[Note that this article is a transcript of the video embedded above.]
If you’ve ever ridden a bike, driven a car, or operated pretty much any other vehicle on earth, there’s a fact you’ve probably taken for granted: you can see farther than it takes to stop. Within the span between seeing a stationary hazard and colliding with it, you have enough time to recognize it, apply the brakes, and come to a stop to avoid a collision. Your sight distance is greater than your stopping distance; it sounds almost silly, but this is a critical requirement for nearly all human-operated machines. But it’s not true for trains.
Engineers can see just as far as the rest of us, but the stopping distance of a fully laden freight train can be upwards of a mile. That means if an engineer can see something on the tracks ahead, it’s often already too late. So, trains need a lot of safety infrastructure to make up for that deficiency. For one: trains almost always have the right-of-way when they cross a road or highway at the same level, or at grade. The cars have to wait; And, we use a litany of warning devices at grade crossings to enforce that right-of-way and try to prevent collisions. In most cases, these devices have to detect the impending arrival of a train and give motorists enough time to clear the tracks or come to a stop. It sounds simple, but the engineering that makes that possible is, I think, really interesting, and of course, I built some demonstrations to help explain. This video is part of my series on railroads, so check out the rest after this if you want to learn more! I’m Grady, and this is Practical Engineering. Today, we’re exploring how grade crossings work.
It’s inevitable that roads cross railroad tracks, and it’s just not feasible to build a bridge in every case. In the US alone, there are over 200,000 grade crossings where cars and trains must share the same space. A car to a freight train is an aluminum can to a car: in other words, there’s a pretty big disparity in weight. So we’ve put a lot of thought into how to keep motorists, cyclists, and pedestrians safe from the trains that can’t swerve or stop for a hazard. You’ve probably stopped for a train at a crossing, but you may not have consciously added all the safety features up.
Of course, the locomotives at the front of trains themselves have warning devices, including bells, bright headlights, smaller flashing ditch lights, and most noticeably, the blaring horn. The standard pattern at a crossing is two long blasts, one short blast, and one final long blast. But the crossing has warnings too. Passive warning devices don’t change with an approaching train. They include a stop or yield signs, the crossbuck, which is the international symbol for a railroad crossing, and sometimes a plate saying how many tracks there are so you know whether to look for one train or many. Another crossbuck is usually included as a pavement marking to make sure you know what’s coming up. Many low-traffic crossings have only passive safety features, leaving it up to the driver to look out for trains and proceed when it’s safe. But, many crossings demand a little less margin for error. That’s when the active warning devices are installed.
A typical grade crossing features both visual and audible warning signals that a train is coming: red lights flash, a mechanical or electronic bell sounds, and usually a gate drops across oncoming lanes. That seems pretty simple, but there’s quite a bit of complexity in the task and the consequences if anything goes wrong are deadly. And the first part is just knowing if a train is coming.
Detecting a train is important for grade signals (it's also important for signaling trains about OTHER trains, but that's a topic for another video). It can be handled in a bunch of ways, but the simplest take advantage of the electrical conductivity of the steel rails and wheels themselves. A basic track circuit runs current up one rail, through a device called a relay I’ll explain in a minute, and back down the other rail. When a train comes along with its heavy steel wheels and axles, it creates a short circuit, a preferential path for the current in the track circuit. That deenergizes the relay, triggering all the connected warning devices or signals. But why use an ordinary old diagram when you have a model tank car, and an old railroad relay you got off eBay? Let me show you how this works in a real demonstration.
On the left, I’ve hooked up a power supply to the tracks, putting a voltage between the two rails. On the right side, I’ve attached a relay. Let’s take a look inside it to see what it does. I love playing with stuff like this. At its simplest, a relay is just an electromechanical switch: a way to turn something on or off with an electrical signal. When I energize the coil (at the bottom), it acts as an electromagnet, pulling a lever towards it. On the other side of the lever, you can see the movement interacting with several electrical contacts. It’s a little tough to see here, but these contacts are like switches that can control secondary circuits. Some will be switched on when the relay is energized, and others are switched off. When the relay is energized or de-energized, it basically flips the switch on these circuits, allowing various devices, like lights, bells, and gate arms, to be activated or deactivated. In my case, I have a simple battery and LED to indicate whether or not a train is being detected on the rails.
When there’s no train, current passes through the relay from one rail to the other, energizing the coil and holding the switch open so the LED stays dark. When I put a railcar on the tracks, the circuit changes. The wheels and axles create a short circuit (or shunt), a low-resistance path for current to flow, essentially bypassing the relay. The coils in the relay de-energize, closing the switch and lighting the LED to warn any nearby tiny drivers that a train is present on the tracks. It all depends on the train giving a preferential current path, which can be a problem if there are leaves or rust on the rails. You can see how shiny and clean tracks look when they’re in frequent use. Tracks that haven’t seen a train in a day or more often impose a speed restriction on the first train just in case there is rust that could affect the track circuits along the way.
If all this circuitry seems a little convoluted to simply detect the presence of a train, it’s because of how this simple track circuit handles when things go wrong. Let’s say the track circuit loses power; what happens? The relay deenergizes and falls back to the safest condition: assuming a train is occupying the tracks. Same thing if a rail cracks or breaks: the relay deenergizes and the light comes on. This is called failsafe operation, or as the engineers prefer to call it: fail to a known condition. If anything goes wrong, we want the default assumption to be that there’s a train coming because it might be true. Fail safe operation isn’t just in the track circuit but the warning devices too. Gates are actively held up with a powered brake. If power is lost, they fall just by gravity alone. And the bells and lights are usually powered by banks of batteries that can last for hours or days. Most modern train detection systems have moved to more sophisticated equipment, but relays are still used around the world because of their reliability. In fact, this is called a “vital” relay because of all the features that make it extremely unlikely to fail. You can see it acts slowly so that the inevitably noisy signal of a train shunting the tracks can’t cycle it on and off over and over; The armature assumes the de-energized position even if the spring breaks; The contacts use special materials to keep from welding together; And they’re just really robust and beefy to make sure they last for decades.
But even though assuming a train is coming is the safest way to manage problems, it’s not without its own challenges. Warning devices depend on trust, and that’s an extremely tenuous confidence to ask of a motorist. We are naturally dubious of automated equipment. Every time a grade crossing activates and no train comes, that trust is eroded, making motorists more likely to drive around the gates. So failing safe isn’t enough; we also need to make sure that failure is rare. Current leaking between the tracks through water, plant growth, or debris can falsely trigger warning devices. So railroads put a lot of time into keeping tracks clean and the coarse gravel below the tracks (called ballast) freely draining to prevent water from pooling up. In addition, even though maintenance workers can manually trigger devices by shunting current across the tracks, this is done rarely to avoid impacts to road traffic.
But maybe you’ve spotted a flaw in this simple track circuit. If not, let me point it out. It’s all to do with where you put the boundaries. If the circuit is close to the crossing on either side, there’s no warning time. By the time the train is detected, the motorists wouldn’t be able to clear the intersection or come to a stop. But if the circuit extends far enough beyond the crossing to give adequate warning time, motorists will have to sit and wait well after the train is past before it comes off the track circuit and the warning devices turn off. So, instead of a single track circuit, most crossings use three: two approaches and an island. Let me show you how this works with another demo.
Now I have three track circuits set up with power going to each one. The rails are separated by a small gap to avoid an inadvertent connection across the circuits. On actual railroads, you can often identify insulated joints used to isolate the track circuits. They can be hard to distinguish if the insulating material matches the profile of the rail itself, but they’re often painted to be easy to spot. A three-circuit configuration requires a little bit of logic to decide when to turn on the warning devices and when to turn them off. So, despite the fact that I have the coding skills of a civil engineer, I put this demo together using an arduino microcontroller. The model railroad folks are surely shaking their heads at this. You can see my LEDs as I roll the train along the tracks indicating which of the circuits is detecting the presence of a train; from approach to island to other approach. And here’s how the logic works.
When a train is detected on either approach circuit, it immediately activates the warning devices. The lights flash, bell sounds, and gates drop. As the train keeps moving toward the crossing, it’s detected on the island circuit too. The circuit effectively takes over control of the warning devices. They’ll stay on for as long as a train is occupying the island circuit. But as soon as the island is unoccupied, the warning devices turn off (even though one of the approach circuits is still detecting a train). You can see how just a little bit of logic makes it possible to give some warning time for motorists before the train arrives at the intersection without keeping them stuck behind gates after the train has passed. But, how much warning time is enough?
In the US, the minimum requirement is 20 seconds between activation of the warning devices and the arrival of a train, but it’s typical to see 30 or 45 seconds. You might think that the more warning time the better, but it’s a balance. Too much warning time, and motorists might become impatient and drive around the gates, so more time can actually make crossings less safe. For the three-circuit example in the demonstration, the only control you have over warning time is where to start the approach circuit. The farther away from the crossing it begins, the more warning time you get. But the exact time depends on the speed of a train. Since the approach is fixed in place, a slow train will provide lots of warning time, and a fast train will provide less. And a train stopped on an approach circuit before it even reaches the crossing will hold the gates down indefinitely. So the next step in grade crossing complexity takes speed into account.
I put a little acoustic distance sensor on my arduino so I can try to estimate the speed of an oncoming train. The large cardstock cutout just helps my sensor to ‘see’ the train a little better. The arduino measures the distance over time, converts that to an approximate speed, and guesses how long it will take the train to arrive at the crossing. If the expected arrival time is longer than the warning time I programmed in, nothing happens. But if an arrival is expected within the warning time, the devices are activated.
You can see if I approach the intersection slowly, the gates don’t drop until I’m relatively close to the crossing. And if I speed things up, the gates drop when I’m farther away, anticipating the faster arrival of the train. In theory, this type of sophistication means that the warning time at a crossing will always be the same, no matter the speed of the train. But it doesn’t just solve that problem. If you have ever sat at a railroad crossing while a train is stopped on the approach circuit, you know the frustration it causes. A grade crossing predictor avoids the issue. You can see as I move my train toward the crossing, the devices activate assuming the train will cross. But when I stop short, the predicted arrival time goes effectively to infinity, and the controller opens the gates back up.
Of course, actual crossings don’t use sonar to predict the speed of a train. In most cases, they use track circuits with an alternating current. A train interacts with the frequencies of the circuit as it travels along the rails, giving the sensors enough information to detect the presence and speed. Sometimes you can even hear these frequencies since they’re often in the audible range. AC track circuits are also used for electric train systems because they are less susceptible to interference from the traction currents in the rails used to drive the trains.
Another challenge with grade crossings happens in urban areas where signalized intersections are present near the railway. Red lights form a line of vehicles that can back up across the tracks. You should never drive over a railway until you know it’s clear on the other side. But, if you’re not paying attention, it can be easy to misjudge the available space and find yourself inadvertently stopped right on top of the tracks. Traffic signals near grade crossings are usually coordinated with automatic warning devices. When a train is approaching, the signal goes green to clear the queue blocking the tracks.
Equipment for the most basic track circuits to the most sophisticated, including relays, microcontrollers, backup batteries, and more are usually housed in a nearby bungalow or cabin that is easy to spot. In the US, every grade crossing has its own unique identifier, and they all have a phone number to call if something isn’t working correctly. Railroads take reports seriously, so give them a call if you ever see something that doesn’t look right. If you want to see a lot of these grade crossing systems in action, check out my friend Danny’s channel, Distant Signal, for some of the best railfan videos out there. We depend on trains for a lot of things, and in the opinion of many, we could use a few more of them in our lives. Despite the hazard they pose, trains have to coexist with our other forms of transportation. Next time you pull up to a crossbuck, take a moment to appreciate the sometimes simple, sometimes high tech, but always quite reliable ways that grade crossings keep us safe.
