Disk Read Error Correction Using Majority Voting and ECC
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data read operations in disk drives often introduce errors due to surface abnormalities and obstructions, limiting the effectiveness of error correction coding (ECC) as the error zone size is large, making it difficult to correct soft errors.
Innovation Solution
The method involves obtaining multiple data sets through repeated read operations, tallying logic states at each position, and deriving a new data set based on the majority logic state to reduce the error zone size, allowing ECC processors to correct remaining errors effectively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If data is read once from the disk, then the read operation is fast, but errors are more likely to occur due to surface abnormalities and obstructions
Solution Approach 1:
The system performs preliminary read operations to obtain multiple data sets before final error correction is applied. By reading the data multiple times in advance and storing these data sets, the system prepares multiple versions of the data that can be compared and processed to correct errors, thus improving reliability without sacrificing the speed of the final data retrieval operation.
2Reliability
If multiple read operations are performed to correct soft errors, then data accuracy improves, but the time required increases
Solution Approach 1:
Multiple read operations are performed in advance to obtain constituent data sets, which are then stored for later use. This preliminary action allows the system to prepare corrected data versions beforehand, so that when data retrieval is needed, the pre-processed data can be used immediately, reducing the time penalty of multiple reads.
Solution Approach 2:
The system creates multiple copies of the data through repeated read operations, storing these as separate data sets in memory. These copies serve as redundant information that can be compared and processed to identify and correct errors, with the corrected result being a single derived data set that replaces the need for repeated reads during actual data retrieval.
3Measurement precision
If the error zone size is large, then more errors can be detected, but ECC correction capability is exceeded
Solution Approach 1:
The system extracts only the essential information from multiple data sets by comparing corresponding positions and identifying the majority logic state. This extraction process isolates the correct data values from the noise of errors, creating a derived data set that contains only the corrected information, thereby reducing the error zone size to a level that ECC can effectively handle.
Solution Approach 2:
The system applies local quality by processing each bit position independently through majority logic voting. Each position in the data set is evaluated separately, and the correct value is determined by the majority state at that specific position. This localized approach allows error correction to be applied precisely where needed, rather than treating the entire data set as a single error zone.
4Reliability
If repeated read operations are performed, then soft errors can be identified, but the complexity of the system increases
Solution Approach 1:
The system uses simple copying of data through repeated reads, storing multiple data sets in memory without requiring complex hardware modifications. The complexity is managed by using standard memory storage and comparison logic, rather than requiring specialized error correction hardware, thus maintaining relatively simple system architecture while achieving soft error correction.
Data Source
AI summary
An apparatus and method are disclosed for correcting errors in data obtained from read operations on a storage medium. Errors that occur in a minority of read operations for the data are corrected by a voting technique. The data may then be processed with error correcting code to correct errors that occur in a majority of read operations.