Journal of Advanced Mechanical Design, Systems, and Manufacturing

Transcription

1 Controlling Vibration of HDD Actuator by Using Dummy Heads * Noritaka OTAKE**, Keiko WATANABE**, Toshihiko SHIMIZU**, Kenji TOMIDA*** and Toshihiro ARISAKA*** **Hitachi, Ltd. Mechanical Engineering Research Laboratory 832-2, Horiguchi, Hitachinaka, Ibaraki , Japan noritaka.otake.nn@hitachi.com ***Hitachi Global Storage Technologies, Japan, Ltd. 1, Kirihara, Fujisawa, Kanagawa , Japan Abstract Some hard disk drives (HDDs) have fewer than the maximum number of magnetic heads. Such HDDs need to be re-designed to keep their position accuracy. We focused on adding dummy heads instead of re-designing the whole actuator. A dummy head is put on the edge of an arm model, and it influences the vibration characteristics of the model. Because the vibration characteristics of the arm model are related to those of the whole actuator, we optimized the dummy head design by focusing on the vibration characteristics of a local arm model instead of focusing on the whole actuator. We created a dummy head that would enable HDDs to keep the gain of their vibration characteristics around the sampling frequency as low as those of an HDD with the maximum number of mountable magnetic heads. Key words: Actuator, Vibration, Dummy Head 1. Introduction *Received 14 July, 29 (No ) [DOI: /jamdsm.4.82] Copyright 21 by JSME The storage capacity of hard disk drives (HDDs) has increased as the positioning accuracy of their magnetic heads has increased. However, there is still strong demand for HDDs with smaller than the maximum storage capacity because they are more affordable for the average consumer. To meet this demand, we have created HDDs with fewer magnetic heads than found in HDDs with the maximum storage capacity. In this paper, we distinguish between these two types of HDDs; HDD-f will stand for an HDD having the maximum number of magnetic heads and HDD-d will stand for an HDD having fewer magnetic heads. A decrease in the number of magnetic heads causes a change in the actuator dynamics in the HDD, especially in the vibration characteristics of the whole actuator. To keep the positioning accuracy at a high level, we need to re-design the actuator of the HDD-d in the same way that an actuator is designed for the corresponding HDD-f; the usual practice is to re-design the servo and the mechanics of the actuator [1][2]. Occasionally, this means re-designing the actuator dynamics for the sake of adequate positioning accuracy [3] or for shock robustness [4]. Some modifications to the VCM excitation force [5][6] might be also needed. When fewer magnetic heads are used, a dummy head is sometimes put on the edge of an arm or somewhere else so that it acts as a weight [7]. In many cases, designers try to keep the entire mass balance. In particular, they often make a suspension structure that can control the vibration of the whole actuator [8]. In this paper, we describe a dummy head shape that can keep the gain of the HDD-d actuator s vibration characteristics around its sampling frequency as low as those of the HDD-f actuator. We designed an optimal dummy head by analyzing the vibration 82

2 Dummy head z Middle arm y Head gimbals assembly End arm Dummy head x Head gimbals assembly Measurement point (a) HDD-f actuator (b) HDD-d actuator Fig. 1 Example of actuators used in calculating vibration characteristics Gain [db] Mode 1 HDD-f upward HDD-f downward HDD-d upward HDD-d downward -1 Mode Fig. 2 Vibration characteristics of magnetic head position in HDD-f and HDD-d (a) Mode 1 (b) Mode 2 Fig. 3 Actuator modal shapes located between 16. khz and 21. khz characteristics of a local arm model instead of analyzing that of the whole actuator model, based on the relationship between the frequency of the whole actuator model and that of the local arm model. We used the same actuator parts as in an HDD-f, except for replacing some of the magnetic heads with dummy heads. 2. Factor Analysis Figure 1 shows example actuator models of the HDD-f and the HDD-d. These actuators have two middle arms and two end arms; a middle arm can have two head gimbals assembly (HGA) and end arm can have one HGA in HDD-f. The HDD-d actuator has two magnetic heads on the middle arms and four dummy heads on the other arms. The dummy heads are designed so as to keep mass of the whole actuator balanced. The magnetic heads vibration characteristics in the direction of the y-axis were calculated with the finite element method (FEM). The actuator was fixed on both sides of its pivot. Excitation force was input around the coil, and the severity of the force was based on the results of magnetic field analysis (Elf/Magic). Figure 2 shows the vibration characteristics. The dashed line shows the results for the HDD-f actuator, and the solid line shows the results for the HDD-d actuator. Although the HDD actuator has several magnetic heads, Figure 2 shows results for only upward and 83

3 Fixed face Fixed face (a) Arm-A Measurement point Measurement point (b) Arm-B Fig. 4 Analysis setup in Arm-A model and Arm-B model Gain [db] S1 Middle Arm-A End Arm-A Fig. 5 Vibration characteristics of Arm-As and Arm-Bs downward heads. For the sake of convenience, we define that a frequency band with a low gain is such that it has a gain under the criteria line of -8 db. An acceptable range is shown by arrow line. The vibration characteristics of the two actuators become different around 2. khz. The HDD-f actuator has a frequency band with a low gain from 18. khz to 2. khz. The HDD-d actuator has a frequency band with a low gain from 17. khz to 18.5 khz. The sampling frequency range for the HDD-f will be set from 18. khz to 2. khz because the range tends to be set at a higher frequency band with low gain in the HDD-f. But we cannot choose the same range for the HDD-d actuator that has a frequency band with a high gain from 18. khz to 2. khz. Figure 3 illustrates the mechanical resonance modes of HDD-f actuator that we can see between 16. and 21.5 khz. Mode 1 is around 16. khz, and Mode 2 is around 21.5 khz. Both modes are coupled modes involving an arm sway mode and an arm torsion mode, and they have different direction of deformations at the coil holder. In the HDD-d actuator, Mode 1 is around 16. khz, and Mode 2 is around 2. khz. To determine what caused the difference between these two actuators, we examined local models. Figure 4 shows two local models. One is an end arm model with a magnetic head (Arm-A) and the other is an end arm model with a dummy head (Arm-B). In fact, there are four local models; Arm-A as a middle arm, Arm-A as an end arm, Arm-B as a middle arm, Arm-B as an end arm. We calculated the vibration characteristics of the magnetic head and the edge of a dummy head when the root of the arm was fixed and excitation force was applied to the root. Figure 5 shows the vibration characteristics of Arm-A and Arm-B. Each plot has several peaks. Arm-A s results show that the first sway (S1) mode appears around 11. khz, the second torsion (T2) mode appears around 16. khz and the second sway (S2) mode appears around 18. khz. Arm-B s results show that the S1 mode appears around 12.5 khz, the T2 mode appears around 25. khz and the S2 mode appears around 46. khz. We supposed that the differences in the vibration characteristics of the actuator models were due to the different vibration characteristics of the arm models, and tried to control the vibration characteristics of the actuator by modifying the vibration characteristics of the Arm-B model. T2 S2 84

4 Mode1 [khz] T2 [khz] Mode1 [Hz] S2 [khz] (a) T2 mode and Mode 1 (b) S2 mode and Mode 1 Fig. 6 Relationship between arm model mode and Mode 1 Mode2 [khz] T2 [khz] Mode2 [khz] S2 [khz] (a) T2 mode and Mode 2 (b) S2 mode and Mode 2 Fig.7 Relationship between arm model mode and Mode 2 3. Relationship Between Vibration Characteristics of Arm Model with Dummy Head and Those of HDD-d Actuator We investigated the relationship between the modes of the Arm-B model and the modes of an HDD-d actuator model by testing dummy heads with various shapes. Besides varying their length or thickness in their heads, we prepared such a dummy head that has a neck. Figure 6 (a) shows the relationship between the T2 mode and Mode 1, and Fig. 6 (b) shows the relationship between the S2 mode and Mode 1. Figure 7 (a) shows the relationship between the T2 mode and Mode 2, and Fig. 7 (b) shows the relationship between the S2 mode and Mode 2. In these figures, the horizontal axis is the frequency at which the T2 mode or the S2 mode appears. The vertical axis shows the frequency at which Mode 1 or Mode 2 appears. Circles show the results for the end Arm-B model and squares show the results for the middle Arm-B model. Figures 6 (a) and (b) show that Mode 1 appears around 16. khz wherever the T2 mode or the S2 mode is located. It seems that none of dummy head can shift the frequency of Mode 1. Figures 7 (a) and (b) show that there is a relationship between the T2 (S2) mode and Mode 2. Mode 2 appears around 23. khz when the T2 mode is in the range from 15. khz to 2. khz and the S2 mode is from 2. khz to 25. khz. End and middle Arm-Bs show a similar tendency. Hence, we used end Arm-B for optimizing dummy head. We found that the HDD-d actuator will have its Mode 2 over 21. khz by using such dummy heads that put the T2 mode from 15. khz to 2. khz and the S2 mode from 2. to 25. khz. If we set Mode 2 over 21. khz, the HDD-d actuator will have a frequency band with a low gain around 2. khz. That means that we can use the same sampling frequency for HDD-f and HDD-d, since HDD-f has a frequency band with a low gain from 18. khz to 2. khz. 4. Dummy Head Design To control modal frequency, we focused on a dummy head with a neck shown in Fig.8. For it has a small stiffness part in the middle, we thought that the dummy head enabled us to control modal frequency efficiently. Using experimental design and the parameters shown in Table 1, we investigated the effects of some different dummy heads on the T2 mode and the S2 mode with the FEM analysis. Parameters a and d are the length and width of the dummy head s edge. Parameters b and c are the width and length of the neck. Figures 9 and 1 show the results of experimental design. In both figures, the horizontal axis is the parameter value. For example, a1 means the parameter a and #1 value in Table 1. 85

5 Vertical axis is a relative contribution; if a parameter has high relative contribution under certain condition, modal frequency will be at high frequency under the condition. For example, if the a1 shows a higher relative contribution than that of a2, we can set the mode at a higher frequency by using a s #1 value than by using a s #2 value. Figure 9 shows that a larger value for any parameter lowers the frequency of the T2 mode. In particular, parameters c and d have a large influence. Figure 1 shows that smaller b decreases the frequency of the S2 mode. Larger c and larger d put larger mass at further position from the fixed point and the torsion mode will appear at lower frequency. Smaller b reduces the in-plane stiffness around the dummy head and the sway mode will appear at lower frequency. For these reasons, parameters c and d have a large influence on T2 mode and parameter b has a large influence on S2 mode. Using these results, we designed a dummy head so that the end Arm-B model has its T2 mode around 16. khz and its S2 mode around 25. khz. The dimensions are a=2. mm, b=.5 mm, c=1. mm and d=4. mm. Though it is necessary to design dummy head with considering product tolerances, we c a b d Fig. 8 Shape of dummy head Table 1 Parameter values of dummy head Parameter #1 #2 #3 a b c d Relative cotribution Relative cotribution Gain [db] a1 a2 a3 b1 b2 b3 c1 c2 c3 d1 d2 d3 Fig. 9 Effect on T2 mode a1 a2 a3 b1 b2 b3 c1 c2 c3 d1 d2 d3 Fig. 1 Effect on S2 mode Middle Arm-A End Arm-A T2 S2 S Fig. 11 Vibration characteristics in Arm-B having optimized dummy head 86

6 Gain [db] HDD-f upward HDD-f downward HDD-d upward HDD-d downward Fig. 12 Vibration characteristics in HDD-d actuator having the optimized dummy head use these dimensions here as a representative. The vibration characteristics of end and middle Arm-B with this dummy head are shown in Fig. 11. We can see that end Arm-B and middle Arm-B have T2 modes around 18. khz and S2 modes around 25. khz. Figure 12 shows the vibration characteristics of the HDD-d actuator with this dummy head. Mode 2 is around 23. khz. We found that the HDD-d actuator has a wide frequency band with a low gain from 18. to 23. khz. The HDD-d actuator s frequency band has as low a gain as the HDD-f actuator around the sampling frequency. As for the bend mode of this actuator, the FEM analysis has shown that dummy head edge doesn t contact on the disk by shock input within specifications 5. Conclusion We developed a way to control modal frequency by using dummy heads. Numeric calculation confirmed that there is a relation between the vibration characteristics of the Arm-B model and those of the HDD-d actuator model, and that dummy heads influence the vibration characteristics of Arm-B. In our sample, it seems that if we put the T2 mode from 15. khz to 2. khz and the S2 mode from 2. khz to 25. khz, the HDD-d actuator will have Mode 2 around 23. khz. According to these findings, we designed a dummy head that gives Arm-B a T2 mode around 18. khz and an S2 mode around 25. khz. We found that the dummy head can control the HDD actuator s vibration characteristics and that the optimized dummy head enables the HDD-d actuator to keep its gain around the sampling frequency as low as that of the HDD-f actuator. References [1] Atsumi, T., Arisaka, T., Shimizu, T., and Masuda, H., Head-Positioning Control Using Resonant Modes in Hard Disk Drives, IEEE/ASME TRANSACTIONS ON MECHATRONICS, vol.1, no.4, Aug. 25, pp [2] F-Y., Huang, Semba, T., Imaino, W., F., Lee, Active Damping in HDD actuator, IEEE/TRANSACTIONS ON MAGNETICS, vol.37, no.2, Mar. 21, pp [3] F-Y., Huang, Tshuda, S., Ohta, M., Actuator dynamics optimization through FAAM and beyond, ASME/JSME Joint Conference on Micromechatronics for Information and Precision Equipment (MIPE26), Jun. 26. [4] Nishida, T., Tominaga, J., Hanya, M., Hayashi, T., Dynamic Damping of Arm 1 st Bending for HDD Shock, ASME/JSME Joint Conference on Micromechatronics for Information and Precision Equipment (MIPE26), Jun. 26. [5] H-S., Lee, Y-H Kim, T-Y., Hwang, C-S., Kim, VCM Design to Improve Dynamic Performance of an Actuator in a Disk Drive, IEEE/TRANSACTIONS ON MAGNETICS, vol.41, no.2, Feb. 25, pp [6] Suzuki, K., Tsuda, S., Deguchi, T., Shimizu, T., VCM Design with Round Coil and Axe-Shaped Magnet for Hard Disk Drive Actuator, ASME/JSME Joint Conference on Micromechatronics for Information and Precision Equipment (MIPE26), Jun. 26. [7] F-Y., Huang, Ohta M., Takahashi, K., Takeuchi, K., US Patent Application [8] Shinada, H., Ono, K., R., Evans, Yamaura, H., Study of a Head Suspension with Dynamic Absorber, Conference on Information, Intelligence and Precision Equipment 25, 25,pp