Memory Controller Read Voltage Inferrer for Dynamic Read Quality Evaluation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current memory devices face challenges in ensuring the quality of read operations, particularly in inferring optimal read voltages, which can lead to read failures and data inaccuracies due to inappropriate voltage levels.

Innovation Solution

A storage device with a memory controller that includes a read voltage inferrer to evaluate read operations and infer secondary read levels based on read quality evaluation results, using machine learning models to determine optimal read voltages for improved data retrieval.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If fixed read voltage levels are used in memory devices, then the device structure remains simple, but read failures and data inaccuracies occur due to inappropriate voltage levels

Engineering Contradiction:
Improveread operation reliabilityVSAvoidread voltage control complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent implements a feedback mechanism where read operation results are evaluated and used to adjust subsequent read voltage levels. The memory controller monitors read quality metrics and dynamically modifies read voltages based on this feedback, creating a closed-loop system that continuously optimizes read operations while maintaining reliability.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The patent transitions from static fixed read voltage levels to dynamic adjustable read voltage levels. The read voltage is no longer a fixed parameter but becomes a dynamic variable that adapts based on read operation outcomes, memory cell state, and evaluated quality metrics, enabling the system to respond to changing conditions.

Inventive Principle:
Principle #15Dynamics

2Measurement precision

If read voltage levels are adjusted dynamically based on read quality evaluation, then data accuracy improves, but the control mechanism becomes more complex

Engineering Contradiction:
Improveread data accuracyVSAvoidread voltage control mechanism
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The memory controller performs self-adjustment of read voltage levels by evaluating its own read operation results. The system uses its internal resources to monitor read quality, infer appropriate voltage adjustments, and apply corrections without requiring external intervention, thereby improving data accuracy while managing complexity through autonomous operation.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent replaces manual or external voltage adjustment mechanisms with an automated electronic control system. Machine learning models and algorithms substitute for traditional mechanical or manual voltage setting methods, enabling precise, dynamic voltage adjustment based on real-time read quality evaluation.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

3Reliability

If machine learning models are used to infer optimal read voltages, then read operation quality improves, but processing time increases

Engineering Contradiction:
Improveread operation qualityVSAvoidvoltage inference processing time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent performs preliminary training of machine learning models during manufacturing or initial operation phases, storing pre-computed voltage inference parameters and decision rules. During actual read operations, the system applies these pre-trained models quickly without requiring extensive real-time computation, thus improving read quality while minimizing additional processing time.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent implements a hybrid approach where machine learning models provide voltage recommendations but final voltage selection may use simplified rules or thresholds for common cases. This partial application of complex ML inference reduces processing overhead while maintaining improved read quality for critical operations where the full ML model is applied.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentUS12068046B2Memory device and operating method thereof
Publication Date: 2024.08.20 SK HYNIX INC
  • US12068046B2 patent drawing
  • US12068046B2 patent drawing
  • US12068046B2 patent drawing

AI summary

A storage device includes: a memory device including a plurality of memory cells, the memory device performing a read operation of reading data stored in selected memory cells among the plurality of memory cells; and a memory controller for receiving a read request from a host, and controlling the memory device to perform the read operation corresponding to the read request. The memory controller includes a read voltage inferrer for, when the read operation is completed, receiving read information on the read operation from the memory device, performing a read quality evaluation operation of evaluating the read operation based on the read information, and performing a read voltage inference operation of inferring a secondary read level corresponding to the read information according to a result of the performing the read quality evaluation operation.