Methods of Improved Servo Head Passive Noise Cancellation by Placement and Orientation of Multiple Servo Transducers

This invention relates to servo heads in disk file data storage systems. In particular, the present invention relates to an improved passive noise cancellation in the servo heads by the use of multiple servo transducers.

Disk file data storage systems contain magnetic media disks on which data is stored in concentric tracks by means of data heads which write and read magnetically encoded data. These data heads are mechanically linked to a servo head located on an electromechanical actuator. The servo head and the actuator are part of a negative feedback, closed loop servo system which is used to move the data heads (and the servo head) from one track to another for track seek operations and to hold them directly over a track for track following.

The normal output signal of the servo head is encoded radial position information, which it reads from the servo disk. However, the servo head is also susceptible of picking up external electromagnetic radiation. The main source of this external electromagnetic radiation is the data head located across from the servo head. When that data head is writing data, it is also radiating noise to the servo head. The noise corrupts the normal radial position information for the servo system, thus, reducing the track following precision. Consequently, the data written at that time may not be completely within the actual data track boundaries. Later, when this data is being read, the normal track following precision has returned. However, the data head is then flying within the track boundaries trying to read data which may not be written completely within the track boundaries. This increases the probability that data errors will occur during reading.

A passive noise cancellation technique for servo heads is described in a commonly-assigned, co-pending patent application entitled Cross Talk Cancellation for Magnetic Read Head by L. A. Wilcox, Ser. No. 159,878, filed Feb. 24, 1988, and which is hereby incorporated by reference. In this co-pending application, the servo head includes two servo transducers. The first servo transducer reads the servo information stored on the servo disk. The second servo transducer is located on the same slider as the first servo transducer, but has its pole tips recessed slightly from the servo disk surface to prevent it from reading a signal from the servo disk. The second servo transducer is connected to the first servo transducer with opposite coil polarity in either a series-opposing or parallel-opposing configuration. The noise induced in the second servo transducer (which is generated by the adjacent data head when it is writing data) tends to cancel similar noise which is induced in the first servo transducer. Therefore, during normal operation, the net output voltage of this combined servo head, even in the presence of noise, is the signal read from the servo disk by the first servo transducer.

The present invention further improves passive noise cancellation in servo heads of a disk file storage system. The present invention is based upon a recognition of limitations in the passive noise cancellation technique which uses two servo transducers.

First, the difference in distances of the first and second servo transducers from the data transducer of the data head can limit the effectiveness of the passive noise cancellation. Although the difference in distance is small (typically 0.015 inches to 0.020 inches), it has been found that this difference is enough to cause up to 20% difference in the amplitude of noise induced in the two servo transducers. Because the second servo transducer is more distant from the data transducer than the first servo transducer, lower amplitude noise is induced in the second servo transducer. This results in incomplete cancellation of noise from the servo output signal. In addition, as the space between disks is decreased, the difference in distance between the first and second servo transducers and the data transducer becomes even more pronounced.

Second, although the second servo transducer is recessed slightly from the disk surface, it can still read signals from certain types of servo track patterns. In particular, the second servo transducer has been found to be responsive to signals from the edges of gap fields which are found in sectored servo tracks. Examples of gap fields are index fields, that are blank in every sector but one per revolution, and track ID fields that are essentially blank for certain digital track ID codes. These are low spatial frequency structures which are more readily picked up by the recessed second servo transducer. The flux produced at the edges of gap fields induces a low amplitude read signal in the second servo transducer. This read signal corrupts the servo output signal and decreases the accuracy of position information used by the servo system.

The present invention increases track following precision and accuracy of the servo system by changing the placement and orientation of the second servo transducer.

In order to eliminate incomplete noise cancellation, the servo head of the present invention locates the second servo transducer on the servo slider in a position so that the first and second servo transducers are essentially at equal distances from the data transducer.

In order to reduce susceptibility to signals from edges of gap fields, the pole tips of the second transducer are preferably oriented away from the servo disk surface. This increases the distance of the pole tips from the servo disk surface, and reduces the ability of the second servo transducer to read any unwanted signals from the servo disk.

FIG. 1 is a block diagram of a servo system of a rotating disk data storage system.

FIG. 2 is a perspective view of a typical servo or data head with a thin film magnetic transducer.

FIG. 2A is a diagrammatic view of a thin film transducer of the type shown in FIG. 2.

FIG. 3 is a diagram showing the relationship of the transducer to the track and to the transducer output signal.

FIG. 4 is a diagram showing wave forms of data transducer write current and noise induced in a servo transducer as a function of time to show the relationship of data transducer write current to noise induced in the servo transducer.

FIG. 5 is a graph showing amplitude of noise induced in a first servo transducer as function of magnitude of write current in the data transducer for two different disk spacings.

FIG. 6 shows a diagram of a servo head having passive noise cancellation, but without the improvement of the present invention.

FIG. 7A shows the connection of first and second servo transducers in a series opposing configuration.

FIG. 7B shows first and second servo transducers connected in a parallel opposing configuration.

FIG. 8 is a graph showing amplitude of noise induced in a servo transducer as a function of magnitude of write current showing the effect of disk spacing on the amplitude of noise induced in the first and second servo transducers of the servo head shown in FIG. 6.

FIG. 9 is a graph of amplitude of noise induced in a servo transducer as a function of the magnitude of write current in the data transducer showing the effect of disk separation on the amplitude of noise induced in series opposing connected servo transducers.

FIGS. 10A-10D illustrate the effect of disk separation on the distance of the first and second servo transducers from the data transducer.

FIG. 11 is a diagram showing the relationship of the second servo transducer to the track and the second transducer output signal at the edges of gaps.

FIG. 12 is a diagram showing the placement and orientation of the first and second servo transducer in accordance with the present invention for use with a single disk spacing distance.

FIG. 13 is a diagram showing two different disk spacings and the placement and orientation of the first and second servo transducers in accordance with the present invention to accommodate both disk spacing distances.

FIG. 1 shows a disk file data storage system 10 which includes magnetic disks 12, 14, 16 and 18 which are mounted on spindle 20. In the embodiment shown in FIG. 1, data head 22 writes data to and reads data from lower surface 12b of disk 12. Data head 24 reads data from and writes data to upper surface 14a of disk 14, while data head 26 reads data from and writes data to lower surface 14b of disk 14. Data head 28 reads data from and writes data to lower surface 16b of disk 16, and data head 30 reads data from and writes data to the upper surface 18a of disk 18.

The positioning of data heads 22, 24, 26, 28 and 30 is controlled by a servo system which includes servo head 32, actuator 34, servo read amplifier 36, servo position demodulator 38, position error signal generator 40, and actuator driver 42. In this particular embodiment, actuator 34 includes an actuator spindle 44 and three actuator arm assemblies 46, 48, and 50. Data heads 22 and 24 are supported by flex arms at the distal end of actuator arm assembly 46. Data head 26 and servo head 32 are supported by flex arms at the distal end of actuator arm assembly 48. Data heads 28 and 30 are supported by flex arms at the distal end of actuator arm assembly 50.

Servo head 32 is positioned adjacent upper surface 16a of disk 16. This upper surface 16a acts as the servo disk surface, and has stored track information which is read by servo head 32 and supplied to servo read amplifier 36 and then to servo position demodulator 38. Servo position demodulator 38 produces an actual transducer position signal based upon information read by servo head 32. The actual transducer position signal is compared to a desired transducer position signal, and a position error signal is generated by position error signal generator 40. The position error signal is supplied through servo driver 42 to control operation of actuator 34. The servo system shown in FIG. 1 provides a negative feedback, closed loop control so that the data heads are moved generally in a radial direction with respect to disks 12, 14, 16 and 18 in order to move the data heads to a desired track during a track seek operation, and to hold the data heads over the desired track during a reading or writing operation.

In the preferred embodiments of the present invention, data heads 22, 24, 26, 28 and 30, and servo head 32 use thin film magnetic transducers. FIG. 2 shows a typical prior art servo head, which includes slider 60 and thin film transducer 62. The construction of the servo and data heads is generally similar.

Transducer 62 is shown in further detail in FIG. 2A. Basically, transducer 62 includes a flat coiled conductor 64 and upper and lower pole pieces 66 and 68 which are formed in a U or horseshoe configuration. Upper and lower pole pieces 66 and 68 terminate in a pair of pole tips 70 and 72 with a gap 74 between them. Opposite ends 76 and 78 of coil conductor 64 are connected to transducer terminals 80 and 82, respectively. External leads can be connected to transducer terminals 80 and 82.

The pole pieces 66 and 68 loop through coil 64, and pole tips 70 and 72 are oriented toward the disk surface. As the disk spins, the pole tips concentrate the low level magnetic flux from the track into coil 64, thus generating an output voltage signal.

FIG. 3 shows the operation of servo transducer 62 as a servo track on upper surface 16a of disk 16 moves relative to transducer 62. In FIG. 3, the transducer output voltage, which is derived from transducer terminals 80 and 82, is shown as a function of time as the servo track moves past pole tips 70 and 72.

The primary source of undesired electromagnetic radiation to the servo transducer 62 of servo head 32 is the data head 26 which shares the same actuator arm assembly 48 with servo head 32. As shown in FIG. 1, data head 26 and servo head 32 are both located between disks 14 and 16, with data head 26 adjacent lower surface 14b of disk 14, and servo head 32 adjacent upper surface 16a of disk 16. The undesired electromagnetic radiation to servo head 32 occurs when data head 26 is writing data. As illustrated in FIG. 4, noise is induced in servo transducer 62 of servo head 32 with each transition in the data transducer write current. The amplitude of the noise induced in servo transducer 62 is proportional to the magnitude of the write current in data head 26, and also is related to the distance between the servo transducer 62 in servo head 32 and the data transducer in data head 26. These effects are illustrated in FIG. 5. The further the servo transducer is located from the data transducer, the lower the amplitude of the noise induced for a given write current.

FIG. 6 illustrates a passive noise cancellation technique which is described in further detail in the previously mentioned co-pending patent application. In this embodiment, servo transducer 32 includes first servo transducer 62 and a similar second servo transducer (or "inactive transducer") 90. Second servo transducer 90 is located on slider 60, but has its pole tips recessed slightly from surface 16a of disk 16. Second servo transducer 90 is connected to first servo transducer 60, but with opposite coil polarity in either a series opposing configuration as shown in FIG. 7A or a parallel opposing configuration as shown in FIG. 7B. In either case, the noise currents i.sub.N1 and i.sub.N2 induced by a write operation of data head 26 tend to cancel one another. Therefore, during normal operation, the net output voltage of the combined servo transducers 62 and 90, even in the presence of noise, more precisely reflects the signal read from the servo surface 16a of disk 16 than with a similar signal derived solely from first servo transducer 62.

As shown in FIG. 6, data head 26 includes a slider 100 and thin film magnetic data transducer 102. The construction and positioning of data transducer 102 is similar to the construction and positioning of first servo transducer 6. As shown in FIG. 6, first servo transducer 62 and data transducer 102 are generally aligned with one another and are closer to one another than are second servo transducer 90 and data transducer 102.

The noise cancellation by the use of first and second servo transducers 62 and 90 is directed specifically to noise induced by the write operation of data transducer 102 of data head 26. Disk 12, 14, 16 and 18 are typically metal disks and, as a result, servo transducer 62 is effectively shielded from all of the other data heads 22, 24, 28 and 30 by a metallic disk. That is not the case, however, with data head 26 which, like servo head 32, is positioned between disks 14 and 16.

Although the passive noise cancellation provided by the structure shown in FIG. 6 represents an improvement over the use of only a single servo transducer, there are limitations to the effectiveness of the passive noise cancellation technique. First, the arrangement of servo transducers 62 and 90 as shown in FIG. 6 does not take into account the difference in distances between first and second servo transducers 62 and 90 and data transducer 102. Although the difference is small, it is enough to cause up to 20% difference in the amplitude of the noise induced in servo transducers 62 and 90. Because second servo transducer 90 is further from data transducer 102, lower amplitude noise is induced in second servo transducer 90. This is illustrated in FIG. 8. For disk spacing of 0.150 inches, there is a noticeable difference in the amplitude of the noise induced in the first and second servo transducers. For disk spacing of 0.120 inches, the difference between the amplitude of noise induced in the first and second servo transducers becomes more pronounced, and the difference increases as the magnitude of write current increases. Incomplete cancellation of noise from the servo read output signal affects the precision of the servo system during track following.

FIG. 9 shows the effect of disk separation on the amplitude of noise induced when servo transducers 62 and 90 are connected in series opposing arrangement. From FIG. 9, it can be seen that net amplitude of noise induced becomes more pronounced as spacing between the disks is reduced.

FIGS. 10A-10D illustrate the effect of disk separation on the distances of the first and second servo transducers 62 and 90 from data transducer 102. In each FIGS. 10A-10D, the spacing between first and second servo transducer 62 and 90 is unchanged (0.070 inches) while the distance from first servo transducer 62 to data transducer 102 is reduced from 0.150 inches in FIG. 10A, to 0.140 inches in FIG. 10B, to 0.130 inches in FIG. 10C, to 0.120 inches in FIG. 10D. It can be seen that the difference between the distances becomes more pronounced as the distance between first servo transducer 62 and data transducer 102 is decreased.

The second deficiency in the prior passive noise cancellation technique is that the second servo transducer 90, even though it is recessed slightly from the surface of disk 16, is still close enough to the surface so that it can read a signal from certain types of servo track patterns. This effect is illustrated in FIG. 11. In particular, second servo transducer 90 can read a signal from the edges of gap fields which are found in sectored servo tracks. Examples of gap fields are Index fields (that are blank in every sector but one per revolution) and Track ID fields (that are essentially blank for certain digital track ID codes). These are longer wavelength structures than the normal servo pattern. In general, the signal amplitude from a transducer is a function of not only the space or distance between the transducer and disk, but also the wavelength of the magnetic pattern. This relationship is described in the following approximate formula for spacing loss, in dB: ##EQU1## where d is the distance between the transducer pole tips and the disk surface, and .lambda. is the wavelength of the pattern. Thus, for a given distance, d, a longer wavelength structure, such as a gap field, has less signal loss. The distance, d, used in the prior passive noise cancellation technique is not large enough to eliminate the read signal from the gap fields. As illustrated in FIG. 11, the flux from a gap induces a low amplitude read signal in second servo transducer 90. The second read signal corrupts the normal servo read signal generated in first servo transducer 62, and decreases the accuracy of the position information used by the servo control system.

FIG. 12 illustrates a first preferred embodiment of the present invention. In FIG. 12, servo head 32' includes a servo slider 60', first servo transducer 62', and second servo transducer 90'. Servo head 32' of the present invention achieves an increase in track following precision by improved passive noise cancellation. In this embodiment, second servo transducer 90' is fabricated on servo slider 60' in a position such that first servo transducer 62' and second servo transducer 90' are an equal distance from data transducer 102 of data head 26. In addition, the pole tips 110 of second servo transducer 90' are oriented away from disk surface 16a to more effectively prevent second servo transducer 90' from reading a signal from disk 16.

When servo transducers 62' and 90' are connected with opposing polarity (as discussed previously), the output noise is completely canceled only if the amplitude of the noise induced in each separate transducer is equal. The amplitude of noise induced in servo transducers 62' and 90' is a function of the distances of those transducers from data transducer 102. For maximum noise cancellation, therefore, servo transducers 62' and 90' are located at an equal distance from data transducer 102.

The distance between first servo transducer 62' and data transducer 102 is determined by the disk spacing between disks 14 and 16. Second servo transducer 90' is located in a circular zone of equidistance 120 which is drawn around the center of data transducer 102. This zone of equidistance is defined by first radius R1 which extends from the center of data transducer 102 to the uppermost portion of first servo transducer 62', and a second radius R2 which extends from the center of data transducer 102 to the lower most portion of first servo transducer 62'. As shown in FIG. 12, this zone of equidistance shows the zone on servo slider 60' where second servo transducer 90' can be positioned in accordance with the present invention.

The location of the zone of equidistance 120 is a function of disk spacing. The sensitivity to disk spacing, however, can be minimized by placing second servo transducer 90' as close as possible to first servo transducer 62'. FIG. 13 shows the intersection of two zones of equidistance for two typical disk spacings (0.150 inches and 0.120 inches). The area of intersection 130 is widest near first servo transducer 62'.

Other constraints (such as the position of transducer output terminals on slider 60') can effect the exact position of second servo transducer 90' within the zone of equidistance 120 shown in FIG. 12, or the intersection of two zones of equidistance 130 as shown in FIG. 13.

Another important feature of the present invention is that pole tips 110 of second servo transducer 90' are oriented generally away from the servo disk surface 16a. Second servo transducer 90' is essentially identical in construction to first servo transducer 62', so that they both have essentially the same inductance, resistance, and physical size. The amplitude of noise induced in both transducers 62' and 90' by a given electromagnetic field, therefore, should be equal. However, the orientation of pole tips 110 of second servo transducer 90' prevents second servo transducer 90' from reading an unwanted signal from servo disk surface 16a.

The improved noise cancellation provided by the present invention is important because servo head 32' can be placed in the center of the disk stack and still have data head 26 across from it. This permits the utilization of all data disk surfaces, which maximizes the data capacity of the disk drive. In addition, the mechanical and thermal drift tolerance stack-up between the outer data heads and the servo head is minimized, which improves track following accuracy. This also allows the possibility of inclusion of more disks in the disk stack, which again increases overall data capacity.

Although the present invention has been described with reference to preferred embodiments, workers skilled in the art will recognize that changes may be made in form and detail without departing from the spirit and scope of the invention.