Flash Memory Test Apparatus Error Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional semiconductor memory test apparatuses incorrectly identify flash memory failures due to non-permanent software errors that can be corrected by error correcting codes, and fail to classify flash memories based on their error-prone characteristics, affecting their use and pricing.
Innovation Solution
A test apparatus with a logical comparator, data error counting section, and classifying section that compares data from a memory-under-test with expected values, counts errors, and classifies the memory based on predefined upper limit values to distinguish between correctable and non-correctable errors, thereby accurately assessing memory quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the test apparatus detects failure when any data inconsistency occurs, then the detection sensitivity is improved, but false failure detection increases due to correctable errors
Solution Approach 1:
The patent changes the detection parameter from binary pass/fail to quantitative error counting. By counting the number of inconsistent data points and comparing against a threshold, the system transitions from sensitive but inaccurate detection to both sensitive and accurate detection, filtering out false failures caused by correctable errors.
Solution Approach 2:
The patent introduces an intermediary error counting mechanism between the data comparison and failure detection. This intermediary component aggregates error information across multiple data points and uses a threshold criterion to determine whether to report failure, thereby mediating between raw detection signals and final failure conclusions.
2Measurement precision
If the test apparatus classifies memory based on error characteristics, then the memory quality assessment is improved, but the device complexity increases
Solution Approach 1:
The patent segments the memory classification task into distinct functional modules: data comparison unit, error counting unit, threshold comparison unit, and classification unit. Each module performs a specific function in the quality assessment process, making the overall complex task manageable and systematic while improving measurement precision.
Solution Approach 2:
The patent changes the classification approach from complex multi-parameter analysis to a simplified threshold-based system. By using a single error count threshold parameter, the system achieves effective memory quality classification without requiring complex multi-dimensional assessment criteria.
Data Source
AI summary
There is provided a test apparatus for testing a memory-under-test for storing data strings to which an error correcting code has been added, having a logical comparator for comparing each data contained in the data string read out of the memory-under-test with an expected value generated in advance, a data error counting section for counting a number of data inconsistent with the expected value, a plurality of registers, provided corresponding to each of a plurality of classes, for storing an upper limit value of a number of errors contained in the data -under-test to be classified into the class, comparing sections for comparing each of the plurality of upper limit values stored in the plurality of registers with the counted value of the data error counting section and a classifying section for classifying the memory-under-test into the class corresponding to the register storing the upper limit value which is greater than the counted value.