Blacklisting Equalization Coefficients for Serial Interface Training
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High-speed serial interfaces face challenges in ensuring reliable data transmission due to variations in circuit design, component manufacture, and environmental conditions, leading to inefficient transmitter and receiver equalization mechanisms that require extensive iterative processes for optimal compensation.
Innovation Solution
A method for determining the best impedance tuning setting in high-speed serial interfaces by using a blacklist of equalization coefficients to skip known unacceptable settings, thereby reducing the number of combinations to test and speeding up the link training process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If iterative equalization processes are used to achieve optimal compensation settings, then transmission reliability is improved, but link training time increases significantly
Solution Approach 1:
The patent pre-calculates and stores equalization coefficients for various channel conditions before actual link training. During link training, the system retrieves pre-computed coefficients based on measured channel characteristics, avoiding iterative optimization. This preliminary preparation enables immediate application of optimal settings, resolving the contradiction between achieving reliable transmission and minimizing training time.
Solution Approach 2:
The patent pre-calculates and stores equalization coefficients for various channel conditions before actual link training. During link training, the system retrieves pre-computed coefficients based on measured channel characteristics, avoiding iterative optimization. This preliminary preparation enables immediate application of optimal settings, resolving the contradiction between achieving reliable transmission and minimizing training time.
2Reliability
If comprehensive equalization settings are tested to ensure optimal performance, then transmission quality is improved, but the complexity of the calibration process increases
Solution Approach 1:
The patent pre-calculates and stores equalization coefficients for various channel conditions before actual link training. During link training, the system retrieves pre-computed coefficients based on measured channel characteristics, avoiding iterative optimization. This preliminary preparation enables immediate application of optimal settings, resolving the contradiction between achieving reliable transmission and minimizing training time.
Solution Approach 2:
The patent transforms the complex iterative optimization problem into a parameter lookup task by pre-computing equalization coefficients for different channel conditions. The system measures channel characteristics (such as loss and dispersion parameters) and selects appropriate pre-computed coefficients, converting a complex calibration process into simpler parameter measurement and selection.
Data Source
AI summary
A serial interface comprises a receiver including a first input compensation module with a first setting that selects a first value from among a plurality of first values for a first input characteristic of the receiver, a memory to store a first blacklist value from among the first values, and a control module to select each of the first values, except for the first blacklist value, to evaluate an indication of a performance level of the receiver for each of the selected first values, and to select a particular first value based upon the indications of the performance level of the receiver.


