§ 5. Siemens's Armature.-We may here mention that the system of core and coil, in which the E.M.F. is induced, is called an armature. In order to get the coil into a powerful field close to the magnets, Siemens invented a very compact form of armature.

In this form the core is a long cylinder of soft iron in which are two deep grooves cut longitudinally on opposite sides. The wire is then wound longitudinally in the grooves. The general appearance of this armature is shown in the figure.

§ 6. The Self-Exciting Principle.-It soon occurred to makers that electro-magnets would give far more powerful fields than could be obtained from permanent steel magnets. (We may here mention that the magnets giving the field in which the armature moves are called field-magnets.)

These electro-magnets could be excited by the whole or part of the current given by the armature of the machine itself.

It is found that the (so-called) soft iron cores of the electromagnets always retain enough magnetism to give a current in the armature when rotated. This current, or part of it, passing round the field-magnet coils, increases the magnetism of the cores, and therefore the strength of the field. In consequence of this the current induced in the armatures becomes stronger. The reciprocal action proceeds until the field-magnets acquire some maximum strength depending on various conditions not specified here.

Mr. Ladd was the first to construct a 'self-exciting' machine of this sort. In his machine there were two armatures; one served to excite the field-magnets, and the other to give the external

current.

§ 7. Continuous Current' Machines. In the machines described above we do not obtain a continuous current. In the coils there is in one half-revolution induced an E.M.F. that begins, rises to a maximum, and then ceases; in the other half-revolution an E.M.F of similar rise and fall, but contrary to the former in direction. This gives in the coils two currents in opposite directions, separated by an instant of zero current; and, if there be a

commutator, this gives in the external circuit two currents for each revolution, these currents being in the same direction, each rising to a maximum and falling again, separated from each other by an instant of zero current.

For many purposes it is desirable to have a current more continuous in nature. This can be effected in more than one way; we will describe very briefly one of these ways, and more at length another.

(a) The first method is to have an armature in which there are more than one pair of opposite coils, these pairs being arranged round the circumference of a circle. Thus we may conceive of a Clark's machine in which there are four, six, eight, or more pairs of coils; each pair is wound with one continuous wire whose extremities terminate in two metallic segments fixed to opposite sides of the axle. All these segments are insulated from one another; and therefore the circuit of each pair of coils is closed only when they come into contact with the two metallic springs or 'brushes' by means of which (as in the simple case of § 4) the circuit is completed through the external circuit. Each pair of coils thus feeds the external circuit in turn; and, by so arranging the springs that they come into contact with one pair of segments before they have quite broken contact with the preceding pair, we can insure that the current will never fall to zero. It is, however, of an undulatory character.

(3) The second method cannot be understood until some description has been given of quite a different form of armature. As this method is of great interest, we shall describe at some length one form of machine that is constructed on this principle; the machine, viz, that is called the Gramme, after the name of its inventor. His armature, however, is not quite original in construction, being similar to an earlier form, invented in 1862 by Prof. Pacinotti.

§ 8. The Gramme.' Construction of Armature.-In fig. i. we give a general view of a hand-model of the Gramme machine. Here we can see the ring-shaped armature turning between the pole-pieces a and b of the steel magnet A; the arrangements for giving a rapid rotation to this armature; and the collecting springs or brushes c and i, in contact with the axle on opposite sides of it, by means of which a current is sent through the external circuit.

a great difference between this kind of armature and that of Clark or Siemens. In the upper part of fig. ii. we see this wire

as it is actually wound; and in the lower part of the figure there are shown a few sections of the coil, to indicate the mode of winding.

This wire is, as we said, continuous. But at regular and close intervals the wire is brought out, is laid bare of insulating material, is soldered to an insulated copper segment m n (a few only of these copper segments are given in the figure), and is then led back to continue the winding. From what we have said it is clear that we have a continuous wire with which we can, by means of the insulated segments m n, &c., make metallic contact at regular and frequent intervals all round the ring. These segments are arranged all round the axle, and it is with them that the brushes of the external circuit make contact. In fig. ii. O represents the solid body of the axle; it is made of hard wood or other insulating material, there being an inner core of metal, insulated from the segments m n, to give necessary strength.

§ 9. The Gramme. The E.M.F.s Induced in the Ring.-Let us now consider the condition of this ring-armature, with its continuous wire, when it is rotated between the poles ab of the magnet; and let us at present suppose that there are no springs or brushes in contact with the segments. Were it not for the core, the magnetic field would consist of lines of force running nearly straight across from one pole piece to the other. These lines would pierce the coils of the armature, the number piercing each coil depending upon the position of that coil. Thus, a coil which is at the top of the ring depicted in § 8, fig. i., or which is situated equatorially, embraces the maximum of lines, supposing the field to be uniform. As the ring turns, this coil embraces fewer and fewer, until, when it lies edgeways to the lines or is situated 'axially,' it embraces none. (If the field be not uniform this statement must be modified; but, in any case, the coil embraces zero lines when situated axially.) As the ring turns still further, our coil begins to embrace lines in the other direction; and these reach a maximum when it is at the bottom of the ring, i.e. when it is again situated equatorially, 180° from its initial position. All this will give an induced E.M.F. in one direction, as has been already explained in Chapter XXII. § 4, and elsewhere. As the coil turns through the other 180° back to its initial position, there will, it is clear, be induced an E. M.F. in the opposite direction.

Thus, as the ring rotates, all the coils that are to the one side of the equatorial diameter give an E. M. F. in the one direction, and all those lying to the other side give an equal E. M. F. in the opposite direction. The result will be that no current will flow, but that the different parts of the wire, and so also the copper segments connected with them, respectively, will be maintained at different potentials. Thus, in the figure here given we shall have

the wire at A and B, i.e. at the extremities of the equatorial diameter, maintained at the greatest AV, the potential falling from (say) A symmetrically along the wire, through each route, down to B.

Thus, though the individual coils assume all positions in turn, the ring, as a whole, maintains a character that is constant as long as there is constant field-strength and velocity of rotation. The revolving ring has been not inaptly compared to a stationary system of two equal batteries (see fig. ii.) set against one another; these batteries giving no current in their circuit, but maintaining the poles A and B at a constant AV. We can evidently obtain a current in an external circuit by connecting A and B.

The core. -We have hardly mentioned the core in what we have said. The fact is that the core only modifies the field; its presence does not affect the general theory of the ring as given above. Its action is mainly to concentrate the field upon the coils. The lines of force cannot now, to any considerable extent,