JPH02208874A - System for retrying readout error - Google Patents

Abstract

PURPOSE:To perform retrying more surely than ever by performing readout by setting a voltage higher and lower by a constant value decided in advance than a slice level voltage at the time of performing the readout when a readout error occurs. CONSTITUTION:A host controller 2 changes the generating voltage of a variable voltage generation circuit 12 by issuing a command to change the voltage to a voltage control means 13 at the time of performing the retrying for the readout error, and performs the readout of a memory device 1 by changing the slice level of a peak detection circuit 11. In other words, it is possible to perform the readout without performing extra readout by setting the slice level higher by the constant value, and also, to relieve the error due to missing by setting the slice level lower by the constant value. In such a way, the retrying can be performed for not only the error due to the defect of a medium, but also, a noise or off-track.

Description

[Detailed Description of the Invention] [Summary] Regarding retry when a read error occurs in a storage device, an object of the present invention is to provide a retry method that is effective even for read errors caused by minute defects in the medium. A variable voltage generation circuit that supplies a DC voltage that can be changed as a slice level voltage to a peak detection circuit that detects the peak of a read waveform and converts it into a data pulse, and a voltage control that controls to change the voltage generated by the variable voltage generation circuit. means, when a read error occurs in the data reproduced from the recording medium, the readout is performed by setting the slice level higher by a predetermined value and lower by a predetermined value than the slice level during normal readout. Configure it to do so.

[Industrial application field]

The present invention relates to reading data in a magnetic storage device,
In particular, it relates to a retry method when a read error occurs.

[Prior Art] In a device for recording and reproducing data such as a magnetic storage device, it is impossible to completely eliminate read errors when reproducing data.

Possible causes of read errors include noise, media defects, and interference with adjacent data on the media.

Normally, a host device (magnetic storage control device, etc.) checks the error detection and correction code contained in the recorded data.
If a read error is detected, retries are performed until the read is successfully performed.

Conventional retry methods include: 1. Repeating data without doing anything; 2. Reading by changing the data strobe position; 2. Reading by offsetting the magnetic head.

To read data in such a storage device, as shown in FIG. 5, a read waveform obtained by amplifying the voltage read by the read/write head is input into a peak detection circuit, the peak of the waveform is detected, and the peak position is determined. Obtain data pulse ■ with rising edge. This data pulse (2) is input into a demodulation circuit, and is strobed with a window of a constant width using a data strobe pulse created from a synchronizing signal to determine data "1191."0" and convert it into binary data (2).

Method (2) performs rereading without making any changes to the readout circuit. Method (2) is to read data by changing the data strobe position back and forth. In method (2), reading is performed by offsetting the position of the magnetic head to the left and right at right angles to the track direction.

In reality, a combination of these is often used. That is, for example, if a read error is detected, first read several times without doing anything, and if it still cannot be read normally, change the position of the data strobe forward and backward and read several times each, and if it is still normal. If reading is not possible, the magnetic head position is offset to the left and right and reading is performed several times each.

[Problem to be solved by the invention]

In the above-mentioned conventional three-glue try method, method (2) is a problem of the probability of read error occurrence. For example, if there is a minute medium defect and there is a probability that an error will occur 3 times out of 10 reads, then If you retry several times, you will be able to read normally without doing anything, and this is the most basic and simple method.

Method (2) is effective against peak shift (a phenomenon in which the peak position of the readout waveform moves), but it also causes "data leakage" (a pseudo phenomenon called extra due to micro defects in the medium) as shown in Figure 6. There is no effect on "missing" (a phenomenon in which a peak appears) or "missing" (a phenomenon in which the height of a peak in a read waveform becomes low due to a defect in the medium, and it is determined that there is no data pulse without exceeding the slice level).

Method (2) is effective against misalignment caused by thermal off-track (a phenomenon in which a magnetic head located particularly far from the servo head deviates from the track due to temperature changes due to differences in the coefficient of thermal expansion of various parts), etc. There is a problem in that it has little or no effect on errors caused by minute defects.

The problem to be solved by the present invention is to provide a more reliable method of retrying in combination with the conventional retrying method which has such problems.

[Means to solve the problem]

FIG. 1 is a diagram showing the configuration and principle of the present invention.

In FIG. 1(a), 1 is a storage device such as a magnetic disk device, and 2 is a host control device that controls the storage device 1. In FIG.

A peak detection circuit 11 detects a peak in a read waveform from a magnetic head or the like and converts it into a data pulse.

Reference numeral 12 denotes a variable voltage generation circuit, which generates a DC voltage that can be changed by external control and supplies it as a slice level to the peak detection circuit 11.

Reference numeral 13 denotes voltage control means, which controls changes in the voltage generated by the variable voltage generation circuit 12.

In the present invention, when retrying for a read error, the host controller 2 instructs the voltage control means 13 to change the voltage, changes the voltage generated by the variable voltage generation circuit I2, and changes the slice level of the peak detection circuit 11. is changed to cause the storage device 1 to perform reading.

[Operation] The voltage read out from the recording surface of the medium by the read/write head is an analog waveform, and the peak position is detected by the peak detection circuit and converted into a data pulse.

However, in actual readout waveforms, due to the effects of medium defects, waveform interference, etc., extra data (extra) and missing data occur as shown in FIG. Therefore, in order to distinguish between extra peaks and regular peaks, a method is used in which a slice level is provided and only peaks exceeding this slice level are treated as regular pulses.

Therefore, when the slice level is raised, it becomes stronger against extras and becomes weaker against missing parts. Conversely, if the slice level is lowered, it becomes stronger against missing, but becomes weaker against extra. Typically, this slice level is set midway between the respective limit values.

However, depending on the size of the defect and the delicate positional relationship during writing, the extra may change this slice level or may not reach this slice level, resulting in missing data.

If the waveform in FIG. 6 is caused by off-track of the magnetic head, retrying by offset of the magnetic head will be effective, but if it is caused by a media defect, it will have little effect. Also, changing the strobe position has no effect at all.

Therefore, in the present invention, at the time of retry, a voltage higher by a predetermined constant value and a voltage lower by a predetermined constant value than at the time of normal reading is set as the slice level by human control from the voltage control means 13 to the variable voltage generation circuit 12. , to the peak detection circuit 11.

As shown in (■) in Figure 1(b), by slightly increasing the slice level, reading can be performed without extra data, and as shown in (■), by slightly lowering the slice level, errors due to missing can be avoided. It can be rescued.

In actual equipment, media defects are detected in advance.
Since the data is stored in an alternate recording location, the read error is due to a minute defect that was not detected. Therefore, errors often occur at delicate positions relative to the slice level.

As described above, the retry by changing the slice level according to the present invention is particularly effective against minute defects in the medium, and is also effective against noise and off-track.

If you use this retry method alone or in combination with the above-mentioned methods of ``Rereading without doing anything'', ``Reading by changing the data strobe position,'' and ``Reading by offsetting the magnetic head,'' all causes of errors can be eliminated. You can retry effectively.

(Examples) Hereinafter, the present invention will be explained in more detail with reference to Examples shown in FIGS. 2 to 4.

FIG. 2 is a diagram showing the main circuit configuration of an embodiment of the present invention.

In the figure, 1 is a magnetic disk device, and 2 is its upper control device.

A microprocessor (MPU) 13° controls the interface between the magnetic disk device 1 and the upper control bag W2, and controls the positioning of a magnetic head (not shown).

11 is a peak detection circuit.

12 is a variable voltage generation circuit, R1-R4 and SWI
, SW2 are resistors and switches that constitute a variable voltage generation circuit.

The voltage supplied to the peak detection circuit 11 as a slice level (A) is created by dividing the power supply voltage VCC by resistance.

The slice level voltage is the switch SWI and 5W2S
The signal ■ that controls the switch SWI and the switch S
MP
"Up", "Normal", "Down" from U13'
can be controlled. The MPU 13' controls the signal and
control.

That is, in the normal read state (normal), the signal (2) is set to 1, the signal (2) is set to "0", and the switch SW1 is turned on and the switch SW2 is turned off. Thereby, the slice level becomes a voltage determined by the voltage division ratio of the parallel resistance value of R2 and R3 and the resistance value of R1.

If you want to raise the slice level (up) at the time of retry, set both signals ■ and ■ to "0゛°" and use SWI.

Leave SW2 both off. As a result, the slice level becomes a voltage determined by the voltage division ratio of R2 and R1.

If you want to lower the slice level at the time of retry, set both signals ■ and ■ to "1" and use SWI.

Turn on both SW2. As a result, the slice level becomes a voltage determined by the parallel value of R2, R3, and R4 and the voltage division ratio of R1.

FIG. 3 is a flowchart showing a retry operation according to an embodiment of the present invention.

The operation will be explained below according to the processing steps of the flowchart.

(2) If the host control device 2 detects the occurrence of an error that cannot be corrected even by ECC (error detection/correction code), the process proceeds to step (2). If not detected, return to step ① and continue monitoring.

(2) The host controller 2 issues a slice-up command.

The MPU 13' of the magnetic disk device 1 receives this and changes the signal from "1" to "0".

(2) The host controller 2 issues a set sector command to the same sector where the error occurred, and issues a read command in response to a sector arrival interrupt from the magnetic disk device 1. Reading is performed in the magnetic disk device 1 by raising the slice level to "up".

(2) The host control device 2 determines whether an error that cannot be corrected by ECC has occurred or not. If an error occurs, proceed to step ■; if an error does not occur, proceed to step ■.

(2) The higher-level control device 2 issues a slice down command.

The MPU 13' of the magnetic disk device 1 receives this and changes both the signals ``1'' and ``1'' to "1".

(2) The host controller 2 issues a set sector command to the same sector where the error occurred, and issues a read command in response to a sector arrival interrupt from the magnetic disk device 1. Reading is performed in the magnetic disk device 1 with the slice level "down".

(2) The host controller 2 identifies whether an error that cannot be corrected has occurred using ECC. If an error occurs, proceed to step ■; if an error does not occur, proceed to step ■.

(2) The host control device 2 determines whether or not to try again. Check the flag for retrying/not retrying, which can be set manually or by the host device (main body), and if it is "1", return to step ■; if "OII", proceed to step [phase]. If a retry is in progress, this judgment is not made and the process advances to step [phase].

■The read data is processed normally and the host device (main unit)
Transfer to.

[Phase] The higher-level control device 2 reports the occurrence of an uncorrectable read error to the higher-level device (main unit).

■The host controller 2 issues a slice normal command,
Terminate retry processing. MPU of magnetic disk device 1
13", the signal is "1°", and the signal (2) is "O".

FIG. 4 is a flowchart showing a retry operation according to another embodiment of the present invention. This embodiment is a combination of glue tries with various state changes, and the details of each glue try are omitted in this flowchart.

The operation will be explained below according to the processing steps of the flowchart.

@The host controller 2 detected the occurrence of an uncorrectable error.

■The host controller 2 simply issues the set sector command and read command again without making any changes.
Let them lead.

(2) If the host control device 2 detects the occurrence of an uncorrectable error, it proceeds to step @, and if no error is detected, it proceeds to step (2).

(2) After issuing a command to offset (forward or backward) the data strobe, the host controller 2 issues a set sector command and a read command to read the data strobe. When reading is finished, return to the normal data strobe position.

[Phase] If the host controller 2 detects the occurrence of an uncorrectable error, it proceeds to step @, and if no error is detected, it proceeds to step 0.

■The host controller 2 offsets the track (inward, outward)
After issuing a command to read the data, issue a set sector command and a read command to read the data. When reading is finished, return the offset to zero.

(2) If the host control device 2 detects the occurrence of an uncorrectable error, it proceeds to step (2), and if no error is detected, it proceeds to step 0.

[Phase] After issuing a command to offset the slice level, the upper control device 2 issues a set sector command and a read command to perform a read. When reading is finished, return to the normal slice level. The detailed contents are the same as steps ① to ② and ② in Fig. 3.

If the host controller 2 detects the occurrence of an uncorrectable error, it proceeds to step @, and if no error is detected, it proceeds to step (2).

■The read data is processed normally and the host device (main unit)
Transfer to.

(2) The higher-level control device 2 reports to the higher-level device (main body) the occurrence of an uncorrectable error.

Note that the above-described successive processing by changing the slice level of the present invention can be used when detecting a medium defect.

Methods for detecting medium defects include a method of directly detecting extras and missing from analog waveforms, and a method of detecting medium defects by actually performing running (recording and reproducing operations of data in an actual device). By changing the slice level during this running, it is possible to reliably detect minute defects near the slice level.

〔Effect of the invention〕

As is clear from the above explanation, according to the present invention, since the slice level of the peak detection circuit that affects extra and missing is changed or tried, it is possible to prevent not only errors caused by media defects but also noise and off-track. is also effective.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing the configuration and principle of the present invention, FIG. 2 is a circuit diagram of a main part of an embodiment of the present invention, and FIG. 3 is a flowchart showing a retry operation according to an embodiment of the present invention. FIG. 4 is a flowchart showing retry winter production according to another embodiment of the present invention, FIG. 5 is a diagram showing reproduction of magnetic recording, and FIG. 6 is a diagram explaining extra and missing. 11 is a peak detection circuit, 13 is voltage control means, and R1-R4 are resistors. 12 is a variable voltage generation circuit, and 13' is an MPU. SWI and SW2 are switches; (b) (a) Flowchart showing retry operation according to one embodiment of the present invention; FIG.

Claims (1)

[Claims] In a storage device (1) that records and reproduces digital data and a host control device (2) that controls the storage device (1), a peak detection method that detects the peak of a read waveform and converts it into a data pulse. a variable voltage generation circuit (12) that supplies a DC voltage that can be changed as a slice level voltage to the circuit (11); and a voltage control means (13) that controls to change the voltage generated by the variable voltage generation circuit (12). , and when data reproduced from a recording medium causes a read error, reading is performed by setting the slice level higher by a predetermined constant value and lower by a predetermined constant value than the slice level during normal reading. A retry method for read errors characterized by: