Digital Pre-Distortion Architecture for Wideband Doherty Amplifiers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional digital pre-distortion architectures do not provide adequate linearization for certain amplifier designs, particularly under specific signaling conditions such as extremely wideband signals and Doherty amplifiers.
Innovation Solution
A digital pre-distortion architecture that generates a pre-distorted signal by combining frequency-dependent pre-distortion signals through products of derivative pre-distortion functions and input signals, using multiple signal paths and summation to improve linearization for non-linear amplifiers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If conventional digital pre-distortion architecture is used, then the system is simple to implement, but it does not provide adequate linearization for certain amplifier designs under specific signaling conditions
Solution Approach 1:
The pre-distortion function is segmented into multiple parallel paths: a first path processes the input signal through a first pre-distortion function, a second path processes the input signal through a second pre-distortion function, and a third path processes the derivative of the input signal through a third pre-distortion function. Each path contributes to the overall pre-distorted signal, allowing the system to achieve better linearization accuracy by combining multiple simplified processing branches rather than using a single complex processing chain.
Solution Approach 2:
The patent introduces a new dimension to the pre-distortion processing by incorporating the derivative of the input signal as a separate processing path. This adds temporal dynamics (frequency-dependent behavior) to the traditional amplitude-dependent pre-distortion, effectively moving from a static correction approach to a dynamic one that better models the amplifier's non-linear behavior across different signal frequencies.
2Measurement precision
If conventional pre-distortion is applied, then the processing is computationally simple, but it fails to accurately model non-linear amplifiers under wideband signaling conditions
Solution Approach 1:
The computational task is segmented across three parallel processing paths, each handling a specific aspect of the pre-distortion function. This segmentation allows the system to distribute the computational load and better model the amplifier's non-linear behavior by combining the outputs of multiple specialized processing branches rather than using a single monolithic computational structure.
Solution Approach 2:
The patent adds a temporal dimension to the pre-distortion computation by including the derivative of the input signal in the third processing path. This enables the model to capture frequency-dependent non-linear effects that are present in wideband signals, improving modeling accuracy without requiring a complete redesign of the computational architecture.
Data Source
AI summary
An input signal is pre-distorted to reduce distortion resulting from subsequent signal amplification. Frequency-dependent pre-distortion is preferably implemented in combination with frequency-independent pre-distortion, where the frequency-dependent pre-distortion is generated by expanding the derivative of a product of a pre-distortion function and the input signal and then relaxing constraints on the pre-distortion function and/or on frequency-dependent filtering associated with the frequency-dependent pre-distortion. In one implementation, four different frequency-dependent pre-distortion signals are generated for the expansion using up to four different pre-distortion functions and up to four different frequency-dependent filters.