PR-MLSE Waveform Equalization for Eye Diagram Measurement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The discrete output of a Maximum Likelihood Sequence Equalizer (MLSE) prevents the construction of an eye diagram, limiting quality metrics evaluation and leading to unstable and inconsistent results due to incorrect symbol estimates.
Innovation Solution
A method and system that generate a continuous signal using Partial Response Maximum Likelihood Sequence Equalization (PR-MLSE) by applying inter-symbol interference (ISI) to a feed forward equalizer, characterized by the equation yˆk=rk+αrk+1-α(mk-1+mk+1)1+α2, allowing for eye diagram generation and stable measurements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If discrete symbol sequence output from MLSE is used, then MLSE performance can be evaluated with threshold testing, but eye diagram construction is prevented and quality metrics evaluation is limited
Solution Approach 1:
The patent introduces an intermediary continuous signal representation that bridges the discrete MLSE output and the continuous eye diagram requirements. This intermediary signal allows eye diagram construction while maintaining correlation to MLSE performance through a mathematical relationship, effectively mediating between the discrete decision output and continuous visualization needs.
Solution Approach 2:
The patent transforms the discrete symbol sequence into a continuous signal by changing the parameter domain from discrete symbol values to continuous amplitude values. This parameter transformation enables the signal to be displayed on eye diagrams while preserving the essential MLSE performance characteristics through a defined mathematical relationship.
2Reliability
If feedback of estimated symbols is used in MLSE, then symbol estimation can be performed, but bursts of errors occur when estimates are incorrect leading to unstable results
Solution Approach 1:
The patent creates a copy of the MLSE decision process in the form of a continuous signal model. Instead of directly using the discrete feedback that causes error bursts, the patent models the decision process continuously, allowing error analysis without the destabilizing effect of actual error bursts in the measurement system.
Solution Approach 2:
The patent prepares for potential estimation errors by using a continuous signal model that doesn't propagate errors in the same discrete feedback manner. This beforehand cushioning prevents the cascade of errors that occurs in traditional MLSE when incorrect estimates are fed back, providing more stable measurements.
Data Source
AI summary
A test system implemented method of equalizing a digital signal under test (SUT) results in a waveform representation of the output of a Partial Response Maximum Likelihood Sequence Equalizer (PR-MLSE). The method includes applying the SUT, having greater than one symbol period of inter-symbol interference (ISI), to a feed forward equalizer to obtain a Partial Response equalized sequence rk having an ISI characterized as (1+αD), wherein rk denotes the kth equalized sample of the waveform, sampled at the symbol rate, and α is a programmable variable. The method further includes obtaining a known symbol sequence, mk, as a user input to the system, and determining y(capped)k, the kth sample of the PR-MLSE equivalent waveform, in accordance with the following equation,yˆk=rk+αrk+1-α(mk-1+mk+1)1+α2.