Signaling of the New York City Subway

Block signaling

The New York City Subway system has, for the most part, used block signalling since its 1904 opening. As of May 2014, the system consists of about 14,850 signal blocks, 3,538 mainline switches, 183 major track junctions, 10,104 automatic train stops, and 339,191 signal relays. Trains used to be controlled by signal towers at interlockings, but this was eventually phased out in favor of master towers. Eventually, these master towers were replaced by a single rail control center: the New York City Transit Power Control Center in Midtown Manhattan.

The New York City Subway generally distinguishes its current signals into:

Common automatic and approach signals consist of one signal head showing one of the following signal aspects:

Where different directions are possible, the subway uses both speed and route signalling:

The system is old, and some replacement parts must be custom built for the MTA, as they are no longer available from signaling suppliers. Additionally, some subway services have reached their train capacity limits and cannot operate extra trains with the current Automatic Block Signaling system. Changes in assumptions about train performance, and changes in operational rules also had notable impact on system throughput, such as the prohibition of movement of trains past permissive signals at restricted speed (No Key-By without permission from Control Center.) As of May 2014, the system consists of about 14,850 signal blocks, 3,538 mainline switches, 183 major track junctions, 10,104 automatic train stops, and 339,191 signal relays.

Types of block signals

The system currently uses block signalling, which is used in other systems such as the Toronto subway and RT. The block signals that the New York City Subway currently uses is identical to those on the RT's signaling system.

The system also has automatic and manual key-by red lights. They involve the operation of an automatic stop with an automatic or manual release, then a procedure with caution, with preparations to stop in case of debris or other obstructions on the track.

Speed control on the subway is ensured by "Time Signals". A timer is started as soon as the train passes a certain point and will clear the signal ahead as soon as the predefined time elapsed; the minimum time is calculated from the speed limit and the distance between start of timer and signal. "Time Signals" are distinguished into "Grade Timer" for speed supervision at grades, curves or in front of buffer stops, and in "Station Timer" to allow a following trains enter a station as another is leaving as long as they are going at a reduced speed. There are two types of Grade Time signals. The first time, "two-shot timer," are generally used on down grades where the train must be under a set speed for a longer length of track. "One-shot timers," on the other time are found at sharp curves.

Another addition to the system's transit signals are "Wheel Detectors." First introduced in 1996 at interlockings, they further enforce the speed in which a train is traveling through an interlocking, and they are only active when a switch is set to the divergent route.

"Gap Filler Signals" are used at the 14th Street–Union Square and South Ferry Loop stations, where gap fillers extend out from the platforms to bridge the space between the platform and the car body and door at the curved stations. When the signals are red, the gap fillers are extended, and when the red light is no longer lit up, the gap fillers have been retracted, and the train operator can increased the speed of the train and leave the station.

Aside from some parts of the original IRT system, the entire subway uses this signalling system. On the only line that still has the IRT-type signals, the IRT Dyre Avenue Line, the lights, from bottom to top, are yellow, red, and green. On the rest of the system, the lights from top to bottom, are red, yellow, and green.

These signals work by preventing trains from entering a "block" occupied by another train. Typically, the blocks are 1,000 feet (300 m) long, although some highly used lines, such as the IRT Lexington Avenue Line, use shorter blocks. Insulators divide the track segments into blocks. The two traveling rails conduct an electric current, as they are connected to an electric current. If the circuit is closed and electricity can travel across the rails without interruption, the signal will light up as green, as it is unoccupied by a train. When a train enters the block, the metal wheels interrupt the current on the rails, and the signal turns red, marking the block as occupied. The train's maximum speed will depend on how many blocks are open in front of it. However, the signals do not register the trains' speed, nor do they register where in the block the train is located. If a train passes a red signal, the train stop automatically engages and prevents the train from moving forward.

Until 1970, when a train paused at a red block signal, the train operator was allowed to pass the red signal using a procedure called "keying by." The train operator would pause at the red signal, and then the operator would get out of the cab, descend to the trackside, and crank down the track trip with a key-like device. After a series of accidents, in which train operators keyed by and and crashed into trains in front, the procedure was made against the rules, unless permission was granted by the train dispatchers.

Types of interlocking signals

Interlockings are places where tracks are connected by switches and where routes may conflict. They are arranged in a specific way, with switches and signals that prevent conflicting movements once one route is set up. Interlocking signals are fixed signals within an interlocking, containing two separate red-yellow-green signal heads and often other indications. A home signal is defined as an interlocking signal at the entrance to a route or block to control trains entering that route or block. These terms are used interchangeably, as most interlockings only have one controlled signal on each track.

They are controlled by human operators in a signal tower near the switches, not by the trains themselves. A train operator must use a punch box, which is located besides the cab window in the station closest to the interlocking, to notify the switch operator of which track the train needs to go. The operator has a switchboard in their tower that allows them to change the switches.

Interlocking signals also tell switch operators which way switches on the subway are set. The top part of an interlocking signal indicates the condition of the block ahead, while the lower part indicates the route selected. The following interlocking signals are used on the New York City Subway:

Until the DeKalb Avenue reconstruction on the BMT in 1958, there was a unique situation in which there was a three-way switch. As there were three possible options, a special blue signal was used.

Another type of signals are dwarf signals, which are often used at switches to allow for occasional movements against the usual direction of train traffic. Controlled manually from a tower, they are not part of ordinary operations, and they generally do not include trip arms.

Chaining

To precisely specify locations along the New York City Subway lines, a chainage system is used. It measures distances from a fixed point, called chaining zero, following the twists and turns of the railroad line, so that the distance described is understood to be the "railroad distance," not the distance by the most direct route ("as the crow flies"). This chaining system differs from the milepost or mileage system. The New York City Subway system differs from other railroad chaining systems in that it uses the engineer's chain of 100 feet (30.48 m) rather than the surveyor's chain of 66 feet (20.1168 m). Chaining is used in the New York City Subway system in conjunction with train radios, in order to ascertain a train's location on a given line.

Automatic Train Supervision

The New York City Subway uses a system known as Automatic Train Supervision (ATS) for dispatching and train routing on the A Division (the IRT Flushing Line, and the trains used on the 7 <7> services, do not have ATS for two reasons: they are isolated from the main-line A Division, and they were already planned to get communications-based train control (CBTC) before the ATS on the A Division, or ATS-A, project had started). ATS allows dispatchers in the Operations Control Center (OCC) to see where trains are in real time, and whether each individual train is running early or late. Dispatchers can hold trains for connections, re-route trains, or short-turn trains to provide better service when a disruption causes delays. ATS is used to facilitate the installation of train arrival displays, which count down the number of minutes until a train arrives, on the A Division and on the BMT Canarsie Line. ATS was first proposed for the BMT Canarsie Line and the A division in 1992, after a 1991 derailment killed five people on a 4 train that derailed near the 14th Street – Union Square station. CBTC for the Canarsie Line was proposed two years later.

The deployment of ATS-A involved upgrading signals to be compatible for future CBTC retrofitting, as well as consolidating operations from 23 different master towers into the Power Control Center. Parsons Corporation helped the MTA install the system on the 175 miles (282 km) of A Division track, as well as did some preliminary planning for ATS on the B Division. The project, valued at $450 million, cost $200 million. The project's completion was delayed by five years, and it ultimately took 14 years to implement ATS-A. The long duration of ATS-A's deployment was attributed to bad communication between workers and contractors; because of the uniqueness of the New York City Subway's interlocking equipment, which necessitated extra workarounds; because of the MTA's use of its own workers rather than of contractors' workers; because of the poor training that contractors had; and because of bad communication interfaces. In addition, the MTA kept missing deadlines for testing ATS. The single biggest issue during the project, however, was that MTA and the contractors did not cooperate well.

In 2006, 2008, and 2010, the MTA considered upgrading the B Division to ATS, but dismissed the proposal because it was too complex and would take too long. However, the MTA stated that due to high customer demand for train arrival displays, it would use a combination of CBTC and a new system, named the "Integrated Service Information and Management" (abbreviated ISIM-B). The simpler ISIM-B system, started in 2011 would essentially combine all of the data from track circuits and unify them into digital databases; the only upgrades that were needed were to be performed on signal towers. Originally slated to be completed by 2017, ISIM-B was later delayed to 2020.

Automation

Trains using CBTC locate themselves based on measuring their distance past fixed transponders installed between the rails. Trains equipped with CBTC have a transponder interrogator antenna beneath each carriage, which communicates with the fixed trackside transponders and report the trains' location to a wayside Zone Controller via radio. Then, the Controller issues Movement Authorities to the trains. This technology upgrade will allow trains to be operated at closer distances, slightly increasing capacity; will allow the MTA to keep track of trains in real time and provide more information to the public regarding train arrivals and delays; and will obviate the need for complex interlocking towers. The trains are also equipped with high-tech computers inside the cab so that the conductor could monitor the train's speed and relative location. The wayside controllers themselves are located in enclosed boxes that can withstand floods and natural disasters. The traditional block systems will remain on these lines despite the installation of CBTC.

The block system handles all control and supervision of routes through interlockings including switch (point) control and switch status, for broken rail protection, and tracking of trains with failed (or not equipped for) CBTC. CBTC lines are fully track circuited with power frequency, single-rail track circuits. (Broken-rail protection is only guaranteed on one of the two rails, however.) Equipment aboard every train identifies the location of the train using wayside transponders as a basis. Once the Zone Controller has determined, based on track circuit information and train localization, that the CBTC train is a single discrete train ("sieved"). The Wayside Zone Controller uses this information to grant movement authorities based on conditions ahead. The CBTC Zone Controller functions then as an overlay which only provides safe separation of trains and cannot do so without interaction from the Wasyide (Legacy) Signaling system. Trains, with CBTC, can then operate closer together, although as before, platform dwell times and train performance are the true limiting factors in terms of headway performance. With the new system, signals and interlockings are still absolutely required, their job being done better by relay interlockings or Solid State Interlocking controllers. The ATS system at the Control Center is not a vital (life-safety) system and serves only to automate the routing of trains based on the overall timetable. The location of the train is also used to inform passengers of arrival times. The MTA's form of CBTC uses a reduced form of the old fixed-block signaling system, requiring that both be maintained at high cost.

Only newer-generation rolling stock that were first delivered in the early 2000s—the R143s, R188s and 64 R160s (8313-8376)—are equipped for CBTC operation. Future car orders, specifically the R179 and the R211, will also be designed to be CBTC compatible. After the retirement of the R68 and R68A cars, all revenue cars, except those on the G, J, M, Z and S trains, will be equipped with CBTC. The BMT Canarsie Line was the first line to implement the automated technology, using Siemens's Trainguard MT CBTC system.

Most subway services are already at capacity, in terms of train spacing, during rush hours, except for the 1, G, J/Z, L, and M trains (the L service already is automated with CBTC). Therefore, transit planners are viewing the installation of CBTC as a way to free up track capacity for more trains to run, and have shorter headways between trains. However, installing CBTC in the New York City Subway is harder than in other systems due to the subway's complexity. The MTA hopes to install 16 miles of CBTC-equipped tracks per year, while the Regional Plan Association wants the MTA to install CBTC signals on 21 miles of tracks per year.

However, even without CBTC, the system is currently retrofitted to operate at frequencies of up to 60 trains per hour (tph) on the IND Queens Boulevard Line (30 tph on each of the local and express pairs of tracks made possible by the Jamaica–179th Street terminal, which has four sidings past the terminal for each set of tracks) and 33 tph on the IRT Flushing Line. The BMT Canarsie Line is limited to a 26 tph frequency due to the bumper blocks at both of its terminals; however, the IRT Lexington Avenue Line operates at frequencies of 27 tph without CBTC. By contrast, lines on the Moscow Metro can operate at frequencies of up to 40 tph, since lines in the Moscow Metro, unlike most of the New York City Subway (but like the Jamaica–179th Street station), typically have four sidings past the terminals instead of bumper blocks or one or two sidings.

42nd Street Shuttle automation

The 42nd Street Shuttle, which runs from Grand Central to Times Square, was briefly automated from 1959 to 1964. The chairman of the Board of Transportation, Sidney H. Bingham, in 1954, first proposed of a conveyor belt like system for the shuttle line. Charles Patterson, a few years later, as the President of the newly formed New York City Transit Authority (NYCTA) told of a vision of automated mass transit, without relying on the use of motormen. General Electric responded to Patterson's speech, stating that this technology was feasible, and that the company was interested in the idea of automating the New York City Subway.

The idea of automation at that time relied on commands that were sent to the train while the train is at a station, to keep its doors open. When the commands cease, the doors would promptly close. A new series of commands would start the train and gradually accelerate it to 30 miles per hour (48 km/h), maintaining that speed. This is only under the condition that no other command overrides it. When approaching the next station, there was an insulated rail joint, where if the train had passed it, new command would come to slow it to 6 miles per hour (9.7 km/h). Inside the station, new commands at another insulated rail joint would command the train to stop. At the station, the train would have opened its doors, reversed course (as this is a two station shuttle line) and the lighting for the directional signs would be changed to match its new destination.

Sea Beach Line test track

Starting in December 1959, the fully automatic train was tested on the BMT Sea Beach Line express tracks between the 18th Avenue and New Utrecht Avenue stations. The train was equipped with a telephone system to keep voice communication with human dispatchers at the two shuttle terminals. At each station there was a cabinet that housed 24 relay systems that made up electronic dispatchers. The relays controlled the starting of the train, the acceleration, braking, and stopping of the train, as well as the opening and closing of the car doors. The relays were operated by electrical impulses initiated by a punched tape. When operating at full speed, the train operated at 30 miles per hour, and the train slowed to 5.5 miles per hour when coming into the two stations. When entering stations, the train passed through a series of detectors, which caused a series of tripper arms at trackside to go into the open position if the train was going at the speed. If the train was going too quickly, the tripper arms would stay upright and the train's brakes would automatically be set.

The equipment was built and installed by the General Railway Signal Company and the Union Switch and Signal division of the Westinghouse Air Brake Company, after several years of research and development. The NYCTA contributed between $20,000 to $30,000 on the project and supplied the three R22 subway cars to be automated. The bulk of the money, between $250,000 and $300,000, was contributed by the two companies, which paid for installation, maintenance and technological oversight of the automation process, including signaling. The automation of the shuttle was opposed by the president of the Transport Workers Union, Michael J. Quill, who pledged to fight the project and called the device "insane."

Implementation and demise

The R22s were fitted with different types of brake shoes, to see which one would negotiate the rail joints better. It was eventually found that the automated trip took 10 seconds longer than manual operation (about 95 seconds, compared to 85 seconds). As the tests on the Sea Beach line progressed, grade time stops were added to ensure safety on the line, and on the 42nd Street line. The train was dubbed SAM, and was to operate on Track 4 of the shuttle line.

In the afternoon of January 4, 1962, the three-car automated train began service, with a ceremony. The trains carried a stand-by motorman during the six-month trial period. The train had scheduled to begin service on December 15, 1961, but Quill threatened to strike all city- and private-owned transit in the city if the train ran. Under the new contract with the TWU, the NYCTA agreed to put a motorman in the train during the experimental period. While in its experimental period, the automated train was only operating during rush hours. In July, the test was extended for three more months, and in October the test was extended for six additional months. The chairman of the NYCTA, Charles Patterson, was disappointed by the automated shuttle train, doubting that the train could be operated without any transit personnel on board.

Initially, the automation of the shuttle was expected to save $150,000 a year in labor costs; however, with one employee still required on the train, there would essentially be no savings. If the test succeeded, it was planned to automate the IRT Flushing Line, the Franklin Avenue Shuttle, and the Culver Shuttle. However, the NYCTA did not have plans to automate the whole system.

A severe fire at the Grand Central station on April 21, 1964, destroyed the demonstration train. The fire began under a shuttle train on track 3, and it became larger, feeding on the wooden platform. The train on Track 1 was saved when the motorman saw smoke, and reversed the train. The basements of nearby buildings were damaged. Tracks 1 and 4 returned to service on April 23, 1964, while Track 3 returned to service on June 1, 1964. The reinstallation of Track 3 was delayed because of the need to replace 60 beams that were damaged in the fire. Reconstruction of the line persisted until 1967.

The automation test provided the framework for automated rapid transit technology on Bay Area Rapid Transit (San Francisco) and PATCO Speedline (Philadelphia to Lindenwold). After the fire that destroyed the automated shuttle subway cars, ideas for New York City Subway automation lay dormant for years, until an intoxicated motorman caused a train crash at Union Square station that killed 5 people and injured 215. The collision was a catalyst to a 1994 business case outlining arguments for automatic train operation (ATO) and CBTC, which led to the automation of the BMT Canarsie Line starting in the early 2000s.

CBTC test cases

The first two lines, totaling 50 miles (80 km) of track miles, got CBTC from 2000–2017. The two lines with the initial installations of CBTC were both chosen because their respective tracks are relatively isolated from the rest of the subway system, and they have fewer junctions along the route.

Canarsie Line CBTC

The Canarsie Line, on which the L service runs, was chosen for CBTC pilot testing because it is a self-contained line that does not operate in conjunction with other subway lines in the New York City subway system. The 10-mile length of the Canarsie Line is also shorter than the majority of other subway lines. As a result, the signaling requirements and complexity of implementing CBTC are easier to install and test than the more complicated subway lines that have junctions and share trackage with other lines. Siemens Transportation Systems built the CBTC system on the Canarsie line.

The CBTC project was first proposed in 1994 and approved by the MTA in 1997. Installation of the signal system was begun in 2000. Initial testing began in 2004, and installation was mostly completed by December 2006, with all CBTC-equipped R143 subway cars in service by that date. Due to an unexpected ridership increase on the Canarsie Line, the MTA ordered more R160 cars and these were put into service in 2010. This enabled the agency to operate up to 26 trains per hour up from the May 2007 service level of 15 trains per hour, an achievement that would not be possible without the CBTC technology or a redesign of the previous automatic block signal system. The R143s and R160s both use Trainguard MT CBTC, supplied by Siemens. In the 2015–2019 Capital Program, funding was provided for three more electrical substations for the line so that it could accommodate even more trains per hour. Also included in the Capital Program, automatic signals will be installers on the line to facilitate the movement of work trains between interlockings.

Flushing Line CBTC

The next line to have CBTC installed was the pre-existing IRT Flushing Line and its western extension opened in 2015 (served by the 7 <7> trains). The Flushing Line was chosen for the second implementation of CBTC because it is also a self-contained line with no direct connections to other subway lines currently in use. The 2010–2014 capital budget provided funding for CBTC installation on the Flushing Line, with scheduled installation completion in 2016. The R188 cars were ordered in 2010 to equip the line with compatible rolling stock. This order consists of new cars and retrofits of existing R142A cars for CBTC.

In late winter 2008, the MTA embarked on a 5-week renovation and upgrade project on the 7 <7> trains between Flushing–Main Street and 61st Street–Woodside to upgrade signaling and tracks for CBTC. On February 27, 2008, the MTA issued an Accelerated Capital Program to continue funding the completion of CBTC for the 7 <7> trains and to begin on the IND Queens Boulevard Line (E F trains).

The installation is being done by Thales Group. CBTC, as well as the new track configuration added in the line's 2015 extension, allow the 7 <7> services to run 2 more tph during peak hours (it currently runs 27 to 30 tph, but has a built-in capacity for 33 tph).

Wider installation of CBTC

As part of the 2015–2019 Capital Program, there will be 73.2 miles (117.8 km) of lines that will get CBTC, at a cost of $2.152 billion (part of a $2.766 billion automation/signaling project that is being funded in the Capital Program). Another $337 million is to be spent on extra power substations for CBTC. This installation of CBTC would require Siemens and Thales to cooperate on the installation process for all of the lines; they had worked separately in installing the Canarsie Line's and Flushing Line's CBTC systems, respectively.

Culver Line CBTC

Funding is allocated for the installation of CBTC equipment on one of the IND Culver Line express tracks between Fourth Avenue and Church Avenue. Total cost is $99.6 million, with $15 million coming from the 2005-2009 capital budget (phase one) and $84.6 million from the 2010-2014 capital budget (phase two). The installation is a joint venture between Siemens and Thales Group. The estimated completion date was scheduled for March 2015, and the installation is expected to be permanent. Should Culver Line express service be implemented, the express service will not use CBTC, and testing of CBTC on the express track will be limited to off-peak hours. Test trains on the track were able to successfully operate using the interoperable Siemens/Thales CBTC system. That system is now the standard for all future CBTC installations on New York City Transit tracks as of 2015. A third supplier, Mitsubishi Electric Power Products Inc., was given permission to demonstrate that its technology could be interoperable with the Siemens/Thales technology. The $1.2 million Mitsubishi contract was approved in July 2015.

The local tracks would also get CBTC as part of the 2015–2019 Capital Program, as well as the entire line between Church Avenue and West Eighth Street–New York Aquarium. The Ditmas Avenue, Kings Highway, and Avenue X interlockings would be upgraded.

Queens Boulevard Line CBTC

The MTA is planning to implement CBTC on the IND Queens Boulevard Line. CBTC is to be installed on this line in five phases, with phase one (50th Street/8th Avenue and 47th–50th Streets–Rockefeller Center to Kew Gardens–Union Turnpike) being included in the 2010-2014 capital budget. Following the success of the Culver Line CBTC test track, the MTA awarded a $205.8 million contract for the installment of phase one to Siemens and Thales in 2015. These two suppliers were the only MTA-qualified suppliers that could install CBTC on New York City Transit tracks. Of this initial cost, the MTA will pay $156.2 million to Siemens and $49.6 million to Thales. Both of these suppliers would install the new signaling system on the tracks themselves. Seven of 8 interlockings in the first-phase section of the line would be upgraded under this contract. In addition, 309 four- or five-car sets of R160s will receive Trainguard MT CBTC, the same CBTC system installed on trains assigned to the BMT Canarsie Line, which will be compatible with the SelTrac CBTC system installed on the tracks. Of the 309 units set to be converted for CBTC-compatibility, 305 will receive new onboard equipment, which NYCT contractors will install in 301 of these 305 units.

Planning for phase one started in 2015 and was complete by February 2016, with major engineering work following in November 2016. In January 2017, L.K. Comstock & Company Inc. was selected to fulfill a $223.3-million contract to upgrade existing signals and install communications, fiber-optic, and CBTC infrastructure for the new signal system. The total cost for the entire Queens Boulevard Line is estimated at over $900 million. The automation of the Queens Boulevard Line means that the E F services will be able to run 3 more trains during peak hours (it currently runs 29 tph). This will also increase capacity on the local tracks of the IND Queens Boulevard Line. However, as the line hosts several services, installation of CBTC on the line can be much harder than on the Flushing and Canarsie lines.

Eighth Avenue Line CBTC

Funding for CBTC on the IND Eighth Avenue Line from 59th Street–Columbus Circle to High Street is also provided in the 2015–2019 Capital Program, along with the modernization of interlockings at 30th and 42nd Streets. Originally, this funding was to be for the IND Sixth Avenue Line, but since the Eighth Avenue Line is more heavily used, it is prioritized to be the next corridor to be retrofitted with CBTC, and the Eighth Avenue Line automation project will probably be done concurrently with the Queens Boulevard Line automation. It will get two new electrical substations to support CBTC upgrades. Construction is scheduled to begin in October 2018.

Other lines

As of 2014, MTA projects that 355 miles of track will receive CBTC signals by 2029, including most of the IND, as well as the IRT Lexington Avenue Line and the BMT Broadway Line. The MTA also is planning to install CBTC equipment on the IND Crosstown Line, the BMT Fourth Avenue Line and the BMT Brighton Line before 2025.

On the other hand, Regional Plan Association prioritizes the Lexington Avenue, Crosstown, Eighth Avenue, Fulton Street, Manhattan Bridge, Queens Boulevard, Rockaway, and Sixth Avenue subway lines as those in need of CBTC between 2015–2024.