Adaptive Memory Read and Write Systems for Multi-Level Cells
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Multi-level memory cells, such as flash memory cells, face challenges in precise programming and data retrieval due to Gaussian distribution of charge levels, leading to errors in read and write operations, especially after cycling, which requires adaptive adjustment of detection thresholds and mean values to mitigate retention loss.
Innovation Solution
An adaptive memory read and write system that designates pilot memory cells to estimate mean and standard deviation values, computes optimal or near optimal detection threshold values, and uses these values in look-up tables to facilitate accurate reading and writing, minimizing errors caused by degradation over cycling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If multi-level memory cells store varying amounts of charge to indicate multiple logic levels, then storage density increases, but programming precision deteriorates due to Gaussian distribution of charge levels
Solution Approach 1:
The system dynamically adjusts detection thresholds and programming means values based on estimated distribution parameters (mean and standard deviation) of charge levels. By changing these parameters adaptively, the system compensates for Gaussian distribution variations and maintains programming precision despite inherent physical variations in charge storage.
Solution Approach 2:
The system uses pilot memory cells to estimate the actual mean and standard deviation of charge level distributions, then feeds this information back to adjust detection thresholds and programming parameters. This feedback mechanism enables the system to adapt to actual hardware variations and maintain precision in multi-level storage operations.
2Ease of operation
If detection thresholds are fixed for reading multi-level memory cells, then read operation simplicity is maintained, but read accuracy deteriorates due to distribution changes after cycling
Solution Approach 1:
The detection thresholds are transformed from fixed static values to dynamic adaptive values that change based on estimated distribution parameters. The system continuously updates thresholds using pilot cells to track distribution shifts after cycling, enabling accurate reads while maintaining operational simplicity through automated adaptation.
Solution Approach 2:
The system performs preliminary estimation of distribution parameters using pilot memory cells before actual data reading. By pre-characterizing the current state of the memory cell distribution, the system prepares accurate detection thresholds in advance, ensuring high read accuracy without adding complexity to the actual read operation.
3Reliability
If the system tracks and adapts to distribution changes in memory cells, then reliability improves, but device complexity increases due to additional estimation and computation blocks
Solution Approach 1:
The system uses pilot memory cells as copies of the actual data storage cells to estimate distribution parameters. These pilot cells replicate the same physical characteristics and cycling history, allowing the system to infer the state of data cells without directly measuring them, thus achieving reliable adaptation with minimal additional hardware complexity.
Solution Approach 2:
Pilot memory cells serve as intermediaries between the physical memory cells and the detection/programming system. They provide a manageable interface for estimating distribution parameters, enabling the complex adaptation task to be performed through a simplified intermediary structure rather than direct complex measurement of all data cells.
4Measurement precision
If multiple detection thresholds are used for multi-level memory cells, then read precision improves, but computational complexity increases for threshold determination
Solution Approach 1:
The system determines detection thresholds as functions of the estimated mean and standard deviation parameters rather than using fixed complex lookup tables. By expressing thresholds in terms of these two key parameters, the system achieves high read precision while reducing computational complexity through simpler parameter-based calculations.
Data Source
AI summary
An apparatus including: a plurality of multi-level memory cells configured to store data, wherein one or more of the multi-level memory cells are designated as pilot memory cells, and wherein each pilot memory cell is configured to store known, pre-determined data; an estimation block configured to, based on the known, pre-determined data, determine (i) estimated mean values of level distributions of the multi-level memory cells and (ii) estimated standard deviation values of level distributions of the multi-level memory cells; and a computation block configured to compute at least optimal or near optimal detection threshold values of level distributions of the multi-level memory cells based, at least in part, on (i) the estimated mean values and (ii) the estimated standard deviation values, wherein the optimal or near optimal detection threshold values are to be used in order to facilitate reading of the data stored in the multi-level memory cells.