Microwave Power Amplifier Predistortion for Long-Term Memory Effects
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional predistorter algorithms for microwave power amplifiers are sensitive to input signal bandwidth and amplitude distribution, requiring continuous recalculations of coefficients and being ineffective for varying modulation formats and power levels, especially due to long-term memory effects from dynamically changing internal biasing and temperature.
Innovation Solution
A dynamic predistortion method using a system model with an inverse memory model is implemented, characterized by X-parameter kernels that predict and counteract amplifier distortion, allowing for predistortion of signals across a wide range of modulation bandwidths and amplitude distributions without feedback loops.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional predistorter algorithms with constant coefficients are used, then the predistorter works well for signals with the same spectral and statistical characteristics as the test sequence, but the performance degrades quickly when input signals have different modulation bandwidth or amplitude distribution
Solution Approach 1:
The patent transforms the static predistortion coefficients into dynamic time-varying coefficients that adapt to changing signal characteristics. The system models the amplifier's long-term memory effects as a time-varying system and uses dynamic predistortion coefficients that evolve over time to match the changing operating conditions, thereby maintaining predistortion accuracy across different signal types and bandwidths.
Solution Approach 2:
The patent changes the parameters of the predistorter from fixed constant coefficients to time-varying coefficients that adapt to different signal conditions. By modeling the amplifier characteristics as time-varying and相应地 adjusting the predistortion parameters, the system achieves versatility across different modulation formats, bandwidths, and power levels while maintaining high predistortion accuracy.
2Measurement precision
If the complexity of the function F(.) is increased to improve predistortion performance, then the number of coefficients increases, but the coefficients cannot be directly measured and must be determined by a difficult fitting procedure
Solution Approach 1:
The patent replaces the difficult nonlinear fitting procedure with a more straightforward identification method based on measured amplifier characteristics. Instead of performing complex iterative fitting to determine coefficients, the system uses direct measurement and modeling approaches that simplify the coefficient determination process while maintaining high predistortion accuracy.
3Adaptability or versatility
If new sets of coefficients are continuously calculated using feedback loops, then the predistorter can adapt to changing input signal characteristics, but the system requires many digital and analog parts, is challenging to implement, and requires significant time to achieve accuracy
Solution Approach 1:
The patent performs preliminary modeling of the amplifier's long-term memory effects during an identification phase, capturing the time-varying characteristics in advance. This preliminary action allows the system to use pre-characterized time-varying models for predistortion without requiring complex real-time feedback loops, thereby reducing implementation complexity while maintaining adaptability to changing signal conditions.
Data Source
AI summary
A method and system for predistorting signals provides a test signal to model a non-linear component. Model kernels representative of static and dynamic parts of the model are extracted from an output of the non-linear component responsive to the test signal. The dynamic part represents memory effects of the non-linear component. The model kernels are then used to calculate an inverse memory model component model. An input signal is predistorted using the inverse memory model.