Conditional Drift Cancellation in Memory Cell Programming
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing methods for programming memory cells consume significant energy due to the necessity of applying drift cancellation pulses, which involve polarity transitions, and there is a need to reduce energy consumption while ensuring accurate data storage.
Innovation Solution
A predictive model is used to determine whether the drift cancellation operation can be skipped for memory cells based on attributes such as electrical distance, time since the last write operation, and location, allowing for selective omission of drift cancellation pulses in cases where it does not affect the programming process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If drift cancellation pulses are applied to memory cells during programming, then data storage accuracy is improved, but energy consumption increases
Solution Approach 1:
The patent changes the parameter of drift cancellation operation from mandatory to conditional by introducing a predictive model that evaluates drift risk based on memory cell attributes. This parameter change allows the system to adjust the drift cancellation application status dynamically, skipping operations when drift risk is low and applying them when drift risk is high, thus resolving the contradiction between data storage accuracy and energy consumption
Solution Approach 2:
The predictive model enables the memory system to self-evaluate and self-regulate the drift cancellation operation needs based on intrinsic attributes of memory cells such as electrical distance, time since last write, and location. This self-service mechanism allows the system to autonomously determine when drift cancellation is necessary without external intervention, optimizing the balance between reliability and energy consumption
2Reliability
If drift cancellation operations are performed for all memory cells, then programming reliability is improved, but processing time increases
Solution Approach 1:
The patent transforms the processing parameter from uniform drift cancellation application to selective application based on predicted drift risk. By changing the operation status parameter conditionally, the system reduces unnecessary processing time for memory cells with low drift risk while maintaining programming reliability for cells with high drift risk
Solution Approach 2:
Instead of applying drift cancellation to all memory cells (excessive action), the patent applies drift cancellation only to memory cells where it is predicted to be necessary (partial action). This partial action approach maintains programming reliability for critical cells while avoiding time consumption on cells where drift cancellation would be redundant
Data Source
AI summary
A memory device having memory cells, voltage drivers, and a controller configured to determine, based on an attribute of a memory cell, whether to apply a drift cancellation pulse that is in the opposite polarity of a programming pulse configured to place the memory cell in a state to represent a bit of data. If the drift in the state of the memory cell from a previous programming operation to write data into the memory cell is predicted to be insufficient to prevent the selection of the memory cell during the application of the programming pulse, the drift cancellation pulse is skipped. Otherwise, the drift cancellation pulse is applied in the opposite polarity of the programming pulse.