Memory Die Temperature Compensation Across Program-Erase Cycles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional memory sub-systems face challenges in adequately compensating for voltage shifts due to temperature variations across program erase cycles, leading to increased raw bit error rates beyond the error correction capability.
Innovation Solution
The memory sub-system generates and updates a fitting equation to determine temperature compensation values based on program erase cycles, reducing the size of the lookup table and accounting for die-to-die variations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a conventional lookup table is used for temperature compensation, then temperature compensation can be provided, but the memory footprint is large and cannot adequately account for die-to-die variations across program erase cycles
Solution Approach 1:
The patent segments the temperature compensation data into two parts: (1) a small set of representative temperature compensation values stored in a compact lookup table, and (2) measured temperature compensation values from actual dies that are used to generate fitting parameters. This segmentation allows the system to store only essential reference data while deriving specific die compensation characteristics through mathematical fitting, thereby reducing memory footprint while maintaining compensation accuracy across program erase cycles.
Solution Approach 2:
The patent transforms the temperature compensation approach from storing complete lookup tables for each die to storing fitting parameters (such as coefficients of a polynomial function) that can generate temperature compensation values dynamically. This parameter change reduces the memory footprint from storing extensive lookup table data to storing compact fitting parameters, while still enabling accurate temperature compensation adaptation for each die across different program erase cycles.
2Reliability
If temperature compensation values are updated for each die, then die-to-die variations are accounted for, but the device complexity increases
Solution Approach 1:
The patent implements a self-service mechanism where each die's temperature compensation characteristics are automatically measured and used to generate fitting parameters during initial operation or calibration phases. The system autonomously captures temperature compensation values from actual dies, performs fitting operations to derive parameters, and stores these parameters for ongoing use. This self-service approach eliminates the need for manual configuration or complex external intervention, reducing operational complexity while maintaining accurate die-to-die variation compensation.
Solution Approach 2:
The patent performs preliminary measurement and fitting operations to establish temperature compensation fitting parameters before normal operation begins. During initial calibration, the system measures temperature compensation values from each die, generates fitting parameters through regression analysis, and stores these parameters for subsequent use. This preliminary action ensures that accurate die-specific compensation is ready in advance, eliminating the need for complex real-time measurement and adjustment during normal operation, thereby reducing operational complexity.
Data Source
AI summary
A first analysis of each respective die of a multi-die memory device is performed. An equation to determine a respective temperature compensation (tempco) value for each respective die based on a number of program erase cycles (PECs) of the respective die based on the first analysis s determined. The equation for use in processing memory access requests directed to the respective die is stored. Whether to update the equation directed to the respective die based on a second analysis of the respective die is determined.