JP2000173035A - Head ginbal assembly and magnetic disc drive - Google Patents

Abstract

PROBLEM TO BE SOLVED: To achieve a high data transfer rate at high recording density by roviding a suspension with a flexure part for connecting an electromagnetic conversion element, an active circuit part for processing I/O signal of the electromagnetic conversion element, and a load beam part disposed between the flexure part and the active circuit part thereby reducing the parasitic impedance due to wiring between a magnetic head and the active circuit part. SOLUTION: A head ginbal assembly is provided with a silicon suspension 50 comprising a load beam part 1, a flexure part 2 and an active circuit part 3. The suspension 50 is made of single crystal silicon and a read/write IC is formed in an active circuit region 53. A wiring pattern 51 is formed on the surface of the flexure part 2 and the load beam part 1 through an insulation layer and a pad 52 is provided at an end for fixing a slider while the other end is connected electrically with the I/O end of the read/write IC. According to the structure, wiring length is shortened between a magnetic head and the read/write IC while reducing parasitic impedance.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a head gimbal assembly for a magnetic disk drive, and more particularly, to a structure of a front end of a base plate. In addition, the present invention relates to a magnetic disk drive on which this is mounted.

[0002]

2. Description of the Related Art With the recent progress of multimedia in the information-oriented society, there has been an increasing demand for smoothly processing enormous information such as voices, still images, and moving images. For this reason, microprocessors and peripheral logic have been required to increase the processing speed, but more importantly, large-capacity, high-speed data storage devices.

[0003] Magnetic disk drives, which are non-volatile data storage devices, are superior to other storage devices in terms of capacity, high speed, unit cost per bit, etc., and thus range from large computers to consumer electronic devices such as personal computers. It has been used in a wide range of fields. 2. Description of the Related Art In recent years, the areal recording density of a magnetic disk medium has been rapidly increasing at an annual rate of 60% due to advances in recording media and electromagnetic transducers (recording and reproducing heads). Also, the data transfer speed is increasing year by year due to an increase in the number of disk rotations.

[0004] In the current magnetic disk device, a floating recording system is generally adopted. In this method, when the recording medium rotates, the flying force of the air fluid generated by the high-speed rotation and the load of the suspension pressing the slider against the medium are balanced, and the slider flies above the medium with a small interval.

It is well known that in order to increase the surface recording density, the flying height of the slider, that is, the distance between the head and the magnetic disk medium is reduced as much as possible, and the bits formed on the medium are miniaturized. However, a decrease in the reproduction output due to the miniaturization of bits must be compensated for. Therefore, in recent years, a magnetoresistive (MR) sensor is often used instead of a conventional inductive head that senses a magnetic field by a coil. The MR sensor utilizes a change in electric resistance due to the magnetization of a material, and provides a higher output signal than an inductive head. Therefore, the signal-to-noise ratio at the time of reproduction is higher, and a high areal recording density of the magnetic disk medium can be achieved.

On the other hand, in addition to a method of increasing the output signal to increase the signal-to-noise ratio, a method of reducing noise is also important for increasing the signal-to-noise ratio. FIG. 22 is a schematic view showing the periphery of an actuator of a conventional general magnetic disk drive.

As shown in FIG. 22, in a conventional magnetic disk drive, a signal reproduced by a reproducing head (not shown) mounted on a slider 54 includes a flexure 211, a load beam 212, a base plate 45, and a carriage arm 2.
A read / write IC 21 having a built-in amplifier for amplifying a reproduction signal which passes through a lead wire 213 arranged along the line 08 and is connected to a flexible printed circuit 214 or the like.
5 is connected. The flexure 211, the load beam 212, and the base plate 45 are made of stainless steel or the like, and are each mechanically connected by welding or the like.

[0008] Flexure 211, load beam 212,
The gimbal assembly including the base plate 45 and the slider 54 is called a head gimbal assembly. Thermal noise is generated due to the impedance of the resonance circuit formed by the inductance, capacitance, and resistance distributed in the system from the head to the amplifier.

In order to suppress this thermal noise, it is effective to increase the resonance frequency of the resonance circuit. As is well known, the MR head can greatly reduce the inductance as compared with the conventional induction type head, so that the effect is also very high in this respect. However, when the inductance of the head is reduced by the MR head, the head and the read / write IC 2 which are not problematic in the conventional inductive head are not required.
The influence of the parasitic inductance due to the wiring between the wires 15 increases.

[0010] This parasitic inductance is
3, Flexible printed wiring circuit 216, IC 215
The sum of the parasitic inductance of each of the wirings inside and the wiring inside the slider 54 is obtained.

For example, the inductance of the conventional inductive thin film head is about 120 nH, and the total parasitic inductance is about 50 nH. That is, the parasitic inductance of the wiring is less than 30% of the total inductance component.

On the other hand, when an MR head is used, since the head inductance is 1 nH or less, 98% or more of the total inductance component is the parasitic inductance of the wiring.

Therefore, in order to suppress thermal noise, MR
It is apparent that even when a head is used, the effect of suppressing thermal noise due to the low inductance of the head is not sufficiently exhibited due to the parasitic inductance of the lead wiring.

On the other hand, at the time of recording, there are the following problems. In order to increase the data transfer speed, it is important to reduce the inductance of the inductive recording head as much as possible and to shorten the rise time of the data signal. However, when the number of turns of the coil of the induction type recording head is reduced and the inductance is reduced, the write magnetic field does not have a magnitude that can sufficiently saturate the recording medium.

[0015] Thin-film heads are well known and are particularly useful as heads for increasing the data transfer speed. Since a thin film head can make a magnetic circuit smaller than a conventional bulk head, it is possible to reduce inductance while securing a sufficient write magnetic field. However, when the recording head inductance is reduced by the thin film head,
The influence of the parasitic inductance due to the wiring between the read / write IC 215 and the head, which has not been a problem in the conventional bulk head, increases. This problem is similar to that of the above-described reproducing head, but is more serious because the inductance component has a large effect on the rise time of the signal.

To solve the above problem, Japanese Patent Application Laid-Open No. 8-25545
No. 0 discloses one solution. JP-A-8-255450 discloses a flexure 211, a load beam 212, a base plate 45, and a carriage arm 20 shown in FIG.
8. An assembly including the actuator coil 207 is integrally formed. When the assembly is formed of silicon, the read / write IC 215 is integrated integrally around a pivot shaft for rotating the assembly. According to this method, of the wiring from the magnetic head to the read / write IC 215, the wiring after the flexible printed wiring circuit 214 can be omitted, so that the parasitic component due to the wiring can be reduced.

As described above, the MR head is a reproducing head useful for increasing the recording density. However, when an overcurrent flows into the MR head due to electrostatic discharge (ESD), there is a high risk that the MR element is very easily destroyed. For example, an MR head having a DC resistance of about 30Ω may be destroyed or change in characteristics by applying a voltage of only 120V. ESD is mainly caused by the discharge of the accumulated electrostatic charge due to the presence of a high resistance material such as a human body or plastic at the time of assembling the carriage assembly.

As described above, in the manufacturing process of the carriage assembly, sufficient care is required for handling the carriage assembly. If a defective head occurs despite careful handling, the defective head is replaced. Repair work needs to be performed.

In this case, in order to suppress an increase in cost caused by the repair work, it is preferable that the repair work can be performed in the smallest possible unit of parts. The most preferable method is to replace only the defective slider. However, since the slider is usually firmly fixed to the extremely deformable flexure at the tip of the suspension using an adhesive,
It is extremely difficult to remove the slider without plastically deforming the gimbal shape and completely removing the adhesive. Therefore, in the integrated assembly structure disclosed in Japanese Patent Application Laid-Open No. 8-255450, the entire assembly from the flexure including the read / write IC to the actuator coil must be replaced in the slider repair work. That is, in the disclosed structure, the unit to be repaired becomes large, resulting in an increase in cost.

In order to solve this problem, Japanese Patent Laid-Open No. 10-124839 discloses a head slider assembly in which a read / write IC in the form of a bare chip is laminated on a slider to reduce the wiring length between the read / write IC and the slider. It is disclosed in the gazette. In JP-A-10-124839, a bare chip IC is stacked on a slider via a resin layer, and a head in the slider is electrically connected to an integrated circuit in the IC using bump electrodes and the like. According to this method, since the wiring length between the read / write IC and the slider is only the height of the bump electrode, the parasitic inductance of the wiring can be extremely reduced.
Since the replacement of the suspension portion in front of the carriage arm is sufficient, the unit of repair can be reduced.

However, according to this method, the mass of a portion that can be moved by an actuator, which can greatly affect the time required for the slider to move between data tracks (seek time), increases by the amount of the read / write IC chip. Would. Increasing the mass of the moving part increases the moment of the entire actuator and acts to lower its mechanical resonance frequency. Further, since the IC is located at the tip of the movable portion, the increase in the moment is remarkable. Therefore, in order to increase the mechanical resonance frequency of the movable part and to achieve a shorter travel time between tracks, the method has a completely unfavorable effect.

In recent years, sliders have been reduced in size for the purpose of reducing the weight of movable parts and improving the productivity of sliders. On the other hand, it is difficult to drastically reduce the size of the read / write IC chip unless the configuration of the internal circuit is simplified. Therefore, it is also a serious problem that the weight of the read / write IC chip reduces the effect on the seek time due to the weight reduction of the slider which is becoming smaller year by year.

When a plurality of magnetic heads are mounted on the slider, the data transfer speed can be dramatically increased by a method of simultaneously recording or reproducing data on a plurality of data tracks. This is a method similar to a well-known disk array technology in a large-sized computer or the like for striping data to a plurality of magnetic disk devices, and it is understood that a disk array is constructed in one magnetic disk device. For example, if four sets of recording and reproducing magnetic heads are mounted on one slider and 4-bit data can be simultaneously recorded and reproduced, the data transfer speed can be simply quadrupled. This method can increase the transfer rate without reducing the length of the data bits on the data disk, so that the reduction in coercive force, that is, the phenomenon of thermomagnetic relaxation, which occurs when the bit length is reduced in the longitudinal recording method. Transfer speed can be increased without any problem.

However, this method is disclosed in
When a bare chip read / write IC disclosed in Japanese Patent No. 24839 is applied to a head-slider assembly in which the read / write ICs are stacked on a slider, the read / write ICs are required by the number of heads to which recording circuits and amplifier circuits are connected. Therefore, the size of the read / write IC is inevitably increased. Therefore, the weight of the tip of the head gimbal assembly also increases, and as a result, there is a problem that high-speed access is hindered as described above.

As shown in FIG. 22, the conventional head gimbal assembly has a slider 54 and a flexure 21 as shown in FIG.
1 and the load beam 212, and the assembling process is very complicated. JP-A-10-12
In the method disclosed in Japanese Patent No. 4839, there is a problem that productivity is extremely reduced because a step of connecting a bare chip IC is further added.

[0026]

As described above, in order to increase the signal-to-noise ratio during reproduction and achieve a high areal recording density of a magnetic disk medium, a head for reducing the inductance component of a reproducing head is required. However, thermal noise due to the parasitic impedance distributed in the system from the head to the amplifier could not be reduced.

It is also effective to reduce the inductance of the head in order to increase the data transfer speed at the time of recording, and a head has been developed to reduce the inductance component of the recording head. The parasitic inductance distributed in the system leading to the head could not be reduced.

In the structure disclosed in Japanese Patent Application Laid-Open No. 8-255450, in which an assembly including a flexure, a load beam, a base plate, a carriage arm, and an actuator coil is integrated, a read / write IC can be integrated into the assembly. Therefore, slider-IC
The wiring length between them can be shortened, and the parasitic impedance can be reduced.
However, with this structure, if the magnetic head is damaged at the time of assembling and repairing occurs, the read / write IC
It is necessary to replace the entire assembly including the above, and there is a problem that the manufacturing cost increases.

Japanese Patent Application Laid-Open No. 10-124839 discloses a read / write I / F comprising an amplifier circuit and a write circuit.
By stacking C on the slider, the slider-IC
Also disclosed is a method of shortening the wiring length between them and reducing the parasitic impedance. However, in this structure, by stacking the read / write IC on the slider, the moment of the movable portion is increased, and it is difficult to increase the resonance frequency of the movable portion by the actuator. As a result,
There has been a problem that it has become difficult to shorten the important access time with the improvement of the recording density and the data transfer speed of the magnetic disk device.

The problem is that the effect of reducing the weight of the slider is reduced when the slider is reduced in size, and a magnetic disk drive having a plurality of magnetic heads mounted on the slider and capable of high-speed data transfer. Therefore, a fundamental solution is strongly desired because it becomes a bigger problem. Further, at the same time, the number of assembling steps of the head gimbal assembly is increased, and there is a problem that productivity is reduced.

[0031]

SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and in particular, has been made to reduce a parasitic impedance caused by wiring between a magnetic head and a read / write IC.
Achieving a high data transfer rate at a high recording density, reducing the unit of repair when repairing the slider, suppressing an increase in the moment of the actuator movable part, shortening the access time, and manufacturing a head gimbal assembly Provided is a head gimbal assembly which can reduce the number of assembling steps at the time and dramatically improve productivity, and a magnetic disk device equipped with the same.

To achieve the above object, the present invention comprises a base plate, an integral suspension adhered to the base plate, an electromagnetic transducer adhered to the suspension, and a wiring connected to the suspension. The suspension has a flexure unit for connecting the electromagnetic transducer, an active circuit unit for processing input / output signals to the electromagnetic transducer, and a load beam unit between the flexure unit and the active circuit unit. A head gimbal assembly is provided.

Here, the suspension may be made of silicon. Further, a reinforcing layer can be provided on the active circuit section and the load beam section.
Further, the electromagnetic conversion element and the flexure section may be connected by a bump electrode.

Alternatively, the load beam portion may have a constricted portion. Here, the suspension is composed of a slider support for mounting the slider and an elastic flexure that allows the slider to rotate in the pitch and roll directions, with adhesive, welding, and rivets attached to each part and its boundary. , A spiral, etc., and is made of a continuous and homogeneous material. It is also possible to add a load beam part that applies a load to the slider integrally. In this case as well, the boundary between the load beam part and the flexure part is made of any connecting material such as adhesive, welding, rivet, spiral, etc. It is composed of a continuous and homogeneous material without the use of

Further, an active circuit for transmitting a recording signal to the magnetic head or an active circuit for receiving a reproduction signal from the magnetic head connects a separately formed circuit with a connection material such as an adhesive, welding, a rivet, or a spiral. Not a thing,
The suspension arm can be formed directly on the surface thereof based on a continuous and homogeneous material constituting the suspension arm. With the above configuration, a high-performance suspension arm on which an active circuit is formed can be realized with a very simple structure and light weight.

Here, the silicon thickness of the flexure portion of the suspension arm reduces the total mass and moment of the actuator movable portion comprising the carriage arm,
As a result, the seek speed is increased, and at the same time, the roll rotation and pitch rotation movement of the slider are enabled. In order to secure a suitable spring property, that is, a suitable displacement amount for a proper load, about 100 μm or less. It is desirable that Further, in consideration of acceleration applied during seeking, impact resistance required for a disk device, and handling during assembly, it is desirable that the silicon thickness be about 20 μm or more.

The slider and the read / write IC are electrically connected at a very short distance, which is the height of the bump electrode. It is well known that the height of the bump electrode has a strong correlation with the environmental resistance after connection, and the height is determined by the required connection reliability, the material of the bump electrode, and the shape of the bump electrode. In order to increase the parallelism of the slider to the flexure and accurately control the attitude of the flying slider, the bump electrode height distribution is ± 1 within the same IC plane.
% Is desirable.

The self-alignment effect that allows the different termination pads to be aligned with high precision in a self-induced manner when the bump electrodes are melted and connected is well known in the art.
In particular, it is useful as a method for mounting a semiconductor device that requires high-precision alignment. For the purpose of fixing the slider with high accuracy using the self-alignment effect, a structure in which these termination electrodes are melted by heat and connect different termination pads to each other is used. When the bump electrode is melted by heat, the melting point of the bump electrode material is lower than the maximum temperature in the head and IC manufacturing process in order to protect the elements constituting the head and the read / write IC from thermal destruction or irreversible characteristic fluctuation. It is desirable that
This temperature is approximately 250 ° C. Further, the melting point of the bump electrode material is desirably 100 ° C. or higher for the purpose of ensuring sufficient environmental reliability.

The present invention further provides a carriage arm,
A base plate supported by the carriage arm;
An integrated suspension bonded to the base plate, an electromagnetic transducer bonded to the suspension,
A wiring connected to the suspension, and control means for controlling the base plate, wherein the suspension has a flexure unit for connecting the electromagnetic transducer, and an active circuit unit for processing input / output signals to the electromagnetic transducer. ,
A magnetic disk drive comprising a load beam section between the flexure section and the active circuit section.

[0040]

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, the present invention will be described in detail.
The present invention reduces the parasitic impedance due to the wiring between the magnetic head and the read / write IC, achieves a high data transfer rate at a high recording density, and can reduce the repair unit when repairing the slider, and can reduce the mass of the head gimbal assembly. The present invention provides a magnetic disk drive capable of reducing the access time and reducing the access time, and further simplifies the assembly process at the time of manufacturing the head gimbal assembly. Focusing on the structure of the head gimbal assembly, an assembly structure in which a read / write IC is formed integrally with the suspension arm and the slider is electrically and mechanically connected by bump electrodes is effective.

An embodiment of the present invention will be described below with reference to FIGS. The present invention can be used with various modifications without being limited to the following embodiments. As shown in FIG. 20, the magnetic disk drive according to this embodiment mainly includes a magnetic disk 201, a spindle motor 202, an actuator mechanism 203, and a head gimbal assembly 20.
4, the control circuit board 205 and the like.

The magnetic disk 201 is rotated by a spindle motor 202 and the head gimbal assembly 20 is rotated.
Data is read from and written to the magnetic disk 201 by a magnetic head (not shown) fixed to the end of the magnetic disk 201. Although FIG. 20 shows an example in which the number of the magnetic disks 201 is one, it is needless to say that a plurality of magnetic disks and a plurality of head gimbal assemblies can be provided corresponding to the magnetic disks.

The head gimbal assembly 204 is attached to the carriage arm 208 and the actuator mechanism 2
03, the target track on the magnetic disk (not shown) by moving on the magnetic disk 201
Align the head with.

As the actuator mechanism 203, a voice coil motor can be used. The voice coil motor is a linear motor having a movable coil 207 arranged in a magnetic field generated by a fixed permanent magnet 206, and its operation is precisely controlled by a positioning control circuit (not shown) built in a control circuit board 205. Is controlled.

As the positioning control method, a closed loop control method in which position information is recorded at the same time as data recording at the time of data recording, and the head is positioned by driving the voice coil motor while reading the position information at the time of reproduction can be used.

The control circuit board 205 includes a read channel IC (not shown) for converting a data signal reproduced from a magnetic head into a digital signal, and an interface control IC (FIG. 10) for transmitting the digital signal to a host interface. (Not shown), a cache memory IC (not shown) for temporarily storing data before transmission to the interface, and the like.

The present invention is particularly characterized by the structure of the head gimbal assembly 204. FIG. 1 shows a schematic diagram of a head gimbal assembly. As shown in FIG. 1, the head gimbal assembly includes a base plate 45, a suspension 50 including a load beam and a flexure, and a slider (not shown) for mounting a recording head and a reproducing head. FIG. 2 is a plan view of the suspension 50, FIG. 3 is a side view of the tip of the head gimbal assembly, and FIGS. 4 to 6 are straight lines AA ', BB', and CC 'in FIG. Are respectively shown in FIG.

The base plate 45 functions as a connection when the head gimbal assembly is mounted on the carriage arm, and is made of a stainless steel plate (SUS304 or the like) having a thickness of about 300 μm. The wiring pattern 14 is formed on the surface of the base plate 45 via the insulating layer 13 which is thin over the entire length. The end of the wiring pattern 14 to which the suspension 50 is fixed is the pad 43.
And the other end is provided with a pad 15.

The suspension is formed of a single crystal silicon having a thickness of about 30 μm, and has a read / write I / O.
This is the silicon suspension 50 on which C is formed. Young's modulus of silicon single crystal is 1.9 × 1012 dyne /
cm ² is close to stainless steel and nickel, Knoop hardness is 850 kg / mm ^2, which is harder than nickel and ordinary glass, and the flexural strength is 6.9 × 109 dyne / cm ^2, which is more than 3 times that of stainless steel and is extremely superior in mechanical properties. Has characteristics.

Therefore, a suspension comparable to a conventional suspension using stainless steel can be formed. The load beam portion 1 of the silicon suspension 50 has a sufficiently high rigidity required to apply a slight load in the direction in which the slider is pressed against the magnetic disk. This embodiment has a structure in which a reinforcing layer 47 is formed on a single-crystal silicon having a thickness of about 30 μm, as shown in FIGS.

The flexure section 2 fixes the slider at the tip of the load beam section 1 and has a sufficiently low rigidity required for the slider to follow the undulating vibration of the magnetic disk surface. And a follow-up performance with respect to a rotation operation along the pitch rotation axis. In this embodiment, the flexure section 2 has a thickness of about 30.
It is made of a silicon single crystal of μm, and there is no reinforcing layer 47 formed in the load beam portion. Therefore, the flexure portion is flexible because there is no backing by the reinforcing layer.

The active circuit section 3 of the silicon suspension 50 is connected to the load beam section 1, and a reinforcing layer 47 is formed on the surface thereof in the same manner as the load beam section 1. The active circuit section 3 can also function as a load beam. On the silicon surface layer of the active circuit section 3, an active circuit area 53 having a read / write IC function mainly consisting of a recording switching circuit and an amplifier circuit is formed by a well-known integrated circuit forming method.

The load beam section 1, flexure section 2 and active circuit section 3 are all formed on the basis of a uniform silicon single crystal having a thickness of about 30 μm, and are continuously integrated.
Therefore, there is no mechanical connection between them, and the structure is extremely simple and lightweight as compared with a conventional suspension in which stainless steel is assembled by welding or the like.

A wiring pattern 51 is formed on the surface of the flexure portion 2 to the load beam portion 1 of the silicon suspension 50 through a thin insulating layer over substantially the entire length thereof.
Pads 52 are provided at the ends to which are fixed. On the other hand, the other end of the wiring pattern 51 is electrically connected to the input / output end of the circuit via a via hole (not shown) formed in the insulating layer 49 of the active circuit area 53 of the active circuit section 3.

In order to supply power to the read / write IC circuit and electrically connect with the read channel IC, pads 56 electrically connected to the input / output terminals of the built-in circuit are formed on the active surface of the silicon suspension 50. Then, a bump electrode 46 is provided thereon.

As the insulating layer 49 formed on the silicon suspension 50, a polyimide resin or the like having a low dielectric constant is used, and its thickness is about 5 μm. For example, a laminated film of a titanium film / copper film / titanium film is used for the wiring 51 with low resistance copper as the center. Here, the titanium film functions as an adhesive layer with the resin, and its thickness may be about 50 nm. Further, the thickness of the copper film is set to a thickness that does not impair the flexibility of the flexure. In this embodiment, the thickness of the copper film is about 2 μm.

Further, the insulating layer 48 functions as passivation, and similarly, a polyimide resin is used.
It is about μm. The insulating layer 48 has pads 56 and pads 5.
An opening is provided to expose 2.

The reinforcing layer 47 serving as a silicon reinforcing material is made of, for example, a copper-tin alloy having a thickness of about 10 μm. The copper-tin alloy is suitable for the present reinforcing layer because it has a feature that a thick film can be easily formed by a plating method, has high corrosion resistance, and has a small internal stress of the film. By setting the composition ratio of tin in the alloy to 5 to 50% by weight, the film thickness distribution,
An alloy film having excellent composition distribution and surface smoothness can be obtained.

As shown in FIG. 2, the silicon reinforcing material 4
Since the copper-tin alloy layer serving as 7 covers all of the active region 53 and a part of the wiring pattern 51, if the alloy layer 47 is connected to the ground potential, external electromagnetic waves that can affect the circuit and the input / output wiring can be obtained. Also works effectively as a shield.

Here, since silicon has conductivity, the wiring pattern 51 forms a strip line structure in the load beam portion 1 and forms a microstrip line structure in the flexure portion 2. Therefore,
It is important for high-speed data transfer to calculate the characteristic impedance of the line by each part and change the line width of the wiring pattern 51 so as not to form a discontinuous part of the characteristic impedance.

FIG. 7 is a perspective view of the slider of this embodiment. The slider 54 is made of a ceramic material such as aluminum oxide and titanium carbide, and has a head 72 on one end of which a recording head and a reproducing head are stacked. An inductive thin film head may be used for the recording head.
The reproducing head is an MR head or a giant magnetoresistive (GM)
R) A head composed of elements may be used. The slider 54 faces the magnetic disk and has an air bearing surface 73 on which a protruding rail is formed and a back surface 74 facing this surface.

A lift is generated on the air bearing surface 73 by the air flow generated by the rotation of the magnetic disk, and the lift and the load by the load beam are balanced, and the slider 54 flies a certain amount from the magnetic disk surface. A taper 75 is formed at an end facing the end face on which the head 72 is formed, for introducing air necessary for floating the slider. Slider back surface 74
Are formed with pads 76 electrically connected to the input end of the recording head and the output end of the reproducing head.

As shown in FIGS. 4 and 6, in this embodiment, the silicon suspension 50 and the base plate 45 are used.
Is the silicon suspension 5 by the bump electrode 46.
0 and the slider 54 are mechanically and electrically connected by a bump electrode 55. The bump electrode 46 connects the pad 56 on the silicon suspension 50 and the pad 43 on the base plate, and the bump electrode 55 connects the pad 52 on the silicon suspension and the pad 76 on the slider.

These pads 43, 56, 52, 76
Serves as a base for a bump electrode described later.
It is constructed by laminating metal films that mainly have the effect of preventing metal diffusion and improving adhesion. In this embodiment, a stacked film of nickel and titanium is used for copper wiring. When a tin-lead eutectic solder having a high tin content is used as a bump electrode, the nickel film exhibits wettability when the solder is melted and acts as a tin diffusion preventing layer. The thickness may be about 1 μm or more in consideration of the diffusion rate of tin. Nickel reacts relatively slowly with tin and has been shown to exhibit high reliability in environmental reliability tests (S. Honma, et.
l., "Effectiveness of thin film barriermetals for
eutectic solder bumps ", Proceedings of ISHM 96, p
p.87-92, 1996 and K. Tateyama, et. al., "Barrier
metals for use with high melting point solder bum
ps in flip-chip interconnections ", Proceedings of
IMC96, pp. 318-323, 1996). The titanium film acts to increase the adhesion strength between the underlying wiring and the pad, and may have a thickness of about 50 nm. Further, a gold film or a platinum film having a thickness of about 50 nm may be laminated on the nickel film as a corrosion prevention layer for preventing oxidation of nickel.

The bump electrodes 46 and 55 are made of a material which can be melted at a temperature at which the magnetic head element and the integrated circuit element in the IC are not destroyed or whose characteristics are not changed, for example, indium, tin-lead alloy, tin-bismuth alloy, tin-gold alloy, tin A solder material such as a zinc alloy or a tin silver alloy can be used. In this embodiment, a tin-lead alloy is used as the bump electrode, and in particular, the bump electrode 55 for connecting the silicon suspension 50 and the slider 54 is used.
213 ° C with a composition ratio of 90% by weight of tin and 10% by weight of lead
Using a silicon suspension 5
A tin-lead alloy having a composition ratio of 63% by weight of tin and 37% by weight of lead and a melting point of 183 ° C. was used as the bump electrode 46 connecting the 0 and the base plate 45. These melting points are determined by the maximum temperature (approximately 300 degrees) added in the manufacturing process after IC device formation.
° C) or the maximum temperature (approximately 250 ° C) applied in the manufacturing process of the magnetic head, and the time required to maintain the melting point temperature during connection is at most about 1 minute. It can be understood that there is almost no influence.

Similarly, for example, when a tin-bismuth-based alloy is used as the bump electrode, the silicon suspension 50
The composition ratio of the bump electrode 55 connecting the slider 54 and the slider 54 is tin 91.9% by weight, bismuth 4.8% by weight, and silver 3.3.
The bump electrode 46 connecting the silicon suspension 50 and the base plate 45 is made of an alloy having a composition ratio of 42% by weight of tin and 58% by weight of bismuth and having a melting point of 138 ° C. be able to.

Hereinafter, a method of manufacturing the head gimbal assembly of this embodiment will be described. First, with a thickness of about 625 μm,
A 5-inch diameter silicon wafer having a read / write IC formed on its surface is prepared.

The read / write IC comprises a recording switching circuit for generating a recording signal from an input data signal and a head amplifier circuit for amplifying a reproduction signal. However, in the present invention, these circuit configurations and devices are not particularly limited, and necessary circuits can be additionally provided as appropriate. The read / write IC is manufactured by a method well known in the art. That is, pattern transfer by photolithography, reactive ion etching or wet etching, formation of a silicon oxide film by thermal oxidation, formation of a silicon oxide film or polycrystalline silicon film by a CVD method or the like, aluminum or gold film by a sputtering or vapor deposition method A desired IC can be manufactured by combining unit processes such as formation of a semiconductor and doping by ion implantation.

As shown in FIG. 8, the IC is designed in advance so that a metal layer other than the silicon oxide film, a polycrystalline silicon layer, a doping layer, and the like are not formed in a region other than the IC region 53 of the silicon wafer 81.

FIG. 9 is an enlarged view of one chip area surrounded by the dicing line 82 in FIG. 91 in the figure
Reference numerals a to 91k denote input / output pads of the IC. FIG.
FIGS. 0 to 15 are partial cross-sectional views of the wafer for explaining the manufacturing process of the silicon suspension 50. FIG. In the figure, the left side of the dashed line shows a cross section of DD ′ in FIG. 9, and the right side of the dashed line shows a cross section of EE ′. In FIGS. 10 to 15, active elements and wirings in the active region 53 are omitted except for input / output pads and a passivation film near the uppermost layer of the IC.

Aluminum is used as the input / output pad 91 in FIG. 10, and its peripheral portion is partially covered with a passivation film 101 such as a silicon oxide film. In this embodiment, the pad 91a is a pad connected to a magnetic head, and the pad 91b is a pad connected to a read channel IC on an external substrate. The dimensions, the number of electrodes, and the electrode pitch of the semiconductor element are appropriately selected.

The thickness of the passivation film 101 is small,
If the insulation and flatness can adversely affect the wiring, pads and bump electrodes formed on it,
Different insulating films can be stacked over the passivation film 101. Here, the different insulating films are those containing silicon nitride, silicon oxide, polyimide, epoxy, benzocyclobutene, or the like as a main component.

Next, as shown in FIG. 11, a first insulating layer 49 is formed on the wafer 81. As the first insulating layer 49, a photosensitive polyimide resin or the like is used. This resin varnish is applied on the wafer by spin coating or curtain coating to form a coating film having a thickness of about 7 μm. Thereafter, baking is performed at 80 ° C. for 20 minutes, and drying is performed.

Further, a via hole 111 for electrically connecting the input / output pads 91a and 91b of the IC and the wiring layer 51 is formed in the polyimide film by exposure and development, and thereafter, at 350 ° C. for 30 minutes in nitrogen. Cure and cure the resin. The film thickness after curing is about 5 μm. As the first insulating layer 49, a non-photosensitive polyimide resin can be used instead of the photosensitive resin. In this case, after coating the varnish, curing is performed at 350 ° C. for 30 minutes, a carbon dioxide laser is irradiated, and the resin is ablated to form a via hole 111.

The surface of the first insulating layer 49 formed as described above is subjected to a plasma treatment in order to increase the adhesion to the wiring layer 51. This treatment method is known as a treatment for increasing the adhesion between the metal and the resin. The resin surface is exposed to a plasma such as oxygen or nitrogen gas to form fine irregularities on the surface and to form a carboxyl on the resin surface. This is a method in which a polar group such as a group is formed to increase affinity with a metal and enhance adhesion. Further, by this processing method, the polyimide thin film remaining at the bottom of the via hole after the development can be removed at the same time, and as a result, the electrical connection between the pads 91a and 91b and the wiring layer 51 can be further ensured. The first insulating layer 49
No 2 exists in the flexure section on the right side of FIG.

Next, a wiring layer 51 is formed. Wiring layer 51
Has a laminated structure of, for example, a titanium film / copper film / titanium film. The thickness of this laminated film is set to be about 2 μm in total. Here, the copper film functions as the center of the wiring, and the titanium film functions as an adhesive layer for improving the adhesion between the copper film and the resin surface. Therefore, the thickness of the titanium film may be small, and about 50 nm is sufficient. in this case,
Since the titanium surface is easily oxidized, it is preferable to continuously form an upper copper film without breaking vacuum. By forming a titanium film and a copper film in this way,
Intervention of a natural oxide film can be prevented, and a wiring having high adhesion and low resistance can be obtained.

Next, the laminated film is processed into a wiring pattern by a photolithography technique. A resist (not shown) is applied on the wafer by spin coating, baked, and processed into a wiring pattern by exposure and development.

The titanium film other than the wiring pattern is subjected to reactive ion etching with a mixed gas of carbon tetrafluoride and oxygen, and the copper film is etched with a mixed solution of ammonium persulfate, sulfuric acid and ethanol.

Finally, the wiring layer 51 can be formed by dissolving and removing the resist film with acetone. The wiring connected to the pad 91a is the flexure section 2 on the right side of FIG.
To form a termination portion 112. The wiring connected to the pad 91b is the same as that of the left active circuit unit 3 shown in FIG.
It is terminated above to form a termination 113.

Next, as shown in FIG. 12, the second insulating layer 4
In step 8 and the subsequent steps, a barrier metal layer 122 for forming a termination pad is formed. The second insulating layer 48 can be formed using a polyimide resin in exactly the same steps as the first insulating layer 49.

A via hole 121 is formed in the termination portions 112 and 113 of the wiring layer 51. The barrier metal layer 122 is formed of, for example, a stacked film of a titanium film and a nickel film.
First, a titanium film and then a nickel film are sputtered.
The layers are continuously laminated by a vapor deposition method or the like. The thickness of each film is about 50 nm and about 1 μm. Further, these laminated films simultaneously act as base electrodes for electroplating described below. If an oxide film is present at the film interface, the adhesion of the film is significantly reduced. Therefore, it is preferable to form these films continuously without breaking the vacuum for the purpose of preventing the natural oxide film from intervening.

Subsequently, a solder bump electrode is formed on the termination 112. The bump electrode is formed by depositing a metal such as tin or lead using a metal mask and selectively depositing it on the pad, bonding a metal wire to the pad,
A well-known method for mounting a semiconductor chip, such as a method of cutting a wire or a method of arranging a metal ball on a pad, can be appropriately selected and used for the present invention. In this embodiment, in order to obtain high productivity, a case will be described where an electroplating method is used in which bump electrodes can be collectively formed in a wafer state and the film forming speed is high.

The method of forming the bump electrode by the electroplating method is as follows.
K. Tateyama, et.al., "Barrier metals for use with
high melting point solder bumps in flip-chip inte
rconnections ", Proceedings of IMC 96, pp.318-323,
It is known as disclosed in 1996, for example by the following procedure.

As shown in FIG. 13, the barrier metal layer 122
A solder plating resist 132 is formed on the
Patterning is performed so that the upper surface of the substrate 12 is opened. As the solder plating resist 132, for example, a high-viscosity positive resist capable of forming a thick film can be used. In this embodiment, the thickness of the resist 132 is about 50 μm.

The resist pattern is formed by applying a resist by a spin coating method, performing pre-baking, and performing a known exposure and development process. Next, solder plating is performed on the wafer on which the solder plating resist 132 has been formed.
Solder plating is faster than electroless plating.
An electroplating method that allows easy management of the plating solution is preferred.

As the electroplating apparatus, it is desirable to use a jet-type plating apparatus or the like that makes the concentration of ions to be plated uniform on the wafer surface as shown in FIG. The solder plating 160 flows in from the lower part of the apparatus and circulates. For example, a sulfonic acid-based plating solution is used as the solder plating solution 160, and the composition ratio of the solder film after plating is 90% by weight of tin and 1% of lead.
The amounts of tin ions and lead ions are adjusted so as to be 0% by weight. As the anode 163, a metal plate having the same material and composition as the desired film to be plated is used, and a plurality of holes are formed so as not to obstruct the flow of the plating solution.

The cathode electrode 161 and the anode electrode 16
4 is connected to the current supply 162. By bringing the cathode electrode 161 into contact with the nickel film of the barrier metal layer 122 and energizing in the plating solution 160,
As shown in FIG. 7, a solder film 131 is formed in the opening of the plating resist 132. The thickness of the solder film 131 is about 50 μ
m.

In FIG. 16, the cathode electrode 161 is installed around the wafer 81 in order to provide a larger usable area of the wafer. The barrier metal layer 122 made of a nickel film and a titanium film serves as a base electrode in plating, but nickel having a lower resistivity than titanium is formed over the entire surface of the wafer in a thickness of about 1 μm. The resistance will be low enough. Therefore, the current concentration in the vicinity of the cathode electrode is reduced, and the plating film thickness distribution in the wafer is within ± 10% for a 5-inch diameter wafer, and the plating film thickness distribution in the range of each IC chip is ± 10%.
Less than 1% is possible. After the plating is completed, the resist 132 for solder plating is sufficiently washed with water, and the solder plating resist 132 is dissolved and peeled off with an organic solvent such as acetone or an alkaline solution.

Next, as shown in FIG. 14, a solder plating resist 132 is formed on the wafer by a method similar to the method shown in FIG. 13, and patterning is performed so that the termination portion 113 of the wiring 51 is opened. Using an electroplating apparatus shown in FIG. 16, a solder having a composition ratio of 63% by weight of tin and 37% by weight of lead is plated on the opening. The thickness of the plating film is about 50 μm. A sulfonic acid-based plating solution is used as the solder plating solution 160, and the amounts of tin ions and lead ions are adjusted so that the composition ratio in the solder film after plating is 63% by weight of tin and 37% by weight of lead. Subsequently, the solder plating resist 132 is dissolved and peeled with an organic solvent such as acetone or an alkaline solution.

Further, as shown in FIG. 15, a reinforcing layer 47 is formed by electroplating in the same manner as for the solder bump electrodes. The reinforcing layer 47 is made of copper on the barrier metal layer 122.
It is formed of a tin alloy. A resist for solder plating (not shown) is formed on the wafer, and patterning is performed so that a portion where the reinforcing layer 47 is formed is opened, and a composition ratio of copper 90% by weight, tin 10% by an electroplating apparatus shown in FIG. A weight percent of the alloy is plated in the openings to a thickness of about 10 μm. As the plating solution 160, a mixed solution of copper pyrophosphate, stannous pyrophosphate, and sodium pyrophosphate is used. Copper ions and tin are mixed so that the composition ratio in the film after plating is 90% by weight of copper and 10% by weight of tin. Adjust the amount of ions. Subsequently, the plating resist is dissolved and peeled with an organic solvent such as acetone or an alkaline solution.

Further, the barrier metal layer 122 other than the lower portions of the bump electrodes 55 and 141 and the reinforcing layer 47 is removed by etching using the solder bump electrodes 55 and 141 as a mask.
The etching solution for removing the nickel film of the barrier metal layer 122 has a hydrochloric acid mixing ratio of 2% or less and a nitric acid mixing ratio of 13%.
Hereinafter, a mixed solution having an acetic acid mixing ratio of 85% or more is used. Acetic acid has the effect of passivating tin contained in the solder to reduce the etch rate, and nickel has less of this effect, increasing the nickel / tin etch rate ratio.
Since the etching rate ratio of nickel / lead is sufficiently larger than 1, only the nickel film around the bump electrode can be selectively removed by using the bump electrode made of a tin-lead alloy as an etching protection mask.

Further, as an etching solution for removing the titanium film of the barrier metal layer 122, a mixed solution composed of ammonia, ethylenediaminetetraacetic acid, and hydrogen peroxide is used. In an etching solution for removing a titanium film, titanium /
Since the etching rate ratio of tin and titanium / lead is sufficiently larger than 1, only the titanium film around the bump electrode can be selectively removed by using the bump electrode made of a tin-lead alloy as an etching protection mask. However, since the etching rate ratio of nickel to the copper-tin alloy is 1 or more in the present etching solution, the reinforcing layer 47 made of the copper-tin alloy is etched simultaneously with the nickel film, but is sufficiently thicker than the nickel film. There is no problem because the required pattern dimensional accuracy of the reinforcing layer is about 10 μm.

Further, by polishing the back surface of the wafer thus formed, the thickness of the wafer 81 is reduced by about 30.
The thickness is set to μm. The polishing is performed by a usual lapping apparatus using diamond, silicon carbide, aluminum oxide, cerium oxide or the like as abrasive grains. At the time of polishing, in order to protect the bump electrodes 46 and 55 formed on the wafer surface,
A resist (not shown) having a thickness of about 50 μm is applied to the entire surface of the wafer. As the resist, a thick resist similar to the plating resist can be used.

After the polishing, the resist is dissolved and peeled off with an organic solvent such as acetone or an alkaline solution. Through the above steps, a wafer 81 having the active area 53 of the read / write IC, the wiring layer 51, the bump electrodes 46 and 55, the reinforcing layer 47, and the like shown in FIG. 15 is obtained.

The wafer 81 thus formed is processed into a suspension shape. First, a resist is applied to the wafer 81 by a spin coating method, and prebaked.
A resist pattern is formed by a known exposure and development process. The resist pattern is processed according to the suspension outline 92 shown in FIG. As the resist, a thick resist similar to the plating resist can be used.

Thereafter, dicing is performed according to a dicing line 82 shown in FIG. 9 to process the wafer into a pellet. For dicing, a dicing apparatus using a general diamond blade used in a general IC manufacturing process can be used.

Next, the cut silicon pellet is processed by an etching process. Hydrofluoric acid for etching,
It is performed by wet etching using a mixed solution of nitric acid and acetic acid as an etchant. It should be noted that this etching solution proceeds isotropically with respect to the plane orientation of the silicon single crystal, so that the silicon side wall is also etched and an undercut occurs in which the processing proceeds inward from the resist pattern. Therefore, the pattern formed by the resist must be designed in consideration of the undercut.

Through the above steps, the silicon suspension shown in FIG. 2 can be formed. In the present invention, the silicon suspension formed as described above is connected to the slider and the base plate to form a head gimbal assembly. The steps of assembling the head gimbal assembly will be described with reference to FIGS. 1 to 4, FIGS. 6 to 7, and FIGS.

As shown in FIGS. 3 to 6, the connection between the silicon suspension 50 and the slider 54 is made by a bump electrode, and this connection is made by mounting a semiconductor element provided with the bump electrode on a substrate. It can be done by a method.

That is, first, a high-viscosity flux was applied on the pad 76 formed on the back surface 74 of the slider 54 shown in FIG. 7, and then the surface of the silicon suspension 50 on which the bump electrode was formed was applied with the flux. It is made to face the slider back surface 74.

FIG. 17 is a diagram schematically showing a process of assembling the silicon suspension 50 and the slider 54.
As shown in FIG. 17, the silicon suspension 50 is fixed to a two-dimensionally movable stage 171 by a jig (not shown), and the slider 54 is fixed by a vacuum chuck 172 from the back surface thereof. The CCD camera 173 is inserted between the opposing silicon suspension 50 and the slider 54 to image the bump electrode 55 on the silicon suspension 50 and the termination pad 76 on the back surface of the slider. The stage is moved while performing image processing for superimposing the images, and the bump electrodes and the pads are accurately positioned. Although the positioning accuracy depends on the accuracy of the apparatus, an accuracy of approximately ± 10 μm is sufficient.

After the alignment is completed, the vacuum chuck 17
2 is brought closer to the stage on which the silicon suspension is fixed, and the bump electrodes 5 on the silicon suspension 50 are moved.
When the pad 5 is brought into contact with the pad 76 on the slider 54, the slider is temporarily fixed on the silicon suspension due to the viscosity of the flux.

In this state, while the silicon suspension is fixed to the jig, the silicon suspension is moved onto a belt (not shown) of an infrared belt furnace to perform reflow. The reflow is performed at a temperature at which the bump electrodes 55 on the silicon suspension 50 are uniformly melted, for example, at 240 ° C. In the heating state of this reflow process, the bump electrode 55 melts and becomes a natural hemisphere due to its surface tension.
Pad 76 on slider where bump electrode is not formed
Wet the surface and connect the different pads with solder bridges. If the different pads are connected by a solder bridge and the heating continues for a certain period of time, the solder moves the slider 74 to an appropriate position by its surface tension. This effect is known as a self-alignment effect, and the alignment accuracy between the pads is ± 1 higher than the accuracy at the time of temporary fixing.
An accuracy of about μm can be obtained.

Further, the relationship between the bump electrode height distribution in the IC and the roll angle or the pitch angle after the connection of the slider in this embodiment is geometrically expressed by the following equation. = tan ^-1 (h / 100l) where is the roll angle or pitch angle of the slider 54, is the percentage of the bump electrode height distribution, h is the average height of the bump electrodes, and l is the distance between the bump electrodes 55, respectively. Is shown.

From this equation, l is 0.7 mm and h is 50 μm.
In the case of the present embodiment, the bump electrode height distribution in the IC is ± 1% or less, so that connection with extremely high parallelism of ± 0.040 ° or less is possible.

After the connection, the residual flux is sufficiently dissolved and removed in an appropriate organic solvent using an ultrasonic cleaner or the like, and the connection between the slider 54 and the silicon suspension 50 is completed. In practicing the present invention, not only an infrared belt furnace, but also a more advanced laser reflow device can be used.

Subsequently, the silicon suspension 50 to which the slider 54 is connected and the base plate 45 are connected. As shown in FIGS. 3 and 4, in the present invention, the connection structure using the bump electrode 46 is also used for the electrical and mechanical connection between the silicon suspension 50 and the base plate 45 in the same manner as the connection between the silicon suspension 50 and the slider 54. Can be used.

As shown in FIG. 1, a wiring pattern 14 is formed on the surface of the base plate 45 via a thin insulating layer 13 over substantially the entire length thereof. Insulating layer 13 and wiring pattern 1
The wiring layer 4 can be formed by bonding a known flexible printed wiring board (FPC) on the base plate 45.

FIG. 18 is an enlarged plan view showing the FPC on the side to which the silicon suspension 50 is connected, and FIG. 19 is a sectional view taken along line FF 'in FIG. FPC
Has a wiring pattern 14 formed by etching a copper foil formed on an insulating film 13 made of a polyimide film via an adhesive layer 191, and a coverlay 192 made of polyimide is formed on the uppermost layer. It is normal.

The thickness of the insulating film 13 is about 100 μm, and the thickness of the wiring pattern 14 is about 18 μm. Pads 43 are formed at the ends of the wiring pattern, and a nickel film / gold film laminated film 193 formed by electroless plating serving as a barrier metal is formed on the surface. The thickness of the nickel film is about 4 μm, and the thickness of the gold film is 50 n
m.

The FP fixed to the base plate 45
The connection of the silicon suspension 50 to C can be performed in the same manner as the method in which the slider 54 and the silicon suspension 50 are connected. At this time, it is necessary to change the set temperature of the belt furnace so as not to re-melt the bump electrode 55 already connected to the slider 54 and the silicon suspension 50.

As described above, the bump electrode 46 connecting the silicon suspension 50 and the base plate 45 shown in FIG. 2 uses solder having a composition ratio of 63% by weight of tin and 37% by weight of lead. The upper bump electrode 46 is uniformly melted, and the slider 54
Is set at a temperature at which the bump electrode 55 connecting the electrodes is not melted, for example, 200 ° C.

The positional accuracy and the roll pitch angle after connection are the same as in the case where the slider and the silicon suspension are connected. The roll pitch angle of the silicon suspension with respect to the base plate is the sum of the respective angles. The angle is 0.080 ° or less, and connection with extremely parallelism can be achieved.

In the present invention, as shown in FIG. 3, the bump electrode 46, 55 connects the silicon suspension 50, the slider 54, and the base plate 45. An appropriate sealing resin 31 may be arranged in the gap between the silicon suspension 50 and the slider 54 and in the gap between the silicon suspension 50 and the base plate 45 as necessary for the purpose of improving the reliability of the connection. As the resin to be sealed, for example, bisphenol-based epoxy and imidazole effect catalyst,
An epoxy resin or the like containing an acid anhydride effect agent and a spherical quartz filler can be used. The viscosity of the resin can be adjusted by changing the diameter and the amount of the filler to be mixed.

In disposing the sealing resin in the gap,
For example, the gap can be impregnated from the end of the silicon suspension using the well-known capillary phenomenon. Thereafter, it can be cured by heating at 100 ° C. to 160 ° C. for 2 to 6 hours.

When the magnetic disk drive is started and stopped, the air bearing surface of the slider and the surface of the magnetic disk medium tend to adhere to each other, and a large shearing force is applied to the slider. A good connection strength is obtained. Further, in a temperature cycle as an environmental reliability test, the breaking life of the bump electrode due to the difference in the coefficient of thermal expansion between the silicon suspension 50, the slider 54, and the base plate 45 can be significantly extended.

Before disposing the sealing resin in the gap between the silicon suspension 50 and the base plate 45, the pad 15 of the wiring pattern 14 shown in FIG.
It is preferable to perform a write and read operation test. By this test, deterioration of the magnetic head characteristics due to the assembling process up to that time can be grasped. If this operation test reveals that the magnetic head characteristics have deteriorated, repair is performed.

In the present invention, the repair can be performed on a silicon suspension basis or on a head gimbal assembly basis. When replacing the silicon suspension, the connection between the silicon suspension 50 and the base plate 45 is uniformly heated by an infrared heater or the like, and the bump electrode 46 is re-melted to separate the silicon suspension 50 and the base plate 45.

FIG. 21 is a view of the tip of a head gimbal assembly according to another embodiment of the present invention as viewed from the side. The input / output pad 223 is formed on the head mounting surface of the slider 54 as in the conventional slider. The slider 5 on the silicon suspension 50
A pad 224 for making an electrical connection with the slider 4 is formed at the tip of the silicon suspension 50 so as to correspond to the input / output pad 223 of the slider 54, and the connection is made by a gold ball 221 by a well-known ball bonding method. Do.

The pad 22 on the base plate 45
5 and the silicon suspension 50 are electrically connected by a gold wire 222 by a well-known ball bonding method. For the mechanical connection between the slider 54 and the silicon suspension 50 and between the silicon suspension 50 and the base plate 45, a method via a conventional adhesive resin layer 226 can be used.

According to this method, unless a plurality of sets of magnetic heads are mounted on the slider 54, as long as the number of electrical connection points between the slider 54 and the silicon suspension 50 is small, the slider 54 as shown in FIG. The slider having the conventional structure can be used without preparing a slider having the pad 76 formed on the back surface 74 as shown in FIG. Therefore, there is no need for a step of forming the pad 76 on the back surface 74 in the manufacture of the slider 54,
Consequently, high costs are not required. Furthermore, since assembly can be performed using widely used wire bonding equipment, there is an advantage that equipment investment can be minimized.

The present invention is not limited to the above-described embodiment, but can be modified and implemented without departing from the gist thereof. In the above embodiment, the case where the base plate to which the FPC is bonded is used is described. However, the present invention is not limited to this structure. For example, a well-known structure in which an insulating layer and a wiring layer are directly formed on a base plate is formed. May be used. Further, the wiring layer 51 shown in FIG. 11 may be formed on a silicon wafer at the same time as the step of forming an IC in advance. This makes it possible to further reduce the number of insulating layers.

Further, the type of head mounted on the slider is not limited to the above-described embodiment. In addition, plating conditions, etching conditions, reflow conditions, and resin curing conditions are appropriately determined according to the material, shape, function, etc. of the sample. It is desirable to change.

Further, the wafer, slider, silicon suspension, insulating layer, wiring layer, bump electrode, sealing resin,
The plating solution and the etching solution can be used with various changes in the material, composition, dimensions, and the like, and it goes without saying that the bump manufacturing method and connection method are not limited to the above examples.

[0125]

As described in detail above, the present invention provides an assembly structure in which a read / write IC is formed integrally with a suspension comprising a flexure and a load beam, and a slider is directly connected to this suspension by bump electrodes or the like. By using this, the wiring length between the magnetic head and the read / write IC can be reduced and the parasitic impedance can be reduced without laminating the read / write IC on the slider, so that a high recording density and a high data transfer speed can be achieved.

In the case where the slider is repaired at the time of assembling, the present invention only requires replacement of the suspension, and the repair unit does not increase. further,
According to the present invention, it is possible to avoid an increase in the moment of the movable portion of the actuator, and it is easy to increase the resonance frequency of the movable portion by the actuator. As a result, it becomes easy to improve the recording density and the data transfer speed of the magnetic disk device and to shorten the important access time. This effect is effectively exhibited when the slider is further downsized.

Further, in the present invention, in order to form a read / write IC into a suspension larger than that,
For example, in a magnetic disk drive capable of forming a large-scale circuit and capable of high-speed data transfer in which a plurality of magnetic heads are mounted on a slider, the same number of heads mounted on the slider is required. Since the read / write circuit can be accommodated in the suspension, a higher-speed magnetic disk drive can be constructed.

Further, in the present invention, since the read / write IC, flexure, and load beam can be integrated, the number of elements constituting the head gimbal assembly can be reduced, and the number of steps of assembling the head gimbal assembly can be reduced. It also has the effect of improving productivity.

[Brief description of the drawings]

FIG. 1 is a schematic diagram illustrating the structure of a head gimbal assembly according to the present invention.

FIG. 2 is a schematic diagram illustrating a structure of a suspension arm in the head gimbal assembly of the present invention.

FIG. 3 is a schematic diagram for explaining the structure of the distal end portion of the head gimbal assembly of the present invention.

FIG. 4 is a schematic cross-sectional view illustrating the structure of the distal end portion of the head gimbal assembly of the present invention.

FIG. 5 is a schematic cross-sectional view illustrating the structure of the distal end portion of the head gimbal assembly of the present invention.

FIG. 6 is a schematic cross-sectional view illustrating the structure of the tip of the head gimbal assembly of the present invention.

FIG. 7 is a schematic diagram illustrating the structure of a slider in the head gimbal assembly of the present invention.

FIG. 8 is a schematic diagram illustrating a manufacturing process of a suspension arm in the head gimbal assembly of the present invention.

FIG. 9 is a schematic diagram illustrating a manufacturing process of a suspension arm in the head gimbal assembly of the present invention.

FIG. 10 is a schematic diagram illustrating a manufacturing process of a suspension arm in the head gimbal assembly of the present invention.

FIG. 11 is a schematic diagram illustrating a manufacturing process of a suspension arm in the head gimbal assembly of the present invention.

FIG. 12 is a schematic view illustrating a manufacturing process of a suspension arm in the head gimbal assembly of the present invention.

FIG. 13 is a schematic diagram illustrating a manufacturing process of a suspension arm in the head gimbal assembly of the present invention.

FIG. 14 is a schematic diagram illustrating a process of manufacturing a suspension arm in the head gimbal assembly of the present invention.

FIG. 15 is a schematic diagram illustrating a manufacturing process of a suspension arm in the head gimbal assembly of the present invention.

FIG. 16 is a schematic diagram of an electroplating apparatus used in manufacturing the head gimbal assembly of the present invention.

FIG. 17 is a schematic diagram illustrating an assembly process of the head gimbal assembly of the present invention.

FIG. 18 is a schematic diagram illustrating a wiring structure on a base plate in the head gimbal assembly of the present invention.

FIG. 19 is a schematic sectional view illustrating a wiring structure on a base plate in the head gimbal assembly of the present invention.

FIG. 20 is an exploded schematic view of the magnetic disk drive of the present invention.

FIG. 21 is a schematic diagram illustrating the structure of the tip of the head gimbal assembly of the present invention.

FIG. 22 is a schematic diagram illustrating the structure of an actuator arm in a conventional magnetic disk drive.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 13 ... Insulating layer 14 ... Wiring pattern 15 ... Pad 31 ... Sealing resin 43 ... Pad 45 ... Base plate 46 ... Bump electrode 47 ... Reinforcement layer 48 ... Second insulating layer 49 ... First insulating layer 50 ... Silicon suspension 51 ... Wiring pattern 52 ... Pad 53 ... Active circuit area 54 ... Slider 55 ... Bump electrode 56 ... Pad 72 ... Magnetic head 73 ... Air bearing surface 74 ... Slider back surface 75 ... Taper 76 ... Pad 81 ... Silicon wafer 82 ... Dicing line 91a-91k ... · IC input / output pads 92 ··· Lines indicating suspension outline 101 · · · Passivation film 111 · · Via holes 112 · · · Termination 113 Terminator 122 Barrier metal layer 121 Via hole 131 Solder film 132 Resist 141 Solder film 160 Plating solution 161 Cathode electrode 162 ... Current supply source 163 ... Anode 164 ... Anode electrode 171 ... Stage 172 ... Vacuum chuck 173 ... CCD camera 191 ... Adhesive layer 192 ... Coverlay 193 ... Barrier metal layer 201 Magnetic disk 202 Spindle motor 203 Actuator mechanism 204 Head gimbal assembly 205 Control circuit board 206 Permanent magnet 207 Coil 208 .. Carriage arm 211 Flexure 212 Load beam 21 … Wiring 214… Flexible printed wiring circuit 215… Read / write IC 216… Printed wiring circuit board 221… Gold ball 222… Gold wire 223… Pad 224… Pad 225 ... pad 226 ... adhesive layer

 ──────────────────────────────────────────────────続 き Continued on the front page (72) Inventor Takashi Tsugasaki 33, Shinisogocho, Isogo-ku, Yokohama-shi, Kanagawa Prefecture Inside Toshiba Production Technology Research Institute (72) Inventor Kazuki Tateyama Shinisogocho, Isogo-ku, Yokohama-shi, Kanagawa 33 Toshiba Production Technology Laboratory Co., Ltd. (72) Inventor Hiroaki Yoda 72 Horikawa-cho, Saiwai-ku, Kawasaki-shi, Kanagawa Prefecture F-term in Toshiba Kawasaki Office 5D042 NA02 SA02 TA09

Claims (6)

[Claims] 1. A base plate, an integrated suspension bonded to the base plate, an electromagnetic transducer bonded to the suspension, and a wiring connected to the suspension, wherein the suspension connects the electromagnetic transducer. A head gimbal assembly, comprising: a flexure section for performing the operation, an active circuit section for processing an input / output signal to the electromagnetic transducer, and a load beam section between the flexure section and the active circuit section. 2. The head gimbal assembly according to claim 1, wherein said suspension is made of silicon. 3. The head gimbal assembly according to claim 1, wherein a reinforcing layer is provided on the active circuit section and the load beam section. 4. The head gimbal assembly according to claim 1, wherein said electromagnetic transducer and said flexure are connected by a bump electrode. 5. The head gimbal assembly according to claim 1, wherein said load beam portion has a constricted portion. 6. A carriage arm, a base plate supported by the carriage arm, an integral suspension adhered to the base plate, an electromagnetic transducer adhered to the suspension, a wiring connected to the suspension, A control unit for controlling a base plate, wherein the suspension is a flexure unit for connecting the electromagnetic transducer, an active circuit unit for processing input / output signals to the electromagnetic transducer, the flexure unit and the active circuit unit. And a load beam portion between the magnetic disk devices.