Ethernet PHY Delimiter Randomization to Reduce Spectral Spikes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Fixed stream delimiters in Ethernet communication systems cause correlation and spectral spikes, leading to increased noise emissions and training errors due to their constant nature, which affects signal-to-noise ratio and filter training accuracy.
Innovation Solution
Randomizing delimiters independently of data transmission and selecting disparity reset values based on randomized delimiters to generate an initial running disparity, ensuring that delimiters are not constant and do not adversely impact running disparity, thereby reducing noise emissions and training errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If fixed stream delimiters are used in Ethernet communication, then frame synchronization and boundary detection are improved, but spectral spikes and noise emissions increase
Solution Approach 1:
The patent applies dynamics by making the delimiter pattern changeable rather than fixed. The transmitter dynamically switches between different delimiter patterns (e.g., +1,+1,-1 and -1,-1,+1) based on a randomization bit, transforming the static harmful delimiter into a dynamic element that prevents spectral spikes while maintaining synchronization functionality.
Solution Approach 2:
The patent implements periodic action through the systematic alternation between different delimiter patterns at regular intervals. The delimiter pattern changes periodically based on the randomization bit sequence, creating a rhythmic variation that eliminates continuous spectral spikes while preserving the periodic synchronization function needed for frame detection.
2Reliability
If fixed delimiters are transmitted continuously, then frame boundaries are clearly defined, but training filter accuracy deteriorates due to correlation
Solution Approach 1:
The patent makes the delimiter dynamic by introducing randomization bits that switch between different delimiter patterns. This dynamic variation breaks the correlation between consecutive delimiters, allowing training filters to learn from diverse patterns while still maintaining clear frame boundary definitions through the systematic alternation between standardized patterns.
Solution Approach 2:
The patent changes the parameter of the delimiter pattern based on the randomization bit. By switching between different delimiter patterns (changing the signal parameters) while maintaining the functional requirement for frame boundary detection, the system improves training filter accuracy by providing varied training data without sacrificing boundary clarity.
3Object-generated harmful factors
If delimiters are randomized independently of data, then spectral spikes are reduced, but device complexity increases
Solution Approach 1:
The patent segments the delimiter processing into independent components: a simple randomization bit generator and a delimiter pattern selector. This segmentation allows the complex task of delimiter randomization to be broken down into manageable parts, reducing the overall device complexity while maintaining the benefit of reduced spectral spikes through the coordinated operation of these segmented components.
Data Source
AI summary
Systems and methods are provided for performing operations comprising: accessing, by a transmitter physical layer (PHY) controller, data for transmission to a receiver PHY controller over a network connection; randomizing a delimiter independently of randomizing the data; selecting a disparity reset value based on the randomized delimiter to generate an initial miming disparity; encoding the randomized data based on the initial running disparity; generating a frame that includes the encoded randomized data and the randomized delimiter; and transmitting the frame from the transmitter PHY controller to the receiver PHY controller.