Amplifiers – Hum or noise or distortion bucking introduced into signal...
Reexamination Certificate
1999-10-26
2001-06-12
Pascal, Robert (Department: 2817)
Amplifiers
Hum or noise or distortion bucking introduced into signal...
C330S136000, C375S296000, C455S126000
Reexamination Certificate
active
06246286
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Technical Field of the Invention
The present invention relates in general to the field of communication systems, and in particular, to adaptive linearization of power amplifiers in such communication systems.
2. Description of Related Art
In order to keep pace with the ever increasing demand for higher capacity wireless and personal communication services, modern digital communication systems have become increasingly reliant upon spectrally efficient linear modulation schemes, such as Quadrature Phase Shift Keying (QPSK), Quadrature Amplitude Modulation (QAM), and recently 3&pgr;/8-8PSK used in the Enhanced Data Rates for GSM Evolution (EDGE) system. Unlike conventional digital modulation techniques which utilize a constant envelope, linear modulation schemes exploit the fact that digital baseband data may be modulated by varying both the envelope (e.g., amplitude) and phase of an RF carrier. Because the envelope and phase offer two degrees of freedom, digital baseband data may be mapped into four more possible RF carrier signals, enabling the transmission of more information within the same channel bandwidth than if just the envelope or phase were varied alone. As a result, linear modulation schemes provide significant gains in spectrum utilization, and have become an attractive alternative to conventional digital modulation techniques.
The variation of both the envelope and phase of the RF carrier, however, causes linear modulation schemes to be highly susceptible to the inherent nonlinear distortion associated with power amplifiers. Although conventional digital modulation techniques are less susceptible to such distortion due the use of a constant envelope, the non-constant envelope utilized by linear modulation schemes causes the gain and phase-shift of the power amplifier to vary as a function of the input signal. This non-constant gain and phase-shift, in turn, causes two types of nonlinear distortion. The first type of nonlinear distortion, known as AM/AM distortion, occurs when the input power and the output power depart from a linear relationship. The second type, known as AM/PM distortion, occurs when the phase-shift of the power amplifier varies as a function of the input power level.
If the power amplifier used to amplify linearly modulated signals fails to compensate for both types of nonlinear distortion, the power amplifier will generate unwanted intermodulation products and cause an accompanying degradation in the quality of the communications. When intermodulation products occur outside the channel bandwidth, for example, an effect known as spectral regrowth or widening causes increased interference with communications in adjacent channels. Furthermore, intermodulation products occurring within the channel bandwidth may distort the modulated signal to such an extent that it cannot be properly reconstructed or detected at the receiver, resulting in increased bit error rates. Therefore, in order to prevent unwanted intermodulation products and avoid the accompanying degradation in the quality of communications, linear modulation schemes require a linear power amplifier with a constant gain and phase-shift for all operating power levels.
Unfortunately, because power amplifiers are inherently nonlinear devices, the gain and phase-shift of power amplifiers vary in a complex, nonlinear manner depending on such variables as component aging, component variation, channel switching, power supply variation, component drift, temperature fluctuations, and the input signal itself. Existing approaches, such as continuous feedback, feedforward networks and conventional predistortion techniques, have attempted to compensate for these nonlinear characteristics by utilizing some form of continuous feedback loop or fixed pre-processing or post-processing network. These approaches, however, either fail to adaptively compensate for time-varying fluctuations in nonlinear characteristics or prove difficult to implement at RF frequencies. For example, continuous feedback approaches, such as negative feedback or Cartesian feedback, typically require a high loop bandwidth and could cause stability problems when operated at high frequencies. Feedforward networks, on the other hand, cannot adaptively compensate for variations in distortion characteristics due to the fixed nature of the feedforward network, and require precise matching and scaling of components in order to avoid inadvertently introducing additional nonlinear distortion. Conventional predistortion techniques similarly fail to adaptively compensate for variations in nonlinear characteristics due to the use of a fixed set of predistortion coefficients.
One existing approach that adaptively compensates for variations in nonlinear distortion is an approach known as adaptive predistortion. In contrast to the conventional predistortion technique mentioned earlier, traditional adaptive predistortion periodically senses the output of the power amplifier and updates the predistortion coefficients for time-varying nonlinearities in the forward path. These updated predistortion coefficients are then used to predistort the input signal in such a manner that a linear amplified signal is produced at the output of the power amplifier.
Although traditional adaptive predistortion provides adequate linearization of a power amplifier, the traditional adaptive predistortion technique places significant processing demands on the digital signal processor used to implement this technique. Typically, the look-up table that stores the predistortion coefficients must be updated several times per symbol (e.g., five times per symbol) depending on the oversampling rate. Moreover, a typical “burst” in a Time Division Multiple Access (TDMA) system may include as many as 100-200 symbols. As a result, this example would require the digital signal processor to update the lookup table 500-1000 times per burst. This places a significant burden on the precessing requirements (and corresponding cost) of the digital signal processor and increases current consumption.
A further disadvantage of the traditional adaptive predistortion technique is that it requires a quadrature demodulator in the feedback loop. This quadrature demodulator is required in order to enable the digital signal processor to compare the data stream detected at the output of the power amplifier with the input data stream. In addition to the increased costs and current consumption, the quadrature demodulator can also introduce errors which will be reflected in the updated predistortion coefficients and will adversely affect the ability to compensate for nonlinear distortion in the power amplifier. Therefore, in view of the deficiencies of existing approaches, there is a need for an adaptive linearization technique that can effectively compensate for time-varying nonlinearities of the power amplifier and at the same time relax the processing requirements of the digital signal processor and decrease current consumption.
SUMMARY OF THE INVENTION
The deficiencies of the prior art are overcome by the method and apparatus of the present invention. For example, as heretofore unrecognized, it would be beneficial to measure distortion characteristic(s) across a power amplifier during amplification of a distortion detection signal. This distortion detection signal comprises, for example, a ramp-up signal or ramp-down signal. Preferably, however, the distortion detection signal comprises a burst up-ramp or burst down-ramp signal commonly used before or after a communication burst in a Time Division Multiple Access (TDMA) system. Using a burst up-ramp or burst down-ramp signal offers the added advantage of allowing the principles of the present invention to be easily incorporated into existing TDMA communication systems.
Based on the measured distortion characteristic(s) and known characteristics of the distortion detection signal, a relationship between the measured distortion characteristic(s) and input power is calculated. A predistortion lookup table is updated in acco
Choe Henry
Jenkens & Gilchrist PC
Pascal Robert
Telefonaktiebolaget LM Ericsson
LandOfFree
Adaptive linearization of power amplifiers does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Adaptive linearization of power amplifiers, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Adaptive linearization of power amplifiers will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2478171