25 March 2009

Focus on: SF Transbay Transit Center

San Francisco's Transbay Transit Center, planned to be built in the years 2012-2018, will replace the decrepit Transbay Terminal and become the new transit hub of downtown San Francisco. The TTC includes a new underground rail station, which Caltrain and HSR will use as their northern terminus. The station will be connected to the tracks that currently terminate at 4th & King by a new downtown tunnel extension, also known as the DTX.

The DTX and revamped Transbay Terminal were planned since long before the high speed rail project, to better connect Caltrain and the peninsula to San Francisco and the east bay. Long starved of funds, the project was finally approved by San Francisco voters in 1999, with the creation of a new Transbay Joint Powers Authority (TJPA) charged with designing, building, operating and maintaining the DTX and the new Transbay Transit Center.

Fast-forward a decade, and Caltrain is playing second fiddle to the California High Speed Rail Authority, which (despite some past theatrics) fancies the Transbay Transit Center and its all-important underground train box as a northern terminus and veritable Grand Central of the West. Officials from the respective agencies are now squabbling over the design of the rail station, with the CHSRA's chairman Quentin Kopp calling the TJPA's design inadequate. (As it turns out, he may be right!) The disarray threatens the project's prospects for receiving economic stimulus funds, which has California politicians in Washington increasingly concerned that the appearance of "shovel readiness" is eluding their grasp.

The TJPA and CHSRA have developed a memorandum of understanding to govern future collaboration on the design of the DTX and Transbay Transit Center--not a moment too soon, since the TJPA plans to award construction contracts for Phase 1 of the project (possibly including the $390 million train box) starting in October 2009.

In this post, we'll discuss the rail component of the Transbay Transit Center and the DTX tunnel. We'll begin with a summary of the current train station design, as presented at the TJPA's March 12th board meeting (the source of most of the illustrations in this post), followed by some thoughts on what makes a good train station, and conclude with a discussion of shortcomings in the current TJPA design.

2013 Update: since this post was written in early 2009, the design of the TTC has evolved in small ways.  These are discussed in a 2010 update and 2012 update.

Many thanks to Richard Mlynarik for his advice and enlightening discussion of these issues.

Current Train Station Design

A cross section of the TJPA's design for the Transbay Transit Center (as of late 2008) is shown at right. It is a long, narrow building that occupies approximately the same footprint as the Transbay Terminal and its ramps (see location map).

The underground train station is reached via a tunnel which starts a bit before the existing Caltrain terminal at 4th and King Streets. After a ramp down and a sharp right turn onto Townsend Street, the tunnel takes a sharp left onto 2nd Street, then a sharp right into the terminal building.

Station Layout. The Transbay rail station has two levels: a mezzanine level (immediately below grade) and a second basement level with platforms and tracks, as shown in the plan at right by TTC architects Pelli Clarke Pelli. The mezzanine level does not extend for the entire length of the station; the south end contains a Greyhound bus terminal, and a large portion is reserved for 'program space'. What remains of the mezzanine level is further subdivided: a central area concentrates passenger access and contains ticketing facilities and some retail shopping, and is flanked by a Caltrain waiting area on one side and a larger HSR waiting area on the other side, which will likely feature airline-like passenger security screening.

Track Configuration. The station has six tracks, tentatively numbered 21 through 26, from south to north. The two southernmost tracks (furthest from downtown) are dedicated to Caltrain. The remaining four tracks are dedicated to HSR. Due to differing platform height and passenger security requirements, the tracks are permanently assigned to Caltrain or HSR and cannot be switched. Adjacent tracks are placed on 16 ft 6 in (5 m) centers. Two tail tracks were originally planned to extend beyond the end of the terminal; their layout has changed several times, but they are now "deferred until operationally required." Of the many tail-track options evaluated (including a loop tunnel back to 4th & King), one configuration required the southern track to pinch inwards to avoid building foundations, cutting the Caltrain platform short. The only vestige of the tail tracks is a bulge in the end of the train box, with knock-down panels where a future tail track tunnel might be built, if they ever turn out to be "operationally required."

Platforms. The station is planned with three island platforms flanked by a track on each side, 1300 ft (400 m) long and 30 ft (10 m) wide, curving sharply at the entrance to the station. The curvature (500 ft or 150 m radius) opens gaps in excess of one foot (0.3 m) between the platform edge and train doors. The platform height is 3 ft 6 in (1067 mm) for the level-boarding HSR tracks, and 2 ft 1 in (639 mm) for Caltrain-- the latter value seemingly picked to match the floor height of a Bombardier car. A row of columns supporting the terminal building above runs along the length each platform, spaced on 42 ft 6 in (13 m) centers. The assumed HSR schedule turn-back time is 30 minutes, with a nominal platform dwell of 20 minutes, giving the four HSR tracks a total throughput of 6 high-speed trains per hour (after accounting for the delay required for another train to re-occupy the platform). The Caltrain schedule turn-back time is 20 minutes (nominal dwell of 18 minutes), giving the two Caltrain tracks a throughput of about 4 trains per hour (or 6 tph with tail tracks). The TJPA claims it can accommodate Caltrain's planned TTC ridership of 31,700 passengers per day (versus ~8000 per day at 4th & King today).

DTX Tunnel. The downtown extension tunnel is designed with three tracks. The portions of the tunnel close to grade level are built using cut-and-cover methods, with the middle portion mined under Rincon Hill, until Folsom St. (curved section shown in TJPA drawing at right). The tunnel includes a new underground Caltrain station along Townsend St. between 4th and 5th, adjacent to the existing 4th & King Caltrain terminal. This station has two short 800-ft (240 m) outside platforms, with the center track used for overtaking trains stopped at the station. Since rolling stock choices for Caltrain and HSR are yet to be determined, the engineering drawings show clearances sized for Bombardier cars. Like the TTC, the new 4th & Townsend station features a mezzanine level above the tracks, requiring the entire alignment to be built at sufficient depth to allow for the requisite clearances.

4th & King Station. The DTX project preserves the existing 4th & King station location as a Caltrain terminal and train storage yard, since the TTC's two platform tracks will not support Caltrain's anticipated service levels or allow diesel trains (diesels may be used for Dumbarton service). Caltrain is planning for eight platform tracks at 4th & King. Their Caltrain 2025 plan, Appendix D, issued before Caltrain was assigned just two platforms at the TTC, assumes that fully 60% of Caltrain services will terminate at 4th & King. None of the platforms at 4th & King will accommodate HSR, due to the difference in platform height.

Curve Radius. The 1.4-mile DTX features three extremely sharp curves that allow the tracks to conform as best they can to the San Francisco street grid. The minimum curve radius is just 500 feet (150 m), the sharpest allowable radius for European off-the-shelf trains, and too sharp for some Japanese high speed trains. In theory, such curves can be navigated at a maximum safe speed of 35 mph (55 km/h); in practice, slower speeds of 25 to 30 mph are likely. Regardless of the exact speed, such tight curves will elicit deafening screeches from the stiff, long-wheelbase bogies that are a technical necessity for high speed rail. The DTX is likely to set California-wide technical requirements for high speed trains.

Station Throat. The track schematic above shows the configuration of the tunnel and underground station tracks, with the new underground Mission Bay station (under Townsend St.) also featured. The throat of the station, where the three tracks of the DTX tunnel divide into the six platform tracks of the TTC, begins all the way back at Bryant St., about 3800 ft (1150 m) before the bumpers at the end of the station platforms. At 25 mph, a train takes nearly two minutes to snake its way through the throat of the station. Most of the track switches are according to US freight (i.e. Caltrain) standards, AREMA #14's.

Passenger Access. The TJPA is sizing the station for simultaneous peak passenger flows of 5000 pax/hour for Caltrain and 4000 pax/hr for HSR. All vertical access (escalators and stairs) is oriented lengthwise along the station axis. Rather than accessing platforms directly from the street, passengers are funneled through a grand entrance and waiting areas in the mezzanine level. After accessing the platforms, passengers must walk some distance along the quarter-mile platform to their correct boarding location.


Grade Separations. Between the existing Tunnel #1 and the entrance to the new DTX tunnel, the Caltrain tracks cross 16th St. and Common St at grade. After the latter grade crossing, the tracks descend underground via a 1960 ft (600 m) open cut. A later grade separation of these streets would require demolishing this U-shaped open cut ramp and extending the DTX tunnel.

So What Makes a Good Train Station, Anyway?

Designing a train station for efficient operations is not rocket science, although space constraints can considerably complicate the exercise. The extremely constrained location of the TTC, combined with the very high level of train traffic envisioned by 2030, demand an intelligent station design.

We can draw up a simple list of three key requirements for a good terminal station design:

  1. High train throughput. By definition, all trains arriving at a terminus must reverse direction. The station must be able to turn back trains at the highest possible rate; otherwise, the station itself becomes a worse bottleneck than the tracks leading to it.
  2. High passenger throughput. The flows of passengers into and out of trains must be made as simple and direct as possible. This involves an optimization of street access, ticketing facilities, and vertical access to station platforms. Limited and circuitous passenger throughput can become a bottleneck at peak hours.
  3. Operational flexibility. The station must provide flexibility to adjust service patterns in response to real-time conditions. It must degrade gracefully in the event of a service disruption, minimizing the possibility of cascading delays.
As we re-examine the TTC design, keep these three requirements in mind.

Shortcomings of the TJPA Rail Design

Throat Cancer. The TJPA station throat design is a bottleneck that promises low speeds, long traversals and frequent conflicts (when a key piece of track is occupied by one train, preventing other trains from passing.) The train throughput of a station throat, which sets a limit on the throughput of the station itself, is related to its overall length (shorter is better), train speeds (higher is better) and most importantly, topology (track switches thoughtfully laid out to allow conflict-free parallel moves.) The station throat begins a full 3/4 mile (1150 m) from the end bumpers, causing trains to occupy the throat for far longer than necessary and increasing the delay required for a platform to be re-occupied by the next arriving train, thus limiting
train throughput. The station throat does not allow simultaneous arrivals and departures from most platforms (such conflict-free parallel moves are the holy grail of throat design because they minimize the re-occupancy delay), severely limiting train throughput and operational flexibility.

Different Platform Heights. With Caltrain's fleet due for a complete replacement with the advent of electrification, and with Caltrain's platforms due for complete reconstruction with the track expansion for HSR, the opportunity exists to make HSR and Caltrain share the same standard platform height. (The shared use of the same platform by HSR and commuter trains is common practice in Europe.) Designing the two systems for different platform heights makes this already-constrained station design even less operationally flexible because platform tracks cannot be assigned as needed to (a) minimize conflicts in the station throat, (b) accommodate service peaks, and (c) recover from disruptions. While this isn't TJPA's decision to make, the TJPA, CHSRA and Caltrain should agree on a single existing standard that is based neither on California's outdated safety clearance regulations, which were written for freight trains in 1948, nor on the floor height of Bombardier cars. Inventing new platform standards will hinder the ability of both operators to procure trains
off-the-shelf and will create an operational mess.

Mis-Oriented Passenger Flows. The dominant pedestrian flows, especially for Caltrain, come from the center of San Francisco, which lies along the flank of the station (unlike traditional European terminals, which are often oriented toward the city center). Vertical access to the station platforms should be oriented perpendicular to the station, towards Market Street, to make use of the inherent horizontal reach of stairs and escalators to bring passengers closer to their destinations. Funneling all passengers through a grand entrance, concourse, fare gate, etc., all along the skinny axis of the building, while perhaps architecturally spectacular, does nothing for passenger throughput. Caltrain monthly pass holders need to get from the street to the platform, pronto, without any scenic detours.
Homeland Security Theater. The total segregation of passenger flows for HSR and Caltrain, to accommodate airline-style HSR security screening, constricts passenger throughput and prevents the flexible re-allocation of tracks. As the Madrid train bombings of 2004 demonstrated, securing HSR facilities does nothing to prevent mass carnage on the rails. Securing HSR without securing Caltrain is no more effective than building a moat only halfway around a castle. If security theater cannot be avoided, then adopt a security plan that does not operate the TTC as two separate mini-stations. With only two available tracks, Caltrain would find itself just one minor breakdown away from a total disintegration of their rush hour schedule, a disaster for operational flexibility.

Extremely Tight Curves. While curve radii at the location of the TTC itself are dictated by the street grid and surrounding building foundations, the two curves in the DTX tunnel approach (to turn from 7th onto Townsend and from Townsend onto 2nd) are much sharper than they need to be, and beyond the capability of some off-the-shelf high speed trains. High speed trains aren't MUNI street cars and won't turn on a dime; they are longitudinally stiff to provide good stability at high speeds. At these two curves on the 1.5-mile approach to the station, the radius could be
easily increased from 150 m to at least 200 m, saving 30 to 40 valuable seconds in the long, slow crawl into the TTC, and increasing the train throughput of the station throat. If the throat can be straightened as well, a 200 m minimum radius would allow off-the-shelf procurement of trains without the need for expensive modifications to fit the extreme curves of the DTX, which threaten to set a standard for the entire California high speed rail system.

Odd Tail Track Arrangement. Among the various design options considered before trail tracks were de-scoped, no clear function can be ascribed to the tail tracks under Main Street. They could theoretically have been used to increase HSR train throughput by (a) performing cleaning and resupply of trains without tying up a platform and (b) switching departing trains to the northern platform, to minimize conflicting moves in the station throat during departure. In practice, none of the TJPA's attempted tail track layouts exploited these possibilities; they amounted to very expensive train parking.
Inefficient 4th & Townsend Layout. The new underground station at 4th & Townsend should have four tracks to allow overtaking stopped trains in both directions (to improve train throughput) and its platforms should be lengthened to 400 m to allow HSR trains to use them in the event of a disruption at the TTC, thus improving operational flexibility to HSR and Caltrain's benefit.

Oversized DTX Tunnel. Despite the high level of train traffic, a simple two-track tunnel into the station would do just fine. Because trains are limited to the same homogeneous speed by the sharp curves, a two-track tunnel could easily support 15 to 20 trains per hour, each way, or far more than the terminal could feasibly handle. The third track makes the tunnel needlessly expensive, and requires fancy construction techniques like the New Austrian Tunneling Method (NATM).

Suggestions for Improvement

All of the design issues enumerated above have been analyzed by San Francisco rail advocates, who have suggested numerous possible improvements for the DTX and TTC rail station in order to exploit the constrained site as efficiently as possible for both Caltrain and HSR. These detailed suggestions will be the subject of an upcoming post.

The rail infrastructure of the Downtown Extension and Transbay Transit Center is not nearly as sexy as sweeping glass towers and roof gardens: it's invisible, and for now, it's just a box. We can only wish that the architects realize the extent to which that simple, invisible box will determine the operational efficiency of San Francisco's train station for decades to come.


150 comments:

  1. Has there been any confirmation that there will need to be security theater at the HSR stations?

    ReplyDelete
  2. Can the box just be made longer in either direction? Its not clear from google maps which foundations nearby are problems and which are not.

    Regardless, it's absurd that this station's shortcomings are going to set design constraints for the rest of the line. That's not a recipe for success. In terms of changing floor heights, wouldn't that be an issue for caltrain stations that need to share with the old diesels for gilroy service (not to mention added costs across the line, although I'd guess those are minimal compared to the overall electrification and grade separation costs)

    It seems like all these problems are solvable, we can only hope that they'll actually be solved rather than ignored until its too late.

    ReplyDelete
  3. The train box needs to be wider so more platforms and tracks can be laid..And I bet it cant for one reason..and that is they want all that space for these office towers
    the real drive behind this project

    ReplyDelete
  4. The other alternative to get more platforms is to go vertical. With the substantial idle space on the Mezzanine and the desirability of increasing convenience of Caltrain access and egress, one option would be to bring one side platform on each side of the Mezzanine, with the primary access to the Mezzanine level re-configured, as suggested, to escalators aligned along the width axis, to make contiguous spaces in the (now narrower) Mezzanine that are not split in the middle by lengthwise escalators.

    Another would be to leave the Mezzanine as is, narrow the train box footprint to one island platform and two side platforms, and put in two levels.

    In either case, the vertical separation offers opportunities to use "half-dives" to get better parallel access.

    ReplyDelete
  5. I've heard the TBT folks discuss two level train boxes and they are terrified of the added costs.

    ReplyDelete
  6. @Clem

    Outstanding work as usual, thanks for your ongoing efforts and attention to all the details. I know it's not rocket science, but I'm amazed you have the time to do this! Thanks.

    ReplyDelete
  7. Clem,

    Really interesting informative read.

    Question: Do you ever get a chance to go to public meetings and give voice to this stuff?

    Or at least, do you know if anyone with any actual power over Caltrain and CalHSR reads the blog?

    I am worried that good work like this goes unnoticed.

    ReplyDelete
  8. @Alex

    Without naming any names, it is my understanding that some CHSRA engineering managers do read Clem's blog (they are better informed than they typically get credit for). I do not know about the SF TBT folks, but I hope they, too, are aware of Clem's efforts since it is very difficult to find free advice that is this good and well intentioned.

    ReplyDelete
  9. @Alton, several past iterations of the mezzanine layout showed security screening facilities and waiting areas. I think it is safe to assume that security theater will occur, although perhaps not to the same barefoot extremes as air travel. After all, trains are far less vulnerable than airplanes.

    @David S, the train box could go further east, but it would have to curve slightly to avoid building foundations, offending the CHSRA's stated requirement for straight platforms. In practice, some platform curvature (e.g. 1000 m radius) is no problem; the CHSRA is just struggling with its own poorly developed specification requirement. With a staff of six, I'm not sure you can blame them, although it's still a shame.

    @Bruce, if a 2-level train box is $1 billion more expensive (as the TJPA estimates), then why not build the loop tunnel and turn the TTC into a through-station? Besides, there is room for improvement in their turnback times.

    @Alex, I'm afraid most of the decisions are political, not technical. I'm not very savvy politically.

    @Andrew, please note, many of the creative ideas in this post come from Richard Mlynarik. Credit where credit is due!

    ReplyDelete
  10. @ Andrew, the only version of a two level train box that I have seen proposed is a 2x6.

    Bringing Caltrain up to the mezzanine would be a slightly deeper box, since it would have to have headroom for the trains ... the more expensive of the two eight-platform-track designs would be a 2x4 design, but even that might be substantially cheaper than a 2x6 box, since the outer support columns might not be over the box, substantially reducing the load bearing through the box itself.

    But untangling the access, even a bit, is a big point. The point about the very big station throat and the consequent time required for actually completing a movement that blocks access or egress for other platforms. Simple parallel bi-directional operation does not have adequate capacity, while getting operations where a Caltrain access is not blocked by an HSR egress, or visa versa, is not a straightforward problem to solve.

    And there are statutory requirements. San Francisco 1999 Prop H requires the TBT to be the HSR station, and its arguable therefore that it requires the TBT to satisfy the requirements placed upon the HSR system. CA 2008 Prop 1A requires network headways of 5 minutes.

    Even if that is interpreted for a terminal station as a "unload and send the train elsewhere", as when services have been backlogged by some interruption and the backlog must be cleared ... its not straightforward that there is an operation table that gives 12 HSR trains at the headways they will require through the four platforms together with planned service levels for Caltrain services.

    Its one thing to divide trains by platforms to say trains per hour per platform and just say "get them turned around that fast" ... its another thing to actually be able to one train from a platform track to a tunnel track without being blocked by some other train movement.

    ReplyDelete
  11. I find something that no one seems to be talking about..The out right UGLY design of this place!Enough of the Grand Central of the West
    It is nothing more than a 21st Century version of 1965 PennStation
    All rail elements are below ground and there is nothing more than a small light well. We are building something that the city of New York has been trying to correct for the last 20 years..enough of this poor design..rebuild 4th and King as a real skylite station at half the cost and trouble

    ReplyDelete
  12. This comment has been removed by the author.

    ReplyDelete
  13. Clem: "if a 2-level train box is $1 billion more expensive (as the TJPA estimates), then why not build the loop tunnel and turn the TTC into a through-station? Besides, there is room for improvement in their turnback times."

    That's the 2x6 train box. I think 1x6 is undercapacity, but 2x6 takes the cross-over problems of the 1x6 and complicates it further.

    Bringing the Caltrain up to the Mezzanine is a bit of a bag on a box, with the future option of a loop back or continuing line closed off completely for those two platforms. But it'd be the cheaper of the two.

    The better 8 platform track solution is two levels with two side platforms and a central island each ... the top level with the Caltrain on the two side platforms with the option of continuing (tail, loop-back, or through) using at least the train box side of the tail track footprint.

    That puts the outside load bearing columns of the main building outside or on the outer walls of the train box rather than in columns continuing through the box ... and for the dig, while it is around 35% to 40% deeper, it is also 30% narrower below the Mezzanine.

    ReplyDelete
  14. If Transbay Terminal has security theater, does that mean every other HSR station has it too?

    Note that original plans for Eurostar had trains visiting many European capitals (including hotel train service). Those plans were curtailed due to security theater requirements: the cost to build dedicated, secure "international" platform was way too expensive for a 1-train per day service. This had a big impact on Chunnel ridership, and one reason why Eurostar failed ran into financial problems.

    It seems unlikely that a minor location (like Visalia) will ever see HSR as a result.

    ReplyDelete
  15. Can you imagine security theater in Redwood City and Palo Alto stations?? Millbrae station can use some of its massive swathes of empty parking for security screening. Security screening will also blow the operating budget. If you are just trying to get an express commute from Palo Alto to SF, going through a security screening is going to be a real disincentive.

    Those TBT tailtracks might have been intended to be the hopeful beginnings of a second transbay tunnel. BART likes to use tailtracks as the starter pieces of extensions: "Well, you see we have now built these expensive, under-utilized tailtracks here, so we might as well build further extension." It is annoying that our incompetent regional planners didn't seriously consider rail service continuing on from the TBT to cross the Bay Bridge. Then again, this threatens BART, and we all know that BART cannot tolerate direct competition. MTC calls it "redundancy"; I call it "demonstrably debunking BART's myth of transit superiority".

    Given that the TBT is a stub-end terminal in a city of only 800,000 people, six platforms seem to be plenty, especially with 4th and King still functioning as its own terminal. The notion of 12 inter-regional trains per hour has always been crazy and unfounded.

    ReplyDelete
  16. I'm wondering what the argument is for security theater on the HSR. Japan doesn't have security theater on the Shinkansen. Acela doesn't have security theater either. It's my understanding that the only HSR service that has security theater is the Eurostar through the Chunnel. In that instance, it seems that the asset really being protected is the Chunnel itself.

    If there are good arguments for having security theater on HSR, I would be interested to hear them. Given that freedom from security theater enhances HSR's time-competitiveness with air travel, I would be curious to know who is arguing we need security theater at TBT, and what the rationale is.

    ReplyDelete
  17. Why is the tunnel planned to follow Townsend and 2nd, instead of taking a more direct diagonal route?

    ReplyDelete
  18. While I will never discount the ability of state planners to implement stupid things, I do think Kopp "gets" that one of the time advantages of rail is no security theatre. Trains can't diverted to rogue nations with hostages nor turned into bombs like planes. I think its pretty telling that in Japan there was tons of airline security theatre, yet very little rail security theatre. I think they recognize the cost/benefit there.

    However, all it takes is one state assembly person from a smaller community like Palo Alto to raise a stink about the possibility of "high speed drug train" to get some of this stuff implemented.

    Yet another reason not to put stations in communities that don't want them.

    @Jason - I believe the reason is because going under streets is far easier to manage. No building foundations - if you look at the planned route with google maps overlayed, you see the route avoids the larger buildings. This is why I was curious just how much "wiggle room" there was to move tracks around down there. I'm no engineer; I have no idea how deep the foundations of your average 10 story building go.

    ReplyDelete
  19. @ Jason ... what David said. Only a part of the tunnel is bored, most of it is cut and cover. Digging into the street and then covering it avoids quite a lot of property condemnation.

    Why its 2nd instead of 3rd, which would make for a curve in the tunnel and then the fan-out on a straightaway ... I don't know.

    ReplyDelete
  20. As for security screening on HSR:

    Security is the enemy of efficiency.

    There is essentially no precedent for extensive security for trains and hopefully there never will be. The Eurostar exception is because of the length and cost of the Chunnel and its crossing of international borders.

    ReplyDelete
  21. CHSRA Implementation Plan:

    The high-speed train facilities at each station will consist of tracks, controlled access platforms, full access for disabled passengers and ticketing/waiting/passenger service areas.

    Request for Expressions of Interest
    for Private Participation in the Development of a High-Speed Train System in California, Exhibit B "The Project, including preliminary engineering and EIR/EIS documents, cost estimates, ridership projections and operational information"


    The high-speed train facilities at each station will consist of tracks, controlled access platforms, full access for disabled passengers and ticketing/waiting/passenger service areas. Many stations along the route will have platform tracks off the main high-speed line to allow express trains to pass unimpeded. Stations are expected to have 1,300-foot-long platforms allowing level boarding of the train.

    ReplyDelete
  22. "controlled access" just means that you have to show your ticket/swipe it through a reader to get onto the boarding platform. In other words, no different than BART, or the MUNI stations downtown. "Controlled access" doesn't automatically imply "security theater."

    ReplyDelete
  23. I agree with Bianca's interpretation here, nearly all trains that sell expensive tickets have "controlled access platforms". In fact, JR sells a very inexpensive meet & greet platform-only ticket to allow non-passengers to meet family (like children traveling alone) right at the door of the shinkansen. Many tourists buy one of these just to go see the "bullet trains" without having to pay the fare to ride on them. No security screening, but definitely controlled access.

    ReplyDelete
  24. It would be incredibly dumb to prevent CHSRA from choosing a Japanese design just because the city of SF can't figure out how to make the curves big enough. IMHO, that's a dealbreaker right there. Major case of tail wagging the dog.

    The number of platforms at the TBT is quite adequate. A 30-minute turnaround with 4 available platforms (incl. exit into tunnel) equals 8 tph for HSR.

    The problem is the design of the throat, something I have underestimated in the past. The inbound track needs to split six ways and the outbound one combine six ways, all in a tight curve.

    ---

    Alternative 1:

    Make do with dual track tunnel. Run it up 3rd Street (not 2nd), with the single inbound track veering east under Folsom. Loop this around such that all trains (HSR + Caltrain) enter the train box from the east. Instead of one throat, you now have two, with one-way traffic across all switches and run-through platform tracks.

    All trains exit to the west via a single track under Natoma (not Minna) so the curve back onto 3rd can avoid impacts on the MOMA building. A single track tunnel with ~25' overburden is much easier (= cheaper) to construct underneath existing buildings than one 55-65' wide.

    Result: no traffic jams trying to get into or out of the station. The 3rd St alignment means all platforms can full length at perfectly straight. It should be much easier to avoid the extremely loud screeching of wheel flanges against rail flanges, which will resonate in the train box - some passengers will stay away for that reason alone!

    If need be, decades from now, add an auxiliary terminal (underground only) with platforms between Folsom and Clementina, with moving walkways linking its concourse level to that of the main terminal.

    ---

    Alternative 2:

    make do with a dual track tunnel under 2nd St. Fork it at Tehama, with two tracks continuing straight toward Market with a crossover X. Use these for Caltrain.

    The other fork continues to the TBT with a crossover X, followed by a pair of three-way points to support six platforms used exclusively for HSR.

    All of these should be dead straight and pushed as far east as possible - even 100' is valuable - to enlarge the curve radius in the throat as much as possible. Discard the option of future tail tracks down Main.

    ---

    Alternative 3:

    Split off the two-track DTX tunnel at Mariposa, veer toward the Bay under China Basin, dive under Mission Creek, stay east of the ball park, run under Embarcadero and up Main.

    Fork the tracks to terminate Caltrain at Market, intermodal with Embarcadero BART.

    The other fork enters TBT train box from the east, where there's more room for crossover X + triple points. Platform tracks again dead straight but pushed west as far as possible.

    ReplyDelete
  25. Following up on Bianca and Andrew's points, the platforms for Amtrak intercity services at Boston South Station, New York Penn Station, Philadelphia 30th St. Station, and Washington Union Station are generally controlled access platforms. But there are no metal detectors or security screenings.

    Given that Acela Express, which runs through both the largest city in the country (NYC) and the nation's capital (DC), has no security theater, it would be rather absurd to include it on the California HSR.

    ReplyDelete
  26. @mike, I remember being rudely shooed off the South Station Acela platform where I took my infant son to look at the trains. I started arguing with the conductor, he went on about September 11th, I went on about pushing a stroller for cryin' out loud, and he threatened to call the cops. I decided it was curtain time for that little bit of security theater!

    Mind the gap (between perceived risk and actual risk)

    ReplyDelete
  27. "Controlled access" implies rider-hostile stations, increased accessed time, reduced accessibility, and massively increased station capital costs.

    Check out the horrible mess they're making of Amsterdam (screwing up public access routes in order to turn stations into prisons -- as in California, it is the prison guards (fare system vendors) who run the show), or compare metros and stations in German-speaking parts of the world with those in the will-never-get-it Anglophone world.

    Look at the scale and cost and urbanist brutality and user hostility of BART stations and compare them with open, community-accessible, transfer-time-minimizing S-Bahn/RER stations pretty much anywhere in central or northern Europe.

    Mezzanine levels! Fare gate banks! Multiple fare gate banks! Shoulder-height walls around all stations! Razor wire! More and deeper excavation! Parallel and redundant and expensive "air-side" and "land-side" circulation areas. Ceiling-height prison bars! One-way turnstiles! One underpass/overpass for passengers, another for The Community. Sharply limited access points into the surroundings, each with the full fare barrier apparatus and complete with a full-time do-nothing "security" nanny "station agent".

    Pork pork pork pork pork pork pork all for no negative benefit.

    And as a bonus you get to build completely useless "waiting areas" on the airside. (In advanced civilized industrialized first world democracies, they have these things called "cafes" and "restaurants" and "bars" and "shops" which serve this purpose.) And then we need separate "waiting areas" for HSR and Caltrain. And separate fare barriers for HSR and Caltrain! This is just awesome! Construction and program management budgets on steroids!

    All in all, this is a great deal if you're paid to build this worthless and hostile junk (especially if you get to set the desgin specifications of just how much of it is "necessary"), but it's simply a terrible deal for human beings who wish to ride the trains or even just visit the stations. Or have to pay to fund this contractor welfare scam.

    ReplyDelete
  28. I share Richard's skepticism about security pork, but the real risk is not the short-lived benefit to construction contractors for yet another fence, it is the long term employment offered by useless "security" personnel that the unions all love. Jobs are only good for the economy if they are productive, most security jobs are not.

    The TSA, which will no doubt be salivating about all those potential extra screening jobs, needs to be kept firmly away from this project. They bring extraordinary costs and almost no benefit (unless you mean pension benefits).

    ReplyDelete
  29. High Tech Crossings27 March, 2009 17:04

    This is similar to the story of BART. Like CHSRA, BART started with a tiny skeletal staff with virtually no technical experience. BART was all hype and PR. The consultants (surprise, surprise: PBQD-Tudor-Bechtel) did all the designing with no oversight, and they created a bloated customized, expensive system. Great for the contractors; bad for the public! Now, BART has something like 2800 well-paid employees for only 100 miles of transit service, and their unions are running the show. See BART the institution in action with their handling of their recent murder.

    ReplyDelete
  30. So what does anyone think of the actual look and design of the building? or does it not matter?

    ReplyDelete
  31. I think we can all agree that security screening is pointless, but I don't understand why you think it is a bad idea to separate caltrain from HSR.

    After all, in Japan the Shinkansen tracks are separate from regular JR tracks.

    ReplyDelete
  32. I don't understand why you think it is a bad idea to separate caltrain from HSR.

    Separating different classes of traffic is fine when (a) one has the luxury of real estate to do so (tracks, platforms, junctions, flyovers...) and (b) it makes any sort of operational sense.

    For example, high speed traffic generally gets separated from lower speed as soon as possible after leaving a constrained, quasi-uniform-speed urban setting (which on the SF Peninsula would mean at Redwood Junction, not Gilroy) in order to maximize capacity and throughput.

    But Caltrain+HSR are inherently mixed traffic to the northern terminal. Unmixing them just to make one stop can only consume capacity, because doing so involves unnecessary crossing moves to place trains at or extract them from non-optimal platforms.

    Imagine that a slightly delayed HS train is arriving on the right-hand track and there are two empty platforms at the terminal (platform tracks 3 and 6, arbitrarily numbering from north to south, ie left to right on approach), and meanwhole a Caltrain is departing form track 5 (as one is every 7 minutes). In a competently configured station throat it would be possible to berth the incoming train at track 1 without delaying the departure. (Of course the later departure from track 1 will involve crossing all the way over, but we deal with that when it comes, and it comes inevitably.) But if HS trains were arbitrarily forced to not use track 1, there's an unavoidable delay to either the arrival or departure. (Or if TJPA, Caltrain and PTG design your station, there's a delay anyway, because tracks 1/2, 3/4 and 5/6 each pass though unnecessary single-track bottlenecks.)

    Now for customer convenience (and operating predictability) it certainly is nice to be able to say "the Caltrain local departs every 15 minutes from track 1" (or "from track 1/2", on opposite sides of the same platform island), but at facility as space constrained and as over-subscribed as the Transbay Terminal station you need every possible piece of operating flexibility because there will be nearly zero spare capacity that can be chewed up through the luxury of operating stupidly if there's every any sort of delay. (Modest 12tph = an arrival or departure every 2.5 minutes, for hours at a time. Schedule non-conformances of over two minutes are going to happen, guaranteed. Then what?)

    In and out as rapidly as possible, stopping as little as possible (at stations or elsewhere), and having as many trains as possible on the move at once (rather than getting in each others' way) should be the name of the game. If one is paying nose-bleed prices to build infrastructure it should be in support of revenue service, not of trains parked idle, or trains stopped waiting.

    Of course this model of throughput and efficiency maximization is the exact opposite of what our US Transportation Professionals are up to.

    ReplyDelete
  33. @Alex: in Japan this is done by necessity because the high speed trains run on standard gauge tracks, and the regular trains run on narrow-gauge tracks (for historical reasons). In San Francisco there's no need to do that, and I think I explained fairly thoroughly why it is important to have operational flexibility when a site is so constrained.

    @yeson1a: form follows function. If the design doesn't yet meet the basic function of the building (from a rail perspective) there is little point in worrying about the "look" of things before it's fixed.

    Regarding security, here is the clearest evidence I have of the plans for security theater. It comes from a late 2008 floor plan of the mezzanine, one option that was part of the 90% design package. Notice airport-like screening and waiting areas.

    The TJPA's floor plans have been all over the map; it's hard to identify a design baseline, let alone a clear path of design evolution that got them there.

    ReplyDelete
  34. @ Clem -

    if there is a real security threat to HSR, it is to the rails and perhaps, to the stations. Long-distance trains simply don't have the passenger density of subways or commuter trains during rush hour, so a single small bomb can't hurt or kill a lot of people.

    France has had to deal with a number of bomb threats against its TGV lines in recent years, e.g. by one from an radical animal rights group. By contrast, the last time someone (Carlos the Jackal) brought a bomb on board a high speed train was in 1982.

    It would be foolish to pretend that terrorists cannot or will not attack a high speed train. But it's equally foolish to assume they will do so in the same way they attack airplanes. Airport-style security is already practiced at Eurostar terminals and also on Spain's AVE network, but that's more to provide some visible means of "doing something" - a placebo, if you will - than actually achieving a significantly higher level of security.

    What HSR needs is fences (or sound walls that double as fences) + CCTV + other measures. Those are less visible to the public but they are standard features of many HSR lines.

    Indeed, CCTV + microphones could also be very helpful in addressing UPRR's concerns about detecting derailments of their freight trains early enough to avoid follow-on accidents involving high speed trains carrying hundreds of passengers.

    ReplyDelete
  35. @Clem

    While I strongly oppose airport style security for HSR (and for airplanes for that matter), I actually think it is good that some of the floor plans for TBT have tentatively included security screening areas. The only thing worse than having security screening is having security screening added to a facility that is not designed for it. Dallas Fort Worth Airport (DFW) is an absolute design disaster with regard to having added airport security after the design of the facility. Some other older US airports are little better (like EWR and JFK).

    ReplyDelete
  36. Richard's design is actually a mass transit design as opposed to a modified Pennsylvania Railroad PTG Utility Relocation Design Build Fiesta like the current DTX LPA. His design could actually work, but to make powerful statement it would require a nice scale model or software simulation to validate its assumptions and inform actual decision or policy makers of the benefits.

    One potential chess move to be on the look out for is that Richard's design will require some sort of compatibility with a future Transbay Crossing. A lot of the funding for Transbay stipulates this in the legislation. The current loop track could address this but in my opinion it is not buildable and would require absurd, life-span shortening interaction with Caltrans staff for eons due to the Bay Bridge footings. Watch out for MTC & BATA Friends on that funding issue.

    I would imagine that there is a very significant cost savings in the 2-track tunnel. One of the things I have heard from the engineers working on this project is that there are a lot of 'mystery' utilities under Second Street. PG&E doesn't even know exactly what they are linked to and some of them are pretty major. Any reduced scope in this area would seem to reduce a lot of risk and is just common sense.

    ReplyDelete
  37. @Clem
    Well if form follows function then we have A MESS on all levels

    ReplyDelete
  38. @rafael
    By making the TBT a through station for HSR, your are in essence eliminating tail tracks and decreasing the min max dwell time to 18 min for a 4 track station and 28 min for a 6 track station. Without tail tracks, a through TBT only makes sense if the majority of trains are proceeding across the bay into Oakland. Such short turnaround times for a HSR terminal station just isn't logistically reasonable.

    Alt 1: Using Folsom St. doesn't help as the maximum curve radius remains about the same, 500ft (150 m).

    Alt 3: If you are going to go to such lengths to make the TBT site work, then the idea of using the TBT as the HSR terminus needs serious rethinking.

    If we're going to think crazy, then how about this: 1) make Caltrain the only rail serving TBT (requires wriggling out of 1999 Prop H), 2) lower 4th & King station 2 levels with 2 island platforms for Caltrain and 2 island platforms for HSR, 2) add 1 island platform underneath Townsend for Caltrain and tunnel underneath Townsend to Essex, First, or Beale to the TBT, 3) add 2 island platforms underneath King for HSR with provisions to extend past the Ballpark and underneath the bay, 4) new Geary BART line via Brannan and 5th or 7th with 600 ft underground passageway underneath 4th connecting it with 4th & King station.

    ReplyDelete
  39. @Herbie, Rafael

    I don't think there's a need to discard the existing concept. It just needs to be improved and optimized, because as currently planned, everyone can agree that it just won't work.

    Going to a bi-level station or additional tunnels is the most expensive answer. The least expensive answer is to modify the design along the lines Richard is suggesting (wider radius for faster speeds, longer train box to relieve the throat curvature, much shorter throat with minimal platform re-occupation delay, etc.) Like I said, it's not rocket science, but it will require thinking outside of the box of Caltrain design standards and US freight rail traditions-- and perhaps introducing such exotic European devices as curved turnouts. (a hushed murmur descended upon the shocked audience...)

    We need to do another post on this.

    ReplyDelete
  40. Pennsylvania Railroad PTG Utility Relocation Design Build Fiesta

    The PRR designers would have done a much better job. Penn. Station in New York is a horror show but considering it's 100 years old, the designers expected to need a major overhaul by 1960 and that it still works, not well but it still works, they did a pretty good job, They didn't expect that the major overhaul would be to rip it all down and build an office building either. Other places where they had more space and better budgets did very well. Penn. Station in Newark is holding up reasonably well. I'm sure commuters who use it would have suggestions but implementing those would inconvience other riders or make the station a block wider. 30th Street in Philadelphia solves lots of problems they were facing and does it reasonably well. Compromises were made as in Newark but things work well for most people.

    His design could actually work, but to make powerful statement it would require a nice scale model or software simulation to validate its assumptions and inform actual decision or policy makers of the benefits.

    Everything he says makes sense. Even his suggestion to put the local tracks in the middle makes sense. I think he would agree that having the locals in the middle isn't going to happen on the Peninsula because to implement it would cost too much, a lot more land would have to be condemned. Railroad designers don't have to see computer simulations of what he describes, he's describing things that have been worked out over the past 150 years. Commuters recognize it because they experience the principals in use everyday. People who have only seen trains in movies need the computer simulations. Should be reasonably easy to do because railroad designers and managers have simulators to check their work with.. ones that produce nice 2D or even 3D representations of it. The trouble would be convincing a consultant, who has that kind of software, that you really want to blow $50,000 dollars on having him produce one... that 20 trains an hour in each direction on a mostly 4 track railroad... is the kind of things railfans sketch out on the back of envelopes over a cup of coffee.... that make sense to people who have ridden on a railroad

    exotic European devices as curved turnouts. (a hushed murmur descended upon the shocked audience...)

    Why not,? They are going to be using exotic electric trains without locomotives! With exotic standard gauge track! With level boarding and all sorts of other 20th Century innovations. ,,,though aren't curved turnouts, level boarding, electric trains etc 19th Century innovations?

    ReplyDelete
  41. At the risk of revisiting old battles, one of the reasons I preferred the Skidmore, Owings and Merrill proposal for the Terminal was that they seemed to pay much more attention to the train box, giving it much more space and light, rather than seemingly tacking it on.

    (http://www.som.com/content.cfm/transbay_presentation -- You will have to select the "Terminal" portion of the presentation)

    ReplyDelete
  42. How does Transbay compare to Los Angeles Union Station as a potential HSR station? I know this blog focuses on Caltrain, but what kind of adjustments are we to see at LAUS besides run-through tracks?

    ReplyDelete
    Replies
    1. Spokker- I know that this is a belated response, but here it is nonetheless. The Transbay Redevelopment Project is actually the more ambitious of the 2 projects, according to my sources at Caltrans. The LA Master Plan calls for an addition of 3.2 million square feet of mixed use housing, office, and retail space. This includes hotels, towers, and retail adjoining Union Station as well as expanding Union Station to just under 500,000 square feet with the addition of a lower concourse featuring 144,000 square feet of retail space, and a bus ramp above ground. At last count, the Transbay Redevelopment Program which began a few years ago is currently at 9.5 million square feet of retail, office, housing, open space, and counting. This includes Salesforce Tower, 181 Fremont St. Tower, Oceanwide Center Towers ! & 2, other highrises, Transbay Park + surrounding condominiums, and of course, the demolished and nearly re-built Transbay Transit Center. When completed, it will be more than twice the height and twice as long(1500 feet) as the old terminal. It will feature 166,000 square feet of retail and dining, of which 100,000 will be leasable. In addition to the new train platforms, a bus ramp w/wide overhead roadway featuring a mock bridge span(an ode to the area's 2 famous bridges) will whisk buses onto and off of the freeway. The most anticipated attraction of all will be the rooftop park. Spanning a quarter of a mile, it will feature gondolas that will transport visitors from the street onto the park, cascading waterfalls, trees from all over the world, a small track for running, a play area for children, 2 rooftop eateries including a 14,000 square feet restaurant, and a 300-feet long, 800-seat open air amphitheater for live concerts and performances. When completed, it will be the largest rail station west of the Mississippi at 1.5 million square feet. This monstrosity of a structure will dwarf post-renovation Union Station by over 1 million square feet.

      Delete
  43. At the risk of revisiting old battles, [...]

    Note that the teams (a whole two of them, out of an earth-shattering total of three competitors) that attempted to do any design whatsoever were actively punished by the "architectural" "design" "competition" "jury".

    What a tragedy. What an immense squandered opportunity. How perfectly typical of how we do things around here.

    Slap some lipstick on that pig!

    ReplyDelete
  44. LA Union will still be a grand station with HSR..even better..The people here are debating the stupid tracks and such..3 inches above that does not matter..kow quality forward thinking

    ReplyDelete
  45. @ Rafeal: "The number of platforms at the TBT is quite adequate. A 30-minute turnaround with 4 available platforms (incl. exit into tunnel) equals 8 tph for HSR."

    I still think you are lapsing into assuming sufficient operational flexibility so that the capacity is for a 30-minute average dwell, with exceptional delays being caught up with incrementally by rushing the turn-around of the next few trains.

    Given the operational inflexibility implied by crossing over incoming and departing trains, if 30 minutes (four minutes faster than the fastest European HSR turn-around that DoDo at EuroTrib is aware of) is an average, than 30+ minutes would be the required maximum dwell. 35 minutes to capture a majority of the extraordinary delays and avoid interfering with the operations, add three or more minutes between platform clearance and platform arrival, and you are right about at the 40 minutes that CHSRA requested.

    IOW, 40 minutes is by no means an unreasonable request for the type of service where passengers must finish debarking before passengers can embark, including the servicing to prepare for a three to four hour service. Obviously when Caltrain electrifies, they can adopt 3 door bi-levels and have the two outer doors dedicated to peak flow and the central door dedicated to counter-peak and get very fast turn-over ... and they have less turn-around servicing required and, given the length of their route, there's no reason that the full servicing has to occur at both ends of the route. Indeed, if the SF-origin trains stable elsewhere, there would be no need for any full turn-around servicing of Caltrain services inside the TBT train-box.

    Two different eight platform train boxes.

    ReplyDelete
  46. @ BruceMcF -

    I don't think cleaning and provisioning Caltrain locals at SFTT was ever on the menu.

    And as I said, other than prop H (1999) there is no reason for Caltrain to be at the SFTT at all. The only other option voters were presented with was to keep Caltrain at 4th & King.

    Caltrain locals are commuter trains, they need to be at Market St - preferably intermodal with BART, SF Muni subway and SF Muni streetcars. Market and 2nd would do nicely, giving HSR six full-length platform tracks if they're smart enough to push them east beyond Beale.

    The throat into the SFTT would need to have a crossover X (4 regular points) and two three-way points.

    The throat into the Caltrain terminal would need least two tracks and one crossover X.

    Having two X's would ensure that all trains are on the correct track in the section of the two-track DTX shared by both services.

    ReplyDelete
  47. @ Rafeal: "And as I said, other than prop H (1999) there is no reason for Caltrain to be at the SFTT at all."

    Given the transit interconnections that will be put in place (some sooner, some later) for the HSR terminus, there will be a transport demand for at least some Caltrain services going to the TBT.

    And in any event, given that the application for funds for the TBT train box is going to go in very shortly, and other than that they do not have the funds nailed down to build either the undersized train-box or over-sized tunnel that they have planned, a resolution of the problem by giving short shrift to those transport demands and ignoring SF'99 Prop H and taking Caltrain out of the TBT train box is not remotely plausible.

    And what main stakeholders are you going to have put their hand up for that solution? It sets the Caltrain services at two transfers from the the various bus services pooled at the TBT, TBT are going to simply stand on Prop H, and given the amount of cooperation that the CHSRA needs from Caltrain, its not likely they will do something that can be so easily portrayed as stabbing Caltrain in the back.

    ReplyDelete
  48. [...] the capacity is for a 30-minute average dwell, [...] 40 minutes is by no means an unreasonable request [...]

    At some point you have to decide whether the goal is to build an underground parking lot for out of service trains, or whether you're trying to maximize public benefit by building an efficient system for providing transportation services to human beings.

    If a passenger train can't be reliably turned in well under 10 minutes that simply suggests that either your train operator is so egregiously incompetent that it can only be operating as an contractor welfare operation (crews get paid whether trains are moving or not, companies get paid to build extra trains to sit idle rather than provide service, and taxpayers get soaked for the lot) or that your station design (passenger circulation, platform widths, etc) is so incompetently inadequate that you shouldn't be allowed anywhere near the public's money and should under no circumstances be allowed to advertise youself as an architect.

    Visit Leipzig or Frankfurt a.M. or München or Zürich or Luzern or Den Haag CS or Fenchurch Street or Waterloo or Göteborg or ... sometime.

    What possible justification can there be for wasting four billion (or more) of public money on a rail line that is operated like something about of the 19th century?

    Time is money.

    ReplyDelete
  49. "If a passenger train can't be reliably turned in well under 10 minutes ... Visit Leipzig or Frankfurt a.M. or München or Zürich or Luzern or Den Haag CS or Fenchurch Street or Waterloo or Göteborg or ... sometime"

    What is the HSR route in Europe that turns around in 10 minutes? I noted that in the DTX/trainbox presentation to the board, there were lots of international comparisons on a wide range of points ... the justification of the turn radius by citing existing HSR sets for which they are in spec ... but none for the issue of HSR platform dwells.

    DoDo at Eurotrib is aware of one turn-around of 34 minutes, the former Metropolitan between Cologne and Hamburg. None of the other European turn-arounds seem to be as short as the time CHSRA is requesting.

    And of course, given the operational inflexibility of the TBT design, a targeted turn-around of 34 minutes would require some slack to prevent any minor delay from causing a break in the planned operational cycle.

    To get inside 30 minute HSR platform dwells, you've got to go to Japan, and a design that targets matching the outlier is a design more likely to fail than to succeed.

    ReplyDelete
  50. Bruce: I don't know about HSR, but in the US, half an hour is considered enough time to turn around an Amtrak corridor train (such as the Surfliner in San Diego), and the stop in New York Penn is 20 minutes for Regionals and 15 minutes for Acelas, which is enough time for about half the train to disembark and be replaced with new passengers. It's also considered to be enough schedule recovery time, even given the two very tightly scheduled commuter corridors on either side of New York, one of which is owned by Metro North (and if the Amtrak is late arriving on their territory, it might well get put behind a commuter local). I'd say about three HSR trains per platform per hour is reasonable, and at least four Caltrains. With a sane track layout (look at Fenchurch Street, for example, or even the existing 4th/King terminal), the proposed six tracks should be plenty to meet the needs of both Caltrain and HSR for at least a few decades. And if they aren't, well, that's why 4th/King will stay around.

    ReplyDelete
  51. What is the HSR route in Europe that turns around in 10 minutes?

    Where do you come up with these bizarre and counter-factual but assertively stated suppositions?

    Start with "pretty much any ICE or EC stopping at Leipzig Hauptbahnhof."
    There are thousands of other examples.
    eg ICE 692: ... -- Stuttgart 08:47 08:51 -- Frankfurt(Main) 10:08 10:13 -- ...

    I noted that in the DTX/trainbox presentation to the board, there were lots of international comparisons on a wide range of points ...

    "Cherry picking" and "garbage in, garbage out."

    Ask a deliberately misleading question and you can solicit the desired answer.

    Don't ask: "Could the TTT rail alignment have been engineered to have higher capacity and fewer constraints?" but instead ask "Is it possible to run a train into TTT the way we've designed it?" and present a "yes" answer as a full endorsement of every aspect of your highly professional work.

    DoDo at Eurotrib is aware of one turn-around of 34 minutes, the former Metropolitan [...]

    Where do you find these foamers? Oh right: inside the borders of the USA, with non-existent or very lightly used passports.

    To get inside 30 minute HSR platform dwells, you've got to go to Japan [...]

    Like, whatever.

    Just FYI, here's an actual quotation from a highly-paid professional rail engineering consultant with "international experience" credentials who worked for the TJPA, explaining half-hour-plus dwell times and the irrelevance of any non-Anglophone precedent:
    "Asians don't value life the same way we do."

    And no, I am not making this up.

    ReplyDelete
  52. ""DoDo at Eurotrib is aware of one turn-around of 34 minutes, the former Metropolitan [...]"

    Where do you find these foamers? Oh right: inside the borders of the USA, with non-existent or very lightly used passports.
    "

    Hungary, USA, little difference. In either case you can go quite a ways without using your passport much, though of course from Hungary you don't have to fly to connect with an HSR.

    He's the only fellow working in rail in Europe I had on hand to ask.

    ReplyDelete
  53. Richard, DoDo is Hungarian and works in Germany, I believe.

    On the other hand, there really is no excuse for 30-minute turnarounds, at least for Caltrain. The New York City Subway takes 2-3 minutes to turn around its mainline trains and 1 minute to turn around its shuttles. Intercity trains might need to be cleaned every time they reach a terminal, but commuter trains don't.

    ReplyDelete
  54. There are thousands of other examples.
    eg ICE 692: ... -- Stuttgart 08:47 08:51 -- Frankfurt(Main) 10:08 10:13 -- ...


    For those unfamiliar with Germany, it's useful to point out that Stuttgart and Frankfurt are both stub terminal stations, where the ICE high speed trains reverse direction without going out of service after four- to five-minute platform dwells.

    That's eight to ten times quicker than the 40-minute timing that is considered such a terrible limitation at Transbay.

    California HSR services could be operated LA-SF-LA, for a total journey time of about six hours. That sort of service duration is not a big deal for the equipment or crews, from the standpoint of resupplying the cafe car, emptying the toilets, cleaning the seats, or whatever. All those routine maintenance operations can and should be done elsewhere than San Francisco, where time and space are at a hefty premium due to the highly constrained location.

    A two-level train box belongs to a class of solutions involving lots of excavation, concrete, and astronomical expense--and by implication dismisses the much simpler and smarter solution of dealing with the capacity constraint through efficient operations, just like the Germans do. As Richard said, work smarter, not harder.

    The notion that a train must necessarily go out of revenue service upon "terminating" at the transbay "terminal" is causing the mental block.

    ReplyDelete
  55. To do other people's homework for them as usual:

    Here are a just few hours of scheduled ICE and IC (InterCity Express high speed trains and InterCity loco-hauled ~200kmh long-distance trains) that pass though the stub-end Frankfurt (am Main) main station (Hauptbahnhof) in the first few hours of Tuesday 31 March 2009
    (train number, platform, arrive, depart)

    ICE 826 6 05:35 05:44
    ICE 521 9 05:46 05:51
    ICE 672 7 05:50 05:55
    ICE 1092 9 06:00 06:14
    CNL 420 3 05:56 06:15 (two halves of sleeper train...)
    CNL 40400 3 05:45 06:15 (... join at platform)
    ICE 523 7 06:48 06:54
    ICE 774 8 06:52 06:58
    ICE 824 10 06:56 07:02
    ICE 874 8 07:08 07:13
    ICE 2273 13 07:33 07:38
    ICE 2028 7 07:36 07:42
    ICE 591 9 07:44 07:50
    ICE 525 6 07:48 07:54
    ICE 670 8 07:52 07:58
    ICE 822 7 08:05 08:10
    ICE 694 9 08:08 08:13
    ICE 23 6 08:13 08:18
    IC 2376 13 08:18 08:23
    IC 79676 6 08:41 08:42
    ICE 373 9 08:44 08:50
    ICE 527 7 08:48 08:54
    ICE 772 8 08:53 08:58
    ICE 517 6 09:00 09:05
    ICE 720 7 09:05 09:10
    ICE 872 7 09:08 09:13
    ICE 1557 9 09:15 09:21
    ICE 1097 9 09:28 09:34
    IC 2275 13 09:33 09:38
    ICE 1026 6 09:36 09:42
    ICE 1091 9 09:42 09:50
    ICE 529 7 09:48 09:54
    ICE 78 8 09:52 09:58
    ICE 71 6 10:00 10:05
    ICE 728 7 10:05 10:10
    ICE 692 9 10:08 10:13
    ICE 25 6 10:13 10:21
    IC 2374 13 10:18 10:23
    IC 79674 6 10:41 10:42 (1 minute!)
    ICE 375 9 10:44 10:50
    ICE 621 7 10:48 10:54
    ICE 770 8 10:53 10:58
    etc etc etc.

    There are no 40 minute dwells anywhere to be seen, not even for the two different loco-hauled sleeper trains City Night Line) trains that arrive from Wien and Milano, joined up on platform 3 of Frankfurt HB, and then headed out with a final desination of Dortmund.

    The normal scheduled ICE dwell time, including reversing, is six minutes at one of the most important and heavily-used stations in the continent (meaning lots of people getting on and off each train.)

    ReplyDelete
  56. "A two-level train box belongs to a class of solutions involving lots of excavation, concrete, and astronomical expense--and by implication dismisses the much simpler and smarter solution of dealing with the capacity constraint through efficient operations, just like the Germans do."

    SO under that scenario, a final route to SF could run a boomerang route to, say, Merced and terminate (especially convenient if Castle is a maintenance center), making the TBT a through station without the through tracks.

    But could the current TBT design maintain 5 minute headways for the HSR without interfering with Caltrain services? If its acting as a through station, then it cannot bow out of the network requirement on the excuse of being an origin/terminus of the network.

    That is 20 minutes dwell plus headway, which sounds plausible for a through station, even one operated by Americans and including slack for contingencies, but can the DTX and the station throat support those headways?

    Or, if it cannot, can the design of Richard's?

    If the former cannot and the second can, that would be a point of leverage for the second.

    ReplyDelete
  57. Alon Levy said..."On the other hand, there really is no excuse for 30-minute turnarounds, at least for Caltrain."

    Where does 30 minutes come from? Even at the 6tph Caltrain is suggesting, with 2 platforms that is a maximum 20 minutes less time between clearing and occupying the platform ... and if they are running one class of service to the TBT, 6 services per hour is certainly a reasonable frequency, putting them over the threshold from a memory timetable to an arrive and go timetable.

    ReplyDelete
  58. Guess I'm starting to see why CAHSR was so against Altamont alternative.

    If they are too retarted to figure out how to turn around trains on just a single line, imagine how much more difficult an exercise this would be for simple minds to grasp running more than a single HSR line into the Transbay terminal.

    ReplyDelete
  59. I think the sane thing to do is have almost all (LA-SF) HSR trips have LA as both origin and destination, with just a quick (15 minute?) turnaround at SF. The only case where that won't work is at the beginning and end of the service day, and maybe at the beginning and end of the peak (if traffic is peaky enough). In those cases, the trains would be serviced wherever they are stored overnight or between peaks (4th/King yard, or Bayshore, or whatever), and still would do a fairly quick turnaround at the SF terminal itself. As for Caltrain, 10 minute turnarounds are considered normal for many commuter agencies, so two platforms for 8 tph should be enough. With a sane track layout and adequate passenger access to the platforms, it seems to me like the number of platforms at SFTT is adequate.

    ReplyDelete
  60. arcady said...
    "I think the sane thing to do is have almost all (LA-SF) HSR trips have LA as both origin and destination, with just a quick (15 minute?) turnaround at SF."

    Continuing services through the TBT eliminates the platform bottleneck, but that strengthens the argument that it falls under the institutional requirement from Prop 1A for 5 minute headways through the HSR network. And of course the TJPA has a strong financial incentive for the DTX to qualify as part of the HSR network, since that's how they hope to get the train-box funded as part of the original foundation of the TBT.

    Also, it would seem from the powerpoint presentation to the TJPA Board that they made claims on meeting the requirements of SF 1999 Prop H while avoiding mention of the Prop H mandate of the TBT as the San Francisco HSR station ... and so meeting Prop 1A mandates regarding the HSR system, once they were established, might be argued to be an implication of that local mandate.

    Under the TBT design, can an HSR service enter the DTX every five minutes without interfering with the ability of Caltrain services to use their platforms? This post suggests that the station throat is a bottleneck, with a tight curves in the throat implying relatively slow operating speeds, and all crossover switches in the tunnel itself, behind the 2:1 fanout of the two sides of the island platforms, 3/4 of a mile or more from the platforms.

    Looking from the platform to the tunnel, as near as I can make out the TJPA proposal, the leftmost island will be the Caltrain services. If Caltrain services cannot slot in or cross over while maintaining 5 minute headways for the HSR services, then that would relegate Caltrain to use one tunnel track as bi-directional track (pushing platform capacity even further away from being a constraint).

    While the TJPA can claim to have met statutory requirements, they need to be able to provide operational details that prove it. The performance at the Senate hearing sounded much more like bluffing with bullshit than proving it.

    ReplyDelete
  61. @ Richard Mlynarik: "Ask a deliberately misleading question and you can solicit the desired answer.

    Don't ask: "Could the TTT rail alignment have been engineered to have higher capacity and fewer constraints?" but instead ask "Is it possible to run a train into TTT the way we've designed it?" and present a "yes" answer as a full endorsement of every aspect of your highly professional work.
    "

    But if the TBT train box is to be within the HSR network, the acid test is whether its is possible to actually operate the train box to bring HSR services in on 5 minute headways while also support Caltrain services.

    I of course don't know that they haven't answered that, I just know that I haven't seen a clear and persuasive positive answer.

    If their layout can't and your layout can, it would seem that CHSRA would have substantial leverage under SF 1999 Prop H and California 2008 Prop 1A to shift them to your design.

    (Although I realize that you are as scathing of CHSRA as of the TJPA.)

    ReplyDelete
  62. BruceMcF: these mandated 5 minute headways, are they for the entire HSR network, or only for the "trunk" line, between LA and the Central Valley Wye where the SF and Sacramento branches will split. Because if you have 5 minute headways on the core, that means fewer trains for each branch. Even with a 1:2 split, that means only 8 tph for SF. In terms of platform space, that should be doable. In terms of interlocking capacity, you'd really need to do simulations to see if it all works out, but for example Flatbush Avenue on the LIRR accommodates 7 inbound movements in the peak half hour (most of which presumably turn around to either go back out or to the yard), so I'm optimistic that a properly designed SFTT can handle 16 tph.

    ReplyDelete
  63. @ arcady, the entire HSR network: "2704.09.(pdf) The high-speed train system to be constructed pursuant to this chapter shall be designed to achieve the following characteristics:
    ...
    (c) Achievable operating headway (time between successive trains) shall be five minutes or less.
    "

    In terms of trunk capacity versus branch capacity, both trunk and branch have to have 5 minute headways ... even if a branch would not likely every receive 12tph because of the maximum capacity of the trunk, it still has to be able to accommodate two successive trains at 5 minute intervals.

    I'm not a lawyer of course, but it seems arguable that is doesn't really mean a 12tph requirement for a terminal, if there are not 12 trains that need accommodating ... but certainly, if there are six trains or eight trains per hour to accommodate (*frequency note), it does seems to say that they have to be able to be accommodated as:

    :00, :05, :10, :15, :20, :25 ...

    What that means for mixed traffic is really something for lawyering ... it might mean anything from:
    * (1) five minute headways for whatever trains are in the mix
    * (2) a segment is able to accept HSR trains at five minute headways and deliver HSR trains at five minute headways, and if it uses shorter headways (say, 3 minutes) to shuffle additional traffic in within the segment, that's OK
    * (3) five minute headways for HSR trains must be maintained throughout

    (NB. I take 3 minute headways as a crude benchmark because the Cityrail system in Sydney can reach 3 minute headways, and they always whinge about it being crappy old signaling, switching and etc. that needs to be upgraded. If a new design cannot match the long under-funded and increasingly decrepit Cityrail system, its being designed broken from the start.)

    (1) is most convenient to the TJPA strategy of getting recognized as a bona fide part of the HSR network with the minimal possible train capacity, (3) would give the most leverage to the CHSRA, but might be inconvenient in terms of inter-operation with Caltrain elsewhere in the corridor, and (2) sounds to me like a plausible reading of the original intent of the clause.

    And it should be noted that certainly under reading (2), the Prop 1A requirement is a good thing ... if you allow local pollies of various sorts to hire engineers to meet the minimal possible requirements, you get all sorts of local limitations that hobble the scheduling that best builds and then taps market demand.

    (Or in the terms of the profession, there are massive principle/agent problems in building a system like this, and if we allow each local authority to optimize for local goals, we are guaranteed to be far from optimal at the system level.)

    "so I'm optimistic that a properly designed SFTT can handle 16 tph."

    If the Caltrain peak is 1 train every 10 minutes, (2) boils down to the ability to bring 2 HSR services out, 2 HSR services in, 1 Caltrain service out and 1 Caltrain service in a 10 minute interval, in the context of some actual workable rotation of platforms.

    The topic of the original post, way up there at the top ... is whether there is a proper rail design access for the SFTT train-box. If its designed badly enough, it might run afoul of the headway requirement.

    So how is the TJPA proposing to meet that requirement? Do they even have a proposed operating schedule to meet that requirement? And if they do, what are the hidden gotchas attached? That needs to be flushed out into the open.

    Working through a variety of platform combinations, it seems that Richard's layout (pdf) would not be strained in meeting (2), much less (1).

    Obviously with a two track tunnel for most of the length, it would have a lower project cost ... and pushing for the design that costs less "to the taxpayer" would put CHSRA in a much better position in the public perception war than they are at in the moment.

    So while on the one hand I wonder whether the CHSRA is making an unsustainable ambit claim when using that provision to demand 12tph ... on the other hand, it seems as if they may have leverage there, if they would use it wisely.

    Remember that the twist in the tale here compared to the usual situation is the HSR clause in the SF 1999 Prop H ... we are used to seeing old systems granfathered in, but seeing not yet existing systems grandchilded in is unusual.

    (* frequency note: 8 tph is what is in the Caltrain-financed HSR ridership modelling, but that involves a complex station skipper schedule, and a simpler Express / Limited / All-Stations structured schedule seems like it would require fewer trains.)

    ReplyDelete
  64. Under the TBT design, can an HSR service enter the DTX every five minutes without interfering with the ability of Caltrain services to use their platforms?

    It depends. The DTX tunnel is effectively two-tracked. Three tracks are equivalent to two since running the trains 2/1 into a terminal doesn't increase capacity without a lot of parking space for trains.

    Now, a two-track tunnel at low to medium speed can support 24 tph comfortably, and 30 tph with a lot of headaches. Whether 12 tph (or 18 with headaches) is good enough is something Caltrain has to decide for itself. I'd say no because it serves a far larger market than that, but that's because I live in New York and expect rail systems to have decent modal shares. If Caltrain is willing to live with a 30% market share, 12 tph is fine.

    Working through a variety of platform combinations, it seems that Richard's layout (pdf) would not be strained in meeting (2), much less (1).

    Richard's layout only has a single-track 4th and King bypass. Does that mean that some HSR trains will have to stop at 4th and King?

    ReplyDelete
  65. Alon: Just because there's a platform, doesn't mean trains have to stop there. Consider Secaucus. As for Caltrain's market share, it's not that huge of a market to begin with. According to Caltrans, the peak hour traffic volume on the busiest part of the 101 is about 17,000, which is equivalent to about 12 ten-car trains with all seats filled. Assuming 8 tph is reasonable, as that's the point where service frequency basically stops mattering in terms of convenience, on both locals and expresses, and it becomes better to increase capacity by lengthening trains. Only once they get to 10 cars or so would they have to start worrying about running more trains and, as mentioned above, that's already a very major shift in market share. And Caltrain would have many other things to worry about before they can even reach that level of ridership, such as how all those riders are going to get to the stations.

    ReplyDelete
  66. Alon Levy: "It depends. The DTX tunnel is effectively two-tracked. Three tracks are equivalent to two since running the trains 2/1 into a terminal doesn't increase capacity without a lot of parking space for trains."

    Normally a non-engineer thinking in terms of whether trains can get to where they need to get to in order to meet and build demand for transport would assume the tunnel is the bottleneck.

    But the post makes the point that it is quite a long ways between the platform and the 1:2 switches in the tunnel itself, before the tunnel gives way to the turn into throat. And whatever switching there is to cross-over egress and access to each pair of platform tracks will be further down the tunnel.

    The longer a platform egress blocks a platform access, the greater the likelihood that there is a bottleneck. And as far as I understand, an HSR trainset will not be like a Cityrail V-Set when negotiating a 500ft to 600 ft. curve radius ... it has a longer truck wheelbase and stiffer coupling, for lateral stability at high speeds.

    So if an HSR will not be accelerating up toward the 30mph maximum tunnel speed until it has reached the tunnel itself, the station throat would seem to emerge as a possible bottleneck.

    A maximum length HSR train is 1/4 mile long. If it traverses 3/4 mile between the switch and the platform, that is a mile after it begins moving to clear the switch. That is an appreciable period of time ... if it accelerates smoothly from a dead stop to 20mph as the tail clears the switch, that is an average of 10mph.

    12 HSR tph headways with the HSR all passing over one of two platform switches is 12 HSR train movements per switch in the original design, as each trainset has to pass over the same switch twice. An average of 10mph for the 1 mile between the platform track switch is 6 minutes from platform to switch. Add the time to do the crossover for the departing train to make way for the arriving trains, and that is, obviously, under 10 train movements per hour per switch.

    Now, if the HSR can accelerate smoothly to 30mph as the tail crosses the switch, that is an average of 15mph, 4 minutes per access, if the crossover plus crossing headway, the crossover switching and crossing headway combined would have to be under a minute to hit 5 minute headways. If it could hit and maintain 30mph 2/3 or less of the way to clearing the track, that's be 20mph or better, 3 minutes per access, crossover switching and crossing headway would have to be under two minutes per movement to hit 5 minute headways.

    "Richard's layout only has a single-track 4th and King bypass. Does that mean that some HSR trains will have to stop at 4th and King?"

    I was looking at the DTX/throat/platform movements, I hadn't looked at 4th and King.

    "Whether 12 tph (or 18 with headaches) is good enough is something Caltrain has to decide for itself. I'd say no because it serves a far larger market than that, but that's because I live in New York and expect rail systems to have decent modal shares."

    There is lots of room to expand capacity of the Caltrain sets, especially when they move to EMU's. A three door bi-level, with 4 across seating on top and metro style seating and standing room below would have crush capacity over 1,000 for an eight car set, and a substantial share of end-to-end San Jose to TBT passengers would be taking the HSR for the 1/2 hour transit time.

    ReplyDelete
  67. @ Alon Levy -

    Caltrain is running 5tph peak right now and hopes to go to 10tph by 2025. Since they're looking to triple ridership with just twice the train frequency, they're looking at substantially longer trains as well.

    Of the 10tph, only a subset (basically, as many as possible) will go to the Transbay Terminal, the rest will stick with 4th & King - especially any remaining diesel trains.

    @ BruceMcF -

    Richard's proposal for the throat looks reasonable, though it still calls for a three-track DTX tunnel. I think this whole capacity discussion would become a moot point if the Transbay Terminal were set up as a through station with a single one-way loop track supporting all six platforms. The additional cost is compensated by making do with a 1x2 or 2x1 DTX tunnel, with the loop sections actually down to one track.

    Constructing a narrow tunnel 25' under existing buildings is much easier than constructing a wide one.

    As Richard points out, it's perfectly feasible to keep dwell times at a major through station down to 6-15 minutes. Jim, who works for Amtrak, tells us significant cleaning and housekeeping don't need to happen at both ends of a trip.

    And if you're still fretting about how Caltrain and HSR will co-exist, keep in mind that there will most likely be a single dispatcher for the Caltrain ROW for safety reasons. That means operators will co-ordinate their schedules to minimize any conflicts during rush hour.

    ReplyDelete
  68. "Richard's proposal for the throat looks reasonable, though it still calls for a three-track DTX tunnel."

    No, it doesn't ... the central track switches with the extension of the access track, then switches onto the extension of the egress track, and then there's a crossover between the two, and for most of the length of the DTX its normal dual track. I am not 100% where the bored section is, but I'm guessing that the three wide section is restricted to part of the final cut and cover section of the tunnel.

    "And if you're still fretting about how Caltrain and HSR will co-exist, keep in mind that there will most likely be a single dispatcher for the Caltrain ROW for safety reasons."

    You've got me confused with someone else, I'm talking about operational capacity, when the highest frequency services are relegated to one of the side islands. Richard's layout has enough opportunities for parallel movement that a train can be either arriving or departing the farthest platform track (whichever is the cross-over) without blocking other platforms for five or more minutes at a time.

    I haven't seen what the switch layout is supposed to be in the present tunnel, so I don't "know that it can't". Rather, based on what I have seen, I don't find any reason to put any faith in fact-free re-assuring noises from the TJPA.

    As to what could be done with a proper loop-back, of course, but that's neither here nor there given hundreds of millions of dollars of money on the table for an application in the near term future. If an alternative cannot be argued to fall within the envelope of the existing EIS clearance, its not a live option.

    As far a I understand it, the section of Richard's layout from two tracks turning up 2nd and all the way through to the train box itself gets in as a live option, so if CHSRA wants an option that will give them the best operational flexibility for Stage 1 with room to grow into for Stage 2, they could do worse than demanding that part of Richard's design. Or else, as a "concession" on the number of platform tracks, give an alternative set of design constraints for living with only four HSR platforms ... 5 minute headways, minimum 200m straight platform sections, 1000m platform curve radius, number of parallel movements, 190m minimum curve radius on tracks accessing the platforms, 45mph speed in the tunnel ... that mandates Richard's design.

    Unless the TJPA can demonstrate 5 minute headways through the throat of the present design, the CHSRA would have substantially more leverage along that path than in holding out for a six island two level train box ... that would be the CJSRA relying on a dead center reading of the Prop 1A headways language, rather than a tenuous extension of it, and would be arguing for a less expensive DTX.

    ReplyDelete
  69. As Richard points out, it's perfectly feasible to keep dwell times at a major through station down to 6-15 minutes.

    @Rafael, Richard has pointed out that you could easily achieve these dwell times at a major terminal station.

    I haven't seen what the switch layout is supposed to be in the present tunnel

    @Bruce: the diagram in the post is to scale and was drawn from TJPA engineering drawings dated early 2009.

    ReplyDelete
  70. Clem: "the diagram in the post is to scale and was drawn from TJPA engineering drawings dated early 2009."

    But where are the switches?

    It shows no switches at all except for branches. Indeed, it appears to show the Mission street side island platform sharing a single tunnel track with one platform on the central island, the other platform on the central island sharing a tunnel track with the inside track of the Howard Street side island and the outer tunnel track of the Howard Street side island with a tunnel track of its own.

    And no cross over until the other side of 4th and King.

    So it may be from 2009, but the engineering drawings either left the switching out, or it is three bi-directional tracks, one to one exclusive to Caltrain, one to two shared between Caltrain and HSR, and one to three exclusive to HSR.

    I would wish the TJPA the worst possible luck in trying to demonstrate that such a system could conceivably offer 5 minute headways to three successive HSR services. If that's the actual design, then y'all need to force them to switch designs are they are at risk of getting squeezed out of the HSR money pie altogether.

    ReplyDelete
  71. Clem: "@Rafael, Richard has pointed out that you could easily achieve these dwell times at a major terminal station."

    Yes, if the terminal is not a terminus, platform capacity is not an issue.

    But if its not a terminus, it needs to have 5 minute headways. The only wiggle room would be what a 5 minute headway means for a shared line.

    ReplyDelete
  72. @ Clem, oh, that diagram. Yes, lots more switches on that diagram. Odd shortage of switches across from the TBT side of MT-4 to the 4th and King side of MT-5, though.

    I'm assuming that the 4th and King platform on MT-5 is the platform for inbound Caltrain services and the 4th and King platform on MT-4 is the platform for outbound Caltrain services.

    In the diagram, a Caltrain service departing either T-21 or T-22 will either crossover MT-2 at around +1050m or use the next switch to occupy MT-2 from there to +1720m, to allow an outbound HSR (not stopping at the local platform) to pass. In that case, an inbound HSR express can bypass a Caltrain inbound at the inbound platform on MT-2, then use the next switch to MT-5.

    That is setting up that HSR to occupy MT-5 until the only switch back to the HSR platforms at about +900m. And of course the Caltrain local has to stay on MT-5 until the end.

    I guess that means that MT-5 with a Caltrain inbound is a critical path. After the passing HSR uses the 2/5 switch at +1800m, how many minutes until the Caltrain local at the platform gets its green light and then passes by that switch, how many minutes until a following HSR can use / pass by that switch, and then how many minutes until an HSR passing the next local at the platform.

    On the strictest reading of Prop 1A, the passing HSR and a following HSR can be five minutes apart and the Caltrain local has to fit in between ... on a more flexible reading, if two HSR are through and its clear for a third to pass inside of ten minutes, that would be close enough.

    Now, an HSR outbound passing a Caltrain local outbound so the Caltrain local can use the outbound platform ... I got that. But an HSR outbound five minutes later would want to be on MT-2 all the way to pass. That's going to occupy MT-2 either all the way or from +700m to the 2/4 switch at +2450.

    So that pins down the inbound pattern on MT-5. An inbound HSR uses MT-2 to pass a Caltrain local at the 4th and King platform on MT-5. Then the local pulls out of the platform. Then the following HSR uses the 2/5 switch at 2450 so that MT-2 is open for an outbound HSR passing a Caltrain local on the outbound platform, and chases the Caltrain Local up MT-5.

    And ten minutes or less after the first passing HSR switched onto MT-5, another passing HSR must be clear to switch onto MT-5.

    That's an awfully rigid pattern of operations to swap the single passing track back and forth to be used by alternate inbound and alternate outbound HSR services ... a timing diagram is needed to work out whether there are any conflicts up at the TBT side as the services cycle through the four HSR platforms.

    And that diagram makes it a lot clearer why Caltrain has been pushed to say they only need to bring 6 trains per hour into the TBT ... those MT-5 and MT-4 platforms at 4th and King need to be clear on a regular basis for HSR services running through, if its going to have a prayer of meeting all statutory requirements.

    ReplyDelete
  73. Arcady: my market size comparison is based on the number of people who live in San Mateo and Santa Clara Counties and work in SF. It works out to somewhat more than half the number of people who commute from Long Island to Manhattan, and two thirds the number of people who commute from the areas served by Metro-North to Manhattan. Both of those markets have 80% rail modal share, which is what Caltrain should be aiming for.

    Caltrain is different from the LIRR and MNRR in that it has only one line, so higher-capacity trains are more viable. The LIRR already uses some long bilevel trains, which is how it achieves higher ridership than the MNRR with lower tph count.

    Bruce: I'm not sure how what you said relates to my point that 3 tracks offer the same capacity as 2. To run trains 2/1, you need to have a source of sink of trains. MNRR can run the trains 3/1 in Manhattan because Grand Central has 67 tracks for train parking. Between 8 and 9 am on a weekday, 50 trains enter GCT and 10 leave, but the station has more than enough capacity for that. CAHSR and Caltrain, which will have a grand total of 6 tracks at SFTT, will not have the privilege of asymmetric services.

    Rafael: for the average Caltrain commuter, frequency will equal frequency of trains into SFTT. It's one thing for CAHSR and Caltrain to declare 4th and King their major terminal and encourage CBD development nearby; it's a whole other thing to leave some trains at an inferior location. The lines in New York that do that, for example the Hoboken Division of the New Jersey Transit, are underused. And Hoboken has better rail connections to Midtown and Lower Manhattan than 4th and King does to Market Street.

    ReplyDelete
  74. @Bruce: I just realized the blog software cut down the size of the graphic, making it poorly legible. I just fixed that.

    The official nomenclature is 4th & King for the existing surface terminal, and 4th & Townsend for the new underground station, although I've seen the latter referred to as Mission Bay, the name of the neighborhood.

    The problem you identified is the lack of a fourth track at 4th & Townsend, something Richard has been rightly complaining about.

    Once that is added, the only conceivable reason for a third track in the tunnel (besides enriching construction companies) is redundancy. I don't think it's really needed, since NJT and Amtrak seem to pull off 24 tph in a two-track tunnel under the Hudson without breaking much of a sweat.

    ReplyDelete
  75. Alon Levy said... "Bruce: I'm not sure how what you said relates to my point that 3 tracks offer the same capacity as 2."

    If that throughput on two way track is not the binding constraint ... and the TJPA design creates a local/express interference problem, where throughput of trains in motion is NOT the binding constraint ... then using additional infrastructure to increase raw throughput would be a waste.

    I am certainly unsure whether I understand how the DTX as the TJPA has designed it would be used to meet statutory requirements, but if my guess is right, it seems like a grotesque waste of public resources to have a third track all the way through the tunnel for that.

    Its like its designed for a private freight rail operation on the surface in a four-wide right of way to balance property taxes on a third track against property taxes per switch.

    ReplyDelete
  76. @ Clem, yes, take Richard's design of the fan-out to the TBT (and the platforms), then simple local siding platforms 4th and Townsend and through running for the HSR services would eliminate the need for the third track and save quite a lot of money in the DTX.

    ReplyDelete
  77. since NJT and Amtrak seem to pull off 24 tph in a two-track tunnel under the Hudson without breaking much of a sweat

    They sweat it everyday and if there is a problem trains are delayed for a very long time afterwards. Not that Transbay is ever going to have 24tph from the Peninsula

    ReplyDelete
  78. Clem said ... "since NJT and Amtrak seem to pull off 24 tph in a two-track tunnel under the Hudson without breaking much of a sweat"

    Adirondacker said...
    "They sweat it everyday and if there is a problem trains are delayed for a very long time afterwards. Not that Transbay is ever going to have 24tph from the Peninsula"

    Yes, running a line for a long period at its maximum capacity makes the logistics brittle ... but if the system permits 2.5 minute headways, and every second slot is open to being an HSR, without crossing bottlenecks, that would certainly meet the statutory CA-2008-Prop1A and SF-1999-PropH requirements under even the strictest interpretation.

    OTOH, does the Hudson Tunnel have those tight "underground streetcar" curves built into it? What is its maximum speed ... 30mph as the TJPA claims for its design, or 45mph as in Richard's design?

    ReplyDelete
  79. The speed limit in the Hudson and East River tunnels is 60, but the speed through the station throat at NY Penn is 15 or even 10 mph. At least the track layout is fairly flexible. By the way, one aspect that I'm not sure Richard considered in his design is speed restrictions due to the buffer stops at the end of the tunnel. Depending on details of the signal system, it might mean a speed limit of 30 over the final curve into the terminal. A design speed of 45 makes sense for the rest of the tunnel though.

    ReplyDelete
  80. Bruce, none of the underwater tunnels in New York has underground-streetcar curves. The really sharp curves are only in very old cut and cover sections that need to turn from one street to another. The underwater sections are straight, and tend to have the highest speed limits in the subway system.

    And Adirondacker is right - 24 tph is crush capacity for mainline track, and commuter rail agencies try not to get there. The 2-track LIRR Main Line, which tops at 26 tph at Mineola, has to run trains one-way during rush hour.

    ReplyDelete
  81. What a HORRIBLE design.

    It's looking like the "New Transbay Tube" would be better value for money!

    ReplyDelete
  82. Alon Levy said... "none of the underwater tunnels in New York has underground-streetcar curves. ... The underwater sections are straight, and tend to have the highest speed limits in the subway system."

    Thanks for that ... that was my impression, but of course I know far more about the NSW train system than trains on the East Coast.

    "And Adirondacker is right - 24 tph is crush capacity for mainline track, and commuter rail agencies try not to get there. The 2-track LIRR Main Line, which tops at 26 tph at Mineola, has to run trains one-way during rush hour."

    For the purposes of leverage given by statutory requirement, 2.5 minute headways is clear even for the stricter reading of 2008-Prop1A. Having used Sydney's Town Hall traveling against the evening rush hour during one semester, I am no big fan of running rail lines with no available spare slots during peak ... its too easy for a single delay to cascade if there's no catch up available until the peak has passed.

    ReplyDelete
  83. arcady said... "By the way, one aspect that I'm not sure Richard considered in his design is speed restrictions due to the buffer stops at the end of the tunnel. Depending on details of the signal system, it might mean a speed limit of 30 over the final curve into the terminal. A design speed of 45 makes sense for the rest of the tunnel though."

    Would the slow down for such a buffer be on the TBT side of the 2:3 fan-out or on the tunnel side?

    If the trains can hit the fan-out at 45mph, that seems like it should make a difference for headway at the 2:2 cross-over before the 2:3 fan-out, and the 2:2 cross-over must be used to either access or egress the extreme platform tracks for both side islands.

    Indeed, leaving those two critical paths on a single track and delaying the cross-over as long as possible would seem to make the most sense if that allows bringing the crossing trains through that switch at a higher minimum speed.

    At that inbound point at that TBT side 2:2 crossover, 2.5 minute headways, if it includes the crossing headway, would seem to permit:

    :00:00 HSR access
    :02:30 crossing Caltrain egress
    :05:00 HSR access
    :07:30 Caltrain access
    :10:00 HSR access
    :12:30 open
    :15:00 HSR access
    :17:30 Caltrain access

    If that's possible, that would satisfy the Prop 1A HSR headway requirement at that point, which is grandchilded into San Francisco by the 1999 local Prop H ... and which the DTX would want to satisfy in any event if it eventually wants to use Federal HSR funds for part of its funding.

    Now, even including the commuter overlay to Sacramento, possible specials, and even an possible Mojave spur to Las Vegas, there won't actually be 12 HSR trains per hour, so meeting the headway requirement with four open slots per hour would mean in practice capacity for a peak of 8 HSR trains of various sorts and up to 12 Caltrain and other Bay regional commuter trains.

    ReplyDelete
  84. The slowdown would most likely be at the start of the final fanout, where the tracks turn from Second St to parallel Mission. Among other things, a lower speed lets you get away with no superelevation and thus reduced constraints on where you can put turnouts. The 2:3 fanout would still be at 45 mph. Indeed, if you look at the current Caltrain layout, the 2:3 fanout is taken at 40 mph, with the final turn to parallel Townsend being 25 mph, and the station throat itself being 10 mph. In the case of the TBT, I'd expect a drop to 15 halfway down the platforms.

    ReplyDelete
  85. arcady said... "The slowdown would most likely be at the start of the final fanout, where the tracks turn from Second St to parallel Mission. ... In the case of the TBT, I'd expect a drop to 15 halfway down the platforms."

    Ah, OK, even for a full length HSR, when the head is halfway down the platform, in Richard's design the tail has reached the four track wide section. No matter what, there's an unavoidable crossover for one in four Caltrain movements (egress from one extreme platform track), and one in eight HSR movements (access to one extreme platform track). And of course, if need be, what I set down as a (statutory) target could turn into:

    :00:00 HSR access
    :02:30 crossing Caltrain egress
    :05:00 open
    :07:30 HSR access
    :10:00 Caltrain access
    :12:30 HSR access
    :15:00 HSR access
    :17:30 Caltrain access

    ... as long as the two successive HSR accesses are headed to the correct platforms.

    Certainly upgrading Caltrain to high platforms is preferable, for better flexibility, but even under the current institutional constraints, it seems plausible to me that Richard's design could have a technical argument for meeting statutory HSR requirements even if CHSRA is putting up an opposing argument.

    Of course, it takes an organization with a stakeholder interest to get an engineering report to confirm that as well as confirming the difficulties that the existing TBT design would have in meeting the statutory requirement of 5 minute HSR headways.

    And in terms of the political argument, "does better for less money" is an awfully easy frame to sell for a publicly funded portion of a project.

    ReplyDelete
  86. A high platform Caltrain could mean potential problems at San Jose and Santa Clara, unless ACE and the Capitol Corridor get compatible trains as well, or else have dedicated platforms just for their own services there. I suppose an acceptable compromise would be having three platforms for the ACE, Capitol Corridor, and Coast Starlight, and leaving the remaining six platforms for Caltrain, with HSR on a new upper level. It might even be possible to do without the separate HSR level if all HSR trains stop at San Jose (or run past the platforms at 40 mph), and some Caltrains continue to Tamien, which they should do anyway, for the light rail connection. As for the Coast Daylight train from LA to SF, it can just use high-platform equipment, as 798/799 already does.

    ReplyDelete
  87. A high platform Caltrain could mean potential problems at San Jose and Santa Clara, unless ACE and the Capitol Corridor get compatible trains as well, or else have dedicated platforms just for their own services there.

    The steam trains can be easily accommodated at one platform. They operate, what, one an HOUR or something?

    Isn't there a theme park or Civil War re-enactment site or something like that where they could stop and hang around instead at their Own Special Platforms for as long as is historically accurate?

    I suppose an acceptable compromise would be having three platforms for the ACE, Capitol Corridor, and Coast Starlight, and leaving the remaining six platforms for Caltrain, with HSR on a new upper level.

    Rod Diridon Edifice Complex! Greated Railroad Station In the World!

    Why on earth would a handful of steam-era trains need 80 plus feet of valuable station right of way width between them? Who's supposed to pay for this nuttiness? And what possible purpose could a cretaceous-era nonsense like the "Coast Starlight" possibly serve, other than to induce copious mouth foam among the hopelessly nostalgic and the generally hopeless?

    It might even be possible to do without the separate HSR level if all HSR trains stop at San Jose

    "Might"? But without at least three levels of trains Rod Diridon Memorial Station will not be the Most Intermodalist Importantist NOT-San-Franciscoist Railroadiest Grand Centraliest Station in the Universe?

    More platforms = more levels = more important. End of story.

    (or run past the platforms at 40 mph), and some Caltrains continue to Tamien, which they should do anyway, for the light rail connection.

    Have you ever seen how many people use this "intermodal" connection? Caltrain is just wasting trains and crews pointlessly running them empty back and forth to Tamien. (And for the most part, for that matter, it is wasting trains and crews running them nearly empty south of Mountain View, to a good approximation.)

    Caltrain should run into the platform in SJ, reverse, and get the hell out of Dodge, just like it should in SF. Why are we supposed to pay to park trains doing nothing, or pay crews to run empty trains around for no purpose? It's not like "making the train go backward" has been any sort of big deal since steam locomotives roamed the primeval forests.

    As for the Coast Daylight train from LA to SF ...

    Good Lord! Words fail!

    How about building a new level at Transbay to make space for The Lark? And the Presidential Train? And for the delivering the beet harvest?

    Up above people were talking about real train operators reversing long distance trains at terminal stations and sending them back out in five minutes or less, and now we're back to parking steam trains at through platforms in San Jose for hours at a time.

    We're doomed.

    ReplyDelete
  88. arcady said... "A high platform Caltrain could mean potential problems at San Jose and Santa Clara, unless ACE and the Capitol Corridor get compatible trains as well, or else have dedicated platforms just for their own services there."

    Does depend on what time frame we are talking about. There is an option value in having all platforms in the TBT train-box with a footprint that permits a future upgrade of Caltrain to high platforms and the improved flexibility of six operationally equivalent platforms at the TBT ... even if that option is not exercised when the train-box is originally populated.

    And there are German tram-trains that have a mechanism that slides a high-floor level boarding ramp out of the way to reveal steps to streetcar platform level ... equipping a relatively smaller number of regional and Amtrak passenger cars to cope with multiple platform heights could be a lot cheaper than building a new auxiliary train-box to cope with artificial capacity constraints imposed by operational rigidity at the TBT.

    ReplyDelete
  89. BruceMcF: the fundamental problem isn't that high-floor trains can't deal with low platforms, because in general they can. The only exceptions are the LIRR and MNR MU fleets, the LIRR double deckers, and the Acela. The rest of the eastern commuter fleets and all of Amtrak's Amfleets all have trapdoors that open to reveal step, which can be used to access low platforms. The real problem is that some double decker trains have doors only on the lower level, where the floor is at a height of some 17 inches above top of rail. That means you'd have to have steps UP to the high platform, which is pretty much impossible. It might be possible to modify the Bombardier cars with doors on the middle level instead of the bottom, but Amtrak's Superliners and California Cars can't be used at all with high platforms.

    And Mr. Anonymous: you may laugh about the "steam trains", but ACE has more modern equipment than most of Caltrain, and while it only runs four round trips, it manages to get an average weekday ridership of 3200, and anecdotally, the trains are packed. The Capitol Corridor also has decent ridership on the southern segment despite relatively poor service. But all together, it adds up to maybe 2 tph at peak, which isn't all that much.

    And just to remind you, Mr. Anonymous who likes to mock the existing train system: it actually exists and is carrying passengers today. The Coast Daylight is already halfway there, with a roundtrip from LA to SLO and a bus connection onward, and you can buy a ticket for it today by going to www.amtrak.com. The HSR does not exist, and, even assuming everything goes well, will not exist for another decade or so. And if things don't go well... who knows.

    ReplyDelete
  90. arcady said... "the fundamental problem isn't that high-floor trains can't deal with low platforms, because in general they can. ... The real problem is that some double decker trains have doors only on the lower level, where the floor is at a height of some 17 inches above top of rail."

    I don't see the fundamental problem ... don't allocate those trains to the Caltrain corridor, problem solved. As (*hopefully) oil prices repeatedly spike, there will be plenty of opportunities for allocations of new rolling stock to suit while reallocating existing rolling stock where its not an issue.

    (*hopefully: well, its possible that the world can stay in an extended global depression so that the oil price merely creeps up as more and more cheap oil is exhausted, but that is a worse scenario than repeated oil price spikes)

    ReplyDelete
  91. future upgrade of Caltrain to high platforms

    Almost all the Caltrain platforms will be upgraded by the time the first HSR rolls into Transbay. The current side platform ones are where the local track will be. I haven't looked to see if there are any island platforms at suburban stations. If there are any, they are where the express tracks will be.

    Amtrak's Superliners and California Cars can't be used at all with high platforms.

    How many Superliners or California cars go to San Francisco?

    Anon, they can't park steam trains at San Jose for hours. They have to leave fast so they can get to a coaling station and someplace with water. But then adding coal and water facilities would make San Jose much more attractive to all the steam trains operating in the Bay Area.

    ReplyDelete
  92. Amtrak's Superliners and California Cars can't be used at all with high platforms.

    And even if they could, the FRA still wouldn't allow compliant Superliners to share track with non-compliant HSR/EMU trainsets.

    ReplyDelete
  93. @bikerider,

    I don't see why not, if the "compliant" equipment is equipped with proper 21st century train control systems.

    I often see this knee-jerk reaction that the FRA can't possibly allow the mixing of "compliant" and "non-compliant" trains. You have to take a step back and consider the whole issue: "non-compliant" trains rely mostly on active safety measures (e.g. automatic train control) to reduce the probability of a collision. "Compliant" trains rely mostly on passive safety measures (e.g. collision posts) to reduce the consequence of a collision. Both approaches are valid ways to increase safety.

    "Compliant" trains could easily share tracks with "non-compliant" trains if they were suitably equipped with modern safety systems. Without those systems, I can see why the FRA doesn't want them mixed-- you have steam-era conductors acknowledging signal aspects by voice (between punching tickets), with a train load of passengers just one engineer's mistake away from disaster. (Chatsworth anyone?)

    With humans out of the safety loop, and with a 100% grade-separated right of way, I think a good regulatory and safety case can be made for allowing mixed operations.

    ReplyDelete
  94. Why would anybody want to permit such dinosaurs on a (modern) Caltrain line?

    Low acceleration, high weight, poor reliability, abysmal time-keeping: these sound like the tenants from hell.

    Fundamentally, any dino-train that operates on the Caltrain line is one that eats significant amounts of capacity and incurs significant capital costs for civil structure and signalling safety compatibility.

    While it may be technically possible to do so, why on earth would anyway want to?

    If the train doesn't have ETCS signalling, if it has an axle load over 17 tonnes, if it can't accelerate and brake at 0.6m/s^2 or better, if its MDBF is less than 100000km, or if its punctuality of entering the Caltrain-controlled ROW within a 2 minute schedule window is less than 95%, then it doesn't meet the requirements and isn't worth pandering to.

    If your infrastructure is optimized for specific timetable operated by trains with specific minimum performance, why would you want to spend hundreds of millions extra to provide extra track to allow one or two long-obsolete dinos to lumber alongside the clockwork trains carrying real passengers?

    Who needs to borrow trouble and great expense for no measurable gain? There are railway museums where you can go to see junk -- uh, make that "valuable historic treasures" -- like the Coast Starlight or SP steamers or "commuter trains" that make one round trip a day.

    Otherwise, get off at Didiron "not yet dead" Memorial Pangalactic Intermodal, jump on Caltrain (one leaves every 15 minutes or so), brush the soot and coal dust from your hair, and enter the 20th (dare we hope for the 21st?) century.

    ReplyDelete
  95. @Richard, the purported impossibility of running "compliant" trains mixed with "non-compliant" trains is routinely brought up in various contexts by detractors of HSR and a Caltrain "non-compliant" solution, to show how this or that operational scenario simply isn't feasible.

    I think your broad-brush dismissal of all existing equipment as "steam trains" may be over the top. If the goal is operational flexibility, why not allow for some mixed operations?

    Taking a hard line such as yours can backfire: would you really want Caltrain to end up with ALP-46's because you let somebody win the inane argument that you cannot ever mix "compliant" and "non-compliant" traffic on the same tracks?

    (The terminology itself is loaded: "non-compliant" wrongly suggests unsafe... which is why I try to use quotes.)

    ReplyDelete
  96. Clem said...
    "I often see this knee-jerk reaction that the FRA can't possibly allow the mixing of "compliant" and "non-compliant" trains."

    The problem is the idea of a single blanket regulatory scheme, which in practice is a heavy freight rail regulatory scheme, and then any operation which is not compatible with the scheme requires case by case waivers/exemption etc. This is classical management by exception, and a symptom of something wrong in a management system.

    In this case, the problem is the idea of a single regulatory scheme. Given a Heavy Rail regulatory scheme and a Rapid Rail regulatory scheme, then instead of the on/off categories of "compliant" and "non-compliant", there would be:
    (1) Compliant with Heavy Rail
    (2) Compliant with Rapid Rail
    (3) Compliant with Both
    (4) Non-compliant

    Under that system, most of the "problems with mixing stock" go away. Rapid Rail stock would run on Rapid Rail systems, Heavy Rail stock on Heavy Rail systems, dual-compliance stock would operate on each system as a full fledged member of that system, and only the residual non-compliant stock would require special case treatment.

    The dual compliance stock would likely be less energy efficient than the stock that is Rapid Rail compliant but not Heavy Rail compliant, but then that means that it would be focused on routes where there was a strong case for running under both regulatory systems.

    ReplyDelete
  97. Hey, there's nothing wrong with ALP-46s. They're a good locomotive, and have been working so well on NJT that they want to order 50 more, and I hear Amtrak is interested too, and the new 46A model will have 7500 hp instead of 7100. And of course unlike Caltrain's imaginary multiple units, they've made thousands of trips and moved millions of passengers.

    That said, as far as I understand it, Caltrain's stated goal has been to allow running of non-compliant trains by special waiver for grade crossings, and with enforced signal separation from compliant passenger trains and time separation from freights. And I do agree with their assessment that MU trains are the way to go, at least for local trains, given that the line has so many stops (which is a GOOD thing, it means more convenient access for more passengers).

    I'd also just like to say that I'm a bit disappointed at the hostility to existing train services. Sure, there's a lot not to like about FRA regulations and crappy American freight-style diesel locomotives. But the Capitol Corridor and ACE and the rest of California's mainline rail services have been doing a pretty good job of attracting ridership, and they provide a useful service to a large and growing number of people, TODAY. HSR, and for that matter the shiny new 2025 version of Caltrain haven't helped anyone at all yet, for the simple reason that they don't exist, not even in blueprint form. It would be incredibly foolish to sacrifice today's success for a dream that may yet fail.

    ReplyDelete
  98. BruceMcF: I really believe that we need to stick to a single unified rail network as much as possible. There's just too many benefits from doing so. I think in my ideal world, the FRA would create a second tier of crashworthiness standards, probably just imported from Europe, which applies when there is an adequate level of signal protection that can provide train separation, at a minimum equivalent to PRR's ATC or German Indusi, as well as more emphasis on things like flank protection.

    ReplyDelete
  99. the purported impossibility of running "compliant" trains mixed with "non-compliant" trains is routinely brought up in various contexts by detractors of HSR and a Caltrain "non-compliant" solution, to show how this or that operational scenario simply isn't feasible.

    Most people who bring this sort of thing up are simply foamers (Amtrak fans, etc) and are of no interest.

    The issue with non-compliant trains (and by this I mean trains that aren't compliant with a base level of performance and reliability) is that there are massive costs involved in pandering to them.

    We go round and round in circles on this, but given that infrastructure, timetabling (operations), and rolling stock are intimately inter-related and together determine the type and quality of service that can be delivered, issuing an edict that non-performing rolling stock must be pandered to can only lead to an increase in infrastructure requirements (= $$$$$$$) and/or a degradation in system capacity and throughput.

    We're already in the state where people are demanding that we retain an entirely separate, eight-track, surface-level station at Fourth and Townsend in addition to wasting four billion on DTX because we "need" to be able to run three diesel trains per direction per day across a (single track, low-level, why bother?) Dumbarton Bridge and into SF. Oh, and do it at peak hours, just to really screw things up.

    This is simply insane! The benefits are orders of magnitude removed from the costs.

    The same goes for some worthless Amtrak junk trundling into San Jose minutes or hours or days late. What possible value can there be in screwing up the entire Caltrain line (infrastructure and operations) in order to have such a thing clang-clang its way up the line?

    It is only going to consume multiple timetable slots (because of low performance) even if it starts off on time, which it won't, and it's going to force the avoidable problem of accommodating multi-hundred-ton (= $$$$) freight-type locomotives and trains, and it's going to screw up the choice of a train/platform system corridor-wide, and it's going to require a dedicated turn-back (= $$$$$$$$$$) terminal since it sure as hell isn't going to TTT, etc, etc, etc, etc.

    Why borrow trouble? The only ones who come out ahead doing this are gold-plating construction contractors and railroad nostalgia fans.

    We're already talking about pretty much saturating some very, very expensive infrastructure with circa 12tph (8 Caltrain, 4 HSR): why drive the costs through the roof to add steam trains into the mix?

    If the train isn't compliant with a base level of performance (acceleration, weight, timekeeping, platform interface, etc) there is simply no possible public benefit in letting it screw up the rest of what will be an intensive operation.

    Other things being equal, segregation of traffic types (HSR from Caltrain at Redwood Junction; HSR from Caltrain in Santa Clara; stream trains from Caltrain+HSR in San Jose) is always the cheapest way to provide capacity and reliability. We mix traffic types as an engineering and economic compromise (either because excess capacity is available for free, or because constructing parallel facilities is unjustifiably expensive), but if the compromise is avoidable and unnecessary it shouldn't be made.

    "Mixed traffic" isn't a goal in itself, it's a means to an end.

    ReplyDelete
  100. An anonymous note to everyone: let's call things by their proper names, shall we? Sure the FRA regulations are antiquated and most of the trains running under them are underperforming. But they're not "steam trains", and they're not that level of outdatedness either. And name-calling is one of the signs of a troll, and makes it hard to take your argument seriously.

    ReplyDelete
  101. arcady: "I really believe that we need to stick to a single unified rail network as much as possible."

    A single unified rail network "as much as possible" translates into multiple, and surely heading to hundreds, of different rail networks under a variety of waivers, track segregations, time segregations, and etc.

    As Richard, interspersed with over the top rhetoric, points out that: "The issue with non-compliant trains (and by this I mean trains that aren't compliant with a base level of performance and reliability) is that there are massive costs involved in pandering to them."

    That's the point. Establish a regulatory regime for the needs of a Rapid Rail system, just as the present FRA regulatory regime is established for the needs of a bulk heavy freight rail system.

    Then if a train can meet the requirements of the Rapid Rail network, whether it can also meet the requirements for running on the bulk heavy rail system is neither here nor there. On track under the Rapid Rail regulatory regime, whether rolling stock also meets

    A very large number of distinct regulatory regimes clearly eliminates substantial opportunities for network economies, but there are trade-offs between the optimum design for a specific transport task and the network economies from operating on a common network, and it does not seem to be the case that the best balance on that trade-off is a single one-size-fits-all regulatory scheme.

    Hauling granite on a "whenever you can get it here at that price" and carrying passengers on schedule at high speed, reliability, and often frequency are distinctive enough transport tasks that a single regulatory regime to try to meet all of those needs will involve massive compromise on the requirements of one side or the other or both ... it will either serve one of the two badly, or will serve both badly.

    Which of course is why we have the proliferation of distinct regulatory schemes in the guise of waivers, time and track separated systems, non-standard-gage Mass Transit systems and etcetera and etcetera.

    ReplyDelete
  102. Richard: "The same goes for some worthless Amtrak junk trundling into San Jose minutes or hours or days late. What possible value can there be in screwing up the entire Caltrain line (infrastructure and operations) in order to have such a thing clang-clang its way up the line?"

    The metric for the value of the "worthless Amtrak junk" would be patronage growth. If a passenger train puts bottom on seats, the fact that it's not more exciting toys for the boys is neither here nor there.

    The problem of consuming multiple slots during peak is straightforward to resolve ... curfew any set that needs a double slot from the core of the morning and evening commuter peak. Any regional rail service with a substantial commuter patronage that can meet its on-peak demand will have the off peak slots available to cater to Rapid Rail inter-regional services.

    If a system as badly run as Cityrail in Sydney can work that out for the regional and intercity smokies running into Sydney, surely the Bay can do that.

    The problem of an expensive surface terminal would also seem to be easy to fix ... find where its going to leave the Caltrain corridor to turn back, the last through station on the Caltrain corridor before that is its northern passenger terminus/origin. Expensive surface terminal station eliminated.

    The problem of the expense of catering to heavy rail is already there for heavy freight rail on the Caltrain corridor. The freight rail operator with rights to use the corridor is not a public service organization or a charity ... given the option value of those rights in the event of continued diesel price shocks, there is no reason to expect them to abandon those rights without a freight bypass of equal or better capacity.

    Given that a portion of the Caltrain corridor track will have to continue to be able to operate as a freight line, the idea that a Talgo should be absolutely forbidden from running into the Caltrain Corridor is just hyperventilation.

    ReplyDelete
  103. BruceMcF: right now we have exactly zero miles of Rapid Rail, and the world's biggest network of "bulk heavy freight rail". We have 29 million passengers a year on Amtrak, and service that actually covers a significant part of the country. It would be stupid to try to build a whole new network from scratch: you have to use what you have. And I really do think that a single unified network is the way to go. I think the way to get there is to have the FRA acknowledge that preventing trains from crashing is just as good as having excessively heavy trains. They're already moving in that direction with Crash Energy Management and the PTC mandate. I think the ultimate solution is to either bring the infrastructure under public ownership or limit the liability of infrastructure owners. Because ultimately, the reason FRA regulations stick around is because the freights are very afraid that a train full of chlorine will derail into a train full of passengers, and the resulting lawsuits will bankrupt them.

    ReplyDelete
  104. They're [FRA] already moving in that direction with Crash Energy Management and the PTC mandate.

    Highly, highly doubtful FRA is really going to move in that direction (though I'd be happy to be proven wrong...)

    Best case scenario is that FRA permits lightweight rains with PTC, but only if every single other train on the ROW is also on PTC. This does not help very much to modernize a service like the ACE -- because it depends on UP also using PTC exclusively on their trains (which they absolutely will not do for cost and operational reasons).

    So, at best, we are left with extremely heavy, slow, unreliable, and infrequent "Amtrak" type trains which might have operational flexibility, but are otherwise totally useless as providing a real transit alternative to the car.

    And since there is no actual Amtrak-type service into SF anyway (well, except for Caltrain itself), what is really being lost by not putting in backwards compatibility?

    ReplyDelete
  105. "right now we have exactly zero miles of Rapid Rail, and the world's biggest network of "bulk heavy freight rail"."

    And by far the majority of long-haul containers go by road, at a penalty of over 10 times the energy consumption compared to electric freight rail.

    Not only does the one size fits all mentality present a bureaucratic roadblock to establishment of new passenger rail lines, forcing each new line to develop their own solution then beg for a waiver ... but its also one of the barriers standing between us and saving 10% of our petroleum imports.

    ReplyDelete
  106. bikerider: the PTC mandate says all lines with passenger service must have it by the deadline (2012 I think). And given that there is a PTC mandate, you have to figure that maybe the freights aren't entirely opposed to the idea. I personally think that they do see some advantages in PTC, perhaps in terms of absolving them of some liability, or maybe they really do want to reduce accidents because they've been getting really expensive for them. Another possibility is that they think it's a step on the way to having just one person in the cab.

    ReplyDelete
  107. the PTC mandate says all lines with passenger service must have it by the deadline (2012 I think).

    So thanks to the 2012 PTC mandate, Metrolink can go out and buy some non-compliant Stadler DMU trainsets with level-platform boarding?

    I think not.

    ReplyDelete
  108. Quite a discussion - lots of good points. Whether they get listened to, and acted on...we can hope.

    As we sell of lots for new skyscrapers along the path of the old Key System loop (which is tight radius but had the one-way station loop down to a science), it pains me to see palm trees on the roof rather than...rail platforms? They'd need to be made longer than now, and it wouldn't be possible to access them via the old viaduct (those were streetcar turns, but the radius can't be too much more than the turn from 2nd St). To save some of the old historic buildings on 2nd Street, maybe tear down some of the Macris era squat skyscrapers-with-hats that litter the end of Mission Street past Main.

    ReplyDelete
  109. So thanks to the 2012 PTC mandate, Metrolink can go out and buy some non-compliant Stadler DMU trainsets with level-platform boarding?
    Thanks to the PTC mandate, Metrolink will at least be able to argue that now there is minimal risk of a train collision, so maybe collision strength requirements could be loosened. As for level boarding, that's a CPUC problem, because they prohibit platforms higher than 8 inches. NJT does in fact run Stadler DMUs, with level boarding, on a line that is shared with freight (by time separation).

    ReplyDelete
  110. bikerider: "So thanks to the 2012 PTC mandate, Metrolink can go out and buy some non-compliant Stadler DMU trainsets with level-platform boarding?"

    arcady: "Thanks to the PTC mandate, Metrolink will at least be able to argue that now there is minimal risk of a train collision, so maybe collision strength requirements could be loosened. As for level boarding, that's a CPUC problem, because they prohibit platforms higher than 8 inches. NJT does in fact run Stadler DMUs, with level boarding, on a line that is shared with freight (by time separation)."

    But improvement on the status quo is not the same as getting where we would like to go.

    This still leaves Metrolink begging for permission, instead of being able to simply verify compliance with a standard, where compliance implies permission to use as laid out in the standard.

    And it still leaves a patchwork quilt of waivers and special exemptions, with no guarantee that what was permitted for one will be permitted for another.

    Finally, note that "shared by freight with time separation" is the "different networks" that having a single standard is supposed to prevent.

    After all, if you count two trains complying with two different regulatory schemes and running on the same track at two different times of day under as "one network", then a distinct Rapid Rail regulatory system, developed by adopting appropriate international standards and putting our own index numbers on them, is not going to be a "different network" at all, since all it needs is a declared time of day separation to also be part of the heavy freight network.

    Rapid Rail, passenger or freight, will run to schedule, so that is quite often the preferable way to pool infrastructure costs with heavy freight ... hand over use of a path during a time of day that it is not required for Rapid Rail. Where that is not workable precisely when either the heavy freight or the Rapid Rail is using the line at a frequency where it makes sense to put in additional track.

    Now, if lightly used single track, is being upgrade for low frequency to be used at low frequency by Rapid Rail, but time of day is not appropriate, adding 10:50 passing sidings and running dual compliance Rapid Rail on that line as a heavy rail train, which can then run into a connecting path during the Rapid Rail time of day ... for instances, an upgraded ACE to the upgraded Caltrain corridor ... that's a perfectly reasonable middle ground.

    ReplyDelete
  111. BruceMcF: I'd much rather have Metrolink beg for permission for an actual specific train and an actual specific operation that they specifically need. Right now, there is no Rapid Rail, and if the FRA were to start setting standards, it would be based on zero experience, and you'd get something like the Acela, where the standard was written first, and the train was made to try and fit the standard. I'd much rather get more experience with various non-traditional rolling stock first to see what would be reasonable to codify. You can draw an analogy with the Internet: the IETF standards require at least two interoperable implementations before they can even be considered for standardization. On the other hand, the ISO wrote a big pile of standards for networking, and once people tried to implement them, it turned out to be a disaster, and nobody uses them anymore.

    I also think that the whole Rapid Rail idea is wrong. We don't need regulator-imposed separation. Sure, it would be a stupid idea to run a gravel train on the Caltrain line in the middle of rush hour, but that's a matter for timetable writers, not regulators. And in the end, I think this urge to separate for safety is misguided, because no matter what you do, you're never going to get absolute safety, you're only going to mitigate risk to one extent or another, so it would be good to start with that idea as a basis.

    ReplyDelete
  112. So, the 'Bombardier height' (640 mm) seems to be pretty nearly standard on the West Coast for commuter lines. The Superliners are about 17", but that design has serious accessibility issues, so it's not really worth thinking about.

    The Talgo cars for the Cascades service offer *level walk-through*, just like "high-level" cars, with a lower floor height. I can't find a single reference which contains the actual Talgo floor height, however. The earliest Talgo designs had 300mm floor heights, but I doubt they stuck with that.

    Europe is "standardizing" on a pair of heights: 550mm and 760mm.

    I would probably go with the lowest height which can support "level walkthrough" in a Talgo design. Those would also support standard bilevel designs without difficulty.

    The two proposed heights for CAHSR and Caltrain are so close that it's exceptionally, shockingly stupid to make them different. The "Bombardier car" height would be fine for a level-boarding, single-level Talgo-style HSR, or a level-boarding, bilevel TGV style HSR.

    ReplyDelete
  113. Hmm. Platform heights are still ticking through my brain.

    How About....
    (1) A new standard for passenger trains outside the "high platform zone" in the Northeast and Mid-Atlantic. A standard which is *WIDER*. This is more comfortable anyway and eliminates *ALL* those freight company complaints about over-wide loads, brakemen, etc. The standard would be designed not to hit the 8" platforms.

    (If other trackside constraints prevent the passenger trains from being wider, then the freight companies have no justification complaining about platforms, because the very same other constraints would cause the same problems for the freight trains as the platforms would. In this case, just order the freight companies to shut up, and install standard Eastern-style high platforms.)

    I suppose passenger-only trackage might impose such constraints, but I'm pretty sure that (for instance) Chicago Union has room to narrow some platforms and widen space for tracks without creating great inconvenience.

    (2) Set this platform height to the lowest height which allows for walk-through trains on Talgo bogies on very high speed trains running around curves and up and down grades. Lower heights make cheaper platforms and less alteration.

    (3) Phase in the implementation of this standard.

    ReplyDelete
  114. "the Cityrail system in Sydney can reach 3 minute headways, and they always whinge about it being crappy old signaling, switching and etc. that needs to be upgraded"

    Cityrail in Sydney has the advantage of an *AMAZING* set of flying junctions just south of Central Station. The amount of flexibility afforded by this is immense, and it's just topped off by the City Loop.

    The flying junctions are arranged like a loom. If all northbound traffic is coming in on the east tracks and all southbound traffic is leaving on the west tracks, any northbound can platform anywhere in the station without interfering with any southbound leaving the station.

    It's an absolutely amazing design, meaning that most trains don't take conflicting paths until far away from the Central Business District.

    The addition of the City Loop means that remaining trains going the same direction which might otherwise conflict (switching from 'fast track' to 'slow track' and vice versa) often sort themselves out by going around the loop.

    The "critical path" in Sydney Cityrail is in outlying junctions! I would not expect normal station throats, lacking a giant "loom" of flying junctions, to achieve the same headways.

    ReplyDelete
  115. neroden: the problem with high platforms isn't a freight railroad problem, believe it or not. It's a state regulation problem, in that the California Public Utilities Commission mandates that nothing be within 7.5 feet of the track centerline and higher than 8 inches. That's why they have the silly mini-not-very-high wheelchair platforms, and why the tracks swing out at new stations to accommodate the inter-track fence. As for having wider passenger trains, I believe Russia does this, but I'm not sure it's such a great idea, given that absolutely everything in the mainline network is built on the assumption of 10'8" wide, 85 foot long cars being the maximum.

    ReplyDelete
  116. arcady: "Right now, there is no Rapid Rail, and if the FRA were to start setting standards, it would be based on zero experience, ..."

    The existing international standards incorporate substantial experience.

    "I also think that the whole Rapid Rail idea is wrong. We don't need regulator-imposed separation."

    Yes, but you are the one advocating the status quo, and the status quo is regulator imposed separation.

    Adherence to the FRA heavy freight rail standard undermines the development of common passenger rolling stock for the US market, because its a crappy standard for passenger rolling stock, and those who think they can succeed try to get out from under it ... and getting out from under it is a shot in the dark, which is no basis for long term investment in plant and equipment.

    Instead of the current system which is based by default on separation of FRA heavy rail compliant from "anything else", set the target to hit, just as is already done for FRA heavy rail compliance, and if you hit the target, you are approved.

    The current regulatory regime is one of the things standing between us and a Rapid Freight Rail network. If we aim to be an energy independent industrial economy and continue to ship freight coast to coast, we have to knock down all of the things standing between us and a Rapid Freight Rail network, and so we have to knock those regulatory barriers down too.

    The benefit for passenger rail is gravy.

    ReplyDelete
  117. This is so far off-topic it's hard to know where to start, but since it was brought up...

    Here is an aerial view of the Sydney Central flying junctions.

    The Sydney Metropolitan network track map (650k PDF) may help in understanding what the flying junctions do.

    This is hardly the most amazing piece of rail engineering in the world, but it does demonstrate a basic level of professional competence, foresight and understanding of rail operations that is manifestly totally lacking from our local "professionals". ("Route conflict? What's that mean?")

    It was designed in the 1910s, by J. J. Bradfield, best known as the designer of the Sydney Harbour Bridge, but also the visionary architect of the (only partially built) Sydney suburban rail system.

    The fact that anybody is holding up Cityrail as an example of anything (other than of some century-old historical engineering foresight) just shows how pathetic/foamy/English-speaking-world-limited we are here. The insane politics of NSW transportation "planning" would be right at home in the Bay Area. Run away!

    ReplyDelete
  118. Speaking of off-topic and Cityrail, observe fast lines outside, local lines inside, island platforms in the centre (just as God intended, even if our local transportation "professionals" didn't get the memo.)

    They're even building more of it.

    ReplyDelete
  119. And Richard, note how they cleverly use curves to make up for the tracks swinging out for the platform (east of where you linked to), and it takes up more space than it has to (west of where you linked to). But note also the fact that it allows for easy turnbacks of local trains. Also note that the express trains take the divergence at the two-four track split. Note that the speed limit on that line can't possibly be as high as 125 mph. Also note that elsewhere, you see a FFSS arrangement. I think they really did that because it made for a convenient upgrade to an existing line with island platforms, and made the flying junction work out well. I don't know how well the FSSF arrangement will fit in the Caltrain ROW.

    ReplyDelete
  120. Also, I hear the speak English in Australia, so by your own standards, we shouldn't listen to anything they have to say about rail engineering.

    ReplyDelete
  121. Richard if She had intended the Peninsula to have local tracks in the center She would have given Southern Pacific the urge to buy a 300 foot wide ROW when they double tracked the line. :-)
    'Murcan Engineers can come up with all sorts of things. They've been fiddling with this since the '20s.

    ReplyDelete
  122. And Richard, note how they cleverly use curves [...]
    Thanks. Who ever would have noticed this incredibly subtle detail?

    I don't know how well the FSSF arrangement will fit in the Caltrain ROW.

    I do know. I've done all the sections and plans and I've used real-world (German standard dynamics, track separation blown out for fat-ass American trains) professional engineering criteria for track alignments and clearances. But if you want to believe that 1.95m lateral slew over several hundred metres run is the end of the earth, geometrically and vehicle dynamically impossible, well, that's a personal decision between you and your spiritual advisor of choice.

    F-S-S-F on open track (two rows of poles, 200kmh safety zones) is 17.3m centreline F-to-F (6.5+4.5+6.4). F-S-platform-S-F for local stops is 21.2m F-to-F (4.5+1.6+9.0+1.6+4.5).
    OOOOOHHHHH no! The curves are going to kill us! 300 foot wide rights of way! The sky is falling! Best just do things exactly the way they're done as set down by the PRR Founding Fathers when they received the stone tablets containing the alignment of the Sacred North East Corridor. (Next up: 25hz, 11kV: holy writ, immutable law, or sacred truth?)

    ReplyDelete
  123. 25hz, 11kV: holy writ, immutable law, or sacred truth

    I go for immutable law circa 1915. Even then it was a compromise. If it's 1915 and you are building an electrical system to drive trains you pick a low frequency. At 16 2/3 Hz the lights flicker unacceptably. At 25 Hz it's not that bad.

    It did give the H&M stations a certain charm. I forget when PATH finally replaced them with flickerless fluorescents for station lighting. They were still flickering away well into the 70s.

    If it's 1980 and you are NJ Transit you go with community power because Amtrak says they are going to replace the 25Hz stuff "soon"

    Does it really matter much today? From what I've read they are moving to converting whatever the power is to three phase.

    21.2m F-to-F (4.5+1.6+9.0+1.6+4.5)

    Sounds perfectly reasonable to me. But that doesn't leave room for brakemen to dangle off the caboose or the 8 platforms they are going to need at the HSR stations ( because they didn't pick the same platform height for Caltrain and HSR ) ...Ten... Because they are going to get the urge to run diesel hauled Superliners to San Francisco. Makes me wonder where they are going to put the ventilation tower at Transbay...

    ReplyDelete
  124. Richard: since you've clearly done the background work on this, what are the curve radii you get on the express tracks at the stations? What sort of superelevation and lateral forces are we talking about here? Also, is it really 1.95m of displacement? I thought in your design the platforms were somewhat wider. Anyhow, I just want to convince myself that the double reverse curves can be done in a reasonable amount of space, and without too much lateral force on the track or the passengers. And by the way, you oughtn't disparage the width of American trains. I bet the European operators really wish they could have a bit more horizontal clearance. Indeed, the Shinkansen is even wider than the US standard, and I think they were experimenting with extra-width trains in Sweden too.
    Adirondacker: who said the superliners had to be diesel-hauled? What's to stop Amtrak from sending a few ALP-46A's over to the West Coast and doing an engine change at San Jose?

    ReplyDelete
  125. who said the superliners had to be diesel-hauled? What's to stop Amtrak from sending a few ALP-46A's over to the West Coast and doing an engine change at San Jose?

    If Wikipedia is correct Amtrak doesn't have any. That might be a problem. Why would they change engines at San Jose? Why not at Gilroy?

    ReplyDelete
  126. Adirondacker: Amtrak has said that they intend to use stimulus funding to buy new electric motive power. Rumor has it that it will be ALP-46As, which makes sense given that NJT is already buying some, so the production line will already be there. In any case, Amtrak has and will continue to have electric locomotives, and hopefully will be able to spare a few for the Peninsula. And about the "actually have any," may I remind HSRA has no trains (or specs for any), Caltrain has no EMUs (or specs for any) or overhead wires, and Transbay Terminal is still a bus station, so everything here is really just speculation about the future.

    As for why San Jose, I just assumed that that would be the end of the Caltrain electrification, and the potential extension to Gilroy seems rather unlikely at the moment, for lack of money or, really, any need. It also makes more sense as most train crews are based at San Jose, and there's a CEMOF where the electric locomotives could be stored and maintained. But if the wire is extended to Gilroy and San Jose becomes too congested to do engine changes, then perhaps the hypothetical engine changes could be moved to Gilroy. I doubt that would be the case though, since any potential intercity trains would be restricted to run outside of rush hour.

    ReplyDelete
  127. Richard Mlynarik said... "The fact that anybody is holding up Cityrail as an example of anything ... just shows how pathetic/foamy/English-speaking-world-limited we are here."

    Recall that the structure of the argument was, "if even a decrepit system like Cityrail can accomplish X, surely X could be designed into a system designed today".

    ReplyDelete
  128. San Jose..end of the Caltrain electrification

    Rumor has it they are going to extend the electrification all the way to Los Angeles.

    ReplyDelete
  129. Adirondacker: yes, there's going to be electrification to LA, and even a whole new line, but I don't think anybody in their right mind is going to propose letting anything but HSR trains on it. And if HSRA's predictions of demand are even remotely right, there isn't going to be room in the timetable even for modern lightweight commuter/regional trains (like the class 395s in the UK).

    ReplyDelete
  130. I have posted an article that addresses many of issues with the Transbay Terminal. I'd really appreciate anyone coming by to read. Thanks!

    ReplyDelete
  131. arcady said...

    "Richard: since you've clearly done the background work on this, what are the curve radii you get on the express tracks at the stations? What sort of superelevation and lateral forces are we talking about here?"

    Say the lateral throw is 2 meters. According to 1950s-era US recommendations (the rules they would have followed if they had decided to run the Twentieth Century Limited at 125 mph) 300 meters of track isn't quite enough for that S-curve-- but we foamers don't know how conservative that is. My guess is your grandmother was supposed to be able to walk thru the diner without worrying about putting a hand in somebody's soup. Nowadays I'm guessing HSR designers accept higher lateral acceleration and (more importantly, if true) higher rate-of-change of lateral acceleration.

    ReplyDelete
  132. Next time you're riding a southward Caltrain express thru San Mateo see how you like the ride thru the S-curve just north of the depot on the usually-southward track. It's a 2.94-meter throw (+/- a couple cm) in maybe 190 meters, and by the 1950s recommendations it's not good for 79 mph. (The track tags say they're 0-45 curves with half-inch cant and 50-ft spirals.)

    ReplyDelete
  133. All off-topic, should be under the Slow Traffic Keep to the Middle entry, but nevertheless...

    ... the S-curve just north of the depot on the usually-southward track. It's a 2.94-meter throw (+/- a couple cm) in maybe 190 meters, and by the 1950s recommendations it's not good for 79 mph. (The track tags say they're 0-45 curves with half-inch cant and 50-ft spirals.)Caltrain engineering: always the gold standard.

    0'45" means (outside the Retarded Zone), r=2329m. Balancing elevation for v=127kmh would be 82mm; the usual mixed-traffic design goal of 0.7 v^2 / r suggests 50mm. Caltrain's operating at a cant deficiency of 70mm (assuming your reported 1/2" cant, which I haven't verified), so passengers would experience 0.45ms^-2 lateral acceleration (~= deficiency/153) in a 2329m radius 13mm cant constant radius curve. Hardly the end of the earth, so that in itself can't the cause of a bad ride.

    On the other hand, one would expect a transition length of 4.5 v delta_cant_deficiency / 1000 = 4.5 x 128 x 2 x 70 / 1000 = 80m between the reverse curves, and a Bloss (quintic parabola) transition spiral with length 40.3m per German track standards. Compare to the 50' = 15.2m curve transition you say Caltrain has.

    If laid out with a Bloss spiral, 2930m minimum radius, an intermediate straight of 70m, and at marginal low superelevation of 13mm, a pair of reverse curves complete with four transition spirals will slew 2.9m in 191m linear run. (Note there's only 10m of constant radius in each curve -- they're nearly all spiral in and out.) A similar run vs slew geometry is possible with nicer superelevations such as 50mm.

    So if Caltrain is slamming you about in this curve -- and THANK GOD I no longer ride Caltrain enough to remember every inch of its hatefulness -- it seems like the problem has to be with Caltrain track layout or Caltrain track maintenance or with Caltrain pre-historic we-hate-our-riders gallery car "suspension", because there seems no inherent vehicle dynamics reason this should happen just to achieve such a lateral track displacement, at least on rail system designed and maintained to first world standards.


    To answer an earlier question from somebody else, I picked 25000m curves through the intermediate stations (meaning zero problems with platform gaps) and 15000m to 25000m curves out of the straightaways. These are arbitrary, but so large that nobody ought to be able to have any legitimate quibble about slalom courses or other innumerate nonsense. They're also so large one can neglect transition spirals altogether at this level of detail.

    ReplyDelete
  134. ... will slew 2.9m in 191m linear run ...ARGH. I stupidly inserted the wrong calculation output.

    Sorry!

    Ending up parallel after spiral-circle-spiral-straight-spiral-circle-spiral requires 231m. (I'd re-used a formula for other purposes which doesn't include an end spiral and parallel straight.)

    So Caltrain would indeed appear to be cutting corners on the curve transitions. Big surprise.

    ReplyDelete
  135. "So if Caltrain is slamming you about in this [San Mateo] curve..."

    I didn't mean to suggest the San Mateo curve is presently bad-- it may be good enough. I'm suggesting people pay attention next time they ride thru there; if it is OK, maybe we don't have be as conservative as we maybe thought.

    ReplyDelete
  136. I've been looking over the TBT EIS, namely the train station project alternatives. Part of the reason they eliminated many of the alternative alignments was their inability to accomodate tail tracks. So if the tail tracks are not built as part of the TBT, would that force the TJPA to revisit their EIS, or is simply the ability to add tail tracks at some unforeseen future date sufficient? I still have a hard time believing the train station -must- be directly under the bus terminal.

    ReplyDelete
  137. Since they have been cleared on the present footprint, if they want to build the train-box as part of the original foundations of the TBT, it has to be within the present footprint.

    That's why I favor Richard's train-box ... it seems to be within the footprint, it is compatible with a more effective switching network in the station throat, and even if used with the TJPA DTX design, it would allow for loosening of the curves.

    ReplyDelete
  138. Aside from all the design challenges, the big question is whether or not a station in the Transbay Terminal makes sense.

    The Transbay Terminal is near but not in downtown San Francisco.

    The terminal is not served by Muni or BART. It is essentially a bus terminal.

    The circuitous and tight route to the terminal will be slow.

    So it doesn't seem to make much sense to extend Caltrain to the Transbay Terminal.

    Instead what about extending Caltrain to the Civic Center Muni/BART station? Such a station would provide a direct connection to the Muni/BART tunnels. This station would also symbolically serve the San Francisco civic center. In addition the station and tunnels could be designed to support eventually extending Caltrain under Van Ness, under the Golden Gate and into Marin.

    ReplyDelete
  139. What are the reasons for limiting a platform radius to 1000m if the minimal turning radius of the train can be much higher? Any information on this topic would be great.

    ReplyDelete
  140. Re : R. Mlynarik's 10 Apr.'09 0151
    Your link to page #469 has died. The current portal is :

    Kingsgrove - Revesby Quad.


    The old pages are available through the web archive :

    TIDC of NSW(AU) page #469

    ReplyDelete
  141. Re: Caltrain / HSR Tunnel Route
    The real problem is that the San Francisco government was blind in the 1980's. Had they incorporated a Caltrain tunnel into the foundations of Moscone Center So. (1981) and No. (1992) there would have been an easy shot into the old TBT. The route would have been :
    a) Up Fourth St. to Folsom;
    b) Diagonal under Moscone to Third + Minna;
    c) East on Minna to the TBT.
    Also, the Planning Commission could have required the SFMOMA (1995) to put a tunnel box under its foundation in return for being allowed to block Natoma at Third St. Then a stub tunnel pointing at Natoma from under Moscone would have a simple cut+cover link up for a Caltrain expansion / HSR tunnel via Natoma.

    Yes, this would have added to the costs of the Moscone Center and the SFMOMA. But the payback would have come from perhaps accelerating the Caltrain extension to the mid-1990's, getting some use out of the old TBT's lower level, and making the new TBT easier to build.

    Moscone Center
    SFMOMA
    Moscone Center area map [PDF]

    P.S. Just think of the construction of the Moscone Center and SFMOMA as cut+cover writ large.

    P.P.S. The Central Subway would NOT be a problem. That slab of pork doesn't need a tunnel until it reaches the Stockton Tunnel. The T-Third would do just fine with a transit hump on Fourth St. and a transit mall on Stockton St. from Market St. to the tunnel. At that point it could tie into a Northeast Tunnel that loops around to Fort Mason. (The above is just a little thinking outside of the box.)

    ReplyDelete
  142. New drawings from the April CHSRA board meeting:

    Caltrain is still single tracked in the throat but there's an "emergency crossover". The trainbox extends all the way to main, underneath that large high-rise. There's no tail tracks at all. Caltrain has two 200m platform tracks.

    ReplyDelete
  143. "It was designed in the 1910s, by J. J. Bradfield, best known as the designer of the Sydney Harbour Bridge, but also the visionary architect of the (only partially built) Sydney suburban rail system.

    The fact that anybody is holding up Cityrail as an example of anything (other than of some century-old historical engineering foresight)"
    I think I was rather holding it up as an example of exactly that.

    There isn't a flying junction of that comprehensive a design anywhere else in the world. It really is the only reason they can manage the tight headways, given that the system isn't exactly operated to the highest standards.

    I think it just goes to show that the good engineers of the 1910s designed things better than the Bechtel types are designing things today. :-( This is the period which got us the four-track IRT and the Manhattan Tunnel and Terminal Project and the land development scheme at Grand Central. (To be fair, there are also examples of terrible design from the period -- the tangled mess made of the Chicago rail system is a good one.)

    ReplyDelete
  144. Why does CalTrain have a double crossover right after the split from one track to two?

    ReplyDelete
  145. As had been mentioned by other people, I think the TTC is a solution in search of a problem. There is a low need for connection between HSR and CalTrain, but a high need for connection to BART, Market St Muni, and Market St shopping and offices. I just spent several days in the CalTrain 4th/King, Transbay, and Market St area, and I saw almost no one, residents or office workers, that would be served by a TTC location. I hope that CalTrain wakes up and stays out of TTC. It is a dead end in more than a physical sense. If there wasn't the piece of land at the TTC location, would anyone even be talking about a station there?

    ReplyDelete
  146. Well it happens to be pretty close to a lot of jobs (you know, those things that people commute to). It's not exactly the heart of the financial district, but it's probably about as close as you're going to get.

    ReplyDelete
  147. http://www.sfcta.org/sites/default/files/content/Executive/Meetings/board/2018/04-Apr-10/ENC%20-%20Peer%20Review%20Final%20Report.pdf

    ReplyDelete
  148. Is there anywhere this debate is being held currently?

    ReplyDelete