Seek Control Apparatus Optimizing Acceleration Profiles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional magnetic disk drives face challenges in short seek control due to modeling errors and mechanical resonance excitation, leading to degraded seek response and write errors, especially when traversing longer distances.
Innovation Solution
A seek control apparatus and method that utilize a head seek controller, position detecting unit, error detecting unit, and feedback controller to generate optimized acceleration data, minimizing the sum of squares of acceleration change rates and reducing mechanical resonance excitation by calculating acceleration inputs that slow down the head's response during settling, using a constrained least-square method and Linear Matrix Inequality (LMI) optimization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If deadbeat control (bang-bang control) is used in the feed-forward input to reduce seek period, then seek time is reduced, but modeling errors and mechanical resonance excitation cause seek response degradation and write errors
Solution Approach 1:
The patent changes the acceleration input parameters from conventional deadbeat control to a time-varying acceleration profile that minimizes the sum of squares of acceleration change rates. This parameter transformation reduces mechanical resonance excitation while maintaining fast seek performance, resolving the contradiction between speed and accuracy.
Solution Approach 2:
The patent implements feedback control by detecting the actual head position, calculating position errors against the target trajectory, and correcting the control commands. This feedback mechanism compensates for modeling errors and prevents write errors during settling, addressing the reliability issue while maintaining reduced seek period.
2Productivity
If the number of traversed tracks is increased to reduce seek period, then productivity is improved, but modeling errors and mechanical resonance excitation become more serious
Solution Approach 1:
The patent transforms the acceleration input from conventional bang-bang control to a optimized time-varying profile that minimizes acceleration change rates. This parameter change enables the system to handle longer seek distances without excessive resonance excitation, allowing increased traversed tracks while maintaining reliability.
Solution Approach 2:
The patent makes the acceleration input dynamic by using a time-varying acceleration profile rather than static bang-bang control. This dynamic approach adapts the control input to the specific seek distance and timing, reducing resonance effects across a wider range of traversed tracks while maintaining high productivity.
3Device complexity
If conventional feed-forward control with simplified VCM model is used, then device complexity is reduced, but modeling errors degrade seek response
Solution Approach 1:
The patent changes the control parameters to minimize the sum of squares of acceleration change rates, which inherently accounts for mechanical resonance effects without requiring a complex model. This parameter optimization achieves high reliability while keeping the control system relatively simple.
Solution Approach 2:
The patent adds feedback control to compensate for the simplified model's inaccuracies. By detecting position errors and correcting control commands in real-time, the system achieves accurate seek response despite using a simplified VCM model, resolving the contradiction between simplicity and reliability.
Data Source
AI summary
There is provided a method for generating acceleration data used in feed-forward control of head seeking, wherein, a seek distance and one of a seek period and the count of samples are received; a weighting factor for the change rate of acceleration between adjacent sample times is determined respectively; and an optimum acceleration at each sample time is obtained by performing an optimization calculation that minimizes an objective function under a constraint condition based on variables representing accelerations at each sample time, either the seek period or the count of samples, and the seek distance.