The insulation of wires

The Insulation of Wires.*

BY THOMAS A. EDISON.

To telegraphy the Insulation of the circuits is so vital a matter that at the risk of making history repeat itself, and of recounting an oft-told tale, a short paper on the subject may be tolerated.

As a class, gases are the best insulators; next, liquids, and solids last, of all. The insulation of gases is so good that no determination appears yet to have been made of any leakage through them. If it were possible to make an experiment at or near the center of the earth, so as to be beyond the reach of gravitational force, an electrified body might lie left for hours in a gas without any visible means of support, and observations then made to determine whether any loss of charge look place through the surrounding gases. Under existing circumstances, the loss of charge which always does take place on electrified bodies, cannot readily be traced beyond that through the suspension or support that holds them. Even in moist air the loss of charge has not yet been brought home to the aqueous vapor. Professor Boys exhibited before the Physical Society of London in April 1889, a pair of electrified gold leaves, suspended in moist air by a short hook of quartz. The loss of charge appeared to be about 25 per cent in live hours. A glass hook, under the same conditions, would have allowed the charge to disappear, it was said, within one minute.

If any perfectly insulating solid could be found, telegraphists would soon discover any surface leakage from their wires through the air, for the surface of a No. 8 B. & S. wire is 224 square feet, or 20.8 square meters per mile. The loss of charge which lakes place convectively into the air over any sharp point in an electrified body, is a phenomenon of a different nature. There the layer of atmosphere over the point is continually being torn by the magnitude of the forces brought locally to bear upon it, and the particles of moving air carry away the charge.

Some liquids have also a very high insulation, notably most mineral oils. Animal and vegetable oils are by no means so good, or, at least, there is greater difficulty in obtaining them in a highly insulating condition.

Water, contrary to prevailing notions, is quite an insulator. In the purest distilled water, resistance as high as seven megohms per cubic centimeter have been recorded; that is to say, a block one centimeter cube of water pressed between two opposite conducting plates as electrodes, would offer a resistance of seven megohms; but the least trace of impurity brings the resistance down. Ice at - 12° C. has been measured at 2,240 megohms per cubic centimeter.

There are two conditions of surface upon which bare wires miles in length, laid on the ground, have been worked telegraphically — one over the dry desert sands of Africa, the other over dry ice in the far north. Sea water has a resistance of about 30 ohms per cubic centimeter at 5° C. It would appear probable that liquids cannot conduct without electrolytic decomposition. A microscope will show that one microampere decomposes the drop of water it traverses. Liquids, have, as a class, the quality of elastic insulation to high tensions, which is an important feature to the electrical engineer. Air for the few centimeters will break and allow a spark discharge to occur when the pressure reaches from 10,000 to 50,000 volts per centimeter (25,000 to 127,000 volts per inch) according to the shape and condition of the electrodes, but rosin oil is said to stand about 75 times the pressure of air per centimeter without disrupting.

The following is a list of actually observed resistance in commercial samples of well-known insulating substances. The results are given in megohms per cubic centimetre at or near 18° C.:

All transparent solids are insulators, but of course the opposite statement is not true, that all opaque solids are good conductors. There is now good evidence for believing that the process by which light is propagated, the mechanism by which it, is transmitted through space, is purely electro magnetic, and the magnetic vibrations passing through a conductor would generate electrical currents and be absorbed in the substance as heat, that is to say, it would be opaque to the light, failing to transmit the energy. The difficulty with solids is not so much to find insulators, for the great majority of solid substances freed from moisture are poor conductors, but to find an insulation of suitable mechanical qualities. Glass, porcelain and mica seem to be almost the only practically available insulators that will support considerable stresses and these for many structural purposes are far weaker than is desired.

In American telegraphy, glass is almost the universal insulator, but in Europe, particularly in the south and west, the atmosphere is so much more humid, and glass so hygroscopic, that no circuit of any length could be operated with glass insulators except in dry weather. Porcelain or vitrified stoneware insulators are used instead, and in quite a variety of forms. Practically speaking, the insulation of a line is never that of the material forming the insulators, but always that of their surfaces, and the films of dust and moisture that may have become encrusted thereon. The most perfect insulators are those which have underneath the hook a cup filled with oil, in such a manner that the current leaking from the wire to the ground has to pass over the oil, or else through the substance of the insulator itself. These insulators are certainly more expensive, and require to be refilled with oil at intervals, but they will defy weather and keep the insulation nearly as high in fog as in sunshine.

When a long leaky wire is opened at the distant end, and tested for insulation, the insulation per mile always appears to be somewhat more than it is for any actual mile, since the more distant portions of the line are tested with a reduced pressure owing to the leakage over the nearer portion. When the wire is grounded at the far end, and its conductor resistance measured, the leakage will on the other hand make the apparent resistance per mile too low. But if the conductor resistance is reduced in a given ratio, say as 100 to 95, the insulation will be over-indicated in the exact inverse ratio of 95 to 100, provided that the insulation of the line is uniform. So that if a wire's conductor resistance at its temperature of observation is known to be 10 ohms per mile, but appears by leakage over the whole length to be 9 ohms, then if the insulation per mile apparently measures 400,000 ohms, it will be really 360,000 and each mile taken separately might be expected to measure 360,000.

As is well known, telegraph lines work better up to a certain point if the insulation is rather low. A wire has to be emptied of its charge between the impulses of the key sufficiently far to keep the relays from sticking. If the insulation is perfect, this quantity has to be cleared through the ends to the ground, but if the line leaks it can escape more readily at all points. The longer the line, the more essential good insulation necessarily becomes. The insulators are perhaps never more carefully tested than those intended for the long circuits on the Asiatic plains. It was at one time customary to test each insulator before it went out, by taste. The insulator was immersed nearly to the rim, head downwards in the water, and a battery of 100 cells was connected with one pole to the water and the other to the insulator stalk through a particular kind of key. The operator first rested his moistened finger on this key and then applied his tongue to it. If he tasted no current, the insulator was passed on as satisfactory. The test was sensitive and expeditious, and saved the care and handling of a galvanometer.

The insulation of a wire is a definite term, and stands for a definite property so long as it is not necessary to measure it very accurately. There is usually no difficulty in finding the insulation of any overhead wire within 5 percent at any one time, and two observers measuring the line from the same end with different instruments would generally agree in their results to that limit; but as higher degrees of accuracy are attempted (a condition fortunately that does not practically occur) the difficulty of obtaining concordant results may increase rapidly. A resistance coil of wire at a uniform temperature can have its resistance measured to within one-fiftieth of one per cent if necessary, but a leakage resistance is essentially liable to variation. The atmospheric conditions may be altering, or there may be polarization, or inductive disturbances from neighboring wires, or a combination of conditions that may set close measurement at defiance. It is generally advisable to employ galvanometers for this purpose, that even if sensitive, are slow of movement and dead beat. The only instances in which accurate measurements of installation are possible and necessary, are in connection with subterranean or sub-aqueous wires.

Jute and paper, dry or saturated with compound, are rapidly coming into use for subterranean wires sheathed in lead, while for long cables under water, india-rubber and gutta-percha are invariably employed. With conductors highly insulated by these methods, the insulation can be measured much more closely, and is definite for a given temperature of the core, but even in this case the current that will flow from a battery into the wire freed at the distant end is not all leakage. A considerable portion may be stored up in the insulating substance, and be returned from the wire to ground after the battery has been removed. This ''polarization'' is particularly noticeable with gutta-percha and india-rubber covered wires, and their insulation may be apparently 50 per cent greater after three minutes of charging than at the end of the first minute. In specifications it is usual to call for a certain insulation per mile at 75 degree C., and after a definite interval of charge. The insulation obtained from a given thickness of covering depends on the diameter of the wire as well as on the quality of the cover, for a large wire supplies a larger leaking surface. Also, if an insulating coating of a certain thickness produces an insulation of 100 megohms per mile, doubling that thickness will not double the insulation because the leakage will take place with greater relative facility through the greater surface of the second coat. The exact increase of insulation will depend upon the diameters of first coating and wire.

When a wire is well and homogeneously insulated, its insulation resistance at a given duration of charge or period of electrification will appear to be the same with different battery powers or voltages applied, but if, on the contrary, it contains any small faults or defects, it will generally show less insulation with increasing testing pressures. This often forms a criterion as to the reliable insulation of a long cable. The insulation in an ocean cable under the pressure of great depth and near the temperature of melting ice that deep oceans approach, might be 15,000 megohms per nautical mile at the fifth minute. With a small incipient fault the insulation might still reach 3,000 megohms per mile and seem excellent, but it is probable that when tested with five cells and with 50 cells, the insulation resistance in the latter case would not appear so high. It would also probably be different with the zinc or copper pole to line.

*Abstract of paper read before the Railway Telegraph Superintendent's Association, Denver, Colo.