02 August 2010

Peninsula (Northeast) Corridor

It's official: as noted on page 7 of the latest program management team monthly progress report, the selected track arrangement for the peninsula corridor, a key architectural decision with extensive ramifications for Caltrain operations, is slow-fast-fast-slow (SFFS). That means the express tracks will be located in the middle of the right of way, with Caltrain local tracks flanking them on each side. This arrangement is similar to the Northeast Corridor throughout New Jersey (photo at right by theahnman). From the report:
While the June schedule does not show any change in the deliverables dates, in July there will be a change in the delivery of the 15% integration package from September 2010 to October 2010 to change the alignment of the Caltrain tracks to the outside and HST tracks to the inside tracks to reduce the footprint at Caltrain station locations (using side rather than center platforms).
The good news is that the previously-favored, segregated SSFF or FFSS arrangements seem to be off the table. Those arrangements would have prevented the express overtakes that are the key to an effective Caltrain timetable, and would have forced wrong-way movements whenever any one of four tracks went out of service. Wrong-way movements vastly diminish track capacity and exacerbate cascading delays, where one late train makes a mess of the whole timetable--as any regular Caltrain rider knows all too well. So that's the good news.

The bad news is that Caltrain will be stuck with side platforms. When a Caltrain track goes out of service, as it inevitably will, routing trains to the opposite platform will now require cutting across the HSR tracks, as shown in the diagram at left. No more central island platforms, which have undeniable operational advantages as well as simplified passenger access.

Given the choice between giving up island platforms (SFFS) or giving up overtakes (FFSS / SSFF), then SFFS wins by a mile because overtakes are key--but that's really a false choice, constrained by the menu of alternatives. The one alternative that is seemingly not being given due diligence is FSSF.

Their logic might go like this:
  1. Caltrain design criteria (Chapter 3, paragraph 1.1.d) and HSR Technical Memo 2.2.4 (Station Platform Geometric Design, section 6.1.3) dictate that platforms must be perfectly straight.

  2. Straight island platforms require a double-reverse curve "wow" around the platform at every station, conflicting with HSR Technical Memo 2.1.2 (Alignment Design Standards for High-Speed Train Operation, section 6.1) that prohibits more than four so-called direction changes per mile.

  3. The reverse curve "wow" beyond each end of the island platform consumes an inordinate amount of land precisely where it is most valuable, in the denser suburban cores where train stations tend to be located.
That's a perfectly logical chain of reasoning, but it is unfortunately based on a poorly conceived requirement that prohibits curved platforms.

As discussed in Football Island, curved platforms allow the "wow" around the center island platform to be much more compact, using barely any more land than the side platform configuration. The diagram at right (do you see the football?) highlights the difference in green, amounting to about a half acre. With the exceedingly generous clearances likely to be used--the diagram shows a mere 75 feet of right of way width!--the difference would be even less. Now imagine a straight island platform, not shown in the diagram, where the green area would need to bulge out over a far greater length; the extra area would amount to the entire area of the platform, 30 x 750 feet or another half acre. That is indeed a waste of valuable land.

Fast-Slow-Slow-Fast isn't some far-fetched concept. It works in Sweden, and it would work here, far more efficiently and flexibly than a carbon copy of the Northeast Corridor. Why not give a careful second thought to curved platforms, unencumbered by rote compliance with ill-considered specification requirements?

96 comments:

  1. Clem, didn't it mention in the report they changed the track configuration to lessen the ROW width and takings? It might even help during construction to have the SFFS. Build the new ROW, bridges, tracks, then finish up with the inner high speed tracks. Might be less moving around of tracks and less shu-fly's

    ReplyDelete
  2. The Northeast Corridor practice with taking a local track out of service is to run the locals on the express track of the same direction, and use bridge plates, either a small number of temporary ones to line up with selected doors of the train, or a large number to line up with all doors. This means you get a three track railway, as opposed to a potentially much lower capacity arrangement with a 2-track railway next to a single-track one. The temporary bridge plates can be deployed and removed fast enough for a mid-day track closure with the track reopening for rush hour. And yes, I agree that not allowing curved platforms at all is probably a bad idea, but I bet Sweden doesn't have an ADA-imposed level boarding mandate.

    ReplyDelete
  3. The ADA issue is spurious given a large enough curve radius, but there's another argument for SFFS, which is that it makes infill stations easier to build. It depends on whether Caltrain expects to decrease station spacing in the future; if it does, then SFFS is better than FSSF.

    ReplyDelete
  4. Another interesting tid-bit of information from the memo: "Key meetings were held this month with BNSF and the UPRR to discuss and resolve right-of-way issues." "On June 16, met with UPRR (Metro/Authority/PMT) to discuss CHSTP proposals to share existing Metro-owned ROW through San Fernando Valley and UPRR-owned ROW entering Palmdale."

    So, in other words, despite what many nay-sayers have claimed, UPRR is in fact engaged in discussions with the Authority to resolve the ROW issues. In other words, UPRR's "line in the sand" stance is simply a basis for negotiation.

    Also noted: "Responses to comments on the Bay Area-Central Valley EIR are being prepared for Board Action on the re-certification of the EIR expected to be at the September 2 Board Meeting."

    So only one more month to wait for the EIR to be recertified. It should be interesting to see how long it takes for the PCC to file a new lawsuit.

    ReplyDelete
  5. And there are definitely high-platform stations on the NEC with less than perfectly straight platforms. South Station, Back Bay, and Providence come to mind. So the "perfectly straight" design standard is spurious and indeed allowing curved platforms could allow the easing of some curvature at Lawrence with the existing SFFS arrangement if that's needed for 125 mph.

    ReplyDelete
  6. From the TWG2010June.pdf of your June 27 post, it looks like the Aerial configuration has FSSF with a center island for the San Antonio station.
    Am I mis reading it?

    Hopefully they will do diligence on both SFFS and FSSF.

    ReplyDelete
  7. @Peter - The PCC has never filed a lawsuit. That was PCL.

    ReplyDelete
  8. I also think that as they move along with actual design, they are realizing that some of their outlandish ideas wont work.

    ReplyDelete
  9. And that was original ideas

    ReplyDelete
  10. @ Anonymous

    You're right, the PCC has not. I fully expect them to file a new one, though.

    ReplyDelete
  11. If not the PCC, then at least a new group of their member cities will likely file a lawsuit.

    ReplyDelete
  12. Amazing job by the Peninsula Rail Program experts.

    There is exactly one argument -- and an irrelevant one, given actual ROW and the costs of misconfiguration -- in favour of SFFS (that being marginally narrower station footprint), and about a thousand against.

    There's just no competition and no question. It's cut and dried.

    There's simply no reason to choose to do things the wrong way, yet they always make idiotic, technically and economically unsupported decisions ... of their own free will.

    CBOSS, FRA, mixed traffic, 1% grades, FFSS/SFFS, Pacheco, incompatible stations and platforms, unworkable Transbay, terminating most trains short of the CBD, choosing to operate without overtakes, etc, etc, etc, .. all obvious, stupid, easily resolved problems, all with obvious fixes, every single one either unattended to, dismissed, ignored or or voluntarily screwed up by the Peninsula Rail Program's railroading professionals.

    Something for all of you to ponder, as a complete serious exercise: name one thing that Caltrain's engineering and planning departments have done competently in the last 5 years. One!

    What a tragedy. And how completely avoidable. It's not like any of this is rocket science. And it's not like anybody's asking for 100% professional competence or 100% outcomes ... just anything, anything, above 0% would be a nice start.

    ReplyDelete
  13. Caltrain First03 August, 2010 11:35

    Clem, methinks you oversell the advantages of FSSF over SFFS. Of all the recommendations on your otherwise laudable Corridor To Do List, the "Keep slow traffic in the middle" recommendation is the weakest.

    FSSF does offer some very mild operational advantages over SFFS, but both are vast improvements over SSFF/FFSS. That's the real distinction to be made.

    The big problem with FSSF is that it assumes massive corridor-wide "Big Bang" construction, and that's highly unlikely on the Peninsula. The reason why SFFS is so much more common on four-track corridors worldwide is because rail corridors tend to be developed incrementally over time, which is precisely the case with the Northeast Corridor. What happens when the stations are built to four-tracks while the mainline remains at two-tracks? This is a likely scenario, either due to lack of available funding or because the PRP adopts Japanese rail operation practices. The fast trains would be looping out at each and every station. FSSF really requires four mainline tracks, and that's just not going to happen overnight. It will be a long, long wait, so SFFS is more robust as the incremental solution.

    Getting freight off the Peninsula is going to remain difficult, even though it is desirable. SFFS obviously works better for the eastside freight sidings.

    ReplyDelete
  14. Caltrain First:

    So what if the express trains are looping out at various stations? They will be looping out somewhat after 4-tracking is complete, but it hardly matters, because the curve radii involved (circa 10km) hardly count as curves at all. FSSF does not in any way require four mainline tracks for the whole length of the corridor. The stations can be built in a phased manner as well

    ReplyDelete
  15. Joey,

    I'm guessing that by "looping out" he means that the express trains would need to go in the diverging direction at each of the crossovers. Amtrak's high-speed NEC crossovers with movable frogs are still only good for 80 mph in the diverging direction. Even the AVE or TGV crossovers are only good for 100 mph diverging, IIRC.

    ReplyDelete
  16. AFAIK there are high speed turnouts these days which support speeds up to 230 km/h (144 mph) in the diverging direction.

    ReplyDelete
  17. DB has installed a number of 200 km/h turnouts.

    ReplyDelete
  18. @Joey I doubt that you'd want to spend so much money investing in those crossovers. They might only be temporary anyway.

    Regardless, I agree with Caltrain First that the FSSF recommendation is the weakest of all of Clem's recommendations. I do mildly prefer FSSF over SFFS, but I don't think it's anywhere near the "slam dunk" case that Clem makes it out to be. Even with FSSF, the best solution for dealing with a stalled local train will often be to just skip that stop (and allow passengers to ride back one stop). Trying to manage bidirectional local traffic on a single track will quickly become a mess with anything more than 2-3 tph.

    In contrast, the advantage of either SFFS or FSSF over SSFF is HUGE. This choice has restored some of my hope that the project will eventually generate huge positive benefits for the Peninsula.

    ReplyDelete
  19. This was good news to hear. Not the greatest, but considering how the prospects were before, I'm relieved. Peninsula rail might end up being somewhat decent, instead of totally fubar'd.

    Of course, FSSF is technically superior on almost all counts, and CHSRA & friends sandbagged it. As Alon pointed out, the only area its lacking is infill stations, which is a legitimate criticism. More stations are viable possibilities if we have a shared 4 track corridor, places that aren't under consideration now. Freight is another issue, but I suspect under either SFFS or FSSF, freight might have to cross fast tracks somewhere.

    We still need harmonized platforms, signaling system, overtakes, etc.

    ReplyDelete
  20. SFFS seems like it could be a little bit quieter for residents living near the tracks.

    Also I think SFFS would require the fast trains to run on less banked tracks around turns, since the smallest radius is on the inside track.

    ReplyDelete
  21. Freight is another issue, but I suspect under either SFFS or FSSF, freight might have to cross fast tracks somewhere.

    I think you hit the nail on the head. The HSR people were said to be adamant that freight trains not touch HSR track anywhere under any circumstance, even with time separation. SFFS allows freight trains to run around a giant loop, with businesses on the east side served on the northbound leg, and the west side on the southbound leg. With the HSR flyover in Brisbane, switching to FFSS (viewed north) freight trains could not only serve the port of SF, but turn back without so much as setting a single wheel on HSR track. In a world of compromises, that certainly starts to make some sense.

    @CaltrainFirst: yes indeed, SFFS is a vast improvement over SSFF or FFSS. I acknowledged that as the good news.

    ReplyDelete
  22. @ RobBob

    I don't think there's much of a difference in terms of noise from a train going 110 mph vs. 125 mph.

    ReplyDelete
  23. RobBob: high-speed trains are built to minimize noise. This means that at any given speed, they're quieter than commuter trains. A high-speed train at 200 km/h may actually be quieter than a commuter train at 177 km/h.

    As for the canting issue, the difference in curve radius is tiny: it's equal to the track center distance, which is about 4.5 meters. A 4.5-meter difference in curve radius makes the difference between 177 km/h and 176.7 km/h.

    ReplyDelete
  24. Adirondacker1280003 August, 2010 22:11

    high-speed trains are built to minimize noise. This means that at any given speed, they're quieter than commuter trains. A high-speed train at 200 km/h may actually be quieter than a commuter train at 177 km/h.

    Either of them are going to be much quieter that the diesel hauled trains now on the tracks.

    Assuming trains make more noise the faster they go, Caltrain expresses will be going faster than Caltrain locals, it puts the Caltrain expresses farther away from people. A decibel here a decibel there can add up...

    ReplyDelete
  25. Caltrain expresses will rarely get to use the fast tracks. They'll probably just get them for overtakes.

    ReplyDelete
  26. OK! So we have the capacity to do overtakes. *Good*. Next design priority: *compatible platforms*. This is *critically* important. Keep banging away at that.

    Given that HSR will be buying more expensive trains and will have a smaller selection of choices, Caltrain needs to commit to using whatever platform height HSR chooses.

    And in order to Keep It Simple, it really ought to be one of the preexisting heights. (There's two "Euro" standards, high and low, there's a "UK" standard, a "Netherlands" standard, and two "US" standards, high and low -- that's more than enough choices for anyone, and I don't recommend the Netherlands standard).

    For ADA and cost reasons the lowest height which allows for roll-through cars is the most desirable choice; I haven't checked which one that is.

    ReplyDelete
  27. Adirondacker1280003 August, 2010 23:33

    Caltrain expresses will rarely get to use the fast tracks. They'll probably just get them for overtakes.

    Switching back and forth between tracks is one of the best ways to kill capacity - one train uses up slots on two tracks. Wreaks havoc on cross platform transfers too.

    ReplyDelete
  28. neroden: the cost of 48 inch platforms vs say 36 inch platforms is not going to be significant

    Adirondacker:

    When you have trains traveling at 3 or more different speeds (CalTrain locals, CalTrain expresses, HSR trains with a couple of different stopping patterns), weaving in and out proves to be the best way to maximize capacity on the smallest number of tracks. And the question is more or less irrelevant to cross platform transfers as long as the correct train is at the correct platform when it is scheduled to be there.

    ReplyDelete
  29. Adirondacker1280004 August, 2010 00:43

    Joey, when the express train on the local track needs to move to the express track or vice versa it needs an empty block to move into, there can't be a train on the express track ready to move into the block it will occupy. It leaves behind a empty block that will remain empty until another train passes the local. It sucks up huge amounts of capacity. When the line is busy it has to happen at exquisite precision, no waiting an extra ten seconds while granny gets all confused by boarding this new fangled electric train without any stairs.

    Having the express on the local track behind the local and doing the overtake in the station means the local has an extraordinarily long dwell time which pisses passengers off and slows the local down.

    Trains don't move between tracks unless there's no alternative on busy systems. ... not that the Peninsula is ever going to busy.

    ReplyDelete
  30. Having the express on the local track behind the local and doing the overtake in the station means the local has an extraordinarily long dwell time which pisses passengers off and slows the local down.

    That's why it makes sense to put the express train on the express track for one station before and one station after the transfer station.

    And you are correct that strategic overtakes require that neither the express or local tracks be saturated at any point in time (if they were we'd be building six tracks instead of four). But we're probably going to have at most 10tph CalTrain and less than 9tph HSR, neither of which would saturate either the express or local tracks, strategic overtakes seem to be the best way to fit the desired three service levels into two tracks per direction.

    ReplyDelete
  31. Adirondacker, you're assuming that the line is anywhere near its total capacity. It's not, the reason for four tracks is to be able to fit in non-stopping trains running at 100 or 125 mph together with locals running at an average of 40 mph with stops. You can have something like 2 minutes between trains with no stops, and there's going to be something like 12 tph total (realistically speaking) on the line: 4 express, 4 local, 4 HSR, so there will be plenty of space for a Caltrain express to get in front of a local, stop at Redwood City and Palo Alto and Mountain View, then get back on the express track behind the HSR train and pass another local on the way to San Jose. Anyway, I suspect that there will be a Caltrain express service that just stops at Mountain View and Palo Alto between SJ and SF (those are the highest ridership stations), and at 100 mph, that might even be faster than an HSR train with a top speed of 125 mph and two stops with longer dwell time.

    ReplyDelete
  32. An alternative way to build a four-track line would be to have the fast tracks on the outside and the slow tracks in the middle. This may go against the prevailing logic - which is highway dominated - but for railways it makes sense. It will of course require island platforms but it makes it easier for slow trains to switch tracks or be turned back at intemediate stations without having to run across the fast tracks. The London Underground uses this arrangement in places and it works very well.

    ReplyDelete
  33. Q: Somebody finally notices that 60% or more of trains to an from San Francisco reverse at Mountain View, particularly at high traffic hours, without affecting total system rider count, while saving >10% of total system annual running cost and saving >15% of the commuter fleet size. How can this be done with a SFFS track configuration without blocking all four tracks -- at peak hour?

    Q: 80% of Caltrain terminate in San Jose and return north. How can this be done without blocking all four tracks at peak hour, and without requiring an extremely expensive, redundant second level built overhead for through trains, just to make up for an extraordinarily stupidly chosen track configuration further north?

    Q: A Caltrain experiences a failure at or near a station. (These are most common of all failures, not just for Caltrain. Door systems, passenger self-injury, brake system, traction control, etc.) How does Caltrain continue to provide any level of service to that station in both directions while the faulty train is cleared ... without blocking all four tracks?

    Q: Note that with no or very few stops, the above scenario is not an issue for High Speed Trains on the corridor. Nevertheless, show how a High Speed train failure anywhere can be routed around only affecting 2 tracks.

    Q: Compute the number of aggregate track-minutes during which tracks are blocked (ie the interlocking occupancy time) for the following maneuvers:
    S->F->F->S
    S<->F
    F<->F
    and then weight them by the number of trains operating on the F and S tracks. You may realistically assume 2x as many Caltrain as High Speed Trains.
    Now compare and contrast with a corridor configured with a more rationally engineered arrangement.

    Q: Stations on very high speed lines are typically configured SFFS or SSFFSS to allow absolutely straight run-through for the highest speed trains, which are the major users of very high speed new lines. Explain the practical effect that not having a dead straight run (all the way from San Jose to San Francisco, guided by laser beams) would have for trains that bypass intermediate stations.

    Q: Freight on the Caltrain line is a requirement that drives the need for quad-tracking everywhere, for CBOSS, for 1% maximum grades, and for SFFS. Show where this "requirement" comes from. Show how it "necessarily" implies all of the other measures.

    ReplyDelete
  34. With the HSR flyover in Brisbane, switching to FFSS (viewed north) freight trains could not only serve the port of SF, but turn back without so much as setting a single wheel on HSR track.

    I thought the decision implied SFFS for everywhere south of 4th & Townsend.

    Or do you think they'll still try to do FFSS north of Brisbane and south of Santa Clara?

    ReplyDelete
  35. @ Samsonian

    South of Santa Clara they will likely be separated at different levels, with HSR elevated.

    We don't know the full extent to which they are implementing SFFS. All that we have to go by is the memo that Clem linked to in the post. Expect more tomorrow.

    ReplyDelete
  36. Clem wrote: "The HSR people were said to be adamant that freight trains not touch HSR track anywhere under any circumstance, even with time separation."

    Hi, I'm missing the part in any MOU, in any public announcement, in anhy policy statement, or in any economic transfer of rights where "the HSR people", minority tenants of a public transportation corridor, not only get their own exclusive dedicated "HSR tracks" over which they have complete policy control.

    Do we take this to mean that CHSRA's special snowflake trains are so special that their delicate wheels cannot even touch rails beaten to crap by a single traversal of mighty UPRR? If so, how do we reconcile this with the stated position (ie "strategic misrepresentation") of the Peninsula Rail Program transportation professionals and the PCJPB executive that SF-SJ will be a shared corridor, any train, any track? And we must assume this means that even in the case of track maintenance, train or systems failure, no Special Snowflake HST may run on the sullied slow snail-train tracks.

    Do we take this to mean that the two shared tracks from Redwood Junction to Mountain View which is proposed in the latest CHSRA federal pork extraction document will be allocated one to UPRR and Caltrain, and one to HSR? There can be no other possible explanation.

    Given that the economically and technically woeful (but consultant bonanza inducing) planning of the Peninsula Rail Program calls for operation of mixed traffic FRA behemoths, including Caltrain and Amtrak locomotives that weigh more than UP freight cars and as much as many freight locomotives, the only possible conclusion is that Caltrain (and all the terribly, terribly important Amtrak Day Starlight runs) will also be 100% excluded from ever touching the HSR Only No Girls Allowed Tracks.

    Another obvious implication is that the HSR-only tracks are to be maintained to different standards from the sullied snail tracks or may use different control systems. Just how this is supposed to be reconciled with automated maintenance machinery, common risks of derailment (snail trains falling off sadly sub-maintained sullied tracks can do a nice job on the special snowflake trains running a couple feet away from them, you can be sure) is something that only the US's finest engineering minds at the Peninsula Rail Program could possibly clarify.

    And speaking of risk, if the Special Snowflake trains are so afraid of sullied UPRR tracks, even 4 hours after the freight train has departed the scene, based presumably on danger from track damage, why is it any way acceptable that twice as many Caltrain passengers be exposed to that risk while riding larger numbers of Caltrain services on the snail tracks? After all, the Peninsula Rail Program transportation professionals assure us that Caltrain will be using special Eurosnowflake trains (well, with the 100% markup for pals at LTK Engineering Services for special local conditions redesign overhead); won't these be just as likely to be assaulted by UPRR as the HST Eurosnowflakes are? Why is bad old UP murdering commuters more acceptable than knocking off people heading to Disneyland or Los Banos?

    In short: what we're seeing here is what we see every time: third-rate engineers proceed with their third-rate fixed ideas, completely impervious to change, logic, or experience, and produce transparent tissues of outright lies (sorry, "strategic misrepresentations") to justify the only thing they ever were going to do, and the only thing of which they have any experience in their tiny incestuous little fraternal world.

    HSR tracks, on your right of way, owned and controlled by ... somebody, not sure who, and not sure why, and not sure by what right, but whatever it is it's fine with the Peninsula Rail Program professionals and you can bet it's Context Sensitive out the wazoo.

    ReplyDelete
  37. HSR tracks, on your right of way, owned and controlled by ... somebody, not sure who, and not sure why, and not sure by what right

    I've been wondering about this for a while.

    Who will own and/or control the new tracks?

    The MoU doesn't really lay out that out. Also the MoU can be terminated by either party with 30 days notice.

    What happens to the new infrastructure if either party terminates the MoU?

    ReplyDelete
  38. Adirondacker1280004 August, 2010 15:05

    Somebody finally notices that 60% or more of trains to an from San Francisco reverse at Mountain View,
    ...


    I dunno how do the world class operators, Amtrak and NJTransit do it at Jersey Ave? The northbound platform at Jersey Ave is on the west side of the southbound tracks. ( or the eastbound platform is north of the westbound tracks using railroad conventions ) They then do it again at the border between Newark and Elizabeth where the Raritan Valley line merges in. The North Jersey Coast line has a duckunder in Rahway and the NEC is six tracks wide. The Morris and Essex lines have a duckunder in Kearney and the NEC is only two tracks wide there. Or the NEC has a flyover whichever way you want to look at it.

    Or they could ask the world class operator Metra how they do it. Or the world class operator Metro North how they do it. Or the LIRR or...

    There's lots of ways to do it. If there's so little traffic that the expresses are bobbing and weaving onto the express tracks and back onto the local tracks in other places it's shouldn't be much of a problem at Mountain View. Or they can have the trains that turn at Mountain View move from the local tracks to the express tracks north of Mountain View and then arrive at Mountain View on the express platform. They can depart from the express platform. If it's too busy for that a pocket track or two should alleviate things. Or if it's too busy for that, which I doubt it will be unless automobiles are outlawed there's those things called flyovers and duckunders....

    Keep in mind that anyone who suggests that the San Jose express run local between San Jose and Mountain View and express between Mountain View and San Francisco or that the Mountain View express run local between Mountain View and Redwood City or Hillsdale and then express between there and San Francisco ....watches heads explode as Californians alternately think "but that's not the way BART runs" and or "but that's the way Metra, Metro North, LIRR, NJTransit, Septa etc run and we have palm trees so that can be done here" Or the way the CTA runs some El lines or the way the MTA runs some subway lines or...

    80% of Caltrain terminate in San Jose and return north. How can this be done without blocking all four tracks at peak hour, and without requiring....

    I dunno maybe they can ask that other world class operator the Long Island Rail Road how they manage to get a train a minute in each direction at Jamaica using 8 platforms. I think there are bypass tracks in Jamaica, I'd have to dig up track maps. There's never going to be a train in minute in each direction in San Jose.

    A Caltrain experiences a failure at or near a station. ...

    The same thing that happens all over the world on busy lines when that happens? Everything moving in the direction with a problem expresses through and the unlucky passengers go to the next express stop and take the local in the opposite direction back to their station. The thing that makes a train an express train is the lack of stopping not the track it's on. Trains can express on the local track. If the line is so lightly used that the Caltrain expresses are bobbing and weaving around the local it can bob and weave like it normally does.

    Compute the number of aggregate track-minutes during which tracks are blocked (ie the interlocking occupancy time) for the following maneuvers:
    S->F->F->S


    And what happens when the fast tracks are blocked in FSSF configuration and the fast train is coming through. Six of one half a dozen of the other for any individual train. And moving from the fast track to the slow track or vice versa will take the same amount of time regardless of whether the tracks are arranged FSSF or SFFS.

    ReplyDelete
  39. Caltrain First04 August, 2010 16:01

    If a four-track Peninsula corridor is genuinely interoperable (as it most certainly should be), the comparison of track blockage scenarios between SFFS and FSSF is moot. Trains get routed to the adjacent same-direction track, regardless of whether it is 'slow' or 'fast'. With interoperability --the real issue here-- a Caltrain local can easily be allowed to go on a 'fast' track to get around a blockage. Why on earth would one cause far more serious capacity constraints by sending, say, a southbound train onto a northbound track?

    The one key advantage FSSF provides is the center platform at local stations, which allows waiting passengers the ease of crossing platforms for another service or a rerouted train, but this isn't a big deal at minor local stops. SFFS can still involve center platforms at the more important Baby Bullet express stations, allowing for easy local-express cross-platform transfers: S-platform-F-F-platform-S.

    A key advantage of SFFS is the ability to easily add infill stations, and this could be a big deal on the Peninsula. A four-track corridor would give the Peninsula all the track capacity it could ever hope to need (especially considering the growth constraints the Peninsula has imposed on itself), and with all this capacity, the 'slow' tracks would use their capacity for genuinely local service, knitting Peninsula communities together with a sort of light rail/metro service. With more capacity dedicated for local service, it makes sense to add more infill stations to deepen the service reach. The 'fast' tracks become tasked with commuter expresses and HSR. (BTW, HSR service will be made to fall in line with the more frequent commuter express services).

    Also, considering that freight movements should be time segregated even if allowed to remain, capacity-eating crossing movements don't really make any difference at 1am.

    ReplyDelete
  40. It's amazing how hard people -- people who have a chance of having a clue and aren't US transportation consultants -- work to misunderstand!

    The key point is that FSSF has zero disadvantages and many significant advantages, so why go a thousand miles out of your way to make excuses for how SFFS could work somehow with enough effort, compromise and expense?

    Most important to understand is that it enables as much as possible of Caltrain operation -- recall Caltrain riders will always be by far the the majority users of the public corridor -- to function independently of HSR and HSR screwups. With two adjacent tracks and with the ability to use either side of island platforms for trains from either direction, the ability to avoid, work around and recover from failures or maintenance is immensely improved. Caltrain becomes master of its own fate to the greatest possible extent, rather than a slave to HSR shenanigans. Yes, cooperation is always needed, but the less it is required the better for everybody in the world.

    "Infill stations". OK. Go ahead. Make my day. Name one location. Talk about clutching at straws! (I'll raise you two anti-infill stations: Hayward Park has got to go, for very obvious geometric reasons, and Atherton is never coming back.)

    The people claiming that mid-line (or San Jose) turnbacks with FSSF are easy as with SFFS? Pure counter-factualism; where to even begin? Please get out a piece of paper and a pencil and draw a little hand animation of what happens, minute by minute. It's instructive, and even sort of fun the first time. Flat junctions (which is what these are, without flyover$ and duckunder$) kill capacity, crossing opposing direction running lines is a total capacity killer, crossing fast tracks kills far more capacity than slotting into slow traffic, crossing two directions of fast traffic and trying to thread the temporal needle is, in practice, almost impossible if the fast tracks are actually used enough to justify having been constructed in the first place! The bottom line is that it can't be done. The good news is that there's no need to do it. The bad news is that Peninsula Rail Program don't give a damn.

    "Flyovers"? "Duckunders? Good God! And you're (Adirondacker1200) the one happy to complain about California burning your federal tax dollars? Why throw concrete (= contractor profit$) at something when having a clue gets you 1000 times further? Words fail.

    There's a slogan used over and over by Swiss railway planners: "Elektronik vor Beton", ("Employ electronics first, concrete last", aka "Work Smarter not harder"). There's a great example near Olten at the critical point of their national network where, by optimizing crossover and signal placement and adding a couple signal blocks, in combination with careful timetabling, they were able to defer the massive expense of a flyover at an important but space constrained junction for nearly 20 years.

    So why a two level station in San Jose? Why elevated flyovers in Brisbane and Santa Clara? Why new tunnels into San Francisco? Why two stations in San Francisco? You know my answer.

    I just don't get all this Stockholm Syndrome BS here. Caltrain is being thrown under the bus ... by Caltrain employees, and you're busy working out ways in which being roadkill might be regarded as pleasant experience after all. If there were any economic, logistical or even political reason to choose to do the wrong thing that would be one thing: but this, like San Bruno, like Transbay, is a blatant example of bloody-minded, outrageous incompetence and unprofessionalism.

    ReplyDelete
  41. Adirondacker, Mountain View is not Jersey Avenue. While the NEC Line gets a lot of traffic, there are relatively few through-trains going to Princeton Junction. At the peak hour, NJT runs 4 tph inbound and 2 tph outbound using the express tracks at Jersey Avenue, and Amtrak another 2 tph per direction. This is different from the expected traffic volume on the Caltrain corridor: there will be a lot more intercity trains, and potentially more Caltrain through-trains to San Jose on account of the line's more symmetric demand.

    And Richard, I think Brisbane could be a good infill station, for one. Oakdale would be another. (Caltrain really needs to let go of the North American concept of commuter rail as being only for the suburbs, making as few stops in the central city as possible.)

    ReplyDelete
  42. Adirondacker1280004 August, 2010 21:25

    Good God! And you're (Adirondacker1200) the one happy to complain about California burning your federal tax dollars?

    What part of not needing either until cars are outlawed wasn't clear?

    I suspect during peak periods Caltrain is going to be running lots of trains express in one manner or another. With SFFS the ones on the express tracks can reverse as easily as the locals can in FSSF. Outside of peak when less trains in general are running, that the locals have to cross over those big scary he-manly express tracks isn't a much of a problem as it during peak.

    FSSF has some advantages. Mostly to do with cheaper stations that are a bit more passenger friendly. It's not something to set your hair on fire over.

    Mountain View is not Jersey Avenue.

    No it's not and probably never will be either. Came out pretty well for a PRR low cost experiment in park-n-rides. Probably will be converted to a conventional two side platform station around the same time North Brunswick is built. Never know though NJTransit may decide to rebuild the branch line that goes west from there instead.

    there are relatively few through-trains going to Princeton Junction.

    They are all through trains in Princeton Junction unless you mean the Dinky which is the only thing that terminates there. The Dinky is on it's own separate single track row so I don't think you had the Dinky in mind.

    ReplyDelete
  43. Locals are A LOT more likely to reverse mid-line than expresses.

    ReplyDelete
  44. If you want to talk about at-grade junctions across busy trail lines, the Hunter interlocking on NEC is probably a good example. The Raritan Valley Line branches off to the west at an at-grade junction. Northbound trains merging into the NEC have to cross over all the tracks, and the crossovers aren't high speed either (45 mph, I believe). There's also CP Shell at New Rochelle, where southbound Amtrak trains cross all the Metro North tracks. That was recently reconfigured so that southbound Amtrak trains cross from the SB express to the NB express before the station, stop on the NB express (if they're stopping), then cross the NB local south of the station. NB trains remain on the local track through the station and cross over to the express north of it on a high speed (80 mph) crossover. It all seems to work out pretty nicely now.

    As for SFFS versus FSSF, there's a slight advantage in pedestrian circulation with the platforms on the outside when your line is at-grade like Caltrain. At a station with the parking lot on the west side (as most stations are), the passengers getting off the southbound train can leave the platform along its entire length, rather than just at some small number of stairways. True, in the other direction they'd still have to take some stairs, but there's fewer people getting off northbound trains in the afternoons and passengers getting on trains tend to arrive over some period of time rather than all at once. And I think the single tracking local operation on FSSF is not such a great idea operationally: it means your local trains are now late in both directions, while bypassing stations on the express means that the other direction is not affected at all, while the affected direction trains can still stay on time, or else the passengers can just switch to the express. There'll be plenty of stations with opportunities to do so: Millbrae, Hillsdale, Redwood City, Palo Alto, Mountain View at the very least, probably Sunnyvale and San Mateo as well, and maybe also Lawrence and South SF.

    ReplyDelete
  45. The likelihood of reversing trains between SF-SJ on a regular schedule is very little. The distance is relatively short (about an hour one way) and both termini have strong ridership markets (including potential markets). San Jose has strong transit connections with light rail and regional buses.

    It is possible that Caltrain may one day have to serve Giants in SF and A's in SJ on the same night! That's already happening with Sharks (although it doesn't draw ridership as much as baseball).

    ReplyDelete
  46. No it's not and probably never will be either.

    You're thinking in the wrong direction. Mountain View has 3,200 weekday boardings and Jersey Avenue 1,600.

    They are all through trains in Princeton Junction unless you mean the Dinky which is the only thing that terminates there.

    "Through-trains to Princeton Junction" in the context of Jersey Avenue means "trains that don't terminate at Jersey Avenue, and therefore run on track that the Jersey Avenue locals will cross at-grade."

    Seriously, I thought Jersey had one of the best public school systems. How come your reading comprehension skills suck so much?

    ReplyDelete
  47. Caltrain First05 August, 2010 00:44

    While it's certainly true that turnback movements can eat up capacity, midline turnbacks on the Peninsula will be very rare. Any midline turnback also encounters a paradox with regards to the resulting capacity losses. One plans a midline turnback precisely because the corridor traffic demand is lackluster or insufficient to justify full operations, so the act of installing a midline turnback implies excess capacity to spare.

    It's also crucial to get away from the idea that the S track is for this service and the F track is for that service, etc. All four tracks should be optimized. One idea to consider is the use of an assortment of local-express hybrid services where a single train runs local along one geographic segment of the Peninsula, picking up riders (eg, all stations between San Jose and Palo Alto), and then switches into a full express with a service pattern matching HSR (eg, Redwood City, Millbrae, and San Francisco stops). This would allow Caltrain expresses to keep time with HSR on the fast track.

    ReplyDelete
  48. Caltrain First, the idea you're proposing for express-local mix exists in Greater New York, but not in areas where commuter rail is run by halfway competent people. The basic problem with it is that it makes local travel annoying, and doesn't improve local-to-express travel over running separate local and express trains with zero-wait transfers (Caltrain's six-minute "timed transfer" is a sad joke).

    The only time this mixture works is when you have all local trains terminating mid-line, and all express trains running local further out. For example, on the Chuo Line, local trains make all stops until Mitaka, whereas express trains only make a few stops, but then start running local from Mitaka until Takao. Even the geniuses of Metro-North default to this service pattern off-peak, when there's not enough need to run 20 trains in one hour, each making only 4 suburban stops.

    ReplyDelete
  49. Adirondacker1280005 August, 2010 09:37

    Seriously, I thought Jersey had one of the best public school systems. How come your reading comprehension skills suck so much?

    Some poor defenseless tourist snags you at the Times Square IRT station and wants to know which train to take to get to 23rd Street. Do you you tell him to take one that goes to Sheridan Square or do you tell him to take one that goes to South Ferry? If service was running normally I'd tell him to take one that is against the wall and point to the local track. Or since I was going to Houston Street I'd tell him "the one I'm taking, follow me" If he wanted to get to Wall Street I'd tell him to take the train going to New Lots or Flatbush. That all of them stop at Chambers Street is irrelevant.

    Service to Princeton Junction except from Princeton itself is provided by trains that go to Trenton or come from Trenton unless you want to wait around for the few trains a day that Amtrak deigns to have stop at Princeton Junction.It used to be provided by trains that went to Philadelphia in addition to the ones that went to Trenton. Many many decades ago local service from New York terminated in New Brunswick and to get to Princeton you had to take a train that went to Philadelphia or points farther. Many many decades ago trains to and from Princeton would go through the junction and wander farther afield.

    The trains to Trenton from New York rarely if ever turn around in Trenton but go all the way to Pennsylvania to do it. Telling someone that the train goes to Morrisville would be very very confusing even to foamers because the trains that stop in Princeton Junction are either going to Trenton or points beyond Philadelphia. The only train that goes to Princeton Junction is the Dinky. Sorry you grasp of that convention is so poor.

    I pity the poor tourist in Times Square who asks you how to get to the Museum of Natural History because in your little convoluted world they have to take a train that goes to Cathedral Parkway. They are going to have a very long wait because trains don't go to Cathedral Parkway.

    ReplyDelete
  50. Peninsula Rail Perjury Program

    "Track Configuration

    The Supplemental AA Report recommends that the design and environmental efforts focus on a horizontal track configuration that has Caltrain predominantly operating on the outside two tracks and HST on the inside two tracks (see figure S-4). This configuration is recommended primarily because it requires significantly less (approximately 20% less) right of way than having both Caltrain tracks on one side of the corridor (see figure S-5). This reduced need for ROW would be particularly significant where Caltrain stations are close together (approximately a mile apart) and there is insufficient distance to narrow the ROW width between stations. This configuration also allows greater flexibility in coordinating schedules and sharing track capacity on the corridor for the reason that it would allow HST trains overtake other trains in certain areas without crossing opposing rail traffic.
    "

    In a better world, there would be penalties -- legal, financial, personal, and professional -- for people who engage in this sort of sleazy and grotesquely unethical mendacity. Not to mention the laughable level of professionalism represented by this extensive "analysis". Like, um, dude, where's, like, the analysis part at, and, the, um, alternatives and costs and the trade-offs and the benefits and like that sort of shit?

    America's finest transportation professionals. Stealing your tax dollars to support lives of habitual disingenuousness.

    ReplyDelete
  51. @anonymous - 04 August, 2010 09:10

    Q: 80% of Caltrain terminate in San Jose
    (and 60% in Mountain View) and return north. How can this be done without blocking all four tracks at peak hour, and without requiring an extremely expensive, redundant second level built overhead for through trains, just to make up for an extraordinarily stupidly chosen track configuration further north?


    A: The question implies certain mindsets, such as 'slow' trains changing directions must operate on the 'slow' tracks for as much as possible, and can never operate on the 'fast' tracks. A solution of having the turnback train shift from southbound 'slow' to southbound 'fast' at speed north of the turnback station, terminating in the southbound 'fast' platform (to support cross-platform transfers with southbound expresses), then heading northwards on the 'fast' tracks to later change to the 'slow' tracks at speed is ruled out by the phrasing of the question.

    Other solutions such as having a stub track south of (or in) the station between the two 'fast' tracks to hold the train (eg, SFSFS) clear of everything while the driver changes ends are also ruled out. (This also implies that the last/first stop of the train is made on the south/north 'fast' platforms)

    As Richard said in his own fashion, a bit of thought towards 'Work Smarter not harder' makes this problem seem like a relatively minor issue.

    ReplyDelete
  52. "Caltrain First" wrote: "... midline turnbacks on the Peninsula will be very rare. ... One plans a midline turnback precisely because the corridor traffic demand is lackluster or insufficient to justify full operations, so the act of installing a midline turnback implies excess capacity to spare"

    I can't imagine where you get your "encounters a paradox with regards to" conclusion out of this.

    Indeed one turns back trains precisely because demand is not equal at all points of the corridor.

    Yes there may be "excess capacity" in terms of open track available which could be filled up with more trains, but the point is that operating those trains isn't free, especially when they're filled with empty seats.

    Here's a exercise: there's lots of "excess capacity" on the empty track from San Jose to Gilroy. Therefore at least two Caltrain per hour, bidirectionally, should run on this segment, because otherwise the track capacity would be going to waste.

    I simply don't get this wholesale advocacy of inefficiency, and worse, this advocacy of forever shutting out future operating flexibility, flexibility for either scheduled or unscheduled operating needs.

    "Anonymous" above gave one of my favourite examples of where Caltrain should consider turning back trains: Mountain View. (Perhaps anon is somebody I know, or perhaps great minds think alike, or perhaps anybody who rides Caltrain with his or her eyes open would soon propose the same thing.)

    As noted, there are significant opportunities to reduce fleet capital cost and fleet operating cost by matching service to demand at the peaks (which is what drives fleet and crewing size) without significantly degrading provided service quality. As one idea among many possibilities, running the non-turnback trains local south of Mountain View, where the number of stops is few and the inter-stop spacing great, could be an excellent trade-off that could still provide excellent service (15 minute headway) and excellent travel times (cross-platform transfers in Mountain View.)

    But you'd have to have any interest in allowing alternatives.

    What if San Jose became do TOD-of-riffic (circa 2070) that the S-Bahn San Jose justified 7 minute headway local trains shuttling Sunnyvale-SJ-Blossom Hill?

    San Francisco to San Jose, at over 80km, is an inter-city corridor. The correct way to think of regional service on the corridor that there are inter-city trains (regional express) and there are intra-regional S-Bahn trains associated with various agglomerations. Now because of the reach of the various S-Bahn it will make operational sense to combine some of them: eg local service on the "S-Bahn Palo Alto" continues through as a local train of the "S-Bahn San Jose", instead of having two different trains turn back from the north and south at, say, Mountain View or Sunnyvale, but the thing to understand is that sub-markets and uneven traffic flows within the corridor, based upon non-uniform demographics, and that allowing the possibility of flexibly allocating costly train service to match that geographical reality is an unalloyed good: all benefit and nearly zero cost.

    Why do things worse for no reason? What's with the Stockholm Syndrome?

    ReplyDelete
  53. "thatbruce" wrote: "A solution of having the turnback train shift from southbound 'slow' to southbound 'fast' at speed north of the turnback station, terminating in the southbound 'fast' platform (to support cross-platform transfers with southbound expresses), then heading northwards on the 'fast' tracks to later change to the 'slow' tracks at speed is ruled out by the phrasing of the question."

    Hi, I suggested the exercise of taking out a pencil and paper and creating a little minute by minute hand animation of what this implies in real life. You should try it. (I could write a program to do it for you, but "teach a man to fish ..." and all that jive.)

    Questions for you to ponder as or after you do the exercise:
    1. Why does nobody else in the world do as you suggest?
    2. Why jump through hoops to do the wrong thing, less efficiently?
    3. It can't be said enough: just slotting a move across two tracks with opposing direction traffic -- and then decelerating to a stop on one of them -- consumes an immense amount of practical, robustly-operable, timetable capacity.

    Do the exercise!

    PS as a pleasant virtual tourism distraction from local unprofessionalism: here's a high speed turnback station done right.

    FS_SS_S_F, where the "F" is really fast, and the downgrades are significant -- and unequal between the S and F tracks. A very nice piece of engineering indeed. And obviously not designed by any American.

    Naturally with interoperable 760mm platforms, ETCS/ERTMS, UIC60 standard turnouts and rail, ... all proven, all interoparable, all went into service flawlessly.

    ReplyDelete
  54. The same quote that Richard just posted from the AA also has this tidbit, which those focused on 'slow must stay on slow, fast must stay on fast' should pay attention to:


    The Supplemental AA Report recommends that the design and environmental efforts focus on a horizontal track configuration that has Caltrain predominantly operating on the outside two tracks and HST on the inside two tracks (see figure S-4).


    ie, Caltrain will normally be operating on the outside 'slow' tracks. However, Caltrain operating select services at select locations on the inner 'fast' tracks (for say, operational requirements to optimise turnbacks or recover from some sort of service interruption) is not ruled out.

    Personally, as long as all four tracks are built and maintained to the same standards, the platforms are compatible with all regular trains operated along the route, and trains can shift between 'slow' and 'fast' tracks at their normal operating speed, SFFS vs FSSF can be changed around depending on the operational requirements of each day.
    Of course, there are some doubts about all of these provisions being met ;)

    ReplyDelete
  55. You can't switch between FSSF and SFFS without building extra platforms at every local station, Bruce.

    ReplyDelete
  56. thatbruce: So they must also shift the platform locations around from one set of tracks (where they are) to another (where they weren't yesterday) in order to make this happy unicorn best of all possible worlds flexibility happen.

    Reality: once the wrong thing is cast in concrete (Transbay, San Bruno, ...) it can't be undone.

    PS Your "shift between 'slow' and 'fast' tracks at their normal operating speed" belief isn't one based on reality. A 130kmh crossover (2500-1.26.5/1.27.85 with 4.5m track spacing) is 130m long. For 200kmh (16000/1600-∞ clothoidal) you're looking at 377m. First you need a place to put them. Then you need to justify paying, a lot, for these puppies. In a phrase: "Not going to happen." Or you could instead go the "Work smarter, not harder" route, just for novelty's sake.

    ReplyDelete
  57. Quoting again:

    thatbruce:
    A solution of having the turnback train shift from southbound 'slow' to southbound 'fast' at speed north of the turnback station, terminating in the southbound 'fast' platform (to support cross-platform transfers with southbound expresses), then heading northwards on the 'fast' tracks to later change to the 'slow' tracks at speed is ruled out by the phrasing of the question.

    richard:
    3. It can't be said enough: just slotting a move across two tracks with opposing direction traffic -- and then decelerating to a stop on one of them -- consumes an immense amount of practical, robustly-operable, timetable capacity.

    Ok, lets work through this. Bear in mind that the solution I proposed (quoted above) does not involve crossing two tracks of traffic at the same time. I'll note that I didn't feel the need to initially state the obvious, that of the fast trains using the 'slow' tracks through the turnback station while the train turning around is occupying the 'fast' tracks.

    In this example, both Local Train A and Express Train B are heading south from San Francisco, and will both arrive at Mountain View (or San Jose) to perform a cross-platform transfer. Northbound Express Train C will also make a stop at Mountain View at the same time. Local Train A will then turnaround at Mountain View (or San Jose) and head north while Express Train B will continue south, and Express Train C will head north.

    The stations used in this example, San Antonio (or College Park) has one side platform on the southbound outer track, and one island platform between the northbound inner and outer tracks, while Mountain View (or San Jose) has two island platforms between the outer and inner tracks.

    At the start, Local Train A has just made a stop at San Antonio station, using the outside southbound track. It then gets back up to speed and transfers to the inner southbound track between San Antonio and Mountain View.

    At the same time, southbound Express Train B is passing through Menlo Park on the inner southbound track. It makes a transfer to the outer southbound track between Menlo Park and California Ave Stations, then proceeds on the outer southbound track through California Ave and San Antonio Stations.

    Also at the same time, northbound Express Train C is passing through Lawrence station on the inner northbound track. It transfers to the outer northbound track between Lawrence and Sunnyvale stations.

    Local Train A, having arrived at Mountain View on the inner southbound track, undergoes turnaround. While the train crew is changing ends, the southbound Express Train B arrives at the outer southbound platform, and passengers perform their cross-platform transfer dance. Northbound Express Train C arrives at the outer northbound platform at the same time. Express Train B departs southwards, while Express Train C departs northwards, having discharged its passengers intending to transfer to the next northbound local.

    Southbound Express Train B transfers from the outer southbound track to the inner southbound track between Sunnyvale and Lawrence. Northbound Express Train C transfers from the outer northbound track to the inner northbound track between San Antonio and California Ave.

    The turnaround procedure complete and passengers boarded, now northbound Local Train A heads northwards from the southbound inner track, and crosses to the northbound inner track between Mountain View and San Antonio stations. It stops at San Antonio on the inner northbound track.

    Local Train A continues northbound from San Antonio, and transfers from the inner northbound track to the outer northbound track between San Antonio and California Ave.

    (cont)

    ReplyDelete
  58. In summary, the solution as proposed doesn't require crossing two tracks at the same time (and thus killing capacity in both directions at the same time). It does require that the first station north of the turnback stations have island platforms in the northerly direction, a point which I wasn't aware of when I proposed the solution initially.

    It also requires a liberal application of 'work smarter not harder', in having the express trains used the normally 'slow' tracks for a brief period, in order to allow time for various blocks to clear etc.

    Regarding your (Richard) other two questions, (2) seems to be a matter of misinterpreting the proposal, while (1), well, your link to Segovia shows a similar concept of having the through trains on the outside while the turnaround trains are on the inside. Better track layout though and with only two tracks each side of the station, no confusion about 'fast' vs 'slow' tracks.

    ReplyDelete
  59. Now that I think about it, short turns aren't necessarily more convenient with FSSF as opposed to SFFS. It depends on what sort of turnback facility you have. If you just want a turnback siding or even just a crossover on directly between the slow lines, then yes, FSSF is the clear winner. But if you want a storage yard as well, then that yard will most likely have to be on the outside of the ROW, which with FSSF means that trains going into or out of the yard will conflict with a fast line unless you build a flyover for the fast line, which only leaves conflicts between yard movements and one slow line. With SFFS, one direction is conflict-free, and building a flyover eliminates all conflicts entirely. Look at Welwyn Garden City for an example of how that could work.

    ReplyDelete
  60. @alon:
    You can't switch between FSSF and SFFS without building extra platforms at every local station

    True, treating FSSF/SFFS as purely a day to day operational shift does assume platforms on every track at every station, space permitting. Island platforms between the outer and inner tracks, with the inner tracks being kept straight and the outer platforms being curved to reduce the station footprint would be my preference.

    That then runs into issues such as maximum number of direction changes per length unit permitted for HSTs, radius of the station curves etc.

    ReplyDelete
  61. Adirondacker1280005 August, 2010 14:21

    You're thinking in the wrong direction. Mountain View has 3,200 weekday boardings and Jersey Avenue 1,600.

    There's more to stations than how many passengers use it. I can't find an authoritative source for the amount of trains they store at Jersey Ave. They store trains there. How many are they going to be able to store at Mountain View? Not that they'd want to store trains in Mountain View. The storage space at Jersey Ave was essentially free, it's an abandoned branch line. So was the station since it's on the abandoned branch line. So storing a few trains there probably makes sense. Has the added bonus of alleviating the storage problems other parts of the system are facing.

    How many locals to San Jose will be able to get to San Jose while they are turning the train on the main tracks? How many Acelas could ( not do, but could ) pass by the train being turned at Jersey Ave? How many NJTransit locals to Trenton?

    Silly human operators need restroom and meal breaks. How many restroom and meal breaks can the operators of train being turned at Mountview take? While their train is blocking the local track?

    Mountain View ain't Jersey Ave. It ain't gonna be either. Not in this lifetime anyway.

    First you need a place to put them. Then you need to justify paying, a lot, for these puppies.

    Richard, haven't you read the above about how the Baby Bullet or however they brand the express service is going to bob and weave back and forth between the local and express tracks every other station? The full interlocking on either side of every itty bitty little Caltain station is going to give a lovely braided effect on the satellite images, I'm sure the foamers will love it.

    ReplyDelete
  62. is sffs any better if you do island platforms at your critical cross-platform transfer stations (i.e. redwood city)

    ReplyDelete
  63. Adirondacker12800 said "maybe they can ask that other world class operator the Long Island Rail Road how they manage to get a train a minute in each direction at Jamaica using 8 platforms. I think there are bypass tracks in Jamaica..."

    Public timetables show 36 trains eastward thru Jamaica between 5 and 6 PM, including 13? that don't stop.

    ReplyDelete
  64. People, you have just to see the SF-SJ Supplemental AA Appendix C - Typical Sections

    The Peninsula Rail Program: America's Finest Transportation Professionals, hard at work on ... something or other, and certainly burning through your money like it's going out of fashion, but definitely not working for you!

    This stuff is beyond belief.

    ReplyDelete
  65. Again, I don't think regular mid point turnback operation between SF and SJ is justified. I understand the loss of operating flexibility, but that marginal benefit doesn't appear to justify the marginal cost and the loss of construction/implementation flexibility.

    Look at the Caltrain ridership count. Not the raw count, but the count split by peak periods. San Jose has the highest northbound AM boarding. PA and MV have higher boarding counts because of reverse commute and regular commute. In the case of PA, more people get off the northbound train there in the morning than getting on, so PA as a destination station isn't just reverse commute.

    It appears that any turnback train will be less productive than those that don't. Morning trains from SF to SJ may carry less riders south of MV, but sure will pick up more northbound riders in SJ than MV.

    ReplyDelete
  66. Adirondacker1280005 August, 2010 20:37

    This stuff is beyond belief.

    There's a few drawings that look reasonable. The four track at grade for instance. Why it has catenary structure designed for ice storms in Quebec is a bit questionable. I like how they get increasing incredible and then finish off with the 13 track extravaganza just off Cahill Street. Do you think they are doing that so it will make Millbrae look crowded?

    ReplyDelete
  67. I especially like the 'Typical' label for the Millbrae prototype, as if that might ever be replicated.

    My guess is one reason they went with Slob Fat Fat Slob was to keep freight off the express tracks. Maintenance starts to get costly above 90 mph and above 110 mph isn't very practical, even with just a pair of freights per day. So will Caltrain limit passenger speeds on the outside track (s) shared by freights to 90 or 110 mph?

    I don't recall the freight clearance envelopes, but are approx 25" platforms low enough to avoid the need for gauntlet tracks? The drawings and AA would suggest the issue was already vetted, but given their past history one can never be sure.

    ReplyDelete
  68. Am I reading this right? Still with the 8" platforms?

    Caltrain just threw ADA under the bus.

    ReplyDelete
  69. @ Drunk Engineer

    "Still with the 8" platforms?"

    Well, given the current stage of the design, I'm guessing they copied and pasted the 8" number. There are a number of "typical" stations shown with Caltrain platforms where the platforms look higher than those labeled with 8" platforms without the platform height being given. Hence, it does not look like they are wedded to any of the design elements just yet.

    But yes, 8" platforms would not work.

    ReplyDelete
  70. Whoa, you're right. The 2' minimum I saw was for HSR platforms, Caltrain is 8". That solves the freight/platform issue.

    Screw the ADA, or rather the inflexible zealots trying to enforce it without common sense. Level-boarding is preferable but there are other reasonable methods to provide good accessibility. Done all over the world, no big deal. If 15" becomes the low-floor standard in new North American equipment then the differences becomes just a single step, so boarding for the masses shouldn't lengthen too much.

    An actual logical compromise. Perhaps this hints that the fed ADA view on platforms is evolving? Or is this just another Caltrain stretch that they hope to get a waiver for?

    ReplyDelete
  71. ADA is a touchy subject. The disabled folks will see lack of level boarding something along the line of "separate but equal." If blacks and gays reject that notion, why would we expect the disabled to feel any different.

    I think a reasonable situation is that people in wheelchair can board the vehicle unassisted (although some may feel they should have access at any door and access the upper level too). Caltrain itself may have more of a flexibility because it is being transitioned from an older standard.

    ReplyDelete
  72. Freight on high speed track is not impractical. After all, freights run on the NEC, including the sections in RI and MA with speeds of 125-150 mph. True, it's not a whole lot of freight, but it's probably about the same amount as on the SF-SJ line overall. And, assuming away FRA regulations for a moment, FSSF would work great for freight: after 9 pm or so, all passenger trains run on the "slow" tracks and freights are free to serve the sidings that are connected to the express tracks. By the way, if you look at the original design concept of the NEC (at least the NYC-Philly part), it's not so much Slow-Fast-Fast-Slow as Passenger-Freight-Freight-Passenger, as evidenced by the layout at the Lane and Morrisville junctions, where freight-only lines branch from the inner pair of tracks.

    ReplyDelete
  73. There's no compromise necessary or needed on level boarding, even if that were legal, which it isn't.

    Even if it weren't a matter of human decency, it's an engineering operational necessity for an optimally functioning train system -- keeping trains moving, not waiting, and avoiding boarding injuries and incidents.

    Moreover, there's no compromise necessary or needed on level boarding for all train types on the corridor all serving the same platforms.

    Anybody who even proposes such a thing should be crucified. And certainly under no circumstance should such a manifestly unqualified, unprofessional, uncreative, unthinking, fiscal fraud perpetuating individual ever be allowed to show his or her face in public ever again.

    This is all very very simple.

    Unfortunately we have America's Finest Transportation Professionals involved, where maximizing cost and screwing the public are always the only criteria.

    ReplyDelete
  74. STA-4A is probably my favorite from Appendix C of the Supplemental Alternatives Analysis Report, which I'll note as pS|PF|FP|Sp. An outside low platform for Caltrain, the slow track, a fence, a high HSR platform, then the HSR track, another fence, then repeat in reverse for the other direction. No cross-platform transfers here (or drainage for the HSR platforms).

    ReplyDelete
  75. This is an interesting conversation. My 2c is that while the merits of FSSF are argued pursuasively here, the Authority's main concern - the width of the ROW - is not sufficiently acknowledged. Given that Caltrain has so many stations and HSR trains will want to be on straight track as much as possible, it seems that FSSF will simply require a wider ROW in much of the corridor. The only way to narrow the ROW between Caltrain stations is to weave all tracks towards the center of the ROW. This is likely not to be feasible for HSR trains. The Authority is designing in a corridor urbanized corridor with many ROW constraints and in a difficult political environment. Their intention to narrow the ROW between stations is born out of necessity in getting the project done. Reducing takings is going to be critical in getting this project completed. Further, consider the impact on cost in trench sections. Maintaining a wider corridor to accomodate SFFS will most certainly increase the costs of trench very significantly.

    ReplyDelete
  76. @anon 409 at 05 August, 2010 21:15
    Both layouts for Millbrae (STA-16A and STA-16B) seem to be missing one of the BART tracks as well.

    Pity there isn't enough room for SpBBpFFpS at Millbrae (with the first 'S' taking over the easternmost BART track).

    ReplyDelete
  77. First off, I agree that the best solution is a single uniform platform and car height with level-boarding for all passenger services, so that HSR and commuter rail can share all platforms and tracks as needed. Plop in gauntlet tracks for the freight tracks, a couple of extra feet of width is no big deal.

    That said, 8" platforms aren't the end of the world. The major equipment manufacturers are attempting to develop 15" commuter trains. The proposed Dallas dmu's (FRA compliant) are an example, though at this point it doesn't mean 15" is assured of becoming standard (I'm going to ignore the 15" giant Amtrak Cal car design as too heavy for widespread adoption compared to better options.)

    The gap between 15" car floors and 8" platforms is a single step, and I hardly think using a ramp with a 7" rise is a return to Jim Crow. Sheesh.

    7" is less than the difference between the dual Euro standards of 550 and 760 mm. Do we thus crucify the Euro's? Have those barbarians no regard for basic human decency? So I guess we have only the Japanese to learn from now?

    Is Zaragoza a failure?

    http://www.gozazaragoza.com/pix/200705/cercanias-deli-0531171642-zoom.jpg

    http://upload.wikimedia.org/wikipedia/commons/f/fa/Zaragoza-Delicias_%28Cercan%C3%ADas_destino_Miraflores%29.jpg

    http://www.flickr.com/photos/7455207@N05/3188086240/#/photos/7455207@N05/3188086240/lightbox/

    http://lh6.ggpht.com/_zakIRZjpe70/SBOYWmKAdvI/AAAAAAAABVQ/R2ItXRdlWsM/s1024/CIMG0255.JPG

    ReplyDelete
  78. BTW, freight trains and cars were not nearly as long and heavy when the NEC was designed. They also had gobs of workers to tinker with the jointed rail of the era. Amtrak does try to restrict NEC freight to outside tracks that in places have lower passenger speed limits. The maintenance costs for shared track do rise significantly with speed, and above 110 mph becomes difficult to justify the costs. 125 mph standard track can see damage after just a single freight train traverses it. Someone might find an outlier example for this or that, but cost effectiveness matters. I can show you mainline switchback operation in N. America, but it would be foolish to adapt such a technique in most situations.

    Hence I suspect that ultimately the local tracks that freight use in this corridor won't exceed 110 mph. And with the local stopping patterns higher speeds aren't needed.

    ReplyDelete
  79. anon 409: Spain (including the Zargonza cercanias) has an extensive
    "installed based" around which to work, as well as European interoperability standards with which to comply on the high speed side.

    The present state of train-platform interface isn't perfect; it's transitional, and a technical compromise.

    What you seem to be deliberately failing to understand is that there is no installed base at Caltrain -- a merely handful of obsolete trains, all of which are due for replacement in the next few years! -- and there is no installed base of HSR.

    We can choose -- and any even marginally intellectually functioning engineer with even the slightest skill would choose -- any train-platform interface we like, with no compromises on level boarding.

    (I happen to believe that ~600mm "low level" level boarding is the optimal solution for both double deck HS trains and for double and single deck regional trains; others here are stuck around the 950+mm of historical precedent.)

    We're starting from scratch, here. Why deliberately fuck up, casting absolutely wretched designs in concrete for 50 years?

    If the morons at Caltrain build new stations at an incompatible height and procure an entire new fleet at an incompatible height, it is game over for decades.

    If you want a local example, look at what the similar morons did in San Francisco, buying high floor streetcars at a time of complete fleet replacement when the installed base on high floor stations was small and manageable and when the entire world was going low floor. Same class of idiot, same wretched public outcome, same doom cast in concrete for 20, 30 or more years.

    What's with the Stockholm Syndrome? Why do people want the wrong thing to happen when better is not just possible but much easier and cheaper? Why does anybody believe a word that the repeatedly proven liars on the agency staff utter?

    Why not even consider the possibility of doing the right thing just for once?

    ReplyDelete
  80. Dan Krause wrote: "the Authority's main concern - the width of the ROW - is not sufficiently acknowledged"

    Just a thought: what if it turned out that they were engaging in strategic misrepresentation? Perhaps stacking the decks the teensiest bit? Not doing any analysis at all as part of an Alternatives Analysis document? Never trading costs against benefits? Lying through their teeth? Being economical with the truth? Exaggerating? Withholding contrary evidence?

    Hard to believe, but rumour has it that this sort of thing has happened in the past.

    Let's all fully "sufficiently acknowledge" that this wasn't touted as THE driving and uncontestable factor determining corridor-wide train operations forever, until yesterday.

    ReplyDelete
  81. Richard says: It's amazing how hard people -- people who have a chance of having a clue and aren't US transportation consultants -- work to misunderstand!

    Richard, for the hundredth time, you have many good ideas, but your insults and hyperbole only serve to alienate many of us who would otherwise agree with you on a lot of points. I would say it's equally amazing how anyone who disagrees with your viewpoint is automatically branded as incompetent, willfully ignorant, or corrupt, if not all three.

    More surprisingly, your solutions are all upside and zero downside, even though almost every other choice in the world involves some type of trade off. FSSF has "zero disadvantages"? You cannot be serious. I can name two disadvantages vis a vis SFFS right off the top of my head: inflexible station placement and greater ROW requirements.

    Station placement:

    The "infill" discussion misses the point. It's about having any flexibility at all with your stations. Want to add a new station? Can't do it. Want to consolidate two low ridership stations into one intermediate station? Can't do it. Want to extend the platforms to accommodate longer single level trains instead of shorter bilevel trains? Can't do it.

    Even if you get all the stations exactly right in 2018, this infrastructure will be around for 100 years. Population and travel patterns will likely change a lot during that time.

    ROW requirements:

    Using Clem's figures, you only need 10 acres of extra ROW for FSSF (0.5 acres times ~20 local stations). But Clem has also calculated that total ROW takings for the entire Peninsula need not exceed 4 acres! So you might still easily double the necessary takings by choosing FSSF.

    Overall, I think FSSF could still be worth doing. But to claim that FSSF has "zero disadvantages" is inexplicable, especially coming from someone as informed as you are.

    ReplyDelete
  82. Thing is, Mike, MLY is correct 90% of the time and the insults are often well earned (not to mention very entertaining.) Thicker skins are good. There is much to criticize in Bay Area transportation planning and execution. For which those of us in states competing with CA for limited federal funds are quite grateful.

    As to:

    The "infill" discussion misses the point. It's about having any flexibility at all with your stations. Want to add a new station? Can't do it. Want to consolidate two low ridership stations into one intermediate station? Can't do it.

    BS, for the most part. Side platforms consume more total ROW width than a single central platform, plus add an extra elevator(s), ramp(s), lights, seating, square feet of concrete, etc. Dual platforms disperse people so that late at night a person may not feel as safe on the lighter used platform, especially with the fence between tracks. Perception can lead to reduced use. A lot of the above are arguments on the margins, but then so are many of the claims against Fat Slob Slob Fat.

    In the simplest version, keep a minimum ROW width of at least 80' and you can add a station almost anywhere. In theory just have the middle tracks on 50' separation and no track relocation is necessary, just plop in a platform. Even that isn't required, stations won't be added that often and when they are you can incur a bit of cost to relocate track and catenary system over a few feet. If you must have a new station in a constrained ROW there is always easements and land acquisition options.

    ReplyDelete
  83. What's really missing from all this "let's turn trains at Mtn View" debate are any serious metrics. Consider Clem's preferred metric. He showed that with FFSS, it took 67% more resources (measured by tph) to provide 1% better service than either SFFS or FSSF. Switching to SFFS was a huge win vs FFSS.

    Now consider modifying Clem's timetable to turn 2 of the 6 tph at Mtn View. For those two trains, this would cut total run times from 100 mins (including 10 min dwell at terminus) to 83 mins. Total resource savings: 6 percent (i.e., 17 percent * 0.33).

    I don't have Clem's raw input spreadsheet, so I can't do the analysis, but it looks like the turnarounds would not drop the service metric by more than 1-2 percent max. So turnarounds would be a clear win by that metric, but nowhere near the order of magnitude of the gain accrued from going to SFFS from FFSS.

    ReplyDelete
  84. @anon409

    Thing is, Mike, MLY is correct 90% of the time and the insults are often well earned (not to mention very entertaining.) Thicker skins are good.

    I agree he's correct 90% of the time, but if you engage in that kind of hyperbole you really need to be correct 100% of the time (which he's clearly not).

    Regarding thick skins: I personally could not care less what Richard calls me or anyone else here. I too find his rants entertaining. But I also care about his own effectiveness at communicating his message, and that is clearly compromised. A lot of people understandably take the attitude of, "Oh boy, there goes crazy Richard again," and so the content of what he's saying doesn't get across. Which is too bad, because he often has good ideas.

    BS, for the most part. Side platforms consume more total ROW width than a single central platform, plus add an extra elevator(s), ramp(s), lights, seating, square feet of concrete, etc.

    Come on. Even Clem, a strong advocate of FSSF, doesn't claim that it takes less space.

    Dual platforms disperse people so that late at night a person may not feel as safe on the lighter used platform, especially with the fence between tracks.

    Of course, but it's not relevant to my points.

    ReplyDelete
  85. Caltrain First06 August, 2010 14:54

    I agree that Richard is correct on many points and is right to point out the piss-poor engineering, lack of professional ethics, and strategic misrepresentation involved in this project, but Richard doesn't always get it right. For instance, ROW width is a genuine strategic concern, and the land required beyond the formal station footprint for any displaced tracks is a flaw with FSSF. More importantly, infill stations are distinct future possibilities on a corridor where stations are currently about two miles apart. Think of all the infill stations BART should be building but can't politically due to the inability of BART service to run expresses. The Caltrain corridor can be designed to accommodate expresses, and that makes future infill stations both possible and desirable with corridor development.

    It would help Richard get his otherwise important messages across if he didn't behave like a tempestuous child that doesn't get exactly his way.

    ReplyDelete
  86. Who and where are all these "other people" who people say/worry are so turned off by Richard's entertaining rants that they cannot grasp the excellent and imminently sensible ideas contained therein?

    I've noticed that it's always some theoretical "other people" people talking about Richard are concerned about. Nobody has yet fessed up to being one of those people who is so "turned off" that they cannot grasp or appreciate the arguments he makes on their merits.

    I love his passion and insights ... and his cynical and passionately expressive style just makes them all the more interesting to read.

    ReplyDelete
  87. Hey, I did the nice thing for more than a decade.

    I've spend about $50k of my own money and many many thousands of hours on various advocacy projects.

    I cooked birthday dinners for board members. I schmoozed. I made political donations to people I hate. I said nice things in public on all occasions about people who were and are dumb as rocks.

    And what we get is the Transbay Terminal, BART to Millbrae, HSR to Los Banos, and the Peninsula Rail Program.

    The people controlling these processes are all of unprofessional, unethical, unintelligent, uninquisitive, xenophobic, lazy, insular and habitually mendacious. Good luck spending years and years maintaining the social fiction that isn't the case: I wish anybody who can make any inroads by being politic and agreeable nothing but the very best.

    I'm only creating a paper trail at this point for "I told you so" purposes; there's no hope of educating the uneducable before the disaster.

    At least this isn't about anything important, like catastrophic atmospheric CO2 concentrations, mass extinctions, or out of control human population.

    ReplyDelete
  88. There are two tracks between Attleboro and Readville. Likewise, for the most part, between Warwick, RI and New Haven (there's a couple of bits of third track around Groton and Guilford, CT). Both tracks have Amtrak trains, they have equal speed restrictions, and both sections include some areas of 125-150 mph running. There is freight service on both sections, so at least one track on each is shared between freight and 125 mph freight trains, and it doesn't seem to be hugely problematic. And what about the WCML and ECML in the UK, both of which have 125 mph running and heavy freight traffic. Or for that matter HS1 and Perpignan-Figueres, both of which are full on high speed lines with 300 km/h TGV service, and freight trains using the same tracks at night.

    ReplyDelete
  89. @Caltrain First:

    If you're running peak 12 TPH, per direction through the crossings (which is a conservative estimate compared to official numbers), they you'll be lucky if the crossing is open for a couple of minutes ever hour. There is functionally little difference between that and closing the crossing altogether during peak hours.

    ReplyDelete
  90. That's 24 activations per hour, which means 12 minutes of gate downtime assuming a 30 second cycle. Probably not so good for a major street, but not disastrous for the likes of Churchill or Sunnyvale Avenue. It would be best, though, to ensure that pedestrians always have a grade-separated path to cross the tracks at stations.

    ReplyDelete
  91. "Joey said...

    neroden: the cost of 48 inch platforms vs say 36 inch platforms is not going to be significant"

    It changes ramp lengths, which is significant. Anyway, we're talking the difference between 48" platforms and 21.6" platforms; the difference in required ramp lengths is substantial (the latter requires only one landing, the former requires three).

    It also changes the suitable train designs, particularly for bilevels, and so it affects the structure gauge height for the entire line. This is much more significant.

    It's not like it's a huge deal either way though. However, doing anything other than standardizing would be horrible.

    ReplyDelete
  92. Regarding turnbacks, turnbacks at Mountain View would be difficult with SFFS. However, I expect San Jose to end up with a flyover/duckunder appearing as a side effect of the new bridge south of the station, regardless of how stupid the bridge design is; I don't see a way to do it without generating an obvious diveunder location for the outside tracks.

    ReplyDelete
  93. High platforms don't cost all that much. In Germany they've built a station for $200,000. I think the boarding height is 550 mm, but even with a factor-of-10 difference in cost between that and an HSR-compliant boarding height, we're talking about $4 million per station, i.e. less than $150 million systemwide. Now compare the cost of this with the cost of procuring low-floor HSR...

    ReplyDelete
  94. What if San Jose became do TOD-of-riffic (circa 2070) that the S-Bahn San Jose justified 7 minute headway local trains shuttling Sunnyvale-SJ-Blossom Hill?

    No need to wait until 2070. It was not too long ago that VTA was talking about paying for additional Caltrain runs, to turn around at Palo Alto.

    Something to keep in mind is that Caltrain is not monolithic entity -- it has no dedicated funding source for operations. Instead, it is up to whims of the individual member counties.

    It is not hard to imagine one county deciding to fund higher level of service than another. In such cases, they are going to want to turn trains around at arbitrary political boundaries.

    ReplyDelete
  95. Adirondacker1280007 August, 2010 17:00

    No need to wait until 2070. It was not too long ago that VTA was talking about paying for additional Caltrain runs, to turn around at Palo Alto.

    Just because San Mateo county runs nearly empty BART trains thither and yon doesn't mean Santa Clara has to run nearly empty trains thither and yon. Anyway once they build BART to San Jose they will be able to run nearly empty BART trains like San Mateo.

    ReplyDelete
  96. Thing is, Mike, MLY is correct 90% of the time and the insults are often well earned (not to mention very entertaining.) Thicker skins are good.

    90% of the time it's just repetitive and bizarre insults. You guys seem to think he's some sort of genius because he knows how to use CAD software.

    ReplyDelete