JPS6120944B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a magnetic head positioning method for positioning a magnetic head on a predetermined track on a data surface with high precision.

In movable head type magnetic disk devices for computers, etc., it is necessary to position the magnetic head on a predetermined track on which information in a large number of tracks on a data surface is to be recorded or reproduced. Therefore, conventionally, as shown in FIG. 1, information necessary for positioning is recorded on a servo surface 1 of a large number of magnetic disk surfaces, and this information is reproduced by a servo head 2 for obtaining position information, and the control circuit The position of the data head 5, which is rigidly coupled to the servo head 2, is determined by detecting the position using the data head 3 and driving the actuator 4 so as to minimize the position error. In addition, in FIG. 1, four magnetic disks are shown, the lower surface of the lower magnetic disk is the servo surface 1, the other surfaces are the data surfaces 6, the servo head 2 is on the servo surface 1, and the data head 5 is on each data surface 6. We are beginning to respond.

However, in this method, the relative position of the data surface 6 and the data head 5 is determined by the position of the servo head 2 and the data head 5, which are rigidly coupled, or the position of the servo surface 1 and the data surface, which are also rigidly coupled. If the position with respect to 6 is shifted due to thermal deformation or the like, it will shift, and if there is a significant temperature rise or local temperature non-uniformity, there is a drawback that precise positioning cannot be performed. This is because the positioning control of the data head 5 does not constitute a complete closed loop with respect to the position of the data surface 6. Therefore, another method is to partially record control information in the data surface 6. In addition, a so-called sector servo type data surface servo has also been proposed, which performs positioning in a closed loop by reading the control information from the data head 5 and performing positioning when the data head 5 moves in accordance with its position. However, in this case, the circuit becomes complicated and the usage of the data surface 6 is limited due to the pre-recorded position information.

In order to overcome these drawbacks, the present invention provides a magnetic head positioning system in which the information necessary for positioning the data head is obtained from the corresponding data surface of the data head. The invention will be explained below with reference to the drawings.

First, the principle of this invention will be explained. Second
As shown in Figure a, a large number of data tracks 7 with track widths TW are arranged on the data surface 6 at track pitches.
They are arranged at TP intervals. TP-TW=x (guard band width), and is usually designed so that TW/TP≒0.6 to 0.8. Of course, information is recorded on the data track 7, and in the case of digital recording, there is no correlation between information on adjacent tracks. Therefore, it cannot be used to detect a control signal using the timing of information between adjacent tracks, such as the conventional servo surface 1. However, if the positions of the data track 7 and the data head 5 shift, that is, go off-track, the amplitude of the regenerated signal S decreases as shown in FIG. 2b. When the amount of deviation exceeds the guard band width x, signals of adjacent tracks are also reproduced, so that the average reproduced signal amplitude becomes constant.

The present invention utilizes the change in the amplitude of the reproduced signal S due to this off-track. In other words, fluctuations in the reproduction output due to off-track occur depending on which side of the track position the data head 5 is located on (a, b in Fig. 2).
Since the same problem occurs even if the deviation is positive or negative, it is only necessary to determine this directionality. Further, since a control signal is also required when recording information, a dedicated head for achieving the above-mentioned purpose may be installed in the vicinity of the data head 5.

FIG. 3 shows an embodiment of the present invention, in which position signal detection heads 8a and 8b for detecting the direction and amount of deviation are arranged rigidly in the vicinity of the data head 5. It is something. When the position signal detecting heads 8a and 8b, each having a core width TW' narrower than the track width TW, are arranged at the very edge of the track width TW, the reproduction output fluctuation will be similar to that shown in FIG. 2b in the direction away from the data track 7. Decrease. However, when moving toward the center of the data track 7, there is no output variation with respect to only the position signal detection heads 8a and 8b. Position signal detection heads 8a and 8b
The outputs of are as shown in Fig. 4a and b, and when these are added via the level detector 9 in the opposite direction as shown in Fig. 3, the data track 7 and the data When there is no positional deviation of the head 5, zero, positive, and negative error voltages can be detected corresponding to zero, positive, and negative deviations. The controllable range is basically ±
x (guard band width), but if the positional deviation exceeds ±x, if it occurs during recording, it will be recorded on the adjacent track, so there is no point in controlling it beyond ±x. , which is within a sufficient range for practical purposes. In order to enable control in the range of at least ±x, the core width TW' must be such that the original signal can be reproduced even if it deviates by x, and that the output does not decrease even if it deviates by x toward the track center. Therefore, it is necessary that x≦TW′≦(TW−x), but usually x/TW=0.25 to 0.7
It is possible to design (possible if x/TW≦1).

Since the position signal obtained in this way uses the end of the information track, the signal quality may be poorer than that of the conventional servo surface method. In particular, if a method with large amplitude fluctuations is selected as the information signal recording method, it will be difficult to distinguish between errors due to positional deviation and signal amplitude fluctuations due to patterns, resulting in a deterioration of signal quality. Position signal detection heads 8a, 8 for detecting the position since there are a large number of information tracks
If multiple b are used at the same time, due to the principle of averaging,
It becomes possible to improve signal quality.

FIG. 5 shows another embodiment of this invention,
An example is shown in which there are three position signal detection heads 8a and 8b. If the number n=9, S/N=
Although improvements of as much as 10 dB are possible, such magnetic heads are extremely simple using integrated head processing techniques.

In the description of this invention, the position signal detection heads 8a, 8b are placed adjacent to the data head 5, but as described above, the position signal detection heads 8a, 8b are placed adjacent to the data head 5.
As long as the positions of the data track 7 and the data track 7 are maintained, they may be separated from each other. For example, position signal detection heads 8a and 8b may be placed on the sliders on both sides of the floating head, respectively, and the data head 5 may be placed between the sliders, for example in the center.
There is no problem in placing the . The distance between the position signal detection heads 8a, 8b and the data head 5 is nTP±(TW-TW')/2, provided that n is an integer. If the position signal detecting heads 8a and 8b are on both sides of the data head 5, the sign of ± is an arbitrary one, and when they are on one side, the sign is + or -, respectively.

As explained above, according to the present invention, since positional information is obtained from the data surface, highly accurate head positioning of the data head is possible, and there is no need to record special positional information to obtain control information. Therefore, in terms of signal processing, there is an advantage that the conventional method can be used as is.

[Brief explanation of the drawing]

Fig. 1 is an explanatory diagram of a conventional magnetic head positioning system, Fig. 2 is a diagram of output change characteristics due to off-track, Fig. 3 is a schematic diagram of a configuration showing an embodiment of the present invention, and Fig. 4 is an implementation of Fig. 3. FIG. 5 is a perspective view schematically showing the configuration of another embodiment of the present invention. In the figure, 1 is a servo surface, 2 is a servo head, 3 is a control circuit, 4 is an actuator, 5 is a data head, 6 is a data surface, 7 is a data track, 8a,
8b is a position signal detection head, and 9 is a level detector.

Claims (1)

[Scope of Claims] 1. In a system in which an arbitrary track of a large number of data tracks with a track width TW arranged at equal intervals TP is selected by moving a data head whose core width is TW, the core width is equal to or smaller than the track width. narrower than TW
2n (n: integer) position signal detection heads each having a TW' are arranged with respect to the data head in a relationship of n TP ± 1/2 (TW - TW'). A positioning method for a magnetic head, characterized in that the signal output of one head group and the signal output of another head group are rectified and summed in opposite directions to obtain a position signal, and the data head is positioned using this position signal. However, the ± sign in the above equation takes either sign when n head groups are placed on both sides of the data head, and when n head groups are placed on one side of the data head. Assume that + and - signs are assigned to each head group.