Patent Grant
7215498
This U.S. patent application incorporates by reference all of the following co-pending applications:
U.S. patent application Ser. No. 10/872,062 entitled “Method for Optimizing Dynamic Stroke in the Self Servo-Write Process,” by Calfee, et al., filed Jun. 17, 2004.
U.S. Provisional Application No. 60/533,292 entitled “Method for Optimizing Track Spacing Across a Stroke,” by Gururangan, et al., filed Dec. 30, 2003.
U.S. Provisional Application No. 60/533,454 entitled “System for Optimizing Track Spacing Across a Stroke,” by Gururangan, et al., filed Dec. 30, 2003.
U.S. patent application Ser. No. 10/733,131 entitled “Methods to Determine Gross and Fine Positioning on a Reference Surface of a Media,” by Richard M. Ehrlich et al., filed Dec. 10, 2003.
The present invention relates to methods to servowrite media for use in data storage devices, and systems for applying such methods.
A hard disk drive typically contains one or more disks clamped to a rotatable spindle motor, at least one head for reading data from and/or writing data to the surfaces of each disk, and an actuator utilizing linear or rotary motion for positioning the head(s) over selected data tracks on the disk(s). The actuator positions the read/write head over the surface of the disk as the spindle motor rotates and spins the disk.
As the head is loaded onto a disk, for example from a ramp, the servo system determines the position of the head on the disk surface by reading servo wedges passing beneath the head. A first track identified by the servo system as the head unloads from the ramp is identified as an acquire track. A first user track can be assigned based on the position of the acquire track, and can define an outer boundary of a data region. The acquire track is some small distance from the ramp, and farther from the outer diameter of the disk than is optimal or desired, wasting otherwise usable space and requiring an increased track density for a given hard disk drive capacity.
Details of embodiments of the present invention are explained with the help of the attached drawings in which:
A rotary actuator 110 is pivotally mounted to the housing base 104 by a bearing 112 and sweeps an arc between an inner diameter (ID) of the disk 108 and a ramp 130 positioned near an outer diameter (OD) of the disk 108. Attached to the housing 104 are upper and lower magnet return plates 118 and at least one magnet that together form the stationary portion of a voice coil motor (VCM). A voice coil 116 is mounted to the rotary actuator 110 and positioned in an air gap of the VCM. The rotary actuator 110 pivots about the bearing 112 when current is passed through the voice coil 116 and pivots in an opposite direction when the current is reversed, allowing for precise positioning of the head 114 along the radius of the disk 108. Each side of a disk 108 can have an associated head 114, and the heads 114 are collectively coupled to the rotary actuator 110 such that the heads 114 pivot in unison. The invention described herein is equally applicable to devices wherein the individual heads separately move some small distance relative to the actuator. This technology is referred to as dual-stage actuation (DSA).
Refinements in disk fabrication have enabled manufacturers to produce disks 108 having ultra-smooth surfaces. Electrostatic forces can cause stiction between the slider 228 and the surface. If the speed of rotation of the disk 108 slows such that the air bearing collapses, the slider 228 can contact and stick to the surface of the disk 108, causing catastrophic failure of the hard disk drive 100. Stiction can cause the disk 108 to abruptly lock in position or stiction can cause the slider 228 to forcibly disconnect from the suspension 226. Thus, when the hard disk drive 100 is not in use and before rotation of the disks 108 is slowed and stopped (i.e., the disks 108 are “spun down”), the heads 114 can be removed from close proximity to the disk 108 surface by positioning the suspension 226 on a ramp 130 located either adjacent to the disk 108 or just over the disk 108 surface.
The slider 228 is removed from close proximity with the disk 108 by pivoting the actuator 110 such that a lift tab 332 extending from the suspension 226 contacts the ramp surface and slides up the ramp 130. The position along the ramp 130 where the lift tab 332 first contacts the ramp 130 can be called the touch-point. As the lift tab 332 slides up the ramp 130 from the touch-point, the ramp 130 opposes the spring force of the suspension 226 and forces the slider 228 (and the head 114) away from the disk 108 surface. The HSA 222 can continue its motion along the stroke by traveling up the grade portion of the ramp 130 to a substantially flat portion that optionally can include a detent for cradling the lift tab 332. The slider 228 can be loaded back onto the disk 120 after the disk spins up to a safe speed. In other embodiments, the suspension 226 contacts the ramp 130 at a location along the suspension 226 between the slider 228 and the pivot point. Unloading the slider 228 from the disk 108 prevents sticking, and reduces a risk of damage from non-operating shock by suspending the slider 228 over a significantly wide gap between the slider 228 and an opposing slider or surface. In still other embodiments in accordance with the present invention, the hard disk drive 100 can include a ramp 130 positioned near the ID, rather than near the OD. In such embodiments, the slider 228 is removed from close proximity with the disk 108 by pivoting the actuator 110 toward the ID such that the lift tab 332 (or suspension 226) contacts the ramp surface and slides up the ramp 130. Such hard disk drives 100 can further include an OD crash stop which can be associated with the housing, and can limit or block a pivoting movement of the actuator 110 at the OD. Methods in accordance with the present invention are equally applicable to such hard disk drives 100 having a ramp 130 positioned near the ID, and optionally an OD crash stop. Systems and methods described below are described with reference to embodiments of hard disk drives 100 having a ramp 130 positioned near the OD and an ID crash stop; however, it will be understood by one of ordinary skill in the art that such embodiments can alternatively include a hard disk drive 100 having a ramp 130 positioned near the ID, and optionally an OD crash stop, and that such embodiments are within the scope of the present invention.
It should be noted, the description herein of the disk surface passing under or beneath the slider is intended to mean that portion of the disk surface that is in close proximity to the slider. It will be understood that when referred to as “beneath” or “under” the slider, the disk surface can be over, or adjacent to the slider in actual physical relation to the slider. Likewise, it will be understood that when referred to as “over” the disk surface, the slider can be beneath, or adjacent to the disk surface in physical relation to the disk surface. By extension, where the slider is beneath the disk surface, the suspension travels down the ramp when the slider is separated from the disk surface.
A flexible circuit (not shown) is connected with the rotary actuator 110 to supply current to the voice coil 116 and to provide electrical connections to the heads 114, allowing write signals to be provided to each head 114 and allowing electrical signals generated during reading to be delivered to pre-amplification circuitry (pre-amp) 448. Typically, the flexible circuit comprises a polyimide film carrying conductive circuit traces connected at a stationary end with the lower housing 104 and at a moving end to the rotary actuator 110. The disk controller 444 provides user data to a read/write channel 450, which sends signals to the pre-amp 448 to be written to the disk(s) 108. The disk controller 444 can also send servo signals to the microprocessor 446, or the disk controller 444 can control the VCM and spindle motor drivers directly, for example where multi-rate control is used. The disk controller 444 can include a memory controller for interfacing with buffer memory 456. In one embodiment, the buffer memory 456 can be dynamic random access memory (DRAM). The microprocessor 446 can include integrated memory (such as cache memory), or the microprocessor 446 can be electrically connected with external memory (for example, static random access memory (SRAM) 454 or alternatively DRAM).
When a slider is loaded onto a disk from a ramp, the servo system must determine the position of the head along the stroke. The HSA is unstable when the slider is initially loaded due to suction forces and the transition from the graded ramp to the disk. Once the slider stabilizes and an air bearing is established between the disk and the slider, the head 114 can determine its position on the disk by reading servo wedges passing beneath the head 114. After some criteria is met—e.g., the track is measured on a predefined number of consecutive servo wedges—the head locks onto a track. The track on which the head locks is called an acquire track.
The information stored on such a disk can be written in concentric tracks, extending from near the ID to near the OD, as shown in the exemplary disk of
Servo information often includes transition pairs called “servo bursts.” The servo bursts 672 can be positioned regularly about each track, such that when a data head reads the servo bursts 672, a relative position of the head can be determined that can be used to adjust the position of the head relative to the track. For each servo wedge, this relative position can be determined, in one example, as a function of the target location, a track number read from the servo wedge, and the amplitudes or phases of the bursts 672, or a subset of those bursts 672. The position of a head or element, relative to the center of a target track, will be referred to herein as a position-error signal (PES).
For example, a centerline 676 for a given data track can be “defined” relative to a series of bursts, burst edges, or burst boundaries, such as a burst boundary defined by the lower edge of A-burst and the upper edge of B-burst. The centerline 676 can also be defined by, or offset relative to, any function or combination of bursts or burst patterns. This can include, for example, a location at which the PES value is a maximum, a minimum, or a fraction or percentage thereof. Any location relative to a function of the bursts can be selected to define track position. For example, if a read head evenly straddles an A-burst and a B-burst, or portions thereof, then servo demodulation circuitry in communication with the head can produce equal amplitude measurements for the two bursts, as the portion of the signal coming from the A-burst above the centerline 676 is approximately equal in amplitude to the portion coming from the B-burst below the centerline 676. The resulting computed PES can be zero if the radial location defined by the A-burst/B-burst (A/B) combination, or A/B boundary, is the center of a data track, or a track centerline 676. In such an embodiment, the radial location at which the PES value is zero can be referred to as a null-point. Null-points can be used in each servo wedge to define a relative position of a track. If the head is too far towards the outer diameter of the disk, or above the centerline, then there will be a greater contribution from the A-burst that results in a more “negative” PES. Using the negative PES, the servo controller could direct the voice coil motor to move the head toward the inner diameter of the disk and closer to its desired position relative to the centerline. This can be done for each set of burst edges defining the shape of that track about the disk.
The PES scheme described above is one of many possible schemes for combining the track number read from a servo wedge and the phases or amplitudes of the servo bursts. For example, U.S. Pat. No. 5,381,281 to Shrinkle et al. describes a PES scheme including a quad-servo burst pattern having first, second, third, and fourth servo bursts distributed in a series along the length of a portion of the data sector such that the center point of each servo burst is offset from adjacent bursts by a radial distance equivalent to one-half of the data track width. A quadrature-based track following algorithm applying a difference of sums of servo burst pair read voltages can minimize track following errors where servo bursts are mispositioned relative to one another. Such a scheme can benefit from embodiments of the present invention, as can many other track following schemes. The schemes described above are only a few of many possible schemes for positioning the head. Hard disk drives using most (if not all) possible PES schemes could benefit from the invention contained herein.
Servo patterns can be written to the disks prior to assembly of the hard disk drive 100 using a media writer. A stack of disks is loaded onto the media writer and servo patterns are carefully written onto the surface of each disk, a time consuming and costly process. Alternatively, a commonly less time-consuming and less expensive method can include writing servo patterns or template patterns on a reference surface of a single blank disk to be used as a reference for self-servo writing unwritten (and written) surfaces of one or more disks of an assembled hard disk drive. In one such self-servo writing method, called printed-media self-servo writing (PM-SSW), a coarse magnetic template pattern can be transferred to a single disk surface (a reference surface) by magnetic printing. A magnetic printing station can be used to magnetically print or otherwise transfer a template pattern using a known transfer technique. One such transfer technique is described in “Printed Media Technology for an Effective and Inexpensive Servo Track Writing of HDDs” by Ishida, et al. IEEE Transactions on Magnetics, Vol. 37, No. 4, Jul. 2001. A blank disk (the reference surface) is DC erased along the circumferential direction of the disk by rotating a permanent magnet block on the disk surface. A template, or “master”, disk is then aligned with the blank disk and the two disks are securely faced with each other by evacuating the air between the two disk surfaces through a center hole in the blank disk. An external DC field is applied again in the same manner as in the DC erasing process, but with an opposite polarity. A number of different transfer techniques exist, and one of ordinary skill in the art can appreciate the different methods for transferring a template pattern to a reference surface.
A completed and enclosed hard disk drive can be assembled with at least one disk 108 having a reference surface, and optionally one or more blank disks. The template pattern 780 is applied by the hard disk drive electronics to self-write highly resolved product embedded servo patterns 670 onto storage surfaces of each disk 108, including the reference surface having the template pattern 780. When the at least one disk 108 is removed from a magnetic printing station and connected with a spindle 120, a shift typically occurs between the axis of rotation and the center of tracks of the template pattern 780. The shift is attributable to machining tolerances of the spindle and magnetic printing station, as well as other variables. The track followed by the head 114 can be displaced laterally in a sinusoidal fashion relative to the head 114 as the disk 108 rotates. This sinusoidal displacement is typically referred to as eccentricity. Firmware executed by the hard disk drive 100 and the hard disk drive electronics enable the head 114 positioned over the reference surface to follow and read the template pattern 780 and enable each of the heads 114 to write precise final servo patterns 670 on each of the respective surfaces of each disk 108. The hard disk drive 100 can compensate for eccentricity, writing tracks that are nominally concentric with the center of rotation of the spindle, or alternatively, having some built-in eccentricity as defined by the firmware, for example. A final servo pattern 670 can be written to the reference surface in any sequence, i.e. prior to, subsequent to, or contemporaneously with writing final servo patterns on some or all of the other surfaces. The final servo patterns can be written contemporaneously to reduce servo write times, and the final servo patterns 670 can be written between pattern wedges of the template pattern 780. The template pattern 780 is overwritten either during the self-servo writing process or by user data. For example during hard disk drive 100 testing data is written to the data fields and read back to test the data fields.
Traces 883 overlay the pulse pairs 882 in
Most commonly-used servo demodulation systems determine the digital content of a servo wedge signal by detecting either the presence or absence of filtered signal pulses at specified times or by detecting the value of the filtered signal at specified times. The signal can be filtered through a low-pass filter, a high-pass filter, or a combination of the two (i.e., a band-pass filter). The amplitude of the filtered signal can be calculated and compared to a threshold. The threshold can vary with an average amplitude of the filtered signal in the vicinity. The location along the stroke where the amplitude no longer exceeds the threshold can be used as a crude position signal indicating a marker-zone edge. A radial position of the head 114 can be known within a distance that is smaller than the size of the read width of the head 114 by detecting the marker-zone edge. The read width of the head 114 is much smaller than the width of the chevron cycle wc. For example, in one embodiment the width of the chevron cycle is 3 μm. The width of the read head 114 is a small fraction of a micron. Therefore, the chevrons can provide fractional positioning of the head 114 relative to the gross positioning provided by the marker-zone edge.
A chevron cycle located at the same radial position as the marker-zone edge can be assigned a designated cycle count from which the head 114 can determine radial positioning along the stroke by the cycle count of the chevron over which the head 114 passes relative to the marker zone edge. If the position of the head 114 is lost, the head 114 can locate the marker-zone edge and the radial position is known to be the designated cycle count. For example, if the designated cycle count is 1000, the radial position of the marker-zone edge is chevron cycle count 1000 (plus a fractional cycle count based on whatever fractional position is measured from the actual chevron angle). Use of this scheme can present a problem if the location of the marker-zone edge nearly coincides with an exact integer chevron cycle count. If one of the chevrons (either the zig-burst 884 or the zag-burst 886) has a phase of very nearly zero degrees at the edge of the marker-zone, then it can be difficult to decide whether to set the integer portion of the chevron cycle count to the designated cycle count or one count less than the designated cycle count. Using the example discussed above, the designated cycle count for the zig-burst 884 at the marker-zone edge is 1000, while the corresponding designated cycle count for the zag-burst 886 is −1000. If the measured phase of the zig-burst 884 at the marker-zone edge is very near zero degrees, for the servo wedge at which the chevron cycle counts are altered to account for the known location of the head 114, where the measured phase of the fractional cycle count is slightly more than zero degrees (i.e., a small positive phase) the integer portion of the zig-burst 884 cycle count can be set to 1000, while where the measured phase of the fractional cycle count is slightly less than 360 degrees (i.e., a small negative phase) the integer portion of the zig-burst 884 cycle count can be set to 999. Thus, a phase of a fractional cycle count near zero degrees (but slightly greater) will result in a total zig-burst 884 cycle count that is slightly greater than 1000, while a phase of a fractional cycle count near to 360 degrees (but slightly less) will result in a total zig-burst 884 cycle count that is slightly less than 1000. The same reasoning can be applied to determine the integer portion of the zag-burst 886 cycle count at the time that both the zig-burst 884 and zag-burst 886 cycle counts are altered to account for the known location of the head 114.
The marker-zone can be positioned anywhere along the stroke. In one embodiment, the marker-zone can be positioned centrally along the data stroke (wherein the data stroke is that portion of the stroke traversing data tracks), bisecting the data stroke and minimizing the maximum distance from any location on the disk to the marker-zone, thereby improving nominal recovery time where the head 114 slips chevron cycles. In other embodiments, the marker-zone can span a defined distance and have a first edge, for example, near the OD and a second edge near the ID. One of ordinary skill in the art can appreciate the myriad different arrangements of the marker-zone on the disk.
Referring to
The data stroke traverses a portion of the disk surface between the first user track 992 and a final user track 994 offset from the ID crash stop 131 by an inner guard band IG. In low-cost designs, the mechanical tolerance of the ID crash stop 131 location and the touch-point 934 location is a significant portion of the data stroke. The location of the average acquire track 990 from the touch-point 934 includes a tolerance that can vary with the criteria for assigning an average acquire track 990; therefore, setting the first user track 992 based on the average acquire track 990 can further reduce the width of the data stroke (and increase the variability). Further, the first user track 992 is typically assigned to a track that is a conservative distance from the average acquire track 990. Typically, a manufacturer will increase the density of the tracks written to the disk 108 surface to produce a hard disk drive 100 having a targeted capacity. An increase in track density can negatively impact hard disk drive 100 performance, resulting, for example, in a reduction in manufacturing tolerance for the width of the head 114, or a degradation in the performance of the servo system.
The touch-point 934 can be more accurately located for defining a first-user track 992 by detecting a dramatic change in an average bias force as the actuator 110 contacts the ramp 130. Electrical bias forces can result from voltage and current offsets in the electrical circuitry and can act on a rotary actuator 110 as a function of the radial position of the head 114 on the disk 108. An average bias force can be measured by the servo system as the head 114 reads servo wedges passing beneath the head 114. The servo system can seek the OD and measure the average control effort (i.e. bias force) required as the head 114 changes radial position.
Alternatively, the touch-point 934 can be located by detecting a dramatic change in a level of gain adjustment in an automatic gain control (AGC) circuit associated with the read/write channel 450. The AGC circuit adjusts the amplitude of a signal received from the current preamplifier 448 within desirable boundaries when converting an analog signal into digital form.
As described above, data tracks written to the disk surface can be formatted in radial zones. For example, the servo pattern of
A method in accordance with one embodiment of the present invention can include determining a final servo pattern to be written to one or more surfaces of a disk during a self-servo write process. The method can be applied to a reference surface having a template pattern, for example as shown in
Referring to the flowchart of
The ID crash stop can be identified in a similar fashion. Referring to the flowchart of
Once the ramp interference point and the ID crash stop interference point have been determined, the mechanical deviation of the ramp and the ID crash stop from a nominal radial position can be calculated. The mechanical deviation of the ID crash stop and the ramp interference point can be used as manufacturing feedback data, and optionally used as failure criteria. In one embodiment, statistical methods are applied to calculate a distribution around a nominal value of radial position for the ID crash stop interference point and ramp interference point. For example, in one embodiment a Gaussian distribution can be calculated and a deviation, e.g. 3 sigma, can be assigned as a failure criteria. Alternatively, a fixed value for a radial position can be assigned as a failure criteria. Assembled hard disk drives that fail one or both of the failure criteria for the ID crash stop and ramp interference points can be binned as lower capacity drives, discarded, or otherwise dispositioned. In other embodiments, a total value of the data stroke is calculated from the ID crash stop and ramp interference points and compared with a failure criteria calculated or determined for the data stroke. Multiple different criteria can be applied to reject hard disk drives having data strokes too small to provide robust performance at the targeted radial density.
If a hard disk drive falls within acceptable criteria, the radial positions of the ID crash stop and ramp interference points can be used to calculate the available data stroke. Referring to the flowchart of
In other embodiments, the final servo pattern can be written so that a variable number of data tracks are accurately placed at the appropriate radial locations, again, according to a single read/write format and radial density. This scheme also assures accurate data frequency at the various radial data zones, and a minimum ID and OD guard-width. However, an increase in the width of the data stroke results in an additional number of data tracks, increasing the capacity of the disk. In this way, hard disk drives can be binned and sold according to capacity, or alternatively customized, having only a minimum capacity and a variable maximum capacity.
In still other embodiments, a minimum ID and OD guard width can be assigned, based on a slider width, or some other criteria, and the remaining data stroke is used to write data tracks having a variable radial density to maximize robustness of the written data for a given capacity. The density of the remaining data stroke is determined by the radial width of the remaining data stroke and the relative proportion of the remaining data stroke within the inner and outer radial zones. For example, where a data stroke of a disk in a first hard disk drive is shifted closer to the ID than a data stroke of a disk having the same radial width in a second hard disk drive, the disk from the first hard disk drive will have a higher radial density. This is because the radial positions of the radial zones are fixed; therefore the size of the inner radial zone, having a lower frequency than the outer radial zone, increases when the ID crash stop interference point shifts toward the ID, while the size of the outer radial zone, conversely having a higher frequency than the inner radial zone, increases when the crash stop interference point shifts away from the ID and toward the OD.
Methods in accordance with the present invention can further be applied to self servo write a plurality of disks or a plurality of disk surfaces connected with a spindle motor. Where a plurality of heads are connected with the actuator, a position of the ramps can be determined relative to a marker zone edge by positioning the plurality of heads over the respective disk surfaces, locating the marker zone edge as described above, and pivoting the actuator toward the OD of the plurality of disks surfaces until the actuator contacts at least one of the ramps. A metric—e.g. an average bias force and/or AGC level—is measured by the heads as the actuator pivots until contact between at least one of a plurality of HSAs connected with the actuator and a corresponding ramp is detected. In one embodiment, the plurality of heads are tied together via the head stack and move together on the actuator. The head closest to a corresponding ramp determines the ramp interference point common to all heads. The bias force will change while servoing on any head when the head nearest a corresponding ramp comes into contact. Once the common ramp interference point is determined relative to the marker zone edge, the actuator can be pivoted toward the ID until the actuator contacts the ID crash stop. As described above, the number of cycles between the common ramp interference point and the marker zone edge, and between the marker zone edge and the crash stop interference point can be measured as the head travels across the reference surface. A final servo pattern can be determined and written to the one or more surfaces of the disk(s) as described above.
In some embodiments, multiple surfaces can include printed reference patterns. In such embodiments, a ramp interference point can be determined for each surface and corresponding head by measuring a metric only from the head associated with the target surface. A final track layout can be determined for each of the multiple surfaces, and a final servo pattern can be written to each of the multiple surfaces in accordance with the final track layout. Such embodiments can provide an advantage in optimizing track layout across the entire drive, particularly where the mechanical tolerance between relative head position is large.
The foregoing description of preferred embodiments of the present invention has been provided for the purposes of illustration and description. It is not intended to be exhaustive or to limit the invention to the precise forms disclosed. Many modifications and variations will be apparent to one of ordinary skill in the relevant arts. The embodiments were chosen and described in order to best explain the principles of the invention and its practical application, thereby enabling others skilled in the art to understand the invention for various embodiments and with various modifications that are suited to the particular use contemplated. It is intended that the scope of the invention be defined by the claims and their equivalence.
| Number | Name | Date | Kind |
|---|---|---|---|
| 4354208 | Maury | Oct 1982 | A |
| 5189571 | Murphy et al. | Feb 1993 | A |
| 5862005 | Leis | Jan 1999 | A |
| 6005725 | Emo et al. | Dec 1999 | A |
| 6069764 | Morris | May 2000 | A |
| 6304407 | Baker et al. | Oct 2001 | B1 |
| 6633451 | Chainer et al. | Oct 2003 | B1 |
| 6738205 | Moran et al. | May 2004 | B1 |
| 20010043428 | Morris et al. | Nov 2001 | A1 |
| 20050275963 | Ehrlich | Dec 2005 | A1 |
| Number | Date | Country | |
|---|---|---|---|
| 20050280916 A1 | Dec 2005 | US |
