JPH06275034A - Head positioning mechanism for magnetic disk device - Google Patents

Abstract

PURPOSE:To improve the positioning accuracy of a head by always controlling an off-track amount to an optimum amount by efficiently calculating an optimum shift amount. CONSTITUTION:A corrected value generating circuit 12 performs correction control so that a reproducing head at a target track position can be off-tracked from a track center oppositely just for the shift amount in the disk radius direction of a recording head to this reproducing head, and an error measuring circuit 17 measures the bit error rate of a regenerating signal in a data zone while changing the off-track amount. Tracking is performed with the off-track amount, for which this bit error rate is made best, as the optimum value of the shift amount. Thus, the positioning accuracy of the head is improved by always controlling the off-track amount to the optimum amount by efficiently binding the optimum shift amount.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a head positioning mechanism of a magnetic disk device, and more specifically, a rotary type support arm is provided with a reproducing head and a recording head separately, and the reproducing head picks up a servo signal of a disk. And a head positioning mechanism of a magnetic disk device for performing track servo.

[0002]

2. Description of the Related Art Conventionally, as a head displacing means of a disk drive device, a linear type in which a support arm is linearly moved to displace the head in the radial direction of the disk, and a rotary type in which the support arm is rotated to be displaced. Therefore, the rotary type is preferable from the viewpoint of downsizing and weight reduction of the device.

On the other hand, although a magnetoresistive head having excellent reproducing sensitivity characteristics has been developed even in a narrow track width in order to cope with the increase in track density, this magnetoresistive head has no recording ability, and therefore another recording head. It is necessary to install two heads together. Therefore, by adopting the rotation type support arm and separately mounting the reproducing head and the recording head on the supporting arm, it is possible to create a small-sized, lightweight, high recording density recording / reproducing disk drive device.

In the above disk drive device, the reproducing head and the recording head are arranged at a constant interval and are displaced on the circumference. Therefore, for example, as shown in FIG. In FIG. 1), the centers of the reproducing head Hp and the recording head Hr are arranged so as to be located on the track center line C2, but at other positions, the centers of both heads Hp, Hr are not located on the track center line C2. It will be.

When the reproducing head Hp executes the track servo to execute the recording mode, the recording head Hr scans the off-track position except the innermost track position. Therefore, the data signal recorded by the recording head Hr is recorded on the track center line. It is recorded at a position offtrack from C2. When the recording signal thus recorded is reproduced, the reproducing head H
Since the center of p is controlled so as to be located at C2 on the track center line, the recording head Hr scans the off-track position of the data signal except for the innermost track Ti, so that noise or a read error occurs. There are drawbacks.

Therefore, in the reproducing mode, the reproducing head scans the just tracks in the data zone on all tracks, and in the recording mode, the recording head is shifted to scan the just tracks in the data zone. In this way, it is necessary to provide a track servo mechanism of a disk drive device that prevents the occurrence of noise, reading errors, etc. due to the head being off-track during recording and reproduction, and as a means for achieving this, For example, there is a “track servo mechanism of a disk drive device” described in Japanese Patent Application No. 1-325215 filed earlier by the present applicant.

[0007]

By the way, the above-mentioned Japanese Patent Application No.
In the track servo mechanism described in 1-325215, it is necessary to calculate the amount of shift of the recording head, but this shift amount is different for each radius and the distance between the recording head and the reproducing head, the head support arm and the disc. It is an amount determined by the distance from the center of rotation.

The radius at which the head is located is uniquely determined by designating the target track number, but an accurate radius position is known in advance due to reasons such as disk eccentricity and track pitch not being completely uniform. Is difficult. Further, the positions of the recording head and the reproducing head change due to variations in the heads. Therefore, it is difficult to know the shift amount accurately in advance. As a result, there is a problem that an optimum shift amount is not required, it is difficult to control the off-track amount to an optimum amount, and the head positioning accuracy of the magnetic disk device deteriorates.

SUMMARY OF THE INVENTION Therefore, an object of the present invention is to provide a head positioning mechanism for a magnetic disk drive which can improve the positioning accuracy by efficiently finding the optimum shift amount and thereby controlling the off-track amount to the optimum amount. I am trying.

[0010]

In order to achieve the above object, a head positioning mechanism of a magnetic disk drive according to a first aspect of the present invention has a reproducing arm and a recording head at the tip of a supporting arm rotatably supported at the base end side. Are provided separately, and the reproducing head and the recording head are moved in the substantially radial direction of the disk by the rotation of the support arm, and the disk has a servo zone for recording a servo signal and a data zone for recording a data signal on the same disk. In a head positioning mechanism of a magnetic disk device, which is provided on a surface, picks up the servo signal by the reproducing head, and displaces the rotational position of the support arm based on the picked-up servo signal to perform track servo, a target track Opposite by the amount of shift of the recording head in the disk radial direction with respect to the reproducing head at the position And correcting means for correcting control in order to off-track from the track center of the read head in direction,
Measuring means for measuring the bit error rate of the reproduced signal in the data zone while changing the off-track amount, and tracking is performed with the off-track amount that gives the best bit error rate as the optimum value of the shift amount. Characterize.

Further, as a preferred mode, in the head positioning mechanism of the magnetic disk drive according to claim 1, means for measuring the optimum shift amount for several tracks and shift amounts at other radial positions are measured shift amounts. It is characterized by comprising means for obtaining the original by interpolation.

In a head positioning mechanism for a magnetic disk drive according to claim 1, a reproducing means for reproducing predetermined recording data recorded on a recording medium by using a partial response system, and a reproducing signal reproduced from the recording medium. An analog-digital conversion circuit for converting the signal level of the signal into a digital signal at a predetermined cycle, and a Viterbi decoding circuit for decoding the reproduction signal by a Viterbi decoding based on output data output from the analog-digital conversion circuit, A level detection decoding circuit for decoding the signal level of the reproduction signal by comparing with a predetermined reference value, decoding data by the Viterbi decoding circuit, and decoding data by the level detection decoding circuit are compared, And an error detecting means for detecting these disagreements.

According to a fourth aspect of the present invention, there is provided a head positioning mechanism for a magnetic disk device, wherein a reproducing head and a recording head are separately provided at the tip of a supporting arm whose base side is rotatably supported, and the reproducing operation is performed by rotating the supporting arm. The head and the recording head are moved in a substantially radial direction of the disc, and a servo zone for recording a servo signal and a data zone for recording a data signal are provided on the same disc surface on the disc, and the servo signal is transmitted to the reproducing head. In the head positioning mechanism of the magnetic disk device for performing track servo by displacing the rotation position of the support arm based on the picked up servo signal, the recording head with respect to the reproduction head at each rotation position of the support arm. Inside the servo track in the opposite direction by an approximately shift amount in the radial direction of the disk. A line is provided to be shifted with respect to the center line of the data track, and correction means for correcting and controlling the reproducing head to be off-track from the track center at the target track position, and reproduction in the data zone while changing the off-track amount. And a tracking means that performs tracking while performing correction control using an off-track amount that maximizes the bit error rate.

In the head positioning mechanism of the magnetic disk drive according to claim 4, means for measuring the optimum shift amount for several tracks and means for obtaining shift amounts at other radial positions by interpolation based on the measured shift amounts. It is characterized by including and.

In the head positioning mechanism of the magnetic disk drive according to claim 4, a reproducing means for reproducing predetermined recording data recorded on the recording medium by using the partial response method, and a reproducing signal reproduced from the recording medium. An analog-digital conversion circuit for converting the signal level of the signal into a digital signal at a predetermined cycle, and a Viterbi decoding circuit for decoding the reproduction signal by a Viterbi decoding based on output data output from the analog-digital conversion circuit, A level detection decoding circuit for decoding the signal level of the reproduction signal by comparing with a predetermined reference value, decoding data by the Viterbi decoding circuit, and decoding data by the level detection decoding circuit are compared, And an error detecting means for detecting these disagreements.

[0016]

According to the present invention, correction control is performed so that the read head is off-track from the track center in the opposite direction by the shift amount of the write head in the disk radial direction with respect to the read head at the target track position, and the data is changed while changing the off-track amount. The bit error rate of the reproduction signal in the zone is measured, and the off-track amount that gives the best bit error rate is used as the optimum shift amount for tracking.

Therefore, the optimum shift amount is efficiently obtained, whereby the off-track amount is always controlled to the optimum amount, and the positioning accuracy is improved.

[0018]

Embodiments of the present invention will be described below with reference to the drawings. 1 to 6 show a first embodiment of the present invention. FIG. 1 shows a schematic configuration diagram of a magnetic disk device. In FIG. 1, the base end of the support arm 1 is the rotary shaft 1.
It is rotatably supported by a. The rotating shaft 1a is rotated by an arm driving means (not shown), and a reproducing head Hp and a recording head Hr, which will be described below, move between the innermost track Ti and the outermost track To in the disk radial direction.

A slider body 2 is attached to the tip of the support arm 1, and a magnetoresistive reproducing head Hp and a winding recording head H are attached to the slider body 2 as shown in FIG.
r and are provided. Playback head Hp and recording head H
r is arranged on the center line C4 of the support arm 1 at regular intervals, and the recording head Hr is provided at the front position and the reproducing head Hp is provided at the rear position. And disk 3
Center line C of the support arm 1 at the innermost track Ti position of
4, that is, the center line C4 of the reproducing head Hp and the recording head Hr is arranged so as to coincide with the track center line C2.

FIG. 3 shows a format diagram of the disc 3. In FIG. 3, track areas having different radii are formed at constant intervals on the disk 3 around the center of rotation, and servo areas Es and data areas Ed are alternately formed in each track area. As shown in FIG. 18 described above, a servo track is formed in the servo area Es and a data track is formed in the data area Ed, and a time for switching the recording / reproducing amplifier is provided between the servo area and the data area. There is a necessary gap to make money.

Each of these areas may be formed by recording magnetization as servo information on a uniformly formed magnetic film by a recording head or the like. Alternatively, the servo track and the data track may be formed on the magnetic layer and other areas. The range may be configured as a non-magnetic layer. Further, the servo track and the data track may be in a convex state, and the other areas may be in a concave state so that recording / reproduction cannot be performed. The servo area Es has a track address portion TA and a clock mark C.
And fine address sections A, B, X and Y.

A track address signal is recorded in the track address section TA, and fine address sections A, B, X,
Fine information for tracking is recorded in Y. In the clock mark C, clock information for reproducing a clock for reading these signals by the PLL is recorded. Here, the fine address part is
The center line C1 is configured to always coincide with the center line C2 of the data track Td.

When data is reproduced from the data area Ed, tracking is performed by the signal picked up from the fine address section so that the reproducing head coincides with the track stop line C2. On the other hand, when recording data in the data area Ed, if the reproducing head is tracked by the signal picked up from the fine address section so as to coincide with the track stop line C2, the radial position other than the innermost circumference of the disk In, the center of the recording head shifts to a position different from the track center line C2.

This shift amount X is x = where the distance between the reproducing head Hp and the recording head Hr is 1, and the skew angle θ between the data track center line C2 and the support arm 1 center line C3.
It is obtained by lsin θ, and becomes a larger value as it goes to the outer track. Here, θ is a quantity that can be calculated depending on the radial position of the head, the length of the support arm, and the like.
The radial position of the head is a value that can be calculated from the track number if the track pitches are equal.

However, the track pitches are not exactly equal, or the distance l between the reproducing head and the recording head is large.
The actual shift amount may be different from the calculated value due to variations in the values, changes over time, and the like. Therefore, in this embodiment, in order to accurately know the shift amount at each radial position, data is recorded in the data area while changing the shift amount little by little, and the optimum bit error is measured by measuring the bit error at each shift amount. Seeking shift amount.

FIG. 4 shows a block diagram of a circuit for obtaining the optimum shift amount. In FIG. 4, the signal reproduced by the reproducing head Hp passes through the amplifier 10, and then the address extracting circuit 11 detects a fine address and a track address. Data track center line C2 calculated from track address and support arm 1 center line C3
From the skew angle θ of x, the rough shift amount x = lsin
Find θ.

On the other hand, in the correction value generating circuit 12, the shift amount is set by adding the value set by the shift amount setting circuit 13. The shift amount set here is supplied to the support arm drive motor 15 in addition to the fine address by the adder 14, so that the recording head is tracked to a target position and data is recorded on the data track.

Next, by turning off the correction value generating circuit 12, the reproducing head is just tracked to the data track, and the data detecting circuit 16 reproduces the data area. Then, the bit error of the obtained reproduction data is measured by the error measuring circuit 17, and the shift amount set at the time of recording the data and the value of the measured bit error rate are stored in the shift amount error rate storage memory 18. .

The above operation is repeated while changing the shift amount. Although the bit error rate changes as shown in the graph of FIG. 5 when the shift amount is changed, it should be the best when the recording head is accurately recorded on-track in the data area. , This shift amount is used as the optimum shift amount.

Here, the bit error can be measured by detecting the error by collating the recorded data with the reproduced data, as is usually done. The collation of data can be performed in real time using hardware, or can be performed by loading the data in a memory and using a microprocessor and software.

FIG. 6 shows an example of a block diagram of a circuit for detecting an error by collating data using hardware. In FIG. 6, a synchronizing signal is detected by the synchronizing circuit 21 from the reproduced signal, and a signal indicating the start point of the sector is extracted. The same signal as the signal recorded on the data track is stored in the comparison data generation circuit 22, and the comparison signal synchronized with the reproduction signal is output by the sector start signal detected by the synchronization circuit 21. Also,
The comparison circuit 23 compares the reproduction signal with the comparison signal,
When a mismatch occurs, it is considered that a bit error has occurred and a mismatch detection signal is generated.

By counting this signal by the counter 24, the bit error is measured. By performing the above operation for each radius of the disc, the optimal shift amount at each radial position can be known, and when actually recording data, the optimal shift amount is used for tracking. It should be noted that these operations may be performed on all tracks of the disc, or some points may be picked up and the shift amount may be calculated by interpolation or the like using the values of the measured points for the tracks in between.

Here, when the head is at the innermost circumference, θ =
Although an example of 0 has been shown, this does not necessarily have to be the case. This is because the center of rotation of the disc and the support arm 1
Changes depending on the distance from the center of rotation, the length of the support arm 1, and the like. If these values change, the position of θ = 0 also changes.

Operation of the First Embodiment During reproduction, the reproducing head Hp moves the servo track Ts.
When the head is scanned, the reproducing head Hp picks up the track address signal and the reproducing head fine address signal. As a result, the reproducing head Hp runs just on the servo track. The center lines of the fine address track and the data track Td are coincident with each other, and the reproducing head Hp runs on the data track Td while being in a jato track.

When the reproducing head Hp scans the servo track Ts during recording, the reproducing head Hp picks up the track address signal and the fine address signal. At this time, the reproducing head Hp is shifted by an optimum shift amount obtained in advance. Track. As a result, the recording head Hr travels on the data track Td by following the just track.

As a result, at the time of reproduction, the reproduction head Hp
However, at the time of recording, the recording head Hr travels with the data track Td just tracked. Since the shift amount at each radial position is determined so that the bit error rate in the data area Ed becomes the best, it is possible to perform tracking so as to always obtain the best bit error rate at any radial position. As a result, the off-track amount can always be controlled to an optimum amount, and the head positioning accuracy of the magnetic disk device can be improved.

Second Embodiment The measurement of the bit error rate described in the first embodiment is performed by detecting the error by collating the recorded data and the reproduced data bit by bit. In order to realize this means, a memory or the like for storing the recorded data is required, and in order to perform matching in real time by hardware, a circuit that outputs the data in synchronization with the reproduction signal. However, since it takes time to perform verification by software, it may not always be possible to perform efficient bit error measurement.

Therefore, in the second embodiment, it is not necessary to reduce the circuit scale and hold the record data by using the method combining the maximum likelihood decoding and the level detection as the bit error detecting method. The following is an example. The parts other than the detection of the bit error are the same as those in the first embodiment, and are not shown here.

First, the partial response of the modulation code in the magnetic recording apparatus which is the object of the present invention will be described. A partial response is used for the modulation code in the magnetic recording device, and as the types of the partial response, those commonly used include an arithmetic circuit 31 shown in FIG. 7A and an arithmetic circuit 32 shown in FIG. 7B. There is a system using 33. PRS (1,1), PRS (1, -1), PR
S (1,0, -1) is the condition judgment of the operation example. These system polynomials are G (D) = 1 + D and G (D) = 1, respectively.
-D ² and the arithmetic circuit 31 is an independent arithmetic circuit 32, 3
3 is considered to be so-called two-nested. D is a delay operator.

That is, the arithmetic circuit 31 shown in FIG.
In (Partial response is PRS (1,0, -1)), the calculation is performed between the input data and the sample two before, so
There is no relationship between the odd-numbered sample and the even-numbered sample, and each is an independent partial response.
It can be regarded as a sequence of PRS (1, -1).

In the arithmetic circuits 32 and 33 shown in FIG. 7B, two series of odd-numbered samples and even-numbered samples with respect to the input data are switched by switches 34 and 35 to be divided into two. The calculation is being performed. That is, the arithmetic circuits 32 and 33 (the partial response is PRS
(1, -1)) and arithmetic circuit 31 (Partial response is PRS
The decoding of (1,0, -1)) is essentially the same, and here, the partial response PRS (1,0, -1) will be described as an example.

The partial response PRS (1,0, -1) itself has a property of propagating an error, and if a 1-bit error occurs under a certain condition, a catastrophic error may occur. Therefore, precoding before recording is performed. You need to do it. For this, it suffices to multiply by the one that performs the inverse conversion of the partial response. In this case, the configuration of the entire device is
It is shown as in FIG.

In FIG. 8, reference numeral 41 is a precoder for executing the processing of (1 / 1-D ² ), and the recording data is subjected to the arithmetic processing of (1 / 1-D ² ) by the precoder 41. It is converted to precode data that changes between the values 1 and -1 of the recording data by utilizing the correlation between the recording data, and is output to the recording channel circuit 42. In the recording channel circuit 42, the arithmetic processing circuit 43 performs the arithmetic processing of (1-D) on the output of the precoder 41, and the arithmetic result is added to the adder 44.
The noise is added in and output to the arithmetic processing circuit 45 in the subsequent stage. In the arithmetic processing circuit 45, (1 + D) arithmetic processing is performed on the output from the recording channel circuit 42, and the arithmetic result is decoded and output by the decoder 46.

The signal obtained from the recording channel circuit 42 has a signal level of ± 2, as shown in FIG.
2, 0, +2} are taken, and a fixed threshold is used to decode this into binary data.
Value level detection and Viterbi decoding, which is maximum likelihood decoding, can be considered.

The ternary level detection sets a threshold level having a fixed value between 0 and +2 and 0 and -2 and decodes it depending on which area the sample point falls in, and the circuit is very simple. On the contrary, the detection ability is not so high. On the other hand, the maximum likelihood decoding (Viterbi decoding) is a method of estimating the most probable sequence as one sequence by using the values of the sample points before and after, and the detection is higher than the ternary level detection. When the same data is decoded, the bit error rate is improved by one digit to two digits as shown in the example of FIG.

Here, when decoding exactly the same data, four kinds of results shown in the following Table 1 can be considered as to what results the two decoders output.

[0047]

[Table 1]

In Table 1, ∘ indicates that the recorded data was correctly decoded, and x indicates that an error occurred. As an example, it is assumed that the bit error rate decoded by ternary detection is P _L and the result of Viterbi decoding is P _V. That is, the probability of occurrence of pattern 3) or 4) in Table 1 above is P _L , and pattern 2) or 4).
The probability of occurrence of is P _V.

That is, it is expressed by the relationship of P _L = P ₃ + P ₄ and P _V = P ₂ + P ₄ . Here, as described above, P _V <<
Since it can be set to P _L , P ₄ ≦ P ₂ + P ₄ = P _V <<
P _L becomes there is a relationship, the P ₄ << P _L. Therefore,
P ₄ has a negligible value with respect to P _L , and can be regarded as P _L ≈P ₃ . Further, from the relationship of P ₂ + P ₄ << P ₃ + P ₄ , since P ₂ << P ₃ , there is a probability that the decoding results of the two become different patterns 2) and 3), that is, an error in ternary level detection. It can be considered as a probability of occurrence.

By utilizing this property, even if the information about what was previously written cannot be obtained correctly, both the ternary level detection and the Viterbi decoder are operated at the same time, and the decoding results of both are compared. The error rate can be measured. In an actual device, the output from the Viterbi decoder is adopted as the decoding result. However, depending on the quality of the signal input to the decoder, the error rate at the time of ternary level detection and the error rate at the time of Viterbi decoding are detected. If the error rates are associated in advance, it is possible to estimate the error rate when the Viterbi decoding is performed using the error rate of the ternary level detection measured by the above method.

Next, a circuit example of the Viterbi decoder will be shown.
Viterbi decoding will be explained as a preparation before that. FIG. 11 shows a trellis diagram for one system (that is, the partial response PRS (1, -1)) taken out every other bit from the system using the partial response PRS (1,0, -1). Here, the branch metric is also displayed. In order to find a path that maximizes the sum of these branch metrics, the path metric L _k up to a certain sample time _k is the value L _k − of the path metric up to the previous sample time k−2.
By using 2, it can be expressed as the following formulas (1) and (2).

[0052]

[Equation 1]

[0053]

[Equation 2]

In order to output the optimum path while calculating this metric, three squarers, six adders and two comparators are required. Furthermore, a serial shift / parallel load register for storing the path is required. Therefore, instead of faithfully calculating the path metric, the algorithm using the differential metric reported by Wood et al. Is used to simplify the circuit.

Now, consider the Viterbi algorithm when there are only two states. The Viterbi algorithm is to determine data for each state at a certain time k while narrowing down one path having the largest likelihood to reach that state. The above-mentioned decoding circuit (decoder) is for realizing it faithfully.

As an example, when there are only two states, the branches that survive at that time can have only the following three patterns. State <-1> → state <-1> and state <-1> → state <+1
> State <-1> → State <-1> and State <+1> → State <+1
> State <+1> → State <+1> and State <+1> → State <-1
>

Therefore, it is easily understood that the pattern of state <+1> → state <−1> and state <−1> → state <+1> is not possible. Each of these patterns →
We will write ↑, →→, → ↓. Then, for each branch, which of these patterns survives is determined by calculating the path metric. Now, there are only two states,
The difference between the respective path metrics is expressed by the following mathematical expression (3).

[0058]

[Equation 3]

Focusing on this equation (3), let us consider whether it can be used to determine which pattern will survive. From the above equations (1) and (2), the following equation (4)
The relationship is established.

[0060]

[Equation 4]

[0061] In this case, 4y _k -ΔL _k - since ₂ is common, by determining the magnitude by comparing this value with 4 and -4, or reveals selected either branch. By calculating this, it is possible to determine which of the above-mentioned pattern branches survives. That is, even if the path metric itself is not calculated, if the differential metric is calculated, the path can be determined in the process. From Equation (3) described above 4y _k -ΔL _k - When thus by case analysis on the three ways by _two values, are expressed by Equation (5).

[0062]

[Equation 5]

Further, if variable conversion is performed with ΔL _k = 4y _p -4β, it can be expressed as the following formula (6).

[0064]

[Equation 6]

Now, let us consider the meanings of β and 4y _p .
β takes a value represented by the following formula (7).

[0066]

[Equation 7]

Β is the immediately preceding state transition candidate (location p)
Represents the transition pattern in. In other words, it represents the type of transition at a point where the transition (→ ↑ or → ↓) other than the first parallel path going back from the current time is considered as a candidate. On the other hand, y _p is the value of y at that time.

For example, if it is assumed that → ↑ has occurred in the previous branch (that is, the last branch that has not been fixed), β =
+1, and the update rule judgment condition and β and y _p at that time is as shown in FIG. In other words, the meaning of β can be regarded as having a role of adding an offset to the threshold value for the determination in the above expression.

In this way, the probability of the previous state transition candidate (location p) and the transition at the current sample point (location k) are compared, and the more probable one is determined as a new state transition candidate. I repeat.
Since the person who loses the judgment is considered to have no transition, the memory for storing the path needs to be randomly accessible so that the information at the p point or the k point can be updated.

When a circuit is realized based on such an algorithm, its block diagram is as shown in FIG. In FIG. 13, the reproduction data from the recording channel is divided into odd-numbered column samples and even-numbered column samples for arithmetic processing, and FIG. 13 shows the case of even-numbered column samples in detail as an example. That is, the even column samples from the recording channel are supplied to the subtraction circuit 52 and the register 53 via the switch 51. The register 53 stores the value of the previous state transition candidate y _p , and the subtraction circuit 52 subtracts the value of the register 53 from the even column sample and outputs it to the comparison circuit (comparator) 54.

The comparison circuit 54 has a threshold value of +2,
0 and -2 are given, and arithmetic processing is performed on the output from the subtraction circuit 52 and the output from the register 55 storing β. The operation of the comparison circuit 54 may be performed as shown in Tables 2 and 3 below, and the comparison circuit 54 outputs the output data shown in Tables 2 and 3.

[0072]

[Table 2]

[0073]

[Table 3]

56 operates based on the PLL clock,
k is an address register, 57 is an address register that stores p, 58 is a selection circuit that selects (selects) k or p, and 59 is output data from the comparison circuit 54 (RAM data) using the output of the selection circuit 28 as an address. ) For storing), a counter that counts up based on a reference clock. Data of the RAM 59 is read by using the output of the counter 60 as an address and sent to the switch 61. The switch 61 returns the odd-column samples and the even-column samples to the original array, and the data output is obtained.

With such a configuration, the squarer is 0
This requires only one adder, one adder, and two comparators. However, in addition to this, RA for storing the path
It is necessary to prepare M59. An example of operation when a certain signal is input to this circuit will be given below.
The RAM refers to the RAM 59.

Operation Example When an input waveform as shown in FIG. 14 is observed, the operation of the comparator (comparison circuit 54) and the change of each parameter are shown below. However, initial values are y _p = -2 and β = -1
And Since k = 0: input k ₀ = 1.6 y _k −y _p > 2, it can be determined that the condition F is satisfied. In other words, the upward divergence (hereinafter, appropriately referred to as divergence) because it is, the β +1, and _{p = 0, y p = y} 0.

Since k = 1: input k ₁ = 0.2 −2 <y _k −y _p ≦ 0, it can be determined that the condition B is satisfied. In other words, it means that the parallel path, β, y _p
Is written as it is, and data 0 is written to address 1.

K = 2: Input k ₂ = −0.2 −2 <y _k −y _p ≦ 0, so it can be determined that Condition B was met. In other words, it means that the parallel path, β, y _p
Is left as it is, and data 0 is written in address 2.

Since k = 3: input k ₃ = 2 y _k −y _p > 2, it can be determined that the condition C is satisfied. In other words, because it is upward diverg-ence, the β +1, and _{p = 3, y p = y} 3. Here, since the previous candidate has been lost, data 0
Write.

K = 4: Since input k ₄ = 0.2 −2 <y _k −y _p ≦ 0, it can be determined that Condition B was satisfied. In other words, it means that the parallel path, β, y _p
Is written as it is, and data 0 is written to address 4.

Since k = 5: input k ₅ = −0.4 y _k −y _p > −2, it can be determined that the condition A was satisfied.
In other words, because it is a downward dive-rgence, the β -1, and _{p = 5, y p = y} 5. Here, since the previous candidate was correct, data 1 is written in address 3 of RAM.

K = 6: Input k ₆ = −0.2 0 ≦ y _k −y _p ≦ + 2, so it can be determined that the condition E was satisfied. In other words, it means that the parallel path, β, y _p
Is written as it is, and data 0 is written to address 6.

K = 7: Since the input k ₇ = −2.0 y _k −y _p ≦ 0, it can be determined that the condition D was satisfied. In other words, because it is down diverg-ence, the β -1, and _{p = 7, y p = y} 7. Here, since the previous candidate has been lost, the data 0 is stored in the address 5 of the RAM.
Write.

K = 8: Since the input k ₈ = 0.2 +2 <y _k −y _p , it can be determined that the condition F was satisfied.
In other words, since it means upward divergence, β = 1,
_{p = 8, y p = y} 8 is intact, the data at address 7 1
Write.

Next, a circuit for ternary level detection will be described. This can easily be implemented in the form of a Viterbi decoder subset. That is, since only the sample value at the present time is focused on for decoding, it is not necessary to store the value of the previous state transition candidate y _p or p. Further, similarly, the value of β becomes unnecessary. Therefore, 3
The value level detection circuit is configured as shown in FIG.

In FIG. 15, the reproduction data from the recording channel is divided into odd-numbered column samples and even-numbered column samples for arithmetic processing, and FIG. 15 shows the case of even-numbered column samples in detail as an example. That is, the even column samples from the recording channel are supplied to the comparison circuit 62 via the switch 61. Thresholds +1 and −1 are given to the comparison circuit 62, the even-numbered column samples are compared with these thresholds +1 and −1 to perform arithmetic processing, and the comparison result is stored in the RAM 63 as RAM data. To be done.

64 operates based on the PLL clock,
The RAM 63 is an address register that stores k, and the RAM 63 uses the output of the address register 64 as an address.
The output data from 2 (RAM data) is stored. 65
Is a counter that counts up based on the reference clock, and the data of the RAM 63 is read by using the output of the counter 65 as an address and sent to the switch 66. Switch 66 puts the odd and even column samples back into the original array and the data output is obtained.

Here, a part for dividing into an even-numbered sequence and an odd-numbered sequence and a part for returning these to one sequence again can be shared with the above Viterbi decoder. Also, an address register 64 for writing the decoded result in the RAM 43.
By sharing (the one for storing k) and the control circuit necessary for reading it, the function as the three-valued level detection circuit can be added to the Viterbi decoder without increasing the circuit scale.

As described above, the number of errors can be counted by simultaneously implementing the Viterbi decoder and the ternary level detection circuit, comparing these decoding results, and adding a circuit that regards different ones as errors. become able to. This block diagram is shown in FIG.

In FIG. 16, a reproduction signal obtained by reproducing predetermined recording data recorded on, for example, a magnetic recording medium as an analog signal by using the partial response method is given to an A / D conversion circuit 71 which performs analog-digital conversion. ,
The A / D conversion circuit 71 converts the analog reproduced signal into an A / D signal.
Viterbi decoder 7 for conversion and digital data input
The 2- and 3-value level detection circuit is output to 73.

Specifically, the A / D conversion circuit 71 has a cycle in which the signal level of the reproduced signal rises and falls,
The signal level of the output signal is converted into a digital value, and the resulting input data is output to the Viterbi decoder 72 and the ternary level detection circuit 73. The Viterbi decoder 72 and the ternary level detection circuit 73 can be configured as one circuit that simultaneously realizes these functions by the method described above.

Comparison circuit (corresponding to error detection means) 74
Is a Viterbi decoder 72 and a ternary level detection circuit 73.
The decoding results are compared with each other, the decoding result is output, and the error detection result in which a different one is regarded as an error is output to the counter 75. The counter 75 counts the number of errors based on a command from the control circuit 76, and outputs the count result as the number of errors.

Operation of the Second Embodiment In the second embodiment, a method combining maximum likelihood decoding and level detection is used as a bit error detection method, and the bit error is measured by this method to record. A memory for storing the data is unnecessary, and the bit error can be measured with almost no additional circuit to the normal disk drive body, so the cost does not increase.

Therefore, since the shift amount can be optimized by the disk drive alone, for example, the shift amount is obtained at the time of assembling the disk drive, and at the time of initialization after the power is turned on, for example. It becomes possible to update the shift amount to an optimum value at the time of self-diagnosis. That is, even when the optimal shift amount has a different value due to a change in ambient temperature, a change over time, or the like, the optimal shift amount can always be obtained. As a result, the optimum shift amount is efficiently obtained as in the above-described embodiment, so that the off-track amount is always controlled to the optimum amount, and the head positioning accuracy can be improved.

Third Embodiment In the examples described in the first and second embodiments, the center line C2 of the data track Td and the center line C3 of the servo track are arranged so as to coincide with each other. That is, at the time of reproducing data, the reproducing head Hp performs tracking so that the reproducing head Hp just tracks the center line C3 of the servo track by the reproduced signal picked up, so that the reproducing head Hp also moves to the center line C2 of the data track in the data track. When the data is recorded, the reproducing head Hp tracks the data track at a position shifted from the center line C3 of the servo track by the shift amount when the data is recorded. I was trying to just track to the center line C2 of the track.

That is, since tracking is performed at a point shifted by the shift amount when recording data, in order to obtain the optimum shift amount while shifting the shift amount, "set shift amount → data Record →
It is necessary to repeat the procedure of "data reproduction and error count".

Therefore, as shown in FIG. 17, the center line C2 of the data track and the center line C3 of the servo track are displaced in the opposite directions by the shift amount Scomp calculated in advance from the information such as the head-to-head distance and the radius. Keep it.
At the time of recording data, the reproducing head Hp tracks so as to trace the center line C1 of the servo track.
Since the center line C1 of the servo track is displaced from the center line C2 of the data track by the shift amount Scomp obtained by the above calculation, the recording head Hr traces a point substantially coincident with the center line C2 of the data track Td.

At the time of reproducing the data, the center line of the data track recorded by the reproducing head Hp can be accurately traced by shifting the amount of shift obtained by the same method as in the first embodiment to perform tracking. .
That is, in order to accurately know the shift amount at each radial position, data is recorded in the data area and the bit amount at each shift amount is measured while gradually changing the shift amount to obtain the optimum shift amount. .

The bit error rate shows the change as shown in the graph of FIG. 5 when the shift amount is changed, but it should be the best when the reproducing head Hp accurately on-tracks the recorded data track. Therefore, this shift amount can be used as the optimum shift amount. By performing the above operation for each radius of the disk, the optimum shift amount at each radial position can be obtained.

As a preferable bit error rate measuring method, by using the means for combining the ternary value detection and the maximum likelihood decoding described in the second embodiment, what kind of data string the data recorded in the data zone is. Error can be detected without even knowing what it is, so if the data zone is already in use at the radial position being measured (that is, if some signal has already been recorded). For example, the step of recording any signal prior to measuring the error rate may be omitted.

Operation of the third embodiment When the bit error rate is measured in order to obtain the optimum shift amount, the data need only be recorded once, and the shift amount can be changed with respect to the data once recorded. While repeating the playback. Therefore, it becomes possible to more easily find the optimum shift amount in a shorter time.

By using a means for combining ternary detection and maximum likelihood decoding as a bit error rate measuring method, bit errors can be detected without knowing what kind of data string the data recorded in the data zone is. Since it can be detected, even if some data is already recorded in the data zone after the disk device actually starts operating, the optimum shift amount in the data zone can be calculated, so the shift amount is always optimized. It is possible to maintain.

In the examples shown in the first and second embodiments,
Since the bit error rate was measured by repeatedly recording and reproducing some data in the data zone, it is impossible to use the already used data zone. That is, for example, the usage is such that the optimum shift amount at each radial position is obtained before shipment, and tracking is performed using this shift amount during actual recording / reproduction. It is difficult to correct such.

On the other hand, in the method shown in the third embodiment, since the bit error rate can be obtained by using the reproduced signal from the data zone in which the data is actually recorded, the bit error rate may change depending on the secular change or the environmental change. Even if the shift amount changes, there is a feature that the optimum shift amount can always be obtained. Therefore, for example, it is possible to always maintain the optimum shift amount by updating the shift amount at each radial position each time the power is turned on, or by updating the shift amount every several hours of actual operation. Become.

[0105]

As described above, according to the present invention,
Bit error of the reproduction signal in the data zone while changing the off-track amount by correcting and controlling the reproduction head to be off-track from the track center in the opposite direction by the amount of shift of the recording head from the reproduction head at the target track position The rate is measured and the tracking is performed with the off-track amount that gives the best bit error rate as the optimum shift amount. Therefore, the optimum shift amount can be obtained efficiently, and the off-track amount is always optimized. Can be controlled to any amount. As a result, the positioning accuracy of the head can be improved.

[Brief description of drawings]

FIG. 1 is a diagram showing a schematic configuration of a first embodiment of a magnetic disk device according to the present invention.

FIG. 2 is a diagram showing a disc format in the embodiment.

FIG. 3 is a diagram showing calculation of a shift amount in the same embodiment.

FIG. 4 is a block diagram of a circuit for obtaining an optimum shift amount in the embodiment.

FIG. 5 is a diagram showing an influence of a shift amount on a bit error rate according to the embodiment.

FIG. 6 is a block diagram of a circuit for detecting a bit error by the hardware of the embodiment.

FIG. 7 is a diagram for explaining a partial response which is the principle of the second embodiment of the magnetic disk device according to the present invention.

FIG. 8 is a diagram showing an example of an apparatus for performing inverse conversion of a partial response in the same embodiment.

FIG. 9 is a diagram showing an aspect of a signal level in the example.

FIG. 10 is a diagram showing a decoding mode result for information reproduction in the example.

FIG. 11 is a diagram showing a trellis diagram for information reproduction in the example.

FIG. 12 is a diagram illustrating a Viterbi algorithm for information reproduction according to the embodiment.

FIG. 13 is a block diagram showing an example of a circuit that realizes a Viterbi algorithm for reproducing information in the embodiment.

FIG. 14 is a diagram showing an example of an input waveform for information reproduction in the example.

FIG. 15 is a block diagram showing an example of a circuit for ternary level detection in the same embodiment.

FIG. 16 is a block diagram showing an example of a circuit for reproducing information in the same embodiment.

FIG. 17 is a diagram showing a third embodiment of the invention and a track pattern.

FIG. 18 is a diagram showing a track pattern of a conventional magnetic disk device.

[Explanation of symbols]

 1 Support Arm 1a Rotation Axis 2 Slider Main Body 3 Disk 11 Address Extraction Circuit 12 Correction Value Generation Circuit 13 Shift Amount Setting Circuit 15 Support Arm Drive Motor 16 Data Detection Circuit 17 Error Measuring Circuit 18 Shift Amount Error Rate Storage Memory 31-33 Arithmetic Circuit 34, 35, 51, 61 switch 41 precoder 42 recording channel circuit 43, 45 arithmetic processing circuit 46 decoder 52 subtraction circuit 53, 55 register 54, 62 comparison circuit (comparator) 56, 57, 64 address register 28 selection circuit 59, 63 RAM 60, 55, 65 Counter 71 A / D conversion circuit 72 Viterbi decoder 73 Three-value level detection circuit 74 Comparison circuit (error detection means) 76 Control circuit Hp reproducing head Hr recording head Td data track E s Servo area Ed Data area C1 Servo track center line C2 Data track center line C3 Support arm center line

Claims (6)

[Claims] 1. A reproducing head and a recording head are separately provided at the tip of a supporting arm whose base end side is rotatably supported, and the reproducing head and the recording head are rotated substantially in the radial direction of a disk by the rotation of the supporting arm. The disc is moved, and a servo zone for recording a servo signal and a data zone for recording a data signal are provided on the same disc surface on this disc, and the servo signal is picked up by the reproducing head,
In a head positioning mechanism of a magnetic disk device which performs track servo by displacing the rotational position of the support arm based on the picked-up servo signal, a shift amount in the disk radial direction of the recording head with respect to the reproducing head at a target track position Comprising a correction means for correcting and controlling the reproduction head to be off-track from the track center in the opposite direction, and a measurement means for measuring the bit error rate of the reproduction signal in the data zone while changing the off-track amount. A head positioning mechanism for a magnetic disk device, wherein tracking is performed by using an off-track amount having the best rate as an optimum value of the shift amount. 2. The apparatus according to claim 1, further comprising means for measuring the optimum shift amount for several tracks, and means for obtaining shift amounts at other radial positions by interpolation based on the measured shift amounts. 2. A head positioning mechanism of the magnetic disk device according to 1. 3. A reproducing means for reproducing predetermined record data recorded on a recording medium by using a partial response method, and a signal level of a reproduced signal reproduced from the recording medium, into a digital signal at a predetermined cycle. An analog-digital conversion circuit, a Viterbi decoding circuit that decodes the reproduction signal based on the output data output from the analog-digital conversion circuit, and compares the signal level of the reproduction signal with a predetermined reference value. And a level detection decoding circuit for decoding, comparing the decoding data by the Viterbi decoding circuit, and the decoding data by the level detection decoding circuit, and an error detecting means for detecting these inconsistencies, The head positioning mechanism for a magnetic disk drive according to claim 1, wherein the head positioning mechanism is provided. 4. A reproducing head and a recording head are separately provided at the tip of a supporting arm whose base end side is rotatably supported, and the reproducing head and the recording head are rotated substantially in the radial direction of the disk by the rotation of the supporting arm. The disc is moved, and a servo zone for recording a servo signal and a data zone for recording a data signal are provided on the same disc surface on this disc, and the servo signal is picked up by the reproducing head,
In a head positioning mechanism of a magnetic disk device for performing track servo by displacing a rotation position of the support arm based on the picked up servo signal, a disk radial direction of the recording head with respect to the reproducing head at each rotation position of the support arm. And a correction means for correcting and controlling the center line of the servo track to be shifted from the center line of the data track in the opposite direction by approximately the shift amount of the above, and for correcting the reproduction head to be off-track from the track center at the target track position. Measuring the bit error rate of the reproduction signal in the data zone while changing the off-track amount, and performing tracking while performing correction control using the off-track amount that gives the best bit error rate. Characteristic magnetic disk equipment Head positioning mechanism. 5. A means for measuring the optimum shift amount for several tracks, and a means for obtaining shift amounts at other radial positions by interpolation based on the measured shift amounts. Item 4. A head positioning mechanism for a magnetic disk drive according to Item 4. 6. A reproducing means for reproducing predetermined recording data recorded on a recording medium by using a partial response method, and a signal level of a reproduction signal reproduced from the recording medium is converted into a digital signal at a predetermined cycle. An analog-digital conversion circuit, a Viterbi decoding circuit that decodes the reproduction signal by a Viterbi decoding based on output data output from the analog-digital conversion circuit, and compares the signal level of the reproduction signal with a predetermined reference value. Then, the level detection decoding circuit for decoding, the decoding data by the Viterbi decoding circuit, and the decoding data by the level detection decoding circuit are compared, and error detection means for detecting these inconsistencies, The head positioning mechanism for a magnetic disk drive according to claim 4, wherein the head positioning mechanism is provided.