Memory Read Voltage Search for Low-Latency Data Recovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing memory systems face inefficiencies in searching for an optimized read voltage for memory cells due to shifts in threshold voltages caused by factors like charge loss and temperature variations, leading to long latency and inefficiencies in data recovery through multiple read retries and calibration.
Innovation Solution
A memory sub-system efficiently searches for an optimized read voltage by measuring signal and noise characteristics of memory cells within a test voltage range, estimating the next range based on these characteristics, and calculating the voltage using count differences to minimize latency and improve data recovery.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple read retries and calibration operations are performed to recover data from memory cells with shifted threshold voltages, then data recovery reliability is improved, but latency increases significantly
Solution Approach 1:
The system performs preliminary calibration operations during idle periods or before data storage to establish baseline threshold voltage characteristics. This preliminary action allows the system to have reference data ready, reducing the need for extensive calibration during data recovery operations and thereby reducing latency while maintaining reliability
Solution Approach 2:
The system implements feedback mechanisms where read results from initial attempts are analyzed to dynamically adjust subsequent read voltage levels. This feedback loop enables the system to converge on optimal read voltages more quickly, reducing the number of retry attempts needed and thereby reducing latency while maintaining high data recovery reliability
2Measurement precision
If traditional calibration methods are used to determine read voltage, then measurement accuracy is improved, but the complexity of the calibration process increases
Solution Approach 1:
The calibration process is segmented into multiple stages: initial rough calibration to establish approximate voltage levels, followed by fine-tuning stages that progressively refine the read voltage. This segmentation allows the system to achieve high measurement accuracy through incremental improvements rather than requiring a single complex calibration operation, thereby reducing overall process complexity
Solution Approach 2:
The system changes calibration parameters dynamically based on observed memory cell characteristics. Instead of using fixed calibration sequences, the system adjusts voltage steps, measurement points, and iteration counts based on real-time feedback from memory cell responses. This adaptive parameter changing simplifies the calibration process by focusing measurements only where needed, reducing complexity while maintaining high accuracy
Data Source
AI summary
A memory device to search for a voltage optimized to read a group of memory cells. In response to a read command, the memory device measures first signal and noise characteristics of the memory cells by reading the memory cells at first test voltages. Based on the first signal and noise characteristics, the memory device may determine that the optimized read voltage is outside of a range of the first test voltages. In response, the memory device determines, based on the first signal and noise characteristics, an estimate of the optimized read voltage, and measures second signal and noise characteristics by reading at second test voltages configured around the estimate. The optimized read voltage can be computed based at least in part on the second signal and noise characteristics. The memory device retrieves data from the memory cells using the optimized read voltage.