Adaptive Control for Multilayered Nonvolatile Memory

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Multilayered nonvolatile semiconductor memory devices face challenges in maintaining optimal threshold voltage distribution profiles across different layers, leading to wider distribution profiles and varying characteristics between memory cells in different layers.

Innovation Solution

Implementing an adaptive control scheme that divides memory cell arrays into predetermined block groups, using a lookup matrix to store characteristic information and a state machine to control operating conditions based on this data, optimizing program, read, and erase operations by adjusting voltages and timing in incremental steps.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If multilayered memory structure is used to increase storage capacity, then storage density is improved, but threshold voltage distribution becomes wider and cell characteristics vary between layers

Engineering Contradiction:
Improvestorage capacityVSAvoidthreshold voltage distribution
Core Design Contradiction:
Quantity of substanceVSManufacturing precision

Solution Approach 1:

The memory cell array is divided into multiple block groups, with each block group containing blocks from different layers. This segmentation allows independent optimization of operating parameters for each block group, enabling precise control of threshold voltage distribution while maintaining high storage density through the multilayer structure.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Different operating conditions (voltages, timing parameters) are applied to different block groups based on their specific layer composition and characteristics. This local quality approach ensures that each block group operates under optimal conditions for its specific configuration, maintaining narrow threshold voltage distribution across all layers simultaneously.

Inventive Principle:
Principle #3Local quality

2Manufacturing precision

If adaptive control scheme is implemented to optimize operating conditions, then threshold voltage distribution is maintained, but device complexity increases

Engineering Contradiction:
Improvethreshold voltage distribution controlVSAvoidcontrol mechanism complexity
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

A lookup matrix is pre-populated with optimal operating conditions for different block group configurations during manufacturing or initialization. During operation, the appropriate parameters are quickly retrieved from this pre-computed table, avoiding complex real-time calculations and reducing control logic complexity while maintaining precise threshold voltage control.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The lookup matrix acts as an intermediary between the control logic and the memory arrays, storing pre-determined optimal parameters. This intermediary structure simplifies the control logic by replacing complex real-time optimization algorithms with straightforward table lookups, reducing device complexity while maintaining control precision.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP2144251B1Multilayered nonvolatile memory with adaptive control
Publication Date: 2016.08.31 SAMSUNG ELECTRONICS CO LTD
  • EP2144251B1 patent drawingFigure 1
  • EP2144251B1 patent drawingFigure 2
  • EP2144251B1 patent drawingFigure 3

AI summary

A method and device for adaptive control of multilayered nonvolatile semiconductor memory are provided, the device including memory cells organized into groups and a control circuit having a look-up matrix for providing control parameters for each of the groups, where characteristics of each group are stored in the look-up matrix, and the control parameters for each group are responsive to the stored characteristics for that group; the method including organizing memory cells into groups, storing characteristics for each group in a look-up matrix, providing control parameters for each of the groups, where the control parameters for each group are responsive to its stored characteristics, and driving each memory cell in accordance with its provided control parameters.