sqrt[0.5^*(K-A_se ²)]

In step 406, the pseudo signal generator 110 modulates a carrier signal to provide a pseudo signal representing the calculated pseudo envelope. In the presently illustrated example, FIG. 5B shows the pseudo signal as 506. The generator 110 may generate the carrier signal, for example with an oscillator, or it may receive the carrier signal from elsewhere. The frequency bandwidth of the carrier signal (representing its single frequency, or frequency range if frequency/phase modulation is used) is different from the frequency bandwidth of the source signal 102. The frequency bandwidth of the carrier signal is referred to as a "pseudo-frequency". In addition, the operation of the downstream filter (discussed below) may be simplified by selecting the frequency bandwidth of the carrier signal in advance so that post-nonlinear circuit products are readily discernable from the source signal 102. To further simplify the process of later removing artifacts of the spurious signal from the output 114b of the non-linear circuit 114, one approach is to avoid any frequency/phase modulation of the carrier signal.

The carrier modulation of step 406 may be accomplished by quadrature modulation, polar modulation, or one of many techniques familiar to those of ordinary skill in the art.

In step 408, adder 112 adds source signal 502 (present on input 102) to dummy signal 506 (present on output 105 of dummy signal generator 110). The output of the adder 112 may also be referred to as the "conditioned" signal 113, or the "combined" signal. Since its amplitude is now specified, the signal 113 is ready to be processed by the non-linear circuit 114. Thus, in step 410, the adder 112 sends the adjusted output to the nonlinear circuit 114.

In step 411, the nonlinear circuit 114 applies its function to its input 114a, i.e. the conditioned signal 113. For example, if circuit 114 is an amplifier, it amplifies input 114 a. If the circuit 114 is a filter, it filters the input 114 a. However, since the source signal 502 has been adjusted by the addition of the dummy signal 506, the circuit 114 is prevented from producing any amplitude-dependent changes ("non-linear behavior"). In this way, the output 114b of the circuit 114 is linearized.

However, output 114b still contains artifacts of artifact 506. Thus, in step 412, filter 116 is used to remove any signal in output 114b that corresponds to the carrier signal of generator 110, i.e., spurious signal 506. The filter also removes any "intermodulation products", which refer to signals resulting from joint interaction of the input signal with spurious signals having non-linear characteristics. Thus, after filtering, only the remaining signal in the output 118 may be attributable to the source signal 102. However, since the source signal 102 (combined with the dummy signal) is processed linearly by the circuit 114, the output 118 is linearized. The output 118 of the filter 116 represents the final linearized output of the linearizer 101.

Multiple dummy signal embodiments

The above technique adds a single spurious signal, which is directed to the mainly third order non-linear characteristic. Even order non-linear characteristics (e.g., two, four, six, etc.) are not problematic. However, for the fifth order non-linear characteristic, the linearizer 101a is used because it comprises two pseudo signal generators. The higher the order of the non-linear characteristic, e.g. seventh order, ninth order, the higher the number of pseudo signal generators used.

To operate the linearizer 101a, many of the same operations 400 are performed as described above. The following description explains the differences in terms of their appearance and description of the requirements. First, although the envelope complement calculator 108a performs step 404 in a manner substantially similar to the calculator 108 of fig. 1A, the calculator 108 performs additional responsibilities. That is, the envelope complement calculator 108a calculates two pseudo envelopes (instead of one) in step 404, where the combination of the two pseudo envelopes is complementary to the source envelope 504. This concept can be extended to three, four, or any number of pseudo envelopes complementary to the source signal 102. The multiple dummy envelope calculation will be explained in more detail in the appendix included below, showing an example of two dummy envelopes.

Another difference in the multiple dummy signal embodiment is in step 406. That is, each generator 110a, 110B (FIG. 1B) modulates a different carrier signal to provide a signal representative of the calculated pseudo-envelope (A)_de1And A_de2) A false signal of one of the signals. The carrier signals of the generators 110a, 110b have different frequency bandwidths from each other because their two envelopes are added at a common frequency, whereas the application calls for two different signals (having defined envelopes) to be added. The frequency of each carrier signal (or frequencies if phase modulation is used) is different from the frequency bandwidth of the source signal 102 to simplify subsequent removal of the corresponding spurious signals. To further simplify the process of removing artifacts of the spurious signals from the output 114b of the non-linear circuit 114, each carrier signal may appear at a single frequency, i.e., without any phase modulation. Like the single dummy signal embodiment, each dummy signal generator 110a, 110b utilizes modulation such as polarization or quadrature modulation.

Another difference in the multiple dummy signal embodiment occurs at step 408. Here, the adder 112 combines the outputs 105a, 105b from the plurality of different pseudo signal generators 110a, 110b with the source signal 102. Another difference is that the filter in step 412 must filter out all spurious signals, i.e., spurious signals from each frequency of each generator 110a, 110 b. Any applicable intermodulation products are also filtered according to the single-spurious embodiment.

OTHER EMBODIMENTS

Those of skill in the art would understand that information and signals may be represented using any of a variety of different technologies and techniques. For example, data, instructions, commands, information, signals, bits, symbols, and chips that may be referenced throughout the above description may be represented by voltages, currents, electromagnetic waves, magnetic fields or particles, optical fields or particles, or any combination thereof.

Those of skill would further appreciate that the various illustrative logical blocks, modules, and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, modules, circuits, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system. The skilled person will recognize the interactivity of the hardware and software in these cases and how best to implement the described functionality for each particular application. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.

The implementation or execution of the various illustrative logical blocks, modules, and algorithm steps described in connection with the embodiments described herein may be implemented or performed with: a general purpose processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination designed to perform the functions described herein. A general purpose processor may be a microprocessor, but in the alternative, the processor may be any conventional processor, controller, microcontroller, or state machine. A processor may also be implemented as a combination of computing devices, e.g., a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a DSP core, or any other such configuration.

The steps of a method or algorithm described in connection with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. A software module may reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor such the processor can read information from, and write information to, the storage medium. In the alternative, the storage medium may be integral to the processor. The processor and the storage medium may reside in an ASIC.

The previous description of the preferred embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without the use of the inventive faculty. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

The word "exemplary" is used herein to mean "serving as an example, instance, or illustration. Any embodiment described herein as "exemplary" is not necessarily to be construed as preferred or advantageous over other embodiments.

Appendix

1. Summary of the implementations

When the circuit gain depends on the signal amplitude, the circuit increases the bandwidth of the signal. Adding an amplitude-correlated but phase-uncorrelated linearized signal to the input reduces this correlation. Since the phase is uncorrelated, the linearized signal will be at any convenient frequency. The linearized signal causes intermodulation products, but these can be filtered out if the signal frequencies are chosen well. This approach is "blind" in that no non-linear behavior is assumed except for the usual AM-AM (amplitude modulation to amplitude modulation) and AM-PM (amplitude modulation to phase modulation) assumptions.

This report examines the rationale. In view of the practical aspects, detailed implementation issues are not emphasized.

2. Brief introduction to the drawings

Considering an input signal to a non-linear circuit (e.g., a transmitter amplifier), the following is given:

x＝A(t)cos(φ(t)+ω₀t). (1)

where A (t) is amplitude modulation, phi (t) is phase modulation, omega₀Is the center frequency. Using the complex low-pass form yields:

the output signal is given as follows:

y＝G(A)A(t)cos(φ(t)+ω₀t+γ(A))， (3)

here, g (a) dominates the amplitude response (AM-AM), while γ (a) dominates the phase response (AM-PM). The output can be expressed in complex low-pass form as:

wherein the complex gain is:

CG＝G(A)expiγ(A). (5)

the complex gain can be represented by a simple polynomial (the basis function should probably be chosen [ Blachman79] with respect to the input signal):

where k is₁、k₃And k₅Associated with the gain g (dB) and the third and fifth order intercept points of the output (in dB for x in volts):

|k₁|＝10^G/20， (7)

wherein R is₀Is the input/output impedance [ Ha81]. Similar relations for higher order intercept points can be derived. The model assumes that the gain and non-linearity are not a function of frequency-a reasonable assumption is made as long as the frequencies involved are not too extensive.

3. Linearization by intermodulation

In this section it will be explained how the gain of the desired signal t (t) is linearized by adding one or more linearized signals at the input of the non-linear characteristic. The linearized signal is related in amplitude to t (t), but not in phase. Furthermore, it is assumed that the location of the summed signals in frequency is able to filter out various intermodulation products, even though t (t) is not affected.

For most of the cases discussed herein, the combination of the three signals constitutes the input of the nonlinear characteristic. First of all, the first step is to,is the signal for which linear amplification is desired. Secondly, the first step is to carry out the first, is a signal for a linearization amplification process-in principle, reducing third-order non-linear characteristics. Finally, can addIn order to reduce third and fifth order non-linear characteristics. In this way it is possible to obtain,

using equation (6), the third order output term (neglecting the coefficient k)₃) Comprises the following steps:

the term for implementing the amplification of T (t) is exp (i φ)_T) Coefficient of (d) -no exp (i φ)_L) Or exp (i phi)_M). At this time, let A_MThe complex gain of t (t) is therefore:

A_Lthe effect of (t) is due to intermodulation. To linearize the gain of t (t), let:

this process is "blind" in that the third-order nonlinear coefficient k need not be known₃。

Although equation (13) eliminates k₃The effect on the third order nonlinear output, however it also amplifies the effect of other coefficients, so the new complex gain of t (t) (e.g. 7 for N) is

Wherein the content of the first and second substances,

new linear gain k in equation (15)₁' subject to all original coefficients k₁、k₃、k₅、k₇The influence of (c). New third order coefficient k₃Without reference to the original third order coefficient k₃Is influenced by k₅、k₇Influence. Similarly, k₅' now under k₇And (4) acting. New seventh order coefficient k₇' unaffected except for the symbol.

In order to improve the linearity of the method, despite the higher-order coefficient K₁The degree of influence of the increase should be as small as possible, whereas the non-linear behavior should be mainly third order.

For example, consider the following often cited nonlinear characteristic [ Saleh81 ]:

the gain and phase of the CG appear in fig. 1(a) and 2 (a). About A at Taylor series expansion_TThe linear gain is 15dB at 0, which can be found as:

k₁＝5.623

k₃＝-5.623+j5.888

k₅＝2.540-j11.777

k₇＝3.627+j16.589 (20)

it is assumed here that K₁0.2, which corresponds to a_THas a maximum peak amplitude of 0.447 volts.

Fig. 1 complex gain magnitude: (a) equation (19), (b) fifth order non-linear characteristics the modified non-linear coefficients from equations (15) to (18) can be:

(21)

for non-linear behavior with gain compression, the number of times the signal is linearized is added reduces the overall gain. Therefore, | k₁' | is slightly less than | k₁L. The improvement of the third order intercept point was calculated directly to be 6.2 dB. To ensure that other terms are excessively degraded when improving the third-order nonlinearity, the nonlinearity of equation (19) IS modeled with an IS-95 CDMA waveform.

Fig. 2 complex gain phase shift: (a) equation (19), (b) linear phase.

Consider a typical digital radio transmitter architecture. Random binary data is mapped into symbols, filtered, and converted into an analog signal. Although OQPSK is used for the transmitter signal, the technique is applicable to any signaling format. Length of 2¹⁹The independent "m" sequences of-1 provide the original in-phase and quadrature binary data. The data is mapped to the OQPSK constellation and then interpolated by a "pulse shaping" filter (for simplicity adopted from TIA-EIAIS-95 "Mobile Station-Base Station Compatibility Standard for Dual-ModeWideband Spread-Spectrum Cellular Systems", telecommunication industry alliance 1993, month 7) at four times the input rate. The Zero Order Hold (ZOH) is followed by 128 samples per "hold" to represent the simulationAnd (4) waveform. The output of ZOH is passed to a reconstruction filter (Chebyshev, type II 5 th order, stopband 80dB below 3/256 of the sampling frequency).

The figure shows the linearized output and the non-linearized output of the non-linear characteristic. The output power in both cases was 11 dBm. For the case of linearization, the input power is increased by 1dB because the linearization signal reduces the gain of the non-linear characteristic. The output power of the linearized signal is 15.2 dBm. The power in the first adjacent channel is improved by 10-15dB and the power in the second adjacent channel is degraded by 4-5 dB. In practice, whether this is beneficial will depend on the non-linear characteristics in the adjacent channels and the requirements.

Fig. 4 shows an expanded view of the power spectral density. The linearized signal ("L") is placed relatively close to the desired signal ("T") in order to minimize the analog sampling requirements. In practice, the frequency of the linearized signal should be chosen to facilitate the filtering requirements.

Fig. 3 shows the power spectral density of the nonlinear characteristic output with and without the added linearized signal.

Fig. 4 power spectral density of nonlinear characteristic output: a broad view of the desired signal (T) and the linearized signal (L) is shown.

With the aid of the second linearization signal, m (t) ═ a_M(t)expiφ_M(t), it is possible to further improve the linearity. Just as the first add signal removes the effect of the third order coefficients, the second add signal may also be added to remove the effect of the fifth order coefficients.

Equation (13) for A_LAnd A_MThe constraint is generated:

to eliminate the fifth order coefficient k₅The function of, require

Equations (22) and (23) must be satisfied simultaneously. Can be conveniently defined therefor

From which two linearized signals can be solved

Due to A_LAnd A_MMust be greater than or equal to zero at any time, and therefore also requires:

and

P₂＞0 (29)

from which also can be derived

And

it is for all A_TThe values must be true. A group K satisfying the formulas (30) and (31)₁、K₂The solution of (a) is:

the method is suitable for nonlinear characteristics with gain expansion, and is also suitable for nonlinear characteristics with power compression, wherein the power (K (t)) in L (t) and M (t)₁And K₂Determined) reduces the overall gain and reduces the output power. The gain can be further reduced if the input power is increased to compensate.

Fig. 5 shows the power spectral densities of the non-linear characteristics of two summed serial signals: close-up of desired signal

Fig. 5 shows the power spectral density at the nonlinear characteristic output, where both l (t) and m (t) are added at the input. The power of the desired signal is 12 dBm. The linearization signals L (t) and M (t) are not shown, but the output power is 16.6dBm and 20.6dBm, respectively. In this case, the non-linear characteristic (shown in fig. 1(b) and 2 (b)) has only first, third and fifth order terms — so the linearized spectrum is nearly ideal. The complex gain of equation (19) is not used because there is both a high order non-linear characteristic and an input signal at three frequencies, so the intermodulation products are very extensive and difficult to manage within the limited analog bandwidth. Thus, FIG. 5 is ideal because it does not include the effects of seventh and higher order terms.

In principle, more signals are added to achieve higher order compensation. However, this benefit quickly disappears because the power required to linearize the signal increases.

4. Conclusion

A technique has been described for linearizing a circuit by adding one or more signals. This method assumes that the strength of the non-linear characteristic is unknown. However, in practice, a suitable constraint on the non-linear characteristic (making it mainly third order) may still be required.

5. Indexing

[ Blachman79], Blachman, N.M. "The Output Signals and Noise from a non-linear with Amplitude-Dependent Phase Shift", IEEE journal of information theory, Vol.IT-25, first No. 1 month 1979, pages 77-79.

[ Ha81], Ha, T.T. "Solid State Microwave Amplifier Design", John Wileyand Sons Press, 1981.

[ Oppenheim89], Oppenheim, A.V. and Schafer, R.W., "Discrete-Time Signalprocessing", Prentice-Hall Press, 1989.

[Saleh81]，Saleh，A.A.M.，“Frequency Independent and Frequency-dependent models of TWT amplifiers”，IEEE。

Claims

1. A linearization method, comprising operations of:

receiving an amplitude modulated source signal having a source frequency bandwidth and exhibiting a source envelope;

calculating a pseudo-envelope, which when combined will result in a predetermined constant;

generating an amplitude modulated pseudo signal exhibiting a pseudo envelope and appearing over a pseudo frequency bandwidth, the pseudo frequency bandwidth being different from the source frequency bandwidth;

adding said source signal and said pseudo signal to form a combined signal;

directing said combined signal to a processing module whose output exhibits amplitude dependent non-linear characteristics;

signals including signals having spurious frequency bandwidths are filtered from the output to provide a linearized output.

2. The method of claim 1, wherein the computing operation comprises: a dummy envelope is calculated, which if combined will result in a predetermined constant.

3. The method of claim 1, wherein the computing operation comprises: a dummy envelope is calculated, and a predetermined constant is derived if the source envelope and dummy envelope are to be processed by a predetermined formula and the processed envelopes are added.

4. The method of claim 3, wherein the predetermined formula comprises:

squaring the source signal envelope;

the pseudo signal envelope is squared and then the squared pseudo signal envelope is doubled.

5. The method of claim 1, wherein:

calculating a pseudo-envelope comprises calculating a plurality of parameters including deriving a predetermined constant when the source envelope and all pseudo-envelopes are to be combined;

the operation of generating the amplitude modulated pseudo signal includes: generating a plurality of amplitude modulated pseudo signals each exhibiting a respective one of a plurality of pseudo envelopes, each pseudo envelope occurring at a pseudo frequency bandwidth different from the source frequency bandwidth;

the adding operation comprises adding the source signal and all of the pseudo signals to form a combined signal;

the filtering operation includes filtering out signals from the output including signals having all spurious frequency bandwidths to provide a linearized output.

6. An apparatus comprising a plurality of circuits interconnecting conductive elements configured to perform operations to linearize an output of a processing module having a non-linear characteristic that depends on magnitude, the operations comprising:

calculating a dummy envelope, deriving a predetermined constant if the source envelope and dummy envelope are to be combined;

adding the source signal and the dummy signal to form a combined signal;

directing said combined signal to an input of a processing module, an output of said processing module exhibiting amplitude dependent non-linear characteristics;

7. The apparatus of claim 6, wherein the computing operation comprises computing a dummy envelope, and deriving a predetermined constant if the source envelope and dummy envelope are to be added.

8. The apparatus of claim 6, wherein the computing operation comprises computing a dummy envelope, and deriving a predetermined constant if the source envelope and dummy envelope are to be processed by a predetermined formula and the processed envelopes are to be added.

9. The apparatus of claim 8, wherein the predetermined formula comprises:

squaring the source signal envelope;

10. The apparatus of claim 6, wherein:

calculating a pseudo-envelope comprises calculating a plurality of pseudo-envelopes that when combined derive a predetermined constant;

the operation of generating the amplitude modulated pseudo signal includes: generating a plurality of amplitude modulated pseudo signals, each of said pseudo signals exhibiting a respective one of a plurality of pseudo envelopes, each of said pseudo envelopes occurring over a pseudo frequency bandwidth different from the source frequency bandwidth;

11. A linearization apparatus for processing a signal input, the signal input being an amplitude modulated source signal having a source frequency bandwidth and exhibiting a source envelope, the apparatus comprising:

an envelope calculator for calculating a dummy envelope using the source signal, the source envelope and dummy envelope being derived from a predetermined constant when combined;

a pseudo signal generator providing an amplitude modulated pseudo signal showing the calculated pseudo envelope, wherein the pseudo signal occurs at one or more predetermined pseudo frequencies, the predetermined frequencies being different from the source frequency bandwidth;

an adder coupled to the signal input and the pseudo signal generator for adding the source signal and the pseudo signal to form a combined signal that is available for processing by the module having the amplitude dependent non-linear characteristic;

at least one filter for filtering out signals including signals having a spurious frequency bandwidth from the output of the module to provide a linearized output.

12. The apparatus of claim 11, wherein the apparatus further comprises:

the module;

wherein the module is coupled to the adder to receive the combined signal as an input and to provide an output exhibiting a non-linear behavior dependent on the amplitude.

13. The apparatus of claim 11, wherein the envelope calculator is configured such that the calculating operation comprises calculating a dummy envelope that yields a predetermined constant when the source envelope and dummy envelope are to be added.

14. The apparatus of claim 11, wherein the envelope calculator is configured such that the calculating operation comprises calculating a dummy envelope that derives a predetermined constant when the source envelope and dummy envelope are to be processed by predetermined formulas and the processed envelopes are added.

15. The apparatus of claim 14, wherein the predetermined formula comprises:

squaring the source signal envelope;

16. The apparatus of claim 11, wherein:

an envelope calculator for calculating a plurality of dummy envelopes, which when combined derive a predetermined constant;

said pseudo signal generator comprising a plurality of pseudo signal generators for generating a plurality of amplitude modulated pseudo signals exhibiting a respective one of a plurality of pseudo envelopes, each of said pseudo envelopes occurring at a pseudo frequency bandwidth different from said source frequency bandwidth;

the adder is coupled to each pseudo signal generator;

the filter filters out signals from the output including signals having spurious frequency bandwidths.

17. A linearization apparatus for processing an amplitude modulated source signal having a source frequency bandwidth and exhibiting a source signal envelope, the apparatus comprising:

envelope calculation means for calculating a dummy envelope using the source signal, the source envelope and dummy envelope being derived by a predetermined constant when combined;

pseudo signal generating means for generating an amplitude modulated pseudo signal exhibiting the calculated pseudo envelope, the pseudo signal occurring over a pseudo frequency bandwidth different from the source frequency bandwidth;

adding means for adding said source signal and said pseudo signal to form a combined signal, said combined signal being available for processing by a module having amplitude dependent non-linearity;

filtering means for filtering out signals including signals having a spurious frequency bandwidth from the output of the module to provide a linearized output.