Hard Disk Drive Fly Height Estimation Using Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current hard disk drive (HDD) systems face challenges in accurately estimating and controlling the fly height of recording heads in real-time, leading to potential write failures and damage due to open-loop control methods that are not sensitive to instantaneous changes in head-media spacing.
Innovation Solution
A machine learning-based system, utilizing a neural network, is employed to estimate fly height by processing multiple parameters such as servo address mark data, temperature readings, and radial position, enabling real-time adjustments and closed-loop control of the recording head's clearance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If open-loop control methods are used for fly height control, then device complexity is reduced, but measurement precision and reliability of fly height estimation deteriorate
Solution Approach 1:
The patent implements closed-loop control by continuously monitoring multiple parameters (servo address mark data, temperature readings, radial position) and using machine learning models to estimate fly height in real-time. The system feeds this estimated information back to adjust head positioning, creating a feedback loop that significantly improves measurement precision and reliability compared to open-loop methods.
Solution Approach 2:
The patent replaces traditional mechanical measurement systems with machine learning-based estimation. Instead of relying on direct mechanical sensors to measure fly height, the system uses ML models that process multiple indirect parameters (servo data, temperature, radial position) to estimate fly height, achieving higher precision without additional mechanical complexity.
2Ease of operation
If traditional control methods are used, then ease of operation is maintained, but productivity and reliability deteriorate due to inability to detect instantaneous changes
Solution Approach 1:
The patent transitions from static, predetermined control settings to dynamic, real-time adjustments. The machine learning model continuously processes incoming sensor data and adjusts fly height estimates based on instantaneous changes in operating conditions, enabling the system to adapt dynamically to varying conditions while maintaining operational simplicity through automated control.
3Device complexity
If open-loop control is used, then device complexity is lower, but reliability deteriorates due to lack of sensitivity to instantaneous changes
Solution Approach 1:
The closed-loop control system continuously monitors multiple parameters and adjusts fly height in real-time based on feedback from the machine learning model. This feedback mechanism detects instantaneous changes in head-media spacing and corrects them before they cause write failures, significantly improving write operation reliability compared to open-loop control.
Solution Approach 2:
The system performs preliminary estimation of fly height using machine learning models before actual write operations occur. By predicting potential fly height issues in advance based on current sensor readings, the system can take preventive actions to maintain reliable write operations, rather than reacting after failures occur.
Data Source
AI summary
Two or more data values are received from one or more sensors of a hard disk drive. The two or more data values are indicative of a fly height of a recording head of the hard disk drive. The two or more data values are input into a machine-learning processor during operation of the hard disk drive. A fly height of the recording head during the operation of the hard drive head is adjusted based on an output of the machine learning processor.