Link Equalization Training for High-Rate Mobile Interconnects

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The existing UniPro and M-PHY standards lack a defined link equalization training process for data rates above HS-G6, which is necessary for high-speed communication in mobile devices, leading to inefficiencies in data transmission.

Innovation Solution

A link equalization training method is introduced, involving power mode changes, training sequences, and decision mechanisms to ensure reliable link establishment at higher data rates, using techniques like Continuous-Time Linear Equalizer (CTLE), Feed-Forward Equalizer (FFE), and Decision Feedback Equalizer (DFE) to compensate for inter-symbol interference.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If link equalization training is performed for high-speed gear configurations above HS-G6, then data transmission reliability is improved, but the complexity of the training process increases due to the need for multiple power mode changes and decision mechanisms

Engineering Contradiction:
Improvedata transmission reliabilityVSAvoidtraining process complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The link equalization training process is segmented into multiple distinct phases: start-up link establishment, power mode changes to achieve target data rate, link equalization training execution, and decision mechanism for error handling. Each phase is independently defined with specific control steps, making the complex high-speed training process manageable and systematic.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Before performing link equalization training at high data rates above HS-G6, the method preliminarily establishes a start-up link and performs power mode changes to configure the target data rate. This preliminary setup ensures that all necessary conditions are met before the actual equalization training begins, preventing errors and improving reliability.

Inventive Principle:
Principle #10Preliminary action

2Speed

If link equalization training is performed at data rates above HS-G6, then communication speed is improved, but the risk of error events and incomplete training increases

Engineering Contradiction:
Improvedata transmission speedVSAvoidtraining completion reliability
Core Design Contradiction:
SpeedVSReliability

Solution Approach 1:

A decision mechanism is implemented that continuously monitors training status and error events during link equalization training at high data rates. Based on feedback from training completion status and error detection, the method dynamically adjusts by either proceeding with link-up or resetting to start-up link, ensuring reliable training completion even at speeds above HS-G6.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The method prepares error handling strategies in advance by implementing a decision mechanism that is ready to handle error events and incomplete training scenarios. This cushioning approach ensures that if errors occur during high-speed training, the system can promptly reset and retry, preventing complete training failure.

Inventive Principle:
Principle #11Beforehand cushioning (Prior cushioning)

3Adaptability or versatility

If power mode changes are performed to achieve target data rate, then adaptability to different speed requirements is improved, but the time required for link setup increases

Engineering Contradiction:
Improvedata rate adaptabilityVSAvoidlink setup time
Core Design Contradiction:
Adaptability or versatilityVSLoss of time

Solution Approach 1:

Power mode changes are performed as periodic, discrete actions to transition the link to the target data rate. Each power mode change is a distinct, timed operation that systematically adjusts the data rate in controlled steps, allowing the system to adapt to different speed requirements while maintaining predictable setup timing.

Inventive Principle:
Principle #19Periodic action

Data Source

PatentUS12627535B2Link equalization training method for mobile device
Publication Date: 2026.05.12 MEDIATEK INC
  • US12627535B2 patent drawing
  • US12627535B2 patent drawing
  • US12627535B2 patent drawing

AI summary

The disclosure provides a link equalization training method for a mobile device, comprising: setting to start-up link; performing a first power mode change to change a link rate to a first data rate; performing link equalization training on a receiver of a second node in response to a plurality of training sequences from a first node; when link equalization training in the receiver of the second node is not completed and there is no any error event, making a decision mechanism to perform link equalization training in the receiver of the second node; when link equalization training in the receiver of the second node is not completed and there is any error event, making the decision mechanism to reset for start-up link; and when link equalization training in the receiver of the second node is completed, setting link-up in the first data rate for data transfer.