Non-Volatile Memory Link Retraining From Transmission Parity Errors

The patent performs re-training operations before actual data communication to establish proper signal alignment. The controller transmits training sequences and adjusts sampling timing in advance, so that when real data transmission occurs, the signal alignment is already optimized for high-speed communication without requiring continuous monitoring during data transfer.

The patent implements a feedback mechanism where the non-volatile memory device reports its operational state to the controller, and the controller adjusts sampling timing based on this feedback. This allows the system to maintain signal alignment at high frequencies by continuously adapting to temperature and voltage changes through controlled re-training operations triggered by state information.

Instead of continuous monitoring, the patent performs re-training operations periodically based on triggered conditions. The controller monitors state information from the non-volatile memory and initiates re-training only when necessary (e.g., after temperature or voltage changes), rather than continuously adjusting signal alignment, thereby reducing resource overhead while maintaining reliability.

The non-volatile memory device autonomously monitors its own operational state and generates state information that triggers re-training when needed. This self-service mechanism eliminates the need for constant external monitoring and control, reducing the computational burden on the controller while ensuring signal alignment is maintained through event-driven re-training operations.