Adaptive Pattern Filtering for Clock and Data Recovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing clock and data recovery systems face challenges in minimizing the interaction between clock and data recovery and decision feedback equalization, leading to corrupted timing recovery and reduced bandwidth due to the application of pattern filters, especially in low-loss transmission channels.
Innovation Solution
Adaptive pattern filtering is applied to discard edges only when DFE feedback levels can corrupt the timing recovery, using a phase detector that receives in-phase and quadrature samples, as well as samples representing the sum and difference of the received signal and equalization coefficients, to selectively suppress timing information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If pattern filtering is applied to remove edges corrupted by DFE feedback, then timing recovery reliability is improved, but bandwidth is reduced due to fewer updates
Solution Approach 1:
The system dynamically adjusts the pattern filtering based on detected interaction levels between DFE feedback and timing recovery. When interaction is detected, pattern filtering is applied to protect timing recovery; when interaction is minimal, filtering is reduced or removed to maintain bandwidth. This dynamic adaptation resolves the contradiction by making the filtering behavior conditional rather than static.
Solution Approach 2:
The system changes the parameter of pattern filtering application based on channel conditions and DFE feedback strength. By monitoring the interaction level and adjusting whether pattern filtering is applied, the system optimizes the balance between timing recovery reliability and bandwidth utilization according to current transmission conditions.
2Measurement precision
If pattern filtering removes edges to avoid DFE corruption, then timing information accuracy is improved, but device complexity increases due to filtering logic
Solution Approach 1:
The system extracts and removes only the specific harmful component (corrupted edges identified by pattern filtering) from the timing recovery input, while preserving the useful timing information from non-corrupted edges. This selective extraction approach improves timing accuracy without requiring complete filtering of all edges, thereby limiting the increase in complexity.
Solution Approach 2:
The pattern filtering logic acts as an intermediary between the DFE feedback path and the timing recovery path. It selectively blocks corrupted timing information while allowing clean timing information to pass through, mediating the interaction between these two paths and resolving the contradiction between accuracy and complexity.
3Reliability
If dual path architecture is used to separate clock and data recovery, then interaction between timing recovery and DFE is reduced, but circuitry complexity increases
Solution Approach 1:
The system segments the timing recovery input stream by identifying and separating corrupted edges from non-corrupted edges using pattern filtering. This segmentation allows selective processing where only corrupted portions are filtered out, achieving timing recovery independence without duplicating the entire clock and data recovery path.
Solution Approach 2:
Instead of creating a complete duplicate dual-path architecture, the system uses pattern filtering to selectively copy and process only the necessary timing information while filtering out corrupted portions. This approach achieves the goal of reducing interaction without requiring a full duplicate path, thereby limiting complexity increase.
Data Source
AI summary
Systems and methods disclosed herein provide for adaptively applying pattern filters so that the edges are discarded only when the DFE feedback has adapted to levels that can corrupt the timing recovery. Embodiments of the systems and methods provide for a phase detector that selectively suppresses timing information based on the logic level states of the Qp and Qm data samples associated with the received signal.