diagrammatically in fig. 5, consists of a ring-shaped iron Fig 5. N core, revolving in the magnetic field, and having a series of coils, A, B, C, etc., wound upon it. These are joined to one another in a continuous series, also to the insulated seg. ments of a commutator, a, b, c, which revolves with the ring, and from which the current is taken by brushes, H, K. Consider now the action of the field in producing electromotive force in any one of the coils, such as A. Near the place in which it is sketched, the coil A is moving in a direction parallel, or nearly parallel, to the lines of force, therefore is having little or no electromotive force induced in it. But by the time the ring has made half a revolution, the same coil will have the lines of force within it reversed. Between these two positions, therefore, there must have been a generation of electromotive force, and this will in fact be going on most actively half-way between the two places. The coil C is at present the most active contributor of electromotive force, but B and D, the coils lying in front of and behind it, also are contributing a share; and the whole electromotive force between A and E, so far as that side of the ring is concerned, will be the sum of the several effects due to all the coils from A to E. A little consideration will show that the same action is going on on the other side of the ring, so that if the brushes be applied at a and e they will take off to the external portion of the circuit a current, half of which is contributed by one side, and half by the other side of the ring, the two sides acting like two groups of battery cells arranged in parallel and of equal resistance and equal electromotive force. The whole electromotive force in the armature is the same as that produced by the coils on one side alone, but the internal resistance is halved by the division of the current between the two sides. In actual Gramme armatures, the number of coils on the ring is very much greater than the number shown in the sketch, and each brush is made wide enough where it presses on the commutator to touch two of the segments at once. Hence the current is never interrupted, and the fluctuations in its strength, which occur as one segment passes out of contact and another comes in, may be made almost indefinitely small. As each coil passes, it is for the instant short-circuited through the brush, and this would give rise to a waste of energy in the coil and to sparking at the brushes, were it not that the brushes are set to bear on the commutator at the points where the development of electromotive force in the corresponding pair of coils is a minimum. These neutral points, as they are called, are not exactly midway between S and N, but are in advance of that posi tion in consequence of the magnetic field within the ring being distorted through the action of the currents in the armature coils. Hence the brushes require to have what is called 'lead,' and this lead has in general to be adjusted whenever the output of the machine is considerably varied, more lead being needed if it happen that the armature current is increased while the field magnets remain of constant, or nearly constant, strength. The adjustment of the brushes is of much practical importance in management of a dynamo; for the sparking to which faulty adjustment gives rise speedily wears away the commutator bars as well as the brushes themselves. A small practical Gramme dynamo of an early form is shown in fig. 6. In this example two field-magnets conspire to produce a n. pole at N, and other two to produce a s. pole at S. The commutator is a series of copper bars mounted on an insulating hub fixed to the shaft, and separated from one another by thin stripes of mica or other insulating material; these bars have radial projections, soldered to the junctions of successive armature coils. Each brush consists of a flat bundle of copper wires pressed lightly against the commutator by a spring. The core of the armature is a ring made of many turns of soft iron wire, on which insulated copper wire is wound to form the coils. It is essential that the core of the armature should not be solid, for in that case currents would be developed in the substance of the moving iron itself to such an extent as very seriously to impair the efficiency of the machine. Hence the core of dynamo armatures is always sub-divided, by being made either of wire, or more usually of thin plates more or less carefully insulated from one another. Fig. 7 shows the armature of a small Gramme dynamo, removed from its place between the pole-pieces. Two years after the introduction of the ring armature by Gramme, it was shown by Von Hefner-Alteneck that the Siemens armature (fig. 4) might be modified so that it also should give continuous currents of practically constant strength. In the original Siemens armature there was but one coil, all wound parallel to one plane, and the current fluctuated from nothing to a maximum in every half-revolution. In the modified form the coil is divided into many parts, which are wound over the same core, but in a series of different planes, the plane of each successive coil being a little inclined to the plane of the coil before it. The coils all are joined in series, and their junctions are connected to the bars of a commutator as in the Gramme ring. The Siemens-Alteneck or drum armature may, in fact, be compared to a Gramme armature, in which the coils, instead of being wound on successive portions of a ring, are all wound on one piece of core, preserving, however, the angular position that they would have in the ring. Their action depends on their angular motion, and is therefore the same in both cases. Fig. 7. As the drum revolves, that coil which is passing the neutral plane (viz., the plane perpen dicular to the lines of force) is for the moment inoperative, and the brushes are set to touch those bars of the commu tator that are connected with it. The other coils are more or less operative, the most active contributor of electromotive force being that one which is for the moment perpendicular to the neutral plane. The electrical effects in drum and in ring armatures are the same. Nearly all continuous current dynamos have one or the other; most makers prefer the ring type, mainly from considerations of convenience in construction; but the drum type holds its place in some of the best modern machines. An important element in the classification of dynamos is the manner in which magnetism is induced in the fieldmagnets. These may of course be excited from an independent source of electricity; but when the machine is self-exciting, there are three important alternative methods. In the early machines the coils on the field-magnets were connected in series with the external part of the circuit: Consequently the whole current produced by the machine passed through both. This arrangement is distinguished as series winding, and is shown diagrammatically in fig. 8. It was first pointed out by Wheatstone, 1867, that the magnet coils, instead of being put in series with the external conductor, might be arranged as a shunt to it, thereby form ing an alternative path through which a portion only of the current would pass. In this arrangement, which is called shunt winding (fig. 9), the magnet coils consist of many turns of comparatively fine wire, so that they may not divert an excessive quantity of current from the external circuit. Finally, in compound winding (fig. 10) the two previous methods are combined. The field-magnets are wound with two coils; one of these (short and thick) is connected in series with the external circuit, and the other (long and fine) is connected as a shunt to it. This plan appears to have been used first by Varley, 1876, and afterward by Brush, who pointed out that it, with simple shunt winding, has the advantage of maintaining the magnetic field even when the external circuit is interrupted. It has, however, when properly applied, another and more important merit, as will appear below. In a series-wound dynamo the magnets do not become excited if the external circuit is open, and become only feebly excited when the external resistance is high. Let the external resistance be reduced, while the armature is forced to turn at the same speed. The current will now in Fig. 10. crease, producing a stronger magnetic field; the electromotive force is therefore greater than before. A curve drawn to show the relation between the current and the difference of potential between the terminals of the machine (which is a little short of the full electromotive force, in consequence of the resistance of that part of the circuit which is within the machine it self) will in its early pertion rise fast as the current increases, in consequence of the rapid augmentation of the magnetic field. Such a curve is called the characteristic curve of the machine, and is shown at AA in fig. 11. If we continue to increase the current by further reduc. ing the external resistance, the magnets tend to become saturated, and finally even have their magnetism somewhat weakened on account of the influence of the currents in the armature coils. Further, the loss of potential, through internal resistance, becomes more considerable. The difference of potential be tween the terminals accordingly passes a maximum, and becomes considerably reduced when the current is much augmented, as appears in fig. 11. The characteristic curve for a shunt-wound dynamo is shown at BB in the same figure. Here the strength of the magnetic field is nearly constant, but decreases a little when the machine is giving much current, partly because the current in the shunt circuit is then somewhat reduced, partly because the current in the armature coils tends to oppose the magnetization. Hence the potential falls off as the current increases. This fall will, however, be slight if the resistance of the armature is very low and if the field-magnets are very strong, and under these conditions a shunt-wound dynamo will give a nearly constant difference of potential whether much or little current be taken from it, provided, of course, that the speed remain unchanged. To make the difference of potential more exactly constant, it is necessary that the magnetic field should become stronger when the machine is giving much current, and compound winding achieves this. A compound-wound dynamo may be regarded as a shunt machine in which the action of the shunt winding is supplemented by that of a series coil on the magnets. When the machine is running on open circuit, the shunt coil alone is operative; as the current taken from the machine is increased, the series coil produces a larger and |