Digital Predistortion Training for Power Amplifier Linearization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional digital predistortion (DPD) techniques for power amplifiers face challenges in achieving accurate linearization due to manufacturing variations, aging, and temperature changes, leading to inefficiencies and inaccuracies in error vector magnitude (EVM) and adjacent channel leakage ratio (ACLR) performance, particularly with traditional learning architectures like Direct Learning Architecture (DLA) and Indirect Learning Architecture (ILA) that require additional memory and training, and suffer from non-unique solutions and interpolation inconsistencies.
Innovation Solution
A hybrid learning architecture (HLA) that trains DPD parameters over input signal values outside the forward path without a PA model, using a combiner to generate an error signal based on the difference between the predistorted and adaptation output signals, and a Dual Adaptive Interpolated Lookup (DAIL) table that separates error terms for LUT entries, improving linearization performance and reducing computational complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional DPD techniques are used with traditional learning architectures (DLA or ILA), then linearization performance can be improved, but manufacturing variations, aging, and temperature changes cause inaccuracies in EVM and ACLR performance
Solution Approach 1:
The patent implements a feedback mechanism where the output of the power amplifier is fed back through a digital downconverter and comparator to generate an error signal. This error signal is used to continuously update and adapt the predistortion lookup table, allowing the system to compensate for manufacturing variations, aging, and temperature changes in real-time, thereby maintaining accurate EVM and ACLR performance
Solution Approach 2:
The system performs self-calibration by using its own output to generate the error signal that updates its predistortion parameters. The adaptive predistortion controller automatically adjusts the lookup table entries based on the measured difference between desired and actual output, enabling the system to self-correct without external intervention
2Reliability
If Direct Learning Architecture (DLA) is used to train DPD parameters, then linearization can be achieved, but additional memory and training requirements increase device complexity
Solution Approach 1:
The patent extracts the training function from the forward signal path and places it in a separate adaptive predistortion controller that operates in the feedback path. This separation allows the main DPD datapath to remain simple while the training function is performed independently using the error signal, reducing the memory and computational requirements in the critical signal path
3Reliability
If Indirect Learning Architecture (ILA) is used to train DPD parameters, then linearization can be achieved, but non-unique solutions and interpolation inconsistencies arise
Solution Approach 1:
The patent uses feedback from the actual power amplifier output to directly update the lookup table entries. The comparator generates an error signal based on the difference between the desired and actual output, and this error signal is used to update individual LUT entries. This direct feedback approach avoids the interpolation inconsistencies of ILA by directly measuring and correcting the actual system behavior
Solution Approach 2:
Instead of training the predistortion parameters by modeling the inverse of the PA characteristics (as in ILA), the patent directly measures the actual output error and uses that to update the LUT. This inverted approach of directly correcting based on measured error rather than mathematical inversion eliminates the non-unique solutions problem
4Use of energy by moving object
If power amplifiers are operated close to saturation to maximize efficiency, then power efficiency is improved, but nonlinearity increases causing unacceptable EVM and ACLR performance
Solution Approach 1:
The patent applies preliminary anti-action by predistorting the input signal before it reaches the power amplifier. The lookup table pre-compensates for the known nonlinear characteristics of the PA, so that when the PA operates in its highly efficient saturated region, the predistortion counteracts the nonlinear effects, maintaining both high efficiency and acceptable EVM/ACLR performance
Data Source
AI summary
A circuit for use with an amplification circuit having a predistortion datapath portion, a power amplifier portion and a gain portion. The predistortion datapath portion can output a predistorted signal based on the input signal. The power amplifier portion can output an amplified signal based on the predistorted signal. The gain portion can output a gain output signal based on the amplified signal. The circuit comprises a digital predistortion adaptation portion and a combiner. The digital predistortion adaptation portion can output a predistortion adaptation portion output signal. The combiner can output an error signal. The predistortion adaptation portion output signal is based on the input signal, the gain output signal and the error signal. The error signal is based on the difference between the predistorted signal and the predistortion adaptation portion output signal.