LRTA logo Light Rail Transit Association
Light Rail for better public transport

Electrolysis from railway currents

By Arthur Vaughan Abbott, C E. — August 1899

Note: This article is accompanied by a diagram which will not be easy to follow on screen. You may therefore wish to print it out prior to reading the article. Diagram.

SEVERAL years ago much excitement and apprehension were aroused by the announcement that the underground metallic structures in many cities were menaced with destruction from the corrosive action of electric street railway currents. The extensive pipe plants of gas and water works were, it was averred, doomed to immediate disintegration, with the inevitable result that the inhabitants would be drowned, if they were not previously blown up. The lead sheaths of electric cables that, after a stubborn fight, had been torn from unsightly pole lines and thrust into subterranean conduits, were discovered to be not only in particular danger, but, when perforated, would furnish the igniting spark that would precipitate the general holocaust. The illustrated scientific press teemed with pictures of pipes in the most advanced stages of decomposition, and the opponents of the "deadly trolley" were filled with satisfaction at the efficiency of the new casus belli. Indeed, it seemed as if, in spite of its generally acknowledged convenience and utility, "the broomstick train" was doomed, and must, with the witch who hauled it, be relegated to the past. But the inevitable reaction ensued. All the pipe systems did not immediately go to pieces, but, on the contrary, evidence from one town where the gas mains had been severely exposed, showed little injury, so that the railway people took heart of grace, and, emerging from the odium with which the "anti-trolleyites" had covered them, serenely proceeded on their way.

At present comparatively little is heard of electrolysis, but still electric currents from hundreds of power stations steadily creep through the earth, and while they may not inflict the predicted wholesale destruction, they are far from being harmless. Unless carefully controlled, they may, indeed, inflict considerable injury.

When electricity passes from place to place, a portion is absorbed in the bodies which form its path. In solids this dissipated energy is transformed into heat and expended, sometimes disastrously, in raising the temperature of the conductor, while in most liquids it is chiefly occupied in decomposing the fluid into its component chemical elements. Thus, if water form a portion of the circuit, it will be separated into oxygen and hydrogen, the oxygen appearing at the positive pole, or anode, and the hydrogen at the negative pole, or cathode. The elemental gases only make their appearance at the surfaces of the conductors extending into the liquid, and, at the instant of decomposition, seem to be endowed with extraordinary activity, readily corroding all but the most resisting substances. This process of decomposing a fluid conductor, or electrolyte, is termed electrolysis, and. used in the art of electroplating, ante dating a few months the introduction of the telegraph, is claimed to be the first practical application of electricity.

Like fire, electricity is a good servant, but a bad master, and so long as it was usefully employed in coating spoons and forks, electrolysis proved itself a most valuable discovery. With the advent of the electric railway, however, a loophole of escape appeared, and the lightning that had been harnessed to deposit metals, carry messages, and drag cars, relieved from a confining path, wandered into the ground with mischievous intent, for in every trolley road electrical engineers had set up a gigantic electrolytic cell. With its engines and dynamos, the power station furnishes an almost inexhaustible supply of current that, passing through the trolley wire and car motor to the rails as one pole, finds in the moist earth a conducting fluid, and an opposite pole in the sundry metallic structures it may encounter on its return.

In an electrolytic cell corrosive action is confined to the anode. If the arrangment of railways is such as to make the rails the positive pole, disintegration would be confined to this portion of the circuit and cause dissatisfaction only to the railway owners. Such would be the case provided all subterranean metallic structures presented a path of 'less resistance than any other route. They would then be universally electro-negative, and no damage could result; but this is not the condition found in practice. A pipe line may parallel a track and then branch off. The joints between pipe lengths may be of comparatively high resistance, and at all such places electricity will seek the earth, with ensuing corrosion. So, while there cannot be the slightest ground for opposition on the score of possible injury to putting an indefinite amount of electricity into the underground systems, there is the most serious objection to its getting out again, excepting in so orderly a manner as never to make spots which are electro-positive, and the real problem is,- How shall all underground metallic plants be always maintained electro-negative?

Expedients galore have been proposed as solutions. To provide a complete metallic circuit, and, by thus preventing any current from entering the ground, secure complete immunity, is at first the most obvious and most radical proposition, but one which necessarily inflicts considerable hardship on railway interests. Immense sums are invested in existing single-trolley roads using the ground as a return circuit. To remodel these, to operate by some form of insulated circuit, would require an enormous outlay, regardless of mechanical difficulties. Certainly from the railway manager's point of view, the double-trolley is impracticable, evidence of which is the survival of the single-trolley which has displaced the older metallic system, with which all early roads were equipped, and it is urged that, as electrolytic action can be confined to relatively small areas, it is unjust to require the reconstruction of entire railway systems, until it is shown that the danger spots cannot be suitably protected.

The alternating current has been held to provide a perfect relief and to obviate the constructive difficulties of the double-trolley; but this proposition, in avoiding a mechanical Scylla, steers into an electrical Charybdis, for there are practically no alternating current motors suitable for street railway work, nor, from the present outlook, is the advent of such apparatus likely. It seems hardly safe to assert that corrosive action would, by such substitution, be always entirely averted. Ordinarily chemical decomposition does not appear to attend the passage of alternating currents through liquid conductors, but the numerous devices for producing unidirectional currents from alternating circuits by the use of electrolytic cells, arouses a doubt as to the thoroughness of the promised relief.

Railways on the three-wire system, with the earth as the neutral conductor, furnish a solution particularly attractive to the railway manager, for under cover of an endeavour to prevent corrosion, the voltage may be doubled, and, theoretically, the cost of the feeder system reduced to one-fourth, while the earth is relieved of any currents but those which might pass laterally between the rails. But the three-wire system can be applied only to double-track roads, and excessive earth currents can be avoided only when both tracks are close to each other, and when the demands of the cars upon opposite portions exactly balance, - conditions that in practice can never, even approximately, be continuously realised. The hazard of two naked wires, differing by 1000 volts in potential, is incurred, and the currents in the earth as a neutral conductor constantly shift over the whole rail system, following changes in car load, rendering the location of danger areas difficult, if not impossible. Yet in favourably arranged roads, and with intelligent design, this plan may prove valuable.

To obviate the mechanical difficulties of the double-trolley, it has been proposed to erect a complete feeder system of both outgoing and return conductors, properly insulated, and to connect the track at frequent intervals with the return wires, in such a manner that all paths to the station should be of equal resistance. The dynamos would be joined to proper feeder conductors, and no grounds allowed. Then the rails would carry current merely between any adjacent connections to the feeder system, and by making these sufficiently frequent, and by paying attention to good rail-bonding and adequate insulation, earth leakage may be completely avoided. While this plan is efficacious in preventing electrolysis, the cost of the conductor system is quadrupled for the same fall of potential, and there is a constant additional expense for the energy dissipated in the compensating resistance. This it is possible to reduce to any desired medium, but only by again increasing the expenditure for the conductors, which on any but very short roads would soon become prohibitive.

Such have been the principal plans for preventing electrolysis by keeping railway currents away from subterranean metallic structures; but as no harm can accrue from the entrance or passage of electricity, and as corrosion occurs only when the current leaves the metal for the earth, it seems pertinent to concentrate attention upon such points with a view to providing, if possible, efficient local protection. The first step is such a study of the territory as will enable its electrical condition to be ascertained. This information is best gained by an electrical survey. Where the area to be examined is small and the railway and underground systems simple, the early method of making voltmeter measurements between the water hydrants, lamp-posts and the rails and adjacent ground usually indicates with sufficient accuracy the points in the pipe systems which are positive to earth; but where the territories extensive, and railway and. underground systems are complicated, more comprehensive measurements are advantageous, and a thorough knowledge is best secured by drawing upon a map of the city in question a series of equipotential lines giving electrical contours. By selecting some desirable central point as a datum and taking advantage of the network of conductors that cover all towns, as electric light and power circuits, telegraph, telephone and fire alarm lines, the difference of potential between the datum and a sufficient number of points may be rapidly measured. With the positive poles of station dynamos connected to the trolley wire, each station becomes the centre of a negative area that gradually diminishes in intensity. If the ground were of uniform conductivity, and if the tracks radiated equally in all directions, the equipotential curves would be a series of concentric circles with each station for a centre; but chiefly owing to eccentricities in the track system, the curves are expanded into irregular ellipses, the longer axes of which are essentially paraliel to the rails. Proceeding from each station the potential gradually decreases to zero, and then increases to a positive maximum, so that the electrical condition may be compared to a rolling country in which each station occupies a deep hollow and is surrounded by a series of circular hills.

It is also desirable, with a view to estimating the probable amount of electrolysis, to determine the changes in earth potential that are likely to occur from time to time. This information is easily secured by a series of consecutive observations between convenient points which may be plotted as a curve showing the time-potential relation. Given an equipotential map, it is easy to determine the condition of the earth in any direction, and so ascertain probable danger points. If an imaginary plane be passed through the earth, its intersection with the equipotential curves will give a series of points forming a curve that represents the electrical condition along the plane.

Metallic structures that run parallel to equipotential curves lie in regions of constant potential and are unaffected; but if they intersect different equipotential lines, they will pass through territory of differing potential and the number of curves intersected in a given distance, or the steepness of the electrical gradient, is a measure of the probable tendency for the current to enter or leave the pipe line. Wherever the earth potential is falling or becoming more and more negative, the metallic structures will tend to be more and more electro-positive, for the current will endeavour to flow toward regions of least potential, and, conversely, in areas of rising potential, the pipe lines will tend to be electro-negative.

When railway electrolysis was first discovered, sweeping assertions were made as to the omnivorousness of the current in attacking, without respect, all metallic street structures, and laboratory experiments were cited to show that most minute differences of potential were sufficient to excite action. Practical experience has not fully sustained these early gloomy prophecies; contrarywise, it has been demonstrated that lead is the most sensitive metal, is attacked by the smallest difference of potential, and yields most readily and rapidly to corrosion. Wrought iron comes next, while cast iron, particularly those varieties that contain large amounts of carbon and silicon, known as white cast iron, are so little affected as to be almost exempt. The lead service pipes, therefore, from the water mains, and the sheaths of underground electric cables will be the first and chief sufferers. The wrought iron gas and water services, and the wrought iron pipes of the newer gas companies stand next, while the cast iron mains of the older installations, particularly those pipes made of chill iron, will practically escape injury.

The degree and rapidity of the corrosive action depends upon the quantity of current flowing, and the nature of the soluble salts contained in the surrounding soil. Chlorides or nitrates from the street wash in a clayey or loamy soil favour action, while in clean, dry sand, corrosion is a minimum. If corrosion were uniformly distributed over the entire surface of an exposed pipe, a measurement of the current and an analysis of the soil would enable a fairly accurate prediction as to the probable rate of injury, but as disintegration always proceeds by pitting, the damage is concentrated on particular spots, and the metal is perforated with comparative rapidity. As lead is most rapidly attacked, it is easy to test the probable maximum rate of action at any place by, putting weighed test pieces of sheet lead in the street manholes and reweighing them after a lapse of time.

The greatest economy in a local protecting system is attained by making it as small as possible, and the system is a minimum under the following three conditions:-

  1. The danger areas must be reduced to their least dimension.
  2. They must be brought as close as possible to the power station.
  3. The current escaping from electro-positive points must be reduced to a minimum.
Danger areas from electrolysis extended to maximum limits by the old plan of connecting positive poles of station generator to the rails
Figure 1

It was formerly the practice to connect the positive pole of the station dynamo to earth, and the negative pole to trolley wire, as shown in Figure 1 with the result that the portion of the underground systems remote from the power house, covering a widespread territory, were rendered electro-positive, and the danger areas extended to the maximum limits, so the first step towards protection was taken by reversing the generator poles, rendering all the territory remote from the generators electro-negative and concentrating the electro-positive points in the immediate vicinity of the station.

Generator poles reversed in accordance with later practice bringing danger areas near the power station
Figure 2

By this simple expedient, illustrated in Figure 2, the first two of the three conditions have been fulfilled with the most gratifying results. Compared to a metallic conductor, the earth, even in a most favourable condition of moisture, offers a considerable resistance, and it is only with great reluctance that currents of magnitude will leave even a naked metal path for the ground. Now in the metal of the track the railways have a conductor of magnificent proportions, far exceeding in electrical conductivity any feeder system in use, requiring only adequate electrical continuity at the joints to form a path that will permit only a minute fraction of the current to leak through the earth, thus tending to fulfil the third condition.

From the specific gravity and conductivity of steel it is easy to calculate the resistance of rails in terms of the weight per linear yard, and in the diagram an attempt has been made graphically to set forth all information pertaining to the use of the rails as a return path. Curve No. 1 gives the resistance per foot of two rails in parallel, while Nos. 2 and 3 are the resistances per mile of two and four rails in parallel (single and double track). The left-hand scale is devoted to rail weight in pounds per yard, while on the lower horizontal margin are two scales. The upper one applies to Curve No. 1 and reads in micro-ohms per foot, while the lower one is for Curves 2 and 3, and reads in ohms per mile.

Thus, the resistance per foot of single track of 60-pound rail is found by selecting 60 on the left-hand scale and following a horizontal to Curve No. 1, then a vertical to the upper bottom scale, finding 4 micro-ohms. The resistance per mile of the same weight of double track is found in a similar manner, by using Curve 3, to be 0.0107 ohms.

By the diagonal lines radiating from the lower left-hand corner, and the two inner scales on the right-hand, the resistance of any number of feet or miles may be at once read. For eight miles of single track of 80-pound rail the resistance is found to be 0.125 ohms, by following a horizontal from 80 on the left-hana scale to Curve No. 2, thence a vertical to diagonal 8, thence a horizontal to the middle right-hand scale. As all scales are decimal, any range may be obtained by multiplying or dividing by ten, one hundred, etc.

Also, with the same diagonals and the second scale on the right-hand side. the fall of potential per mile maybe calculated. To ascertain the drop per mile in a double track of 90-pound rails carrying 900 amperes, follow a horizontal from go on the left-hand to Curve No. 3, thence a vertical to the diagonal 9, thence a horizontal to the second scale on the right-hand, and multiply by 100, obtaining 6.3 Volts.

From the upper left-hand corner a series of diagonals are drawn for calculating the energy expended in the track portion of the circuit, and are to be used with the outer scale on the right hand. For example:-The power lost in a mile of single track of 70-pound rails carrying 30 amperes is found by following a horizontal from 70 on the left-hand to Curve NO. 2, thence a vertical up to diagonal 30, thence a horizontal to the outer scale, giving 15 watts as the dissipated energy. As the amount of energy consumed is in proportion to the square of the current, the outer scale must be multiplied or divided by 100, 10,000, etc., when the diagonals are multiplied or divided by 10, 100, etc.

Curve No. 4 enables the weight of copper per yard of equal conducting power to any rail combination to be calculated. To obtain the amount of copper that must be supplied to be equivalent to a single track of 60-pound rails, follow a horizontal from 60 on the left-hand to Curve NO. 2, thence a vertical to Curve NO. 4, then a horizontal back to the left-hand scale, finding 21.8 pounds per yard as the required amount.

On the top of the sheet two scales will be found for use with the diagonals radiating from the lower right-hand corner. The lower one, " Weight per mile in tons," accompanies the three dotted diagonals, while the upper one, " Area in square inches," is used with the three full lines. With these scales the weight per mile or area in square inches of any rail combination, or area and weight per mile of a copper conductor having the same resistance, may be found by following a horizontal from the weight per yard on the left-hand scale to the proper diagonal and then a vertical upward to the top of the sheet and reading the desired amount on the appropriate scale.

Only a slight study of the diagram is needed to discover the value of the track as a return circuit. A double track road equipped with 80-pound rails would show a resistance of 0.008 ohms per mile. The fall of potential with a current of 500 amperes would be 3.9 volts per mile, while the energy dissipated would amount to 2006 watts, or less than one per cent. To supply a copper conductor of equal resistance, would require an area of 4.20 .square inches, weighing 58 pounds per yard, or 25.3 tons per mile, costing about USD8000, or GBP1600.

But in order to realise the full value of the rails, it is necessary to insure perfect electrical continuity, - that debatable ground upon which has been fought the battle of the bonds. While electrical discontinuity at the joints has always been recognised, its importance has always been underrated, owing to the exaggerated estimate of the value of the earth as a conductor. In early electric railway days the fish-plates merely were held to make a fair connection. Standard construction was a bit of No. 1 wire, riveted to a hole in each rail, with the addition, in gilt-edged work, of a No. 0 ground wire. Small wonder is it that electrolytic action caused the bonds to disappear so quickly as to make railway managers declare their maintenance an impossibility, or that the ground wires sometimes ran so hot as to actually burn the ties in two, and derail the cars by allowing the rails to spread. Could the record of the mountainous scrap heaps of burned out motors be correctly interpreted, the true epitaph of more than one defunct railway would be "Poor Bonding."

Recent experience has shown that in a street railway track properly supported by adequate paving, no allowance need be made for expansion, and, that, except for convenience in handling, the rails might be endless. A continuous track would solve some of the most difficult problems, both electrical and mechanical, that confront the tramway engineer, and as a step toward such continuity, there seems to be little reason why track building of the future may not be done with 60-foot rails, thus reducing by one half or two-thirds the present number of joints.

For those that must remain, electrical welding from the conductivity standpoint presents the best solution, but mechanically the joint thus made is open to criticism.

The cast weld joint, from a mechanical point of view, seems at present the most perfect device for supporting rail ends, but experience has demonstrated that there is not a perfect metallic contact between the rail and the casting, owing to the presence on the rail of an inevitable film of oxide, and the chill experienced by the molten metal when it first encounters the cold steel. A combination of the two methods would seem to attain the perfect joint, for the electrical weld, if properly made, would secure electric conductivity, while the casting would reinforce a system otherwise mechanically weak.

All other methods of bonding are open to the difficulty of making and maintaining two electrically perfect contacts, at each joint between twn dissimilar metals. In order that such a bond should be efficacious, it is requisite that the strip forming the conducting link shall have no greater resistance than an equal rail length, and that the area of contact shall not only be equal to the cross-section of the rail, but enough greater to provide a reasonable factor of safety against possible future impairment. In these particulars rail bonds in common use have utterly failed. The usual form of bond is a piece of No. 0000 wire, upset at the ends and riveted into the rail webs. Such a wire has an area of 0.166 square inch, and is equal in conductivity to only a 10-pound rail. An 80, 90 or 100-pound rail would require 1.25, 1.41 or 1.56 square inches of copper or eight, nine or ten No. 0000 wires. Between the rail web and a No. 0000 bond wire the surface contact is about 0.7 square inch, or a little less than the area of a 10-pound rail, while 80, 90 or 100-pound rails have areas of 7.68, 8.65 or 9.61 square inches, so that, without allowing any factor of safety, the bond should have ten and one-half, twelve, or thirteen times as much contact as is at present customary, for it is perfectly obvious that, if at any point in the return circuit the resistance is increased, a portion of the current will seek the earth.

Even by restricting the danger areas to a minimum, and with the aid of a perfect rail-return to reduce the earth currents to their lowest terms, the electrical survey will show some points where corrosive action is to be expected. As disintegration takes place only where current leaves the underground structures for the earth, the last link in the protective chain is forged by attacking the electro-positive spots directly, and preventing the passage of electricity into the earth. For this purpose it has been proposed to place an additional dynamo in the station, the negative pole of which is to be connected to the pipe lines by extra wires. If the special machine be operating a few volts higher in potential than the rest of the station, its action, to use rather an unscientific hydraulic analogy, is that of pumping the electricity out of the pipe lines, and so keeping them constantly electro-negative.

Doubtless this method is efficacious if a sufficient number of conductors are installed to reach from the special dynamo to all the electro-positive spots on the pipe lines; but if this additional set of wires be installed, the extra dynamo is superfluous, for the conductors may, at a very slight expense, be made of such low resistance compared with the soil as to completely lead away from each electro-positive site all the current that would otherwise escape into the ground.

While a protective system may thus be easily made at a slight expense and completely fulfill the object for which it was designed, it is necessary to keep a close watch over its operation by means of repeated electrolytic surveys.

If the danger to which underground metallic structures are subjected from parasitic railway currents were irremediable, or if an efficacious protecting system could be secured only by incurring an expense out of proportion to the benefits secured, there might be a valid argument for the present state of the railway return; but when the protective system is easily and economically within reach, when its adoption largely contributes to the benefit of railway companies by reducing the expense for motive power, failure to adopt it is in the nature of a deliberate infringement of the rights of others, for which there can be neither legal nor moral excuse.


Note from 1999: This article was originally published in the Eletric Railway Number of Cassier's Magazine August 1899, which was reprinted in book form by the LRTA (then LRTL) in 1960.

Despite having been written 100 years ago it is still relevant today, especially with the current concern (in the UK at any rate) with leakage currents. With the abandonment, in Croydon, of the use of grassed track, despite being in common use on mainland Europe, due to concern about this encouraging leakage current it is felt that a look at the basic principles would be useful. One must wonder, that with the use of lead pipes being discouraged due to health risks and the employment of modern materials, whether our present concerns are slightly excessive.


Technical Article: top