Method of and apparatus for positioning a magnetic head

The present invention relates to a method and a device for positioning a magneto-resistant head and a memory disk of a device for positioning a magneto-resistant head in relation to a data carrier in data storage systems.

A data storage system is known which comprises a magnetic medium for storing data in magnetic form and a converter for reading and / or writing magnetic data on the data medium. For example, a disk storage device includes one or more data storage disks coaxially mounted on an axis of a rotary motor.

182 005 which spins the disks at a speed of several thousand revolutions per minute. Digital information, representing various types of data, is written to and read from the data storage disks by one or more transducers or read / write heads which are attached to an actuator and passed over the surface of rapidly rotating disks. The actuator includes a coil assembly and a plurality of outwardly extending arms having flexible suspensions with one or more transducers and sliders mounted on the suspensions. The suspensions are stacked with the rotating discs by means of an arm assembly attached to an actuator. The coil assembly interacts with the permanent magnet structure and responds to the controller. Also attached to the actuator is a motor with a voice coil VCM diametrically opposed to the arms of the actuator.

In a known digital data storage system, digital data is stored as magnetic transitions in a series of concentric, spaced-out tracks on the surface of magnetizable, rigid data storage disks. The tracks are divided into a plurality of sectors and each sector contains a number of information fields, one field of which is dedicated to data storage, while the other fields e.g. Data is moved and retrieved from specific track locations and sectors by transducers which follow a given track and move from track to track under servo control by the controller.

The head slider is constructed as a body that rises aerodynamically and lifts the head from the surface of the disk as the rotational speed of the rotary motor increases and causes the head to rise above the disk on an air bearing cushion produced by the rotation of the disk at high speed. The head-disk distance of 0.1 micrometers or less is referred to as the head-to-disk distance.

Writing data to a data storage disk causes a current to flow through the recording element of the transducer assembly to produce magnetic flux lines that magnetize a specific location on the disk surface. The reading of data from a specific disk location is accomplished by the reading element of a transducer unit sensing the magnetic flux field or lines originating from the magnetized disk locations. As the reading element passes over the surface of the rotating disc, the interaction between the reading element and the magnetized locations on the surface of the disc causes electrical signals to be generated in the reading element. The electrical signals correspond to the transitions of a magnetic field.

It is known, for example, from European Patent Specification No. 0 353 852 that friction between the magneto-resistant head and the surface of the data carrier causes temperature changes of the head, which cause disturbances in the read signal from the head.

Known data storage systems use a closed-loop control servo to position actuators and read / write converters at specific memory locations on the data storage disk. During normal operation of the data storage system, a servo sensor, mounted near the read / write converters or inserted as the reading element of the transducer, is used to read information to track a specific track and look for specific track and data sector locations on the disk.

A servo write procedure is known for initially writing pictorial information on the surface of one or more data storage disks. A servo writer is used in data storage systems to facilitate the transfer of image data to one or more data storage disks during the manufacturing process.

In servo engineering, embedded pictorial information is recorded on the disk along segments extending outward from the center of the disk. An embedded servo image is then formed between the data memory sectors of each track. The servosector contains a data image, often referred to as a pulse servo image, used to maintain the optimal alignment of read-write transducers over the centerline

182 005 track when reading and writing data in specific data sectors on a track. The servo information may also include sector and track identification codes which are used to identify the sector and track which are used to identify the transducer position. The built-in servo provides much higher track densities than the dedicated servo because servo information is co-located with the output information and servo information can be taken from one single surface of the disk.

A data carrier containing profiled servo information is known from the European patent specification No. 0 590 676. There is also a known solution for increasing the capacity of the disk, which uses a servo-information format called the embossed rigid magnetic disk technique. It is known from the publication of Tanaka et al. under the title "Characterization of the magnetization process of pre-embossed servo-pattern plastic hard drives", IEEE Transactions on Magnetics 4209, Vol. 30, No. 2, Nov. 1994, an extruded, rigid magnetic disk containing servo information in a plurality of servo zones arranged radially across the disk. Each servo zone is embossed with indentations and ridges to form an exact pattern, clock mark, and address code. The exact formula and address code are used to generate servo information signals. In order to produce servo signals, the directions of magnetizing the peaks and pits must be opposite. The magnetization process involves first magnetizing the entire disk in one direction using a strong magnet. Then, the write head is used to magnetize the raised areas in the opposite direction.

Although the use of an extruded rigid magnetic disk increases the disk capacity, servo information is injected onto the extruded rigid magnetic servo in a two-step magnetization process, greatly increasing the time required to write servo information to the disk. In addition, during the second step of the process, the servo information is not yet available on the disk, so an external positioning system is used, which complicates the write servo process. In addition, the extruded rigid magnetic disk technique is not very durable. An embossed rigid magnetic disk, like other embossed servo technology, still remembers information in the disk space that could otherwise be used to store data, with the result that the embossed rigid magnetic disk technique has not been widely applied.

The method according to the invention consists in inducing a thermal signal in the head by a data carrier and moving the head in response to the thermal signal.

Preferably, the thermal signal is separated from the read signal of the head.

Preferably, the thermal signal of the head is selectively applied during the dispensing.

Preferably, the feed is carried out by means of an adaptive inversion filter.

Preferably, moving the head is performed in response to a first output signal corresponding to a first frequency component of the thermal signal and in response to a second output signal corresponding to a second frequency component of the thermal signal.

Preferably, when moving the head, a first value representing the first and second outputs of the comparators is added to a second value representing the head axial offset and memory disk progression, and the sum of this addition is entered into the controller for head positioning.

Preferably, the read signal is mixed with the first oscillator frequency and the second oscillator frequency and the mixed read signal is filtered to obtain the first and second frequency components of the thermal signal.

Preferably, the first and second frequency components of the thermal signal are compared with the first and second threshold values ​​for producing the first and second output signals.

Preferably, the read signal is amplified, the amplified signal is compensated and the read signal is obtained from the compensated signal before mixing.

Preferably, the reading signal is amplified by a lever electronic module.

182 005

Preferably, the amplified signal is compensated by an HR inversion filter.

Preferably, the reading signal is obtained by means of a FIR lowpass filter shifting the mean value.

Preferably, when moving the head, the radial speed of the head is assessed relative to the required radial speed of the head, the difference between the estimated speed and the required speed is determined, and the head is moved in response to the determined difference.

Preferably, when evaluating the radial speed of the head relative to the required radial speed of the head, the time difference between the tracks containing the servo information of the memory disk is determined and the time difference is divided into a jump of the information track of the memory disk.

Preferably, when evaluating the radial velocity of the head, the estimated radial velocity for the course of the path is adjusted.

Preferably, when evaluating the radial speed of the head relative to the required radial speed of the head, the number of information tracks between the current track and the required track is calculated.

Preferably, the difference between the estimated speed and the required speed is amplified before moving the head.

Preferably, the head of the data storage system having at least one memory disk and at least one head is calibrated, and during the calibration, the head is moved to a memory disk calibration zone having a given surface profile, thermal signals are induced in the head by changes in the surface profile and determined characteristics of the head and memory disk using thermal signals.

Preferably, a first frequency surface profile and a second frequency surface profile are used in the calibration zone, and during calibration, thermal frequency signals are induced at the head through the first and second frequency profiles and the thermal threshold levels of the head are determined using the thermal frequency signals.

Preferably, thermal threshold levels are stored in the random access memory.

Preferably, when determining the characteristics of the head and the memory disk, the axial displacement of the head and the eccentricity of the memory disk are determined.

Preferably, during the determination, the absolute values ​​of the eccentricity of the memory disk are averaged over one revolution of the memory disk.

Preferably, the calibration steps are repeated for each head in the data storage, a reference position of one of the heads with respect to the physical track is zero for cylinder zero of the data storage chip, and heads other than the reference head are mapped to the physical tracks based on an axial offset of the reference head.

Preferably, the head with the greatest axial displacement is used as the reference head.

In the device according to the invention, a controller system for controlling the relative motion between the head and the data carrier is connected to the head, including a servo-modulator connected to the magneto-resistive element of the head, sensitive to the thermal response of the head.

Preferably the servodemodulator is connected to a seek controller which is connected to the speed profile driver connected to the first adder and the coupling generator is connected to the second adder, and a multiplier is connected between the two adders, the second adder is connected via the servo compensator to the D / A converter. attached to the drive and the second totalizer is connected via switches to the servo modulator, which is connected to the write read driver and is sensitive to the thermal response of the movable head with respect to the data carrier.

Preferably, the servo-modulator comprises a demodulation circuit for mixing the frequency signals of the thermal response.

Preferably, the data carrier comprises a memory disk.

182 005

Preferably, the memory disk includes sector markers and track marks containing servo information.

Preferably, the memory disk comprises information tracks separated by track marks containing servo information.

The memory disk of the invention comprises a plurality of relief information tracks, and each track has a profiled inner edge and a profiled outer edge, and these track edges are thermally distinguishable by the head.

Preferably, the edges of the tracks have serrations that occur at a first frequency and a second frequency different from the first frequency.

Preferably, the serration on the inner edge of one information track and the serration on the outer edge of adjacent tracks are of equal frequency.

Preferably, the first frequency is equal to twice the second frequency.

Preferably, the serrations are square.

Preferably, the serrations are sinusoidal.

Preferably, the disk includes track marks containing servo information having a topographic variation in the height of the memory disk.

Preferably, the first and second frequencies have a tooth cycle that is a power of 2.

Preferably the serrations are radially aligned.

Preferably, the disk comprises a head-readable Gray code calibration zone for determining the memory disk eccentricity and head axial displacement.

Preferably, the Gray code comprises recesses in the memory disk.

Preferably, the calibration zone comprises a dual frequency surface profile readable by the head to determine the thermal sensitivity of the head.

Preferably, the dual frequency surface profile has a first equivalent frequency to the first tooth frequency and a second equivalent frequency to the second tooth frequency.

Preferably, the dual frequency surface profile comprises radial grooves.

Preferably, the calibration zone is at least as wide as the eccentricity of the memory disk plus a maximum head alignment offset.

An advantage of the invention is that it enables the production of data storage systems with a servo information format that is economical and optimizes disk capacity.

The subject matter of the invention is illustrated in the exemplary embodiments in which Fig. 1 shows a top view of a data storage system with the upper housing cover removed, Fig. 2 - a side view of a data storage system comprising a plurality of data storage disks, Fig. 3 - in an enlarged side view, a data storage disk having various defects and surface features, and the thermal and magnetic response of the head to such defects and features, Fig. 4 - cross-sectional view of the transducer, shown positioned on the track above the centerline of the disc track, fig. 5 - top view of the disc according to the invention, fig. 6 - perspective view of two adjacent tracks on the disc, divided by a track marker, fig. 7 is a plan view of a series of disc paths and markers, FIG. 8 is plots of thermal frequency responses of the head as a function of the head position above the disc track, FIG. 9 - in a block diagram, a head reading signal demodulation circuit, Fig. 10 - in a general block diagram, a servo positioning circuit of the head using the thermal frequency response signals of the head according to the invention, Fig. 11A - segment of the calibration zone according to the invention, Fig. 1BB - in detail the segment of the calibration zone, Fig. 12 is a diagram of mapping heads to physical paths for a cylinder, 13 - in a block diagram a device for extracting a thermal signal from a read signal induced from the head, Fig. 14 - a read signal induced in the head having a disturbed DC baseline, Fig. 15 - the read signal of Fig. 14 having a restored DC baseline for processing by the signal splitter-modulation module, Fig. 16 - in the scheme a signal splitting-modulation module for extracting a thermal signal and a magnetic signal from the read signal induced in the head, Fig. 17A - thermal signal extracted from the read signal, induced in the head

182 005 at a specific track location, Fig. 17B - read signal obtained from the same track location after AC erase, Figs. 18A, 18B and 18C - read signal induced in the head, restored magnetic signal component for the read signal and non-restored magnetic signal component for read signal, fig. 19 and 20 - phase and amplitude response of FIR filter with finite impulse response and FIR filter with window, which are used in the signal splitting-restore module, Fig. 21 and 22 - conventional head, Fig. 23 - amplitude and phase response with high-pass filtering of a typical the lever electronic module, fig. 24 and - comparing the amplitude and phase response in highpass filtering of the lever electronic module and an inverse filter having an inverse transfer function to the effective highpass filter of the lever electronic module, Fig. 26 - waveform representing the inversion filter of Fig. 24 and Fig. 25, Fig. 27 - in the diagram, another embodiment of the signal splitting-restore module using the HR filter with infinite impulse response, fig. 28 (a) - 28 (c) - three waveforms produced at different processing points in the signal splitting-restore module of Fig. 27 and Fig. 29 - comparative plots of the restored magnetic signal and the thermal signal, indicating the presence of protuberances on the disk surface.

Figures 1 and 2 show data storage arrangement 20 with a cover (not shown) removed from base 22 of housing 21. Data storage arrangement 20 includes one or more rigid data storage disks 24 that rotate about an axis of rotary motor 26. Actuator 37 includes a plurality of actuator arms 30 stacked and each arm has one or more suspensions 28 and transducers 27. The converters 27 include a magneto-resistive reading element and a thin film magnetic head for writing information to data storage disks 24. The transducer 27 is, for example, a magneto-resistive head having a write element and a reading element. An actuator 37 including a coil assembly 36 cooperates with the permanent magnet structure to form a voice coil motor 39 responsive to control signals produced by the controller 58. The controller 58 includes a control circuit that coordinates the transfer of data to and from data storage disks 24 and cooperates with a motor 39 with a voice coil to move the actuating arms 30, suspensions 28 and transducers 27 to a given information track 50 and sector locations 52 when reading data from the disks. 24 and writing data to 24 disks.

Figure 3 shows an enlarged side view of the head shoe 79 floating near surface 24a of the magnetic data storage disk 24. Disc surface 24a has a topography that varies at the microscopic level and contains various surface defects such as a cavity 122, a ridge 124, or a surface gap 126 of the magnetic material. The thermal response of the head 80 varies as a function of the spacing y between the MR element 78 of the head 80 and the surface 24a of the disk. Variations in the head-disk spacing alter the heat transfer between the MR element 78 and the disc 24. This heat transfer causes the temperature of the MR element 78 to change. Changes in the temperature of the MR element 78 cause changes in the electrical resistance of the MR element 78, and therefore the output voltage of the MR element 78.

In the event that the instantaneous head-disk gap y increases, it increases the insulation of the air space between the head 80 and the disk surface 24a, causing an increase in the temperature of the magneto-resistive element 78. An increase in the temperature of the magneto-resistive element 78 causes an increase in the resistance of the head 80, related to the positive temperature coefficient of the material used to manufacture the magneto-resistive element 78. A preferred material for the fabrication of the magneto-resistive element 78 is, for example, permalloy, which has a temperature coefficient of +3 x 10 '.³/ ° C. For example, the head 80 extending over the protrusion 124 on the disk surface 24a causes the increased heat transfer between the magneto resistive member 78 and the disk surface 24a, thereby cooling the magneto-resistive member 78. Such cooling of the magneto-resistive member 78 reduces its resistance.

182 005 element, which in turn causes a voltage drop v_TH on the magneto-resistive element 78 at a constant polarization current.

With reference to the recess 122 shown in the disk surface 24a, it can be seen that the voltage thermal signal v_m 119 on the resistive element 78 increases the amplitude as a function of increasing the head-disk spacing y. Then, with reference to the relief 124 shown on the disk surface 24a, it is seen that the voltage thermal signal v_ra 119 reduces the amplitude as a function of decreasing head-to-disk spacing. The thermal component of the read signal is thus in fact an information signal that is used to detect the presence and relative value of topographic changes on the surface of the magnetic data storage disk 24.

Figure 3 also shows a spacing magnetic signal 121 that is adapted to respond to changes in disk surface 24a. The gap magnetic signal 121 incorrectly indicates the presence of certain surface features, such as surface gaps 126 and variations in disk surface topography 24a. The spacing magnetic signal 121 gives inferior indication of other surface features, such as protrusions, compared to the disk surface image information provided by the thermal signal 119.

The thermal component of the read signal of the resistive element 78 is extracted to obtain information regarding the surface properties of the rotating disk 24. The servo information is encoded in e.g. the surface profile of the disk 24 and is read using a transducer having e.g. Since servo information is provided in the disk profile and read simultaneously with magnetically stored data, an additional 15% -20% of the disk is available for data storage, i.e. the portion of the disk previously used for embedded magnetic servo information.

Figure 4 shows the resistive element 78 of the head 80 positioned above the centerline 51 of the information track 50. The head 80 is, for example, of the type used in conventional data storage systems. As the magnetoresistive element 78 passes over the information path 50 of the rotating disk 24, the magnetic transitions obtained on the disk surface 24 cause the generation of a signal induced in the head 80, which is preferably a voltage signal.

Figure 5 illustrates an exemplary disk 24 having pre-embedded or embossed track marks 108 containing servo information and sector marks 106 for providing servo information on the disk in the form of surface profile changes, e.g., head-to-disk spacing. Disc 24 is provided with concentric information tracks 50 used to store data. Each information track 50 is divided into a plurality of sectors 52 identified by sector tags 106. Adjacent track information 50 is separated by track markers 108 containing servo information. The servo information markers 108 and the sector markers 106 are formed as changes to the disc 24 that can be identified using the thermal component of the head reading signal.

Figures 6 and 7 show the track markers 108 including servo information circumferential grooves providing head-to-disk spacing variations between adjacent data tracks 50, and sector markers 106 are radial grooves providing variations of head-disk spacing between adjacent sectors 52. Track markers 108 including servo information. and sector tags 106 are used to provide servo information. Disk 24 is also provided with a calibration zone 110 and an index marker 112, which are formed by a closely spaced pair of sector markers 106. The purpose of the calibration zone 110 and index marker 112 will become clear later.

Figure 7 shows information tracks 50 on a disc 24 having serrated edges 50_id and 50_fromcorresponding to the inner and outer edges of the track. 50 inner edge serrations for each track_id has a different frequency of teething than that of the outer edge 50_from to provide servo radial direction information. The serrations are rectangular or sinusoidal. The teething frequencies have frequencies ή and f₂which differ by a factor of two, although the serrations can have many

182 005 different frequencies, assuming that each track has 50 edges_ID and 50q_D with different profiles. Serrated edges 50_ID and 50_FROM information tracks 50 may have an alternating frequency of teething, as shown in the figure. For example, some information tracks 50, referred to as odd tracks, have an inner edge 50_ID with a frequency of ή serration and an outer edge of 50_FROM with frequency f₂ serrations, while even information tracks 50 have outer edges 50_FROM with a frequency of f] serrations and 50 inner edges_id with frequency f₂ teething. By varying the teething frequency of adjacent information tracks 50, the teething edges of adjacent information tracks 50 correspond. This makes the track marks 108 containing servo information that separate adjacent information tracks 50 easier to detect and the disks 24 easier to manufacture.

Moreover, the serrations are radially oriented, i.e. the serrations are further apart when moving radially outward. This would be useful in a constant angular velocity system where the rates of teething relative to head 80 are constant over the entire surface of disk 24. Also, the number of teething cycles around information track 50 is a power of two such that the oscillator frequency of the data memory chip 20 is divided for precise determination. every frequency ή if₂ teething.

Figure 8 shows the frequency responses t (fj) and t (f₂) the thermal component of the readout signal of the head 80 as a function of the position of this head on even information path 50, i.e. information path 50 having the frequency f₂ at the inner edge 50_IDand frequency f, at the outer edge 50_FROM. When the head 80 is positioned over the center 51 of information path 50, the thermal frequency responses t (f 1) and t (f₂) should be close to zero. As head 80 moves towards edge 50_ω inner even track 50 senses edge serration and thermal signal t (f₂) increases. Likewise, as the head 80 moves toward the outer edge 50_FROM, the thermal signal t (fj) increases. As the head continues to move away along the path, the thermal signals i (f) and t (f₂) have horizontal characteristics. By examining the thermal frequency component of the read signal, the direction and amount of separation of the head 80 is determined, and an appropriate control signal is provided to position actuator 30 over the track centerline.

Thermal frequency responses t (f]) and t (f₂) from which the servo information is obtained are obtained from periodic changes in the spacing between the disc 24 and the head 80. The servo information is obtained from sensed changes in the thermal frequency response that result from the head 80 passing over the serrated circumferential grooves. Varying the disk-head spacing may be alternatively achieved by providing raised portions between the tracks. However, in a typical environment, grooves are preferred because they allow the data storage device 20 to operate with a minimum spacing between heads and data tracks. Servo information is also obtained from other changes in the characteristics of the disk, which are reflected in the thermal component of the read signal. For example, the servo information track markers 108 may differ from the information tracks 50 as to thermal emissivity or other parameters that are reflected in the thermal component. Similar changes to the disk characteristics are applied to the markers 106 of the sectors.

Figure 9 shows a block diagram of a read signal demodulation and generation of servo-positioning outputs 221 and 223. While the demodulation circuit shown in Figure 9 is made as a digital circuit, an equivalent analog circuit is also used. In operation, the demodulation circuit amplifies the read signal 81 obtained using the head 80, and the amplification is performed, for example, using a lever electronics module 202. The amplified read signal 203 is sampled by the sampling circuit 204 at a sampling rate to produce read signal 205. A typical sampling rate is over 100 MHz. The sampled read signal 205 is provided to the data channel for normal processing, and also provided to the track intersection detector 230 for the purpose described later. Readout 205 is also provided to an inverse filter 206, e.g.

182 005 in the lever electronic module 202. The output 207 of the inversion filter 206 is passed through a filter 208, e.g., a moving average finite FIR filter, to recover the thermal component of the reading signal. The output of the filter 208 is coupled to the mixing demodulation circuit 210, which determines whether thermal frequency components ¢ (^) and t (f₂) of the reading signal 81 exceed the respective threshold values ​​t_and it_b. Operation of the agitation demodulation circuit 210 is further explained below.

During a read operation, the read signal 209 received by the agitation demodulation circuit 210 is supplied directly from the filter 208. However, during the write operation, the writer generates heat that is transferred to the magnetoresistive head. The transfer of heat from the write head to the magnetoresistive head distorts the thermal component of the read signal. The dynamics of heat transfer from the write head to the magnetoresistive head is approximated by the transfer function of the first order low pass filter. Interference caused by heat transfer from the write head to the magnetoresistive head is substantially reduced by passing the output of the filter 208 through an adaptive inversion filter 240 having a transfer function inverse to the transfer function of the lowpass filter. A signal 233 indicating the read and write operation is applied to selectively coupling the adaptive inversion filter 240. For example, during a write operation, signal 233 is applied to the adaptive inversion filter 240 coupling such that the signal leaving filter 208 is passed through the adaptive inversion filter 240 before being received. by a mixing demodulation circuit 210. During the read operation, signal 233 is used to pass the output of filter 208 directly to the mixing demodulation circuit 210.

The polar position of the transfer function of the adaptive inversion filter 240 varies for different magneto-resistive heads. The polar position for a given magneto-resistive head is evaluated using the thermal cooling response of the magneto-resistive head after the start of the write operation. For example, the cooling curve is a simple exponential decay function with a time constant that is determined from the time it takes for the cooling curve to reach approximately 36.8%, i.e. e '= 0.368 of the maximum value. The time constant or polar position value, which is the reciprocal of the time constant, is evaluated on site for each magneto head and stored in random access RAM. The transfer function of the adaptive inversion filter 240 is then updated based on the time constant of the magnetoresistance head that is selected. The transfer function of the adaptive inversion filter 240 for each MR head is updated with time to count changes in heat transfer dynamics from the write head to the MR head and from the MR head to the disk caused by contaminants accumulating on the heads.

The adaptive inversion filter 240 restores the disturbed thermal component to one that would be present during a read operation. The use of a read signal having a thermal component restored by the adaptive inversion filter 240 as an input to the hash demodulation circuit 210 reduces track misregistration during writing. In addition, the continuous nature of servo-positioning sensing, that is, sensing during both read and write operations, eliminates the need for accelerometers or other external sensors to control shock and vibration in the drive.

The mixing demodulation circuit 210 includes first and second multipliers 212 and 214, filters 216-218, and first and second comparators 220 and 222. The mixing demodulation circuit 210 produces signals t (f) and t (f).₂) of the thermal frequency response from read signal 209 and compares these signals with threshold values ​​t_and it_b to produce servo positioning control outputs 221 and 223.

In operation, the mixing demodulation circuit 210 receives the read signal 209 and supplies it to the first and second multipliers 212 and 214. The multipliers 212 and 214 multiply the read signal 209 by two or more divided oscillator signals 225 and 227. The oscillator signals 225 and 227 have waveforms and frequencies similar to frequencies

182 005 edge serrations as the disks 24 rotate at rated speed. The multiplication circuits 212, 214 output read signals 213 and 215 which have amplified frequency components with frequencies ή and f₂. Read signals 213 and 215 are lowpass filtered by filters 216 and 218, thereby rejecting the high frequency components of read signals 213 and 215 to produce low frequency thermal response signals 217 and 219 ί (ή) and t (f₂). Thermal signals 217 and 219 are provided to comparators 220 and 222 for comparison with threshold values ​​t_and it ,,.

Figure 8 shows that the t thresholds_and it_b correspond to the threshold amplitudes of thermal signals t (f,) and t (f₂) at which the head 80 has moved away from the path and should be repositioned. The threshold values ​​for t_and it_b are preset values ​​for each head 80 and are stored in random access RAM. The threshold values ​​for t_and it_b are determined taking into account differences in heat sensitivity along the width W of the resistive element of different heads 80 and may be different for different heads.

The outputs 221 and 223 of comparators 220 and 222 may be logic 0 or 1. When the head 80 is centered over the data track 50, both outputs 221, 223 are, for example, logic 0. If the thermal frequency response signal 217 exceeds the threshold t_andthen comparator output 221 is logic 1. Similarly, if signal 219 exceeds threshold L, comparator output signal is logic 1. If signals 217 and 219 have frequency components with frequencies f or f₂ When serrations below each threshold value, then both comparator outputs 221 and 223 will be zero. The comparator outputs 221 and 223 are used by a path tracking control servo system to control the position of the head 80.

Figure 10 shows a control servo system using outputs 221 and 223 of comparators to position the head 80 on an information path 50. The control servo system includes a servo-modulator 200, a write-read controller 340, a seek controller 302, a speed profile controller 304, a coupling generator 306, first and second. adders 308 and 310, multiplier circuit 312, servo compensator 318, digital-to-analog converter 314, and drive 316. In operation, the read-write controller 340 provides a signal 341 to the control servo system, e.g., servododulator 200. Signal 341 is used to control the type of operation or the mode of operation, e.g., read, write or track track, search or set. For example, in tracking a path, the outputs 221 and 223 of the comparators control the switches 320 and 322 to input the digital values ​​+ I₀ and -I_about to adder 310. When comparator output 221 takes a logic value of 1, switch 320 closes so that adder 310 is supplied with + I₀. Likewise, the value of -I_about is provided to adder 310 when comparator output 223 is logic 1. Digital values ​​+ I₀ and -I_about are provided as pulse injection values ​​for controlling the movement of the actuator. If both comparator outputs 221 and 223 are zero, both switches 320 and 322 remain open and no pulse injection value is provided to the adder 310. Note that if both comparator outputs are logic 1, an error occurs and the signals occur. comparator output can be ignored.

An adder 310 adds the impulse injection value + I₀ and -I_aboutif present, with the value 307 of the coupling generator 306 and provides the summed signal 311 to the servo compensator 318. While the adder 310 receives the signal 313, this signal may or may not be provided in path tracking operations as it is mainly used in path finding such as this will be described below. The value 307 represents the axial offset of the head and the path course as a function of the rotation of the disk 24 and is provided to the adder 310 by the coupling generator 306. The coupling generator 306 remembers the predetermined axial offset of the head 80 and the path course for each head 80 in the random access RAM of the storage device. . The path course and offset of each head 80 are determined using the following calibration procedure.

182 005

A servo compensator 318, typically a microprocessor, processes signal 311 and given the type of information path 50, be it odd or even, produces a servo positioning control signal 319. The control signal 319 is converted to an analog signal 315 by a digital to analog converter 314 and supplied to the drive 316, which in response provides a current 317 to the voice coil motor 39 to move actuator 30. In this way, the movement of the actuator 30 is controlled such that the head 80 follows the given information path 50.

The servo control system is also used for track search, i.e., moving the head 80 to a position above the predetermined information track 50 on the disk 24. In a track search operation, the control servo system uses the speed difference between the estimated head radial speed 80 and the required radial speed to produce a radial speed control signal. actuating arm 30 and therefore head 80. For example, if the estimated speed is less than the desired speed, servo compensator 318 produces a signal to increase the radial speed of the head 80 by increasing the rotation speed of the actuator arm 30. Conversely, if the required speed of the head 80 is less than the estimated radial speed of the head 80, the servo compensator 318 generates a signal to decrease the radial speed of the actuator arm 30. radial speed of actuating arm 30.

The lookup controller 302 produces an estimated head speed EV signal 303, which is subtracted by adder 308 from the required head speed DV signal 305 produced by the speed profile controller 304. The speed difference signal 309 DV-EV is multiplied by the speed gain + K or -K. in a multiplier circuit 312, and a product 313 is provided to adder 310. Search controller 302 may also provide a search direction signal 300 for use by a multiplier 312 to determine whether a speed gain -K or + K is used. The adder 310 adds the speed difference signal 313 to the value 307 of the coupling generator 306 and provides the sum value 311 to the servo compensator 318 which processes the value 311 and produces the servo positioning control signal 319. The control signal 319 is converted to an analog signal 315 by the DAC 314 and supplied to the drive 316. The drive 316 produces a current 317 for the motor 39 with the voice coil of the actuator 37, increasing or decreasing the radial speed of the head 80 so that the head 80 can efficiently search and be positioned over the required location.

The seek controller 302 produces the judged head speed by measuring the time between gaps in the magnetic component of the read signal 81 that occur as the head 80 passes over sector markers 106, and dividing this time into fixed track jumps using oscillator pulses. Speed ​​estimation is further refined by compensation for path travel through head 80. Such further refinement is made at low head speeds, such as when the head 80 is positioned on a specific information track 50. The estimated speed may also be determined using the thermal component peaks, which, like magnetic gaps, if disk 24 is magnetically recorded. appear above 106 sector markers. However, the magnetic components provide a higher signal-to-noise ratio.

In addition to evaluating the speed of the head 80, the lookup controller 302 maintains a record of the number of the current data path over which the head 80 is positioned and also counts data path intersections. Data path intersections are counted in a path intersection detector 230 as shown in Fig. 9, by amplitude detection of thermal peaks or gaps in the magnetic component in the read signal that occurs as the sector marks 106 are traversed across the write head. Each time the head crosses sector marker 106, a pulse is generated on line 201 at the output of track intersection detector 230. Search controller 302 uses the current path number and path count to determine the difference between the number of the current data path over which the head 80 is located and the target number of data path that the head 80 is searching for. This track number difference is referred to as the TTG number of tracks to be traversed and is provided by the search controller 302 to the speed profile driver 304.

182 005

During the search operation, the lookup controller 302 counts the path intersection pulses occurring on line 201 and continues to update the TTG count signal 353 of the number of TTG paths to be traversed. When the TTG signal reaches zero, that is, as the head 80 approaches the target path, the radial velocity decreases sharply and the control of the actuator may switch to a set-up mode in which the outputs 221 and 223 of the path tracking demodulators may be used in conjunction with the signals at the outputs of the detectors. path intersection to evaluate speed. The setting mode switch may be toggled if the TTG is below a count number, such as TTG <1-2 tracks.

The speed profile controller 304 receives the signal TTG 353 and uses it to determine the required head speed 80. The required speed is determined based on the relationship between the required head speed and the signal TTG. This relationship is predetermined and can be stored in a lookup table in RAM attached to the speed profile controller 304.

The servo control system provides continuous servo control as the thermal component of the head 80 is sampled simultaneously with the magnetic component of the signal. Unlike traditional magnetic servo, thermal servo can achieve sampling rates in excess of 100kHz. Higher sampling rates for thermal servo techniques provide lower actuator coil currents. Moreover, due to the mechanical inertia of the actuator, these small currents are integrated to produce very smooth movements, thereby reducing actuator jerks. More accurate control of the actuator allows for higher track densities of approximately 6300 track per cm for the dual actuator, and greatly improves the detection of shock and vibration of the disk 24. In addition, for predictive fault analysis, any defect is detected on the disk 24 that could potentially produce track misregistration. Likewise, any random modulation of the head 80 is detected, for example a modulation caused by the head 80 not rising from the disk 24.

The use of prebuilt disks 24 requires that servo information be generated on the disk prior to assembly in the disk drive. When multiple disks are used, consideration should be given to adjusting the servo information for different disks. For example, due to the mechanical tolerances in manufacturing the disks 24 and mounting the disks 24 in the storage device, each disk 24 in the disk stack is likely to have a different eccentricity or course. In addition, mechanical tolerances cause axial shifts between the heads 80. Thus, a calibration procedure is needed to determine a particular disc path, head axial shifts, and rotational position adjustment differences with respect to radial index marks 112.

In order to calibrate and format the disks 24 in the storage device, each disk 24 is provided with an annular calibration zone 110 which is positioned anywhere on disk 24, for example at the outer diameter stop of the disk 24 or the inner diameter stop of the disk 24. The calibration zone 110 has a width at least as large as the worst case eccentricity of the disc 24 plus the maximum axial displacement of the head. An example width is 32 tracks.

Figures 11A and 1 IB show that the calibration zone 110 consists of a repeating Gray code pattern that may be evenly spaced and have six to eight bits. The Gray pattern alternates with the blank area or the surface profiles 384 and 386. Gray code pattern 382 and surface profiles 384 and 386 consist of variations in disk characteristics that are heat readable. In the example shown, Gray code pattern 382 is encoded as embedded recesses, and the surface profiles 384 and 386 include radial grooves. The head 80 thermally responds to a recess or groove with a signal in the form of a short positive pulse.

The profiles shown in Figures 11A and 11B have a linear density across the disk so that their frequency at the rated disk speed is commensurate with the frequency range of the thermal signal. For example, while magnetic frequencies are

182 005 of the order of 100-200 MHz, the thermal frequencies are much lower, for example profiles can have a frequency of 100-200 kHz.

Figure HA shows an exemplary calibration zone segment 110 that shows eight paths n-2, n-1, n, n + 1, n + 2, ..., n + 5, four bit Gray code {b₃, b₂, b_n b₀} and profiles 384, 386 surfaces. Gray code, which is stored at track width increments and an offset of path width%, is denoted by ... j-2, j-1, j + 1, .... In Figure HA, head 80 is located in path n + 2 of the calibration zone. As the head traverses the (j + 4) Gray code / 2 paths, it produces two short positive pulses {b₃, b₂}. If this head portion passes through the (j + 3) Gray code '/ 2 paths, a smaller short burst {b₀}. Short pulses are compared to the threshold and if the threshold is exceeded a logic 1 may be output, otherwise the output is logic 0. For example, if only short pulses {b₃, b₂} exceed the threshold value, {1, 1, 0, 0} will be decoded.

Gray code 382 is processed to determine the axial and relative head offset and track trajectory for each head 80 in the storage device. The head offset and track waveforms can be stored in RAM connected to the feedforward generator 302 and used in the servo control system for track tracking and track search. The axial offset of the head 80 is determined by rotating the disk 24 one complete revolution and the reading of the Gray code pattern 382 by the head 80 followed by signal averaging. A heat Gray code signal is formed in the read signal of the head 80 as it passes over the pits and peaks of the Gray code pattern 382. The heat Gray code signal is processed to determine the absolute average measurements of the waveform or eccentricity of the disk. The head axial displacement is determined by averaging the absolute measurements of the disk eccentricity over the total revolution. By doing this, the position of each head 80 relative to the other heads 80 can be determined.

The head axial movements may be used in multiple disk systems to define the physical path of each head 80 in the cylinder. The cylinder represents the physical paths over which the heads 80 lie. In the exemplary arrangement, the zero physical path of the head 80 with the greatest axial displacement is mapped to cylinder zero. The heads 80 with less axial displacement are mapped to larger physical path numbers based on the greatest axial displacement.

Figure 12 shows an example mapping for a memory system having three disks and six heads 80. The continuous circles represent the position of each head 80 in the calibration zone 110. The distance between the continuous circles and the horizontal axis represents the axial offset of each head 80 from the physical path zero. Since head 80 number 5 has the greatest axial offset of 5.5 tracks, it is mapped to physical track zero. The remaining heads 80 are mapped to physical paths that are 5.5 paths apart from their positions in the calibration zone 110. By mapping cylinder zero to the physical path zero of head 80 with the greatest axial displacement, the search required when switching heads is minimized.

Profiles 384 and 386 of the surface are used to measure the thermal sensitivity of the 80 heads to determine the appropriate t threshold servo levels_and it_b for a given head. Servo threshold levels t_and it_b compensate for the different thermal sensitivities between the heads 80 resulting from differences in dimensions, retraction, magneto-resistive screen distance, and thermal conduction of the leads, etc.

In order to determine the t threshold servo levels_and it_b for each head 80 in the memory system, surface profiles 384 and 386 have serration frequencies that may, for example, be equal to the frequencies ή and f₂ edge serrations 50_ro and 50_FROM. The threshold level calibration procedure for a multiple disk device includes the following steps. First, all heads 80 of the memory system are moved to the calibration zone 110 of the individual disk 24. The head 80 is selected and the individual disk 24 is rotated. Thermal frequency responses t (f) and t (f₂) are produced as the head 80 extends over the surface profiles 384 and 386. Threshold levels t_and and L are computed using

182 005 thermal frequency responses t (f) and t (f₂). In Fig. 8, the ratio of the threshold levels to the amplitudes of the horizontal characteristics can be constant assuming that the thermal sensitivity along the width W of the magnetoresistive element is constant or the threshold levels are determined from the relationships: t_and = k ^ fj and tb = k_bt (f₂). Threshold levels t_and and for the selected head 80 are stored in RAM and the process is repeated for the other head 80 until the threshold values ​​for all heads 80 in the memory system have been calculated.

An exemplary calibration procedure for determining the position of all index marks 112 on the surfaces of the disk 24 in the disk assembly has the following steps. The index mark 112 of the reference disc 24 is identified and the time difference including the head switching time between the index mark 112 of the reference disc 24 and the index mark 112 of the second disc 24 is measured. This is repeated until a time difference is established between all index marks 112 and therefore sectors 52. A relative calibration of the timing is performed simultaneously with the calibration of the thermal signal.

Figure 13 shows an apparatus for reading an information signal having magnetic and thermal components from a magnetic data carrier and separating thermal and magnetic components from the information signal. The head 80 is shown in close proximity to the surface of the data storage disk 24. The read signal induced in the head 80 is amplified by the lever electronics 202. The reading signal filtering by the lever electronics 202 is also performed. Graphically, at the output of lever electronics 202, the analog read signal 460, including a relatively high frequency magnetic component 46la, has a disturbed DC baseline due to the presence of a low frequency modulation component. Modulated read signal 460 or more specifically modulated magnetic component 46la of read signal 460 has long been identified as the source of many disadvantages of a disk storage system, including servo errors and inaccuracies, data storage reduction and data recovery reliability, and in some cases irreversible data loss.

Read signal 460 is a composite signal having independent magnetic and thermal components, and the baseline of the low-frequency modulating read signal is actually the independent thermal component of the read signal 460. Unwanted modulation of the read signal 460 is eliminated or substantially reduced in value, providing a pure magnetic signal. representing data or servo information.

Figures 14 and 15 are sequentially the noisy read signal and an undisturbed read signal restored by the signal splitting-restore module 476 as shown in Figure 13. The signal splitting-restore module 476 processes the read signal 460 to restore the baseline of the read signal as shown in Fig. 15, by eliminating undesired baseline modulation, thereby producing a clean, undistorted magnetic signal 46lb. Signal splitting-restoring module 476 represents a device filtering the read signal of the servododulator 200 of FIG. 9, needed to extract the thermal signal from the read signal.

Figure 17 shows the waveforms determining the independence of the magnetic signal and the thermal signal. The waveform shown in Fig. 17A shows the thermal signal extracted from the composite read signal using a magnetoresistance head and a digital filter constructed as a low pass filter. Upon receipt of the waveform shown in Fig. 17A, the path through which the waveform was produced was subjected to AC erasure. The same magneto head was moved to the same location of the erased path to obtain the waveform shown in Fig. 17B. The extracted heat signal of Fig. 17A and the read signal obtained from the erased track shown in Fig. 17B are substantially identical. The two waveforms shown in Fig. 17 confirm that the two simultaneously read thermal and magnetic signals are independent and separable.

Figure 16 shows an embodiment of the signal splitting-restore module 476 previously discussed with respect to Figure 13. The signal splitting-restore module 476 is used to perform the sole task of separating independently.

182 005 of the magnetic signal from read signal 460 to remove the low-frequency modulation component of read signal 460, contributing to thermal signal effects. In another embodiment, the signal splitting-restore module 476 is used to accomplish the dual task of separating the magnetic component from the read signal 460 to remove the low-frequency thermal component and otherwise extract the thermal signal, making both pure magnetic and pure magnetic signals available for the following processing. pure heat signal in independent form.

Figure 16 shows that the read signal is sensed by a head 80 located in close proximity to the magnetic data storage disk 24. The read signal received from the lever electronics 202 from the head 80 is converted from analog to digital form by an analog-to-digital converter 204. The digital read signal is then routed to delay device 486 and to programmable filter 488. Programmable filter 488 is a finite impulse response FIR filter having a length N, where N represents the number of impulse response coefficients or taps of programmable filter 488. The read signal provided to the input of programmable filter 488 is subjected to a complete signal delay corresponding to length N of programmable filter 488 when the read signal passes through programmable filter 488.

The programmable filter 488 is programmed with the proper tap factors and weighting factors such that it passes the thermal component at a relatively low frequency of the read signal and filters out the magnetic component at a relatively high frequency. Programmable filter 488 is designed as a low pass filter and programmed to pass a thermal signal that is characterized as a medium frequency signal with most energy in the frequency range of approximately 10 kHz to approximately 100-200 kHz. The magnetic component of the reading signal has a frequency in the range between approximately 20 MHz and 100 MHz. The thermal signal 480 at the output of programmable filter 488 is applied to signal summing device 490. From the output of programmable filter 488, the thermal signal 480 may be sent to other components in the data storage system, such as servo control, for controlling track tracking and track searching operations.

Delay device 486 receives read signal 460 from analog-to-digital converter 204 and delays transmission of the read signal to signal combiner 490 for a duration equivalent to the delay time required for read signal 460 to pass through programmable filter 488. The read signal 460, including both magnetic and thermal components, and the thermal signal 480, separated from the read signal by the programmable filter 488, arrive at the summing device 490 at substantially the same time. The signal combiner 490 performs a demodulation operation of the read signal 460 and the heat signal 480 to produce a restored read signal 478. Thus, the signal splitting-restore module 476, illustrated in the embodiment shown in Fig. 16, provides separation of the magnetic and thermal components of the composite read signal and further produces an undisturbed, restored magnetic read signal 478.

In Figs. 14 and 15, the modulated read signal 460 of Fig. 14 represents the occurrence of a read signal prior to signal splitting-restore module 476 processing. The read signal in Fig. 15 shows the read signal of Fig. 14 after the signal splitting-restore module 476 has processed the signal. The undesired influence of the thermal component of the noisy reading signal shown in Fig. 14 has been eliminated by the use of a 9-tap FIR filter in the signal splitting-restore module 476 to produce the restored magnetic read signal 478 of Fig. 15.

Figure 19 shows the amplitude and phase characteristics of a 9-tap FIR filter used to produce the restored magnetic read signal 478 shown in Figure 15.

182 005

In Fig. 19B, the 9-tap filter has an excellent linear phase response over the tested frequency range. The efficiency of a 9-tap FIR filter in eliminating baseline shift or read signal modulation is shown in Fig. 18.

Figure 18A shows a read signal representing a baseline unstable or time-varying. In Fig. 18B, the modulation baseline of the read signal shown in Fig. 18A is eliminated after the noisy read signal has passed through a properly programmed 9-tap FIR filter. The tap weighting factors for the 9-tap filter used to restore the baseline of the read signal are defined to include the tap weighting factors:

B (i) = (l / 9) * (-1, -1, -1, -1, 8, -1, -1, -1, -1), or

B (i) = (-.111, -. 111, -. 111, -111, -. 889, -. 111, -. 111, -. 111, -. 111)

The waveform shown in Fig. 18C was generated by passing the modulated read signal shown in Fig. 18A through a conventional Butterworth highpass filter, which is a monopolar highpass filter. The undesired modulation baseline of the read signal is not significantly reduced in value after the read signal has passed the conventional high pass filter.

Figures 19A and 19B show the amplitude and phase characteristics of a 9-tap FIR filter used to restore the baseline of the read signal as shown in Figure 18B. Fig. 19A shows that there is some degree of ripple in the filter passband which is eliminated by providing a window function for the tap weighting factors of the 9 pin FIR filter. For example, a Hamming window may be used for tap weighting factors of a 9-tab FIR filter to produce a restore with a filter window having the following tap weighting factors:

b (i) = (-.089, -., 0239, -. 06, -. 0961, -. 8889, -. 0961, -. 06, -. 0239, -. 0089)

Figure 20A shows the output of a 9-tap window FIR filter having tap weighting factors that eliminate ripple. In Fig. 20B, the 9-tap window FIR filter maintains an excellent linear phase response. Applying a window function, such as a Hamming window, to the tap weighting factors of the programmable FIR filter 488 allows non-zero DC gain and some increase in the low frequency response.

Figures 23-29 show another embodiment of the signal splitting-restoring module 476. In the lever electronic module 202, as shown in FIG. 13, it is often desirable to include a high-pass filter in conjunction with the preamplifier in order to discard the relatively low frequency component from the read signal produced by the head 80. The high-pass filter of the lever electronic module 202 disturbs both the amplitude and the phase of the thermal component of the composite reading signal. The magnitude of the thermal signal disturbance due to the high pass filter varies considerably depending on the frequency and phase response of the particular high pass filter used.

For example, a high-pass filter for use in a lever electronic module 202 may have a cutoff frequency of approximately 500 kHz and exhibit a non-linear phase response. The frequencies associated with the head-to-disk gap vary, but are in the range below 200 kHz. Moreover, the thermal signal of the read signal has a frequency in the range between 10 kHz to approximately 100 kHz. A high-pass filter, having a cut-off frequency of approximately 500 kHz, significantly distorts the amplitude and phase of the thermal component of the read signal. However, the magnetic component of the read signal remains unaffected by the high pass filter because the frequency range of the magnetic signal is approximately 20 to 40 times the cutoff frequency range of the high pass filter.

Figures 23 (a) and 23 (b) are sequentially shown graphs showing the amplitude and phase response of the highpass filter for a conventional lever electronics module 202. The highpass filter has a cutoff frequency of approximately 500kHz. Function

182 005 high-pass filter, having one pole at 500 kHz and an amplitude response! phase shown in Fig. 23, can be defined as:

_n _ b_k(\) + b_h(2) -z ~^NS_(1] + ^(2)-7-¹ where: b_h(l) = 0.9876, b_h(2) = 0.9876 a_h(2) = 0,9752.

The amplitude and phase distortion of the thermal signal introduced by the high pass filter of the lever electronic module 202 is effectively eliminated by using an inversion filter having a transfer function inverse to that of the high pass filter. Passing the readout output from the lever electronics 202 through the inversion filter restores the thermal signal to its original form, both in amplitude and phase. For example, the inversion filter transfer function for determining the state of the reading signal passed through the high-pass filter having the above-described transfer function from equation [1] is:

H ~ '= 1 + ^ (2) -7-¹

Ą (l) + Ą (2) · ^

Figures 24 and 25 show the amplitude and phase responses of the highpass filter of the lever electronics module 202 and the inversion filter. In particular, the amplitude response of the inversion filter and highpass filter of the lever electronic module 202 is sequentially shown as curve 570 and 572 in Fig. 24. The phase response of the inversion filter and highpass filter is shown as curve 576 and 574, respectively.

The IIR infinite impulse response filter is programmed to respond as an inverse filter to restore the thermal signal of the highpass-filtered reading signal. Although an analogue filter is used in a different embodiment, the IIR filter has many advantages when used as an inverse filter to restore the amplitude and phase of the thermal signal disturbed by the highpass filter characteristic of the lever electronic module 202.

Figure 26 shows the signal waveform of a first order IIR filter constructed as an inverse filter. The coefficients associated with the signal of Fig. 26 for a first order inverse IIR filter having a transfer function given by equation [2] above are: a! = 0.9876 and? = 0.9876, b] = 0.1 and b₂ = 0,9752.

Figure 28 shows three waveforms that determine the efficiency of the inversion filter to restore the original amplitude and thermal phase of the read signal that was passed through the high-pass filter. In Fig. 28A, a read signal detected from a recess in the surface of the data storage disk is shown. The read signal, shown in Fig. 28A, is detected from a track recorded at a write frequency of 20 MHz. The read signal is sampled at 100 MHz and 8 bit resolution. The graph in Fig. 28B is the calculated peak-to-peak value of the reading signal of Fig. 28A. The signal shown in Fig. 28B accordingly represents a magnetic spacing signal 560 which clearly shows the loss of the magnetic signal due to the passage of the magneto-resistive reading element over the cavity. Figure 28C shows the thermal component of the read signal after passing through the highpass filter 550 of the lever electronics module 202. By comparing the waveforms of Figures 28B and 28C, it can be seen that the magnetic gap information and the thermal gap information do not closely correspond due to thermal component distortion caused by by a high-pass filter 550 that differentiates the heat signal.

In Figure 29, thermal gap signal 562 processed by inverse filter 556 and digital filter 558 is shown with magnetic gap signal 560 passed through digital filter 552 and log element 554. Linearized magnetic gap signal 560 is computed by log-peak signal and then multiplying by the known sensitivity the change in output voltage relative to the change in interval

182 005 magnetic according to the well-known Wallace equation. In FIG. 29, in addition to the signal height difference and the slightly larger time constant associated with the thermal gap signal 562, the magnetic gap signal 560 and the thermal gap signal 562 describe a recess in the surface of the disk. Thus, the effect of integrating through the inversion filter 556 of the disturbed thermal signal shown in FIG. 28C ensures that the correct thermal gap signal 562 is produced.

Figure 27 is a block diagram showing a read signal processing system for obtaining magnetic and thermal head-to-disk distance information. The read signal is detected from the disk surface 24 by the head 80. It is assumed that the read signal is a composite signal including both magnetic and thermal components. The read signal detected by the head 80 is sent to the lever electronics 202 and then to the highpass filter 550. The high pass filter 550 is shown external to the lever electronics 202. However, in practice, the high pass filter 550 is inserted into the lever electronics 202 and the transfer function H is denoted._about high-pass filter.

The output of the highpass filter 550 is sampled by an A / D converter 551 to produce digital samples of the highpass filtered read signal. The digital read signal is then sent to the inversion filter 556 which corrects for noise introduced by the high pass filter 150 of the lever electronics module 202. The transfer function of the inversion filter 556 is denoted by H._about'. The average value of the signal passed through the inverse filter 556 is obtained by digital filtering using the digital filter 558 to produce a thermal signal that is linearly related to the head-to-disk separation.

The read signal provided at the output of analog-to-digital converter 551 may also be applied to a digital filter 552, such as an FIR filter that extracts the peak-to-peak amplitude of the read signal to extract a magnetic component from the read signal. The log of the magnetic signal is obtained by passing the magnetic signal through a log element 554 which produces a magnetic signal linearly related to the head-to-disk gap. After extracting both the magnetic and thermal spacing signals 560 and 562, the thermal signal can be calibrated because the magnetic calibration is known and depends only on the recorded wavelength of the signal. Both the magnetic and thermal spacing signals 560 and 562 are linearly proportional to the head-disk spacing y.

Figures 21 and 22 show the general arrangement of the basic elements in a typical combined head 600. The elements are not drawn to scale, but rather indicated to show the relative positioning of the various elements of the head including the pair of screens 601 and 603. Magneto resistive element 602 is disposed between screens 601 and 603 and acts as the reading element of the head 600.

Screen 603 in conjunction with element 604 forms a thin film magnetic head acting as a write element for head 600. Screen 603 and element 604 act as first and second magnetic poles of the thin film recording element. The dual function of the shield 603, that is, acting as the first pole of the recorder and as the second shield, gives the combined properties of the head 600. Insulating layers (not shown for simplicity), such as glass, are formed between the various components of the head 600.

In Figures 21 and 22, a first shield 601, a magneto-resistive element 602, and a second shield 603 extend upwardly from the surface 501A of the disk 501 in individual vertical planes. The second pole element 604 is not shown in Fig. 21 for the sake of simplicity. Part planes are shown parallel to the page plane. In the drawing, the plane of the first-pole-second shield 603 is the closest, followed by the magnetoresistive element 602, and the first shield 601 is the farthest. Also shown is the negative and positive leads 701A and 70IB formed in the plane between the first shield 601 and the first pole-second shield 603. The leads 701A and 701B are electrically connected to the magneto-resistive element 602. Extended leads 705A and 705B are attached to leads 701A and 70IB. . Extended

182 005, leads 705A and 705B have connection points 707A and 707B that are connected to lead leads 709A and 709B, which in turn are connected to preamplifier module 711.

The physical phenomenon of producing the thermal voltage response v-_no on the magneto-resistive element 602 of the magnetic head is that as the instantaneous head spacing of the disc increases, there is more air space between the head 600 and the disk surface 101A, causing the magneto-resistive element 602 to heat up. This heating causes an increase in the resistance of the head 600 due to the positive temperature coefficient of the material. which the magneto-resistive element 602 is made of. For example, the permalloy has a temperature coefficient of +3 x 10 '³/ ° C as mentioned previously. At a constant bias current, the voltage thermal response v_n! on the resistance of the magneto-resistive element 602 will increase. If the resistive element 602 comes close to the disk surface 501A, greater heat transfer will occur between the resistive member 602 and the disk surface 501A, causing the magneto-resistive element 602 to cool. The resulting reduction in the resistance of the head 600 reduces the thermal voltage response v._TH on the magneto-resistive element 602 at a constant polarization current.

182 005

<img file="PL182005B1_D0001.tif" />

FIG. 2

182 005

<img file="PL182005B1_D0002.tif" />

FIG. 3

182 005

<img file="PL182005B1_D0003.tif" />

FIG. 4

182 005

<img file="PL182005B1_D0004.tif" />

FIG. 5

182 005

<img file="PL182005B1_D0005.tif" />

182 005

<img file="PL182005B1_D0006.tif" />

FIG. 7

182 005

<img file="PL182005B1_D0007.tif" />

FIG. 8

182 005

<img file="PL182005B1_D0008.tif" />

210

182 005

SCI

<img file="PL182005B1_D0009.tif" />

316 ----- FIG. 10

182 005

DISK ROTATION

<img file="PL182005B1_D0010.tif" />

182 005

Ρ% (Λ&gt; ι Ο

<img file="PL182005B1_D0011.tif" />

182 005

<td>DISC 1</td><td colspan="2">DISC 2 DISC 3</td>

<td>PAGE AB er</td><td>ABA</td><td>B</td>

<td colspan="3">2 it ¹ and ¹ |</td>

<td colspan="3">C&gt;] 1 · '. ¹ and</td>

<td colspan="3">u C _____________'______________! ______________! ______________! _____________ I _____________! ______________</td>

<td colspan="3"> &lt; ^υ 1 1 ₍ ’ 1</td>

<td colspan="3">H n li! '!!</td>

<td colspan="3">ZH 5 ------------! -! -------------! ------------: ----- -------!------------and------------</td>

<td colspan="3">In [Łt * ¹ * ¹ » 1</td>

<td colspan="3">4 ---------! ---------! -------- Λ -----------------! - -------! ---------</td>

<td colspan="3">¹ * V i ¹ · rM ł-H ij। lt 1</td>

<td colspan="3">«Ix and __________।_________________ &lt; km 3--0-- - 7 '------------ .- --O ---------- -.....</td>

<td colspan="2">zh: /; ; &lt;</td><td></td>

<td colspan="2">&lt;ί UO ________ / ।________________ 1 «___________ and ________ /</td><td></td>

<td colspan="2">² Λ ...... ~ .......; /</td><td> 1</td>

<td colspan="2">¹ ri — η ~; - ^ “r</td><td> 1 1 1</td>

<td colspan="2">Path at ---? —— ΐ --- \ - ΐ --- Ζΐ --— • I \! &gt;</td><td> 1</td>

<td colspan="2">-i - -i — 4 — A — V</td><td>J J ... ..</td>

<td colspan="2">s -2 —V — M — ΐ — L · -</td><td></td>

<td colspan="2"></td><td></td>

<td colspan="2">g -3 -; - V - Mh--</td><td>And ł</td>

<td colspan="2"></td><td></td>

<td colspan="2">H -----! -----? -----; --- V -----</td><td> \ 1</td>

<td> &lt; &lt;</td><td></td><td> \ 1</td>

<td> « -5 ------!-------</td><td>f |</td><td> \ 1</td>

<td> &lt; .</td><td>t ł</td><td></td>

<td>W -6 _______! _______ -CS3 ^AT .</td><td>1 t 1 1</td><td> ......1 </td>

<td>To me</td><td>1 t</td><td> 1</td>

<td>H ________! _________</td><td>and 1</td><td> 1</td>

<td>θ '1</td><td>'2 3 i</td><td>I 5 '</td>

HEAD NUMBER

FIG. 12

182 005

ΓΌ iZ

461a ^ A60

<img file="PL182005B1_D0012.tif" />

HEAT SIGNAL

182 005

<img file="PL182005B1_D0013.tif" />

AMPLITUDE (VOLTS)

TIME (MICRO:

0.4

182 005

<img file="PL182005B1_D0014.tif" />

AMPLITUDE (VOLTS)

FIG. 15

182 005

<img file="PL182005B1_D0015.tif" />

182 005

0.2

<img file="PL182005B1_D0016.tif" />

AMP LITUDA (WOLTY)

<img file="PL182005B1_D0017.tif" />

AMPLITUDE (VOLTS)

TIME (MICROSECONDS)

182 005

<img file="PL182005B1_D0018.tif" />

<img file="PL182005B1_D0019.tif" />

FIG. I8B

AND

AMPLITUDE (VOLTS)

AMPLITUDE (VOLTS)

182 005

<img file="PL182005B1_D0020.tif" />

AMPLITUDE (VOLTS)

182 005

<img file="PL182005B1_D0021.tif" />

<img file="PL182005B1_D0022.tif" />

PHASE (RADIANS)

FREQUENCY (RADIANS)

182 005

<img file="PL182005B1_D0023.tif" />

<img file="PL182005B1_D0024.tif" />

FREQUENCY (RADIANS)

AMPLITUDE

PHASE (RADIANS)

182 005

<img file="PL182005B1_D0025.tif" />

FIG. 21

182 005 ο tD

<img file="PL182005B1_D0026.tif" />

CM CM

182 005

<img file="PL182005B1_D0027.tif" />

<img file="PL182005B1_D0028.tif" />

FREQUENCY (Hz)

AMPLITUDE

PHASE ANGLE (DEGREES)

182 005

<img file="PL182005B1_D0029.tif" />

VALUE

182 OOS

<img file="PL182005B1_D0030.tif" />

PHASE ANGLE (RADIANS)

FIG. 25

182 005

<img file="PL182005B1_D0031.tif" />

ω cu b

182 005

<img file="PL182005B1_D0032.tif" />

182 005

<img file="PL182005B1_D0033.tif" />

AMP. (mV)

182 005

<img file="PL182005B1_D0034.tif" />

AMP. (mV)

<img file="PL182005B1_D0035.tif" />

ΆΜΡ. (mV)

FIG. 28C

182 005

<img file="PL182005B1_D0036.tif" />

MAGNETIC DISTANCE (NANOMETERS)

182 005

<img file="PL182005B1_D0037.tif" />

Publishing Department of the Polish Patent Office. Circulation of 70 copies

Price PLN 6.00.