Subway Signals: Interlocking
Interlocking is defined as the, or an, "arrangement of signals and switches such that they are constrained to be operable only in a safe order." An interlocking is a place where tracks join (switches), and denotes the switches, the surrounding signals, and the control machinery which connects them and enables their operation (by a tower operator (formerly towerman), or sometimes automatically) and ensures that that operation is safe, regardless of the operator's actions. Interlockings are designed such that there is no way the controls can be operated, by mishap, malice, or contrivance, to effect an unsafe situation on the tracks. One speaks of "Coney Island Creek Interlocking", or "Grand Central Interlocking", to denote the complex arrangement of tracks at those places, the signals in and around it, the tower there with the interlocking machine that operates it, and all the electrical logic that makes them work together as an interlocking.
Interlocking is the most complex area of the science and technology of signalling. A very large part of the cost of designing, constructing, and maintaining signalling, in New York or anywhere else, is that of its contained interlockings, and a large amount of what a train operator has to learn about signals concerns the signalling at interlockings.
Up until the 1950's, almost all the interlockings in the New York system were controlled from interlocking machines that had mechanical levers physically interlocked with each other (hence the name) to prohibit unsafe configurations. It is best to start learning about interlocking by describing such machines, for not only do many still remain in use, but their underlying models of safety and operation apply to all railroad and subway interlocking, including the latest computer-controlled interlocking machines, to this day. (There are also machines with small electrical switches for levers, which are not mechanically constrained at all, but whose operation is merely ignored under unsafe circumstances).
The levers controlling switches and signals are arranged in a row, and each has a number, which also appears on the actual signal or switch, and serves to identify it when discussing the interlocking. Above the machine is a model board showing the track layout, with red lamps at various points in the tracks to indicate track occupancy. The levers may be equipped with lights indicating when they are free to be operated (not locked).
To the left is a photo of such a machine, at Myrtle Avenue on the BMT Broadway-Jamaica line. Click here it to see it in full. (Photo by Peter Dougherty). The interlocking machine is at the left of the image. The model board, which has been removed from the top of the machine because this tower is no longer in service, is on the floor on the right. Please click here for a photo album of this interlocking.
On a lever machine, each switch (or pair of switches constituting a crossover) is controlled by a lever that can (under permissible conditions) cause that switch to move. A switch lever can be in the "normal" or "reverse" position, which moves the switch to its own normal ("straight track") or reverse ("switch tracks") position respectively. As with all levers, the the switch lever is mechanically interlocked with other levers to prevent their or its own operation under unsafe conditions.
Each signal that is part of the interlocking also has a lever (sometimes, especially on the IRT, more than one signal is controlled by a single lever). A signal lever, however, cannot make a signal clear (be other than red, and thus allow trains to pass): it can only permit it to clear when and if track conditions permit. The lever can be in the "reverse" position, permitting the signal to clear, or in the "normal" or "restored" position, which forces the signal to indicate "stop" unconditionally. That is, in addition to the track ahead being clear and possible time controls, interlocking imposes additional constraints on a signal's clearing.
Consider the accompanying illustration-qua-plan of a hypothetical, very small interlocking. The thick lines represent tracks. The colored circles represent signals, their lever numbers being shown adjacent to them. The signals with lever numbers and two heads (circles) are home signals, those with one head and lever numbers approach signals, and the small ones with no lever numbers automatic signals. The thin lines represent the control lengths of the signals, i.e., the distance downtrack from the signal within which the presence of any obstruction (including both trains and switches set the wrong way) will force the signal to be red. The box with the dot and the line represents the tower -- the line is the interlocking machine, and the dot is the operator (sometimes it is important to know which ways are the operator's left and right). The other numbers along the tracks identify the track circults (block numbers), between the insulated joints represented by the vertical ticks subdividing the tracks.
Simple Ave. Interlocking diagram. Please note that even this is a simplification -- were this interlocking to exist with the signals shown, even more home signals, and/or coordination with the next towers down-track in each direction (traffic locking) would be needed to ensure safety.
We can see from the state of the signals and switches that the levers for signals 4 and 8 are reversed (signal may clear), and all others restored to normal. The lever for switch 7 must be in the reverse position, as switch 7 is shown reverse. In this situation, none of the levers for signals 2, 6, 10, and 12 would be free to move -- they would be locked in their normal positions. So would the lever for switch 7, because signals 4 and 8 have cleared movement over it. Were the tower operator to restore the levers for signals 4 and 8, forcing those signals to be red, the lever for switch 7 would free up and could be moved. So would the levers for signals 10 and 12. But the levers for signal 2 and 6 would still be locked, because switch 7 is still set against them. When switch 7 is moved, signals 2 and 6, or 10 and 12 can be cleared, but no longer 4 or 8. Only one of the levers for the home signals 6, 8, and 10 can be out of the restored position at once (As a matter of fact, on the "old IRT", more specifically, with electropneumatic lever machines manufactured by Union Switch & Signal, what are here 6, 8, and 10 would all be controlled by one lever, which could move right or left ("restored" in the center), and would be labelled 6Ra, 6Rb, and 6L, respectively, what is here 10's two heads being 6La and 6Lb.)
That is just the bare-bones basics. In fact, the levers have positions other than normal and reverse: for instance, when you start to move a switch's lever, it will not go "all the way", unlocking other levers, until the switch has actually moved, and proven it is not broken or stuck. The signal levers, too, can be pushed in at any time (forcing the signal to stop), but will not fully restore (and thus unlock other levers) if there is a train coming at the signal, or the signal, for whatever reason (e.g., a stuck semaphore or relay) is not actually indicating "stop". And the presence of a train on any switch will lock its lever completely. Train stops are involved in locking the switches, too. And there is much more --- the design of real interlockings is terrifically complicated, even today. In fact, railroad interlockings, like telephone switches, were one of the early ancestors of the modern digital computer.
New York Subway interlockings often feature Train Identification Pushbuttons, trackside, train operator-level boxes at the station before an interlocking, with large, round, red buttons, more or less menus by which the train operator identifies his or her train, that is, requests its route at the interlocking it is approaching. The interlocking maintains the correspondence of route requests and trains via careful monitoring of block occupancy in such a way that the tower operator, or an automatic interlocking, can know the identity of the next train as soon as possible (even if there are many trains or miles between the buttons and the interlocking).
Instead of machines with physical levers, all subway interlockings installed or modernized since the 1950's use the Entrance-Exit (NX/UR) scheme, whereby pushbuttons at the locations on an enlarged model board corresponding to signals on the real track replace the aforementioned levers. To cause the signals and switches to set up to route a train from point A to point B, the tower operator need only press the button at point A (the entrance) and then the button at point B (the exit), and the interlocking automatically figures out and does the rest. Although physical levers are not employed in the NX/UR scheme, the internal logic of the interlocking, and the tower operator's conception of it as well as that of the interlocking itself, is still in terms of levers which still must be operated (albeit automatically) according to rules such as those illustrated above. All of the lever machine principles apply to this day without exception. Switches and signals still have and are referred to by lever numbers.
NX/UR interlockings facilitate central traffic control, which means one very big NX/UR control panel (examples: Grand Central (Lexington), City Hall (BMT Broadway)) can control many interlockings up and down the line via remote control. At left is a photo (by David Pirmann) of a very large NX panel at Coney Island Yard. Click here for a full-size version.
At some places in the system (e.g., DeKalb Avenue in Brooklyn), there are modern digital computers responding to train identification pushbuttons by operating the interlockings, scheduling trains according to a predefined timetable. But even this is but one more added "layer" on top of the two just described, NX/UR operation, and the classic lever interlocking model under that: the computer literally simulates pushing of entrance and exit buttons on the NX/UR panel, which, in turn, operates imaginary levers simulated by relays.
If the subject of interlockings, classical relay-logic electrical computers that implement railroad signalling, fascinates you, or you just wish to learn more about more details of NYC Subway Interlocking, including things glossed over above or not mentioned at all, please check out NXSYS. The software available there (for free download) allows you to operate and experiment with fully-simulated New York-style NX/UR interlockings.
Subway Signals: A Complete Guide
Approach, Automatic, and Marker Signals | Train Stops | Time Signals | Interlocking | Home Signals
Sign Signals | Holdout Signals and Bidirectional Traffic | Single-Line Signal Diagrams
NXSYS, Signalling and Interlocking Simulator
Les signaux du New York City Subway
Descriptions and graphics on this page (unless otherwise noted) are Copyright © 1997-2002 Bernard S. Greenberg (contact).