verted, and secured to the platform, or, better still, special oil insulators may be used, and form a very effective mode of insulation.

Jacob's Method is illustrated in Fig. 82, where D and D1 represent two metal sheathed drums, one of which, D, is directly connected to earth, and the other, D1, is

E

DI

Free

FIG. 82.

insulated in itself, but connected to earth through the galvanometer G. E is the battery, one pole of which is connected to the extremity of the conductor as shown, and the other to earth. The core is wound over from

D to D1, and, while the fault is on D, all the leakage current from it passes to earth via the sheathing of D, but, as soon as it arrives on D1, the leakage current, or at least part of it, passes through the galvanometer G, and a deflection results. The exact position of the fault can, as in the previous case, be located by passing the galvanometer connection, in the form of a moist rag, over the surface of the core in the immediate vicinity of the fault, it being arranged between the drums for that purpose.

A similar method is depicted in Fig. 83, where D and D1 represent two insulated drums as before, E the battery connected between the conductor under test and earth, and G the galvanometer connected between the drum D1 and earth. R is a moist rag or sponge connected directly to earth. The core is wound over from D to D1, all leakage from the fault on D going to earth viâ R, until the fault passes over, when part of the current passes to earth through G, producing a deflection; the winding is then stopped, and the fault located by sliding R slowly along the surface of the insulation.

Another method is illustrated in Fig. 84, and consists, as usual, in winding the core on the two metal or sheathed drums D and D1, both of which are insulated and connected with a switch S, by means of which they can be put to earth, either independently or in conjunction with one another, as required. T is a tank of water through which the core is passed, its position during working being maintained by passing it under a smooth grooved pulley (a shackle insulator answers the purpose admirably) at the centre of the tank. The galvanometer G and the battery E are connected between the tank and the extremity of the conductor. The mode of procedure is as follows: D and D1 are earthed independently in turn, and the resulting deflection from one or the other is an indication of the drum on which the fault lies. Thus, if D1 be earthed, and a deflection results, it is caused by a leakage current from the fault on D, which, travelling by way of the moist surface of the tape or other external coating, returns to the battery via the tank and galvanometer, its only course, since D is insulated, and vice versa. The choice of drums having been made in this manner, both D and D1 are earthed,

and, the battery and galvanometer being still connected as shown, the core is wound off that drum on which the fault lies until a deflection is obtained on G. This indicates that the fault has just left the drum, and, by cleaning and drying a portion, it can be more exactly located by drawing it slowly through the water until the deflection again occurs. Having been thus located and wound on to the opposite drum, its effects disappear as the leakage current goes to earth, and any further faults

FIG. 84.

behind can be localised in like manner without removing the coil for repairs.

It will be noted in the diagrams illustrating the foregoing methods of fault localisation between drum and drum, that the opposite extremity of the conductor to that connected to the remainder of the apparatus is marked." Free." This means free, not only in the ordinary acceptation of the term, but also electrically free from surface leakage-i.e., the surface of the insulation to the extent of three or four inches from the point where the conductor is bared must be cleaned, dried, and, if need be, waxed, in order to eliminate any possibility of surface leakage, which latter, if it exist, tends to give rise to secondary deflections which may be regarded as due to the proximity of the fault, and, in this manner, destroy the thoroughness of the test.

For a similar reason the insulation of the drums and tanks, where indicated, should be as perfect as possible.

As regards the battery power to be employed in these tests, there is no hard-and-fast rule, as it depends upon circumstances, such as the resistance of the fault, resistance to leakage current of the surface of the insulation, degree of dampness of the latter, etc., etc., but, as a rule, a voltage almost, if not quite as high as that employed for the measurement of insulation resistance will be required. If a primary battery such as the Leclanché type be used for the purpose, the cells may be protected from short circuit consequent on the passage of dead earth faults by the insertion of a high resistance, such as 10,000 ohms, in series with the battery.

The galvanometer used should be fairly sensitive; a Thomson reflecting instrument, duly provided with a shunt box and short circuit key, answers very well, and an instrument of the D'Arsonval type even better in that its movements are dead beat, and unaffected by the movements of the drums, etc., if of iron, in its vicinity.

To pass on to the subject of fault localisation by the application of a more or less powerful generator current. This is a practice sometimes resorted to in cases of emergency, and usually consists in the actual burning out of the fault by the application of a current of sufficient magnitude for a short period. Thus, an earth or partial earth fault on an electric light or power circuit may

often be localised in cases of emergency by passing a current from any convenient generator, such, for instance, as one of the dynamos supplying that particular circuit, through the fault, one pole of the machine being connected to the faulty circuit and the other to the metallic sheathing or armouring of the cable if it be of that description, or, failing this, to an effective earth in the immediate neighbourhood, which, needless to state, should not take the form of a gas pipe. A suitable fuse should be inserted to protect the machine, and prevent undue rush of current on the actual breaking down of the fault, and, above all, the test should not be applied except in cases of emergency, where time is an object, and, even then, only when the faulty section is under the more or less immediate surveillance of a responsible individual competent to check at once any undue combustion of insulation, woodwork, or other inflammable material in the immediate neighbourhood of the fault. The drawbacks to this crude system of fault localisation are several in number, chief amongst which may be mentioned (1) the actual destruction, by the heating effects of the current, of most evidence as to the cause of the fault; (2) the destructive effect of the current upon the immediate surroundings of the cable; and (3) the risk of fire consequent upon the test. The only merit of such a system is its time-saving quality, in that it affords immediate visual evidence of the locality of a fault.

Short-circuits, due to temporary accidental contacts between two cables, may often be localised by this method, the second cable being connected in place of the earth, whilst faults in concentric cables between the inner and outer conductors, or between the outer conductor and earth, may be found in like manner. In the former case, if a partial contact exist between the two conductors, it may often be located by passing a steady current of some ten to twenty ampères, regulated by a suitable resistance, through the fault, the said current being maintained for a definite period, will cause an appreciable rise in the temperature of the cable at the fault owing to the imperfect contact at that point, and this rise will ultimately extend through the whole substance of the cable for a short distance on either side of the fault, which may then be located by passing the hand lightly