Memory Read Voltage Calibration Using Threshold Distribution Metadata
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing memory device calibration techniques suffer from inaccuracies and high latency due to the inability to effectively track and compensate for shifts in threshold voltage distributions caused by factors such as charge loss, temperature, and physical defects, leading to uncorrectable read data.
Innovation Solution
Memory devices generate metadata characterizing the applied read voltage level with respect to voltage distributions, allowing for iterative calibration of read voltage levels using failed byte and bit counts to adjust read operations, minimizing latency and improving accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing memory device calibration techniques are used, then read operations can be performed, but inaccuracies and high latency occur due to inability to track voltage distribution shifts
Solution Approach 1:
The patent implements feedback by having the memory device generate metadata characterizing the applied read voltage level with respect to voltage distributions, and using this metadata to iteratively adjust read voltage levels. The controller receives the metadata, determines adjustments based on the characterization, and applies corrected voltage levels in subsequent read operations, creating a closed-loop system that continuously improves calibration accuracy while minimizing latency through iterative refinement.
2Reliability
If read voltage levels are not calibrated, then read operations are fast, but voltage shifts cause uncorrectable read data errors
Solution Approach 1:
The patent applies preliminary action by performing read level calibration before actual data read operations. The controller receives metadata characterizing the voltage distributions, determines the appropriate read voltage levels in advance, and stores these calibrated levels for use in subsequent read operations. This preliminary calibration ensures high reliability in data retrieval while maintaining productivity by preparing the system beforehand rather than during data access.
3Measurement precision
If iterative calibration using metadata is implemented, then calibration accuracy improves, but device complexity increases
Solution Approach 1:
The patent implements self-service by having the memory device itself generate the metadata characterizing the applied read voltage level with respect to voltage distributions of target memory cells. The memory device performs self-diagnosis and self-characterization of its voltage states, providing this information to the controller. This self-service approach improves measurement precision while minimizing device complexity by leveraging the memory device's own capabilities rather than requiring external characterization equipment.
Data Source
AI summary
Described are memory devices producing metadata characterizing the applied read voltage level with respect to voltage distributions. An example memory sub-system comprises: a memory device comprising a plurality of memory cells; and a controller coupled to the memory device, the controller to perform operations comprising: performing, using a read voltage level, a read strobe with respect to a subset of the plurality of memory cells; and receiving, from the memory device, one or more metadata values characterizing the read voltage level with respect to threshold voltage distributions of the subset of the plurality of memory cells, wherein the one or more metadata values reflect a conductive state of one or more bitlines connected to the subset of the plurality of memory cells.