JP2001266491A - Average sample determination for preamble detection - Google Patents

Abstract

PROBLEM TO BE SOLVED: To efficiently detect a preamble in data. SOLUTION: A preamble detector determines a preamble bit corresponding to a received sample z(n) by slicing the received sample z(n) at a threshold. The preamble detector is provide with an option filter, averaging circuit and comparator. Before generating a determination d(n) by the comparator concerning the input sample z(n), an average sample zfa(n) is generated by averaging filtered samples zf(n) and zf(n-k) by adopting the averaging circuit so that reliability in the determination to be performed concerning the correspondent sample can be improved. As one implementation style, as 4-sample delay bank provided with four flip-flops, adders and option output buffers to be used for storing the average sample zfa(n) to be provided to the comparator is adopted for the averaging circuit.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to detection of data in a communication signal, and more particularly, to detection of a preamble in the data.

[0002]

2. Description of the Related Art Many digital transmission systems generally employ techniques for detecting digital data represented by a sequence of symbols, each symbol consisting of a group of bits. The symbol bits are transferred as a signal through a transmission (ie, communication) channel, where noise is usually added to the transmission signal. For example, a magnetic recording system first encodes data into symbol bits that are recorded on a magnetic medium. Writing, recording, and reading data from magnetic media can be thought of as occurring over a transmission channel that has an associated frequency response. The signal can then be read from the magnetic medium as a sampled signal (ie, a sequence of output samples) representing stored data or stored symbol bits. A magnetic recording system for a disk drive reads and detects data from a track on a magnetic medium such as a magnetic disk. Each track includes user (ie, “read”) data sectors, as well as system-specific control (eg, “servo”) data sectors embedded between the read data sectors. Servo data sectors store servo data in the form of control data used by the recording system to 1) search for tracks during the seek mode and 2) position the read head on a particular track on magnetic media. I do. Some prior art magnetic recording systems employ digital signal processing to detect stored servo data,
Some use analog technology.

[0003] Techniques for reading data often employ an easily detected sequence called a preamble sequence at the beginning of the data sequence. For example, a primling sequence is used for servo data processing in a magnetic recording and reproducing system. FIG.
Shows the servo data processing of the magnetic recording system 100. Part of the servo data is received by a servo data encoder 101 (shown as a 1 / N encoder) that encodes the part using a rate 1 / N code. The resulting encoded portion of servo data and the remaining unencoded portion are further processed by magnetic write head 102 and recorded on magnetic media 110. The magnetic read head 103 reads information from the magnetic recording medium 110 as an analog signal.

[0006] The magnetic read head 103 includes a medium 110.
The analog signal may be sampled to provide data read from the as output channel samples.
The term "output channel samples" indicates that data has been passed through a transmission channel (eg, magnetic media 110) having a form of frequency response (and possibly having memory). This type of transmission channel (possibly including the frequency response of a subsequent equalizer) can be referred to as a partial response (PR) channel. The signal representing the encoded servo data has additional noise components and additional signal distortion caused by passing the signal through the frequency response of the channel. Output channel samples can be applied to the equalizer 104 to partially correct for variations in the frequency response of the channel, or frequency response characteristics of the circuitry of the magnetic read head 103.
The equalized output channel samples are then applied to a partial response maximum likelihood (PRML) detector 105.

[0005] The PRML detector 105 provides the detected encoded servo data to a servo data decoder 106. The servo data decoder 106 performs the reverse of the encoding process, and decodes the decoded servo address mark (SA
M) provide data and gray data. FIG.
Shows a burst demodulator 107 for demodulating uncoded burst data read from the medium 110. The burst data is transmitted to the magnetic read head 103
Can be used by the magnetic read head 103 to detect whether it is positioned directly over the center of a track on the medium 110.

FIG. 2 shows a format 200 for recording servo data in a servo data sector of the magnetic recording medium 110.
It is shown. The servo data may include a preamble 201, which is a bit sequence from which timing and gain information is recovered. The timing and gain information allows the magnetic read head 103 to obtain gain and phase lock on incoming analog signals coming from tracks on the magnetic medium 110. FIG. 2 shows a burst demodulation field 20 including burst data.
4 is also shown.

[0007] The preamble 201 may be followed by an encoded servo address mark (SAM) 202, followed by encoded gray data 203 for servo sectors. The SAM 202 includes a predetermined bit pattern to identify the sector as containing servo data, and resets the frame indication clock used by the magnetic read head 103 to read a track / sector from the magnetic recording medium 110. This can be adopted. The gray data 203 indicates the track number and cylinder information of the magnetic recording medium, and can be used to prevent an error when the magnetic read head 103 reads an adjacent track during the seek mode. The SAM 202 and the gray data 203 are portions of servo data that are usually encoded as a symbol bit sequence before being recorded on the magnetic recording medium 110.

[0008]

Usually, the preamble 2
01 is a short sequence having a repeating bit pattern. Since the preamble 201 is usually a short sequence, the maximum likelihood detection technique is not generally used for detecting the preamble 201. This is because such detection introduces a relatively long delay. In addition, some channels have multiple signal paths (ie,
Parallel averaging detection may be employed to collect the same signal detected through the channel) and average the signal together. Averaging the signals may allow for improved detector performance when the noise added in each signal path is not correlated. However, preamble 201
For many applications, such as reading control data from a recording medium, the PR channel does not include a plurality of detected signal paths.

[0009]

SUMMARY OF THE INVENTION The present invention relates to a technique for detecting a preamble in data received from a medium as a sample sequence. According to an embodiment of the present invention, the repetition pattern characteristic of the preamble sequence allows the detector to use a plurality of samples. The plurality of samples can be averaged to provide an average sample, from which a determination is made for one or more input samples.

According to an exemplary embodiment of the present invention, a preamble of data including a value of a periodic pattern generates a current sample of a sample sequence representing the preamble and, based on the periodic pattern, one of the sequences.
It is detected by selecting one or more additional samples. The current sample is combined with one or more selected samples to form an average sample, and the average sample is compared to a threshold to make a determination about the value of the current sample. The preamble is detected based on the decision. For a further embodiment, the process is repeated to generate a decision sequence corresponding to the sample sequence, and the presence or absence of a preamble is detected based on the decision sequence.

[0011] Other aspects, features, and advantages of the present invention include:
The following detailed description, the appended claims, and the accompanying drawings will provide a better understanding.

[0012]

FIG. 3 illustrates a detection system 300 comprising a preamble detector 304 operating in accordance with an exemplary embodiment of the present invention, such as may be employed on the reading side of a magnetic recording system. The read head 301 receives or reads information from the medium 303 and generates a corresponding analog signal. An optional equalizer 302 may be employed to correct for channel variations and memory. The PR channel is the medium 30
3 and a read head 301, and an optional equalizer 302 response. The analog signal is sampled to provide a data sample component y (n) corresponding to the data read from the medium, as well as a noise sample component w (n) corresponding to the additive noise power. Such sampling is performed by the read head 30
1. It can occur in the optional equalizer 302 or in the preamble detector 304. Thus, preamble detector 304 receives sample z (n) = y (n) + w (n), where n represents the nth sample value. The output sample sequence is used by a preamble detector 304 to detect a preamble pattern, and thus a preamble, in the control data read from the medium 303.
Provided to When the control data preamble sequence is detected, the encoded data symbol and the encoded servo address mark (SA) read from the medium are read.
M) and maximum likelihood (eg, Viterbi) detector 305 employed to detect and decode the encoded gray data.

Next, a preferred embodiment employed for detecting a preamble in control data for servo processing in a magnetic recording system will be described, but the present invention is not limited to this. Therefore, to those skilled in the art,
It will be apparent that preamble detection according to the exemplary embodiments described herein may be extended to other systems employing preamble detection.

For the exemplary embodiment described herein, the pattern used in the preamble is P
This is a 2T pattern known in the field of magnetic recording using the R channel. The 2T pattern consists of a bit sequence of the preamble (ie, “001100110011.
0011 ”) is periodically repeated (period K
At) the bit sequence "0011". The received sample value z (n) = y (n) + w (n) represents one bit of the preamble, which value depends on the type of PR channel. For an exemplary embodiment, see PR4
The channel adopts a 2T pattern “0011” for the preamble pattern, and this bit pattern is composed of the (normalized) sample sequence {1, 1, −1, −
1}, where the period of the preamble pattern sequence is K = 4. As will be appreciated by those skilled in the art, other longer or shorter repeating sequences may be employed in the preamble pattern, and the period and bit values of the preamble sequence will depend on the particular implementation.

The preamble detector 304 slices the received sample z (n) at a threshold,
The preamble bit corresponding to the received sample z (n) is determined. The preamble detector 304 operating in accordance with the exemplary embodiment of the present invention is shown in FIG. 4 and includes an optional filter 410, an averaging circuit 411, and a comparator 412.

For the exemplary PR4 channel, equalizer 3
The output sample z (n) of 02 (FIG. 3) is a positive or negative peak value, and the threshold “0” is used for the slice operation. (For example, if the sample value z (n) is greater than 0, the preamble decision d corresponding to the sample y (n)
(N) is “0”, but if the sample value z (n) is less than 0, the preamble determination d (n) is “1”. Such determination of the slicing operation can be achieved using the comparator 412. For other PR channels, where the sample value z (n) does not consist solely of the positive and negative peak values, the preamble detector 304 employs an optional filter 410 to provide the sample z
(N) may be filtered. The filtered samples, represented by zf (n), are each compared with a threshold “0” to determine a “1” or “0” d.
It has only positive and negative peak values that can result in (n). Filtered sample zf
(N) can also be expressed as yf (n) + wf (n),
Here, yf (n) and wf (n) are obtained by filtering the preamble data and the noise sample component, respectively. For a particular implementation, sample z
(N) is that each sample z (n) is equalized and contains N bits, since the analog / digital converter (ADC) is generated by sampling the analog signal generated from the medium 303. Add additional equalization noise associated with the sampling process.

According to the present invention, the averaging circuit 411 is employed to average two or more filtered samples to generate an average sample zfa (n) prior to the slicing operation, thereby obtaining a corresponding sample zfa (n). Can improve the reliability of decisions made on Average sample zfa
(N) is an input sample z by the comparator 412.
Regarding (n), the decision d by the comparator 412
Adopted to generate (n). In the PR channel, multiple versions of the signal are not available, but due to the periodic (repeated sequence) nature of the preamble pattern, each sample corresponding to the same bit value is determined and detected by the preamble detector 304 at each step. To be available in series.

For any preamble sequence having a pattern repeated with a period K, the signal component yf (n) of the sample z (n) separated by K bits (or symbols) represents the same value of the preamble sequence. Therefore, two or more separated samples K (zf
(N), zf (n-4),. . . ) Is averaged,
The average sample z corresponding to the first input sample z (n)
generate fa (n). If K = 4 and a 2T pattern is used in the channel preamble,
The average sample zfa (n) is given by Equation 1.

[0019]

(Equation 1)

An exemplary implementation of the averaging circuit 411 employing the averaging provided by Equation 1 is shown in FIG.
The averaging circuit 411 includes buffers (for example, flip-flops) 501a to 501d and adders (for example, (N +
1) Bit adder) 502 and average sample zfa
Employ a four-sample delay bank with an optional output buffer (eg, flip-flop) 503 used to store (n). As shown in FIG. 5, the adder 502 combines the sample zf (n) filtered in the buffer 501d with the sample zf (n−4) filtered, and stores the result in the buffer 503. The arithmetic average is buffer 50
Sometimes divided by 2 in 3, this division is
As long as slicing at the zero crossing, the comparator 412
Is a constant that does not affect the slicing operation. Thus, while the preferred embodiment described herein ignores the divider, other embodiments set the threshold of comparator 412 to take into account the divider corresponding to the averaged number of samples. It may be adjusted.

As a measure of the improved performance of a preamble detector that employs averaging of periodic sample values, reliability can be employed. The reliability in generating the decision d (n) from z (n) depends on the distance between the positive and negative peak values of the (filtered) data sample component yf (n) and the filtered noise sample component wf (N) as a function of the standard deviation σ _wf .
For any initial gain or phase error introduced, the reliability of the decision d (n) is reduced. However, implementations typically employ signal feedback for synchronization, which removes gain and phase errors, to restore the reliability performance of the decision process (ie, feedback depends on the noise component only for decision errors). Tend to be).

For a partial response channel employing a 2T pattern for a preamble having a period K = 4, the reliability is:
It can be calculated as follows: The sequence of filtered data samples yf (n) is, for example,
2 distance dist _yf (ie, “1” and “−1”
The magnitude of the difference between them has positive and negative peak values separated by 2). The standard deviation of the noise component is σ _wf . The reliability of the decision d (n) when only one sample zf (n) undergoes a slicing operation can be expressed as a bit error rate (BER) given by equation (2).

[0023]

(Equation 2) Where Q is a Q function well known in the mathematical arts.

For the average sample zfa (n), since the data sample yf (n-4) represents the same value as yf (n), the positive and negative peaks of yfa (n) have a distance dist _{yf of} 4 (Ie, the magnitude of the distance between “2” and “−2” is 4). The standard deviation σ _wfa of the additional noise sample component is √2σ _wf . Therefore, the reliability of the decision d (n) when the average sample zfa (n) undergoes the slice operation can be expressed as a bit error rate (BER) given by Expression 3.

[0025]

(Equation 3)

When the BER value is expressed in dB, the decision reliability for the average sample is improved by 3 dB over the reliability of the decision using only a single sample. The improvement in the reliability of the decision adds complexity to the circuit shown in FIG. 5: four N-bit delay elements for storing four samples;
At the cost of an (N + 1) -bit adder that combines the two filtered samples; and a buffer that stores the result.

As will be apparent to those skilled in the art, the averaging is
It can be extended to an average sample generated from a larger number of samples. For the exemplary embodiment described,
Each time the number of samples to be averaged doubles, the decision d
A 3 dB improvement in reliability of (n) can be realized. The average sample zfa (n) using the four filtered input samples is defined by Equation 4 and the reliability improvement increases the complexity of the averaging circuit (for example, the implementation of averaging circuit 411 includes: A delay bank of 12 buffers is needed instead of 4).

[0028]

(Equation 4)

The 3 dB improvement in reliability is the calculated ideal value, and the calculated reliability is that the noise sample component wf (n) is uncorrelated (ie, the spectrum of the additional noise is white). Assume. However, implementations may generally correlate the noise sample components wf (n) since the noise introduced into the PR channel may be filtered by the equalizer 302 (FIG. 3) as well as the filter 410 of the preamble detector 304. Therefore, improvement in reliability can be reduced.

Referring back to FIG. 3, the preamble decisions generated by the preamble detector 304 are collected within a time window for the overall preamble sequence decision. In the PRML servo system, the (Viterbi) detector 305 uses the encoded servo address mark (S
AM) 202 and encoded servo gray data 203
(FIG. 2) is detected. SAM detection performance is characterized by a SAM miss rate (SMR), with smaller SMR values indicating better performance. Since the operation of the detector 305 is synchronized with the received servo data samples based on the preamble detection, the SMR depends on the performance of the preamble detection. The preamble detection performance is characterized by a preamble detection miss rate (PMR).
The smaller the value, the better the performance.

FIG. 6 shows simulation results for the SNR for a preamble detector that produces a decision on single and average samples for SMR and PMR. As shown in FIG. 6, an average sample is used to generate a preamble detection decision that provides a relatively good PMR value given a relatively poor signal to noise ratio. With respect to the simulation results shown in FIG. 6, a preamble sequence having a period K = 4 is
Used for the preamble sequence. The two wavy curves are SMRs obtained with two different encoding methods for SAM and gray data: Code 1 and Code 2
Represents Code 1 has less coding gain than code 2 (ie, "less powerful"). The solid curves represent PMRs obtained with different detection schemes. S-PM
The curve labeled R shows the PMR obtained by using a preamble detection decision based on a slicing operation using a single sample zf (n). The S-PMR curve is
It is well below the curve labeled Code 1 SMR. Therefore, preamble detection based on a slice operation using a single sample zf (n) does not limit the performance of the system when Code 1 is used.

However, a more powerful code, Code 2
, The benefit of the more powerful code, Code 2, is limited by the PMR performance of preamble detection, since the S-PMR curve is above the SMR curve of Code 2. The average PMR curve is labeled AVG-PMR, uses two samples, and the average sample zfa (n) employed in the slicing operation for determination, such as that described with respect to the exemplary embodiment of FIGS. 3-5. Generate Averaging according to an exemplary embodiment of the present invention is 14d
Improving the decision reliability in B's SNR by a power of 10;
The higher the SNR, the greater the improvement. Thus, the PMR performance of the AVG-PMR curve is
Better than MR performance, allowing the system to use all the benefits of the more powerful Code 2.

While the exemplary embodiments of the present invention have been described with respect to encoding and decoding methods, the invention is not so limited. As will be apparent to those skilled in the art, various methods may be implemented as functions of circuit elements, and may also be implemented as processing steps in a software program. Such software can be employed, for example, in a digital signal processor, microcontroller, or general-purpose computer.

The invention can be embodied in the form of methods and apparatuses for practicing these methods. The present invention also provides a floppy disk, CD-ROM,
It can also be embodied in the form of program code embodied in a tangible medium, such as a hard drive or any other machine-readable storage medium, wherein the tangible medium is
When the program code is loaded on a machine such as a computer and executed by the machine, the machine becomes an apparatus embodying the present invention. The present invention can also be embodied in the form of program code transmitted over a transmission medium, such as, for example, via electrical wiring or cable, optical fiber, or electromagnetic radiation, where the storage medium Stored in, loaded into, and / or executed by a machine, or when the program code is loaded and executed on a machine, such as a computer, the machine becomes an apparatus embodying the present invention. When implemented on a general-purpose processor, the program code segments combine with the processor to provide a unique device that operates analogously to specific logic circuits.

Those skilled in the art will appreciate that the details, materials, and parts described and illustrated to illustrate the nature of the invention without departing from the principles and scope of the invention as set forth in the appended claims. It will further be appreciated that various changes in the sequence may be made.

[Brief description of the drawings]

FIG. 1 shows the servo processing of a prior art magnetic recording system.

FIG. 2 shows a format of servo data recorded in a servo data sector of a magnetic recording medium for the system of FIG.

FIG. 3 illustrates a data detection system including operation of a preamble detector, according to an exemplary embodiment of the present invention.

FIG. 4 illustrates an exemplary embodiment of the preamble detector of FIG.

5 shows an exemplary implementation of the averaging circuit of the preamble detector shown in FIG.

FIG. 6 shows simulation results of the preamble detector of FIG. 5 generating a decision on the signal and the average sample.

Claims (13)

[Claims] 1. A method for detecting, in a sample sequence, a preamble including a value of a periodic pattern, comprising: (a) based on the periodic pattern, a current sample and one or more in a sequence corresponding to the preamble. Selecting a plurality of additional samples; (b) combining the current sample with the one or more selected samples to form an average sample; and (c) comparing the average sample to a threshold. Forming a decision on the value of the current sample; and (d) detecting the preamble based on the decision. 2. (e) repeating steps (a) to (c) to generate a decision sequence corresponding to the sample sequence; and (f) detecting the preamble sequence based on the decision sequence. The method of claim 1, further comprising: 3. The method of claim 1, further comprising filtering each sample sequence to bias each corresponding value as a positive or negative peak value. 4. The method according to claim 1, wherein the method is used in a magnetic recording / reproducing system to detect the preamble of servo data read from a medium. 5. The method of claim 1, implemented in an integrated circuit as a processor step. 6. A circuit for detecting, in a sample sequence, a preamble including a value of a periodic pattern, comprising: a current sample from the sample sequence representing the preamble and one or more selected samples; A circuit comprising: an averaging circuit that forms an average sample; and a comparator that compares the average sample with a threshold to detect the presence or absence of the preamble. 7. The delay bank for storing one or more of the sample sequences representing the periodic pattern of the preamble, wherein the one or more additional samples in the sequence are: A delay bank provided from the delay bank based on a periodic pattern; and an adder for combining the current sample with the one or more selected samples from the delay bank to form an average sample. The circuit of claim 6, comprising: 8. The circuit according to claim 7, wherein a decision sequence corresponding to the sample sequence is generated, and the presence of the preamble is detected based on the decision sequence. 9. The system of claim 7, further comprising a filter for biasing each corresponding value to a positive or negative peak value.
The described circuit. 10. The circuit according to claim 7, wherein the circuit is used for detecting the preamble of servo data read from a medium in a magnetic recording / reproducing system. 11. The circuit according to claim 7, implemented in an integrated circuit. 12. A computer readable medium having stored thereon a plurality of instructions which, when executed by a processor, cause the processor to detect a preamble including a value of a periodic pattern in a sample sequence. Instructions, the method comprising: (a) selecting a current sample and one or more additional samples in a sequence representing the preamble based on the periodic pattern; and (b) the current sample. Combining with the one or more selected samples to form an average sample; and (c) comparing the average sample to a threshold to form a decision on the value of the current sample. (D) detecting the preamble based on the determination. Computer-readable media. 13. (e) The steps (a) to (c)
14. The computer-readable medium of claim 12, further comprising: repeating (c) generating a decision sequence corresponding to the sample sequence; and (f) detecting the preamble sequence based on the decision sequence. .