A DSP controlled adaptive feedforward amplifier linearizer

Abstract

Currently, the only available wideband (multiple MHz), accurate, amplifier linearization method is feedforward. Feedforward, though, requires automatic adaptation of key parameters for reliable distortion cancellation as operational and environmental conditions vary. In this thesis, previous analysis on adaptive feedforward linearization is extended to include an _alternative placement of the adaptive signal cancellation coefficient.In contrast to previous analysis, this placement results in a non-quadratic error surface. Consequently, two available criteria for the optimization of the signal cancellation coefficient result in different optimal values. This result can have practical implementation consequences under certain operating conditions. A new analysis is presented that shows that various inaccuracies & the implementation of baseband correlation, such as frequency and phase oeets, filter mismatches, and incomplete image suppression, do not affect the final converged coefficient values. With a novel and appropriate use of DSP, a feedforward linearizer has been implemented with adaptation driven by easily computed gradient signals. This overcomes the dficulties, such as DC -. offsets at the output of analog mixers and masking of weak signals by stronger ones, that slow and/or cause incorrect convergence of many previously reported implementations. The result is 40 dB reduction of intermodulation spectra over a bandwidth of 7 MHz. Coefficient convergence occurs within 50 msec of start-up, and following a 6 dB change in input power, reconvergence occurs \in 3 msec with no loss in distortion suppression. ACKNOWLEDGEMENTSNothing is ever done alone; thus, I would like to thank certain people who helped me along the way with this project. Thanks to my Senior Supervisors, Jim Cavers and Paul Goud, who, I feel, were responsible for advancing my knowledge and confidence in the area of wireless communications to a new and hopehlly much higher level than before.Deserving of thanks as well, is my wife, Irma, for her constant support the whole way through the project, fiom start to finish.

Linearization of Rf Power Amplifiers Using Adaptive Kalman Filtering Algorithm

“…If there is a signal with non-constant envelope at the PA's input, each of its samples will be amplified with different gain and the introduced phase shift will differ according to the input signal amplitude. This nonlinear distrotion is illustrated in the case of 64-QAM modulation as shown in figure (6). The constellation point near the saturation will be more deformed in the case of PA without linearizer.…”

“…To adaptively compensate the amplitude and delays imbalance, an on-line correction of the gain and phase weights can be added to the original structure. An example of adaptive DSP controlled with feedforward technique using LMS algorithm was published in 6 . Feedback control is extensively studied in automatic domain and particularly appreciated in the control of systems with low frequencies dynamics (electrical machines, audio amplifiers, ...) 7 .…”

Steady-State Performance Analysis and Step-Size Selection for LMS-Adaptive Wideband Feedforward Power Amplifier Linearizer

“…The behavior of gradient based adaptive feedforward linearizer in case of a memoryless PA has been studied to some extent in [7], [8], and [14]. Even though the least mean square (LMS) adaptation of coefficients has been presented, the impact of step-size on the final IMD suppression has not been addressed.…”

A wideband architecture for adaptive feedforward linearization

“…In the neighbourhood of the optimum parameter settings, performance and adaptation behaviour in the two locations are very similar. It is shown in [3], [4], though, that placement in the upper (PA) branch causes the minimum power and zero correlation criteria to have slightly different solutions, in contrast to placement in the reference branch, where they are identical.…”

“…This eliminates the DC offset problem. It is shown in [4], [ 101 that filter mismatches in the two downconversions may slow the convergence slightly, but do not bias it. In the second step, the downconversions are made tunable, so they can select specific frequencies over which the minimization is performed.…”