JP2003317206A - Magnetic disk unit - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a read operation technology for achieving the high efficiency of a retry method in a read operation in a magnetic disk unit. <P>SOLUTION: This device is provided with a normal read means for validating a read gate for all necessary sectors, and batch-fetching sector information to decode it, and a retry means for executing a read operation again when an error is detected in the decoded signals. The retry means is provided with a plurality of kinds of retry modes. Thus, the read operation is carried out by a minimum number of read times, and a minimum number of retry times. <P>COPYRIGHT: (C)2004,JPO

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 and, more particularly, to error retry when reading data.

[0002]

2. Description of the Related Art Conventionally, when a read error occurs at the time of reading information stored in a magnetic disk medium, that is, when reproducing data, a controller (HDC, Hard Disk Co) that controls information processing inside the magnetic disk device.
After a part of the operation of the controller is stopped and the magnetic disk medium makes one revolution and returns to the original state, after waiting for so-called rotation,
It tried to reproduce the data again from the sector where the error occurred. Such an operation is called a retry, but if this retry is performed, the rotation waits for one rotation more than if it is not performed, so that the performance seems to be deteriorated accordingly.

In the above example, there is one error, but when an error occurs at a plurality of places or there are a plurality of error factors, 1) the read channel or preamplifier located in the information reproducing system, respectively. Change settings,
2) It is necessary to offset and read the magnetic head in the radial direction of the magnetic disk medium for the read target sector in which the magnetic head is positioned. For this reason,
When reproducing and transferring information at high speed, the set time and rotation waiting are major obstacles.

JP-A-11-7706 and JP-A-5-325
As shown in 434, when a read error occurs during the reproduction of information from a plurality of sectors, the subsequent data is prefetched and then a read retry is performed. It cannot handle error modes that require multiple retries.

As described in Japanese Patent Laid-Open No. 2001-236744, a method of reading another data of a different track and then returning to an error sector for read retry in order to eliminate the waiting for rotation, If there are different second error sectors on the track, or if there are sectors that require multiple read retries, there are inconveniences such as increased seek operations and complex algorithms.

As described in Japanese Patent Application Laid-Open No. 11-327797, in the method of storing the data of the read-retry sector in a cache memory or the like and taking out the data when necessary, the cache memory is finite and all the errors occur. It is impossible to save sector information.

FIG. 2 shows the state of retry when an error occurs during the read operation in the conventional magnetic disk device. The magnetic data recorded on the concentric data tracks 10 are schematically shown by straight lines. Truck 10
The smallest accessible unit is the data sector 11
Say. In a normal data read operation, the read gate 12 is validated from the read start sector, the normal read operation is interrupted in the sector in which an error has occurred, and then the read gate is revalidated from the sector in which the error occurred (retry), and data read To finish. Figure 2 shows 2 error sectors
An example that exists in several places is shown. Data track 1 shown in FIG.
In order to read the data of 0, except for the waiting time for rotation, seek time, and other processing times, three rotations are required even for the number of turns for reading.

[0008]

The cause of the read error is, for example, a defect in the magnetic film of the medium, a thermal aspirity generated when the magnetic head collides with a protrusion on the surface of the medium 2, and the temperature rises. There are off-track lights that are recorded off the center of the track. Particularly, in order to increase the recording density in the circumferential direction of medium rotation and in the track width (medium radius) direction, the S / N ratio of the signal reproduced by the magnetic head is increased.
The ratio and the magnetic characteristics of information recording / reproduction are deteriorated. As a result, read errors due to off-track writing also tend to increase.

At present, since a read error is dealt with by a retry operation, when there are a plurality of error sectors that cause a read error, the cause causing a reproduction error of each error sector is different, and each error sector is different. When it is necessary to perform a plurality of types of retry operations corresponding to the above, there is much waste in the retry operations.

Therefore, in the embodiment of the present invention, a new device is incorporated into a normal read operation in which a normal sector is read as a read target, so that the read operation is less wasteful and the performance is deteriorated due to the error occurrence during the read. Provide read control and retry operation to minimize.

[0011]

[Means for Solving the Problems] The normal read means for enabling the read gate for all necessary sectors and collectively fetching and decoding the sector information, and the error in the decoded signal. Upon detection, a retry means for performing the read operation again is provided. The retry unit has a plurality of types of retry modes. In other words,
1) A read operation for validating a read gate is performed for a plurality of peripheral sectors including a read target sector regardless of whether or not an error has occurred (hereinafter, referred to as a batch normal read). 2) When there is an error in the data due to the batch normal read, the retry operation is performed by repeating the batch normal read operation.

As a result, the read operation can be achieved with the minimum number of reads and the minimum number of retries.

[0013]

1 shows the configuration of a magnetic disk device used in the present invention. The magnetic head 1 records information on the surface of the magnetic medium of the magnetic disk medium 2 facing the magnetic head 1.
Information is reproduced from the surface of the magnetic medium. Magnetic disk medium 2
Is driven to rotate by a spindle motor 3 controlled by a spindle driver 4. The reproduction output from the magnetic head 1 is N by the preamplifier 5 and the read channel 6.
It becomes RZ data and is sent to the hard disk controller 8. These spindle driver 4, preamplifier 5,
A drive CPU 7 is present to control the read channel 6. Further, the connection between the magnetic disk device and the host device is performed via the interface control unit 9. Further, the magnetic head 1 is positioned at an arbitrary position on the magnetic disk medium 2 by an actuator (not shown).
An actuator driver circuit for supplying drive power to the actuator, an electronic circuit for controlling the actuator driver circuit, and an electronic circuit for controlling the magnetic disk device are also built in the magnetic disk device.

Data that is information is the magnetic disk medium 2
It is recorded by electromagnetic conversion in a concentric circle shape or a spiral shape on the surface of. The magnetic head 1 reproduces these data as an electric signal by moving relative to the magnetic disk medium 2, but in this process, a plurality of types of read errors may occur.

In the read operation implemented by the present invention shown in FIG. 3, the read gate is made effective for all data sectors to be read. Once the read gate is enabled, the read channel demodulates the data for that sector and transfers the data to the hard disk controller. The hard disk controller distinguishes a sector that can be normally read from a sector that has an error from the data. Next, a retry operation is performed to make the read gate valid only in the error sector that caused the error.

At this time, the read gate is made effective not only in the error sector but also in the same sector as in the first round, as in the normal read operation in the first round, and the information is fetched.
Only the necessary data (error sector) may be selected.

By such a read operation and a retry operation, the number of laps required only for the read is two laps except the processing time such as waiting for rotation and other processing.

The read retry may be a simple repetition, but various settings may be changed to perform a plurality of types of retry operations. That is, the error retry for the error that is expected to occur frequently is preferentially executed. When there are a plurality of error factors, a read retry corresponding to each of them is necessary, and the above-described conventional retry operation causes a large performance deterioration. The most prominent example is shown in FIG.

In high-density recording, when the track spacing is narrow and the following accuracy is poor, recording is deviated from the track center (hereinafter referred to as off-track write). To read the data, it is effective to retry by shifting the read position. FIG. 4 shows a conventional magnetic disk device (a) in which data recorded by meandering is recorded.
And a magnetic disk device (b) of the present invention is read.

The plus sign + in the data track diagram is
Sectors that are written along the radius of the magnetic disk medium in the outer peripheral direction (inner peripheral direction in a magnetic disk device that is recorded from the outer peripheral direction of the recording medium) are written, and the minus sign − indicates the inner peripheral direction (recording direction). In the magnetic disk device that is recorded from the outer peripheral direction of the medium, the sector is written with being shifted in the outer peripheral direction.

FIG. 4A illustrates an error in which the read gate needs to be re-enabled each time a data sector written with a shift is read. The number of laps required for the read retry is 7 laps. However, in reality, 1) the time required to set the positioning of the magnetic head so as to match the shifted data, 2) wait for the rotation, and 3) it is uselessly repeated until the type of retry suitable for the error sector is reached. Since invalid retries etc. will be received, it is usually expected that the number of laps will be more than double.

On the other hand, in FIG. 4B, the read gate is asserted for all the data sectors that need to be read (collective normal read). Since this is also applied to the retry operation, it is not necessary to perform the same retry many times, and the necessary number of read retry cycles is two.
Further, as the retry operation, not only the data sector in which the error has occurred but also the read gate is asserted for all the read target sectors, and only the necessary data may be selected.

When the reliability of data recording / reproduction is not so demanded, when it is important to transfer the data to the host device without interruption, and when the data transfer speed is important, the read operation shown in FIG. In operation, all the required sectors can be read without reasserting the read gate, resulting in high efficiency.

FIG. 5 shows an outline of dummy data transfer. This is because if the hard disk controller or interface controller transfers data to the host device (host) without interruption when the reproduced data is erroneous data, the magnetic disk device has a function of sending dummy data. Is the explanation.

The data sector and the error sector mean that the read channel cannot correctly demodulate the data () or cannot demodulate itself due to defects in the magnetic film or other defects ( )It is shown that. With respect to such a data string, it is possible to transfer the data to the host device without delay by adding dummy data in the hard disk controller or the interface control unit.

[0026]

The number of read retries can be reduced, and the read operation can be efficiently performed. Further, there is an effect that the data transfer rate can be easily transmitted to the host device without interruption.

[Brief description of drawings]

FIG. 1 is a configuration diagram showing an outline of an entire magnetic disk device.

FIG. 2 is a diagram for explaining a normal read operation in a conventional magnetic disk device.

FIG. 3 is a diagram for explaining a normal read operation in the magnetic disk device to which the present invention is applied.

FIG. 4 is a diagram for explaining a difference between a conventional read operation for off-track written data and a read operation of the present invention.

FIG. 5 is a diagram for explaining a dummy data transfer mode which is an example of a read operation without performing read retry.

[Explanation of symbols]

1 ... Magnetic head, 2 ... Magnetic disk medium, 3 ... Spindle motor, 4 ... Spindle driver, 5 ... Preamplifier, 6 ... Read channel, 7 ... Drive CPU, 8 ... Hard disk controller, 9 ... ... interface controller, 10 ...
… Data track, 11… Data sector, 12…
… Read gate.

─────────────────────────────────────────────────── ─── Continued Front Page (51) Int.Cl. ⁷ Identification Code FI Theme Coat (Reference) G11B 20/18 552 G11B 20/18 552F 572 572B 572F

Claims (2)

[Claims] 1. A magnetic disk medium having a plurality of sectors for storing information, a magnetic head provided so as to face the magnetic disk medium, a preamplifier for amplifying a signal from the magnetic head, and a read channel for processing the signal. A hard disk controller and other electronic circuits; an actuator for positioning the magnetic head on the magnetic disk medium; an electronic circuit for controlling the actuator; and a plurality of target sectors when reproducing information from the magnetic disk medium. A magnetic disk device having a function of collectively reading the same and a function of reading a sector other than the sector in which an error occurs when an error occurs when reproducing information from the magnetic disk medium. 2. A magnetic disk medium having a plurality of sectors for storing information, a magnetic head provided so as to face the magnetic disk medium, a preamplifier for amplifying a signal from the magnetic head, and a read channel for processing the signal. A hard disk controller and other electronic circuits, an actuator for positioning the magnetic head on the magnetic disk medium, an electronic circuit for controlling the actuator, and a plurality of target sectors when reproducing information from the magnetic disk medium. And a function of transmitting data prepared in advance in place of the information of the sector in which an error occurs when an error occurs when reproducing the information from the magnetic disk medium. A magnetic disk device having.