NAND Flash Temperature-Assisted Voltage Sweep
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing NAND flash devices face challenges in efficiently managing temperature changes, leading to increased bit error rates and reduced device lifetime due to the need for brute force voltage sweeps, which are time-consuming and may fail to converge to optimal reference voltages.
Innovation Solution
A temperature-assisted hardware technique that actively monitors and models on-die temperature using a multivariable regression model to predict real-time temperature and adjust threshold voltages, allowing for a local sweep of reference voltages to achieve a minimum read bit error rate without sacrificing data bus bandwidth.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a brute force voltage sweep approach is used to find optimal reference voltage, then the solution is simple to implement, but the time consumption increases significantly (5 to 20 read retry iterations)
Solution Approach 1:
The patent applies preliminary action by using a temperature prediction model to pre-determine the optimal reference voltage before actual read operations. The controller predicts the threshold voltage shift based on temperature changes and proactively adjusts the reference voltage, eliminating the need for time-consuming brute force voltage sweeps during read retries.
Solution Approach 2:
The patent replaces the mechanical brute force voltage sweep method with a computational temperature prediction model. Instead of incrementally sweeping voltages to find the optimal point, the system uses a model to calculate the predicted threshold voltage shift based on temperature, substituting iterative mechanical adjustment with a direct computational approach.
2Measurement precision
If the controller reads out the NAND temperature sensor, then the temperature information is obtained, but the data bus usage is limited during read-out period, leading to longer latency and reduced bandwidth
Solution Approach 1:
The patent introduces an intermediary temperature prediction model that translates temperature sensor readings into threshold voltage shift predictions. This model acts as a mediator between the temperature measurement and the reference voltage adjustment, allowing the system to derive actionable information without requiring continuous direct controller intervention on the data bus.
Solution Approach 2:
The system implements self-service by enabling the controller to autonomously predict threshold voltage shifts based on temperature data and automatically adjust reference voltages without requiring extensive data bus transactions. The temperature prediction model allows the system to self-correct for temperature-induced voltage shifts, reducing the need for frequent controller-mediated adjustments.
3Ease of operation
If incremental voltage sweep is used when threshold voltage change is significant, then the approach is simple, but it can be in the wrong direction and fail to converge to a solution
Solution Approach 1:
The patent implements feedback by using temperature sensor readings as input to the prediction model, which then adjusts the reference voltage based on the predicted threshold voltage shift. This closed-loop feedback mechanism ensures that voltage adjustments are directionally correct and proportional to the actual temperature-induced changes, preventing divergence that could occur with simple incremental sweeps.
Solution Approach 2:
The patent applies parameter changes by dynamically adjusting the reference voltage based on temperature predictions rather than using fixed incremental steps. The system changes the voltage parameter proportionally to the predicted threshold shift, ensuring that adjustments remain effective across different temperature conditions and prevent failure to converge.
Data Source
AI summary
Systems, apparatus and methods are provided for temperature assisted non-volatile storage device management in a non-volatile storage system. In one embodiment, a non-volatile storage system may comprise a temperature sensor, a non-volatile storage device and a processor. The processor may be configured to obtain a read-out from the temperature sensor, generate a predicted real-time on-die temperature for the non-volatile storage device based on the read-out, generate an estimated threshold voltage for reading data stored in the non-volatile storage device based on the predicted real-time on-die temperature and conduct a local sweep of a reference voltage using the estimated threshold voltage as a starting point to obtain a final read reference voltage with a minimum read bit error rate.