Memory Failure Reporting with Baseline-Adjusted Error Thresholds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing memory devices inaccurately report failures due to default fail thresholds that do not account for baseline errors, leading to premature failure reporting and resource wastage.
Innovation Solution
Generate an offset fail threshold by summing a default threshold with an initial error count from an ECS procedure to trigger failure reporting, using customized error ranges to control when failures are reported.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a default fail threshold is used for failure reporting, then the failure reporting mechanism is simple to implement, but it leads to inaccurate failure reporting due to baseline errors
Solution Approach 1:
The patent performs preliminary error characterization during manufacturing by executing ECS procedures to establish a baseline error count. This preliminary action captures the inherent error rate of each memory device before it is deployed, allowing the system to later differentiate between baseline errors and actual failures. The baseline error count is stored and used to dynamically adjust the fail threshold, thereby improving failure reporting accuracy without adding complex real-time analysis.
Solution Approach 2:
The patent dynamically changes the fail threshold parameter based on the baseline error count specific to each memory device. Instead of using a fixed default threshold, the system calculates a customized threshold by adding the baseline error count to the default threshold. This parameter adjustment allows each device to have an optimized threshold that accounts for its unique error characteristics, improving measurement precision while maintaining relatively simple implementation through straightforward arithmetic operations.
2Reliability
If a default fail threshold is used, then resource consumption is low, but premature failure reporting occurs leading to resource wastage
Solution Approach 1:
The system performs preliminary error characterization during manufacturing to establish the baseline error count for each device. This advance preparation allows the system to set an accurate fail threshold that prevents premature failure reporting, thereby avoiding unnecessary resource consumption associated with false failure reports while ensuring reliable detection of actual failures.
Solution Approach 2:
The patent implements a feedback mechanism where the baseline error count obtained from preliminary ECS procedures is used to adjust the fail threshold. This feedback loop ensures that the threshold is optimized for each specific device, preventing both premature failure reporting and missed detections, thereby improving overall reliability while minimizing wasted resources on false positives.
3Measurement precision
If baseline errors are not accounted for, then the error counting process is simple, but failure detection precision deteriorates
Solution Approach 1:
The patent performs the error characterization action preliminarily during the manufacturing process rather than during normal operation. By executing ECS procedures and establishing baseline error counts in advance, the system captures the inherent error rate of each device without adding time overhead to operational ECS procedures. The baseline data is stored and reused, allowing rapid and precise error measurement during device lifetime without repeated time-consuming characterization.
Data Source
AI summary
Methods, systems, and apparatuses related to detecting and reporting failures for a memory device are described. When a count of bit-flip errors is above a fail threshold, a memory device can report a failure. Failure reports can indicate a rate at which the memory device is accumulating errors. An offset fail threshold may be applied instead of a default fail threshold, such as a standardized or specified threshold. The offset fail threshold can be a summation of the default fail threshold and an offset determined from an initial error count determined before the memory device has accumulated errors from use.