"Perhaps no single invention in the history of the development of railway transportation has contributed more towards safety and despatch in that field than the track circuit. By this invention, simple in itself, the foundation was obtained for the development of practically every one of the intricate systems of railway block signaling in use today wherein the train is, under all conditions, continuously active in maintaining its own protection.

"In other words, the track circuit is today the only medium recognized as fundamentally safe by experts in railway signaling whereby a train or any part thereof may retain continuous and direct control of a block signal while occupying any portion of the track guarded by the signal."

"To Mr. William Robinson the Patent Office records concede the honor of having devised the first practical track or 'Rail circuit'. This comprised what is termed the closed track circuit. * * * Closed track circuits are very reliable, wholly safe in principle, and simple of application and maintenance."

The above paragraphs, quoted from the third annual report of the Block Signal and Train Control Board to the Interstate Commerce Commission under date of November 22, 1910, ably express in a few words what the invention of the track circuit has meant to the railroads of this and other countries. In order, however, that those who are not familiar with the principles of the track circuit may have some general knowledge of them, a simple, non-technical description is given, as prepared some years ago by Mr. J. P. Coleman, of the Union Switch & Signal Company.

Historical information on the development and use of direct current and alternating current track circuits for roads using electricity for propulsion purposes and those using steam will be found in a report on this subject made by Committee X to the Railway Signal Association in 1910.

The Rail Circuit Principle

By J. P. Coleman.

Assuming that it is clearly understood that the current is generated at the battery; that it flows from thence through the conductors (of which the coils of the magnet form part) and back again to the battery, and that the magnet is simply a device interposed in the circuit for the purpose of transforming electrical energy into mechanical (magnetic) energy, and that the latter can exist in an electro-magnet only with the presence of the former, we are now prepared to make clear the principle of an electric track section.

To assist in this, let us state an invariable law governing the flow of currents: If two or more paths be presented an electric current, it immediately becomes divided, and flows in each in quantities directly in proportion to the conductivity of each.

The unit of electrical resistance, whereby the comparative merits of various materials and sizes of materials as conductors are designated, is called an ohm, (just as the unit of lineal measurement whereby the comparative lengths and sizes of various objects are designated, is termed a foot) and we will therefore use that term in reference to the resistance of a conductor.

Figure 1 represents an ordinary gravity battery, the conductors from it, and the electro-magnet to which they connect; also the armature as attracted by the magnet and overcoming the spring which tends to withdraw it from the magnet.

Now, as long as the current flows through the magnet, this condition of things remains unaltered; but let a second path be presented to the current several hundred times less in resistance than the original one, and the result is that several hundred parts of the current will leave the magnet for the "short circuit," and consequently leave so little remaining in the original one that the effect will be practically to demagnetize the magnets.

Figure 2 will render this very apparent if we will assume the wire of the magnet R to possess a resistance of 10 ohms, and the conductors themselves a resistance so low as to be inappreciable and unworthy of consideration.

Now, assume the current to be flowing and the magnet to be charged, and let us take a piece of metal which has an electrical resistance of 1/100 of an ohm, and lay it across the conductors at any point between the battery and the magnet. The result is, that instead of flowing through 10 ohms resistance via the magnet, it follows the invariable rule, and takes that offering but 1/100 of an ohm; or, more to the point, if we assume the conductors referred to to be one mile of steel rails each (Fig. 3), and again leave their resistance (which would be about one ohm each) out of consideration entirely, leaving that of the magnet as first stated, and assume the bar of 1/100 of an ohm to be an axle and pair of wheels (a) of a train (Fig. 4), which possess the same resistance, we can readily see that the result would be exactly the same, i.e., instead of all the current passing through the magnets, as when the rails were unoccupied, the presence of the wheels upon them would cause 999/1000 of the current to leave the magnet and pass through them; they offering but 999/1000 of the resistance of the magnets, and thus leaving but 1/1000 of the whole current passing through them, which being so small a part of so feeble a current is imperceptible and without sufficient influence to hold the magnet charged. Therefore, it follows that the instant that a pair of wheels enters upon a pair of rails which thus form part of the conductors of an electrical current holding charged a magnet, that magnet becomes practically demagnetized, and consequently loses all power to overcome any opposing force in its armature.

When the armature of a magnet is arranged upon a small lever, by motion of which a second circuit is closed or opened, or two or more circuits are otherwise controlled, the entire device is termed a relay. In all forms of this instrument, as is the case with almost every other electrical instrument, the armature is so arranged as to fall by gravity, or by tension of a small spring suitably arranged, away from the cores of the magnets when they become demagnetized.

When switches are included in a track section (Fig. 5), it becomes necessary for safety to have them control the track section in such a way that unless they are properly set (and locked, if desired) for the main track, the continuity of the rail circuit is interrupted and the signal is thereby held at danger. To render more certain this result, the circuit controller (switch box) at the switch is arranged in such a way that the track circuit is not only interrupted beyond the switch, but is also short-circuited by it when the switch is not properly set. It is also necessary for safety that the side track from the switch points back to the fouling point (Fig. 5) be included in the track section: thus insuring that all trains on these tracks are out of danger of collision with the main track when a "clear" signal is displayed on it.

In dividing tracks into distinct electrical sections, it becomes necessary to insulate the rail ends at the terminal of each, from those of the adjacent sections. If this were not done the current of each section would traverse the next, and continue on indefinitely, influencing each other so as to interfere with or totally prevent the operation of all.

In order that we may fully comprehend the nature of an insulation, let us make clear a few facts concerning conductors in general. All materials conduct electricity to a certain extent; but some with much more freedom than others. Thus, silver, copper, gold, zinc, platinum, iron, steel, mercury, and other pure metals permit the passage of an electric current through them with but slight resistance, (although all offer a certain amount,) and are therefore termed conductors. The following liquids are classed as conductors: concentrated and diluted acids, saline liquids and water, although they are much less efficient as such than the metals.

To this list might be added the earth itself and the various ingredients forming it, the nature of which ingredients determines very much its efficiency as a conductor. Thus at points abounding in mineral deposits the earth would be far superior as a conductor to those parts in which none exist, but at best should be regarded as a poor conductor.

Next comes a class of materials which offer a great resistance to the current, and which from that reason are termed non-conductors, or insulators; of this class, rubber, glass, leather, resin, wood, brimstone and dry air are the most common.

Wood being a non-conductor, it is very evident that the cross-ties under the steel rails form an insulation between them and the ground; also, that if a piece of the same or similar material be placed between the rail ends, and that if two other pieces of sufficient strength be substituted for the iron fish-plates at that point, a secure insulation will be formed between the rails.

It is precisely in this way that the insulation of one rail from another is effected (Fig. 6) and the long, practical use of many hundred joints of this sort, has proved it to be a method both economical and thoroughly efficient.

A much more secure joint, however, is obtained by insulating the existing iron fish-plates from the rails by means of heavy fiber plates, and their bolts from the rails by fiber bushings (Fig. 7). While this method is superior to the first mentioned one in that it makes a more secure rail joint, it is no more efficient as an insulation.

One would naturally suppose that owing to the large surface of contact existing between the rails and their connecting or fish-plates, and from the apparent security of that contact obtained by the bolts through them, no trouble would be experienced by the current in passing from one rail to the other. This, however, is not the case, as the bolts and even the plates themselves frequently become loose, even when provided with the best of nut locks, and the rust and dirt settling between them and the rails oftentimes increase the resistance of a track section to a serious extent. Again, even when tightly bolted and locked, these plates form but an imperfect contact, owing to the scale or rust upon them. Therefore, to insure that the resistance of a track section may be as low and as constant as possible, we have found it absolutely necessary to connect each two adjacent rail ends together by means of a short piece of very strong wire (Fig. 8).

These wires are termed "track wires" (bond wires) and are provided with a button-head rivet at each end, which is securely soldered thereto, for the purpose of securing them to the rails. (Bond wires are now attached to the rails by channel pins or are welded on.) The connections from the rails to the battery and relay of a track section are secured to the rails in the same manner. The battery is usually located in a chute or well sunk in the ground at the terminal of each section, which is provided with an elevator in which the battery is placed and by which it may be raised and lowered at will. All wires when placed underground are run in grooved lumber in order that they may be secure from injury.

Even in very wet or snowy weather a single jar of gravity battery is generally found to furnish sufficient current to properly work the relay at the other end of any section less than three-quarters of a mile in length; although it frequently happens on longer sections and occasionally on those of ordinary length that two jars are necessary. A greater number of jars is never advisable since by increasing the intensity of the current, the liability of its leaking from one rail to the other during wet weather is correspondingly increased, and as this is attended with some uncertainty in the working of the relay of the section—due to the varying intensity of the current—it should be carefully guarded against. As two jars of gravity battery are not sufficient to operate a signal, lock, bell or any similar instrument with any degree of certainty, it becomes necessary to have a second set of batteries of a greater number of jars for that purpose. The armature of the magnet controlled by the track section is therefore made to control a second circuit using a battery of this sort (Figs. 3, 4 and 5) and which includes the magnet of the signal mechanism. The use of a relay on a track section is therefore necessary; and when it becomes necessary to control two or more devices, each requiring independent circuits, by one track, the use of a relay is indispensable.

Track Circuit Characteristics

While the fundamental principles of the track circuit are the same today as they were when it was originally invented by Dr. Robinson in 1872, it has been found that it is not as simple a device as was formerly supposed to be the case and many problems have arisen which have required and is requiring the careful study of the signal engineers. Accordingly, it is well to present briefly some of the track circuit characteristics as they are known today. In the following presentation, information has been collected from many different sources, including abstracts from papers presented on the track circuit by Mr. A. R. Fugina, signal engineer, and Mr. J. B. Weigle, signal inspector on the Louisville & Nashville.

There are two general classes of track circuits, direct current and alternating current, which may be further subdivided between single or double rail circuits. The essential feature of the track circuit is the insulation of each section of track from the adjoining sections. Each rail in the section is connected to the one adjoining by bond wires, for the purpose of making a continuous conductor from one end of the section to the other.

Rail Bonding

Under the present methods of bonding, the angle bar carries the greater part of the current, and bond wires frequently carry as little as 20 per cent. of it and sometimes even less. The rail resistance is lowest with new rails, but it gradually gets higher, due to rust and dirt formations between the angle bar and the rail. But even with new rail, the rail resistance varies greatly at different periods and even at different times during any twenty-four hours. This variation is entirely due to the fact that the angle bars carry more of the current than the bond wires, and that the bond wires under any condition are only large enough to carry the smaller part of the current from the battery. The lower the resistance of the bonds the less variable will be the rail resistance.

The resistance of the angle bars increases greatly as rail resistance increases, as a result of which the angle bars rapidly carry less of the current.

It is not infrequent to find the rail resistance to be as high as 0.20 ohms per 1,000 ft. of track, and we have known it to run as high as 0.264 in new rail, where especial attention had been given to obtaining as good bonding as possible. Under such conditions the angle bar carries very little of the current, the capacity of the bond is not sufficient to carry the current, and the net result is a failing track circuit, which is probably attributed to bad ballast, zinc treated ties or other causes.

The principal defect in the track circuit is that of improper bonding. The only explanation as to why No. 8 iron wires became standard for bonding appears to be that the bond wires were cut from this size iron telegraph wire which was in general use at the time rails began to be bonded. It is important to obtain better bonding to obtain a minimum constant rail resistance.

It has been recommended that:

First—The use of galvanized wire bonds should be eliminated.

Second—Forty per cent. copper clad bond wires should be used as a temporary expedient to replace galvanized bond wires.

Third—Except for theft and crystallization, copper bond wires would be much more advisable.

Fourth—Larger bond wires should be used, these bonds to be at least equal in carrying capacity to two 46-in. No. 6 solid copper or to two No. 2, 40 per cent. copper clad wires.

Until recently it has been the general opinion of all experts on the track circuit that the rail resistance was rather an unimportant factor and that, as a general rule, the change in rail resistance could be disregarded in making track circuit investigations and calculations.

Many bad track circuit conditions have been laid to bad ballast conditions, zinc treated ties, wet track, etc., which, if carefully analyzed, would have shown the trouble to be due to extremely high rail resistance. These faulty conclusions are being drawn nearly every day.

Single rail track circuits, so called from the fact that but one rail is insulated, are also used. Installations of this kind are made to avoid the expense of two insulated joints or where one rail is needed for another circuit. Such track circuits are more liable to failure than those having both rails insulated for the reason that the break-down of one insulated joint will extend the circuit beyond the proper limit and cause interference of neighboring circuits or extended shunting of the relay, due to the presence of a train beyond the insulated joint.

A track circuit may be made to perform two separate functions in which the direction or polarity as well as the presence of current is made use of in the relay, provided the first or principal function actuated by the presence or absence of current does not interfere with the secondary function, actuated both by the presence of current and its polarity.

Where switches occur in a track circuit, special means must be employed to prevent short-circuiting through the switch rods and leakage of current to the turn-out rail. The usual method is the use of insulated switch rods with insulated joints in the leads of the turnout and at the fouling point of the turnout. The switch points are bonded to the stock rails to insure shunting by a pair of wheels on any part of the track.

None of the methods employed in running track circuits through switches show any protection against an open switch. In order to obtain this protection a switch instrument or switch box is used. This consists of a device with electrical contacts, the whole mounted on a switch timber and connected to the switch point by means of a rod so arranged that when a switch slips open or is thrown open the movement of the rod actuates contacts which, on being closed, form a closed path from one rail to the other through wires connecting the rails to the contacts, thus when the contacts are closed by a switch being opened, the same effect is produced as if a train was on the circuit, shunting it out.

On electrically operated roads where tracks are bonded for the return propulsion current with heavy copper bonds, no additional bond wires are necessary.

The Track Battery

The usual form of track circuit has a primary battery at one end of the insulated track section, with the positive terminal of the battery connected to one rail and the negative terminal to the other, while a relay at the other end of the section is connected to the rails in a similar manner. Current flows from the positive side of the battery through the one rail, the relay and the other rail back to the battery, thus keeping the relay energized.

For d.c. track circuits, four types of cells have been used to a greater or less extent, the gravity cell; Lalande (soda) cell; storage cell and dry cell. The gravity cell has a voltage of about 0.8 or 0.9 volts, the resistance varying with the manner in which the cell is maintained and averaging about 3 ohms. It will remain active for long periods on closed circuits without appreciable polarization. Because of this high internal resistance usually no external resistance is necessary to be connected between it and the rail of the track. The e.m.f. of the Lalande (soda) cell may vary from about 0.67 volts to 0.88 volts while the internal resistance will range between 0.019 ohm to 0.4 ohm. Because of the low internal resistance of these cells it is necessary to use an external resistance of the proper value between the cell and the rail. The storage cell is made in various capacities and a fully charged cell on open circuit has a voltage of approximately 2.1 volts which, when placed on discharge, becomes approximately 2 volts and drops to about 1.8 volts when completely discharged. The voltage in this type of cell varies with the density of the electrolyte and to a certain extent with temperature. It has practically a negligible internal resistance and it is also necessary to use an external resistance in the leads between the cell and the track to prevent a flow of excessive current when a train occupies the track. The dry cell is used only in emergency cases or occasionally for open circuit track circuits of 2 or 3 rail lengths, which are sometimes used as annunciator starts to announce the approach of a train to a tower-man. It is designed primarily for open circuit work and will polarize when current beyond a certain figure is drawn continuously from it.

The Track Relay

The track relay is a development of the instrument of the same name used in telegraph service. It consists of an electro-magnet of the horseshoe type with a pivoted armature, carrying one or more fingers for making or breaking electric circuits for the control of signal apparatus.

Track relays with resistances of 2 and 4 ohms are usually employed. From experience with two-ohm relays on the L. & N., covering a great many of them on all kinds of circuits, the following conclusions are reached:

The two-ohm relay is at least as safe as the four ohm. It should be thoroughly understood that it is as important with the four-ohm relay as it is with the two-ohm relay to have not less than the R.S.A. recommended limiting resistance between the battery and track. This is important with any kind of low internal resistance battery, and under certain conditions with gravity battery also.

In one case assume a train to be passing from the relay end to the battery end of a track section and in the other case from the battery end to the relay end. The effect accomplished is the same except that the relay will not release so quickly when the train passes from the battery end towards the relay end, and this is in part due to the self-induction of the circuit through the relay coils, the rails and the axles of the train. It is due more, however, to small current leakage from the adjacent section and the effects of stray currents which are always present to a greater or less degree. A broken rail will also generally open the circuit and de-energize the relay. Circuits for the control of the various signal devices are broken through the contact points of the track relay.

Track Circuit Maintenance

Cross ties have a relatively high resistance to the passage of electric current, but when a large number connect the rails many multiple paths are introduced into the circuit through which the current may flow from one rail to the other, and, considering them as a whole, the resistance they offer to the passage of the current reaches a relatively low value. Consequently there is always a current leakage from rail to rail through the cross ties and ballast. Every effort should be made to secure and maintain the best ballast and drainage possible on d.c. as well as a.c. track circuits. Cinders, dirty sand, soft water-logged ties and ballast not well cleaned away from the base of the rail will produce track circuit trouble, particularly during wet weather, while good rock ballast, sound ties and clean track give the greatest efficiency.

The use of ties freshly treated with zinc chloride also reduces the ballast resistance. If too many such ties are used in a track circuit the current leakage between rails becomes so great that not enough current reaches the relay to hold it closed, the effect being the same as if a train is on the track circuit shunting out the relay. For good results, the number of zinc-treated ties installed per year in any track circuit should not be greater than 15 per cent. of the total number of ties in that circuit.

Track Circuit Troubles

Some of the common track circuit ailments are relay and track battery troubles, defective track connections, poor bonding and broken rails, short circuits or shunts, excessive leakage and defective insulated joints, all of which will cause the signals to be set in the danger position, while defective relays, foreign current and poor wheel contact may result in a false proceed signal indication with a train in the block section.

It was the quite general practice to operate bad track circuits by piling on gravity battery, either in multiple or multiple-series arrangements to obtain operating results without any regard to the safety of the circuit and, no doubt, many false proceed failures were caused thereby.

The effect of temperature changes on track circuit operation are of considerable importance. The track relay, which is generally housed in a cast or sheet iron box, probably is affected more by changes in temperature than any other part of the track circuit. The resistance of a 2-ohm relay, which is 2 ohm at 70 degrees F., will be 2.22 ohm at 120 degrees F., and 1.69 ohm at 0 degrees F., a variation of .53 ohm. The pick up and release of the relay, .2 and .1 volt, respectively, at 70 degrees F., will be .22 and .11 volt at 120 degrees F., and .17 and .085 volt at 0 degrees F. A relay, with a normal resistance of 4 ohm at 70 degrees F., will be 4.45 ohm at 120 degrees F. and 3.38 ohm at 0 degrees F., a variation of 1.07 ohm. The pick up and release, .3 and .14 volt, respectively, at 70 degrees F., will be .33 and .16 volt at 120 degrees F. and .25 and .12 volt at zero.

The point which is intended to be brought out by these figures is that when the temperature of the relay increases, a correspondingly higher voltage is required to pick up the armature, and when the temperature decreases the armature will hold up with lower voltage across the coils. This indicates that a track relay is more liable to fail to release due to an imperfect train shunt in cold weather than at any other time.

Some of the best preventatives that may be provided to guard against false proceed signals due to track relays failing to release with a train in the circuit, are:

1. Use as much resistance as practicable between battery and track.
2. Use low resistance bond wires, and maintain bonding in good shape.
3. Keep ballast well cleared from contact with rails.
4. Maintain insulation in insulated track joints in good condition.

Aside from these simple remedies no definite rule can be given to combat foreign current. If it is so troublesome that these methods do not overcome it, the circuit affected must be carefully studied to determine the source of the foreign current and its path to the rails, then special means can usually be provided to overcome it.

Ballast Resistance and Leakage

The importance of ballast resistance has long been recognized, and this always has been considered the great variable, whereas, investigations show that the ballast resistance is at least no more variable than the rail resistance, and that of the two it is more important to reduce the rail resistance to a minimum, and especially to establish it as a constant.

When the ballast leakage problem was first taken up (on the L. & N.), various kinds of ballast were measured in both wet and dry weather, the intention being to determine the lowest possible resistance per 1000 ft. for each kind of ballast. It was proposed in this way to establish a standard minimum resistance per 1000 ft. for each kind of ballast. For instance, if a number of measurements in wet weather showed 8 ohms per 1000 ft. as a minimum for track circuits with crushed rock ballast, it was the intention to adopt 8 ohms as the standard minimum ballast resistance per 1000 ft. for all track circuits where crushed rock ballast was in use. If a number of wet weather measurements showed 4 ohms per 1000 ft. as a minimum for cinder ballast, it was the intention to adopt 4 ohms as the standard minimum ballast resistance per 1000 ft. for all track circuits where cinder ballast was used. It was the intention to follow out the same process and establish a standard for all kinds of ballast in use. This was soon found to be impracticable.

After making many ballast resistance measurements, it was noticed that the variation of the resistance on any track circuit, as between wet and dry weather, generally followed quite a definite rule. For instance: If the resistance per 1000 ft. of dry ballast was found to be 28 ohms or more, it would be not less than 8 ohms per 1000 ft. when wet; or if resistance of dry ballast was found to be between 22 and 28 ohms per 1000 ft., it would be not less than 6 ohms per 1000 ft. when wet.

Once a relay is picked up or energized, but a small amount of current is required to maintain it in that condition. This is one reason why it is important to keep the ballast clear of the rails and it is because of the condition which may cause a relay to remain energized that rules are in force requiring the signalmen to disconnect a track relay when track forces are changing out rails.

Combined Rail and Bond Wire Resistance

On circuits newly bonded with two 46-in. galvanized iron wires a joint, the combined rail and bond wire resistance was found (on the L. & N.) to vary from .02 ohm per 1000 ft. of track on some circuits to .265 ohm on others, a difference of over 1300 per cent. This was rather puzzling. After a great many measurements had been made on different circuits it was found that no two measurements gave the same results, notwithstanding the fact that in many circuits the size of rail, length of bond wires, and age of bonding were exactly the same. On account of the bonding being new and the channel pins well driven, the contact between the bond wire and rail was above suspicion. The only other part of a track circuit that could possibly be the cause of this difference was in the contact between the angle bars and rails, and this later proved to be the case. Actual measurements made in the field proved that when the rail is new and the joint bolts tight, nearly all of the current flowing from rail to rail passes through the angle bars, whereas when the rails get old a coating of rust and dirt forms between the rail and angle bars, forcing practically all of the current through the bond wires. On most of the circuits measured on the L. & N. the combined rail and bond wire resistance was found to be less than .1 ohm per 1000 ft. of track, although many were found to be between .10 and .30 ohm per 1000 ft. It is interesting to note that two circuits were found bonded with two 52-in. iron wires, for which the combined rail and bond wire resistance measured .410 ohm, and that by adding two 40 per cent. copper clad bond wires to each joint the combined resistance was reduced to .144 ohm.

The Growth of the Track Circuit

Unfortunately there exists little or no data regarding the mileage of track circuits installed from the time the first installation was made by Dr. Robinson at Kinzua, Pa., and Irvineton up to about 1905. During the period between January 1, 1905, and September 30, 1906, the total automatic block signal mileage installed was 1,710.6, which brought the total up to 6,826.9 for the United States. Between September 30, 1906, and January 1, 1908, 3,976.1 miles of automatic signals were installed, which increased the above total to 10,803.0 miles.

The Block Signal and Train Control Board, seeing the need for accurate data in the signal field, started the tabulation of such statistics when it compiled and issued Block Signal Statistics as of January 1, 1908. After this board went out of existence, the Bureau of Safety of the Interstate Commerce Commission continued the collection and publication of these data yearly. Perhaps no better word picture can be given of what Dr. Robinson's invention has meant to the railroads than to present the story in the form of a table showing the miles of road and the track equipped with the track circuit since January 1, 1908. In addition to the table, the accompanying chart presents the information in a graphical form.

Progress Chart of Automatic Signal Installations
Since January 1, 1908.