Multichannel digital recording system with multi-user detection Number:6,826,140 from the United States Patent and Trademark Office (PTO) owispatent A processing scheme for digital storage media using multi-user detection to separate tracks of data or remove interference from neighboring tracks. In one embodiment, data is written on a plurality of tracks positioned sufficiently close together so that multiple tracks are detected simultaneously by the read access sensor. Upon scanning the surface for data, the read element simultaneously receives the data signals from a plurality of tracks. Joint detection signal processing resolves the interference and data bits from the multiple sensed tracks, enabling closer packing of data with minimal guard space.<H Multichannel, detection, Multichannel, di, 6826140.html, Patent, pto, 6826140

Title: Multichannel digital recording system with multi-user detection

Abstract: A processing scheme for digital storage media using multi-user detection to separate tracks of data or remove interference from neighboring tracks. In one embodiment, data is written on a plurality of tracks positioned sufficiently close together so that multiple tracks are detected simultaneously by the read access sensor. Upon scanning the surface for data, the read element simultaneously receives the data signals from a plurality of tracks. Joint detection signal processing resolves the interference and data bits from the multiple sensed tracks, enabling closer packing of data with minimal guard space.

Patent Number: 6,826,140 Issued on 11/30/2004 to Brommer, et al.

Inventors: Brommer; Karl D. (Hampton Falls, NH); MacLeod; Robert B. (Nasuha, NH); Schmidt; Michael P. (Hollis, NH)
Assignee: BAE Systems Information and Electronic Systems Integration INC (Nashua, NH)

Appl. No.: 251187

Filed: September 20, 2002

Current U.S. Class: 369/94 ; 369/124.02; 369/124.04; 369/47.19; 369/59.22

Field of Search: 369/94,95,96,97,124.02,124.03,124.04,124.05,59.13,59.2,59.21,59.22,59.26,59.27,47.16,47.18,47.19,44.37,44.38 360/46,48,63,67,58

References Cited [Referenced By]

U.S. Patent Documents


4549232	October 1985	Axmear et al.
4945536	July 1990	Hancu
5285439	February 1994	Clemens
5307217	April 1994	Saliba
5371638	December 1994	Saliba
5402399	March 1995	Oshiba et al.
5448514	September 1995	Cho et al.
5543861	August 1996	Harradine et al.
5555142	September 1996	Komai et al.
5748592	May 1998	Hino et al.
5818806	October 1998	Wong et al.
5854714	December 1998	Reed et al.
5949604	September 1999	Saliba
5956196	September 1999	Hull et al.
5959953	September 1999	Alon
5995461	November 1999	Sedlmayr et al.
5995561	November 1999	Yamaski et al.
6031672	February 2000	Bergquist et al.
6122134	September 2000	Kaaden et al.
6122269	September 2000	Wales
6137763	October 2000	Dahan et al.
6150964	November 2000	McLaughlin
6181505	January 2001	Sacks et al.
6222833	April 2001	Seo
6247156	June 2001	Kim
6266712	July 2001	Henrichs
6304400	October 2001	Schneider
6318633	November 2001	Drexler
6373648	April 2002	O'Conner
6377529	April 2002	Lee et al.
6381210	April 2002	Alon et al.
6567364	May 2003	Takahashi et al.
2001/0024335	September 2001	Blaum et al.
2002/0037061	March 2002	Learned
2002/0037737	March 2002	Learned et al.

Other References

Gallager, R.G., Low-Density Parity-Check Codes, IRE Transactinos on Information Theory Jan. 1962, pp. 21-28. .
Barbosa, Lineu C., Simultaneous Detection of Readback Signals from Intefering Magnetic Recording Tracks Using Array Heads. IEEE Transactinos on Magnetics vol. 26, No. 5, Sep. 1990, pp. 2163-2165. .
Vea, Mathew P., and Moura, Jose, M. F., Magnetic Recording Channel Model with Interack Interference, IEEE 1991, pp. 4834-4836. .
Tsang, Chen, Yogi Gigabit-Density Magnetic Recording, IEEE vol. 81 No. 9, Sep. 1993. .
Moon, Jaekyun, The Role of SP in Data Storage, IEEE Signal Processing Magazine, vol. Jul. 1998., No. 5 pp. 54. .
Soljanin, Emina & Georghiades, Costas N., Multihead Detection for Multitrack Recording Channels, IEEE Transactions on Information Theory, vol. 44, No. 7, Nov. 1998, pp. 2988-2997. .
Oenning, T. and Moon, J., Modeling the Lorentzian Magnetic Recording Channel with Transition Noise, IEEE Transactions on Magnetics, vol. 37, Jan., 2001, p. 583. .
Kurtas, Erozan, Proakis, John G., and Salehi, Masoud, Reduced Complexity--Maximum Likelihood Sequence Esitmation for Multitrack High-Density Magnetic Recording Channels, IEEE Transactions on Magnetics , vol. 15, No. 4, Jul. 1999. pp. 2187-2192. .
Richardson, Thomas J. & Urbanke, Rudiger L., The Capacity of Low-Density Parity-Check Codes Under Message-Passing Decoding, IEEE Transactions on Information Theory, vol. 47, No. 2, Feb. 2001. .
Richardson Thomas, J., Shokrollahi, M Amin, and Urbanke, Rudiger L., Design of Capacity-Approaching Irregular Low-Density Parity-Check Codes, IEEE Transactions on Information Theory, vol. 47, No. 2, Feb. 2001, pp. 619-637. .
Yeo, Engling, Pakzad, Payam, Nikolic, Borivoje, and Anantharam, Venkat, High Throughout Low-Density Parity-Check Decoder Architectures, IEEE 2001, pp. 3019-3024. .
Yeo, Engling and Nikolic, Borivoje, VLSI Architectures for High-Speed Iterative Decoders, BWRC Winter Retreat, Jan. 2002. .
Roh, Bong Gyun, Lee Sang, Uk. Moon, Jaekyun and Chen, Ying, Single-Head/Single-Track Detection in Interfering Tracks, IEEE Transactions on Magnetics, vol. 38, No. 4, Jul. 2002. .
Chung, Sae-Young, Forney, G David Jr., Richardson, Thomas J. and Urbanke, Rudiger, On the Design of Low-Density Parity-Check Codes within 0.0045 dB of Shannon Limit. .
Xie, Zhenhua, Rushforth, Craig K., Short, Robert T., Moon, Todd K., Joint Signal Detection and Parameter Estimation in Multiuser Communications,IEEE Transactions on Communication vol. 41, No. 7 Aug. 1993, pp. 1209-1216. .
Poor, Vincent H. Turbo Multiuser Detection: An Overview, IEEE 6th Int. Symp on Spread-Spectrum Tech. & Application, NJ, Sep. 6-8, 2000, pp. 583-587. .
Leonhardt, T. D., Van De Veerdonk, R. J. M., Van Der Heijden, P.A. A., Clinton, T. W., and Crawford, T. M., Comparison of Perpendicular and Longitudinal Magnetic Recording Using a Contact Write/Read Tester, IEEE Trans. Magn. vol. 37 pp. 1580-1582, Jul. 2001. .
Roh, Bong Gyun et al., Single-Head/Single-Track Detection in Interfering Tracks, IEEE Transactions on Magnetics, vol. 38 No. 4, Jul. 2002, pp. 1830-1838. .
Abbott, W. L., et al., Offtrack Interference and Equalization in Magnetic Recording., IEEE Transactions on Magnetics, vol. 24, No. 6, Nov. 1988, pp. 2964-2966. .
Saliba, George,Laser Guided Magnetic Recording, Computer Technology Review, vol. XX, No. 7, Jul. 2000, West World Productions Inc. .
Saliba, George, The Anatomy of Super DLTtape, Computer Technology Review, vol. XX, No. 8, Aug. 2000, West World Productions..

Primary Examiner: Tran; Thang V.
Attorney, Agent or Firm: Maine & Asmus

Parent Case Text

CROSS REFERENCE TO RELATED APPLICATIONS

This application claims priority as a Continuation-in-part under 35 U.S.C. Section 120 from a U.S. patent application Ser. No. 10/228,787 filed on Aug. 26, 2002, which is incorporated herein by reference for all purposes.

Claims

What is claimed is:

1. An apparatus for reading data bits from a storage medium using multi-user detection, comprising: a plurality of tracks wherein said data bits reside within a plurality of storage cells on said tracks; at least one read element simultaneously detecting a plurality of said tracks and converting said data bits into a plurality of electrical signals; a front end unit processing said electrical signals and converting said electrical signals into a plurality of digital bits; a parameter estimator coupled to said front end unit for identifying a track transfer function for said plurality of tracks; and a multi-user detector coupled to said parameter estimator and said front end unit for separating said tracks and reading said data bits.

2. The apparatus according to claim 1, wherein said storage medium is selected from the group consisting of: floppy disks, hard disks, cubical disks, linear disks, multi-level disks, drum memory, linear tapes, helical scanned tapes, radial disks, compact disks, digital video disks, magneto optical disks, and rotating magnetic media.

3. The apparatus according to claim 1, wherein said data bits are stored on said storage medium by a storage technology selected from at least one of the group consisting of: magnetic, optical, magneto optical, electrostatic, and quantum.

4. The apparatus according to claim 1, wherein said track transfer function includes envelope information of a shape, amplitude and phase of each of said plurality of data tracks.

5. The apparatus according to claim 1, wherein said digital bits are represented by a Lorentzian pulse shape.

6. The apparatus according to claim 1, wherein symbols on said disk represent a plurality of said data bits.

7. The apparatus according to claim 6, wherein said symbols use codings selected from at least one of the group consisting of: quadrature phase shift keying (QPSK), binary phase shift keying (BPSK), Code Division Multiple Access (CDMA), quadrature amplitude modulation (QAM), Frequency Division Multiple Access (FDMA), and Time Division Multiple Access (TDMA) amplitude modulation (AM).

8. The apparatus according to claim 1, further comprising a guard-track spacing providing a separation between adjacent tracks.

9. The apparatus according to claim 1, wherein said plurality of data tracks are proximate each other without a guard-track spacing.

10. The apparatus according to claim 1, wherein said front end unit comprises a preamplifier, a low pass filter and an analog-to-digital converter.

11. The apparatus according to claim 1, wherein said data tracks are multi-layered.

12. The apparatus according to claim 1, further comprising a temporary storage buffer and an output multiplexor coupled to said multi-user detector.

13. The apparatus according to claim 1, wherein said multi-user detector is selected from at least one of the group consisting of: maximum likelihood MUD, TurboMUD, and linear algebra based multi-user detector.

14. The apparatus according to claim 1, wherein said multi-user detector uses an algorithm selected from at least one of the group consisting of: M-algorithm, T-algorithm, and MT-algorithm.

15. The apparatus according to claim 1, further comprising a filter unit coupled to said multi-user detector.

16. The apparatus according to claim 15, wherein said filter unit is selected from the group consisting of: whitening matched filter bank and matched filter bank.

17. The apparatus according to claim 1, further comprising a sector cache coupled to said multi-user detector.

18. The apparatus according to claim 1, wherein a sector is a plurality of data bits in one of said tracks, said tracks have a main track and adjacent tracks proximate said main track, and wherein said apparatus further comprises a temporary storage memory unit for storing each said sector for each of said adjacent tracks.

19. The apparatus according to claim 18, wherein each said sector from said adjacent tracks are combined with at least one sector from said main track and placed in a sector cache.

20. The apparatus according to claim 18, wherein each said sector from said adjacent tracks are organized proximate each other in a host computer.

21. The apparatus according to claim 1, wherein said data bits are written to said disk using a set of convolutional codes.

22. The apparatus according to claim 1, further comprising a head tracking controller generating head position error information, and wherein said head position error information is communicated to said parameter estimator.

23. The apparatus according to claim 1, wherein a phase of said data bits between said tracks is controlled when written to said disk.

24. The apparatus according to claim 1, wherein apriori information of said data bits is communicated to said parameter estimator.

25. The apparatus according to claim 1, wherein said parameter estimator calculates information about said tracks.

26. The apparatus according to claim 18, wherein said temporary storage unit contains prefetch data from said adjacent tracks.

27. The apparatus according to claim 1, wherein said at least one read element takes at least one pass over said tracks.

28. The apparatus according to claim 27, wherein at least one of said plurality of analog signals are processed from some read element.

29. A method for processing data bits of a storage medium, comprising: reading a plurality of analog signals corresponding to said data bits from a main track and adjacent tracks of said storage medium; digitizing said analog signals into digital data; generating a track transfer function of said digital data, wherein said digital data is in a Lorentzian form; and demodulating said digital data using said track transfer function.

30. The method according to claim 29, further comprising a plurality of read elements reading said plurality of analog signals.

31. The method for according to claim 29, further comprising the at least one additional step of reading said main track and said adjacent tracks.

32. The method according to claim 29, further comprising filtering said digital data.

33. A system for reading data from a storage medium, comprising: a storage surface on said storage medium having encoded data bits defined by in-track spacing and cross-track spacing, wherein said encoded data bits are stored in a plurality of data tracks; a means for positioning at least one read element over said storage surface, wherein said read element simultaneously detects said encoded data bits from at least one of said tracks; a means for conditioning said encoded data bits from said read element; and a means for demodulating said conditioned encoded data bits from said tracks, wherein said means for demodulating discriminates said data track transfer function that includes envelope information of a shape, amplitude and phase of each of said data track.

34. The system according to claim 33, wherein said track transfer function is Lorentzian.

35. The system according to claim 33, wherein said plurality of data tracks are proximate each other without a guard-track spacing.

36. The system according to claim 33, wherein said adjacent tracks are multi-layered.

37. The system according to claim 33, further comprising a temporary storage buffer and an output multiplexor coupled to a multi-user detector.

38. An apparatus for reading and writing digital data, comprising: a storage medium wherein said digital data is represented in a plurality of storage cells on a plurality of tracks, each of said storage cells having an in-track spacing and a cross-track spacing; at least one read sensor oriented to capture said digital data from at least one storage cell from at least one track; a write element oriented to write said digital data to said storage cell; a servo system coupled to said read sensor, said write element and said storage medium; a system controller coupled to said servo system; and a signal conditioner coupled to said read sensor, wherein said signal conditioner comprises a front end unit, a parameter estimator, and a joint detector for processing said digital data from said at least one track and identifying a track transfer function for said track.

39. The apparatus according to claim 38, wherein said read sensor is stationary and said storage medium is moveable.

40. The apparatus according to claim 38, wherein said read sensor is moveable and said storage medium is stationary.

41. The apparatus according to claim 38, wherein said read sensor is moveable and said storage medium is moveable.

Description

BACKGROUND OF THE INVENTION

1. Technical Field of the Invention

The present invention is a system for storage and retrieval of digital data. More particularly, the present invention involves processing digital recorded material with joint signal detection techniques.

2. Background Art

Digital data storage devices, such as computer drives and portable tapes, compact discs and floppy diskettes, are recording components in many electronic devices, and typically provide mechanisms for storing and retrieving large amounts of data quickly and reliably. Digital recorders, as used herein, refer to the many embodiments employed for storing digital information in a variety of digital systems and a multitude of applications. The most common form of digital recorder is a rotating radial magnetic disk. Other digital recorders include but are not limited to optical disks and magnetic tape systems, including linear devices.

The prior art disk drive system is well known in the art. A data storage disk, such as floppy disks, hard disks, and cubical disks as well as linear and multi-level disks all function in a similar fashion. The common radial disk contains a number of concentric data cylinders that contains several data sectors. The sectors are located on an upper side of the disk and additional sectors may be located on a lower side or in multiple layers within the disk. The disk is accessed by a head element mounted on an arm that is secured to the drive. The disk is accessed via photoemitters/photoreceptors for optical systems and with magnetic read/write elements as discussed herein for magnetic systems wherein various accompanying electronic circuits are familiar to those of skill in the art.

Using disk drives as an example, the disk is typically subdivided into one or more partitions by using a partition table that is located on the disk. A wide variety of partitions file systems as discussed in the prior art are not necessary for a proper understanding of the present invention. A given sector on the disk is usually identified by specifying a head, a cylinder, and a sector within the cylinder. A triplet specifying the head number, cylinder number, and sector number in this manner is known as a physical sector address. Alternatively, a given sector may be identified by a logical sector address, which is a single number rather than a triplet of numbers.

In more specific detail, for a data storage device, such as a magnetic disc drive, the recording medium is typically divided into a plurality of generally parallel data tracks. The data is stored and retrieved by a transducer or head element that is positioned over a desired data track by an actuator arm. The head element can be a combined read/write head or separated into a read head and a write head in close proximity.

The actuator arm typically moves the head across the data tracks under the control of a closed-loop servo system based on servo data stored on the disc surface within dedicated servo fields. The servo fields can be interleaved with data sectors on the disc surface or on a separate disc surface that is dedicated to storing servo information. As the head passes over the servo fields, it generates a readback servo signal that identifies the location of the head relative to the centerline of the desired track. Based on this location, the servo system rotates the actuator arm to adjust the head's position so that it moves to the desired position.

There are several prior art types of servo field patterns, such as a null-type servo pattern, a split-burst amplitude servo pattern, and a phase type servo pattern. A null type servo pattern includes at least two fields which are written at a known phase relation to one another. The first field is a phase or sync field which is used to lock the phase and frequency of the read channel to the phase and frequency of the read signal. The second field is a position error field that is used to identify the location of the head with respect to the track centerline.

In a typical prior art embodiment, as the head passes over the position error field, the amplitude and phase of the read signal indicates the magnitude and direction of the head offset with respect to the track centerline. The position error field has a null-type magnetization pattern such that when the head is directly straddling the track centerline, the amplitude of the readback signal is ideally zero. As the head moves away from the desired track centerline, the amplitude of the read signal increases. When the head is half-way between the desired track centerline and the centerline of the adjacent track, the read signal has a maximum amplitude. The magnetization pattern on one side of the centerline is written 180 degrees out of phase with the magnetization pattern on the other side of the centerline, and the phase of the read signal indicates the direction of the head position error.

To control the servo system, a single position error value is normally generated for each pass over the position error field. Typically, the magnitude of the position error value indicates the distance of the head from the track centerline, and the sign of the position error value indicates the direction of the head's displacement. The position error values are typically created by demodulating the read signal associated with the position error field. In a synchronous process, the exact phase of the read signal from the position error field is known from the phase field's read signal because the phase field is written on the storage medium at a known and fixed phase relation to the position error field. A phase-locked loop (PLL) is typically used to acquire the phase of the phase field, and this phase information is used for demodulating the position error field.

Processing of the read signal is generally demodulated by generating a demodulating signal, such as a square wave, having the same phase and frequency as a fundamental component of the read signal and then, with analog techniques, multiplying the read signal by the demodulating signal. The product is integrated over a time window that corresponds to the middle cycles of the position error field. The result is a position error value for the head with respect to a desired position on the storage medium within that servo pattern. This process essentially identifies the amplitude and phase of the read signal at a specific frequency point. The sign of the position error value indicates which direction the head is located with respect to the desired location.

The most common application for digital recorders is the computer disk drive. All sizes of computers including portable laptops, personal computers and mainframes include a digital recording system. Typically the recording device is a magnetic disk drive, but other devices such as optical disks and tape systems are also commonly used. Besides computers, other digital systems also use digital recorders, for example, digital video cameras write data to a digital recorder in the form magnetic tape, magnetic hard disks or optical disks.

Magnetic disc drives, because of their greater speeds, have become the medium of choice for storing frequently accessed data such as application programs and user data which is being created or frequently modified. Conventional magnetic disk drive storage systems have been commonly used and are well known in the art. These storage systems typically use a flying magnetic read/write head, either combined or separate read head and write head, to record and retrieve data from a layer of magnetic recording material on the surface of a rotating recording disk. The capacity of such a storage system is a function of the number of closely spaced concentric tracks on the recording disk that may be reliably accessed by the read/write head. Some parts of the recording disk surface area may be used for purposes other than data storage.

For example, means for assuring the proper selection of a particular track by the read/write head are required for reliable data storage and retrieval. The read/write head is typically aligned and kept centered over a particular track as the recording disk rotates, to prevent accidental over-writing of data stored in neighboring tracks and to minimize inter-track interference. Some systems use nonmagnetic guard rings between discrete tracks on the recording disk to help keep the head from skipping off-track. Gain control references may be placed at different locations on the recording disk to calibrate the electronic amplifiers used to reliably read back data signals. Time delay elements are also sometimes used to allow the magnetic read/write head to demagnetize after recording data to prevent unintentional over-writing of subsequently accessed locations. The prior art designs take up some of the available recording disk surface area, and thus reduce overall system capacity.

In a magnetic hard disk drive, data bits are stored as transitions between ferromagnetic domains, the absence of transitions, or some combination thereof, indicating a one and a zero respectively. When the read/write head floats over the spinning drive, the transition shows up as a pulsed waveform while the absence of a transition shows up as a flat waveform. By synchronizing and detecting pulses, the read head decodes the ones and zeros on the disk. Each symbol on a track is a `bit` and generally takes the form of a Lorentzian pulse. Magnetic media tends to be two state storage devices due to the physics.

The `bits` as referenced herein refer to any of the schemes that allow for a `1` and a `0` to be detectable by the read element and subsequent processing. For example, the data bits interpreted as `1`s and `0`s may be a magnetic transitions such as polarized and not polarized; polarized and reversed polarized; a transition between two states; or absence of transitions. It is presumed that there will continue to be improvements into the manner in which bits can be written and read, all of which are within the scope of the invention.

Signal processing is used to some extent to optimize the storage on the in-track direction. However, no similar technique currently applies to adjacent track interference, which is known to severely limit disk drive performance. There have been attempts to increase the density of disks in order to have more narrow magnetic domains in the cross-track direction. While the write technology has advanced to allow denser writing, the read technology has been limiting factor in data density.

The importance and significance of signal processing is detailed in the article from IEEE Signal Processing Magazine, July 1998, entitled "The Role of SP in Data-Storage" by Jackyun Moon. The problems related to intersymbol interference (ISI) are described along with the prior art processing techniques involving sequence detectors and symbol-by-symbol detectors. The symbol detectors are more likely to be effected by ISI, whereas the sequence detectors make symbol decision based on the observation of signals over many symbol intervals. Examples of sequence detectors include maximum-likelihood sequence detectors (MLSD), finite or fixed delay tree search detectors (FTDS) and partial-response maximum likelihood (PRML) techniques.

However, the prior art is replete with the attempts and problems with decreasing the spacing in order to put more digital information in a smaller space on the recording medium. While the write elements are technically capable of writing in a smaller area, the read element has limitations that restrict the size of the storage cells and the spacing. An example reference would be Roh, Lee and Moon, "Single-Head/Single-Track Detection in Interfering Tracks", IEEE Transactions on Magnetics vol 38 page 1830 from July 2002. This reference and many others discuss how interference from adjacent tracks due to head misalignment or other effects that tend to become the dominant source of read errors.

There has been considerable research in gigabit-density recording, including an article by Tsang, Chen and Yogi, which discusses the "Gigabit research in Gigabit-Density Magnetic Recording", Proceedings of the IEEE, Vol. 81, No. 9, September 1993. This article illustrates the need for advanced processing that can take advantage of the high density disk recording of the data recorded as transitions that relate to the abrupt Magnetization changes on the tracks of the disk.

There are a variety of factors that limit the read/write capabilities of storage media, including various types of noise, inter-track interference, intersymbol interference, and non-linear distortion. Numerous equalization and coding schemes have evolved to provide more accurate determinations and permit greater density storage. One type of noise source for magnetic disk storage is the result of the recording head positioning error. This off-track or inter-track interference (ITI) can be modeled and reduce the associated errors. There are types of noise that are random, such as transition noise that occurs due to random variations in the geometry of magnetic transitions. There are also types of nonlinearities that have a repeatable characteristic and these distortions can be modeled and eliminated. One manner for describing the nonlinear distortions uses Volterra functional series which constructs the nonlinear portion of the signal as the sun of the outputs of nonlinear kernels. Also, the read head is sensitive to magnetic domains and adjacent tracks, and even if perfectly aligned, tracks must be spaced far enough apart to allow distinctions for processing. One reference for this noise model would be T. Oenning and J. Moon, "Modeling the Lorentzian Magnetic Recording Channel with Transition Noise" from IEEE Transactions on Magnetics volume 37 page 583 (January 2001).

Magnetic recording devices, such as magnetic disks and tapes, use heads to read and write information to and from a magnetic surface. In a typical rotating storage system, data is stored on magnetic disks in a series of concentric tracks. These tracks are accessed by a read/write head that detects variations in the magnetic orientation of the disk surface. In most embodiments the read/write head moves back and forth radially on the disk under control of a head-positioning servo mechanism so that it can be selectively positioned over a specific track. Once the head is aligned over a track, the servo mechanism causes the head to trace a path that follows the center line of the selected track. Tracks as discussed herein refer to any segments parallel to relative motion of the sensor.

The recording head induces a magnetic field with sufficient amplitude to record on the magnetic material of the storage device to a sufficient depth. The magnitude and direction of the magnetic flux is modulated to encode information into the magnetic surface of the storage device. A pattern of external and internal fields are created as the head and recording surface are moved relative to each other. The polarity transitions are then readable as transitions in the magnetic flux at the recording surface. In read mode, as the magnetic storage surface moves across the gap in the head, the magnetic field of the storage surface is detected, and a voltage is induced in the head proportional to the rate of change of the flux. The read channel then processes the analog voltage signal to obtain the digital data.

Various types of indexing marks and alignment indicia are also recorded on the recording disk surface for precise position reference and tracking adjustment of the read/write head. These marks and indicia are often recorded in servo sectors, which are angularly-spaced reserved portions of the recording disk surface that extend out approximately radially from the recording disk centers. Track addresses are sometimes recorded in servo sectors. Angular synchronization signals that determine the circumferential location of the magnetic head may also be recorded in servo sectors. Normal and quadrature servo blocks are often recorded in servo sectors for generation of position error signals that are used to keep the read/write head aligned. Servo sectors use recording disk surface area that could otherwise be used for data storage, however, so servo sector information should be stored as efficiently as possible.

A typical prior art read process commences as the analog read signal originates from the read head which is then amplified in the preamplifier and then provided to a filter for the removal of high-frequency noise components. The filtered signal is then provided to a phase-locked loop clock circuit and delay line. The delay line provides the delayed signal for the analog-to-digital converter (ADC) where the signal is digitized. The digitized signal is passed through an equalizer to obtain a more desirable waveform, and the result is provided to a decoder. The decoder implements a decoding algorithm to generate the digital data signal. The analog-to-digital converter and decoder are clocked by a clock signal generated in a phase-locked loop clock circuit.

With respect to in-track processing, it was recognized early on that the single symbol bit processing was not satisfactory in dealing with ISI and noisy signals, and partial response maximum likelihood (PRML) processing provided certain benefits. In the PRML channel characterized by the polynomial (1-D)(1+D), a notch filter is generally used because the frequency response requires a sharp cutoff and the frequency spectrum is very different from that of the channel response in magnetic recording. A variation of PRML is extended partial response maximum likelihood (EPRML) that obviates the need for the notch filter. However, the Viterbi type computations for maximum likelihood detection become a limiting factor in terms of decoding speed and cost. Furthermore, both PRML and EPRML channels are very sensitive to mis-equalization or changes in signal shape due, for example, to component tolerances and to nonlinearities of the magnetic recording process such as caused by pulse asymmetry and the crowding of write transitions on the media. Moreover, the problems associated with cross-track interference still remain.

Early magnetic storage devices used analog peak detection to process incoming read signals. However, as recording density increased, the analog peak detection scheme became unreliable because of the large amount of inter-symbol interference (ISI) between adjacent pulses. The partial response maximum likelihood (PRML) channel has been used to increase the recording density, but the PRML method requires equalization of the read signal, and the code scheme is incompatible with the run-length limited (RLL) code. In addition, the required number of magnetic flux transitions per inch is much higher. Therefore, the magnetic non-linearity problem is more severe for the PRML system, and could even render it unusable at high recording densities.

Run-length limited (RLL) codes are used to place an upper bound on the number of data clock cycles occurring between signal transitions, and the clock recovery is based on the occurrence of these transitions. RLL codes ensure that sufficient transitions occur for the clock recovery circuit to maintain the correct timing phase and frequency. In an NRZI format, each 1 is represented by a transition, and each 0 is represented by the lack of a transition, and the RLL code is sufficient for clock recovery purposes. Also, by maintaining the minimum of one 0 between consecutive 1's, transitions are separated so as to be differentiable from one another.

A signal processing method that uses RLL codes to improve the detection margin at high recording densities is described in U.S. Pat. No. 4,945,538. In U.S. Pat. No. 4,945,538, sample values of an analog signal corresponding to binary data are coded with a RLL code. The coded analog input signal is converted to a sequence of digital sample values and the signal is equalized to correspond to a predetermined analog shape. A sequence detection algorithm is used to decode the digital sample values into the coded binary data.

A different approach to increase the capacity and speed of optical data-storage systems uses multilevel optical recording systems. The term multilevel refers to more than two levels of data recorded on the medium. The density of data recorded on an optical recording medium is increased by modulating the reflectivity of the optical recording medium into more than two states. However, at high data densities, light reflected from one mark will tend to interfere with light reflected from adjacent marks, causing intersymbol interference (ISI). The effect of the ISI is greater when the marks are closer together.

Optical data disc readers primarily have involved analog filtering of the read signal to equalize the frequency response of the system in order to predict how much contrast an optical imaging system will generate when scanning different spatial frequencies. Digital equalization is generally superior to analog equalization, as discussed in U.S. Pat. No. 5,818,806. And, a method for providing digital equalization filters for multilevel data-storage systems and a compensating scheme for intersymbol interference is described in U.S. Pat. No. 6,377,529.

Phase-change technology has been around since 1995, and the PD drive combines an optical disk drive capable of handling high capacity disks along with a multi-speed CD-ROM drive. It uses purely optical technology, and relies on the use of a laser to write new data with just a single pass of the read/write head. In the PD system, the active layer is made of a material with reversible properties, and a high-power laser heats the portion of the active layer where data is to be recorded. The heated area cools rapidly, forming an amorphous spot of low reflectivity. A low-powered laser beam detects the difference between these spots and the more reflective, untouched, crystalline areas, thus identifying a binary "0" or "1". By reheating a spot, recrystallisation occurs, resulting in a return to its original highly reflective state. Laser temperature alone changes the active layer to crystalline or amorphous according to the data required, in a single pass.

Compact discs (CD's) are examples of digitally recorded data, and typically ascribe to the ISO 9660 standard. CD's originated from audio applications, so the amount of information a CD can hold is measured in minutes:seconds:sectors. Each second contains 75 sectors, each of which can hold 2048 bytes (2 kilobytes) of Mode 1 user data. Recordable CD's presently are available in a variety of sizes, namely 21- (80 mm diameter), 63-, and 74-minute sizes (both 120 mm diameter), which can contain the following amounts of data in the CD-ROM format:

Factory-recorded CD's generally hold 74 minutes of audio or 650 MB of data. There are several overhead fields that must be deducted when calculating the total amount of data that you can fit on a CD: Session Lead-In and Lead-Out. The first lead-in and lead-out on a disc are not usually taken into consideration when calculating space available on disc, and they are considered to be outside the usable disc area.

Files on CD do not occupy a space exactly equal to their original size, because the minimum recordable unit on a compact disc is the logical block. Logical block size depends upon the size of the drive and is calculated by an intrinsic formula. The larger the drive, the larger the logical block size, hence the more space a given file will consume.

The more portable recording mediums such as the traditional floppy disks are slowly relenting to other mediums with greater capacity. With state of the art hard disks measured in gigabytes, and with multimedia and graphics file sizes often measured in tens of megabytes, a capacity of 100 MB to 150 MB is required whether moving a few files between systems, archiving or backing up individual files or directories, and sending files by electronic mail.

Magnetic tape data storage devices, also referred to as tape drives, have been used in the computer industry for years for the storage of large amounts of data. Tape drives have achieved preeminence as storage devices for portable storage and long-term and data backup purposes.

With respect to portable storage, devices such as Iomega's Zip drive provide use a technology developed by Iomega that draws the flexible disk upward towards the read/write head rather than moving the head toward the medium. Another portable scheme is LS-120, later termed SuperDisk, which resembles 1.44 MB 3.5 in disk, but uses a refinement of the floptical technology to deliver much greater capacity and speed. Named after the LS-120 laser servo technology it employs, an LS-120 disk has optical reference tracks on its surface that are both written and read by a laser system. These servo tracks are much narrower and can be laid closer together on the disk, wherein an LS-120 disk has a track density of 2,490 tracks per inch (tpi) compared with 135 tpi on a standard 1.44 MB floppy. Another option for portable storage is Sony's HiFD drive, having a capacity of over 200 MB per disk. Compatibility with conventional 1.44 MB floppy disks is provided by equipping the HiFD with a dual-head mechanism. When reading 1.44 MB floppy disks, a conventional floppy-disk head is used and comes into direct contact with the media surface. The separate HiFD head works more like a hard disk, gliding over the surface of the disk without touching it.

With respect to hard drives, most hard drives are multi-GB, whether removeable or fixed. For removable drives, there are various options including the Iomega Jaz drive, SyQuest's 1.5 GB SyJet and 1 GB SparQ. While generally similar to a hard disk the Jaz drive for example employs twin platters that sit in a cartridge protected by a dust-proof shutter which springs open on insertion to provide access to read/write heads. The Castlewood ORB was the first universal storage system to be built using magnetoresistive (MR) head technology, making them very different from other removable media drives that use older hard drive technology based on thin film inductive heads. MR hard drive technology permits a much larger concentration of data on the storage medium.

Optical storage using the blue laser is of considerable interest because the smaller wavelength of the drive's laser light limits the size of the pit that can be read from the disc, thus having a narrow beam that can read smaller dots. The DVD Forum's Steering Committee are promoting a proposed format--dubbed "Blu-ray Disc"--that is capable of providing storage capacities of up to 27 GB and 50 GB on single-layer and dual-layer discs respectively. The driving force behind such huge capacities is the emergence of multimedia applications in relation to both high-quality digital video and audio into the PC mainstream, coupled with the emergence of high-definition TV (HDTV), which is debuting in terrestrial broadcast systems.

Regardless of the technology underlying the portable disks, DVD's, CD's or tapes, or fixed/removable hard drives, including magneto optical, and rotating magnetic media the overall trend is to write/read in a compact format and optimize the space required with a stable and robust system.

Newer magneto-optical technology offers many improvements over conventional magnetic technology, particularly in terms of increased capacity. Magneto-optical storage systems also record data onto a recording material coated onto the surfaces of one or more rotating recording disks, but via different means than conventional drives. The recording material undergoes a sharp increase in magnetic susceptibility when heated beyond its Curie point, the temperature at which the magnetic properties of the recording material change from ferromagnetic to paramagnetic. A localized magnetic domain is created by heating a region of the recording material and then applying a magnetic field of a desired orientation to the heated region. When the recording material cools, the localized magnetic domain retains its magnetic orientation and again becomes far less susceptible to applied magnetic fields.

An optical fiber may guide an intense beam of focused laser light to heat a localized magnetic domain to be recorded or overwritten. The data stored in a particular localized magnetic domain may also be read back nondestructively by such a combined laser and optical fiber system. A low-powered, linearly polarized laser beam focused on a particular localized magnetic domain will be reflected with a Kerr rotation of the angle of polarization determined by the magnetic orientation of the localized magnetic domain. The pattern of polarization rotations read back as the low-powered laser beam moves across the recording surface thus represents the pattern of magnetic orientations previously written onto the recording surface. The overall reflectivity of a localized magnetic domain may also be determined via measurement of the relative amplitude of the reflected laser beam.

Magneto-optical storage systems quickly and reliably locate and align to any particular storage location on the recording disk, as with existing storage systems. A scheme for accomplishing these goals that takes advantage of the unique properties of a magneto-optical storage system is needed. An efficient system for encoding servo sector information is therefore important for maximizing the amount of remaining disk surface area available for data storage and retrieval.

The technology of tape drives has evolved from large, expensive open reel machines to the current generations of cassette tape drives, which store large amounts of data in convenient self-contained cassettes. Historically, open reel tape drives recorded data on parallel data tracks which extend along the length of the tape, and utilized fixed data recording/retrieval heads, i.e., one dedicated read/write head for each data track.

The actual recording and recovery of data on the tape medium is accomplished by a gap in the read/write head, and is in the form of magnetic flux reversals formed in the magnetic coating on the tape. To maximize the sharpness of the flux reversals and increase the amplitude of the read data pulses induced in the head during subsequent read operation, the length of the head gaps is aligned as precisely as possible with the direction of tape motion past the heads.

Historically, in order to ensure the integrity of data written on the tape, the tape drives included multi-gap heads, with one gap employed to write data and another gap, immediately trailing the write gap along the direction of tape motion, used as a read gap which could perform a read/verify operation on the data just recorded. If the tape drive was intended to record/recover data with the tape moving in both directions, an additional write or read gap was needed.

Several cassette-type tape drive formats are industry standards, including the format referred to as the QIC, or quarter inch cassette. In QIC format tape drives, data is recorded on a plurality of data tracks which extend parallel with the length of the tape as was typical in open reel type tape drives, but employ only a single recording/playback head which is controllably movable to each of the data tracks. A commonly used mechanism for controlling the movement of the head from track to track employs a worm gear driven by a stepper motor, with the pitch of the worm gear and the radial precision of the stepper motor determining the accuracy of head movement, including the repeatability of multiple head movements to any one given track.

One of the major factors controlling the overall storage capacity of tape storage devices is referred to as track density, typically defined in data tracks per inch of tape width, or how closely the data tracks are spaced. The greater the track density, the greater number of tracks that can be recorded on a given width of tape and the greater the overall cassette data capacity. A known factor limiting track density is referred to as adjacent track interference, which is the corruption or loss of data brought about when data on a given track is written at a location touching or even overlapping the previously recorded data on an adjacent data track. In such a situation, the amplitude of the readback signal can be reduced, and there is a limit to the amount of readback signal reduction which can be tolerated and beyond which data can be corrupted or lost completely.

Another factor controlling the ability of the tape drive to recover previously recorded data is a characteristic of the tape drive referred to as head azimuth, or simply azimuth, which is a measurement of the alignment between the longitudinal direction of the data tracks and the gap of the read/write head.

In the specifications defining the QIC tape drive and tape cassettes, one of the major planar surfaces of the cassette, called the cassette base plate, contains features which define a datum referred to as the tape cassette -B- plane. The tape cassette -B- plane is used, in conjunction with mating features on the tape drive which comprise a tape drive -B- plane, to define a mating surface between the tape cassette and the tape drive, and thus a base datum for defining the locations of both tape cassette and tape drive components and features along an axis normal to the common -B- plane. Because the data tracks extend along the length of the tape, the length of the head gap which accomplishes the recording and retrieval of data on the data tracks is nominally parallel with the length of the data track and thus also nominally parallel to the -B- plane. It is known, however, that small deviations from this nominally parallel relationship are introduced by component and manufacturing tolerances. It is this geometric relationship between the length of the head gap and the -B- plane which is referred to as azimuth. When the length of the head gap is parallel to the -B- plane, or, in other words, when the width of the gap is perpendicular to the -B- plane, azimuth is considered to be zero, with deviations from parallel in a first direction being referred to as positive azimuth and deviations in the opposite direction being referred to as negative azimuth. Non-zero azimuths are typically measured in units of rotation, such as minutes.

Tape drives used for recording video images have made use of this knowledge for several years to reduce intertrack interference and maximize the amount of storage on a given area of tape surface.

U.S. Pat. Nos. 5,307,217 and 5,371,638, for instance, (hereinafter the '217 and '638 patents, respectively) disclose apparatus and methods directed to recording data at opposite azimuth angles on adjacent data tracks in order to minimize intertrack interference, and thus maximize data capacity on tape media. There are, however, several differences in both the type of tape drive in which the disclosed method and apparatus are employed and in the specific apparatus which implements the recording of data at opposite azimuth on adjacent data tracks.

Thus, there is a growing demand for increased storage capacity and increased speed. What is needed therefore, is a method and apparatus that allows for more efficient storage and retrieval of digital data. More particularly, such an invention should allow digital recording with joint signal detection techniques for devices where efficient use of a storage medium is desirable in terms of optimizing device size, access speed and power consumption. Applications can include disk drives, tape systems, storage implementations for digital cameras, PDA's, and any related devices employing digital storage media regardless of whether the technology is magnetic, optical or future storage technology.

SUMMARY OF THE INVENTION

The invention is devised in the light of the problems of the prior art described herein. Accordingly it is a general object of the present invention to provide a novel and useful digital recording implementation that uses multi-user detection techniques that can solve the problems described herein. The digital recording devices of the present invention include, but are not limited to various forms of tapes, disks, disc drives and virtually any device that has recorded data on a medium that is extracted by a sensing element.

Accordingly, it is an object of the present invention to provide a digital recording system consisting of a storage medium, a data writing element, a data reading element capable of reading closely spaced interfering bits written on the storage medium, a read signal processing element and appropriate servo mechanisms and controllers for engaging the read and write elements with the storage medium. Joint signal detection is known to separate interfering digital signals in the same channel, provided that the interfering signals have sufficient power relative to the noise floor in the channel. The present invention applies this technique to mitigate adjacent track interference by simultaneously demodulating signals arising from multiple closely spaced tracks underneath the reading element. It has been contemplated and within the scope of the invention to include read-only devices that would eliminate the writing elements and simplify the invention.

Another object of the present invention is to provide digital recorders that can recover data from media packed so densely that adjacent bits written to the medium interfere when accessed by the read mechanism.

Yet another object of the present invention is to provide digital recorders that can simultaneously demodulate multiple digital bits with a single read sensor.

A further object of the present invention to provide data storage scheme that provide parallel access to related parts of data files stored on the recorder to increase bandwidth.

A still further object of the present invention is to provide digital recorders with higher data stor