LNL Compensator FIR Joint Optimization for Nonlinear Distortion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing signal processing technologies face challenges in optimizing compensators for analog to digital converters (ADCs) and digital to analog converters (DACs) to effectively mitigate non-linear distortion while minimizing complexity and power consumption.
Innovation Solution
A Linear, Memory-less Non-linear (LNL) compensator mechanism is introduced, which jointly optimizes first stage and second stage Finite Impulse Response (FIR) filters using calibration signals and a least mean square formulation, determining optimal filter lengths and coefficients to meet performance thresholds while reducing complexity and power consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If compensation filters are designed to correct non-linear distortion in ADCs, then signal processing performance is improved, but device complexity and power consumption increase
Solution Approach 1:
The compensator is divided into multiple parallel branches, each handling a specific order of non-linearity (second-order, third-order, etc.). Each branch contains its own FIR filter and non-linear filter, allowing independent optimization and processing of different distortion components without increasing overall system complexity
Solution Approach 2:
The patent employs joint optimization of FIR filter parameters (coefficients and lengths) across all branches simultaneously. By changing the parameters of multiple filters together and optimizing them as a unified system, the patent achieves better performance with reduced individual filter complexities compared to separate optimization approaches
2Reliability
If compensation filters are designed to correct non-linear distortion in ADCs, then signal processing performance is improved, but power consumption increases
Solution Approach 1:
The compensator processes different orders of non-linearity in separate parallel branches, allowing power-efficient processing of each distortion order independently. This segmentation enables selective activation of branches based on signal conditions, reducing overall power consumption while maintaining performance
Solution Approach 2:
Joint optimization of FIR filter parameters across all branches enables finding the optimal balance between performance and power consumption. By simultaneously adjusting coefficients and lengths, the system achieves maximum efficiency with minimum power requirements for the given performance level
3Reliability
If FIR filter lengths are increased to improve non-linear distortion mitigation, then compensation performance is improved, but device complexity and computation increase
Solution Approach 1:
Different FIR filter lengths are used in different branches based on the specific requirements for mitigating each order of non-linearity. This segmented approach allows shorter filters where sufficient and longer filters only where necessary, reducing overall complexity while maintaining effective distortion mitigation
Solution Approach 2:
The patent performs joint optimization of FIR filter lengths and coefficients across all branches simultaneously. This unified parameter optimization finds the optimal filter length configuration that achieves the required distortion mitigation with minimum total complexity, avoiding unnecessary length increases in any individual filter
Data Source
AI summary
A mechanism is included for jointly determining filter coefficients for Finite Impulse Response (FIR) filters in a Linear, Memory-less Non-linear (LNL), Linear compensator. Calibration signals are applied to a signal converter input in a test and measurement system. Non-linear signal components are determined in signal output from the signal converter. Non-linear filter components are determined at the LNL compensator based on the calibration signals. The non-linear signal components are then compared to the non-linear filter components. The comparison is then resolved to determine filter coefficients for first stage Finite Impulse Response (FIR) filters and second stage FIR filters in the LNL.