Abstract: A discrete-time sliding mode controller (SMC) controls the motion of mechanical apparatus such as a read head in a disk storage system. The overall control effort is generated by combining a linear control effort with a discrete-time sliding mode control effort generated by switching between gains in order to drive the system's phase states toward a sliding line trajectory. An estimate of a reference is generated and added to the control effort, thereby providing an approximation of the derivative of the reference signal. Estimating the reference signal may be performed by various methods, depending on whether the reference is known, unknown but repeatable, or completely unknown and non-repeatable. A least-mean-square (LMS) algorithm is employed to estimate the reference signal by computing coefficients of a function which minimizes a particular system parameter.

Abstract: A method and apparatus is disclosed for converting raster images into a vector format by identifying and converting the borders of image features into a mathematical format referred to as string sequences. To enable the string sequencing of color images, the present invention identifies and resolves contrast conflicts in the image features in order to avoid misperceiving the vectorized image when converted back into a raster format. A contrast conflict occurs when there is a “contrast tie” between overlapping features of the image. The feature that would normally be perceived as the dominant feature breaks the contrast tie so that when the vector image is reconstructed, the dominate feature appears in the foreground of the image while the recessive feature appears in the background.

Abstract: A sampled amplitude read channel for use in disk storage systems (magnetic or optical) is disclosed comprising a simplified branch metric calculator for use in a trellis sequence detector. Instead of computing the traditional Euclidean branch metric as the squared difference between the actual signal sample and the expected signal sample of the target partial response, the present invention computes a simplified branch metric which is then saturated in order to reduce the number of bits required to calculate and store the branch metrics, thereby simplifying the branch metric calculators as well as reducing the add-compare-select (ACS) circuitry for each state in the trellis. Furthermore, the saturation technique of the present invention is substantially data independent meaning that the saturation threshold is essentially independent from the signal samples used to compute the branch metric.

Abstract: A sampled amplitude read channel for magnetic disk recording which asynchronously samples the analog read signal, adaptively equalizes the resulting discrete time sample values according to a target partial response, extracts synchronous sample values through interpolated timing recovery, and detects digital data from the synchronous sample values using a Viterbi sequence detector is disclosed. To minimize interference from the timing and gain control loops, the phase and magnitude response of the adaptive equalizer filter are constrained at a predetermined frequency using an optimal orthogonal projection operation as a modification to a least mean square (LMS) adaptation algorithm. Further, with interpolated timing recovery, the equalizer filter and its associated latency are removed from the timing recovery loop, thereby allowing a higher order discrete time filter and a lower order analog filter.

Abstract: A sampled amplitude read channel is disclosed for disk storage systems comprising an MEEPR4 equalizer, an MEEPR4 sequence detector matched to an RLL d=1 constraint, and an encoder/decoder for implementing a channel code that codes out (2,4,2) quasi-catastrophic data sequences. A rate n/m finite state encoder encodes n bits of user data into m bits of write data, and a sliding block decoder decodes m bits of read data into n bits of estimated user data. The encoder uses the current n bits of user data as well as a current state of a state machine to generate the m bits of write data, where a state-splitting technique is employed to achieve a high code rate in a practical, cost effective implementation. The decoder decodes the m bits of read data into the n bits of estimated user data by evaluating the current detected codeword in context with the following detected codeword.

Abstract: A sampled amplitude read channel is disclosed for disk storage systems (e.g., magnetic or optical) comprising a sampling device for sampling the analog read signal emanating from the read head positioned over a disk storage medium, a channel equalizer for equalizing the signal samples according to a desired partial response, a trellis sequence detector for detecting a preliminary sequence from the equalized signal samples, and a post processor for correcting errors in the preliminary sequence, including errors caused by the channel equalizers correlating the noise in the read signal. The preliminary sequence detected by the sequence detector is remodulated into ideal partial response samples and then subtracted from the actual signal samples to generate a sequence of sample errors. The sample errors are then filtered by a sample error filter, and the filtered sample errors are correlated with error event sequences corresponding to the most likely error events of the trellis sequence detector.

Abstract: An adaptive, discrete-time sliding mode controller (SMC) is disclosed which detects and adapts to gain variations in the controlled plant. The overall control effort is generated by combining a linear control effort with a discrete-time sliding mode control effort generated by switching between gains in order to drive the system's phase states toward a sliding line trajectory. A sliding mode variable &sgr;k defines the position of the system phase states relative to the sliding line. The SMC controller is designed such that the sliding mode variable &sgr;k crosses the sliding line and changes sign at every sample interval. For the nominal plant gain, the SMC controller is also designed such that the magnitude of the sliding mode variable &sgr;k+1=−&sgr;k will remain constant (&sgr;k+1=−&sgr;k) and substantially constrained to |&sgr;k|=&Dgr;/(1+&lgr;) where &Dgr; and &lgr; are predetermined design constants.

Abstract: A sampled amplitude read channel is disclosed for disk storage systems comprising a encoder/decoder for implementing a high rate channel code that codes out specific minimum distance error events of a trellis sequence detector by enforcing a particular code constraint. The trellis sequence detector comprises a state machine matched to the code constraint which effectively removes the corresponding minimum distance errors from the detected output sequence. Additionally, the channel code encodes redundancy bits into the write data for implementing an error detection code. The redundancy bits are processed during a read operation to generate an error syndrome used to detect and correct other dominant error events, such as the NRZ (+) and (+−+) error events. In this manner, the most likely error events of the trellis sequence detector are either coded out by the channel code constraint, or detected and corrected using the error syndrome.

Abstract: A sampled amplitude read channel for optical disk storage systems is disclosed comprising an all digital timing recovery circuit. The RF read signal from the read head is sampled asynchronous to the baud rate and the asynchronous sample values are interpolated to generate sample values that are substantially synchronous to the baud rate. A data detector, such as a Viterbi sequence detector, processes the synchronous sample values to generate an estimated binary sequence representing the recorded binary sequence. The timing recovery circuit comprises a baud rate estimator for estimating the baud rate relative to the sampling rate, wherein the estimated baud rate is used to initialize a timing recovery loop filter at the end of seek operations. The all digital timing recovery circuit and baud rate estimator enable the storage device to begin reading the user data immediately after a seek operation, rather than wait for the CLV servo loop to acquire the target spindle speed.

Abstract: A partial response class-IV (PR4) sampled amplitude read channel is disclosed for detecting user data and embedded servo data. The detected servo data is encoded using a novel servo code capable of accurately decoding detected codewords representing servo track address during seek operations, even when the recording head flies between two adjacent tracks, and capable of correcting errors in the detected codedwords caused by noise in the read signal, such as inter-symbol interference. In one embodiment, the servo code corrects certain minimum distance error events, such as a bit shift error event, associated with a trellis type sequence detector. To achieve the equivalent effect of a conventional Gray code, the codewords are arranged such that adjacent track addresses differ by a number of bits relative to the minimum distance error events corrected.

Abstract: A sampled amplitude read channel is disclosed for reading data recorded on a disk storage medium by asynchronously sampling an analog read signal, equalizing the asynchronous sample values according to a desired partial response, and interpolating the equalized sample values to generate synchronous sample values substantially synchronized to a baud rate of the recorded data. The read channel further comprises a gain control circuit which generates a gain error for adjusting the amplitude of the analog read signal to a nominal value through a variable gain amplifier (VGA). During acquisition, the gain error is computed from the asynchronous sample values at the output of the sampling device in order to avoid the delay associated with the discrete equalizer filter and the timing recovery interpolation filter. This decreases the acquisition time and the corresponding length of the acquisition preamble, thereby reserving more area on the disk to record user data.

Abstract: A sampled amplitude read channel is disclosed for reading data recorded on a disk storage medium by detecting an estimated binary sequence from a sequence of discrete time sample values generated by sampling pulses in an analog read signal from a read head positioned over the disk storage medium. The read channel comprises a sampling device, such as an analog-to-digital converter (A/D), for sampling the analog read signal to generate the discrete time sample values and for sampling at least one other auxillary analog input signal, such as a servo control signal. In this manner, performance characteristics of the read channel can be measured, such as the driving current applied to the servo control voice coil motor (VCM), without requiring additional hardware.

Abstract: A cost reduced finite field processor is disclosed for computing the logarithm LOG.sub..alpha. (.alpha..sup.j) of an element of a finite field GF(2.sup.n) using significantly less circuitry than that required by a lookup table typically employed in the prior art. The result of the logarithm (i.e., the exponent of .alpha..sup.j) is represented as a binary number computed serially one bit per clock cycle. In one embodiment, combinatorial logic is used to compute bit 0 of the exponent. On each clock cycle, the exponent is shifted once to the right and bit of the exponent is extracted until the entire exponent has been computed. Shifting the exponent of a field element to the right is carried out by taking the square root of the element. The present invention requires at most n+1 clock cycles to compute LOG.sub..alpha. (.alpha..sup.j), with one embodiment requiring n/2 clock cycles.

Abstract: An error correction processor is disclosed for correcting errors in randomized data read from a disk storage medium, where the randomized data comprises ECC redundancy symbols generated over the randomized data and check symbols generated over data before being randomized. The error correction processor comprises an ECC decoder for correcting errors in the randomized data using the ECC redundancy symbols; a syndrome generator, responsive to the randomized data, for generating a validation syndrome; a correction validator for comparing the validation syndrome to a predetermined value to verify the validity and completeness of the corrections to the randomized data; and a derandomizer for derandomizing the randomized data after the correction validator indicates that corrections to the randomized data are valid and complete.

Abstract: A sampled amplitude read channel is disclosed for disk storage systems employing a run-length limited (RLL) d=1 channel code which compensates for partial erasure, and a parity channel code for enhancing the operation of a remod/demod sequence detector. During a write operation, after encoding the user data into codewords comprising the RLL d=1 constraint, the parity over one interleave of a block of NRZI bits is computed and two parity bits appended to form a parity codeword. For an even number of "1" bits in the block, the parity bits are set to "00". For an odd number of "1" bits in the block, the parity bits are set to "10" if the codeword ends with a "0" bit and to "01" if the codeword ends with a "1" bit, thereby maintaining the RLL d=1 constraint. Thus, a parity codeword will always comprise an even number of "1" bits (even parity).

Abstract: A multi-layered error detection and correction (EDAC) system is disclosed for processing an error correction code (ECC) typically employed in optical disk storage devices. A first layer of the EDAC system includes a primary ECC, such as a multiple burst Reed-Solomon code, and a second layer incudes a secondary ECC, such as a CRC code, for use in verifying the validity of the corrections made using the primary ECC. The primary ECC is multi-dimensional and, in the embodiment disclosed herein, it is a two-dimensional P/Q product code typically employed in a CD-ROM storage device. The secondary ECC operates in unison with the primary ECC. As the EDAC system processes and corrects the data using the primary ECC, the EDAC system also simultaneously updates the secondary ECC. In this manner, when the EDAC system is finished processing the data using the primary ECC, the validation syndrome generated by the secondary ECC is available immediately for checking the validity of the corrections.

Abstract: An error correction processor is disclosed for correcting errors in binary data read from a disk storage medium, wherein the binary data comprises a first and second set of intersecting ECC codewords of a multi-dimensional codeword. The error correction processor comprises a data buffer for storing the ECC codewords read from the disk storage medium; a syndrome generator for generating ECC syndromes in response to a codeword in the second set; an error-locator polynomial generator for generating an error locator polynomial .sigma.(x) in response to the ECC syndromes; a selector for selecting between the error-locator polynomial .sigma.(x) and an erasure polynomial .sigma.(x).sub.EP, wherein:(i) the erasure polynomial .sigma.(x).sub.EP is generated while processing the first set codewords; and(ii) the erasure polynomial .sigma.(x).sub.EP is used to correct at least two codewords in the second set; andan error corrector for generating correction values in response to either the error-locator polynomial .sigma.

Abstract: A sub-baud rate write circuit is disclosed which writes RLL encoded channel data to a disk storage medium using a write clock frequency significantly below the baud rate. This allows for a higher channel data rate without increasing the cost and complexity of the write circuitry. The write circuitry operates by re-encoding the RLL encoded channel data according to a particular mapping to generate write data at the write clock rate, and then writing the write data to the disk at appropriate phase delays. The phase delays are generated by passing the write clock through an array of delay circuits. The resulting write signal is the same as if the RLL encoded data were written to the disk using a baud rate write clock. The write circuitry of the present invention is ideally suited for use in a sub-sampled read/write channel where the object is to reduce the cost and complexity by clocking the entire channel at a frequency significantly below the baud rate.

Abstract: In a magnetic disk storage system, a sampled amplitude read channel is disclosed that employs a fault tolerant sync mark detector for detecting a sync mark from the channel samples in order to synchronize a time varying sequence detector. The read channel preferably employs PR4 equalization for timing recovery and gain control, and EEPR4 equalization for sequence detection. The EEPR4 sequence detector operates according to a time varying state machine matched to a predetermined trellis code constraint. Because the state machine is time varying, the data stream must be synchronized at the input of the sequence detector rather than at the output as in the prior art. The present invention provides a fault tolerant sync mark detector that detects a sync mark from the EEPR4 channel samples before being input into the sequence detector.

Abstract: A synchronous read channel is disclosed which samples an analog read signal from a magnetic read head positioned over a magnetic disk medium, filters the sample values according to a desired partial response, extracts timing information from the filtered sample values, and detects an estimated data sequence from the filtered sample values using a discrete time sequence detector. Discrete time timing recovery within the read channel comprises a variable frequency oscillator (VFO) for generating a sampling clock. A center operating frequency of the VFO is adjusted through a programmable register which stores a digital center frequency setting. A phase error is computed from the sample values and combined with the center frequency setting to control the frequency and phase of the sampling clock at the output of the VFO.