Decision Feedback Equalizer for Stable Tap Coefficient Convergence
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Decision feedback equalizers face challenges in stabilizing tap coefficient convergence and reducing convergence time, especially when pre-cursor interference is significant, due to the lack of effective equalization of pre-cursor effects and the instability caused by the least mean square algorithm.
Innovation Solution
An equalizer system incorporating a data sampler, arithmetic circuit, tap coefficient calculation circuit, and determination circuit that adaptively updates tap coefficients based on the relationship between reference data and subsequent data symbols, ensuring accurate convergence and reduced convergence time by determining the necessity of tap coefficient updates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the least mean square algorithm is used to determine tap coefficients in a decision feedback equalizer, then the equalization process can be implemented, but the convergence characteristics become unstable when pre-cursor effects are significant
Solution Approach 1:
The patent applies preliminary action by using data from future symbols (post-cursor data) to guide the current tap coefficient update decision. The determination circuit checks future data values before allowing coefficient updates, ensuring that updates only occur when they will not degrade equalization performance. This proactive approach stabilizes convergence by preventing harmful updates before they can occur.
Solution Approach 2:
The patent implements feedback by using the determined presence or absence of update conditions to control the tap coefficient calculation circuit. The system continuously monitors data patterns and adjusts coefficient updates accordingly, creating a closed-loop control mechanism that stabilizes convergence characteristics while maintaining adaptation to changing channel conditions.
2Measurement precision
If tap coefficients are continuously updated using the least mean square algorithm, then the equalizer can adapt to channel changes, but the convergence accuracy decreases when pre-cursor interference is large
Solution Approach 1:
The patent applies partial action by selectively updating only those tap coefficients for which the update condition is satisfied. Instead of continuously updating all coefficients, the determination circuit evaluates each update opportunity and allows updates only when future data patterns indicate it will improve equalization. This selective approach maintains accuracy while preserving necessary adaptation capability.
Solution Approach 2:
The system performs self-service by automatically determining update conditions based on incoming data patterns without external control. The determination circuit autonomously evaluates whether upcoming data values will benefit from coefficient updates and controls the update process accordingly, enabling the equalizer to self-optimize its convergence behavior.
3Device complexity
If the decision feedback equalizer only equalizes post-cursor effects, then the implementation remains simple, but pre-cursor interference remains unaddressed
Solution Approach 1:
The patent introduces an intermediary mechanism (the determination circuit) that mediates between the simple post-cursor equalization structure and the need to address pre-cursor interference. This intermediary evaluates future data patterns and controls coefficient updates to indirectly compensate for pre-cursor effects without requiring explicit pre-cursor filtering, thus maintaining structural simplicity while improving equalization performance.
Data Source
AI summary
An equalizer includes a data sampler that samples input data and outputs a time-series data string according to the input data, an arithmetic circuit that multiplies a data string output before reference data in the data string output from the data sampler by a tap coefficient and forms the input data by an arithmetic operation of a multiplication result and an input signal, a tap coefficient calculation circuit that updates the tap coefficient based on a data string output before the reference data, and a determination circuit that receives the reference data and data output after the reference data in the data string and controls presence or absence of update of the tap coefficient performed by the tap coefficient calculation circuit.