Patent Application
20070047124
1. Field of the Invention
This invention relates generally to a device for testing magnetic read/write heads and more specifically to such a device in which there is a rotating magnetic disk on which servo information has been written.
2. Description of the Related Art
Before their installation in a conventional direct access storage device (DASD), magnetic read/write heads are first tested in a device specifically designed for the testing process. A magnetic head tester is basically a rudimentary disk drive that includes a drive motor and spindle (called a spin stand), at least one magnetic disk mounted on the drive motor spindle and capable of being rotated thereby, an actuator on which the head to be tested is mounted and an electro-mechanical system, called a servomechanism (or, equivalently, a servo-control mechanism), for accurately positioning the actuator (and its mounted head) relative to the rotating disk. The testing device also includes a micro-positioner and associated circuitry for writing servo-tracks, to be discussed below, on the magnetic disk, so that the head can be accurately positioned at target locations on the disk.
In order to test the read and/or write capabilities of a head, it is necessary to accurately position the head at various places on a disk mounted within the tester. To insure the accuracy of this positioning process, the disk mounted within the tester is furnished with embedded information, called servo information, which is data stored (as “bursts” of magnetic transitions) within sectors of small angular width that are periodically distributed along radially narrow, concentric, annular circular tracks on the disk. This servo information, which is written on the disk while the disk is already mounted in the tester, is written using a micro-positioner affixed within the tester and the head already mounted in the tester. The micro-positioner is a device that can incrementally move the head to proper positions at which to write the servo information.
The servo information identifies radial positions within the annular width of the tracks, so that the track center can be accurately located and it also identifies the angular positions of data-free sectors within the tracks, typically adjacent to the sectors containing the servo information, on which data can subsequently be written and read by the head during the testing process. The role of the servo data is of great importance in accurately positioning the head relative to the center of the track and in maintaining that position during portions of the testing process. Thus, servo data supplies both track identifying information and positioning information
When the actuator mounted head seeks a particular position on the disk (the target position), which is typically the radial center of a track at some angular position along the track, the servo information located adjacent to that position is read by the head being tested and used to determine whether the head is actually located where it is supposed to be. This servo information is read immediately before the head reaches the target position and is transmitted to what is called servo-loop circuitry that is an integral part of the servomechanism of the tester. The difference (if it exists) between the intended location of the actuator (the target) and the actual location of the actuator as indicated by the adjacent servo data, generates a position error signal (PES) that is then used to correct the actuator positioning on the track. The PES, which is digital data, is supplied to a digital-to-analog converter (DAC) that generates a current proportional to the digital PES and, in turn, activates a voice coil motor (VCM), which is a current carrying coil positioned between permanent magnets. The VCM then responds to the DAC generated current and repositions the actuator.
There are many reasons why the actuator may be positioned incorrectly. Random vibrations can easily move the head slightly from its target, but much of the position error is related to the fact that the typical magnetic disk is slightly warped or generally fails to rotate perfectly on its spindle. Thus, even if the actuator is locked at what should be the fixed radial position of the target track's center line, the projection of the head on the rotating disk is not a circle concentric with the disk center, whereupon the head fails to follow the target track's center line and may, in fact, overlap several tracks.
The servo data, because of the process with which it is written, is presumably located on tracks that are essentially concentric circles of small radial width (circular annuli). Therefore, when the disk rotates improperly, there is a discrepancy between the track followed by the actuator (and mounted head) and a circular track produced by the servo data writing process. Nevertheless, the embedded servo data is supposed to rectify the improper tracking by means of the provided PES.
Writing accurate servo data on the disk is an important, expensive and time consuming task that is done within the head tester itself using an open-loop or closed-loop micro-positioner and the tester's head.
The additional elements within the tester needed to write servo data, including the micro-positioner and its associated circuitry, add greatly to the expense of the tester. In addition, the use of these additional elements, as noted, adds to the time required to complete the testing process. Even with the added time and expense, however, the repeatability of the servo pattern on a plurality of disks is often questionable and it is difficult to verify pattern accuracy. This is because the head testing device is not an optimal device within which to write servo data.
Takano et al. (US Patent Application Publication: US 2002/0018314 A1) describes a magnetic-disk evaluation apparatus in which servo information is both written on a disk and tested for its accuracy. In such a disk-evaluation process, servo information is first written on the disk in a temporary form, often by simply copying it from another disk. If this servo information is of sufficient accuracy, the disk can be tested for compliance with certain industry standards and, if those standards are met, the disk is then mounted within its final hard disk drive where more accurate servo information is written upon it. A problem arises if the temporary information is of insufficient accuracy to permit the disk to be mounted for writing of the permanent information. Takano et al. therefore provide a system in which the temporary servo signals can be evaluated using a measuring device and positioner to determine whether the arm of the evaluation apparatus is displaced from a target position by an unacceptable amount.
Shitara et al. (U.S. patent Application Publication No.: US2003/0128456) describes a magnetic head positioning control for a magnetic head certifier and magnetic disk certifier that includes a piezo actuator positioned between a suspension spring and a head carriage. The object of Shitara is to rectify the problem associated with the high speed at which the actuator mounted head must respond to servo signals that are being used to accurately position the head. Shitara's piezo positioner is claimed to be capable of more quickly and more precisely positioning a head in response to servo signals.
Although the present invention is not directed at providing an improved mechanism for writing temporary servo data or for more accurately controlling the position of an actuator, Takano and Shitara indicate the difficulties associated with both processes and, in that sense, provide a further justification and substantiation of our assertion that the elimination of such mechanisms from within the head tester itself will be highly advantageous.
The present invention, therefore, proposes the use of a disk in a head testing apparatus in which embedded servo data has been pre-written outside the testing apparatus by a dedicated servo disk writer. The sole task of a dedicated servo disk writer is to efficiently, accurately and repeatedly write servo information on disks. Such a device can do the job with much greater accuracy and repeatability than can be accomplished using a head tester with an included micro-positioner as a servo track writer. Moreover, by allocating the task of servo track writing to a device that is designed specifically to perform such a task, the necessity of using the head tester to perform the same task in a less exact, repeatable and efficient manner, is eliminated. Thus, the head tester can be simplified and dedicated to doing what it is meant to do, namely to test heads.
A problem arises, however, when a disk that has its servo tracks written in one apparatus is then transferred to another. The problem is a result of the fact that servo data is written on circular tracks that are substantially concentric when originally written on the disk, but these tracks will generally be eccentric when the disk is rotated by the drive spindle of the apparatus to which it is transferred (the host apparatus). This, of course, will be the inevitable result if the drive spindle of the original servo writer has even the slightest wobble or if the disk itself slips, is warped or off-center, or if the writing apparatus was subjected to thermal or mechanical shocks during the writing process or if the host apparatus has similar problems. Given that present 3.5″ disks have a track pitch of approximately 100,000 tracks per inch (TPI), it can be seen that the radial width of each track is miniscule and that the slightest variations in track concentricity will be exacerbated by variability between machines or external perturbations. Thus, if a dedicated servo writer is to be used to write the servo information on a disk that is then mounted in a head tester, the head tester must have the ability to either eliminate or compensate for the eccentricities of the pre-written servo data.
The lack of track concentricity encountered by a head testing device attempting to read what are supposed to be concentric circular tracks is termed “repeatable runout” or RRO and, when it occurs, new writes by the read/write head, if the head is kept at a fixed radial position, can cross over several tracks and can overwrite previously written data. It is to be noted that the RRO problem is not restricted to head testing devices, but is also a common problem in the disk drives of actual DASD's themselves. Wherever it occurs, accurate positioning of the read/write head becomes nearly impossible without additional information being present to guide the head to the correct track positions and, when possible, to correct for RRO in some systematic way. This additional information, which is then used to actuate the locating and position-correcting servomechanisms within the drive unit, is the servo information within the tracks themselves.
Along with the repeatable runout that is associated with off-center drive spindles or warped disks, there is the more difficult problem of non-repeatable runout, NRRO, associated with random mechanical, electrical and thermal perturbations of the drive system and/or disk. In principle, the RRO is a stable periodic effect that does not change during operation of the tester, whereas NRRO can change with the external effects that cause it. Clearly, if the regular effects of RRO can be eliminated from the drive system, then the NRRO can be more easily addressed. In general, the servo data embedded in the disk provides enough information to fully characterize the effects of RRO and, therefore, it provides enough information to eliminate or significantly suppress those effects.
Repeatable runout in DASD disk drives (as opposed to head testing devices) has been addressed in the prior art in cases where the servo information is written by the same machine that subsequently uses it, or when the servo information is written in a different machine than the one that uses it. The most common case in DASD disk drives, is where the drive manufacturer writes servo information on the disk using the same drive mechanism that subsequently drives the disk during regular operation. In this case, the problem of transferring a disk from one apparatus to another does not occur.
Nazarian et al. (U.S. Pat. No. 6,310,742) teaches a method for canceling repeatable runout that is caused by unavoidable imperfections in the servo information writing process. The method “learns,” and stores the actual positions of eccentric target data tracks by repeated sampling of the runout values of all the servo sectors written on the disk, then effectively cancels the effects of repeatable runout by subtracting, within the PES, each sector's measured runout value from the servo information actually written into that sector. In short, the servomechanism can concentrate on only correcting for NRRO, because the effects of the RRO have already been subtracted from the PES.
Sri-Jayantha et al. (U.S. Pat. No. 6,097,565) teach a “No-RRO Servo Architecture,” where the RRO component is “ignored,” ie. not tracked. The method removes the RRO component from the PES by subtracting a “locked arm” RRO prior to generating the servo-controller output. In other words, the actual radial position of the actuator is subtracted from the position stored in the servo location beneath it and this difference is essentially removed from the PES for that position. In this way, the arm of the disk drive is not constantly attempting to track an eccentric path and the PES effectively provides the servomechanism with only NRRO components of the disk motion. The methods of Nazarian and Sri-Jayantha are at least philosophically similar, although they differ in the details of the actual process by which runout data is gathered and subtracted from the PES.
Kermiche et al. (U.S. Pat. No. 6,611,396) teaches the use of a disk that is servo written off the spindle (ie. in an external servo writer) of a host disk drive DASD unit. In order to correct RRO for the disk when it is subsequently mounted in the host drive, a set of virtual tracks are defined on the disk by a set of intersections between the circular loci produced within the host drive and the physical tracks pre-written by the external servo writer. The virtual tracks are defined within storage as locations where the host read/write head intersects the servo wedges on the physical tracks.
It is a first object of this invention to make the writing of servo track data on disks used for testing read/write heads a more accurate, reproducible and efficient process and, thereby, to make the head testing process itself a more accurate, reproducible and efficient process.
It is a second object of this invention to achieve the first object without sacrificing the capability of that servo data to enable the accurate location of data tracks during the normal operation of the head testing device.
It is a third object of this invention to provide a method of eliminating the continual servomechanism corrections of repeatable runout (RRO) associated with a rotating disk mounted within a host head testing device when said disk has had its servo data written in a different device.
It is a fourth object of the present invention to improve servo accuracy in a head testing device by minimizing track following error associated with attempts by the head to follow an eccentric path.
It is a fifth object of the present invention to eliminate problems with mechanical hysteresis and resonance that occur when an actuator-mounted head requires rapid movements to follow track eccentricities that include the effects of repeatable runout.
It is a sixth object of the present invention to provide a less costly head testing device by eliminating the need for servo-writing components such as micro-positioners within the device itself.
The objects of the present invention are achieved by introducing a disk or a plurality of disks, each pre-written with servo data, into a head testing device and thereby eliminating the need for a servo data writing apparatus within the tester itself. The servo data is written onto the disk (or disks) using a dedicated servo writing apparatus, which, thereby, improves the quality of the servo tracks and, therefore, the accuracy, repeatability and efficiency of the head testing process. As a result, the servo-writing elements contained within the dedicated writer will no longer need to be made a part of the head tester, thereby significantly reducing the cost of the head tester as well as reducing the time required to complete the head testing process. The repeatable runout (RRO) associated with the use of pre-written disks is then effectively removed from the position error signal (PES) generated by the servomechanism within the tester, so that the RRO is effectively ignored and the actuator-mounted head does not follow an eccentric path. The RRO is made ignorable by the tester servomechanism by simply obtaining an excellent approximation to the RRO by averaging the runout values of selected servo track sectors (the differences between a fixed head location and a servo track location) while the actuator is operated at a low servo bandwidth. The average so obtained, which is an excellent approximation to the track runout, is then subtracted from the PES of the servomechanism while the servomechanism is operated at full bandwidth.
The objects, features and advantages of the present invention are understood within the context of the Description of the Preferred Embodiment, as set forth below. The Description of the Preferred Embodiment is understood within the context of the accompanying drawings, which form a material part of this disclosure, wherein:
The present invention teaches a method of operating a magnetic read/write head testing apparatus (commonly referred to as a “spin stand”) using disks on which servo track information has been externally pre-written in a dedicated servo track writer. The head testing apparatus, therefore, lacks the servo-writing mechanisms that are now a part of the dedicated servo-writer.
Referring first to
Referring now to
Referring to
(a): MOUNT DISK IN TESTER. The pre-written disk is mounted on the spindle of the head testing apparatus of
(b): SET HEAD TO A SINGLE SECTOR LOCKED POSITION AND SET SERVOMECHANISM TO A LOW BANDWIDTH CONDITION. The head within the testing device is made to follow a substantially circular track on the disk, concentric with the drive spindle, through the use of a single sector lock. The servomechanism is set to a low bandwidth condition, so that the head is in a relatively stable stationary state and exhibits minimal movement. The head is thereby able to take repeated samples of the position of a servo sector at the position of its intersection with a circular track that is concentric with the spindle of the head tester and to average those readings.
(c): REPEATEDLY SAMPLE POSITION OFFSET OF SECTOR. While the head is locked in this manner, the sector containing pre-written servo information moves in an eccentric path beneath the head and the position offsets of the sector location is repeatedly sampled in successive rotations of the disk. As many as 1000 rotations, on average, may be used. Each of these sampled offsets is the difference between the intersection position of the servo sector with the concentric circular track presently being followed by the fixed position of the head and the intersection position as identified by the positional data within the pre-written servo sector as produced by the external servo-track writer.
(d): COMPUTE SAMPLE AVERAGE. After a given number of offset samples are obtained, the average offset for the sector is calculated. This sample average provides an accurate indication of the eccentricity of the servo information and, thereby, a very good approximation to the RRO of the track itself. It is understood that the head testing apparatus includes a means, preferably a RAM within the servomechanism circuitry, but possibly also storage areas on the disk itself, for storing data such as these sector offsets. Because of the low bandwidth setting, the tester (i.e. the actuator and head) does not respond to the RRO or the NRRO. Since the NRRO will average to zero, the resulting average is the RRO.
(e): COMPARE POSITION OFFSETS OF BEGINNING AND END OF SAMPLING PROCESS. The servo offset at the beginning and end of the process is compared and the entire set of collected data is retained only if the beginning and end points of the process fall within a programmable amount of error. If the beginning and ending positions are too different from each other, then the head is presumed to have drifted too far during the revolutions of the disk. We only wish to retain sampled data that is obtained when the head is nearly stationary, because any significant drift will add to the observed RRO and invalidate the accuracy of the result.
(f): STORE AVERAGE SERVO OFFSET. To use the full bandwidth capability of the available servo sectors, the averaged servo offsets are stored in a random access memory (RAM) within the servomechanism circuitry of the head tester and used as a reference locus for the sector eccentricity.
(g): REPEAT (a)-(f) FOR MORE SECTORS IF DESIRED. The process can be used to sample the position offsets of any number of servo sectors, between a single such sector and the maximum number of such sectors formed using the dedicated servo track writer.
(h): SUBTRACT STORED AVERAGE OFFSETS FROM POSITION ERROR SIGNAL (PES) GENERATED DURING FULL BANDWIDTH OPERATION TO CREATE A CORRECTED PES AND USE CORRECTED PES FOR TRACKING AT FULL SERVO BANDWIDTH OPERATION. Servo control at full bandwidth is then applied to the actuator through a corrected PES sent to the DAC and voice coil motor. Because the averaged offset values are subtracted from the full bandwidth PES before the PES is sent to the DAC, the DAC will only correct deviations of target track locations from the averaged offsets. This requires that the DAC drive to the actuator be updated as every servo sector passes beneath the head so that the offset is removed from the signal and only sufficient current to correct for deviation of actual head position from the offset value is applied.
As is finally understood by a person skilled in the art, the preferred embodiment of the present invention is illustrative of the present invention rather than limiting of the present invention. Revisions and modifications may be made to methods, constructs, materials, structures and dimensions by which a head testing apparatus is operated without repeatable runout while using magnetic hard disks on which servo information
