Miscellaneous active electrical nonlinear devices – circuits – and – Signal converting – shaping – or generating – Amplitude control
Reexamination Certificate
1994-12-30
2001-01-09
Zweizig, Jeffrey (Department: 2816)
Miscellaneous active electrical nonlinear devices, circuits, and
Signal converting, shaping, or generating
Amplitude control
C327S561000, C360S067000
Reexamination Certificate
active
06172548
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present application relates to read amplifiers for magnetoresistive magnetic storage systems, particularly for disk drives.
2. Description of the Prior Art
Magnetoresistive Heads
Magnetoresistance is a solid-state phenomenon wherein the resistance of an element is affected by the magnetic field around it. This physical phenomenon has been discussed for some years as a way to read the data which is stored in magnetization patterns in tape and disk drives. Disk drive manufacturers have now developed the fabrication technology to manufacture magnetoresistive drives on a large scale.
In a disk drive, traditionally the head was a coil (or more recently a thin film head which is equivalent to a coil), imbedded in some form of a support that slid across the top of the disk platter, and positioned to create a magnetic field in a small area of the surface of the platter. By controlling the amount of current that flows to the coil, and switching it from one direction to the other direction, a series of magnetic dipoles would be created in the ferromagnetic medium at the surface of the disk. That exact same coil would also be used in the read-back mode, to detect changes in the magnetic field vector of the magnetic medium.
Normally a “1” is indicated, on the disk, by a transition in the magnetic field. No transition would imply a zero. (These transitions are synchronized in ways not relevant here.) In the read-back mode, the coil becomes a BEMF generator, and a change in magnetic flux (due to crossing a domain boundary in the medium) will induce a voltage on the coils, which is sensed and amplified to detect the changes in the magnetic structure of the medium. This sensed voltage therefore provides a readout of the information stored in the magnetic medium, without disturbing it.
A disk drive normally includes multiple head elements each mounted on respective arms. The arms move across the disk and trace out various rings of magnetic data. If we could see the magnetic domain boundaries in the magnetic medium on the disk, we would see chains of overlapping circles, almost like overlapping punch-outs, where the write head changed its magnetic field and pushed out a new flux domain. The written domains are spaced closely enough to overlap (and therefore very few of them are circular), but there is enough remaining area in each one to preserve the written data.
If we are reading back with the same coil, then we have to make sure we have a space between the tracks of data, so that the magnetic flux of the adjacent track cannot interfere with coil readings. This intertrack separation requirement limits the density, so that in effect the density has an ultimate limit defined by the dimensions of our thin film head.
However, if we could somehow read a narrower stripe than the written strip, we could avoid intersymbol interference between adjacent tracks. Thus the limiting factor in traditional systems is not the writing, but the fact that traditional systems have to read back with the same inductive element. If we could read back using a sensor with smaller width (i.e. that magnetically coupled to a narrower area of the disk's magnetic medium), then we could reduce the track-to-track spacing. Even while using the same inductive element for writing, and writing with the same data frequency, we can write adjacent tracks more closely together, so that the circular magnetic domains of adjacent tracks actually overlap.
The magnetoresistive (“MR”) head has a very narrow strip to it. (In fact, MR heads are actually fabricated with integrated circuit fabrication technology.) Thus, MR heads provide a way to read magnetic data which is very closely spaced (with very close track-to-track spacings) without intersymbol interference. By contrast, if we were to try to read this pattern back with a thin film head, we would get a tremendous amount of intersymbol interference, and we would not get a reliable read-back. (An example of a modern MR head design is shown in Saito et al., “Development of a magnetoresistive/inductive head and low noise amplifier IC for high density rigid disk drives,” E76-A IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS, COMMUNICATIONS AND COMPUTER SCIENCES 1167-9 (1993), which is hereby incorporated by reference.)
The MR sensing element, basically, is a strip in the head which has a variable resistance which is affected by the ambient magnetic field. To sense this variable resistance, it is biased with a constant current: changes in the resistance will then appear as a voltage change. Since the MR sensing element is a physically narrow strip, it can ride right down the middle of the track, and avoid most of the intersymbol interference from the overlap at the edges of the tracks. Also, the readout from an MR head is not a function of the transitions (as with a sensing coil), but simply a function of the magnitude of the magnetic field.
BIASING AND INTERFACING TO THE MR HEAD
Thus magnetoresistive (“MR”) head technology promises to produce a new generation of disk drive density. However, the interface requirements to an MR head are significantly different from those of conventional read heads. The bias current through the magnetoresistive element must be set so as to optimize the operating point of the read elements.
State-of-the-art tape drives also employ MR heads, to achieve the advantage of speed-independent signal amplitude as well as increased bit density.) In such systems, since the preamplifier may have to support simultaneous reading from a number of MR elements, power consumption may also be an important consideration.
Background on MR head and preamplifier technology may be found, e.g., in Rohen, “Wave-shaping circuit for a magnetoresistive read head,” 21 IBM TECHNICAL DISCLOSURE BULLETIN 984-5 (August 1978); Jones, “Magnetoresistive amplifier,” 20 IBM TECHNICAL DISCLOSURE BULLETIN 4114-15 (March 1978); van Gestel et al., “Read-out of a magnetic tape by the magnetoresistance effect,” 37 PHILIPS TECH. REV. 42-50 (1977, no. 2-3); Robinson et al., “A 0.8 nV/ square root Hz CMOS preamplifier for magneto-resistive read elements, “1994 ISSCC 252-3; all of which are hereby incorporated by reference.
FIG. 1
shows the conventional signal input connections for a read-head preamplifier. A certain bias current Ihead is applied to the read-head (herein indicated as resistor Rhead), and the magnetoresistive variation in the head induces an AC voltage V
head
across terminals V+ and V−. The AC component V
head
is much smaller than the DC drop I
head
R
head
. The AC component of this signal is therefore coupled through two coupling capacitors C
AC
to the inputs of a first preamplifier stage Al.
FIG. 2
shows the frequency response of a system like that shown in FIG.
1
. Note that the upper corner frequency f
BW
is set by the time constant R
in
C
AC
, where R
in
is the input impedance of the preamplifier stage Al. The attempts to integrate the coupling capacitors C
AC
necessarily result in a small capacitance value (e.g. 100 pF or so), so that it is desirable to have a large input impedance R
in
to maintain good high-frequency response. However, this produces a problem, as will now be described.
For power conservation reasons, it is desirable, in many system implementations, to discontinue the bias current to the read head when write operations are occurring. When the bias current to the read head is restored, much of the resulting change in voltage will be coupled through the coupling capacitors, so that the amplifier inputs will initially be forced to a voltage difference which represents the DC bias across the sensing resistor, which is much larger than the typical magnitude of the AC components. The input amplifier will therefore be overloaded, and clamped in an overload condition until the input impedance R
in
has discharged the DC component of the coupling capacitors. This may be a relatively long time. With the sample values of Rin and C
AC
given above, the RC time constant is only about 1 &mgr;sec;
Alegre de la Soujeole Axel
Cameron Scott Warren
Galanthay Theodore E.
Jorgenson Lisa K.
STMicroelectronics Inc.
Telfer Gordon H.
Zweizig Jeffrey
LandOfFree
Input stage with reduced transient time for use in... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Input stage with reduced transient time for use in..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Input stage with reduced transient time for use in... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2441198