Bellite.-Ammonium nitrate 85 per cent., dinitro-benzene 15 per cent. Faversham Powder No. 2.-Nitrate of ammonium 90 per cent., tri-nitro-toluene 9 per cent., moisture 1 per cent. Roburite.-Ammonium nitrate 86 per cent., chloro-di-nitro-benzene (CH,Cl( NO2)2) 14 per cent. This is di-nitro-benzene, with one atom of chlorine (CI), replacing one atom of hydrogen. Ammonite.-Ammonium nitrate 88 per cent., di-nitro-napthalene 12 per cent. A highly important compound also is ammonium nitrate (NH,NO,). This compound, itself a mild explosive when detonated, is very largely used mixed with other more powerful explosives-firstly, as a reservoir of available oxygen to ensure that the actively poisonous gas carbon monoxide (CO) shall be converted, as far as possible, to the less objectionable carbon dioxide (CO2), which, while it does not support life, is not actively poisonous if inhaled; secondly, to cool down the gases of explosion (as its decomposition absorbs heat), and thus produce a 'permitted explosive,' which is one which passes Of the above, dynamite, blasting-gelatine, and the British Home Office test, prior to use in fiery gelignite are not permitted' explosives; the remines, of not causing, by its detonation, an exmainder are. plosive mixture of coal-gas and air to explode. Mention has been made of picric acid as being used Ammonium nitrate is a constituent of some 70 per by itself for explosive purposes. This is specially cent. of these permitted explosives. As it is highly the case for the filling of shell, for which purpose deliquescent, the mixtures in which it is used must its property of becoming thoroughly liquid at be contained in waterproofed cartridges. A mix-251° F. (122° C.) renders it very suitable, as in that ture of tri-nitro-toluene with ammonium nitrate in condition it can be poured into the interior of proportions varying from 60 to 20 per cent. of the the shell. For shell-filling tri-nitro-toluene is also former to 40 to 80 per cent. of the latter, is now well suited, as it melts at the lower temperature of the high explosive largely used for shell and bomb 177° F. (81° C.), and has none of the objectionable filling in the British service. The mixture is termed chemical activity of picric acid leading to the forma'Amatol.' Usually the mixture with 20 per cent. tion of dangerous picrates. On the other hand, it of tri-nitro-toluene is used, and is described as is slightly less powerful, and needs a more powerful *Amatol 80/20.' When closed up in a shell its detonator to start detonation. hygroscopicity is unobjectionable. With 50 to 60 per cent. of tri-nitro- toluene, and heated up above the melting-point of the latter, and kept well mixed, the mixture can be poured into shell, as can be done with tri-nitro-toluene alone (see below); but with less than about 50 per cent. of tri-nitro-toluene the mixture is put into the shell as a powder, and consolidated by suitable means to a density of about 1:45. The mixture of the two ingredients can be effected by milling cold under rolls, or by mixing with stirrers at a temperature of about 100° C. For mines ammonium perchlorate (NH,CIO), mixed with a combustible such as wood-meal, was used largely in the war of 1914-18. This is a powerful high explosive. Explosives with potassium and ammonium perchlorate bases, mixed with nitrocompounds, or with a combustible only, are used to a considerable extent on the Continent. Nitro-glycerine, gun-cotton, and the other explosives mentioned, except Picric Acid (q.v.), which is usually employed by itself, may be, and are, mixed together to form special explosives, but the following substances are often used in the mixtures, viz barium, potassium, and sodium nitrates, as absorbents of nitro-glycerine, instead of the absolutely inert kieselguhr (see below); and when this is done, charcoal, wood-pulp, or wood-meal is added for combustion by the oxygen they contain. See GUNPOWDER. The composition of some characteristic specially named explosives will now be given. Dynamite.-Nitro-glycerine 75 per cent., kieselguhr 25 per cent. Kieselguhr is a silicious earth formed chiefly of the minute silicious cells originally containing Diatoms (q.v.), a group of plants found in water, especially in cold climates. The cells, being porous, absorb the nitro-glycerine. Although the parent of all high explosives, this form of dynamite has been superseded by the more powerful forms mentioned below. Blasting Gelatine.-Nitro-glycerine 93 per cent., nitro-cotton 7 per cent. This is probably the most powerful explosive in existence. The nitro-glycerine is absorbed by the soluble gun-cotton, and forms with it a substance somewhat like india-rubber. Gelignite.-Nitro-glycerine 60 per cent., nitrocotton 4 per cent., potassium nitrate 27 per cent., wood pulp 9 per cent. Monobel.-Nitro-glycerine 10 per cent., ammonium nitrate 80 per cent., charcoal 9 per cent., moisture 1 per cent. Compared with blasting-gunpowder, the cost of high explosives is much greater, their prices varying from about two to three times more than that of blasting gunpowder; but on the other hand, according to their compositions, they are from two to three times as powerful (see GUN-COTTON). See Treatise on Service Explosives (Official, 1907); Manufacture of Explosives: Twenty Years' Progress, by Guttman (London: Whittaker & Co., 1909); Nitro Explosives, by Sandford (London: Crosby, Lockwood, & Son, 1906); Explosives, by Marshall (London: J. & A. Churchill, 1917). Dynamo-electric Machines are machines for generating electric currents by means of the relative movement of conductors and magnets. Faraday discovered in 1831 that an electric current is induced in a conductor when it is moved across the pole of a magnet, so that it cuts the lines of magnetic force, or (more generally) whenever the number of these lines which passes through the circuit of the conductor is in any way varied. If, for example, a coil of wire, the ends of which are connected so that the whole forms a closed circuit, be suddenly withdrawn from the pole of a magnet, a transient electric current is induced in it, while the lines of magnetic force which proceed from the pole are ceasing to be present within the coil. If the coil be replaced, a current will again be induced, but in the contrary direction. Similarly, a transient current is induced if the coil be held at rest while the magnet is drawn away; or, again, if the coil be turned round so that the direction of the lines of force through it becomes reversed, in which case the effect will be twice as great as before. Any movement which causes an alteration to take place in the amount of magnetic induction through the coil produces a transient current, the electromotive force of which is proportional to the rate at which this alteration takes place. The whole amount of electricity produced is the same whether the movement be fast or slow. When the movement is slow, the current lasts longer in proportion as its strength is less. produce the movement requires an exertion of mechanical work, which finds its equivalent in the energy of the induced current. To Faraday's discovery was immediately followed by the invention of numerous forms of magnetoelectric machines, as they were then called, in most of which a steel horseshoe magnet was made to rotate over a pair of coils wound on a fixed armature, or the armature and coils were made to rotate while the magnet was held fixed. Fig. 1 is an example of one of these early forms, in which the armature, BB, with the bobbins, C, D, which consist of coils wound upon iron cores fixed to the armature, revolves in front of the magnet poles, N, S. In every half-revolution the lines of magnetic force through the bobbins have their direction reversed, and a series of transient currents are consequently produced in the coils. These pass to the external part of the circuit through the spring brushes, H, K, which make contact with a revolving collector, consisting of insulated metallic rings on the axle, to which the ends, m, n, of the coils are attached. If m were always in contact with H, and n with K, it is obvious that each successive transient current would take the direction opposite to its predecessor -the direction of the current would alternate at every half-revolution. On the other hand, it is easy, by splitting the rings, to arrange the collector so that H is in contact with m for half a revolution, and then with n for the other half, while K is in con. tact first with n, and then with m, with the effect that the successive currents all have the same direction in the external portion of the circuit. The collector is then called a commutator. A common form of commutator is shown in fig. 2. K H Fig. 1. An ideally simple form of dynamo Fig. 2. is represented diagrammatically in fig. 3, which represents a conductor consisting of a single loop of wire revolving in the magnetic field between the poles of a magnet, NS, so that at every half-revolution the lines of force have their direction of passing through the loop reversed, and a series of transient currents is consequently induced in the loop. Here, again, a commutator is S Fig. 3. required if the currents are to have one continuous direction in the external portion of the circuit. In the position sketched (by full lines), the side, a, of the rectangular loop is cutting the lines of force in one direction, and the side, b, is cutting them in the other, and both these movements are contributing to produce electromotive force in one direction round the loop; the other two sides (i.e. the front and the back) of the loop do not cut lines of force, and therefore do not contribute to the production of electromotive force. As the loop approaches the vertical position (shown by dotted lines), the component motion of the sides across the lines of magnetic force becomes reduced, and the electromotive force diminishes, till, at the vertical position, it disappears entirely, for there the sides of the loop are moving (at the instant) along the lines of force. After that they begin to cut the lines of force again, but in the reverse direction, and an electromotive force opposite to the last begins to act, which reaches its maximum when the coil is again horizontal. The same variations are repeated as the coil turns through the remaining half of its revolution. The strength of the current follows similar fluctuations, being determined by the electromotive force and by the resistance of the circuit, including the resistance of the revolving loop itself. The effect of the revolving conductor in producing electromotive force may be increased (1) by increasing the speed of rotation; (2) by forming the loop with more than one turn of wire so as to make a coil, the whole effect is then the sum of the effects due to the individual turns; (3) by strengthening the magnetic field. One very important method of doing this is to furnish the revolving coil with an iron core, the effect of which is to increase the magnetic induction through the loop, across the space from pole to pole, by providing an easier path for the lines of magnetic force to cross this gap. In early dynamos the armature (as the revolving-piece is called) frequently consisted of a coil of many turns wound on an iron core, in the manner illustrated by fig. 4, which shows in section the simple shuttlewound armature introduced by Siemens in 1856. The ends of the coil were brought to a commutator like that of fig. 2, and the effect was to produce currents which were uniform in direction. They were, however, very far from uniform in strength, varying from zero to a maximum twice in every revolution of the shaft. S Fig. 4. N In the early dynamos permanent steel magnets were used to produce the field in which the armature moved, but it was soon recognised that electromagnets might be employed instead, and in 1863 Mr Wilde introduced a machine with large electromagnets, which were excited by a small auxiliary armature revolving between the poles of a per manent magnet. Before this it had been proposed in machines with permanent magnets to supplement the magnetism when the machine was in action, by having coils wound upon the magnets, and by allowing the current produced in the machine itself to pass through these coils. It was not till 1867, however, that it became known that steel magnets were wholly unnecessary, and that dynamos with electro-magnets might be made entirely self-exciting. Even when the cores of the electro-magnets are of soft iron, there is enough residual magnetism to initiate a feeble current; this develops more magnetism, which in its turn develops more current, and so the process goes on until full magnetisation is reached. The principle of self-excitation was enunciated independently, and almost simultaneously, by Wheatstone, Werner Siemens, and S. A. Varley; it is now made use of in all except the smallest machines. The term 'dynamo-electric' was at first applied to distinguish those machines which were self-exciting from magneto-electric' machines, which had permanent magnets to give the field; but this distinction is no longer maintained, and the name 'dynamo' is now used in the wider sense defined above. An extremely important step in the development of the dynamo was taken in 1870 by Gramme, who introduced a form of armature which, for the first time, gave a current not merely continuous in direction, but also sensibly uniform in strength. The Gramme ring armature is shown diagrammatically in fig. 5. It consists of a ring-shaped Fig. 5. and ture current is increased while the field magnets remain of constant, or nearly constant, strength. As it often happens, especially in the case of a dynamo supplying power for traction purposes, that the demand for current is liable to severe fluctuations, it is obviously a great advantage to have a machine which will work sparklessly at all loads without the necessity of altering the adjustment of the brushes. Various expedients have been introduced with this end in view, chief among which are (1) carbon brushes, the high resistance of which checks the rush of current in the shortcircuited coils; (2) small auxiliary poles, called interpoles,' midway between the main poles, excited by series coils, which supply a reversing field equivalent to giving lead to the brushes appropriate to every load; (3) compensating series windings, arranged to neutralise the distorting effect of the armature current. A small practical Gramme dynamo of an early form is shown in fig. 6. In this example two field iron core, revolving in the magnetic field, having a series of coils, A, B, C, &c., wound upon it. These are joined to one another in a continuous series, and also to the insulated segments 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, and, 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, are also 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, magnets conspire to produce a north pole at N, the two sides acting like two groups of battery cells and two others to produce a south pole at S. The arranged in parallel and of equal resistance and commutator is a series of copper bars mounted on equal electromotive force. The whole electromotive force in the armature is the same as that produced an insulating hub fixed to the shaft, and separated by the coils on one side alone, but the internal from one another by thin strips of mica or other resistance is halved by the division of the current be-insulating material; these bars have radial protween the two sides. In actual Gramme armatures, successive armature coils. Each brush consists of jections, which are soldered to the junctions of the number of coils on the ring is very much greater a flat bundle of copper wires pressed lightly against than the number shown in the sketch, and each As brush is made wide enough where it presses on the neutral points, as they are called, are not exactly 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 subdivided, by being made up 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 are all joined in series, and their junctions are connected to the bars of a commutator just 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 they would have in the ring. Their action depends on their angular motion, and is therefore the same in both cases. As the drum revolves, that coil which is passing the neutral plane (viz. the plane perpendicular to the lines of force) is for the moment inoperative, and the brushes are set to touch those bars of the commutator 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 modern continuous current dynamos have drum armatures of the kind shown in fig. 8, mainly on account of the convenient mutator segments are of hard-drawn copper, and are insulated from each other by prepared mica of the same hardness as the copper, so that the surface under the brushes wears down evenly. The segments are held in place by V clamping-rings insulated with mica. The bolts for drawing these together can be seen spaced at intervals round the shaft. An important element in the classification of dynamos is the manner in which magnetism is Fig. 9. induced in the field-magnets. 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, and consequently the whole current produced by the machine passed through both. This arrangement is distinguished as series winding, and is shown diagrammatically in fig. 10. It was first pointed out by Wheatstone, in 1867, that the magnet coils, instead of being put in series with the external conductor, might be arranged as a Fig. 8. mechanical construction which can then be adopted. The core consists of a large number of discs of annealed sheet-steel with slots punched out at regular intervals round the periphery. At intervals of every two or three inches, air-ducts some half-inch wide are left between the plates through which the rotation causes air to circulate. These ducts are useless so far as the transmission of magnetic lines is concerned, but are invaluable for keeping the armature cool. The winding consists of a number of precisely similar coils of insulated copper wire, braid, or strip wound on a former, and then taped and varnished. Some idea of the peculiar shape of these coils will be gathered from fig. 9. They are inserted two-deep in each slot in such a way that each coil has one straight limb at the bottom of a slot, and the other straight limb at the top of a slot a certain number of slots in advance of the first slot. The coils are kept in place by wire bound round the periphery, or by wooden wedges driven into grooves at the tops of the slots, and the free ends of the coils are soldered to lugs attached to each commutator segment. The com Fig. 12. the external circuit. Finally, in compound winding (fig. 12) the two previous methods are combined. The field-magnets are wound with two coils; one of these (which is short and thick) is connected in series with the external circuit, and the other (which is long and fine) is connected as a shunt to it. This plan appears to have been first used by Varley in 1876, and afterwards by Brush, who pointed out that it, along 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 increase, 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 itself) will in its early portion rise fast as the current increases, in consequence of the rapid augmentation of the magnetic field. Such a curve is called the charac. teristic curve of the machine, and is shown at AA in fig. 13. If we continue to increase the current DIFFERENO OF POTENTIAL BETWEEN TERMINALS COMPOUND SHUNT SERIES CURRENT Fig. 13. by further reducing 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 between the terminals accordingly passes a maximum, and becomes considerably reduced when the current is much augmented, as appears in fig. 13. The characteristic curve for a shuntwound 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, and partly because the current in the armature coils tends to oppose the magnetisation. 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 compoundwound 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 larger supplementary effect on the magnets, and by choosing a proper number of series windings, their effect may be made to neutralise with great exactness the droop in the characteristic curve which would occur if the shunt coil were the only source of magnetism. Compound machines wound for constant potential give a nearly straight horizontal line for their characteristic; CC in fig. 13 is an actual example. By making the series coil more influential, so that the potential at the terminals rises slightly as the current increases, the machine may be compound-wound to give constant potential at the ends of long leading-wires by which the current is conducted to a distance. Series-wound dynamos were at one time largely employed for electric lighting by arc lamps. Compound-wound machines are especially suitable for incandescent lighting, where the lamps are connected in parallel, and where it is important that the potential shall not vary when more or fewer lamps are in action. Shunt-wound machines are also largely used for incandescent lighting, the potential being adjusted to a constant value by varying the speed of the machine, or by throwing resistance into or out of the magnet shunt circuit. Shunt machines are the most suitable for charging storage batteries and for electro-plating, because of their not being liable to have their polarity reversed by a back current from the battery or bath. Fig. 14 illustrates the front and rear of a Westinghouse generator (with shaft removed) for coupling direct to a reciprocating engine, which gives a good idea of the modern way of constructing such machines. The rear view shows that the magnetic circuit consists of eight poles projecting inwards from a circular yoke. The series winding, consisting of a few turns of flat copper strip wound on edge, is seen enveloping the poles next to the yoke. The large rectangular coils at the tips of the poles, consisting of a large number of turns of comparatively fine wire, constitute the shunt winding. Turning to the front view, we notice a ring carrying the eight brush-arms, which may be |