2. Introduction

Although it has already been known for more than 100 years, there has been an increasing interest in the technical exploitation of the magnetoresistance effect. Thin-film technology has advanced so far that it is now worthwhile to investigate whether the effect could be used for measuring magnetic fields, in particular for information read-out. Here we should not only think of information recorded on magnetic tape - although this is the principal subject of this article - but also information stored in a magnetic-bubble memory.

Fig. 1.
Principle of operation of a reading head based on the magnetoresistance effect. Above the tape T, with track Tr, there is a strip of a ferromagnetic alloy, whose resistance is continuously measured with the aid of a measuring current I. Variations in the component Hy of the magnetic field from the travelling tape change the direction of the magnetization of the strip, and this produces a change in its resistance. A magnetoresistive head therefore records Hy as a function of time, whereas a conventional reading head records the time derivitive of H.

The basic concept can be more clearly explained with the aid of fig. 1. Information is stored on a magnetic tape in the form of a magnetization that varies both in direction and in magnitude as a function of position on the tape. This produces an external magnetic field, whose strength varies with position. If the tape now moves in relation to the reading head, which is a strip of a ferromagnetic alloy mounted vertically, the variation of the magnetic field component Hy (perpendicular to the plane of the tape) causes the direction of the magnetization in the reading head to vary. The moving tape therefore, produces indirectly a variation in the electrical resistivity, which can be readily observed.

It will also be clear that the magnetoresistance effect can also be used for the measurement of constant magnetic fields.

Since it has to be extremely thin, the strip is mounted on a substrate - we shall return to this point later.

Because a high resistance to wear is required, both sides of the device are coated with a layer of wear resistant material; see fig. 2.

Instead of a vertically mounted strip it would also be possible to use a strip mounted parallel to the tape, but although this is an interesting arrangement the problem of wear is almost insurmountable. It cannot be solved by allowing the tape to, run along the substrate because the distance between the tape and the strip then becomes too great.

The most important advantage of reading out magnetic information by means of variations in magnetoresistance lies in the possibility of making the reading head very small without reducing the sensitivity to an unacceptably low value. This is of interest if a magnetic tape with more than one track has to be read out, or if the presence of magnetic bubbles at a particular location in magnetic-bubble material has to be determined by means of a reading head.

The performance of a magnetoresistive head (MRH) optimized for the particular application can certainly compete, in terms of sensitivity, noise, crosstalk, and whether it can be included in a tape recorder, with the conventional reading head, which is based on the observation of the change of the magnetic flux in a high-permeability ferromagnetic yoke. With the conventional inductive method the reading and writing functions can often be combined in one head, which is of course an advantage. On the other hand it is not so easy to miniaturize the inductive reading head as it is a strip like the one in fig. 2, and this must be taken into account when a multiple head has to be designed. For example, a decrease in the number of turns that measure the variation of the flux density in the yoke causes a proportional reduction in sensitivity. This demonstrates an important difference between the two methods of read-out. In fig. 1 the y-component of the field of the magnetic tape is measured directly; however, with an inductive head, it is not the field (or the associated flux density) that is measured, but the rate at which the field varies. It is self-evident that at low frequencies (low tape speeds) the inductive method only produces a weak signal.

In the following section we first analyse the response of a magnetoresistive head to a variation in Hy. Then we consider the problem of the linearization of the response function, with emphasis on a method that has been discovered and investigated at our laboratories. In the remaining sections of the article we look at the frequency characteristic, the noise behaviour and the technology of our experimental MRHs, and also the possibility of using the head for certain special applications.

Fig. 2.
In a reading head based on the magnetoresistance effect the ferromagnetic strip and its supply leads are mounted on a substrate Sub. In practice the thickness of the strip is 0.02 to 0.3m, its width is of the order of 10m and its length is equal to the width of the track (between 20m and a few mm). The substrate and strip are placed between two plates of wear-resistant material Q. The surface of the head next to the tape is made slightly rounded.