Caltrain EMU Cab Video Signal Survey

After the loss of the Caltrain gallery car railfan window I wasn't holding out much hope for having anything worthwhile to do on the Caltrain system as European spec rolling stock typically come with bulkhead style cab doors. However videos have been appearing showing that front facing views are present. This have the added benefit of providing a southbound signaling survey as previously such photos and videos were only possible facing northbound. 

I have already covered changes between San Jose and Tamien and can now confirm that the remaining Southern Pacific style signals in the terminal area have been replaced. The good news is that the few SP target type signals on the north end near the tunnels are still in place although unfortunately the triple unilens stacks at CP-COMMON were replaced by a standard LED modular type in a three virtual head configuration. Another note is how bad the sight lines for some of the wayside mast signals have become. This makes clear the advantage for position lights and cab signals in electrified territory. Although the freight spec PTC system Caltrain uses is kinda like a cab signal, moving to a Rule 562 coded track circuit arrangement might save them a lot of trouble. 

You can watch the video for yourself, but despite the hype the weekend and daytime local service is still highly underwhelming with little in the way of interesting signaling or routing. Although faster than diesel, the speeds are still slow compared to a limited stop run and unfortunately these have been largely curtailed with the fastest peak period trains still 3 or 4 minutes slower than the famed 4 or 5 stop Baby Bullets of the 2004-2024 time period.

Inside the US&S A-5 Switch-Point Machine

This blog devotes a lot of coverage to the plight of the pneumatic switch-point machine and its long decline since its heyday between 1930 and 1955, but apart from the ka-chug--woosh sound, what makes them distinctive? Well its actually their simplicity compared to their electric counterparts and its that simplicity that made them so competitive with electric point machines despite the overhead of making and distributing compressed air. 

Above we see a Union Switch and Signal model A-5 pneumatic point machine. (An earlier A-1 model also existed and you can tell them in photos because the housing looks like a miniature Quonset hut.) At one end we see the double acting air piston, at the other we see a connection for the control circuits and a blind tube without an air piston, and in the middle we have a fairly plain looking box.

Next lets quickly go over a standard power operated switch layout using some casual terminology. Moving away from the camera we have a stretcher bar that provides rigidity. Next is the throw mechanism that transmits force from the point machine to the points via a rod. This uses a slack coupling so the point mechanism can begin to move and break the static friction. Finally we have the detection rods that accurately detect if the points are hard against the stock rail and communicate that state to the signaling system.

Exposed US&S A-5 pneumatic point machine.

The mechanism can be set up for operation from either side.

Here we see a naked A-5 without its cover on showing off the simplicity of the action. The air piston connects to an action bar which runs through a crank mechanism that transmits force to the point rodding and also its positional state to the circuit controller.

The crank (left) and circuit controller (right).

The coupling between the air piston and the action bar.

The air piston, these came in 3 inch and 5 inch diameter models.

The throw rod couples to the crank assembly. In the photo below this connection is the bolt hole on the stubby arm below the pivot pin. The detection rods sit under the circuit controller have the golden coated nuts on them.

The circuit controller pulls double duty, detecting both the position of the points for the signaling system and also the position of the action bar / air piston, which is sent to the style CP air value unit to cut the flow of air and dump the pressure in the piston. We can see the hind end of the action bar that is given room to push outward under the covering.

The upper two rods provide the actual point detection while I believe the lower rod that extends through the mechanism provides a physical locking function in conjunction with the action bar. When this technology was new that rod would serve the role of the familiar facing point lock.

The goal of the mechanism is to lock the switch point securely against the stock rail even when air pressure in the cylinder has been exhausted, but allow smooth movement of the action bar and points when air is admitted to move the piston. A similar concept is present in firearms where the bolt is locked against moving rearward by the force of gases in the chamber, but is able to be easily unlocked via motion in a different axis.

While I have personally captured video of an A-5 point machine performing a throw cycle, it only showed the points moving along with the admittedly cool sound. However I recently came across a different video from a UK rail museum that has their own flavor of a US&S A-series set up for live demonstrations. Still popular on the London Underground, these instances are commonly used without the protective cover which allows visitors to actually see how the mechanism functions. This flavor of Westinghouse point machine is intended to be set between the rails and has the throw rods moving through the mechanism. The extra circuit controller on the back provides feedback to the pneumatic value.

 

MARION (AC) Tower's Missing CTC Machine

A couple years ago the folks at the Railroad Media Archive Youtube Channel posted a video covering the last months of Conrail's MARION (formerly AC) tower in Marion, OH.  Originally built by the Erie Railroad in 1902 and located at the parallel crossing of the CSX Toledo Sub and NS Sandusky Branch with the Conrail Indianapolis Line, the video shows off the tower's GRS/Taylor Model 2 interlocking machine with its proto-pistol grip type levers. At the time, MARION's operator only had control of the local interlocking, which seemed reasonable considering that they had to deal with train movements on separate railroads. However a new video from Railroad Media Archive shows a whole other side to operations at MARION. Sometime after the closure of the Eire main line around 1978, MARION gained CTC control of the Indianapolis Line between Belelfonte and Galleon (CP-BURT) with a 80's "Traffic Master"-style CTC panel and interface situated behind the operator's desk controlled by a compact interface of action keys and a numeric keypad.

In the posted video full attention is given to this CTC panel, with MARION's legendary lever frame barely making an appearance. We watch freight trains slowly cross MATION's territory while the operator discusses his duties and lines routes while text annotations appear in the video's side bars. It is mentioned that the tower would be losing its CTC territory by the end of the year (1989). 

The video captures that fleeting period between classic towers and train order offices, and modern point and click computer dispatch interfaces. By the time MARION was fully closed in 1995, all traces of its 1980's CTC territory had been removed. 

New Pittsburgh Line Interlockings

Norfolk Southern's ability to extort the state of Pennsylvania to add additional passenger frequencies on the old Pennsylvania Railroad Main Line is resulting in a slew capacity expansion projects in the form of "new" interlockings and additional tracks. I use the quotes as some of the interlockings are actually restorations of those that existed within living memory, but were removed for the sake of efficiency. 

Restoration of the old SG and NY interlockings.

The first phase of this work is underway at the set of the old SG interlocking west of Johnstown. The new crossovers will allow Amtrak trains to use both sides of the single island platform that serves tracks 2 and 3. Currently if track #2 is unavailable, trains must use main track #3 which follows the slower route of the old Sang Hollow Extension for 15 miles between CP-CONPIT and CP-C since SG interlocking was removed by Conrail around 1980. 

Returning 3 tracks to the Rockville Bridge

Additional changes include the restoration of CP-WEST PITT, removed by NS around 2007, a new crossover on the West Slope near the old NY interlocking between CP-SO and CP-MO, a new third main track between CP-ALTOON and CP-ANTIS, a new interlocking on the Altoona raceway between CP-ANTIS and CP-GRAY and another third main track between CP-BANKS and CP-HARRIS, which would restore a third track to the Rockville Bridge which was lost around 2000. Summary details of the project can be found here with a 2021 report providing more details of the track and interlocking changes. Note the report is not gospel as photos of the new CP-SG show parallel crossovers instead of a universal crossover.

Although no longer equipped with Conrail/PRR era signals, the NS Pittsburgh Line is still Rule 562 cab signaled with "C" lamps, so more interlockings means adding signals where automatics were previously removed.

Clear to Next Interlocking Rule 280a Displayed at CP-SOLOMON (EAST PITT)

In NORAC-aligned cab signal territory where wayside intermediate signals are not provided, Rule 280a, "Clear to Next Interlocking", allows trains without cab signals to proceed under signal indication instead of needing a track warrant or moving at Restricted speed. Related to the old concept of manual block clear, Rule 280a consists of a flashing lunar white light under the letter 'C' adjacent to a wayside controlled signal. These are normally pretty hard to catch in the wild because they are intended to be used to remedy en-route cab signal failures which are both rare and impossible to predict. For a time it was policy for Norfolk Southern dispatchers to run some Amtrak trains under absolute block protection and I was informed that some would display the  Rule 280a "C Lamp" where available, but I never managed to observe this practice for myself. 

That being said I did stumble upon a scheduled use of Rule 280a that one can catch if they are ever in the Pittsburgh area. After arriving at its Pittsburgh terminus, Amtrak Pennsylvanian Train 43 must reverse about 5 miles to CP-HOME where the closest turning wye is located. Because the rear coach lacks cab signal capability, the "C lamp" is displayed in along side the interlocking signal at CP-PITT, CP-EAST PITT (aka CP-SOLOMON), CP-BLOOM (if necessary) and depending on the order of the wye move, CP-HOME. 

Somewhat ironically the first three interlockings, CP-PITT, CP-EAST PITT and CP-BLOOM, are all back-to-back with no code change points between them. NS could have designated that track segment as normal Rule 261 without the "C" lamps, but their policy followed that of Conrail to provide the "C" even where it is not necessary. (Amtrak had chosen to do the opposite up until about 2010).

Anyway, here we see the 2E mast signal governing track #1 eastbound at CP-SOLOMON (EAST PITT) on the former Conrail Pittsburgh Line displaying Rule 280a in conjunction with a Medium Clear indication reverse move of Amtrak's westbound Pennsylvanian Train 43. CP-EAST PITT is only about half a mile from CP-BLOOM and directly adjacent to office parking making it and ideal spot to Rule 280a in the wild. 

I got this video from the east end of CP-PITT, which a shorter walk from downtown. About 30 minutes after its outbound passage, Train 43 will get more more "C" lamps displayed for a second long reverse move back into Pittsburgh Penn Station however the 2W signal at CP-PITT will display a Restricting into the station track without the "C" indication.

Open Railway Map's Signaling Layer is Hot Garbage

Open Railway Map (ORM) is a great resource that recently came on the scene to allow for fast identification and tracing of railway lines, rights of way and even individual tracks. Unfortunately as project that is based in Continental Europe, its coverage of North America (and the British Isles) has some serious shortcomings, especially when it comes to signaling and train protection systems. The information is incomplete and the presentation is seriously lacking. Ironically we have already seen a better way to present this sort of information in multiple railroad-produced employee timetables. 

In Europe almost all railway lines are considered signaled, with the specific method of block working, (ie manual, automatic, token, bi-directional, etc) not being as important as it is in North America. As a result the signaling layer mostly documents special train protection systems like ETCS, TVM420, LZB, etc. (Even then a lot of "National" ATS-like systems like AWS and Le Crocodile do not make an appearance, although PZB does.) 

 

When translated to North America the result is a generic tag for "PTC" and another for "ATC", even though neither labels refer to specific train protection systems as is the case on the European overlays. One might say the project got a little in over its head trying to color code every protection system, although they do have 40 color codes for speeds. The maps cannot seem to handle situations where multiple gauges, electrification systems or train protection systems are present on the same track segment.

The solution, at least for the North American segment, is to first dump the generic "PTC" and replace it with ETMS and ACSES. Next, drop ATC and replace that with Cab Signal System (CSS) and then add in "ETMS with CSS" where applicable. Since ACSES generally needs CSS there is no reason to restate the presence of CSS in that case. For a stretch goal I would add tags for CTC and ABS, however in a global context indicating bi-directionality is less important, so maybe a single TCB tag for "track circuit block" can be used where ETMS is not present. In the (currently) few cases where ETMS s present without TCB, and "ETMS with TWC" tag could be used. 

In an ideal world ORM could "stripe" colors where multiple tags apply in a similar fashion to the Union Pacific employee timetables which at one point were managing CTC, ABS, TWC, ATS, ATC and CSS systems, but that would probably require significant development time.

If ORM had some clear way to make contributions I'd be down to make the more basic changes myself wiki-style, but I can't see exactly how to do that and e-mails are going unanswered. Leave a message in the comments if you know what I'm missing.

Adios B&O MEXICO Tower

The reason I tend to miss a lot of closures and demolitions is because I don't tend to seek out news that will leave me bummed and depressed. In another example of bad news I missed the B&O's "M" aka MEXICO tower, near Cumberland, MD, was apparently demolished in September of 2023. The three story 1950's brick tower was similar to many across the B&O and, with its CPL bracket masts, had served as a backdrop to the local railfanning scene for decades.

CSX has been engaged in a long term project to convert the former hump classification yard to flat switching and a large obstacle to this was the physical hump and related structures.  MEXICO, built at the same time as the classification infrastructure, was located some 2 miles away at the eastern end of the yard leads and was in good structural condition.  Although its possible that CSX plans to alter the track configuration at MEXICO interlocking at some future date, I suspect that since they had already contracted for the asbestos remediation guys to come out for the yard hump tower, also getting rid of MEXICO could have come with little additional cost. 

Former Cumberland Yard hump infrastructure

Closed as an interlocking station in the mid-1990s, the relay based GRS plant was not re-signaled until the mid-2010's. Unless similar B&O towers at GREENE JCT (boarded up) and PATERSON CREEK (derelict), MEXICO saw use as an active railroad structure, either for maintenance crews or a rest facility for departing road freights.  The implementation of new service plans that generally sidelined the old system of classification yards may have played a role as Cumberland would see less originating eastbound trains and therefore less need for a standing structure to support the crews.

I personally only get to visit the location once on the ground, getting my photos from the road bridge like everyone else, although I had many opportunities for photos from the Capitol Limited with the last occurring in 2019.  Of those B&O towers that still survive, PATTERSON CREEK, only about 4 miles away from MEXICO as the crow flies, would be the most similar example.

NECR Resignaling Project is Complete

I can report that the re-signaling of the New Englande Centrale's main line that hosts Amtrak's Vermonter north of the Massachusetts border has been completed. While I had previously reported on the state of this project, I can confirm that all legacy signaling has now been replaced. While the diamond at Bellows Falls was perhaps the most visible of the final round of changes, the north end of the Putney siding was also re-signaled. 

As always there are a few interesting details I should point out. The previously reported signaling anomaly at Bellows Falls resulting from a old double to single track conversion, has been eliminated. Trains at Bellows Falls now get a Clear signal indication for a straight route at Wapole siding as opposed to the previous Approach Medium to Medium Clear. 

The holdout signal at SOUTH BELLOWS FALLS has been retained, evidently to allow interchange with the Green Mountain Railroad without needing to seize the main line all the way to Putney.

The entire NECR portion of the line has done away with the Boston and Maine / Guilford practice of ultra-long ABS blocks. Now signal blocks will be a more uniform 2-4 miles which limits the impact of track circuit failures and increases capacity.

Having been re-signaled just prior to the start of the most recent project, the South Putney interlocking will be the only non-standard signal location on the line with a poor man's bracket mast for southbound trains and the use of Safetran scallop shell modules as opposed to Progress Rail square modules.

Although an honorable mention goes to the old northbound searchlight signal on the Bellows Falls platform that was so popular as a photo backdrop. That has been moved to the adjacent park around the Bellows Falls Chamber of Commerce.

It will be interesting to see if this affects on time performance or the schedule at all. In theory replacing the TWC with CTC for the 10 or so miles south of Brattleboro could allow a speed increase.

No Approach! The Duality of Uncommon Signals

We all learn about traffic lights in elementary school.  Red means stop, green means go and yellow means slow or caution. These are the basics of automatic block signaling that also tend to get taught in entry level railroading books. Now its pretty logical that a signal might lack the ability to display a clear indication. A permanent stop condition or the end of automatic block territory are both situations where a Clear signal would not ever apply. However there are a few situations where signals in North America can display Stop or Clear, but not Approach on the full speed head. 

The first is at interlockings outside of automatic block territory like a diamond crossing or drawbridge. The signal provides movement authority through the interlocking and without track circuits outside of the interlocking limits there is no "prepare to stop" points an Approach type signal would apply to. Trains will have their own movement authority like a track warrant so a clear signal is basically the interlocking telling the train "you do you". One might expect signals in this case to display a Restricting indication, instead of a clear, and sometimes they do, but that is going to come with a 20mph speed restriction, or less, which can be a time hit on unsignaled lines with higher track speeds. 

The next situation, Manual Block territory, would be very familiar to those in other parts of the world. Although almost extinct in the US, there remains a bit of manual block still in service on the LIRR and the way to identify a manual block signal is the lack of an approach indication. (This also applies in general to historic PRR signaling charts.) Signals will display Stop or Clear Block with any Approach functionality handled by a separate distant signal towards the end of the manual block. The PRR even went one further sometimes substituting the \ "Caution" indication in place of / Approach. 

Sometimes its really not good to have locomotives stopped and waiting at a specific signal,  Maybe there's an issue with a grade crossing, or noise, or diesel exhaust. In this case it would not be desirable for trains to approach that signal prepared to stop, so a preceding signal will either hold trains short or allow them to approach a signal with some sort of proceed indication. The most notable of these is at the entrance to Washington Union Stations 1st St tunnel that the currently diesel hauled northbound trains cannot enter without having a signal displayed further on that allows them to leave said tunnel. Another nearby signal location at CSX's LENFANT interlocking might lack southbound approach signals for a similar reason due to an overbuild.

One somewhat PRR-specific situation is where interlockings are provided with exit signals and because of the short distance between the interlocking's entrance signal and exit signal, Slow Approach is used instead of Approach. This creates an upper head that can only display --- Stop or | Clear. The C&O also made frequent use of exit signals, but seemed to keep their upper head Approach indication, although Medium Clear to exit Approach allowed the C&O to minimize the use of R/Y/Y/ Medium Clear.

A fairly new situation where Approach has been "omitted" involves high speed turnouts. At certain points on the NEC, Amtrak has decided to combine a flashing green "Cab Speed" signal indication with R/*Y* Medium Approach. At the end of a main track that converges into another via a high speed turnout, the result will be an upper head with green and red lamps/positions and no yellow lamp/position.
 

The Enthusiasts Step Up

Unlike Europe, Heritage railroads in the United States never developed a vintage signaling component beyond what one might describe as display items. The big reason for this is that in service signaling and interlocking appliances have significant regulatory inspection, testing and documentation requirements that are beyond the reach of most tourist train operators. The result has been that most preserved towers are completely inert, with just a handful having interactive elements such as operable interlocking machines (SS43 BERK) or full on simulations (HARRIS). However, in recent years it seems that the private collector/enthusiast space has been stepping up to fill the gap as exemplified by the small YouTube channel Laser 711.

Signal and signaling equipment collectors have been around for some time, snapping up such items as model boards, CTC cabinets, full size signals and even interlocking machines. Until recently, railroad signaling departments would tend to keep their retired equipment, as similar vintage hardware was still broadly in service and the parts to maintain them were no longer being manufactured. However, the PTC related re-signaling push rapidly phased out so much of the relay era railroad signaling that the equipment is now being sold for scrap or offered free to anyone with the capacity to haul it away. 

In the context of interlocking towers, model boards and CTC cabinets have been widely collected for some time as switch to video display in a dispatch office had long ago destroyed their reuse value. Somewhat ironically, it was the older mechanical and electro-mechanical interlocking machines that retained the most value as a source of spares for their kin.  (For example a major reason LENOX tower near St. Louis was closed was to supply parts to the CNW LAKE ST machine in Chicago.) However as the number of these in service interlocking machines hit zero, anyone with a large enough truck could cart the survivors away for their own personal amusement. This is where Laser 711 comes in, having apparently set up not just a HARRIS style PLC backed simulation, but the full hardware stack including relays, point machines and signals.

 He has also restored an ex-Erie CTC console from BK tower along with its original code system, and now appears to be in the process of implementing as much field functionality as possible "in relay".

Of course it would be great if more publically accessible museums were able to step up to the plate with functional period correct signaling equipment, but the signal enthusiast community, powered by recent advances in the "maker" scene and high capacity pickup trucks, seem to be taking up a lot of the slack. Just like we have seen with the extensive telecom switch collection of Step by Step Phil, these private efforts stand a good chance of eventually finding their way into durable preservation.

Attempted CTC at the Seashore Trolley Museum

I will also mention that non-FRA regulated streetcar/transit museums have the best capacity to implement "live" historic signaling. When visiting Shoreline Trolley Museum I was shown their "in-progress" CTC system that they were building from thousands of relays donated by Amtrak and Metro North.

 

The fact that I have not heard much if anything about that project in the past 20 years hints at its priority in the greater scheme of things, but all it takes is the right team of enthusiasts to get a project off the ground.

New Signals Up At THORN

 Although not much has changed regarding the ultimate fate of THORN tower, recent photos do show what the new signaling situation is going to look like. In the eastbound direction the triple signal bridge covering tracks 1, 2 and 5 Running, have gotten the Safetrain upgrade with Rule 562 'C' boards turned. Although not yet in photos, it is expected that track 4 will also be getting a PL mast signal to replace the existing reverse direction dwarf. SEPTA is tentatively resuming service to Coatesville upon the completion of new station facilities in 2025 which includes a semi-restoration of #2 track to a point just past the new station for the daily NS regional freight. 

In the westbound direction new PL masts are up on tracks 1 and 2 to replace the reverse direction dwarfs and its assumed that a new gantry mounted signal will be in store for #4 track.  The real surprise is what looked to be a new westbound full PL mast signal on what used to be the Low Grade Freight Line track. The old THORN had maintained a vestigial signal in that same location, but despite the rationalization of the former 6-track full crossover, it seems the low grade stub track has not only been retained, but is being prepared for use. What that use might be remains to be seen, but a mast capable of displaying Medium and Limited speed signals could imply SEPTA trains making some sort of switchback maneuver after discharging on Thorndale's track #4 platform or coming out of some yet to be built storage tracks on the footprint of the old PRR Thorndale facility. 

The THORN interlocking cutover is still planned for later in June with DOWNS to lose its switches before complete removal. GLEN and FRAZER will be turned over to the dispatcher by the end of 2025 closing THORN as an active interlocking station.

NECR / Central Vermont Changes You May Have Missed

While I have been posting much about the loss of former Boston and Maine searchlight signaling on the New Englande Centrale, a couple of other changes on the Central Vermont end of the CT Valley corridor caught my eye. While B&M ownership and CTC ended at Windsor, at some point in the 1990's the Statge of Vermont and/or Amtrak got some money to install about 10 miles of new CTC between there and White River Jct. Unlike the low cost B&M system, the CTC extension had regular 2-3 mile long blocks with vertical color lights, but no new interlockings or passing points.



Upon closer inspection one might notice that the type of color light signal are the short lived mono-housing GRS variety that are also present around LA Union Station, Dallas Union Station and the extreme northern end of the Conrail Lehigh Line.



Anyway in an absolutely baffling decision that was probably the result of generous Federal funding, the NECR also replaced the completely modern CTC extension several years before the 1950's B&M signaling. The new signals consist of hooded LED modular units from Progress Rail The new mast-on-relay-hut signals replaced the older and slightly smaller mast-on-relay-hut signals and feature Canadian style "R" boards in both directions as well as new number places with direction suffixes. The change took place between 2022 and 2024 and mirrors the new signaling installed between EAST NORTHFIELD and Brattleboro, VT.



Another change came in the form of a further CTC extension all the way through the White River Junction area sometime in the 2010's. This change took place between 2012 and 2019 and includes interlockings north and south of White River Jct.



Stay tuned for at least one additional post covering the recent history of signaling in the Connecticut River Valley.

CSX Continues to Attack the Big Four

Turns out that the River Line isn't the only part of Conrail territory currently under attack by CSX signaling crews as I just saw that CSX had replaced the 1990's Conrail vintage signals at CP-37 on the former Indianapolis Line in Wellington, Ohio. This portion of the line had been re-signaled late in the Conrail era and could otherwise be considered "modern" instead of a NY Central legacy holdover. The three track eastbound signal bridge could be considered a local railfan landmark.

The westbound cantilever was likewise replaced. The portion of the Indy Line between Berea and Greenwich is notable for carrying bulk of CSX traffic from the former Conrail territory in NY State before it switches onto the old B&O route to Chicago via Willard. 

To be determined if this is a wide area project or more hen pecking. The signal aspects being Conrail or Seaboard will be a hint. Other signals south of Greenwitch, like these NYC-looking tri lights at CP-189 in Ansonia, OH, are also being replaced.

Unfortunately this is not a part of the country that I have good access to so hopefully some locals are getting all the necessary pictures.

Mr. Beast Fakes the Trolley Problem

For those of you who don't know, "Mr. Beast" is one of the most popular figures on YouTube mostly appealing to 11-14 year old males with formulaic content that tends to involve giving away large amounts of money in novel ways. To get value from this post you do not need to know who Mr. Beast is and I am not recommending you view his content. In 2024-2025 Mr. Beast produced a 10 episode reality TV show for Amazon Prime titled "Beast Games". While watching an in depth third party review/breakdown of Beast Games, I noticed something in the 7th episode the pulled back the whole reality TV artifice a little farther than normal due to my knowledge of how railway signaling and operating rules actually work and I figured I should share it with the class.

The premise of the 7th episode, entitled "Elimination Train" was to set up a trolley problem scenario with full a size AAR locomotive and rolling stock on a Canadian short line. Chosen players would be placed on a surprisingly realistic fake signal bridge with a prop "lever" situated perpendicular to the track. They would then have to choose to direct the train either towards a high value vehicle they could win or representations of their team members facing elimination. (Note, this isn't a faithful representation of the trolley problem which tries to explore the morality of action vs inaction.) While the overall show suffered from a lack of play testing, they did manage to coax one of each outcome for this event.

Anyway this is where the fakery of reality TV becomes visible to anyone with rail knowledge. The sequence begins with the contestant "choosing" the car over the other contestants with the "lever" placed in the reverse position. However to keep up the drama its made clear that he can still change his mind at any time until the train passes the switch like in the conceptual trolley problem. As the full sized locomotive with one car in tow appears out of a smoke effect and approaches the switch the episode went into overdrive with rapid cuts, fast-mo, slow-mo and insert shots trying to build the suspense to see if the contestant will change the direction of switch.

However in every wide shot where the switch and the locomotive is visible, the points are shown to never move from their initial position, even if an insert shot shows the pints moving or the contestant moving the lever. The scenario is run more than once, in one the points are reverse and stay reverse and in the other they are set normal and stay normal. While this might fool the casual viewer, for the rail-informed watcher will know the fix is in once the locomotive appears and is moving faster than a jog, because there is no way in hell some shortline was going to allow the producers (and by extension the contestant) to actually solve the trolley problem and derail their locomotive. To the show's credit the switch mechanism was never shown in detail and the area around where a hand throw switch stand would be was intentionally hidden from view, however I am not sure how much this ultimately mattered for the viewers given the moving points were only shown in extreme close-up.

Here's what I believe went down in the actual production. First, while technically possible, its highly unlikely that a short line would have rigged up a power assist switch to the prop lever when a hidden grip could easily do the same job. The contestant was given some amount of time to make his choice while literal man behind a curtain followed the position of the lever and used the hand throw stand to align the switch points for the benefit of all involved. Then at some point, possibly even before the locomotive emerged from the smoke the choice was set and the train was allowed to do its thing. All of the drama was then manufactured in the editing room.

That's the irony of the situation. The show theoretically sets up a contestant the chance to solve the trolley problem, but doesn't have the balls (or insurance coverage) to actually allow the contestant to solve the trolley problem. What people don't get about the trolley problem is that the choices you are given aren't the only choices that exist. If one puts the points on center, the trolley derails and comes to safe stop. This is how split point derails work after all. (Conversely if you throw the points after the first truck passes over them, but before the second, you can get the trolley to drift and kill both sets of people tied to the tracks.) The real life trolley problem tries to set up a binary choice to explore an moral dilemma, however the mechanics of the problem allow folks with knowledge and the ability to think outside the box to make a choice where everybody lives. That's the important takeaway. If presented with a no-win scenario, you change the conditions of the test. 

BTW, if you want to see how The Simpsons got an NYC Subway interlocking tower surprisingly right, check out this post here.