WO1995035565A1

WO1995035565A1 - Nrz magnetic recording

Info

Publication number: WO1995035565A1
Application number: PCT/US1995/004502
Authority: WO
Inventors: Richard E. Fayling
Original assignee: Minnesota Mining And Manufacturing Company
Priority date: 1994-06-17
Filing date: 1995-04-17
Publication date: 1995-12-28

Abstract

In NRZ binary magnetic recording, each current reversal has a waveform edge that is steep while approaching and passing through zero (I0) but is made substantially less steep after passing a threshold level (It). This is accomplished by repeatedly incrementally delaying the increase in absolute magnitude of write current within the range of It and Imax. Doing so affords jagged trailing f-t-e profiles that separate adjacent flux transitions recorded on the magnetic recording medium. Because those jagged profiles are straighter than curved profiles of the prior art, the noise or intersymbol interference produced in adjacent bit cells upon read-out can be reduced, thus permitting increased linear bit densities.

Description

NRZ MAGNETIC RECORDING

Background of the Invention

Field of the Invention

The invention relates to NRZ (non return to zero) magnetic recording methods and systems in which binary data is recorded onto a magnetic recording medium that moves at a constant speed relative to a recording head. The invention is specifically concerned with modifying the write-current waveform to permit increased linear bit density.

Description of the Related Art U.S. Pat. No. 4,521,816 (Kougami et al.) shows four binary recording schemes (NRZ, NRZI, MFM and M²). In each of these schemes, the write current rapidly crosses zero when switched between positive and negative and thus is a non-return-to-zero or NRZ scheme.

In the following disclosure, the term "NRZ" (non return to zero) is used to include any binary or other magnetic recording scheme (including the NRZ, NRZI, MFM and M² of Kougami) wherein the write current rapidly crosses zero when switched between positive and negative.

The Kougami patent says: "In general, in prior-art methods of recording digital signals, a rectangular wave is used as the magnetic head driving current.... It is well known that the playback signal becomes a mountain-like pulse waveform with respect to one reversal of the driving current" (col. 1, lines 21-31). The

Kougami patent shows that a playback signal which indicates "1" does "not become zero at the centers of the adjacent information bits of '0', as shown at numerals 4 and 5 in FIG. 1(c)" (col. 1, lines 48-54). The Kougami patent says that the "intersymbol interference" at numerals 4 and 5 tends to become so large as to cause playback errors when the write current stays at either the positive or negative level for at least 1.5 times the bit period and that such errors can be avoided, as shown in Fig. 10(b), by momentarily making the current reversal amplitude at part 34 of the write signal smaller than it would have been in the prior art as indicated at part 33. This momentarily reduces the amplitude of the playback signal [see Fig. 10(d)] to level 34" (as compared to level 33" of the prior art). Kougami says that doing so reduces "intersymbol interference" in adjacent bit periods, although no reduction in adjacent bit periods is shown in Fig. 10(d). Any reduction in "intersymbol interference" would permit increased linear bit density.

U.S. Pat. No. 4,786,988 (Kobayashi) concerns a high-speed contact printing technique for duplicating video information recorded on a master tape and specifically concerns control signals on a sub-master tape which is duplicated from the master to obtain a mirror of the master signals from which "slave tapes" can be made by contact duplication. The control signals (which are binary write signals generated from NRZ write current) are re-formed as shown in Fig. 1(B) "to lower the spike noise components Ni, N₂, N3 and N₄ conventionally produced at the leading and trailing edges of the differentiated control signal spikes reproduced from the slave tape as shown in FIG. 1(C)" (col. 4, lines 3-8). The re-formed write signals of Fig. 1(B) are similar to the modified write signal of Fig. 10(b) of the above-discussed Kougami patent, but every leading edge of Kobayashi's write waveform is re-formed, not just those that are at least 1.5 times the bit period. Figs. 3(A) to 3(E) of the Kobayashi patent "show some possible modifications to the waveform of the control signals to be recorded on the sub-master tape" (col. 5, lines 5-7).

Many other patents and publications concern modifications of NRZ write signals that provide playback signals having reduced noise (or "intersymbol interference") in adjacent bit cells and thus permit increased linear bit density. See, for example, U.S. Pat No. 3,503,059 (Ambrico); U.S. Pat. No. 4, 167,761 (Best); U.S. Patent No. 4,845,573 (Hardeng); Kato et al.: "Write-Current Equalization for High-speed Digital Magnetic Recording", IEEE Transactions on Magnetics, Vol. MAG-22, No. 5, Sept. 1986, pp 1212-1214; and Jacoby: "High Density Recording with Write Current Shaping", IEEE Transactions on Magnetics, Vol. MAG- 15, No. 3, July 1979, pp 1124-1130. In spite of what is said in the above-cited patents and publications, NRZ binary magnetic recording systems now on the market, of which I am aware, switch the write current between +I_IMX and -I,„χ as fast as possible, and do not modify the write current except for the effects of damping resistors used to control ringing. Because of inherent inductances and other properties of the recording equipment and because of properties of magnetic recording media, the reversal portions of the playback waveforms are substantially sinusoidal (or "mountain-like" in the words of the Kougami patent).

I am the author or co-author of three publications about the nature of flux transitions produced when binary signals are recorded on magnetic recording media. One of these is Fayling: "Edge Profile Studies of Recorded Flux Transitions", IEEE Transactions on Magnetics, Vol. MAG-16, No. 3, Sept. 1980, pp 1249-1255 (here called "Fayling I"). Fayling I teaches that when a square-wave current pulse is recorded on magnetic recording tape that has a longitudinally oriented easy axis and has been premagnetized uniformly in the direction opposite to that of the head field, each flux transition produced in the medium normally has curved leading and trailing edges like those shown in several photographs.

A second of these is Fayling et al.: "A Model for Overwrite Modulation in Longitudinal Recording", IEEE Transactions on Magnetics, Vol. MAG-20, No. 5, Sept. 1984, pp 718-720 (here called "Fayling II"). Fayling II illustrates in each of Figs. 1, 2 and 3 a series of magnetically recorded bits as viewed from the side; showing a curved boundary at each edge of each flux transition.

A third of these is Fayling: "Studies of the Magnetizing Region' of a Ring- type Recording Head", IEEE Transactions on Magnetics, Vol. MAG- 18, No. 6, Nov. 1982, pp 1212-14 (here called "Fayling m"). Fayling IJJ shows in Figs. 2 and 3 experimentally determined flux transition edge ("f-t-e") profiles for different recording currents on a magnetic recording tape that has a longitudinally oriented easy axis and has been premagnetized uniformly in the direction opposite to that of the head field. Fayling LTJ points out (at page 1214, first column) that the curvature of most of the illustrated f-t-e profiles approximate the form predicted by the Karlqvist equation. "The five pairs of flux transition profiles in Figure 2, and the two pairs of flux transition profiles in Figure 3, show that the separation(s) between flux transitions ... increase with increasing write current amplitude. A nearly linear relationship was observed in Figure 2, between the maximum values of flux transition separation (f) and current amplitude" (p. 1214, first column). Szczech et al.: "Improved Field Equations for Ring Heads", IEEE

Transaction on Magnetics, Vol. MAG-19, No. 5, Sept. 1983, pp 1740-44, shows an equation that provides essentially the same information as the Karlqvist equation but should be more accurate.

Brief Summary of the Invention The invention provides an NRZ binary magnetic recording method and system which, like that of the Kougami patent, employs a magnetic recording medium that moves at constant velocity v relative to a recording head and achieves increased linear bit density by modifying the write-current waveform (that is, the I vs. t waveform, where I = current and t = time). The invention differs from that of Kougami by generating NRZ write current wherein each current reversal (between -Inland +Iπ_« and between +I_nuxand -I„ux) has a waveform edge that is steep (preferably as steep as possible) while approaching and passing through zero (lo) but is temporarily made substantially less steep after passing a threshold level (It) at which the write current begins to have an effect upon magnetic flux in the recording medium. Such temporary reduction in steepness can be achieved by incrementally delaying, for a time ta, the increase in absolute magnitude of write current between +1, and +I_mιx and between -I_t and -Im»χ. Doing so produces flux transition edge (f-t-e) profiles that can be straighter than are those of the prior art, thus affording increased linear bit densities.

During each such incremental time delay tj, the write current can either remain constant, or it can be allowed to increase more slowly than it was increasing while passing the threshold level I_t. The former should be easier and more economical to control.

Hereinbelow, reference to I„ux includes +!„,„ and -!„,„. Description of the Preferred Embodiments The method of determining each incremental time delay tj can be more easily understood in reference to the drawing.

Brief Description of the Drawing In the drawing, all figures of which are schematic: FIG. 1 is an edge view of fragments of a magnetic recording head and medium of an NRZ binary magnetic recording system of the prior art and shows idealized trailing f-t-e profiles of successive binary cells;

FIG. 2 is an enlarged fragmentary edge view of the system of FIG. 1 except showing trailing f-t-e profiles of binary cells as recorded in the prior art at three different (increasing) levels of write current, Iι,I₂and I₃;

FIG. 3 is an edge view of the same system as that of FIGs. 1 and 2 but showing a trailing f-t-e profile that has been created in the practice of the invention; FIG. 4 shows write current versus time of a fragment of write current waveform that produces the trailing f-t-e profile of FIG. 3;

FIG. 5 is a block diagram of circuitry by which the increase in write current can be incrementally delayed to produce the trailing f-t-e profile shown in FIG. 3; and

FIG. 6 is a block diagram of circuitry for controlling the switching of the switching elements of FIG. 5.

In FIG. 1, a magnetic recording medium 10, which has a nonmagnetizable backing 11 and a magnetizable layer 12, is driven at a constant velocity v in the direction of an arrow 17 across the gap between the pole tips 14 and 15 of an ordinary ring-type head and at a head-to-medium spacing 18. Before being driven across the head, the magnetizable layer 12 has been erased to be uniformly magnetized in the direction of a second arrow 19. A write current applied to the head forms a magnetizing region 20 within which the magnetic field component extends in the direction of a third arrow 21. The profile of the magnetizing region 20 is substantially circular as viewed in Fig. 1, typical of a longitudinally oriented, uniaxially anisotropic magnetic recording layer. Different magnetic recording materials afford different f-t-e profiles.

NRZ recording systems can record binary data bits by rapidly reversing the polarity of a substantially constant-magnitude write current, thus creating a transient leading f-t-e profile 22 and a persistent trailing f-t-e profile 23 each time the polarity of the write current reverses. The trailing f-t-e profile 23 is identical to each of the previously recorded trailing f-t-e profiles 23a, 23b, 23c, and 23d which separate adjacent oppositely magnetized regions and are similar to the f-t-e profiles of Figs. 1-3 of the above-cited Fayling II publication.

For simplicity, f-t-e profiles are shown as lines even though they are zones of appreciable width, as can be seen in Figs. 4, 5 and 6 of the above-cited Fayling I publication. Ideally, trailing f-t-e profiles are as straight as possible, because this minimizes noise or intersymbol interference in adjacent bit cells upon read-out, thus permitting increased linear bit densities. Typically, trailing f-t-e profiles are less straight than are the idealized, arcuate f-t-e profiles shown in FIG. 1 and are more likely to be shaped like those of FIG. 2.

In most NRZ recording systems, the trailing f-t-e profiles should extend approximately perpendicular to the surface of the magnetizable layer, although in some systems, trailing f-t-e profiles that are uniformly oblique to the surface may be preferred to optimize other characteristics of the systems.

Elements of FIG. 2 that are identical to those of FIG. 1 employ the same reference characters. FIG. 2 shows trailing f-t-e profiles 25, 26, and 27 as they would be recorded with three different (increasing) levels of write current, Iι,I₂and I₃. Portions of the trailing f-t-e profiles 25, 26, and 27 which are approximately straight and perpendicular to the surface 24 of the magnetizable layer are dotted. The trailing f-t-e profile 34 of FIG. 3 was obtained by

(a) writing up to said first level Ii of write current to generate a segment 34a of a trailing f-t-e profile like the dotted portion 25a of the trailing f-t-e profile 25 of FIG. 2;

(b) delaying the increase in write current Ii for a first increment of time t_dύ (c) writing up to said second level I₂ of write current to generate a segment 34b like the dotted portion 26a of the trailing f-t-e profile 26 of FIG. 2;

(d) delaying thejncrease in write current I₂ for a second increment of time t_d2; and (e) writing up to said third level I₃ of write current to generate a segment 34c like dotted portion 27a of the trailing f-t-e profile 27 of FIG. 2. Said first and second increments of time delay tjj and fø are selected so that the three segments 34a, 34b and 34c are in substantial alignment as shown in FIG. 3 to provide the improved, yet somewhat jagged trailing f-t-e profile 34 that is both substantially straight and substantially perpendicular to the surface of the recording medium.

Before showing how to determine the duration of each time delay tji and fø, reference is made to Fig. 2 of the above-cited Fayling III publication which shows f-t-e profiles for five write current levels. When a tangent for each of those five f-t-e profiles is drawn to be perpendicular to the surface of the recording medium (here called the "perpendicular tangent"), the perpendicular tangents for progressively higher write currents have progressively increased spacings from the perpendicular tangent at the lowest current level. The distance s between the adjacent perpendicular tangents can be determined either experimentally or theoretically, e.g., using the Karlqvist equation, preferably modified as disclosed in the above-discussed Szczech publication. Then the duration of time by which the increase in write current is to be delayed between consecutive levels of write current can be determined using the equation = s/v where v is the velocity of the medium.

To determine the incremental time delays that will provide the trailing f-t-e profile 34 of FIG. 3, reference is made to FIG. 2 which shows the distance Si between the perpendicular tangents to portions 25a and 26a of the trailing f-t-e profiles 25 and 26, respectively, and the distance s₂ between the perpendicular tangents to portions 26a and 27a of the trailing f-t-e profiles 26 and 27, respectively. To allow the magnetic recording medium to travel the distance s_} oτs₂ without any increase in write current, the increase in write current must be delayed for a time ι or tj2, respectively.

The trailing fa-e profile 34 of FIG. 3 is somewhat jagged because it involves only two increments of time delay. Larger numbers of incremental time delays and consequently larger numbers of segments could make a trailing f-t-e profile straighter, but it may be commercially impractical to employ a large number of incremental delays. Noise or intersymbol interference can be significantly reduced upon read-out by as few as two incremental delays in the increase of write current during each current reversal. While the trailing f-t-e profile 34 is substantially perpendicular to the surface of the recording medium, it could be made oblique by increasing the time delay tdi by a time increment 2t_c and increasing the time delay fø by a time increment t_e . Another method of making an oblique trailing f-t-e profile is by decreasing the time delay tdi by a time increment 2t_c and decreasing the time delay ₂ by a time increment t_e .

To form the write current waveform 40 of FIG. 4 to produce the trailing f-t-e profile of FIG. 3, the write current is reversed from an initial level -I₃ at time 42 and rapidly decreases in absolute magnitude until it passes zero and the threshold level I_t. Upon reaching said first level Ii of write current at time 43, the increase in write current is incrementally delayed until time 44 (time delay t_dI = time 44 minus time 43). It then rapidly increases to said second level I₂ at time 45. The increase in write current is again incrementally delayed until time 46 (time delay ₂ = time 46 minus time 45) and then rapidly increases to said third level I₃ (I„uχ) at time 47, remaining at that level until there is a negative-going reversal, during which the same incremental time delays tji and /_<# are made while the write current is increasing in absolute magnitude to -I₃. Each of the positive-going portion of the write current waveform 40 of FIG. 4 and the succeeding negative-going portion produces the trailing f-t-e profile 34 of FIG. 3.

FIG. 5 shows circuitry 50 useful for providing current levels in the coil 51 of a recording head to produce the write current waveform 40 of FIG. 4. Three constant-current sources 52, 53, and 54 are respectively connected to a first pair (55 and 56), a second pair (57 and 58), and a third pair (59 and 60) of switching elements. When only switching elements 55 and 62 are closed, positive write current at a first level Ii flows through the coil 51. When switching element 57 is closed while Ii is flowing through the coil, current from the constant current source 53 increases the write current through the coil to the second level I₂, and when switching element 59 is closed while I₂ is flowing through the coil, current from the constant current source 54 increases the write current through the coil to the third level I₃.

Negative write current through the coil 51 can likewise be stepped up by closing only switching elements 56 and 61 to attain -Ii and sequentially also closing switching elements 57 and 59 to attain -I₂ and -I₃, respectively.

FIG. 6 shows circuitry 64 that produces electrical timing pulses to control switching of the switching elements 55-62. Data pulses are provided to a timing generator 65 through an input port 66. Each data pulse activates a high-frequency clock in the timing generator, which clock is programmed to initiate a series of signals at predetermined intervals that either pass through output ports 67, 69, and 71, first closing switching elements 55 and 62 and then sequentially 57 and 59, or they pass through output ports 68, 70, and 72 to close switching elements 56 and 61, then 58, and finally 60. Although the drawing shows a magnetic recording medium having a single magnetizable layer, also useful are magnetic recording media that have two or more coextensive, contiguous magnetizable sub-layers of differing coercivities. The magnetic recording medium can be a tape or a disk.

Claims

What is claimed is:

1. Method of recording binary data onto a magnetic recording medium that moves at constant velocity v relative to a recording head, which method comprises the steps of a) deteirnining the trailing f-t-e profiles for at least two different increasing levels in absolute magnitude of write current, Ii and I₂ within the range of

b) drawing a tangent to each f-t-e profile to be perpendicular to the surface of the recording medium; c) measuring the distance si between the tangents for Ii and 1₂ ; d) deteπriining a time /_» for incrementally delaying the increase in write current above Ii t_d = s/v where v is the velocity of the medium and S is the distance between adjacent tangents; e) generating NRZ write current, each current reversal of which has a waveform edge that is steep while approaching zero and passing through zero (lo); and f) incrementally delaying the increase in absolute magnitude of the write current above Ii for a time tai .

2. Method as defined in claim 1, step a) of which involves determining said trailing f-t-e profiles experimentally.

3. Method as defined in claim 1, step a) of which involves determining the trailing f-t-e profiles of more than two levels of the write current within the range of I_t and lmax-

4. Method as defined in claim 1, in step d) of which the write current remains constant during each time delay t .

5. Method as defined in claim 1, step b) of which involves allowing the write current to increase at a reduced rate during each time delay td .

6. Method of recording binary data onto a magnetic recording medium that moves at constant velocity v relative to a recording head, which method comprises the steps of a) generating NRZ write current, each current reversal of which has a waveform edge that is steep while approaching and passing through zero

(Io) and b) incrementally delaying the increase in absolute magnitude of the write current at least twice within the range of L and Inux where I_t is the threshold level at which the write current begins to have an effect upon magnetic flux in the recording medium, thus affording substantially straight trailing f-t-e profiles that separate oppositely magnetized zones recorded on the magnetic recording medium.

7. Method as defined in claim 6 wherein prior to steps a) and b) are the steps of 1) deteirnining the trailing f-t-e profiles for at least three different increasing levels in absolute magnitude of write current, Ii, I₂and I₃ within the range of It and Ima ; 2) drawing a tangent to each f-t-e profile to be perpendicular to the surface of the recording medium; and 3) measuring the distances between each pair of adjacent tangents; and in step b), incrementally delaying each increase in write current above Ii for a time tdi and above I₂ for a time delay f_c according to the equation t_d = s/v where v is the velocity of the medium and S is the distance between adjacent tangents.

8. System for recording binary data onto a magnetic recording medium that moves at constant velocity v relative to a recording head, which system comprises a) means for generating NRZ write current, each current reversal of which has a waveform edge that is steep while approaching zero and passing through zero (lo) and b) means for incrementally delaying the increase in absolute magnitude of write αrrrent at least twice within the range of I_t and where is the threshold level at which the write cuπent begins to have an effect upon magnetic flux in the recording medium, thus affording jagged, but substantially straight, trailing f-t-e profiles that separate adjacent oppositely magnetized zones recorded on the magnetic recording medium.

9. System as defined in claim 8 wherein said means for incrementally delaying the increase in write current further comprises means for incrementally delaying the increase in write cuπent at least three times.

10. System as defined in claim 8 wherein said means for incrementally delaying the increase in write current further comprises means for holding the write current constant during each time delay.