Non-Volatile Memory Error Correction After Heat Events
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Non-volatile memory devices, particularly phase change memory (PCM) devices, face reliability issues with information retention due to thermal cycles during manufacturing, leading to elevated bit-error rates and signal corruption during heat events like packaging or soldering.
Innovation Solution
Implementing error detection and correction techniques within the memory device to identify potential heat events and correct errors in the memory array, using components like heat event detection and error correction codes, such as LDPC and BCH coding, to reprogram corrected information and maintain data integrity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If error detection and correction techniques are implemented in the memory device, then data integrity is maintained and bit-error rates are reduced, but device complexity increases
Solution Approach 1:
The patent applies preliminary action by pre-programming known signal patterns into the memory array during manufacturing before the device is shipped. These known patterns serve as reference data that can be later compared against actual stored data to detect errors without requiring complex real-time correction algorithms. The controller is pre-configured with the ability to perform simple comparisons and identify deviations from expected patterns.
Solution Approach 2:
The patent introduces an intermediary approach by using known signal patterns as a mediator between the memory array and the controller. Instead of implementing complex error correction codes that require sophisticated algorithms, the known patterns act as a reference intermediary that simplifies the error detection process. The controller compares actual data against these pre-stored reference patterns to identify errors caused by heat events.
2Reliability
If the memory array is re-programmed with corrected information, then error rates are reduced, but loss of time occurs during the re-programming process
Solution Approach 1:
The patent applies preliminary action by pre-programming known signal patterns into the memory array during manufacturing. These known patterns are stored in advance and serve as reference data for error detection. When errors are detected, the system can quickly compare actual data against these pre-prepared reference patterns and perform targeted re-programming only of affected regions, rather than requiring comprehensive error correction procedures.
Solution Approach 2:
The patent implements local quality by enabling error detection and re-programming operations to be performed on specific regions of the memory array rather than requiring full-array operations. When a heat event is detected, the controller can identify and re-program only the specific memory cells or blocks that are affected, significantly reducing the time loss compared to re-programming the entire array.
3Reliability
If error detection operations are performed on the memory array, then bit-error rates are reduced, but productivity decreases due to additional processing steps
Solution Approach 1:
The patent applies preliminary action by pre-programming known signal patterns into the memory array during manufacturing. These reference patterns are prepared in advance, allowing the controller to perform rapid comparison-based error detection without requiring complex real-time analysis. This preliminary preparation significantly reduces the processing time required for error detection operations.
Solution Approach 2:
The patent implements partial action by performing error detection operations selectively rather than continuously. Error detection is triggered by specific events such as heat events or power-on conditions, rather than being performed on every memory access. This selective approach maintains high productivity while still providing necessary error detection when most critical.
Data Source
AI summary
The present disclosure includes methods, devices, and systems for error detection or correction of stored signals in memory devices. An example method includes determining whether to perform error correction operations on contents of a non-volatile memory array. Determining whether to correct can include determining whether a level of errors in pre-programmed signals in the non-volatile memory array exceeds a bit error rate threshold, where the pre-programmed signals are different from the contents of the non-volatile memory array, and performing error correction on the contents of the non-volatile memory array if the level of errors exceeds the bit error rate threshold.