JPH05325432A - Magnetic disk device - Google Patents

Abstract

PURPOSE:To enhance the reliability of recording/reproducing data in spite of increasing an error rate in accordance with the deterioration of S/N of read signal by a high track pitch corresponding to miniaturization and high capacity, and high recording density. CONSTITUTION:As for an error correction of a data part, an error correction by a strong read Solomon code is executed by an ECC circuit 9, and as for an ID part, an error countermeasure by the triplicity of recording and majority logic is used by an ID data majority circuit 10. Up to the present, only the error correction of the data part is strengthened, and as for read of the ID part, only an error detection by a CRC is executed. Therefore, an error of the ID part is only expected to recover by retry, and when a defect is occurred in the ID part, the defective sector have to be registered as a defect and resigned. However, according to this invention, correct ID information can be obtained by multiplexing, and the recording capacity and the performance are not deteriorated.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a magnetic disk device used as a storage device for a computer or the like, and more particularly to a small-sized high-capacity magnetic disk device for which demand has rapidly increased in recent years.

[0002]

2. Description of the Related Art In recent years, in magnetic disk devices, technological innovations such as the appearance of small slider thin film magnetic heads, advances in high-smoothness magnetic recording media, performance improvement of precision positioning servo technology, and evolution of signal processing / modulation system have been made. The improvement in performance and recording density per unit area is remarkable.

Further, the trend toward miniaturization due to these technological innovations is also rapidly advancing, from 3.5 inches to 2.5 inches.
Inches are emerging, even more recently than the 1.8-inch form factor. However, in such a small magnetic disk device, the recording density per unit area becomes very large, the S / N of the signal processing system becomes low, and the presence of magnetic defects, that is, defects on the magnetic recording medium can be ignored. Disappear. Therefore, error correction processing and data format have become important in order to increase the reliability of the data read / written even if there is a defect on the magnetic recording medium.

The operation including the data format and error correction processing of the conventional magnetic disk device will be described below by taking the data reproducing operation on the recording medium as an example. FIG. 5 is a block diagram showing a signal processing system of a conventional magnetic disk device. In FIG. 5, 1 is the whole magnetic disk device, 2 is a magnetic recording medium on which data is recorded, 3 is a data head for writing and reading data on the magnetic recording medium 2, and 4 is a data head. An amplifier 5 for amplifying the signal of 3 is a pulse discriminator for converting the output of the amplifier 4 into a value, and 6 is a signal for the data head 3 which takes in the signal at the time of reading and outputs the write signal at the time of writing into a predetermined data format. Therefore, HDC for controlling, 7 is a central processing means for controlling the HDC 6, 8 is data stored in the HDC 6 at the time of reading from the magnetic recording medium 2, and data to be written through the HDC 6 at the time of writing to the magnetic recording medium 2. It is a buffer that stores.

Reference numeral 13 is an ECC circuit for detecting the occurrence of a read error in the read data and correcting the error. Usually, a 32-bit or 56-bit computer-generated code is often used. An interface 11 transmits and receives data to be read and written between an external host device 12 and a magnetic disk device 1 described later via a buffer 8. The external host device 12 is an information processing device main body that operates the magnetic disk device 1 as an external storage device, and is not included in the configuration of the magnetic disk device 1.

Here, the format format will be described. In a magnetic disk device, 25 data records
Recording / playback is performed in fixed small block units of 6 bytes, 512 bytes, and 1024 bytes, and the unit is called a sector. The format format determines the method in which an ID part as header information and ECC information for error correction are added to this sector, and a plurality of tracks are sequentially allocated to a number of tracks arranged concentrically on the magnetic recording medium. is there. Normally, several tens of sectors are arranged on one track.

FIG. 6 shows an example of a conventional format format. In FIG. 6, 14 and 15 are tracks T and T + 1 formed concentrically on the magnetic recording medium 2, and the n-th sector 21 on the track T is described here. This sector 21 has an ID section 16 and a data section.
Divided into 17.

Further, in FIG. 6, 101 is a PLO (Ph
9 bytes in case Locked Oscillator, 102 is SYNC (Synchronization)
1 byte for Byte, 103 for ID data (Iden
6 bytes in the title data field, 104 is a CRC (Cyclic Redundancy Cod)
e) 2 bytes, 105 PAD 2 bytes, 106 SP
2 bytes for LICE, 9 bytes for 117 PLO, 1 byte for SYNC, 119 for data field, 256
Or 512 or 1024 bytes, 120 is ECC (Erro
r Correction Code) is 4 or 7 bytes, 121 is 2 bytes for PAD, and 122 is 2 bytes for SPLICE. The number of bytes and the order of formats described here can be arbitrarily changed.

The purpose of each piece of information will be described below. Normally, the pulse discriminator 5 is provided with a data discrimination window of a predetermined width for a bit string recorded on the recording medium 2 at the time of reading data, and if there is a magnetization reversal in this discrimination window, it is 1; We judge that it is information. Therefore, in order to read information without error, a VCO (Voltage Control Os) is read within a predetermined time in the reproduction bit string of the discrimination window including the rotation fluctuation of the spindle motor.
It must be followed. For this purpose, PLO 101 and 117 are provided at the beginning of the ID information and DATA information during formatting so that the reproduced bit string of the discrimination window is followed within the time of PLO 101 and 117. SYNC102 and 11
Reference numeral 8 is used to distinguish the PLO 101 and 117 from the ID information and DATA information that follow. I
The D data 103 indicates where on the recording medium the sector to be recorded and reproduced is allocated. In this part, information for indicating the attribute of the sector may be recorded during recording and reproduction. CRC
104 is added with the detection information of the read error generated from the SYNC 102 and the ID data 103, and the HD is read at the time of reading.
This is for detecting a read error in the internal circuit of C6.

When the (2,7) modulation method is adopted as the data recording method, the PADs 105 and 121 cannot convert the data up to the last bit of the data depending on the data pattern. Therefore, 2 bytes are added for conversion. It is what you are doing. If it takes longer than usual due to fluctuations in the rotation of the motor during the writing operation to the ID section 16 and the previous sector, the SPLICE10
If there is no 6 or 122, the PLO 101 or 117 that is next arranged is recorded, and the PLO 101 or 1 is recorded.
The VCO cannot sufficiently follow the reproduction bit string of the discrimination window, which is the role of 17, so that the data cannot be reproduced correctly. Therefore, SPLICE 106 and 121 are arranged to absorb the time variation. The ECC 120 adds detection error correction information of the SYNC 118 and the data field 119 as redundant bits, and the ECC data generated by the ECC circuit 13 at the time of reading and the ECC 12 read from the recording medium.
This is for detecting / correcting a data reading error by comparing with 0.

The process of recording and reproducing on the recording medium according to the above data format is controlled by the HDC 9 (hard disk controller) according to the procedure programmed by the central processing means 7.

The data recording / reproducing process and the error processing operation of the conventional magnetic disk device will be described below with reference to the flowchart of FIG. First, the central processing means 7 is initialized to
A pointer indicating the number of read sectors and the buffer storage position of the read data is set, and the retry count is set in the retry counter (step 201). Next HD
The cylinder number, the head number, the sector number, and the flag information of the ID section, which is the header information of the read sector, are set in C6, and the HDC 6 starts the read operation (step
202).

First, the HDC 6 issues a read mode command to the circuit of the signal processing system, and takes in the signal recorded on the magnetic recording medium 2 from the data head 3 through the amplifier 4 and the pulse discriminator 5. Thus, the HDC 6 starts reading the data of the current track on the disc, and whether the ID information of the corresponding sector exists on the track,
The comparison of ID information is started. When it is detected that the disc has rotated once while the read ID information is different from the ID information of the target sector (step 203), or an error in the ID information is detected by the CRC provided for error detection. In step 205, the retry counter that is initially set is decremented by 1 (step 206) and retry is started (step 207). If the retry counter is 0
However, if the ID information sought is not found on the current track or the error detected by the CRC is not recovered, the process ends in error (step 208).

When the ID information of the target sector is found (step 204), the succeeding data field 11
The signal 9 is judged to be data, the pointer is incremented in the buffer 8 and fetched, and reading is started. The reading is performed for the total number of bytes of the one-sector data described above and the following ECC bytes (step 209). In this way, the sector data for one sector is stored in the buffer 8, and the reading for one sector is completed.

Further, after the reading of the sector data for one sector is completed, the central processing means 7 makes the E of the ECC circuit 13
Check the CC status and check whether an error has occurred in the read sector data (step 21
0). If an error occurs, the central processing means 7
Uses the ECC hardware of the ECC circuit 13 to detect the position of the error that has occurred and the error bit length (step 211). Since the error correction is possible if the error bit length is within a certain length, the central processing means 7 performs error correction processing on the sector data in the buffer 8 (step 212). If the error bit length is long and it is determined that the error cannot be corrected, the central processing means 7 terminates the reading of the sector in error (step 208).

Thus, when the reading of one sector is completed, the HDC 6 decrements the initially set number of read sectors and increments the sector information of the ID section,
The one-sector read process is restarted, and the read process is repeated until the number of read sectors becomes 0 (step 213).

At this time, the central processing means 7 simultaneously hand-shakes the read data and sends it through the interface 11 while monitoring the state of the external host device 12. If an unrecoverable read error occurs on the way, the central processing means 7 transmits the error status to the external host device 12 and stops the read process on the way.

[0018]

In the above conventional example, the ID is
It was designed considering that the frequency of occurrence of an error in reading ID information of a part and an error in a data part having a long error burst length that makes error correction by ECC extremely low occurs. Moreover, even if an error occurs, it can be dealt with sufficiently by re-reading it again by retrying. However, in recent magnetic disk devices, the capacity has been increased, and both the track density and the linear recording density have been significantly increased. For this reason, the presence of defects on the magnetic recording medium and the reduction of the S / N of the reproduced signal, such as the adoption of the Reed Solomon code, which is a stronger ECC code than the conventional computer-generated code, for the error correction of the data section. The design is changing on the premise that misreading will occur.

However, as is clear from the above-described data reading algorithm, even if only the data part can be error-corrected, if the ID part which is the header part information of the sector fails to be read, the data of the sector will not be read. Can not. Therefore, there is an idea that an ECC code capable of error correction is adopted instead of the CRC having only the error detection function in the ID portion.

However, in this case, since the processing time from error detection to error correction is required, when it is judged that the error correction of the information of the ID part is completed and the ID information of the target sector coincides, the data head is There was a problem that the reading of the next sector had already started and the data of the target sector could not be read in time.

If the ID part cannot be read even after retrying, the host device registers the sector as a defect and registers it as a defect. Then, although only the ID part of the sector cannot be read, the storage capacity of the magnetic disk device is temporarily reduced and the continuity of the sector is lost, so that the performance of the magnetic disk device is lost. I have no choice but to decline for a while.

The present invention solves the above-mentioned problems, and the reliability of the recorded / reproduced data is improved regardless of the increase in the error rate accompanying the decrease in S / N of the read signal due to the increase in the track pitch due to the miniaturization and the increase in the capacity and the increase in the recording density. It is an object of the present invention to provide a magnetic disk device having high performance and high performance.

[0023]

In order to solve the above problems, the magnetic disk device of the present invention uses a Reed-Solomon code, which is stronger than a conventional computer-generated code correction method, as an ECC error correction method for the data section. Is adopted, and a plurality of ID sections of the sector are overwritten, and an ID data majority decision circuit that adopts the one having the same ID section information as the correct one is provided.

[0024]

With the above configuration, the Reed-Solomon code, which is a more powerful error correction method, is adopted in response to an increase in errors in the data section. Therefore, errors occurring in the data section can be corrected by performing error correction. The error rate of can be drastically reduced. In addition, even if the information of the ID section, which is the most important as the header information of the sector, is overwritten, and if a data read error due to a defect of the magnetic recording medium or a decrease in S / N occurs in a part of the ID section information, Since the more correct data is adopted by the ID data majority decision circuit, the data in the sector will not be unreadable. In addition, since the ID data majority decision circuit can be configured by a simple combinational logic circuit, the circuit is simpler and the time delay required for error detection and correction is eliminated compared to the case where the ECC is configured by hardware. , A very big effect can be obtained.

[0025]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing a signal processing system of a magnetic disk device according to an embodiment of the present invention. In FIG. 1, reference numeral 1 denotes the entire magnetic disk device, 2 a magnetic recording medium on which data is recorded, 3 a data head for writing and reading data on the magnetic recording medium 2,
4 is an amplifier for amplifying the signal of the data head 3, 5 is a pulse discriminator for binarizing the output of the amplifier 4, 6 is a signal for the data head 3, the signal is taken in at the time of reading, and the output of the write signal at the time of writing is predetermined. HDC for controlling according to the specified data format,
Reference numeral 7 denotes a central processing means for controlling the HDC 6, reference numeral 8 stores the data taken into the HDC 6 when reading from the magnetic recording medium 2, and HDC when writing to the magnetic recording medium 2.
It is a buffer that stores the data to be written through 6, and is the same as the conventional one up to this point.

Next, 9 is an ECC circuit for detecting the occurrence of a read error in the read data and correcting the error, which uses a Reed-Solomon code having a bit length of 88 bits or more. Reference numeral 10 is an ID data majority decision circuit, which will be described later in detail. An interface 11 sends and receives data to be read and written between the external host device 12 and the magnetic disk device 1 via the buffer 8. The external host device 12 is an information processing device main body that operates the magnetic disk device 1 as an external storage device, and is not included in the configuration of the magnetic disk device 1.

Here, FIG. 2 shows an example of a format format according to an embodiment of the present invention. In FIG. 2, 14 and 15 are tracks T and T + 1 formed concentrically on the magnetic recording medium 2, and the n + 1th sector 21 on the track T is described here. This sector 21 has an ID
It is divided into a part 16 and a data part 17.

In FIG. 2, 101 is a PLO (Ph
9 bytes in case Locked Oscillator, 102 is SYNC (Synchronization)
Byte) is 1 byte, 103 is the first ID data (I
6 bytes for dentity DataField), 10
4 is CRC (Cyclic Redundancy Co
2) for de) and 2 bytes for 105 for PAD, and up to this point constitute the first ID section 18. Also, 106 is PL
O (Phase Locked Oscillator)
9 bytes, 107 is SYNC (Synchro
1 byte for the first byte, 108 for the second ID data (Identity Data Field) 6 bytes, and 109 for the CRC (Cyclic Redundancy).
y Code) is 2 bytes and 110 is PAD, which is 2 bytes, and constitutes the second ID section 19 up to this point. Also,
111 is PLO (Phase Locked Osill)
9 bytes in 112 bytes, 112 is SYNC (Synchr)
1 byte in the animation byte, 113 is the third
ID data (Identity Data Field)
d) 6 bytes, 114 is CRC (Cyclic Red)
2 bytes in the "Undancy Code", 115 is the PAD
2 bytes, and 116 is 2 bytes in SPLICE, and up to this point constitutes the third ID section 20. The ID section 16 is configured by combining the first ID section 18, the second ID section 19, and the third ID section.

Next, 117 is a PLO 9 bytes, and 118 is an S
YNC is 1 byte, 119 is a data field, 256 or 512 or 1024 bytes, and 120 is an 88-bit Reed-Solomon ECC (Error Correcti).
on code) is 11 bytes, 121 is 2 bytes for PAD, and 122 is 2 bytes for SPLICE. Up to this point, the data section 17 is configured. The number of bytes and the order of formats described here can be arbitrarily changed. The meaning of each piece of information is as described in the conventional example.

The process of recording and reproducing on the recording medium according to the above data format is controlled by the HDC 6 (hard disk controller) according to the procedure programmed by the central processing means 7.

FIG. 3 is a block diagram showing an example of the ID data majority decision circuit in the magnetic disk device according to the embodiment of the present invention. In FIG. 3, 301 is a first register that stores the first ID data, 302 is a second register that stores the second ID data, and 303 is a third register that stores the third ID data. 3 registers, these are H
It is built in DC6. Here, the register is shown only for 8 bits, but in reality, it is prepared according to the ID data length. 304 is an ID data majority decision circuit, and the inside thereof is composed of AND elements indicated by 305, 307, 309 and 306,
It is composed of OR elements indicated by 308 and 310. Although only three bits are shown here, the remaining bits have the same structure and are omitted.

The ID data majority decision circuit 304 will be described below. The majority decision circuit is a logic that adopts the data having the larger number of coincidences when the outputs of the plurality of logics are not coincident with each other. Here, the first register
301, the second register 302 and the third register 303 output the bit at the same position, which is 0 or 1, whichever is greater.
For example, if the least significant bits of the registers 301, 302, 303 are 0, 0, 1 or 0, 1, 0, etc., 0 is output, and if 1, 1, 0 or 0, 1, 1, etc. 1 is output. Of course, if all match, the matched value is output.

Here, assuming that the data of the register 301 is A, the data of the register 302 is B, the data of the register 303 is C, and the output is Q, in order to satisfy the required majority logic, the equation (1) is used. It is necessary to combine the logic that satisfies the logical expression of.

[0034]

[Equation 1]

By logically compressing this logical expression, the expression (2) can be obtained. Q = A · B + B · C + C · A (2) The logic equivalent to this logical expression is as shown in FIG.
AND elements of 305, 307, 309 and 306, 308, 310
It can be realized by the OR element shown in FIG. 3, and the ID data majority decision circuit 304 is configured by aligning this logic for desired bits.

Data recording / reproducing processing and error processing operation thereof by the magnetic disk device according to one embodiment of the present invention will be described below with reference to the flowchart of FIG. 4 by taking data reading as an example. First, the central processing means 7
As a default setting, set a pointer that indicates the number of read sectors and the buffer storage position of the read data,
The number of retries is set in the retry counter (step 201). Next, the cylinder number, head number, sector number, and flag information of the ID section, which is the header information of the read sector, are set in the HDC 6, and the HDC 6 starts the read operation (step 202).

First, the HDC 6 issues a read mode command to the circuit of the signal processing system, and takes in the signal recorded on the magnetic recording medium 2 from the data head 3 through the amplifier 4 and the pulse discriminator 5. Thus, the HDC 6 starts reading the data of the current track on the disc, and whether the ID information of the corresponding sector exists on the track,
The comparison of ID information is started. It may be detected that the disk has made one revolution while the read ID information is different from the ID information of the target sector (step 203), and the CRC provided for error detection may cause an error of two or more ID information. If it is detected (step 214), the retry counter that is initially set is decremented by 1 (step 206), and then retry is started (step 207). Even if the retry counter becomes 0, if the ID information sought is not found on the current track or the error detected by the CRC is not recovered, the process ends with an error (step
208). Up to this point, it is almost the same as the conventional example, but since this embodiment has three pieces of ID data of the ID section, up to one CRC error in the ID section can be corrected by the majority circuit described above. Therefore, it is acceptable (step 214).

If the ID information of the target sector is found (step 204), the following data field 119 is set.
It is determined that the signal is data, and the pointer is incremented in the buffer 8 and fetched, and reading is started. The reading is performed for the total number of bytes of the one-sector data described above and the following ECC bytes (step 209). In this way, the sector data for one sector is stored in the buffer 8, and the reading for one sector is completed.

Further, after the reading of the sector data for one sector is completed, the central processing means 7 checks the ECC status of the ECC circuit 13 by the Reed-Solomon code to check whether an error has occurred in the read sector data. Inspection to. (Step 215). If an error occurs, the central processing means 7 controls the ECC of the ECC circuit 13
Using hardware, the position of the error that has occurred and the error bit length are detected (step 211). Since the error can be corrected if the error bit length is within a certain length,
The central processing means 7 performs error correction processing on the sector data in the buffer 8 (step 212). If the error bit length is long and it is determined that the error cannot be corrected, the central processing means 7 terminates the reading of the sector in error (step 208).

Thus, when the reading of one sector is completed, the HDC 6 decrements the initially set number of read sectors and increments the sector information of the ID section,
The one-sector read process is restarted, and the read process is repeated until the number of read sectors becomes 0 (step 213).

At this time, the central processing means 7 simultaneously handshakes the read data through the interface 11 while sequentially monitoring the state of the external host device 12 and sends out the data. If an unrecoverable read error occurs on the way, the central processing means 7 transmits the error status to the external host device 12 and stops the read process on the way.

The data reading of the data field is almost the same as the conventional example, but more powerful ECC error detection and correction can be performed by adopting the Reed-Solomon code of 88 bits or more for detecting the ECC error. it can.

In the present embodiment, the multiplexing of the ID part has been described, but this may be performed for other parts of the sector data such as the ECC of the data part. Further, in the present embodiment, three multiplexing is performed, but it goes without saying that another number may be used.

As described above, according to the present embodiment, the ID portion which is important as the header information of the sector data is multiplexed, the majority circuit for the read ID data is provided, and the long bit read is performed in the error detection / correction method of the data portion. By using the Solomon code, the reliability of the sector data can be greatly improved.

[0045]

As described above, according to the present invention, in the reading and writing of data in the magnetic disk device, the ID portion which is important as the header information of the sector is multiplexed and the majority circuit for the read ID data is provided. Defects on the magnetic recording medium that have become non-negligible due to the increase in the linear recording density and the improvement in the track density accompanying the miniaturization of the disk device and the increase in the recording density, and the reading error of the ID information due to the decrease of the S / N occur. If you do, the sector will no longer become unreadable.

Further, regarding the read error of the data part due to the same cause as described above, since the Reed-Solomon code of 88 bits or more is adopted to detect the ECC error and the more powerful ECC error detection and correction are performed, the error correction is performed. The error can be drastically reduced and the error can be corrected even if the error has a long error burst length.

Therefore, when the information in the ID section cannot be read conventionally, the deterioration of the performance and the temporary decrease of the storage capacity due to the defect registration of the sector and the skip processing of the sector in the subsequent processing are eliminated. Its practical effect is enormous.

[Brief description of drawings]

FIG. 1 is a block diagram of a signal processing system of a magnetic disk device according to an embodiment of the present invention.

FIG. 2 is a diagram illustrating a data format system of a magnetic disk device according to an embodiment of the present invention.

FIG. 3 is a block diagram of an ID data majority decision circuit in a magnetic disk device according to an embodiment of the present invention.

FIG. 4 is a flowchart of a data reading operation algorithm by the magnetic disk device according to the embodiment of the present invention.

FIG. 5 is a block diagram of a signal processing system of a conventional magnetic disk device.

FIG. 6 is a diagram illustrating a data format method of a conventional magnetic disk device.

FIG. 7 is a diagram illustrating a data read operation algorithm by a conventional magnetic disk device.

[Explanation of symbols]

1 magnetic disk device 2 magnetic recording medium 3 data head 6 hard disk controller (H
DC) 7 Central processing means (CPU) 8 Buffer 9 ECC by Reed-Solomon code
Circuit 10 ID data majority circuit 12 External host device 14 Track T 15 Track T + 1 16 ID section 17 Data section 18-20 First to third ID section 21 Sectors 101,106,111,117 PLO 102,107,112,118 SYNC 103,108,113 ID data 104,109,114 CRC 105,110,115,121 PAD 116,119 SPLIC Field 120 Reed-Solomon code ECC 301 to 303 First to third ID data register 304 ID data majority circuit

Claims (6)

[Claims] 1. A rotatable magnetic medium on which a plurality of tracks for reading and writing data information are concentrically formed, a data head for reading and writing information on the tracks, and a finite data record defined on the tracks. A data format having a plurality of sectors having a length, an ID part in which a plurality of ID information is written as header information of the sector, a data field for reading and writing data in the sector, and an error correction code part following the data are provided. A hard disk controller to control, an error correction circuit,
A magnetic disk device comprising an ID data majority decision circuit for comparing a plurality of ID information. 2. The magnetic disk drive according to claim 1, wherein the same ID information is repeatedly written at the beginning of a sector as a data format of an ID section for writing a plurality of ID information. 3. An ID data majority decision circuit, a plurality of registers for storing a plurality of read ID data, AND
2. The magnetic disk drive according to claim 1, wherein a combination logic in which an element and an OR element are combined is used. 4. The magnetic disk drive according to claim 1, wherein whether or not to employ the output of the ID data majority decision circuit is determined based on the number of errors detected by the CRC of the ID section. 5. The magnetic disk drive according to claim 1, wherein a Reed-Solomon code is used as the error correction code in the data format and the code used in the error correction circuit. 6. The data format of an ID section for writing a plurality of ID information, wherein a PLO section, a SYNC section, an ID information, a CRC section, and a PAD section are repeatedly written in this order, and a SPLICE section is provided at the end. 1. The magnetic disk device according to 1.