Memory Error Scrub Sequencing Before Power-Down Self-Refresh
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increased data input/output speed in semiconductor devices like DDR2, DDR3, and DDR4 leads to higher probabilities of data transfer errors, necessitating reliable error detection and correction methods, particularly during power-down operations to prevent data loss in adjacent memory cells.
Innovation Solution
A semiconductor system performs an error check scrub (ECS) operation followed by a self-refresh operation on all memory cells after power-down entry, correcting errors and ensuring data integrity by alternating read and write operations and applying precharge signals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If data input/output speed is increased to improve operating performance, then productivity is improved, but the probability of data transfer errors increases, worsening reliability
Solution Approach 1:
The patent performs error check scrub operations before power-down entry to proactively detect and correct potential errors in memory cells. By conducting ECS operations on all memory cells before transitioning to low-power mode, the system prevents errors from occurring or propagating, thus maintaining reliability while enabling high-speed operation during active periods
2Reliability
If error check scrub operation is performed on all memory cells to improve reliability, then data integrity is improved, but the time required for the operation increases, worsening loss of time
Solution Approach 1:
The patent implements periodic error check scrub operations that are triggered at specific intervals or under specific conditions (such as before power-down entry). Rather than continuously scrubbing all memory cells, the system performs ECS operations periodically, balancing the need for data integrity with the constraint of operational time. This periodic approach ensures reliability while minimizing time loss by only performing thorough checks when necessary
Solution Approach 2:
The error check scrub operation is performed in advance before power-down entry, ensuring that all memory cells are verified and corrected before the system transitions to a state where error detection would be more difficult. This preliminary action approach allows comprehensive checking to be completed beforehand, preventing time loss during critical operational phases
3Use of energy by moving object
If power-down operation is entered to reduce energy consumption, then use of energy is improved, but the risk of data loss in adjacent memory cells increases, worsening reliability
Solution Approach 1:
The patent performs error check scrub operations and ensures all memory cells are properly charged and verified before the power-down operation begins. By conducting ECS operations and applying precharge signals to all bank groups before entering power-down mode, the system proactively identifies and corrects potential errors, ensuring data integrity is maintained even when the system transitions to low-power state where monitoring is reduced
Solution Approach 2:
The system implements a feedback mechanism where the completion of ECS operations and precharge signals is monitored and confirmed before allowing power-down entry. The controller tracks whether all memory cells have been properly checked and charged, and only permits power-down transition when these conditions are satisfied, thereby preventing data loss while managing power consumption
Data Source
AI summary
A semiconductor system includes a controller configured to output a command and address for performing an ECS operation after the start of entry into a power-down operation, receive data and output the data in response to correcting one or more errors occurring in the data, and output a command for performing a self-refresh operation when the ECS operation is terminated, and a semiconductor device configured to output, as the data, internal data stored in multiple memory cells after the start of a read operation of the ECS operation in response to receiving the command and address, receive the data having the one or more errors corrected after the start of a write operation of the ECS operation, store the data having the one or more errors corrected, and perform a self-refresh operation on the multiple memory cells after receiving the command when the ECS operation is terminated.