Soft-Input HDD Decoding for Defect Margin Assessment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Magnetic recording systems face challenges in determining the quality and margin of hard disk drives (HDDs) over time due to defects like amplitude dropout and permanently polarized media, as existing error correction codes, such as bounded distance decoding and LDPC codes, struggle to accurately assess the remaining error correction capability and detect defects effectively.
Innovation Solution
A technique that utilizes soft-information decoders to determine the presence of defects by analyzing log-likelihood ratios (LLRs) and identifying sectors with incorrect, large-magnitude LLRs, providing a figure-of-merit to assess the disk's quality and margin, and a process to determine if a portion of the storage device is defective by writing known data and comparing read-back LLRs with configurable thresholds.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If bounded distance decoding or LDPC codes are used for error correction, then error correction capability is improved, but the ability to accurately assess remaining error correction capability and detect defects is worsened
Solution Approach 1:
The patent introduces log-likelihood ratios (LLRs) as an intermediary metric to assess the reliability of decoded data. LLRs provide a continuous measure of confidence in each bit decision, enabling the system to evaluate remaining error correction capability without relying on discrete error counts. This intermediary measurement allows for more precise defect detection and margin assessment while maintaining the benefits of soft-decision decoding.
Solution Approach 2:
The patent changes the assessment parameter from discrete error counts to continuous LLR magnitude analysis. By examining the distribution and magnitude of LLRs rather than simply counting corrected errors, the system can more accurately assess remaining error correction capability. This parameter change enables detection of degradation trends before actual errors occur, improving measurement precision while maintaining reliability.
2Reliability
If product qualification testing is performed beyond intended use, then product quality assurance is improved, but test time and complexity increase
Solution Approach 1:
The patent performs preliminary defect detection by analyzing LLRs during normal operation before failures occur. Instead of waiting for products to fail under extreme stress testing, the system proactively identifies defective sectors by detecting abnormal LLR patterns during regular use. This preliminary action reduces the need for extensive qualification testing while maintaining high quality assurance standards.
Solution Approach 2:
The system uses its own operational data (LLRs from normal read/write operations) to perform self-diagnosis and defect detection. Rather than requiring external stress testing, the product monitors its own performance metrics and identifies degradation autonomously during normal operation, significantly reducing qualification testing time while maintaining reliability.
3Measurement precision
If soft-information decoders are used to analyze LLRs for defect detection, then defect detection accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent extracts only the necessary information from soft-decision decoding for defect detection purposes - specifically the magnitude and distribution of LLRs - without implementing the full complexity of soft-information decoders. By taking out only the essential LLR analysis components needed for defect detection rather than complete soft-decision decoding functionality, the system achieves improved defect detection accuracy while limiting the increase in computational complexity to only what is necessary for the specific task.
Data Source
AI summary
Determining a parameter associated with whether a portion of a storage device is defective is disclosed. Determining comprises: obtaining known data associated with the portion; reading back from the portion to produce a read-back waveform; decoding the read-back waveform, including producing statistical information; and determining a parameter associated with whether the portion is defective based at least in part on the statistical information.