Matt's Web Page--Signals--How Railroad Crossing Signals Work

Picture of access road to Lincoln Air Base

Figure 1--Access road to Lincoln Air Base. This is the actual location described by the circuit print below.

Introduction and Outline

This document describes how the circuitry works at an active, full size, "real" railroad crossing. The scope is of signal circuits and detection circuits from the earliest forms to the present, with the emphasis on relay circuits. The outline is as follows:

Included is an actual railroad circuit print showing the complete wiring of the location pictured above, which used stick relays at the time the print was current.

Before we begin, it is convenient to observe Figure 2, which illustrates the function of the master relay.

Most crossing installations, unless they are entirely solid state, have a normally energized master relay. On prints, it is often designated XR. In some cases, the configuration in Figure 2 may vary somewhat. For instance, if two tracks both cross the highway near enough each other to a share a fraction of the highway warning devices, there may be more than one interface relay affecting the warning operation. Also, the detection circuits may have other outputs to provide train information to controlled highway intersections that are near the crossing. But in any case, the key concept is that when a train is approaching, the detection circuits "drop" the master relay to activate the signals. (XR actually stands for "Crossing Relay," but it is usually called the "Master Relay" instead because it is more descriptive and reduces confusion with other relays at the crossing.)

The circuits on either side of the XR in Figure 2 may be simple or complex. They may be relays themselves or they may be solid state. In discussing them, the master relay forms a good point of reference.

Signal Circuits

Railroad signaling premises usually require that circuits be designed on closed circuit principles. Normally energized warning circuits at crossings include the "hold-clear" on each of a gate mechanism, rotating banner (i.e., the Griswold rotating STOP sign) or an Automatic Flagman (i.e., US&S type of wig-wag with a disappearing banner). However, some devices require normally open circuits because their fundamental nature is to be normally off, as with flashing lights, bells, and the operating coils of most wig-wags.

Lights

The most formal terms for the lights on the signal mast seem to be crossing flashers or flashing light units. The lights on the gate are simply called gate lamps. A diagram for crossing lights in operation is shown in Figure 3 with a total of eight lamps, as would be used in an installation like the one in Figure 1. It might easily be imagined that the lights would be wired something like the left side of Figure 3, but that would be incorrect. Actually, they are wired as on the right side of the diagram under "correct." The difference is that in the correct scheme, the left and right lamps are basically connected in series to begin with, and the flasher relay contacts merely short or bypass half the bulbs rather than supply them.

Print showing crossing signal lamp wiring schemes in operation.

Figure 3--Crossing signal lamp wiring schemes in operation. Keep in mind that the yellow highlights show current, not voltage. Notice that even though the flasher relay contacts on the correct and incorrect wiring schemes are shown moving in unison with each other, the lamps in the respective schemes glow on opposite sides.

Static, Black & White Version of Above for Printing
File Type	File Size	File Name
PDF	20 KB	lamps.pdf

All the relays shown above reside in the relay cabinet or equipment housing near the crossing with wires running out to the signals; they are not parts of the signals themselves. In the correct wiring scheme, the three wires going out to the lamps are here designated EN, EB, and E. (Another naming scheme uses LL, RL, and FL, respectively.)

While not very intuitive at first, the correct circuit topology has two major advantages. First, the bulbs being in series makes for a more favorable situation in case the flasher relay or wire E should break. In that case, instead of bulbs on neither side glowing, bulbs glow on both sides, although not as brilliantly for being in series. See Figure 4. (That the flasher relay should break in the middle position, however, would be a very unusual failure mode because the flasher relay armature is weighted so that it always tends to fall to one side.

A crossing gate typically has three lamps spaced along its arm. The two innermost gate lamps flash in unison with the heads on the mast, but the outermost lamp burns steadily. Even when the other lamps are flashing, the correct scheme provides a constant supply of voltage across EN and EB to which the end gate lamp may be connected. Therefore, the same three wires that supply the flashing heads on the mast can supply all three gate lamps. This is the second advantage to correct wiring.

As an aside, it may be further noticed that each gate lamp is connected across a unique combination of two out of the three wires: EN-E, E-EB, and EN-EB. At least one manufacturer of gate lamps uses a radially symmetrical connector with three pins spaced 120° apart so that each lamp's function (left, right, or steady) depends on how the cable connector is rotated when plugged in.

Figure 3 and Figure 4 show a battery as the signal lamp power supply. But crossing lamps together can easily draw 10 amps or more. Although batteries may have been used alone at some remote locations, in reality even many older installations had a low voltage a.c. supply for the lamps derived from commercial power, using batteries for the relays and as backup for the lamps. The fail-over from commercial power to batteries for the lamp power supply was handled by a normally energized "power off relay," POR.

Incandescent crossing signal lamps today are usually rated at 10 volts and 25 watts. Ten-volt supplies are common in railroad signaling. However, 12 volts is also used, and dropping resistors must then be placed in series at the relay cabinet to provide 10 volts for the lamps. If these resistors are placed in-line with wire E, this allows the lamps to be a little brighter in the failure case of Figure 4 because the resistor is not then in the current path. However, the lamps are still dimmer than normal for being in series. Another scheme puts resistors in the EN and EB lines downstream of the EOR contacts. Both of these resistor schemes produce a different voltage for the gate tip lamps than for the other lamps. This is acceptable because the lamp voltages are viewed as somewhat nominal, although measurements are required.

Figure 3 and Figure 4 also show all the lamps connected on the same circuit. However, it is good practice and usually necessary to distribute the load over more than one set of contacts. In Figure 18, this is done by jumpering several sets of contacts together in parallel, as implied by the notation "1-2-3-4" next to the contacts of the flasher relay, EOR. Another solution is to have separate light circuits on each set of contacts, which requires more complex wiring. Distributing the load amongst several contacts is not on the flasher relay only, but also on the lamp control relay. Sometimes there are even separate power supplies. When separate circuits are used, the lamps may be grouped a number of different ways. One circuit may represent one pair of mast lamps, a pair of mast lamps with one gate, all mast lamps together, all gate lamps together, a whole signal together, or some other combination. Understandably, the use of lamp resistors seems to limit the number of lamps on one circuit. The additional wires required for separate circuits may have the letter I appended to their designations.

Finally, the lamp control relay of Figure 3 and Figure 4 is an abstraction because there usually isn't a separate relay of that name. At a crossing with no gates like the one in Figure 1, the XR can serve as the lamp control relay by itself. At a crossing with gates, the lights need to be flashing whenever the gates are not up. In that case, the lamp control relay could actually be the crossing gate repeater relay, XGPR or GPR. Although the control of the gate repeater relay is broken by position contacts in every gate at the location, the supply for the gate repeater relay is still derived from the XR so that the lamps begin to flash as soon as the XR is de-energized.

Figure 5 below illustrates most of the circuit details and additional features mentioned above. Keep in mind that wiring and designations can vary largely from railroad to railroad and location to location. This diagram is a kind of collage of crossing lamp wiring based on what the author has seen and heard.

Static, Black & White Version of Above for Printing
File Type	File Size	File Name
PDF	32 KB	lampdet.pdf

All the relays above may be special purpose relays designed with the specific application in mind.

The flasher relay is designated EOR or FR, and is always designed for the purpose. The relay needs power to operate, and a separate contact of the lamp control relay (shown as XGPR in Figure 5) switches DC power to the EOR at the same time it switches the lamp supply so that it does not run except when needed. A photograph of a flasher relay is shown in Figure 6 below. This is an electro-mechanical relay that is designed for use with crossing lamps.

This flasher relay works as follows: The armature holding the moving contacts rocks back and forth. The rocking is caused by two electromagnets arranged so that one tends to pull the armature one way and the other tends to pull the armature the other way. The electrical supply of the electromagnets is controlled by internal contacts similar to the lamp contacts. The coils of the two electromagnets are wired to the internal contacts just as a pair of lamps, with care that at a given time the electromagnet is on which tends pull the armature into the opposite position, resulting in oscillation.

The speed of oscillation is kept relatively low by the inductance of the mechanism, that is, the momentum-like property that resists changes in the magnetic flux. From looking at the figure, it can be seen that the tops of both electromagnets share a common pole piece that channels the flux to and from both coils through a central column in front of the coils: it is part of the magnetic circuit or flux path. Copper washers are stacked on this column to increase the inductance. As the mechanism wears from use (probably at the contacts), it tends to slow down a little. Washers can then be removed in the field to reduce the inductance and speed it up again. The number of washers, then, is used to "trim" the flash rate and keep it within specs.

Crossing Bells

Bells are mostly for the benefit of pedestrians or vehicles stopped on the tracks or under the gates, so crossings at remote portions of the open highway are less likely to have them. The crossing pictured in Figure 1 does not have a bell. Usually there is one bell, but some crossings have more. In some areas, bells operate only when crossing gates are rising or falling.

Figure 7--An otherwise typical crossing configuration with two bells. Location: MSPA (Former CRI&P) near De Witt, Nebr.

Unlike the flashing lights whose oscillation is controlled from the relay cabinet, the crossing bell externally requires a mere power supply to operate. The power for the bell is switched to it by back contacts of the master relay. Where the bell's operation depends on the gates, there is a contact of the circuit controller inside the gate mechanism that makes and breaks the supply of the bell at certain positions.

Most mechanical crossing bells work basically the same way as a standard electric bell, such as a school bell or fire bell. That is, an electromagnet forces the clapper to the gong and in so doing disconnects its own electrical supply internally through a set of contacts, causing the armature to fall back until the contacts close and the cycle repeats. There are about 200 distinct strikes per minute. The crossing bell rings more slowly than a school bell or fire bell because of the relatively large mass of the parts involved, and also perhaps some hysteresis in the contacts. Yet it works on the same basic principle. However, there is such a thing as a motor-driven bell.

Picture of the Inside of a Mechanical Crossing Bell.

Figure 10--The Inside of a Mechanical Crossing Bell. Power is connected to two terminals of the white porcelain block (broken).

Mechanical bells are susceptible to different kinds of failures. I once heard one operate without ringing in winter. Also, because of the inductance of the coil, the bell contacts tend to arc when opened, which can burn them away over time. Snubbing resistors or capacitors can reduce the arcing, but may not eliminate it entirely.

An electronic bell is ostensibly more reliable, but the bell sound they generate is only approximate. One signal maintainer I know refers to them as "humdingers." They also do not finish reverberating when power is removed.

Gate Mechanisms

A gate mechanism is very similar to a semaphore mechanism in that it has a circuit controller, a motor that lifts the gate and resists its fall by the electric retarder principle, and a normally energized hold-clear device (here designated HC). See the article "How Semaphores Work" on this website. It is very different, though, in two main ways: first, the gate motor actually drives the gate down for about the first half (45°) of the descent, and retards only for the final portion of the descent; second, the gate mechanism itself contains a relay called the motor control relay (MCR). Although the gates are counterweighted, they are required to be a little heavier than the counterweights by moments. The gates would still fall the entire distance by gravity, albeit more slowly at first, without the motor's assistance.

The purpose of the MCR is to change the connections to the motor, since it drives both up and down; to reverse its supply polarity and select the appropriate circuit controller segments for the up or down movement. Generally, it is energized when the gate is supposed to rise and de-energized when the gate is supposed to fall.

The control of the gate, then, generally centers around the status of a normally energized "UP" signal from which the supply for the coils of the HC and MCR are derived. However, gate mechanism wirings may vary with the particular design, and the individual supplies of the HC and MCR may be broken by the circuit controller in the gate mechanism when they are not needed. Specifically, the MCR is energized when the gate is rising to the up position, but the circuit controller may or may not cut the MCR out after it gets there. Conversely, the HC does not need to be energized until the gate is fully up, and it may not be energized until it is cut in at that time by the circuit controller.

Sometimes the term "power down" is used in connection with crossing gates. This refers to the situation where the motor drives the gate down under power, as intended. The term is not used in the modern sense where a machine conserves energy by "going to sleep." A less confusing synonym is "motor down."

Detection Circuits

Constant Warning Distance

Interlocking Relays

Perhaps the most primitive active crossings merely gave a warning when a train was in the area without regard for whether it was getting nearer or farther away, much less how soon it was getting there. Any detection circuitry worth mentioning, however, is at least a solution to the problem of making the warning stop after the train has passed, even though the train may still be close enough to be detected.

The first such solution in widespread use appears to be that of the "crossing relay," usually referred to more descriptively as the "interlocking relay" because of its design. It could make train detection effective for only the side of the highway that the train approached from and lock out the other side.

Upon close inspection, the relay is found to be actually more like two relays in a single case that may operate independently of each other except that they are mechanically interlocked with each other so that only the first relay to be de-energized can fully drop and thus close its "back" contacts. The other relay can then drop as well, but only into a "half drop" state where its back contacts do not close. In order to be effective for the intended purpose, the second relay must not be allowed to reach full drop even if the first relay to drop has been re-energized. The system is only reset when both relays are energized at the same time. An interlocking relay is shown in Figure 11. Figure 12 is a close-up of the actual interlocking elements as seen in the middle of the relay by looking through the glass on the front of the relay's body. This is only one design of an interlocking relay.

The operation of the interlocking relay is illustrated in the animation of Figure 13 below. The relay does not require power to operate apart from that applied to the coils shown in the diagram. The crossed lines between the coils simply denote the interlocking relationship between what would otherwise appear to be two separate relays. The track is divided into two long sections called approaches on either side of the crossing. The limits of these sections are set by electrically insulating rail joints which appear orange. Each coil of the interlocking relay is connected to the rails of an approach near the crossing, and a track battery is connected to the rails of each approach through a current limiting resistor at the far end. Thus the coils of the relay are both normally energized. However, when a train is within an approach section, practically all the current from the track battery on that approach is shunted through the train's axles, which shunts the current away from the associated coil and de-energizes it.

If the train enters Approach A first, the warning for Approach B will be locked out by the mechanism as the train passes through, and vice versa. By routing the signal power supply through the contacts as shown above, the interlocking relay may provide all detection logic desired. If a crossing bell is the only warning device as was the case on some early crossings, the interlocking relay can be absolutely the only relay at the crossing, representing by itself every part of Figure 2 above. As such, the interlocking relay serves as the master relay, and there is no separate master relay per se. With the addition of a flasher relay, flashing light signals may also be used. Thus, a crossing that uses an interlocking relay can be quite simplistic in terms of wiring, using only one or two relays. The price for simple wiring is the need for specialized equipment, which the field of railroad signaling seems to have borne quite happily.

Although the back contacts of the interlocking relay do not close at half drop, the front contacts still open at half drop. The interlocking effect is therefore realized only on back contacts, which implies that only back contacts can be used in the crossing circuits, at least with this particular design of relay. This further implies that this particular design of interlocking relay cannot be used with warning devices that require closed circuit controls (hold-clear devices, etc.). For closed warning circuits a special kind of interlocking relay was apparently used which did not open its front contacts before achieving full drop.

Figure 13 has an arrow indicating the point at which the crossing is located along the track in relation to the two approach sections, but it ignores the problem of road thickness. Where are the insulated joints at the crossing placed in relation to the road? Practical considerations require that they not be placed in the road. Some early installations put them both on one side of the road as in Figure 14.

This extends one approach, Approach A, through the crossing and allows the warning to continue for as long a train blocks the crossing if it enters by that approach. But the moderately obvious problem is that trains entering from Approach B will lock out the warning on Approach A, and the warning will disappear before the end of the train clears the crossing. The only consideration given was to determine from which approach the greater current of traffic entered, and of course place the insulated joints to extend that approach through the road.

At the time this was done, active crossing devices were considered supplemental to the motorists' powers of observation and judgment. In fact, even if the signals displayed a warning, it was historically viewed as more of a suggestion than an imperative as long as the train was not blocking the crossing. So, with some historical imagination, we can see why installations like the one in Figure 14 were built even though they would never pass in new installations today.

As an improvement over this design, the joints were staggered across the road as shown in Figure 15. Without any adjustments in the wiring, this arrangement would be worse because neither approach would extend across the road since continuity is needed in both rails.

But fortunately there is an adjustment in the wiring. In Figure 15, a back contact on each half of the interlocking relay bridges the insulated joint on the side nearest the approaching train, effectively extending whatever approach a train enters from through the crossing. The circuitry is otherwise the same as in Figure 13 apart from being drawn a little differently so that the track wires attach to the rails as near as possible to the insulated joints.

The opposite approach is still shunted as the train passes away, but the joint separating that approach from the crossing does not get bridged because that half of the interlocking relay does not achieve full drop. The road portion, therefore, remains a part only of the leading approach until the train clears it, and then the warning stops.

To be honest, although I know for a fact that rail joints were staggered as in Figure 15, I have to admit of the possibility that the wiring shown might contain errors because I gleaned it from a verbal description. It is fairly clear and simple, though, so confidence is high.

Figure 15 probably represents the pinnacle of crossing design with interlocking relays. Although they work neatly under favorable conditions, and are elegant in their simplicity, they have a couple of safety disadvantages compared to the detection circuits with stick relays, which will be discussed in the next section. First, if a train looses shunt over the road after it has cleared the approach proper from once it came, and the interlocking relay on that side has a chance to pick up before shunt is regained, the warning is irrecoverably locked out for that train even if the shunt is re-established over the road. Also, if the train should clear the road, stop in the second approach, and back over the crossing, no warning will be received until after the train has fully blocked the road and finally reaches the first approach. (It's still arguably better than the scheme in Figure 14.) Second, the interlocking relay has a more fundamental limitation in that if a momentary failure occurs anywhere in the normally energized circuit associated with the opposite approach just as the train is about to enter the leading approach, and then the fault disappears after the train has entered the leading approach, it will block the warning from the approach that the train is actually on.

Stick Relays

To reduce the number of "activation failures" within the short section over the road, a third track circuit can be used that is dedicated to covering that short portion of track alone. It is referred to as the "island." Plain track relays are used for the island and both approaches, which are not mechanically interlocked in any way. The track relays now merely provide information about track occupancy and do not perform any special functions. The three track circuits are shown by themselves in Figure 16.

Although these three track circuits could work with the same insulated joint topology provided by Figure 15, having two joints on each side of the road provides better isolation. The joint topology of Figure 16 is more contemporary with the three track circuit arrangement.

The functionality previously achieved by the interlocking relay in earlier designs is covered by two "stick relays" and some fancier wiring. The stick relays are also relatively plain by design, which are named stick relays in their application because of how they are wired. They are controlled by some of the contacts on the track relays and some contacts of themselves.

Below is a PowerPoint slide show presentation contributed by a friend named Jerry showing the sequential operation of a simple crossing with stick relays. In the slides, the circuits for lights, gates, or bells are not shown, but remember they would be controlled by the master relay, XR, in the lower right corner as shown in the first section. The stick relays here are called ESR and WSR while the track relays are ETR, WTR, and XTR.

Slide Show Illustrating Stick Relays in Operation Note: Saving a local copy is recommended rather than trying to open the files through the internet. Choose the smallest file that can be viewed on your system.
File Type	File Size	File Name
PowerPoint Slide Show (Requires MS PowerPoint Software)	444 KB	stickshow.ppt
"Single-File Web Page" (MHTML File--Requires Capable Browser)	1.25 MB	stickshow.mht

In the wiring configuration of the slide show, the stick relays are energized as soon as the train enters the leading approach, and hence warning from the trailing approach is already locked out at that time. This approximates the behavior of the interlocking relay. However, one of the safety advantages of stick relays over the interlocking relays is they can be wired so that they don't energize until the train reaches the island. That way, if there is a momentary failure in the track circuit for the opposing approach just as the train enters the leading approach, the warning is still effective in both approaches because no stick relay has yet been energized.

Figure 17 shows a wiring scheme for not energizing a stick relay until the train is on the island. It is a photograph of an actual railroad circuit print, with the fully scanned print being available for download a little farther below.

Entire Crossing Print from an Actual Railroad Using Stick Relays
File Type	File Size	File Name
Web Page (Use this link to view the entire print without downloading the file.)	232 KB	xingprint.html
PDF (Use this link to download the entire print in reduced size with landscape orientation for easy printing on a single page. Requires PDF reading software.)	880 KB	xingprint.pdf
GIF Image (Use this link to download the raw image of entire crossing print.)	220 KB	xingprint.gif

In the center of the photo (Figure 18) is the master relay, XR, which controls the lamp circuits below it as described in the first section. The master relay is de-energized if any of the track relay contacts open, unless the open track relay contact is bypassed by a stick relay contact. Note that the contact of 7TR is always effective. The warning for trains on the island is therefore much more robust with a separate island track circuit.

Turning our attention toward the top of the diagram (Figure 19), we see the circuits whereby the stick relays are energized and de-energized. The sequence of events is explained below.

Suppose the train enters by the west approach. First, relay 6TR drops, but nothing else changes yet in Figure 19. (The signals are activated in Figure 18.) Next the train hits the island, and 7TR drops. Now current can flow from XB10 through contacts 6TR-1, 7TR-1, 8TR-1 and WXSR-3 to the coil of EXSR. The east stick relay is now energized. On Figure 18, the east approach warning by 8TR-2 is now bypassed by EXSR-4. With 7TR down and EXSR up, current can also flow to the coil of EXSR from another XB10 through contacts 7TR-2 and EXSR-2. When the train reaches the east approach, 8TR drops opening first supply path at for the EXSR coil at 8TR-1; however, a third path is then created from yet a third XB10, through contacts 8TR-1 and EXSR-1. When the train clears the island and 7TR picks up, the second current path to disappears, but EXSR is kept energized through the third path for as long as the train is on the east approach.

WXSR is never energized in the above example because its first supply path never comes for EXSR-3 being already open. The operation is symmetrical so that something similar would happen for a complete movement in the opposite direction, but WXSR would be energized instead.

In both stick relay schemes above, the stick relay remains energized until the train leaves the trailing approach. The system cannot tell the difference between a train that has stopped on the trailing approach and a failure that has occurred in the associated track circuit. Consequently, if that track circuit should fail as a train is leaving, there will be no warning on that approach even when a new train enters from that side because the circuits still think that the previous train is still leaving. To help alleviate this problem, a further safety improvement can be made by the introduction of stick cutout modules not included above. A stick cutout module is a timer that cuts out the stick relay after several minutes (say, 12 minutes) of the train's clearing the island. This does not fix the track circuit problem, but it unmasks it by allowing the signals to activate from the trailing approach after a short time. In normal operation, then, the train must then clear the area before time expires or the signals will re-activate until it does.

If a train were to stop without clearing the trailing approach and then actually reverse toward the crossing before time expires on the stick cutout, there will still be no warning before the train reaches the island. In practice, one would hope that the train crew is aware of this and that they account for it in one of three ways: 1) stop and let one of the crew protect the crossing as a flagman, 2) creep up in the train slowly enough that the island circuit starts detecting before the road is blocked, or 3) confirm that there are no cars near enough to be warned and go about their merry way. (The third solution is my favorite.) An automatic warning in this situation is something that only the constant warning time or motion detector circuits can circuits can provide in the next section.

Constant Warning Time Using Predictors

The relay circuits like the ones above can determine only if track circuits are occupied or not, and which ones are occupied first, etc. They cannot tell at what speed the train is approaching the crossing. They also cannot tell at what direction the train is going without assuming complete through movements. Some special relay circuits were used to get better information, but the information was still very much approximate. Relay circuits are essentially constant warning distance circuits.

It is not ideal to have to have a constant warning distance that a warning will be given within, but a constant warning time so that the signals will be activated no sooner or later than they need to be for all train speeds and directions. The warning should also stop if the train stops or backs away before reaching the crossing.

As a solution, some relatively modern circuitry uses the equipment configuration shown in Figure 20. The circuitry involved is too complex to show or describe in complete detail. Moreover, the author's understanding is limited (so beware). Hence, only the basic facts are given.

Transceivers (combination transmitter/receivers) located inside the Box put low frequency a.c. signals on the track. Some significant distance from the crossing, shunts are connected across the rails, which may simply lie on the track between them. The shunts are tuned shunts that behave approximately as short circuits at the frequency of the electrical signals placed on the track. As such, when no trains are present, practically all the current from the transceiver goes through the shunt that is tuned to it the same way that in a d.c. circuit practically all the current would go through plain wires placed across on the track.

The transceiver measures the voltage on the track and the current through it on a time base, from which the system can calculate the impedance and phase angle of the electrical load as seen from the crossing.

As the train enters the space between the transceiver and the shunt, the impedance drops. And it continues to drop more and more as the train gets closer to the crossing. The voltage decreases and the current increases (unless the current is regulated). In this way, the transceiver knows not only that the train is within the distance to the shunt, but how close to the crossing it is.

This relatively precise position information can be differentiated in a mathematical sense to determine the train's speed, and this in turn can be used to calculate when the train will arrive. The apparatus that performs these calculations is called a crossing predictor. An output from the electronic equipment feeds the coil of the master relay, and when it decides the time is right, it turns off this output to drop the relay and thus start warning the automobile traffic a preprogrammed number of seconds before the projected arrival time. The warning can be canceled if the train stops or reverses, and start again with movement toward the crossing.

Having a train on the island should make the signals active regardless of speed or direction. This is called presence detection. Theoretically, it might be possible to determine the presence of the island region from the same position sensing circuits described above, but there may be another special a.c. track circuit between points B and C to monitor it.

The a.c. track circuits for a predictor crossing do not usually need insulated joints to delimit the island or the approach sections of track. Where turnouts exist in the crossing area, however, an insulated rail joint is still needed in the diverging closure rail. An insulated switch rod is needed for the same reason.

Wayside signals may also exist on the line, whose train detection areas overlap with, but are usually not the same as, the detection area for the crossing predictor. The two types of equipment therefore must each have their own track circuits and yet share the rails by using different frequencies in conjunction with filters where necessary to block each other's electrical impulses. Furthermore, where insulated joints are placed for the wayside signals' benefit between the predictor (the Box) and the shunts, filters may be necessary to pass the transceiver's signals around them. (See Figure 21.)

A predictor or motion detector system has many self sanity checks to make sure everything from the track circuit to the predictor itself OK. For instance, the system knows what phase angle is appropriate for a given impedance. It responds to many faults by activating the signals until they are corrected. Consequently, because it is somewhat sensitive to faults, there is a possibility that crossing gates will be unnecessarily blocking a crossing for hours until the signal maintainer arrives and finds the problem. It is thus desirable to have a redundant system that allows automatic fail-over to the backup equipment when a problem occurs.

Wrap Circuits

The automatic fail-over feature mentioned above is available with new equipment today, but historically it was either unavailable or cost prohibitive. A more accessible but less integrated and less desirable solution is the use of wrap circuits.

Wrap circuits use track circuits that are analogous to the simpler d.c. track circuits used with interlocking or stick relays above, which can tell if a train is anywhere in the track circuit but not how close it is. When a crossing is operating from the wrap circuit, its performance at best reverts to something like constant warning distance circuitry. The warning given due to a train will be longer than usual, but at least it will not persist indefinitely.

A simple manner in which the predictor and the wrap circuits share control the master relay is shown in Figure 22. In this "wired-OR" configuration, either the wrap branch or the predictor branch of the circuit may energize the XR and clear the signals. Conversely, both must be de-energized to allow the XR to drop and a warning to be given.

Diagram Illustrating Wrap Circuit Control Principle

Figure 22--The principle involving control of the XR by either the wrap circuits or the predictor/motion detector circuits.

When there are no faults, the wrap branch stops feeding the XR whenever a train is in the track circuits that are part of the wrap, but the functioning predictor keeps it energized through its branch because it's not necessary to activate the signals quite yet. When the predictor feels that the time is right, it de-energizes its branch of the XR also, allowing the XR to drop out and the warning to be given. In the case where the predictor has failed, the predictor control branch is permanently de-energized, so the XR drops as soon as the wrap circuit branch de-energizes.

Unlike some of the most primitive crossing detection circuits, wrap circuits usually do not even have the logic to stop the warning after the train has cleared the crossing until it also clears the wrap detection area on the opposite side. The warning starts when the train is in the area and stops when it leaves: it's as simple as that, insofar as the predictor is nonfunctional. The motoring public is fortunate if the railroad line in question is dark territory (meaning no wayside signals for the trains themselves), because the track circuits in the wrap need be no longer than the length of the crossing approach on each side (say, 1,000 to 2,000 feet).

However, if the line is signaled, the situation is even worse because the wrap circuits do not use track circuits dedicated to the purpose, but instead rely on information from the track circuits used by the wayside signals, which are grossly longer than the approach length of the crossing and don't begin and end where convenient. (See Figure 23.)

Figure 23 shows two highway grade crossings on a railroad line having wayside signals. The location of track shunts (like those in Figure 20) are marked as "SHUNT," which represent the outer limits of the predictor's train detection on each side of that particular crossing. Within the shunts, a fully functional predictor may decide that the train is coming soon enough to activate the warning. However, since the wrap circuits depend on the track circuits of the wayside signals, we see that a failed predictor at the Spruce Street crossing would force the signals to activate as long as a train is somewhere in track section 4T--regardless of movement, speed, direction, or the fact that the rear of the train may have already cleared the crossing. Another crossing at Maple Boulevard is within the track section 8T. However, 8T does not cover the full length of the west approach to the crossing, and 10T must be included. A failed predictor at Maple Boulevard would force the signals to activate as long as a train is somewhere within either of the track sections 8T or 10T. The wayside track circuits might each be, say, 1-3/4 miles long.

From the illustration, it can be seen that a wrap circuit area wraps around a normal zone of detection, which is probably how they get named "wrap" circuits. It can also be seen that wrap circuits aren't very kind to the public, but they call attention to the predictor failure problem without blocking the crossing as much as it would be without any kind of backup at all. Railroads could use special, shorter track circuits that are designed just for the wrap, in spite of the wayside signals, as well as including some directional stick relays to cut the warning out after the train had cleared the crossing, but the author is not aware that this has ever been widely done. Apparently the fact that the wrap circuit represents a failure mode is enough to keep it from being at all fancy.

Motion Detectors

When an approaching train stops close to the crossing but yet outside the island section, it is appropriate for the warning to stop. However, when the train starts again, it is necessary to activate the warning immediately in order to have enough time for gates to be lowered and for people to get out of the way. (Sometimes, when the train starts very close to the crossing, the gates are not even fully lowered by the time the train gets there.) In this case, it is not useful to calculate the arrival time of the train but only to have the signals respond immediately upon train movement. This is the behavior of a motion detector; it activates the signals when the train is moving toward the crossing and de-activates them when it has stopped on the approach or is moving away.

A good predictor would in the same case calculate that the signals needed to activate immediately, so the motion detector has no theoretical advantage over a predictor that the author is aware of. Probably the only reason why motion detectors are used instead of predictors is that a predictor is more complicated and hence expensive, since a predictor must calculate when the train will arrive. When a train enters the approach, the activation is immediate regardless of train speed. This has the disadvantage that for complete, through movements the behavior is the same as that for a constant warning distance circuitry. The motion detector has great advantages over constant warning distance with stick relays, though, in that it can stop the warning if the train stops or reverses on the leading approach; it also has the safety advantage that it can restart the warning if the train reverses toward the crossing on the trailing approach. For these reasons, a motion detector may be used at a location where most crossing activations are the result of slow speed switching moves where stopping, starting, and reversing near the crossing are likely.

The motion detector uses a setup almost identical to that of a crossing predictor, using track shunts and wrap circuits. Both are boxes about size of an automotive battery charger and full of electronics. Because the main difference between them is behavioral, some things about a predictor installation can be learned from looking at motion detector installation and vice versa.

Below is a simplified diagram of a motion detector setup, which is inspired by prints for an actual motion detector installation. The key point to notice is that the outputs of the box are connected to relays MDR and MD-ILR, which tend to control the master relay, XR. MDR drops out whenever motion toward the crossing is detected inside the distance to the shunts. MD-ILR drops out whenever a train is present on the island, even if no motion is detected. The transmitter and receiver straddle the island. The seemingly redundant track wires are check wires that activate the signals if any of the transmitter or receiver wires to the track should break.

The relay here designated TPR represents the sum total of the wrap circuit. Observe that Figure 24 bears some resemblance to Figure 22. In this particular case the territory is dark, so the track circuits used by the wrap are designed specifically with the crossing in mind, and the detection area is no larger than the approach lengths as defined by the track shunts.

This crossing is unusual because of the situation presented by the stick relay, MDSR, and the 1-minute time element relay, TER. Here's how it works:

When a train enters the approach (here the same as the wrap), TPR drops out and breaks the control of the XR through the top branch of the circuit. At this point, the energization of the XR (or lack thereof) depends on lower branch of the circuit through contacts of MDR and MD-IL. The train causes the motion detector to sense motion and drop MDR, opening the lower branch as well. The crossing is activated. If the train should stop on the approach, the MDR would pick up again and tend to energize the XR. However, in this particular design the supply of battery hasn't been established yet in the lower branch because it comes through contacts of MDSR and TER, so XR stays down.

MDSR will pick up if the train is moving when it reaches the island, although the XR won't be energized immediately because MD-ILR is down at that point for the train being on the island. TER will pick up if the MDSR doesn't energize after one minute of the train being on the wrap. In summary, the wrap basically controls this crossing until the train reaches the island, unless the train has once been on the island already or the train has not reached the island after a full minute of being on the approach. This curious behavior somewhat approximates ESR and WSR stick circuitry for complete through movements, although it behaves as a motion detector when the train is performing switching operations.

Time element relays are sometimes used for momentary loss of shunt protection. Another clue that might help to describe why this circuit behaves the way it does is that the model of motion detector used in this design had trouble detecting trains at speeds less than 3 mph. Train crews complained that they were on the island before the gates came down. They needed to increase speed to increase safety! Anyway, the pickup circuit for MDSR tends to verify that the motion detector is working under the given conditions before allowing the motion detector to energize the XR.


Figure 11--Interlocking or Crossing Relay.	Figure 12--Interlocking elements within relay.

How Railroad Crossing Signals Work