Memory Cell Read Abnormality Detection via Periodic Action
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing memory testing methods struggle to accurately detect abnormalities, such as leakage or poor contact, which can affect the service life and reliability of memory products. These abnormalities often go undetected under normal reading and writing conditions.
Innovation Solution
A method and device for testing memory that involves performing at least two consecutive read operations on a memory cell after activating a word line. The output signal from these operations is then analyzed to determine if a read abnormality exists.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If normal reading and writing conditions are used for memory testing, then the test process is simple and fast, but abnormalities such as leakage or poor contact cannot be detected
Solution Approach 1:
The patent applies periodic action by performing multiple consecutive read operations on the same memory cell after activating a word line. Instead of a single read operation, the system repeatedly reads the memory cell state, which allows abnormalities such as leakage or poor contact to manifest over time. This periodic testing approach increases detection accuracy while maintaining a relatively simple test process structure.
2Measurement precision
If multiple consecutive read operations are performed, then detection accuracy improves, but testing time increases
Solution Approach 1:
The patent applies partial or excessive action by performing a limited number of consecutive read operations (more than one but not excessively many) on memory cells that are suspected or need verification. This approach achieves sufficient detection precision for identifying abnormalities like leakage or poor contact, while controlling the testing time to be acceptable. The method balances the need for accurate detection with the constraint of testing efficiency.
Data Source
AI summary
A method and device for testing a memory are provided. The method includes the following operations. After activating at least one word line, at least two times of read operations are performed on a to-be-tested memory cell connected to the activated word line. Whether there is a read abnormality in the to-be-tested memory cell is determined according to an output signal obtained after the at least two times of read operations.