Power Amplifier Memory Modeling Across Waveforms and Power Levels
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing power amplifier models are limited in their ability to accurately compensate for non-linearity and memory effects across varying input waveforms and power levels, leading to performance degradation in RF signal amplification.
Innovation Solution
A universal memory-based model is extracted using a specific extraction waveform with tones of different frequencies, ensuring the model can be used for a variety of waveforms and power levels without the need for additional signal extraction, implemented in a power amplifier modeling system that includes a signal generator, power amplifier, and model extraction hardware.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a power amplifier model is extracted using a specific input waveform and power level, then the model accurately compensates for that specific condition, but the model cannot adequately compensate for input signals having substantially different waveform and/or power level
Solution Approach 1:
The patent applies universality by designing a single extraction waveform that can extract a model valid across multiple waveforms and power levels. The memory polynomial model with specifically chosen tone frequencies and spacing creates a universal model that functions for different input conditions without requiring separate extraction procedures for each condition.
Solution Approach 2:
The patent uses parameter changes by carefully selecting the frequencies and spacing of tones in the extraction waveform. The non-integer-multiple spacing parameter is specifically chosen to prevent spectral overlap and enable the model to capture nonlinear behaviors across a wide range of operating conditions, making the extracted model adaptable to various power levels and waveforms.
2Measurement precision
If multiple models are extracted for different waveforms and power levels, then each model accurately compensates for its specific condition, but the hardware complexity and power consumption increase
Solution Approach 1:
The patent eliminates the need for multiple separate models by creating a single universal model that handles all waveforms and power levels. This reduces hardware complexity by removing the need for multiple model storage units, selection logic, and associated control circuitry that would be required to manage multiple condition-specific models.
Solution Approach 2:
The patent merges the functionality of multiple condition-specific models into a single unified model. By combining the extraction capability into one universal model that covers all operating conditions, the system reduces hardware resources and simplifies the overall architecture while maintaining compensation accuracy across different conditions.
3Measurement precision
If multiple models are extracted for different waveforms and power levels, then each model provides accurate compensation for its condition, but the power consumption increases
Solution Approach 1:
The universal model reduces power consumption by eliminating the need to store and switch between multiple condition-specific models. The single model approach reduces memory access operations, model selection processing, and overall computational overhead, leading to lower power consumption while maintaining accurate compensation across all waveforms and power levels.
4Device complexity
If a single model is used for all waveforms and power levels, then hardware complexity is reduced, but the model cannot adequately compensate for non-linearity and memory effects across varying conditions
Solution Approach 1:
The patent achieves both simplicity and accuracy by carefully designing the extraction waveform parameters. The specific frequency spacing (non-integer multiples) and tone structure enable a single model to capture the full range of nonlinear and memory effects across different operating conditions, maintaining high compensation accuracy without requiring multiple models or complex hardware.
Data Source
AI summary
A method for modelling a power amplifier, including memory effect modelling, for general input waveforms and power levels involves generating an extraction waveform having a plurality of tones each having a different frequency, a difference between the frequencies of two adjacent tones of the plurality of tones not being an integer multiple of a difference in frequency between any two other adjacent tones of the plurality of tones. The method further involves providing the extraction waveform to the power amplifier, receiving output from the power amplifier generated in response to the extraction waveform, and generating a model of the power amplifier based on the output.