TRANSVERSE fissures in rails are fatigue failures of internal origin, whose rate of growth depends largely upon the density of traffic. The typical air-tight fissure, (Fig. 1), has a characteristically bright, silvery appearance produced by the peening action of the fissure faces when the rails are subjected to the impact of heavy wheel loads.

This type of failure was brought to the attention of railway engineers about 19 years ago when a passenger train was derailed by a rail which broke into 17 pieces, disclosing several transverse fissures.

In the years which followed, rail failures due to transverse fissures were reported in ever increasing numbers. The presence of large transverse fissures in relatively new rail under the best conditions of track maintenance, together with the utter impossibility of seeing or locating such fissures, placed this type of failure in the class with a cyclone, a lightning stroke, or other "act of God." In the early part of 1927, internal fissures were successfully located in rails at the laboratory by the electric method and after several months of experimental work a fissure detector car (Fig. 2) was constructed for the American Railway Engineering Association. · The car passed its acceptance test and was placed in testing service on the railroads of the country in November, 1928.

The fissure detector depends for its operation upon the fact that electric current flowing through a rail is compelled to pass around any fracture, inclusion, or separation in the metal. When a rail is sound the axis of current. is substantially straight, but if a fissure is encountered, the current axis at that point is displaced. An inductive pick-up device is mounted directly over the rail head, where it cuts the magnetic lines of force surrounding the rail. In the case of a sound rail carrying current, the magnetic flux surrounding the rail is uniform -and no e. m. f. is generated as the inductive pick-up passes along, but when the axis of current is deflected, as at a fissure, the flux density around the rail varies, resulting in a generated potential at the terminals of the pick-up. This output from the pick-up is amplified and led through relays connected to recorder pens, thus giving a record of the received impulse (Fig. 3).

In applying the electrical method of fissure detection to inspection of rails in track, the following factors materially affect the final design:

1. Many spots are encountered where locomotive drivers have slipped in starting and gouged out the rail..

2. Worn or flat spots occur frequently, especially opposite the rail joints.

3. In certain stretches of track the rail surface has acquired a pronounced ripple, termed corrugated or washboard rail.

4. The rail is frequently covered with a hard insulating film, the resistance of which may run up to several thousand ohms.

5. The gage of the track varies between certain limits.

6. The rail surface may be level or canted.

7. The size of the rails varies from 70 to 130 lb. Per yd.

The single unit type of detector car (Fig. 4) is 45 ft. long, self-propelled, equipped with extra heavy frame and axles, and capable of a speed of 45 mi. per hr. when running to and from work. Just back of the driver's compartment are located the cook's gallery and comfortable living quarters for the crew. Back of the living quarters is the compartment containing the power plant of the detector equipment. This power plant consists of a specially designed double commutator,

6000-ampere 2-volt d-c. generator, direct connected to a 50-hp. Four cylinder gasoline engine. The exciting generator and air compressor are also driven by this engine. Current from the 2-volt generator is carried by heavy busses to the blush units located on either side of the car. The brush unit assembly (Fig. 5) consists of two sets of eight solid phosphor-bronze brushes mounted about

3 ft. apart and rigidly connected by a cast frame. These brushes are supplied with individual springs which insure a pressure of 50 lb. per brush. Flanged wheels located at the ends of the brush holder engage the rail when the brushes are down and, due to their angular mounting, cling to the gage side regardless of varying track conditions. The brush units are applied to the rail by air-operated pistons electrically controlled from ;the operator's switchboard. A safety switch is connected to the gear shift of the motor-car drive so that the brushes will lift if the driver attempts to back up while they are on the rail. It was found that brush contact resistance was greatly reduced by the use of water on the rail. Provision is there fore made to carry a sufficient supply of water for this purpose and the brush control switch applies the water to the rail automatically when the brushes are lowered. Mounted between the brushes is a separate carriage upon which is supported the pick-up unit. This carriage with its eight wheels and swivel mounting holds the pick-up unit at a fixed distance above the rail and rides over burns and flat spots without interruption. Initially driver burns and flat spots gave indications which tended to slow up the testing.

This source of trouble was materially reduced by the eight-wheel mounting. Recently a new pick-up device of radically different design has been on test on five of the cars. It eliminates the indications due to flat spots and all but the largest of the burns.

The operating room is located at the back end of the car and contains the recording table and control board (Fig. 6) as well as amplifying equipment and batteries. The output from the pick-up

unit is brought up to a four-tube amplifier which is properly shielded and protected from shock. Three recorder pen relays are connected to the output of the amplifier and adjusted to respond to different values of plate current, thus giving an indication of the size of the defect detected. Other relays are provided to operate paint guns so that the fissure will be automatically marked by a spot of paint. The record (Fig. 7) is made up of nine lines; the three pens on either side (numbered 1, 2, 3) give the relative size of the fissure while the lines in the center record the angle bars. Joint cut-outs mounted on the brush holders engage the angle bars and cut out the fissure pens and paint gun at the joint, at the same time operating the joint pens. The line to the extreme right is a land-mark pen operated from the motor-car driver's switch-board. He presses the button whenever a mile post, bridge, signal tower or other land-mark is passed and the operator at the recording table stamps in the proper designation.

In operation, the car passes along the track at approximately 6 mi. per hr. with from 2000 to 3000 amperes flowing through each rail, the current value depending upon the weight of rail being tested. The operator, seated at the recording table in the rear of the car, has the record before him and a clear view of the tested track. When the car passes over a fissure an indication appears on the record and the defect is automatically marked with a spot of paint. The operator notes the record, sees the paint mark on the rail, and stopping the car, backs it up to the paint mark for a

hand test of the suspected spot. To make the hand test, 1500 amperes is introduced into the rail by means of air-operated contactors and the unit I. R. drop is measured with a galvanometer. By means of this test the size of a transverse fissure can be determined within a few per cent. The detector car not only locates transverse fissures, but split heads (Fig. 8), horizontal fissures (Fig. 9), compound fissures, pipes, cracked webs, broken bases, and other defects. When a defect is located by the car, the first question asked by the railroad representative is, will it be safe to leave the rail in track until there is a lull in the day's traffic, or must the next train be tied up while the rail is replaced? The answer to this question lies in the result of the hand test. When the hand test shows no transverse crack, the rail is marked for removal at the next opportune time, but if a large transverse fissure is indicated the railroad takes no chance, removing the rail at once.

Although the testing speed of the car is 6 mi. per hr., there are many elements which enter in to reduce the total mileage tested per day.

1. Traffic delays due to clearing for trains frequently consume considerable time, but are unavoidable.

2. When a fissure is located the railroad representative always makes a record of the manufacturer, the year rolled, ingot letter, and heat number of the rail. If the rail has been in track for several years or has been oiled by the track oiler, it is often difficult to locate and decipher the heat numbers and it is not at all uncommon on some roads to have from five to fifteen minutes wasted in obtaining the rail data.

3. A repeat run over a suspected rail takes 30 seconds and from two to three minutes are necessary for a hand test.

The testing procedure on the ten cars now in service is practically identical, but the railroads differ somewhat in their method of conducting the test. A survey of the various methods employed indicates that the following procedure results in a minimum of time lost.

The railroad has one man on the car and one on a hand car which follows the detector car.

In addition there are, of course, the conductor and flagman. When an indication is recorded the car is stopped, a check run made over the suspected rail, and the car backed up for a hand test. The operator examines the rail and if this visual inspection discloses exterior evidence of a split head, horizontal fissure, or other defect he gives the word to the railroad man, who makes a red paint mark on the tie at the defective spot and chalks the identification number on the rail. If no visible defect is noted, the hand test is made and the rail is marked as before. The car then proceeds with the testing while the man on the hand car stops to take down all of the data together with the location of the rail, and then continues on after the test car. In this way the maximum time taken is the two or three minutes required for hand testing.

Since the first detector car was placed in service two years ago, many interesting fissure facts have been accumulated. 1. Rails have been found containing many defects as, for example, one which had 24 transverse fissures. 2. Fissures have been found in rails 28 years old; and also in rails 5 months old. 3. Many miles of track have been found free of fissures; other short stretches have had 20 or 30 fissures per mile.

While these isolated cases are of interest, the most important point is that the transverse fissures are now being located and removed from track before they can cause any damage, where as prior to two years ago their presence was not known until they had grown to the surface or broken the rail.

BACK TO HARCOURT DRAKE