circuit is over-fused, in the burning out of part of the circuit and possibly the firing of premises in which the breakdown occurs.

It is therefore of the utmost importance to be able to test the insulation resistance of a length of cable, main, or circuit, either when no current is flowing through it or when the supply is in actual progress and the main or circuit " alive," as it is usually

termed.

A number of different methods have been devised and are in general use for measuring the insulation resistance of both "dead" and "live" cables and systems, and in the following pages devoted to this question some of the principal and common ones in use will be considered. Before, however, proceeding with actual methods of measurement it may be profitable to make some general remarks.

Electrical cables and wires are in the first place tested for their insulating qualities by the manufacturer prior to being sent out to the purchaser, but the latter should test them also himself, both before and after laying, to make sure that no faults have developed, and of course periodically during use. Such a mode of procedure is of the utmost importance if efficient working and maintenance is to be obtained, for it is quite possible for a cable to be accidentally damaged during laying and a subsequent fault to develop at this point, due to the strain of working conditions, which will finally break down the cable.

(41) Measurement of the Insulation Resistance of Electric Light Cables by the Direct Deflection Method.

Introduction. The ordinary Post Office form of Wheatstone Bridge will measure resistances up to 1.111 megohms, though even at this maximum limit the measurements are not very accurate, owing to the resistances of the arms of the bridge being so widely different from one another; consequently it is unsuitable for measuring insulation resistance, which almost invariably amounts to much higher values, often of the order of hundreds of megohms. The present method of direct deflections, which is also termed the "simple substitution" method, is the most accurate

in such cases and often the most convenient one to employ. The principle of it consists in comparing the deflection of a galvanometer needle caused by a given E.M.F. through a known standard high resistance in series with the galvanometer, with the deflection produced by the same E.M.F. working through the insulation of the cable to be tested substituted for the standard resistance.

Preparation of Cable for Test. This must be carefully done and is of the utmost importance if the true insulation resistance of the dielectric is to be found, as the difference between the results obtained with properly and improperly prepared ends is very great. The method of doing this should be as follows

(a) For vulcanized india-rubber cables, the braiding, tapes, or other covering should be removed for at least six inches from each end down to the surface of rubber covering, care being taken in doing this not to cut or otherwise injure the rubber covering still left.

(b) Wash this rubber surface with naphtha and scrape with a clean knife to remove any foreign material still left on the surface, and in this way so get a clean, fresh surface.

(c) Taper the rubber with a clean sharp knife for about 1" to 2" from the end, and then carefully dry the whole of the prepared end over a spirit flame without burning the rubber.

(d) Paint or coat the whole of the prepared end with three or four coatings, one after another, of clean paraffin wax, melted to a temperature not exceeding that of boiling water. This can be done by placing the can of wax inside one of boiling water, whereas, if the wax is melted over a flame, it may be allowed to burn, and its insulating properties partially destroyed. As each coating of wax will have set before another can be got on, the whole cable end will be eventually sealed by a considerable thickness of the wax, which being much less hygroscopic than rubber, will not allow moisture to accumulate and so impair the prepared end. In lieu of wax the prepared end may be lapped with pure clean warm rubber tape well stretched, but this is not so good as the wax finishing.

A better and much more expeditious way of eliminating errors due to surface leakage over the ends of a cable which is being tested for insulation resistance, than that just described of carefully tapering the ends and coating them with paraffin wax, is to

employ an ingenious device known as Price's guard-wire. This, when properly applied, gives complete protection from errors due to end leakage in the direct deflection method, where the cable ends are close to the galvanometer, and, consequently, the connecting wire between cable core and galvanometer is air insulated. In the case of other methods, such as the loss of charge test, extra precautions are necessary to avoid errors (vide Phil. Mag. vol. xlix. pp. 343-7, April 1900).

If the cable ends are close to the galvanometer then Price's guard-wire device in its simplest form is shown in Fig. 36 (I). T' is a lead-lined tank of water in which the cable C to be tested, for insulation resistance, is immersed. The ends of C are prepared with a long clean taper (t) from the core W, so as to give a long clean surface of insulation exposed to leakage. A thin copper guard-wire (w) is wound two or three times round the tapered part rather nearer the outer braiding than the core W and connected as shown to the galvanometer G and high voltage battery

B. If now the resistances of the taper surface (t) are large compared with that of G they will all be at the same potential, and we shall have no leakage, but if any leakage exists (w) will tend to keep up the potential of W, the deflection of the galvano

a

meter G being now reduced in the ratio , where a and b a+b

represent the conductivities of G and t respectively. Consequently the correct result will be

a+b

a

x deflection.

It will, however, be evident that in some cases the cable ends cannot be brought up close enough to the galvanometer in order to have an air insulated wire connecting G and W.

The simplest and best way of getting over this difficulty is that suggested by Prof. Ayrton and Mr. T. Mather and represented in Fig. 36 (II). The inner conductor (ii) of a concentric wire (K) is used to connect W and G, the outer (oo) connecting (w) with junction of G and B as before. Now if oo has a high insulation resistance compared with the internal resistance of the testing battery B, complete protection is afforded against surface leakage, even though KK is laying on the ground.

Apparatus.

Sensitive high resistance Thomson astatic flecting galvanometer G with its box of shunts1 S; known standard high resistance R; unknown insulation resistance

1 Or the Ayrton and Mather Universal Shunt-box.

(r) of cable to be tested; well-insulated battery B of either Leclanché or secondary cells capable of giving an E.M.F. of from 100 to 500 volts; two-way highly insulated spring tapping key K ; suitable lead-lined water-tank W. If the "lead" or cable to be tested is small enough, it may be run direct from the key K into the water-tank W (clear of everything) and coiled up under water, the free end being carefully kept dry and left standing upright, about 12′′ out of the water, as indicated at D.

The tank should contain ordinary cold tap-water at a temperature of about 70° F., and the cable to be tested should be allowed to soak in this for 24 hours before the test, with its end trained up in mid-air above the water some 12" or so.

If the cable is too large to be taken up to G, a short well-insulated G.P. covered wire must be tied on to it at C, and the joint insulated as at D. In all insulation tests at least the working pressure which the cable is to be subject to should be used to test it with.

The known standard high resistance may preferably consist of a metal megohm, but in lieu of this costly piece of apparatus, a carbon megohm, checked against a metal 100,000 ohm coil occasionally, will do quite well, and costs only a few pounds. It must, however, be borne in mind that such a resistance slowly alters its value with time and temperature, so that the temperature should be noted each time it is checked by the present method.

Note. To avoid damaging the galvanometer, which is a very delicate one, the shunt-box provided with it must always be used in the way indicated below.

Tests.-(1) With the lever switch or short circuit bar of S down, thus short circuiting the galvanometer terminals, and also with the shunt-plug in the hole, connect S up to G first, and then the rest of the circuit as indicated in Fig. 37, and adjust the spot of light to zero by means of the controlling magnet.

999

1 999

(2) Remove the short circuit in S, and, with the shunt plugged up, gently tap K for a fraction of a second so as to complete circuit through the standard known resistance R; if the deflection is inappreciable, release K to plug up the shunt, and again tap as before, and so on until a convenient steady deflection d is obtained. Note this and the shunt SR in use at the time (if any).

R

H