06 January 2010

The Tao of Timetables

Amid all the hoopla about high-speed rail and quadruple tracking on the peninsula corridor, it's easy to miss the original point: to convey people from point A to point B as quickly and conveniently as possible. The peninsula rail improvements aren't just about pouring concrete and running zoomy trains; they ought to be about providing a service, especially to the dominant users of the corridor, namely local commuters and travelers, people who live and work right here on the peninsula. This important constituency is too often eclipsed by the bright spotlight shining on statewide high-speed rail.

Service is where it all begins, as demonstrated when Caltrain's Baby Bullet started operating in 2004. Because the Bullet increased speed and convenience, ridership soon grew by a third, and Caltrain was featured on the cover of industry journals. But the Baby Bullet was just a baby step compared to what might soon be possible, and Caltrain can't afford to rest on its laurels.

The only three things that will really matter for the future of Caltrain are service, service and service.
(photo by Christophe Dune)

Constructing the Ideal Caltrain Timetable

Timetables are admittedly boring, but they are a compact and efficient way to describe a transit service. Improvements to the service are best understood through their effect on the timetable. Caltrain passengers stand to benefit from several future service improvements that can be sorted into two categories: (a) service improvements that come "for free" as a result of electrification and swift new EMU trains, and perhaps more importantly (b) service improvements that require a little bit of planning, creativity, intelligence, and effort to pull off.

Caltrain is relentlessly focused on (a) and appears distressingly oblivious to (b).

Type (a) service improvements that can basically be taken for granted are:
  1. Quicker start-to-stop times, since EMUs are more powerful and lighter than today's anemic and heavy diesel trains. The time savings from faster acceleration and braking really accrue when a train makes many stops.

  2. Shorter station dwell times, since EMUs will have more doors and level-boarding for rapid passenger loading and unloading.

  3. Faster top speeds, since the peninsula corridor will be grade-separated and use a modern signalling system shared with HSR. While acceleration is far more important than top speed, improving from today's 79 mph speed limit to about 100 mph does yield marginal benefits.
Those are very nice, but they're just the low-hanging fruit. Type (b) service improvements are less obvious, but deliver enormous benefits for passengers. They include:
  1. Clockface scheduling, known in german-speaking countries as Taktverkehr (literally, "traffic to a beat"). In a clockface schedule, every train runs on a periodic timetable with a regular stopping pattern that repeats at some fraction or multiple of an hour (e.g. every 15, 30 or 60 minutes). In Switzerland, the entire country runs on a regular beat. Clockface scheduling eliminates annoying Caltrain questions such as "When is the next train coming?" or "Does it stop where I'm going?" A clockface schedule takes all the guesswork out of taking the train by allowing simple memorization of the timetable (e.g. :07 past the hour, every 15 minutes) Clockface scheduling is also a key enabler of intermodal connections. Connecting feeder services, whether bus, employee shuttle, light rail, HSR, etc. can be synchronized to the "beat" for quick, hassle-free transfers.

  2. Zero-wait transfers between local and express trains, so customers at smaller station stops can benefit from fast express service to the destination of their choice. Zero-wait transfers extend the benefit of express service to a much wider selection of origin and destination pairs. A zero-wait transfer means that an express train overtakes a local at zero speed, on the opposite side of the same platform. Passengers simply walk across the platform to switch to a faster or slower train, as suits their destination. Overtaking at zero-speed enables such transfers in the first place; if one were to wait for the express, the time savings from the express would evaporate before they could provide any benefit.

  3. Two, or maximum three, station stopping patterns, for simplicity. Customers no longer fret whether any given train will stop at their destination; no more agonizing Caltrain questions like "Should I get on this train or the next?" or "Does this train serve the stop where I'm going?," no more squinting at the timetable. Caltrain conductors would no longer make long-winded and repetitive announcements about upcoming stops and which stations will or won't be skipped.
What would an ideal Caltrain timetable look like with all of these improvements? We'll soon find out, but first, a small detour through...

String Theory

To think about timetables without having to stare at big tables of numbers, it helps to use a graphical representation of rail traffic known as a string graph (also known as a stringline diagram). Traffic is displayed as a time-distance graph with time along one axis and stations (distance) along the other. The movement of each train is represented by a string that connects each station according to the time intervals required to move from one station to the next. In effect, a string graph is the visual equivalent of a timetable. For example, the weekday morning Caltrain timetable is shown in the string graph at right. There are many kinds of strings, each corresponding to a different stopping pattern; the irregular pattern makes it difficult to memorize the timetable between any two stations. The red strings represent Baby Bullet express traffic, and the blue strings represent slower trains. When the lines cross (as circled on the diagram), one train overtakes the other. Today this can only happen in Sunnyvale or Brisbane, where Caltrain already has four tracks.

Going back to our list of ideal timetable features, we can construct a string graph that includes them all. In the real world of operations planning, this is done using sophisticated simulation software, and needs to take into account detailed train performance, speed limits, curves, signal block lengths, off-nominal conditions, and knitting together with the HSR timetable (itself a complicated mess; see pages 16-24).

To first order, it can be done with a spreadsheet. Download a Caltrain string graph (385 kB Excel spreadsheet) to play with it.

The resulting Caltrain timetable is shown in the string graph at right, screen-grabbed directly from the spreadsheet. This string graph shows an example timetable, at a relatively sparse 6 trains per hour, with the following features:
  • 15-minute local interval
  • 30-minute express interval
  • Fast acceleration using 90 mph EMUs
  • 45-second platform dwell times
  • Zero-speed cross-platform overtake at Redwood City
It's quite easy to change these assumptions using the spreadsheet to see what happens. At a glance, the timetable certainly looks more symmetrical and orderly, but how does one meaningfully compare old and new?

The Service Improvement Table

The best way to convey an improvement plan in a way that the average person understands is to publish a service improvement table, an example of which is shown at right. (This example table does not show every station pair; for brevity, it shows a sampling of 10 stations.) This table is built by simply comparing the old and new string graphs. For each station pair along the peninsula, the improvement table shows how many more trains per hour connect the station pair, and how much quicker they are than today. This table is constructed by comparing the string graph of today's timetable with the string graph we constructed above.

The table is a quantitative, objective measure of the degree to which service is improved. All capital improvements on the peninsula corridor should be evaluated and prioritized by their effect on the service improvement table.

Marketing 101

Memo to Caltrain: nobody really cares about "growth" or "advanced technology" or "flexible equipment" or "power facilities", or even "less pollution" as trumpeted in the latest electrification update. Those are entirely the wrong selling points. What customers most care about is getting quickly and conveniently from point A to point B.

Damn the pantographs, what you need to sell and deliver is a service plan.

When people ask "what's in it for me", show them the service improvement table for where they live and work. When people worry about the disruption from the peninsula HSR project, they can see that some benefits actually trickle down meaningfully and measurably to each and every city along the Caltrain corridor. Put this table in front of city council members up and down the peninsula. Publish it in local newspapers. Make brochures. Drop leaflets. When Burlingame gets worked up about grade separations, explain to them that service at Broadway will go from none at all to a train every 15 minutes. When Palo Alto frets about community division, show them how California Avenue will see a quadrupling of service, and will be 9 minutes closer to Millbrae.

A Litmus Test

The dirty little secret is that all the service improvements described above could have been realized without HSR, requiring only a few miles of strategically placed track amplification to enable a mid-peninsula overtake, and using capital improvement funds in a targeted fashion that measurably improves service. (Instead, Caltrain is plowing $300 million into the San Bruno grade separation, which provides zero service improvement, while claiming it is fiscally doomed without HSR).

The peninsula HSR project, which continues to be engineered for extremely high levels of service (despite lowered ridership estimates), will consume precious resources that currently belong to Caltrain--in particular, key portions of its corridor. Constructing HSR in a manner that precludes the above schedule, or something similar to it (i.e. clockface schedules with mid-line overtakes) from being reliably operated will forever stunt the future of the peninsula commute. That can rightly be considered totally unacceptable.

The timetabling criteria outlined in this article form an objective litmus test for an important question that should linger in the minds of peninsula residents: MOU notwithstanding, is Caltrain getting screwed over by the HSR project? Armed with this knowledge, you can be the judge.

89 comments:

  1. Let's not forget, though, that that type of timetable would only work with electrification, which is considerably more expensive than just a couple of new tracks in Redwood City. And CalTrain is right: as long as BART has a place in Silicon Valley, CalTrain will get the short end of the stick when it comes to funding.

    ReplyDelete
  2. Two questions, Clem:

    1. Your simulation assumes a top Caltrain speed of 145 km/h. Isn't the top speed planned to be 110 mph, i.e. 176 km/h?

    2. In your simulation, which express stations are configured with two island platforms - all, or just the HSR stations? And can Caltrain express trains generally use the HSR tracks?

    ReplyDelete
  3. @Alon:

    Clem's scheduling does not take HSR into account. The only place train overtakes take place are at Redwood city (with the cross-platform transfer), and Brisbane.

    ReplyDelete
  4. 180kmh (110mph) inter-suburban regional service?
    Not going to happen.. Doesn't happen elsewhere, with lighter, cheaper, quieter equipment; won't happen here.

    200kmh (124mph) Flight Level 0 airline service through developed affluent suburbs and town centres?
    Not going to happen. Doesn't happen elsewhere with lighter, cheaper, quieter equipment; won't happen here.


    ("Oh oh oh oh but they do it in Japan! So you're wrong!")


    Anybody who makes run time claims (they never make service claims, which is part of the point of Clem's post here, just airport-like gate-to-gate times) based on such top speeds is either delusional or, well, lying (for the usual cost-maximizing and political reasons.)


    SF Transbay-SJ Cahill in 30 minutes? Not going to happen (except maybe once, as a political stunt.)

    ReplyDelete
    Replies
    1. I wouldn't say that 180 km/h (110 mph) for a fast regional service or 200 km/h for a high-speed service in a suburban area are unreasonable. 160 km/h (~100 mph) is probably a bit more standard as a top speed for German RE-type trains, however. 200 km/h for high speed trains in the suburbs is definitely reasonable. Check out this Paris Region track map (a pretty amaing resource generally). There are stretches of 200 km/h running through the urban area on the inner ends of TGV lines, but I agree that 160 km/h is a bit more common.

      Delete
  5. "Not going to happen because it doesn't happen elsewhere"

    Right, because California has never been known for innovation, we don't care about transportation and we always look to others and/or the past for what to do.

    I'm sorry people bought $2M bungalows along an active railway and were then SHOCKED to see trains running on it.

    ReplyDelete
  6. "Right, because California has never been known for innovation, we don't care about transportation and we always look to others and/or the past for what to do."

    California would be 50 years ahead of where it is in transportation if didn't "innovate" and simply copied what others do.

    BART still doesn't have a reliable train control 30 years after innovative Californian engineers decided they weren't going to have anything to do with what others did and do.

    Designed in LA, manufactured in Fremont, California-specific HSR trains with CBOSS signals designed in San Carlos running on track designed in Utah on a route rammed through by San Jose? What could possibly go wrong?

    ReplyDelete
  7. In other countries 200 km/h isn't even considered high speed.

    ReplyDelete
  8. This is pretty damning for CAHSR. I wonder if Robert will link to it and refute it.

    ReplyDelete
  9. "Not going to happen."

    The right outcome isn't going to happen either. There will never be true high speed rail or commuter rail done the right way in the United States.

    I'm just going to support the CHSRA and see how bad they screw up. It'll be entertaining at least.

    "Designed in LA, manufactured in Fremont, California-specific HSR trains with CBOSS signals designed in San Carlos running on track designed in Utah on a route rammed through by San Jose? What could possibly go wrong?"

    I hope I die on their trains.

    ReplyDelete
  10. Spokker, you're just going to support CHSRA for kicks, huh? Well, that's interesting. Perhaps you should have listened to the central valley farmers lining up for public comment today. For them, and for many others who's livlihoods are at stake, this is not a f'g joke.

    In the meantime, what did we learn at the meeting today?

    1)Medhi Morshi is out - hard date for retiremnent in March. CHSRA now will spend some unknown amount of $$$ on consultants for exec search to take an unknown amount of time. Interesting note - the board members agreed about not needing to necessarily employee a Californian for this job...

    2)$2B of cost added to the Business Plan for tunneling in So Cal - per Morshed "to be on the safe side". No discussion on how that decision to opt for 'safe side' was reached, or who reached it. Or why the 'safe side' isn't being included in the business plans for other places that might need more costly alternatives implemented in order to minimize environmental and economic damages (like areas of San Jose, like areas of the Peninsula, like areas of the Central Valley, etc etc etc.) Not a single question from a single board member on this preposterous addition of $2B to the plan.

    3) Morshed elaborated on the ticket price/revenue level included in the latest business plan, explaining that they have a revolving door of scenarios that they can pick and choose from, depending on what they're trying to sell at the time. Last time they used very low ticket prices and very high ridership because they were doing a sales job on environemental impacts, (ie; trying to sell net environemental benefit of HSR via big environmental benefits of taking LOTS of cars and plane trips out of the system.) Morshed admits that those ridership projections were probably overstated. But THIS time they were trying to sell HSR to the legislature from an business plan hoping to attract investors perspective, so high ticket prices suited THAT sales presentation. IN other words, they cherry pick assumptions based on the sales presentation of the day. Will legislature be fooled by this dishonest analystic trick tactic? We already know the voters were in 2008.

    4) Not a single board member understands how or when, or by whose authority certain alternatives are eliminated from detailed study. Turns out, its an administrative/engineering decision to stop studying certain alternatives in depth. Engineers consult with Medhi, and they decide to stop studying an alternative. Is the board notified, consulted or presented that info at the time? No. Do local officials or impacted communities have opportunity to argue or influence that decision at that time? No. Do they know when certain alternatives are eliminated? No. Do they necessarily include the economic impact on an area for this decision - (like damage to the business or tax base of that effected area)? Who knows because their reasons aren't presented to the board. Last meeting apparently some "inappropriate" decisions/reasonings were caught by the board for alternatives being elminated in the central valley an the board required them to keep that alternative in study. Morshed says GEE, if you want us to keep bringing to you for review, then the process will take too long. A few board members, demonstrating a tad of conscience, thought maybe perhaps the board had the responsibility to retain the ownership and responsibility to the community for elimination of alternatives. Well, they'll table that minor issue for some other time...

    ReplyDelete
  11. "In other countries 200 km/h isn't even considered high speed."

    In civilized parts of the world 200kmh isn't rammed through the middle of suburbs.

    I don't know how many times this can be repeated, but I'm one of the biggest foamers you'll ever meet. I've ridden trains all over the palce. I've ridden backward and forwards between the same city pair a half dozens times in a day just to gawk at the infrastructure. Multiple times. I've stayed in hotels overlooking train lines and stations. I just love steam trains and Victorian/Edwardian engineering. Choo choos, rule, OK?

    And I still think it is insane -- based on real life experience, and real life precedent, not upon some childish named-calling "NIMBY" or similar crap -- to force aerodynamic bluff bodies through one bar pressure at ground level at high speed anywhere near where people live.

    And I still know it is insane -- based on real life engineering, and the sort of basic analysis that anybody who makes the effort to use primary school arithmetic can make -- to unnecessarily mix different classes of traffic on the same track unless capacity is bountiful or there is no alternative.

    200kmh on corridors shared with regional traffic on a constrained corridor in a nose-bleed-expensive environment in built-out suburbs: it's not going to happen (the 200kmh part, and almost certainly the "shared" part, as we'll see as Caltrain service and Caltrain needs gets reamed to make way for imaginary FL0 airline "demand") and it's going to squander a fortune, to nobody's benefit but the builders and consultants.

    HSR-only over-construction with negative consideration of regional transportation along the same corridor is going to completely screw Caltrain service and peninsula residents, for no good reason.
    Just look at what those HSR/Caltrain context-sensitive half-wits are doing in San Bruno!

    Seeking the semblance of public value for public expenditure? Obviously a train hating NIMBY. All we need to do it get rid of CEQA and all will be well! Suck it up, luddite!

    ReplyDelete
  12. @Richard Mlynarik:

    You seem to think that constructing HSR through suburbs is impossible period. This seems odd if, like you say, you have ridden a variety of rail transport systems around the world. HSR goes to plenty of city centers, and inevitably travels through suburbs to get there. In places like Italy and France, HSR trains even share regional trackage for part of the journey, and seem to have no trouble doing so. I can't really give numbers for what I've observed in HSR traveling through suburbs elsewhere, but I can say pretty decisively that trains still go fast enough. If simulated noise levels are determined to be too high on the peninsula, sound walls are always an option, or as a last resort, slowing down the trains a little (even in this case, the reasoning for using the peninsula corridor remains valid). This leads into my next point - you seem to think that using the CalTrain corridor is a bad idea, and somehow, comparatively expensive. Well let's look at the alternatives:

    -Altamont - basically just moves the problem to the East Bay and you still have to worry about the areas north of Redwood City
    -280: Very curvy; avoids populated areas (read: ridership); no ROW currently available - I wouldn't expect it to be cheap
    -101: Sharp curves; either massive eminent domain or a 50 mile aereal (neither being cheap) required for construction; VERY poor connecting transit options at intermediate stations

    I really don't get how you see any of these as better. I don't get why you're saying that HSR will cripple CalTrain either; unless schedules aren't synchronized (and this is an ENTIRELY different issue), CalTrain will reap the benefits of electrification, grade separations, and a more flexible express service (taking advantage of spare capacity on the express tracks).

    ReplyDelete
  13. " ... I can't really give numbers for what I've observed in HSR traveling through suburbs elsewhere ..."

    I'm convinced! Where do I sign up?

    "-Altamont - basically just moves the problem to the East Bay"

    Yes, who can have failed but notice the heavy conflicting regional rail traffic on the Altamont route ("I can't really give numbers for what I've observed") or have failed to compare the residential development and transportation corridor locations and widths through Dublin, Livermore and Tracy ("I can't really give numbers for what I've observed")? Do tell us more.

    "and you still have to worry about the areas north of Redwood City"

    Really? Thanks for sharing!

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

    ReplyDelete
  15. I can't give numbers because I don't carry around a speedometer every time I ride a train. Do you?

    Altamont may be slightly less developed than the lower peninsula, but you're still going straight through highly developed areas (heard of Fremont or Livermore)? The same speed restrictions (if applicable) and noise levels would exist. Plus all the right-of-way is owned by UPRR which means that you WILL have to bulldoze homes (and plenty of them) in order to build anything.

    ReplyDelete
  16. @Alon: acceleration matters much more than speed. Above 90 mph the returns diminish very quickly as you can clearly see by playing with the spreadsheet.

    @Spokker, this is not necessarily damning of HSR, although they'd better not mess with Caltrain, or else... NIMBY Krakatoa! :-)

    ReplyDelete
  17. "Well, that's interesting. Perhaps you should have listened to the central valley farmers lining up for public comment today. For them, and for many others who's livlihoods are at stake, this is not a f'g joke."

    It is a joke. This is a joke of a state. Everybody should abandon ship instead of trying to change a broken system. I am going to move the first chance I get.

    ReplyDelete
  18. "although they'd better not mess with Caltrain, or else"

    Aren't they already? You have already complained about the wrong type of stations being planned. You have complained about the San Bruno grade separation project. You've talked about curves that are not being straightened. If Caltrain/HSR cannot get these things right, there is little hope that they will get scheduling right either.

    I really think we are at the end of infrastructure construction in the United States. Even freeways can't be built. There are too many stakeholders and they have powerful tools to obstruct and kill projects. Projects that do get built are underwhelming. For example, LA will have a lot of connecting transit to HSR, but it'll be slow and meandering crappy light rail.

    I really do think it's over, and I wanted this stupid train as much as anyone.

    ReplyDelete
  19. OK! Nice work. I sign up to ride it five days a week.

    Now ... how do HSR's NINE high-speed trains per hour during Caltrain peak hours fit in with this?

    (It would be cool if some spreadsheet hacker added a third train type for HSR to Clem's Excel masterpiece.)

    In his scheme 2 or 4 fast Caltrains overtake slow Caltrains each hour, using the "HSR tracks".

    Also, you show that happening twice: once to interchange in Redwood City and *again* at Bayshore. (If you get rid of two of the new extra local stations you have added this doesn't happen.)

    With an average of less than 7 minutes (60/9) between HSR trains, and probably a minimum of 3 minutes between any two trains, keeping the Caltrains and HS trains apart seems like a really hard problem.

    Think of a Caltrain express leaving the Redwood City exchange station.

    It has to start up on the Magic HSR Express Tracks ...
    pull ahead of the local train next to it that is also heading north ...
    pull ahead far enough to safely go back onto Yucky Slow Caltrain Tracks ... some place after the local train San Carlos stop, maybe Belmont ...
    run on the Yucky Slow Caltrain tracks for about 12 miles to San Bruno, where it starts to catch up to the previous Caltrain local train ...
    slide over to the Magic HSR Express Tracks again ...
    blow past the Caltrain local train while that is stopped and w-a-i-t-i-n-g at Bayhore just like today ...
    merge back onto the Yucky Slow Caltrain tracks ...

    ... all while not getting in the way of the many HSR trains stacked up behind it.


    2 or 4 times an hour Caltrain express is going to be clogging up Magic HSR Express Tracks for at least six miles (Atherton-San Carlos) and one stop.
    In your spreadsheet that is 6 minutes each time.

    2 or 4 times an hour Caltrain is going to be clogging up Magic HSR Express Tracks for six+ miles South San Francisco to Oakdale. (Atherton and Oakdale? WTF? WTF????.)
    This is also 6 minutes in your spreadsheet.

    If a HSR super-express is going to overtake the Caltrain express in the 13.9 miles San Carlos to South San Francisco, it has 16 minutes to do it, which is the spreadsheet time for the Caltrain express to go between those points (with one stop in Millbrae.)

    At 125mph (but you've shown it will be slower on some of the bad curves nobody is going to fix) it takes 6 minutes 40 seconds to run 13.9 miles, so there's a window of 16 minutes - 7 minutes = 9 minutes that HSR trains have to hit to get past the Caltrain.

    Real trains seem to run AT LEAST two minutes apart even when they're all slow trains and running at nearly the same speeds, so the window is 5 minutes.

    If you're *really* lucky maybe two different HS trains running 3 minutes apart can squeeze past the Caltrain express, but I bet only one really fits into a 5 minute long window.

    Summary: each time a Caltrain express uses the Magic HSR Express Tracks to stop at Redwood City there is 9 minutes that only one HS can run.


    Next message: Now what could this this mean?
    Is it possible that somebody be misleading us somehow?
    It the Pope Catholic?

    ReplyDelete
  20. So ..........


    If each time a Caltrain express uses the Magic HSR Express Tracks to stop at Redwood City there is 9 minutes in which only one HS can run, what does that mean for Caltrain and HS trains on the "shared" Caltrain and HS tracks on the "shared" Caltrain Corridor from San Jose to San Francisco?



    Well, with just 2 Caltrain express/hour (worse than today), that is 18 minutes of the hour with "only" 2 HS train, so the other 7 HS each hour have to squeeze into 42 minutes: one every six minutes. That's starting to get tight! Maybe in Japan, but I don't think anybody tries that.

    With a nicer 4 Caltrain express/hour (plus 4 Caltrain local that's 8/hour, which is less than the 10 Caltrain says it wants to run), then 36 minutes/hour has 4 HS trains, and the other 5 HS trains have to fit into 24 minutes = more than one HS train every 5 minutes! That's definitely more than I have heard of.

    So it doesn't look like Caltrain will get to "share" Magic HSR Express Tracks after all.
    At least not to me.

    My money is on what predicted way back in June: Caltrain loses expresses, certainly the sort of useful expresses that overtake slower local trains, and HS gets full and exclusive use of 2 or the four tracks that are being built on the Caltrain Right of Way.

    If you look at Spain where they are quadrupling the HSR tracks into Madrid from the south, it doesn't look like other people think that they can fit many more than 9 or 10 HS trains per hour reliably onto one pair of tracks either. Something has to give, so let's take bets on whether it is Caltrain that is the party that takes it in the shorts.


    It makes sense that they are Botching San Bruno and putting the Magic HSR Tracks in the wrong place to be used by Caltrain: Caltrain isn't going to get to use them, in reality, so it doesn't matter.

    Caltrain gets the leftover crumbs.

    Look at the dual (DUAL!) San Francisco terminals (terminal****S***!) to see this again: HSR gets two of the four platforms at Transbay in the CBD, Caltrain gets two -- with a SINGLE TRACK BOTTLENECK -- and most Caltrain terminate out at Fourth and Townsend, right where they are today, but $4 billion dollars on "Caltrain Downtown Extension" money later.

    Caltrain doesn't even get the leftover crumbs.

    So ... in summary ... Suck it up, Bay Area!
    HSR is here to "improve" your service.
    Enjoy the decade of construction with disrupted trains service.
    Enjoy the grade separations, because that's all you're going to get out of this.

    ReplyDelete
  21. Okay, so Spain is quadrupling the tracks. Good for Spain. Japan has two tracks and 13 tph on the Tokaido Shinkansen, and France has two tracks and 10 tph on the LGV Sud-Est. But they're not Germany or Switzerland, so they don't count.

    Richard, I'm glad you've ridden trains that don't run at high speed through affluent suburbs. I've ridden Metro-North multiple times, and it does go at medium speed through suburbs that are every bit as rich as Palo Alto. Further south on the NEC, commuter trains do go at 200 km/h right through the suburbs. Those speeds may be necessary in California if only because it sprawls like the Northeastern US, not like Japan or France or Germany.

    ReplyDelete
  22. Adirondacker1280007 January, 2010 20:57

    In civilized parts of the world 200kmh isn't rammed through the middle of suburbs.

    I realize the suburbs in Maryland and New Jersey grew up around the railroad but trains, commuter trains even, go 110 all the time. The cows don't go dry, the chickens continue to lay and Neiman Marcus still sells overpriced goods to grateful suburbanites. You may not consider it civilized but the good people of Maryland and New Jersey find it pleasant. Suburbs that have have the same kind of demographics as the Navel of the Universe that is the Peninsula.

    Some of the trains, ones that are heavier, more expensive and probably noisier than more rational trains go 135. And while the NEC west of Providence isn't dense suburbia it's not remote farmland either. Where the trains go 150. Past platforms protected with a yellow stripe and a really annoying recorded voice warning pedestrians to stand back. All approved by the scary FRA.

    MARC has locomotives that are capable of 125 and recently had their bilevels rated for 125 between Washington DC and Philadelphia. NJTransit's multilevels are rated for 125 and the prototype ALP46a is on it's way to Pueblo where they will attempt to run it at 135-140 and rate it for 125.

    One of the reasons MARC and NJTransit did that is so they can get some extra revenue on holidays by surprising Amtrak passengers with a cramped commuter car. That and Amtrak has told them that if they want to put their expresses on track 2 and 3 they will have to keep up.

    Back before there was Shirley Time the MUs would toy with 110 in the Meadows - blow past Kearny at 108-109 most of the time. I'm assuming the speedometer in cab cars that were in the middle of the consist were accurate. The engineer on the 5:07 express to Trenton had a bit of lead foot. ... well he liked getting to notch 8 fast.

    Wander railroad.net's forums sometime. The MARC foamers like to whip out their GPS and see how fast the express Penn Line is going. It's not 79 MPH.

    ReplyDelete
  23. so whats the deal with the 'mystery fax' that was spoken about cyptically in hushed tones at today's board meeting? something about TY Lin ownership, something not disclosed properly to the board when the decision was being made? Since board member handed someone a copy of that note in the meeting (but didn't disclose what the note said), does that piece of paper become public record?

    ReplyDelete
  24. I'm trying to understand the diagram. The runtime numbers look very much like today's steam-train schedule, not some futuristic world-class engineering?

    45-90s dwell time?

    20s station padding per stop?

    90 minutes total runtime for all-stop local? Even BART with its 1960s technology does better than that (out in suburban areas with similar station spacings).

    ReplyDelete
  25. @Drunk Engineer, the aim of the exercise was to show how unremarkable performance assumptions could still result in remarkable service improvements.

    Fire up the spreadsheet and cut down the dwell time and padding, and remove a couple of stops to see what this baby can really do.

    I know it's tempting to plug in a higher value of acceleration, but it doesn't really work out that way. With a 4000 kW, 296 tonne EMU, life is mostly power-limited, not adhesion-limited... the train accelerates out of the adhesion-limited regime in just 20 seconds.

    ReplyDelete
  26. "I know it's tempting to plug in a higher value of acceleration, but it doesn't really work out that way. With a 4000 kW, 296 tonne EMU, life is mostly power-limited, not adhesion-limited... the train accelerates out of the adhesion-limited regime in just 20 seconds."

    It depends on whether you decide that You're Incredibly Special and Have Unique Needs -- which include over-weight, over-sized, under-performing trains running too slowly and too expensively on too much over-built and badly-configured infrastructure -- or whether you copy what proven smart people have done and buy the right tool for the job.

    Here's 0-90mph in 57 seconds -- >0.7ms^-2 average -- on a train you can go and buy off an assembly line. Over 540 of these babies have been ordered since the end of 2002, because they're a world class engineering solution to common problems. They're selling like hotcakes.

    Or you could employ the same egregiously unprofessiona rent-seeking types who told SMART they need unique, LTK-designed FRA DMUs for transparently self-serving and bogus reasons to tell you that you need Caltrain unique double deck EMUs for completely self-serving and bogus reasons.

    Hmmm.... go with proven success and copy the operations and procurement of world-leading experts, or do our Own Thing because We're Special? Decisions ... decisions ...

    ReplyDelete
  27. "Even BART with its 1960s technology does better than that (out in suburban areas with similar station spacings)."

    BART's trains have good power to weight and power to passenger ratios. Low weight is about the most important thing for performance and for life-cycle and energy costs, and BART's cars (most of which have solid enough 1990s era traction systems) are nothing to sneeze at in the weight department, at least.

    BART's top speed of 80mph in the exurbs (where a subway technology third-rail gold plated system like BART would of course under no circumstances be constructed by anybody who wasn't motivated by private profit) isn't anything to sneeze at either, the 120mph foamer club notwithstanding. That's at the low end of what seems desirable, but it isn't an outright joke nor is it 50 years our of date.

    The reality of the world is that pushing air out of the way is expensive, noisy and energy-intensive. There are trade-offs. "Run trains as fast as necessary, not as fast as possible" is one way to look at things.

    ReplyDelete
  28. Adirondacker1280008 January, 2010 12:03

    BART's top speed of 80mph in the exurbs ...isn't anything to sneeze at either

    It's something to go "meh" over. We'd have to find an New York Central or LIRR foamer to tell us when the first MUs topped 100. The first M1s were delivered in 1968. The M1/M2/M3s are rated at 100. Metro North's track in NY is 90 MPH track. I assume the expresses go 90 in NY.

    it isn't an outright joke nor is it 50 years our of date.

    If you are planning a system that stretches out into the suburbs with 5 miles between stations it was questionable in 1900 and definitely out of date by 1915.

    ReplyDelete
  29. meanwhile Jimin SF on Roberts Rant, posted some interesting news about attorney general stepping into the TBT discussion, and heavier trains (not lighter) FRA ruling coming out tomorrow.

    Wa, wa, wa wa... (with sitcom tune in the background)

    ReplyDelete
  30. Nice work.

    I always think that current diesel operating still can handle 30 minutes headway 6 train/h, with simplified time table.

    2 train/h Local train only SF-Redwood City only.
    2 train/h SF - (Express)Redwood city - (Local)- San Jose.
    2 train/h SF - SJ Baby bullet.

    Above proposal, there will 6 train north of RWC but 4 train south of RWC. So, can be operate almost same operating budjet.



    I always see express-local train which stops all station RWC-SJ are highly utilized. On the other hand, express-local which stops all station SF-RWC are much less utilized.

    Stations like California Avenue and San Antonio has benefited from this train because of express service between RWC-SF.

    Baby bullet need to adjust stop pattern. Not a current two different express train run each 1 train/h. All train should stop at Sunnyvale(Traditional direction only), Mountain
    View, Palo Alto and Hilsdale. This will add 2~3 minutes of total run time but every 30 minutes should be better than current 10~50 minutes frequency.

    Local train should use 2 or 3 car consist to archieve faster accerelation. As combined with mini-high platform as temporary solution, local train can be operate with one conductor.

    Build additional SB truck at Milbrae for cross transfer between local to express. For NB, it takes some walk, but 4 th platform can be use for local train by passed by express.

    How do you think this idea?

    ReplyDelete
  31. Governors proposed 10-11 budget has CHSRA line item at 50m, down from 139M in 09-10.

    ?

    http://www.capitolweekly.net/fs/global:file/article/jq2q6aq0qasen6_files/file/id/yjkvus72k4l3lu?_c=yjkwrido7758eg&_c=d|x0rye2v8gr4ci1|yjkxcrmbkutbbe&_ce=1262980900.74db87c3169a7ff78ecaa11dbb0cdadb

    ReplyDelete
  32. but it does have $375M in federal funding for capital outlay, and 533 for state bond capital outlay.

    Can they build a usable segment for 900M?

    ReplyDelete
  33. Richard, stop condescending. We all know that train speed should be as fast as necessary. I just contend that the distances of Caltrain are much greater than those of commuter railroads in Europe, requiring higher speeds to compensate. SF to SJ is 80 kilometers. The RER rarely goes more than 50 kilometers out of Paris. This isn't due to American rail operations quirks; it's due to American sprawl. This means that the Northeastern US railroads, which despite your mocking do a good job of getting people to the CBD (and nowhere else), are an instructive example.

    ReplyDelete
  34. "Run trains as fast as necessary, not as fast as possible" isn't my slogan. I don't invent things from outside my field of expertise -- I just parrot those with the proven records of accomplishment.

    I'll admit am deeply concerned, as any remotely rational or ethical person being will be, about the energy intensity of transportation and other human activities; but in this case, as many others, that accords quite well with an interest in cost-effective and economically competitive transportation network planning and operation.

    And believe me, I know quite a lot, in very intimate and repetitive detail, about the distance from San Francisco to San José, Capital of Silicon Valley, thank you very much.

    Have a enjoyable weekend.

    ReplyDelete
  35. The length of the SF-SJ Caltrain line is not due to sprawl. Passenger operations have been continuous on this line since the 1860s, and it was originally intended to position San Francisco (as opposed to Oakland and the East Bay) on the transcontiental railroad. San Jose was also the capitol of California at the time.

    ReplyDelete
  36. It's not the length of the Caltrain line so much as its status as a single commuter district that's due to sprawl. For that to work you really need to have the 60 miles of unbroken suburb from SF to the south edge of San Jose, and that's some pretty epic sprawl by world standards.

    As for timetables, I'd say that in the medium term, 4 tph of local service would be overkill, but 2 tph would be about right. Then you'd have 4 tph of express service, with two stopping patterns, and probably ones that involve crossing over between the local and express tracks to fit in the with presumed 4 tph of HSR service. By the way, in the UK, the new timetable has HS1 shared by 8 tph of peak hour commuter service (running at 125 mph) and 3 tph of Eurostar (running at 186 mph).

    ReplyDelete
  37. Richard, the complaint about your comments here has always been that you equate "proven records of accomplishment" with "German-speaking." Caltrain isn't like the S-Bahn in terms of distance.

    By the way: the way commuter rail works in Japan or France is that short distances are done on 130 km/h trains, long distances are done on TGV or Shinkansen. SF-SJ is somewhere in the middle in terms of distance - the Shinkansen commuter towns are about 100 km out of Tokyo, the TGV commuter towns 150-300 km out of Paris. The range in all cases seems to be about an hour, which is what Caltrain is shooting for for the limited-stop trains.

    ReplyDelete
  38. The complication (or simplification?) of all of this is that the Peninsula looks like the suburbs part but is in fact where most of the jobs are.

    I want to go back to the assumption that Caltrain will be on low platform. If they actually were compatible with HSR, which problems would go away?

    I ask this because we are getting more and more confirmation (I am unsure why no one is not just asking Caltrain this directly) that Caltrain thinks its customers will be fine with abandoning the baby bullet service in lieu of a fast all stops train.

    ReplyDelete
  39. The range in all cases seems to be about an hour, which is what Caltrain is shooting for for the limited-stop trains.

    Shooting for? Caltrain's limited-stop trains (the Baby Bullets) get between SF-SJ in less than an hour already, and that includes generous schedule-padding.

    Platform compatibility between Caltrain and HSR trains should be an obvious goal.

    ReplyDelete
  40. The Baby Bullet makes too few stops, and doesn't always make the same stops. The purpose of higher top speed and better acceleration is to allow trains making predictable express stops to do SF-SJ in less than an hour.

    The purpose of the Baby Bullet isn't to skip as many stops as possible. It's to allow faster trips between the major destinations on the corridor.

    ReplyDelete
  41. "> the complaint about your comments here has always been that you equate "proven records of accomplishment" with "German-speaking.""

    Hello Alon,

    I don't know if you are French or not but I am. Maybe Mr Tillier (TGCWeb!) is also.

    If you live in France, you know that the train is a catastrophe outside Paris and on the TGV routes. And don't even talk about freight !

    If this changes, it changes very slowly, and it only changes because the regions develop RER systems. Those RER? They are copied straight from Germany and Switzerland, just like the new tramways of France are copied. It is no bad thing to copy successes.

    There is talk now in France of making all trains run like in Germany or Switzerland, on a cadence the same every hour. Nobody says this is a bad idea ! Where it already runs people like it. The problem is how we get it more places, not go back to the bad ways.

    Maybe you are not from France and maybe you are a visitor. Then maybe you do not know how much politics is in the TGV, and how much other transports and every type of train that is not the TGV (or Paris RER or Paris metro) is harmed by the great project of TGV TGV TGV and little else. German rail looks veyr good if you are in France and not in Paris.

    In short, more "German speaking" (if it makes it easier, think of it as from the Netherlands or from Switzerland, where they speak French), and less only TGV, please!

    "> Caltrain isn't like the S-Bahn in terms of distance."

    I'm sure there are longer S-Bahn lines around Frankfurt, Rhein-Ruhr and definitely Branderburg (Berlin area.) If not, then very close. Maybe St Gallen-Geneve in Switzerland :-)

    Sure Caltrain is not very like S-Bahn or RER. That is a problem. It should be.

    And maybe you can think of Caltrain like RER San Francisco and RER Palo Alto, with the two ends joined up? Rhin-Ruhr in Germnay looks a lot like that. Switzerland looks a lot like that. Netherlands looks a lot like that. Belgium soon looks a lot like that. And maybe France also will look like that. I hope!

    ReplyDelete
  42. I'm not French, but I visit a lot. What I'm familiar with in France is at the two ends - the PACA TER, and the RER. The PACA TER is actually decent. I know PACA is better off than most of the other provinces, but still, it's not Ile-de-France.

    Even in a 100% French-style system (as opposed to Japanese, or German), the areas in California that would get shafted are the Central Valley towns, rather than the Bay Area. And if anything CAHSR's ideas of how to run a railroad bias it toward the provinces - just look at the unrealistically high number of trains it projects to stop in Fresno and Bakersfield.

    The Rhine-Ruhr S-Bahn is more polycentric and has longer lines than the more monocentric systems in Berlin and Munich (I don't think Berlin has lines that go 80 km out of city center, but I'm not sure - 80 km suburb to city to suburb may happen, though). But I'm not sure it's the best model for Caltrain. The Bay Area is polycentric, but it's not that polycentric. And the Baby Bullet's success shows that there's a lot of benefit to cutting travel times under an hour.

    ReplyDelete
  43. Mr Mlynarik says "BART's top speed of 80mph in the exurbs ...isn't anything to sneeze at either"

    Adirondacker12800 responds "It's something to go "meh" over. We'd have to find an New York Central or LIRR foamer to tell us when the first MUs topped 100. The first M1s were delivered in 1968. The M1/M2/M3s are rated at 100. Metro North's track in NY is 90 MPH track. I assume the expresses go 90 in NY."

    Current schedules

    LIRR NY-Ronkonkoma 50.3 miles 11 stops 82 minutes
    NJT NY-Trenton 58.1 mi 13 stops 89-95 min (tho the 10:31 train takes 101 min)
    BART Pittsburg-SFO 51.4 mi 23 stops 83 min

    MetroNorth NY to Croton-Harmon 32.2 mi 19 stops 68 min
    LIRR NY-Babylon 38.9 mi 14 stops 74 min
    BART Richmond-Fremont 36.0 mi 16 stops 61 min

    The express schedules don't pin you in your seat either:

    LIRR NY-Wyandanch 36.6 mi nonstop 55 min start-to-start
    NY to Croton-Harmon 32.2 mi one-stop 43 min start-to-start
    NY to Trenton 58.1 mi three-stop 69-72 min

    ReplyDelete
  44. "Caltrain's limited-stop trains (the Baby Bullets) get between SF-SJ in less than an hour already, and that includes generous schedule-padding."

    Not that generous. I'd guess their on-time percentage into SF is around 50%, if you define on-time as on time.

    ReplyDelete
  45. Caltrain is operated by Amtrak staff, so I never said they were an efficient outfit.

    The "schedule-padding" suggests that the Baby Bullet could do its run in about 50 minutes with an efficient operation. Sometimes the Baby Bullet is early, sometimes late, but the potential is there for a very efficient service.

    The Baby Bullet has about the right number of stops, and in a beefed up Peninsula Commute Service, local Caltrain EMUs would collect riders for the Baby Bullet express stops.

    BART's top speed is limited by its third-rail technology. I think the light BART trains could fly with an OCS power system. The impediment to BART's service speed is its all-local routes. The average speed is about 33mph.

    ReplyDelete
  46. "The "schedule-padding" suggests that the Baby Bullet could do its run in about 50 minutes with an efficient operation."

    With the same engines, same trains, same 79 mph maximum and same stops? Tell us how.

    ReplyDelete
  47. Ride the Baby Bullet sometime, and when things are going smoothly, you will be pleasantly surprised.

    It truly isn't rocket science. The diesels aren't as fast in acceleration as EMUs, but with limited stops, the diesels don't have to start from stop very often. The diesels are more than capable of 79mph.

    It makes me wonder how fast the old passenger express trains covered the route.

    ReplyDelete
  48. Adirondacker1280011 January, 2010 22:59

    It makes me wonder how fast the old passenger express trains covered the route.

    I have a copy of the Official Guide of the Railroad, March 1956. Northbound Coast Daylight left San Jose at 5:05 and arrived at Third Street in San Francisco at 6:00. Southbound it left San Francisco at 8:15 and arrived in San Jose at 9:06. Local trains took an hour and 20 or 25 minutes most of the day. The rush hour expresses did it in an hour and seven or an hour and nine depending on stops.

    ReplyDelete
  49. The Baby Bullet has about the right number of stops

    No, it doesn't. It's basically skip-stop, with every other Baby Bullet serving half the express stops. If you want to get from San Mateo to Mountain View, you're out of luck.

    ReplyDelete
  50. When the Baby Bullet passing track project was being implemented, the official expectation was that the Baby Bullet train could do the run between SF-SJ in 45 minutes. So that is indeed some service padding!

    Given the slower acceleration of diesels, you don't want to give the current Baby Bullet service too many stops. Caltrain is still a commuter railroad serving rather long commutes. Reverse commutes are increasingly common, but San Francisco is still the main destination by far. Average trip lengths are quite long both ways. The passenger rides on Caltrain are still primarily between Silicon Valley (Palo Alto, Mountain View) and San Francisco, and the local service on Caltrain is not particularly popular. Due to weak local transit connections at stations and limited off-peak service, Caltrain does not yet behave like a proper rapid transit system. Speeding up the Baby Bullets by limiting stops on the long runs between Silicon Valley and San Francisco is the reason for it success. Slowing the Baby Bullet service down for more local service defeats its whole appeal.

    Alon: The purpose of the Baby Bullet isn't to skip as many stops as possible. It's to allow faster trips between the major destinations on the corridor.

    No, no, the Baby Bullet allows faster trips between the major destinations on the corridor (SF, Palo Alto, Mountain View, SJ, and Millbrae) precisely because it skips the stops in between the major destinations.

    http://www.caltrain.com/pdf/annual_ridership_counts/2008_Caltrain_Ridership_Counts.pdf

    The reason for the skip-stops is because Caltrain service frequency is still quite low, although it has greatly improved over the years. Caltrain still has to make Baby Bullets behave like collecting locals while still keeping the stops limited, hence skip-stops at certain mid-line stations. Once overall service is increased with both more locals and Baby Bullets, the Baby Bullet service can settle on 5-6 fixed stops between SF and SJ. These stops should be Mountain View, Palo Alto, Redwood City, Millbrae, and 22nd St (which will continue to grow dramatically with the huge Mission Bay development). Maybe, just maybe Hillsdale and Sunnyvale could be added, but their cases are marginal. Ridership at San Mateo hasn't done that well with Baby Bullet service, so it doesn't deserve an express stop.

    Caltrain's all-local service is its least popular service. EMUs would help to speed up the local service, but Caltrain still needs expresses that skip stops to reach the major destinations as fast as possible for its main clientele. Better local feeder service would enhance the expresses/limiteds, but Caltrain should never seek to end up with BART's all-local pattern.

    ReplyDelete
  51. "It truly isn't rocket science."

    Quite right-- it's easy to learn that a stop in 79-mph territory costs a Caltrain 1.8-1.9 minutes, plus the time spent standing still. It's easy to learn that SF to San Jose is 46.7 miles, and it's easy to get a good-enough idea of the allowed speeds. Pencil and paper will show that a five-stop 57-minute schedule includes a reasonable allowance for delay-- maybe three minutes?

    "the official expectation was that the Baby Bullet train could do the run between SF-SJ in 45 minutes."

    45 minutes would be a reasonable schedule for a nonstop; maybe that's what the official was referring to? There was never any chance of a five-stop 45-minute schedule, of course.

    ReplyDelete
  52. When Caltrain planed additional baby bullet in 2005, thier first proposal was 57 minutes, 6 stops. (SF, Milbrae, San Mateo, RWC, PA and Sunnyvale). This proposed time no longer can be found from web site but was listed in new archive.

    I believe dwell time is too long. If dedicate use for Bombardia car on popular train, they can shortern dwell time as combine with mini-hihgh platform.


    After this 2005 schedule implementation, Sunnyvale jumps twice ridership from 900's to 1800s in 2009. Big parking lot of Sunnyvale station used to be half empty but became full. At the same time, Mountain View still remain same level of ridership.
    On the other hand, San Mateo did not observed such ridership increase.

    This indicates that express service between Santa Clara county to SF still have a lot of potential opportunities.

    Currently, Santa Clara station is under served. In the peak hours, there are two trains. However, both train are subjected to by-passed by next Baby Bullet. This takes 3 more minutes at leadt. Add Santa Clara stop into Original Baby Bullet (A) will create huge additional rider with 2 minutes increases in overall run time, still 59min.

    ReplyDelete
  53. @DrunkEngineer... one more thing I forgot to mention: notice the run time "penalty" for going beyond Mission Bay (4th & King) into Transbay is on the order of 5 minutes, because of the agonizingly slow curves.

    Thanks to TimZ for pointing out an "extra" 0.6 miles that shouldn't have been there between Lawrence and Santa Clara. Spreadsheet was fixed.

    (Note to "advanced" spreadsheet users... you can now enter dwell times by entering a number larger than 1, which will be interpreted as seconds of dwell. This undocumented feature allows you to adjust the dwell on a stop-by-stop basis, without mucking up the simplicity for regular users.)

    ReplyDelete
  54. Let's do that back-of-the-envelope calculation of travel times between SF and SJ. A Caltrain traveling non-stop at 79mph could cover the distance of 46.7 miles in less than 36 minutes. Assuming a time penalty of each stop of 2 minutes to stop and start + 36 seconds of dwell time (assuming efficient operations), a total of 5 stops will add 13 minutes to the 36-minute non-stop run. This arrives at 48 minutes. The Peninsula route is mostly straight, but the few curves will add some extra time, but 50 minutes is very much possible and has been achieved during smooth running.

    Caltrain was promoting 45 minutes during the start of the Baby Bullet project. The Baby Bullet was never conceived as a non-stop service between SF-SJ, so it is safe to assume some stops are included. I don't have time to look for original documents, but this is what was being fed to the press by Caltrain:
    http://www.highbeam.com/doc/1G1-63518368.html
    http://www.planetizen.com/node/5531

    Now 45 minutes was overly optimistic, but it is easy to see how Caltrain could reach that travel time between SF-SJ with relatively simple and affordable improvements. Electrification would speed up the starting and stopping at stations. New, lighter rolling stock would help with acceleration, and efficient boarding access would reduce dwell times. Additional passing tracks can further increase capacity. The mind-boggling expensive grade separations will do relatively little to speed up average train speeds on the Peninsula, but the contractors are drooling over the $$$ in design and construction profit.

    The dirty little secret is that the improvements necessary to achieve fast rail service (both Caltrain and HSR) on the Peninsula don't actually require that much investment, can be efficiently achieved in a timely manner, and won't tear up the whole community in the process.

    ReplyDelete
  55. The above time calculation actually works out to 49 minutes, but with shorter dwell times and improved acceleration, the travel time would drop quickly.

    As for the run time "penalty" of 5 minutes to get from 4th & King into the Transbay Terminal in downtown SF, what's the time "penalty" to take the painfully slow MUNI services? What's the time "penalty" to walk from 4th & King to where the really big buildings are? What's the time "penalty" of delay that drivers experience when they drive into or out of downtown SF during rush hour? Why on earth would people willingly suffer such significant time penalties? Oh yeah, because it is DOWNTOWN. Go where the action is... it's worth it.

    ReplyDelete
  56. > Additional passing tracks can further increase capacity.

    I harbor growing suspicions that Caltrain will never, ever have additional passing tracks. It may even lose those that already exist.

    ReplyDelete
  57. Adirondacker1280015 January, 2010 16:54

    I harbor growing suspicions that Caltrain will never, ever have additional passing tracks.

    Why would they need them when there will be four tracks between San Jose and San Francisco?

    It may even lose those that already exist.

    But they will be connecting them all together into express tracks. Or connecting them all together into local tracks. Run the local trains on the local tracks and the express trains can pass the local trains on the express tracks. Something four track railroads do all over the world.

    ReplyDelete
  58. @Adirondacker12800
    Clem's rightfully worried that Caltrain won't be able to use the HSR express tracks, due to differing platform heights, FRA regulations and general agency disinterest. This would of course be a sin, given that Caltrain's biggest success and innovation, basically ever, was the baby bullets. Caltrain got enormous (well, compared to what they had before) political capital by showing how they could beat car travel times with a little investment.

    ReplyDelete
  59. Adirondacker1280015 January, 2010 18:29

    Clem's rightfully worried that Caltrain won't be able to use the HSR express tracks

    Caltrain will be buying all new equipment. Why would they buy anything other than something that is compatible with HSR? Why would the State of California or the Federal government even entertain the idea?

    ReplyDelete
  60. > Run the local trains on the local tracks and the express trains can pass the local trains on the express tracks.

    That would be innovative, wouldn't it? I was told point blank that the need for a mid-line overtake capability was an "assumption" that I was making.... this was in the context of reviewing some currently emerging corridor configurations, which are FFSS and SSFF.

    We Californians like to innovate; this is Silicon Valley, dammit.

    ReplyDelete
  61. FFSS and SSFF still allow wrong-way overtakes, as with the current Baby Bullet.

    ReplyDelete
  62. wrong-way overtakes are capacity killers and timetable destroyers, especially so in high-speed traffic.

    The Baby Bullet only performs same-way overtakes, in areas with four tracks.

    ReplyDelete
  63. Adirondacker1280015 January, 2010 22:47

    FFSS and SSFF sorta depend on whether you are viewing it from the north or from the south. You can still put the express trains on the express tracks. It doesn't matter if the track order is SFFS or FSSF or SSFF or SFSF. If they put the HSR ROW over 101 the Caltrain express trains could still use the HSR tracks. It would be far less useful but it doesn't mean the Caltrain express couldn't overtake the Caltrain local using the express tracks.

    FFSS kills same direction cross platform transfers. What was the reasoning behind slow traffic to the west ( or east ) and fast traffic to the east ( or west )?

    We Californians like to innovate

    It wouldn't be innovative, it would just be unusual. LIRR does it. I'm sure there's other places where it happens too.

    FFSS and SSFF still allow wrong-way overtakes, as with the current Baby Bullet.

    In normal operation they wouldn't need to, assuming the express trains are on the express tracks. Switching eats up a lot of time or capacity or both.

    ReplyDelete
  64. The incorrect "assumption" that Clem appears to have made is that CHSRA and Caltrain did not over-scrape the very very bottom of the global barrel in order to find the very least qualified and very least conmpetent engineers in the entire world.

    You've not seen anything yet.

    Imagine the the stupidest, most expensive, least workable, most obsolete, most rankly idiotic looney tunes scam possible .. and they'll have exceed your most raving hallucinations.

    Three years ago you might have through that running all trains via Los Banos was outlandishly insane. Two years ago you might have thoought two different train terminals in San Francisco was the sort of thing only a profoundly mentally retarded child with a wooden train set would come up with. A year ago you would have laughed at anybody who said they'd have stations with multiple platform heights. Six months ago you'd reject out of hand the idea that Caltrain could propose with a straight face that it develop in-house a safety critical signal system to utterly unique, idiosyncratic, and technically nonsensical specifications.

    They really are that bad — and it's pretty clear there's no stopping them, because the people who hire them are that bad also. There are no check or balances or reviews anywhere in the system, just checks ... with blanks where the budgets should be filled in.

    HNTB doesn't get a lot of rail corridor work in Japan or Denmark, nobody at Caltrain would be remotely qualified to even land an interview for a job in Hungary or South Korea, and one look at the track record of CHSRA's consultant team would be enough to exclude them from any work in Spain or Sweden.

    We're dealing with sone of the least intelligent, least technically proficient, most dishonest, most hubristic, worst-judgement people in the world here. If you doubt this, look around at transportation system of the Bay Area, California and the rest of the country. These fuck-ups are directly responsible, and they're not going to stop fucking up due to an extra $50 billion lining their pockets.

    ReplyDelete
  65. "50 minutes [SF-SJ with five stops]is very much possible and has been achieved during smooth running."

    I'm not a regular rider, but my guess is the five-stop expresses are not better than even money to complete their runs in the scheduled 59 minutes. (I mistakenly said 57 minutes, above; that's the four-stop schedule.)

    It's puzzling-- why does Caltrain insist on running "inefficiently", when they could make the run in a smooth 50 minutes? Probably the conductor holds the train while he strolls over to the donut shop, and the engineer ignores the highball while he finishes his crossword, and the dispatcher's dozing so the signals are red...

    "The dirty little secret is that the improvements necessary to achieve fast rail service (both Caltrain and HSR) on the Peninsula don't actually require that much investment, can be efficiently achieved in a timely manner, and won't tear up the whole community in the process."

    Happy news! I can see why it's a secret-- GM and the oil barons are suppressing it-- but why is it a dirty secret?

    ReplyDelete
  66. FFSS or SSFF makes perfect sense. You just need to think about it from standpoint of a bureaucrat.

    Draw a line in the ballast, Caltrain has responsibility for everything to the left, CHSRA for everything to the right. Give them separate terminals, ticketing, platforms, and voila! World-class shared inter-modal infrastructure -- with no doubts about who is responsible for what.

    Just imagine, how much simpler this make life for the bureaucrat. No finger pointing, no complexity with keeping to a designated train slot, no worries over how to share revenues. Life couldn't be simpler (well, not for the users...).

    ReplyDelete
  67. Bureaucrats shouldn't be making technical decisions. Come to think of it, they probably shouldn't be making financial decisions either. Actually, maybe they shouldn't be making any decisions at all...

    ReplyDelete
  68. SSFF is the only proposal on the table in the LA Basin, where there's no plan to bring Metrolink into the 21st century as there is for Caltrain.

    ReplyDelete
  69. Timz, the bullets have longer dwell times because so many people are entering and exiting at the stations along with the cyclists and occasional wheelchair user. Since Caltrain doesn't have level boarding, this adds to the dwell time, and if a bullet is stuck with the older, high-floor, single door gallery cars the station stops are even longer.

    ReplyDelete
  70. Schedule and service featherbedding form Amtrak staff??? Cozy railroad union inefficiencies??? Never! Simply unheard of... ;-)

    "The dirty little secret is that the improvements necessary to achieve fast rail service (both Caltrain and HSR) on the Peninsula don't actually require that much investment, can be efficiently achieved in a timely manner, and won't tear up the whole community in the process."

    Happy news! I can see why it's a secret-- GM and the oil barons are suppressing it-- but why is it a dirty secret?


    It's a dirty secret because the contractor-led CHSRA doesn't want the public to know that fast train service can be implemented quite affordably and quickly on the Peninsula without deeply disruptive, decade-long construction scars along the entire corridor. The contractors stand to make many hundreds of $millions in profit by building elaborate four-track grade separations and other overbuilt structures at total public expense. This is why CHSRA seeks to keep the inflated pre-2009 business plan ridership and traffic numbers for the Peninsula EIR: the hypothetical but highly unrealistic prospect of super-frequent trains will require total grade separation of the corridor. The far more realistic scenario -- yet still very optimistic, however -- is that the Peninsula will never have the travel demand for more than 4 HSR tph and 8 Caltrain tph in each direction at the absolute maximum peak hour (not daily average, just the daily peak), and with a more manageable traffic load, grade separations are really only a luxury and can be implemented gradually as necessary. Terminal stations don't need to be so big. Off-peak traffic will be considerably less too, and 2 HSR tph and 4 Caltrain tph will be the typical peak until at least 2030. It will take many decades of growth for traffic volumes on the Peninsula to reach 4 HSR tph and 8 Caltrain tph at the peak hour, but the public is intentionally kept in the dark on this. Improved grade crossings will be quite adequate for traffic levels over the next several decades, as strategic grade separations can be gradually implemented as actual traffic levels justify. Shutting down a minor street crossing for a peak hour rail commute really shouldn't be hard either, but this doesn't translate into profit for CHSRA's for-profit contractors, who are running the show with the help of some sleazy politicians.

    It would be happy news indeed if a brave politician or political movement actually stood up to CHSRA's incompetence and lack of contractor oversight and demanded an efficient, cost-effective project that met the realistic future travel demands on the corridor.

    ReplyDelete
  71. Grade crossings suck anyway. With horns, accidents, suicides, and long closures which will only get worse as traffic volumes increase, you might as well remove them when you can find a funding source to do so. Regardless of how the 4 track shared corridor works out, CalTrain will still reap the benefits of grade separations, electrification, and the DTX tunnel, all paid for by HSR.

    ReplyDelete
  72. Grade crossings suck anyway. With horns, accidents, suicides, and long closures which will only get worse as traffic volumes increase, you might as well remove them when you can find a funding source to do so.

    Let's be careful and not overstate the benefits of grade separation. The only technical problem solved by grade-separation is to permit train speeds greater than 90mph.

    Grade crossings do not prevent suicides. They are not needed to eliminate horn blowing. They are not cost-effective for reducing accidents.

    ReplyDelete
  73. Don't forget about service frequencies.

    ReplyDelete
  74. Grade crossings are compatible with speeds of up to 125 mph according to US law, as long as they're protected by an impenetrable barrier. Up to 110 mph, four-quadrant gates are enough.

    While HSR lines never have grade crossings, HSR trains running on legacy track at lower speed do have grade crossings, occasionally. Rail operators try to eliminate them as much as possible, but they don't always consider grade crossing elimination to be essential to service. On busy lines they just keep the road closed off to cars for much of rush hour. The Chuo Line still has grade crossings in Tokyo, though JR East is currently implementing grade separations.

    ReplyDelete
  75. Don't forget about service frequencies.

    They are not required for higher service frequencies.

    (Let the road users pay for the grade-separation if too impatient to wait at crossing gates.)

    ReplyDelete
  76. Grade crossings are compatible with speeds of up to 125 mph according to US law, as long as they're protected by an impenetrable barrier. Up to 110 mph, four-quadrant gates are enough.

    US law mainly deals with FRA-compliant tanks. What about light-weight EMU, which is what Caltrain would like to run?

    Most likely FRA would require heavy and useless locomotive plonked at both ends of the train. Actually, they may even require this for far lower speeds too...

    ReplyDelete
  77. Caltrain is operated by Amtrak staff, so I never said they were an efficient outfit.

    One example.

    Train 368 (Baby Bullet) arrived Tamien 5:39. This trainset depart to San Francisco as Train 285, from Tamien 6:24. Engine, Bombardia coach and crews stays at the empty station for 45 minutes without generating any revenue.

    ReplyDelete
  78. "When Caltrain planed additional baby bullet in 2005, thier first proposal was 57 minutes, 6 stops. (SF, Milbrae, San Mateo, RWC, PA and Sunnyvale). This proposed time no longer can be found from web site but was listed in new archive."

    A few months (?) before the expresses started in 2004, Caltrain published their "Draft Proposed Service Levels", showing five expresses each way, all stopping at Millbrae-Hillsdale-PA-MV, and some stopping at 22nd St as well. The four-stop schedule was 57 min, the five-stop was 59, same as now.

    ReplyDelete
  79. "When Caltrain planed additional baby bullet in 2005, thier first proposal was 57 minutes, 6 stops. (SF, Milbrae, San Mateo, RWC, PA and Sunnyvale).

    Please see here.
    http://web.archive.org/web/20060218063231/www.caltrain.com/pdf/Planning_for_the_Future/weekday_timetable_7_1_05.pdf

    This was so called "Super 88" schedule. However, this proposal had been modified to 96 train when most of station saw reduced service.

    ReplyDelete
  80. Thanks for that link, which got clipped

    http://tinyurl.com/ybfsor4

    Indeed it does propose to schedule trains making five stops between San Jose and SF in 57 minutes-- which isn't ridiculous, but allows for almost no delay. Far as we know they never actually tried to run that schedule?

    ReplyDelete
  81. DE:

    I don't know what the FRA rules for noncompliant trains are. I believe compliance is only required for lines sharing tracks with freight, though. Non-compliant light rail lines do have many grade crossings, so the rules may be the same.

    The compliant EMUs of the LIRR and Metro-North have some grade-crossings as well. They're third rail-powered so the speed is inherently limited, but I don't think there's a special rule for EMUs, again.

    The four-quadrant gate rule is there to protect drivers, not trains. If the train is supposed to withstand an impact with a freight train, it should withstand an impact with a car.

    ReplyDelete
  82. Adirondacker1280018 January, 2010 21:38

    The four-quadrant gate rule is there to protect drivers, not trains.

    Yes, even "lightweight" trains will win any argument they have with an automobile. The main difference between a collision at 30 MPH and one at 60 or 90 or 125 is how long the debris trail along the tracks is going to be. Even trolley cars....

    http://www.youtube.com/watch?v=CV2rdGX4JYc

    ReplyDelete
  83. Richard Mlynarik said...
    "In civilized parts of the world 200kmh isn't rammed through the middle of suburbs."

    Wrong. Unless you consider the United Kingdom uncivilized.

    It's practically impossible to build a railway line anywhere in England without going through the middle of suburbs, and the trains do hit 125 mph *in* London suburbs on what are essentially 19th century alignments.

    Of course they didn't have to widen the ROWs because they were very wide to start with.

    But also, they were sometimes rammed right through housing originally in the 19th century. Though perhaps they weren't civilized back then either?

    ReplyDelete
  84. Spokker:

    "For example, LA will have a lot of connecting transit to HSR, but it'll be slow and meandering crappy light rail."

    It'll give you better trip times than the network of super-meandering undersized tube trains and half-hourly-or-worse surface rail in London.

    LA's been doing quite well with its rail design really, and will end up with an excellent network. Much better than most of the US -- look to San Francisco for real incompetence. Or Boston. Or Chicago.

    That said, until the California state government is fixed to eliminate the 2/3 rule and make government *possible*, California is really in trouble. And until the god-damned US Senate is fixed (eliminating the filibuster is a bare minimum), the US as a whole is in serious trouble.

    LA City government? Doing fine, really.

    ReplyDelete
  85. " Alon Levy said...

    Okay, so Spain is quadrupling the tracks. Good for Spain."

    Actually, they're *sextupling* them. But this one set of three tunnels is planned to carry every high-speed service in the entire *country* plus all the "conventional" intercity trains going through Madrid plus the majority of the Madrid commuter network -- and Spain has *two different track gauges*, which makes the problem a lot more difficult.

    Anonymous wrote:
    "I want to go back to the assumption that Caltrain will be on low platform. If they actually were compatible with HSR, which problems would go away?"

    1 -- bottlenecks approaching the Transbay Terminal. This could be managed with three platforms and two tracks on the approach. This would not solve the curve radius problem.

    2 -- ability to practically run semi-expresses with more stops than the HSR super-express trains and fewer than the Caltrain locals. This becomes very difficult, and causes track obstructions and bottlenecks, if the express and local platforms are entirely separate heights. This is really very valuable.

    ReplyDelete
  86. Mr Mlynarik says "BART's top speed of 80mph in the exurbs ...isn't anything to sneeze at either"

    Adirondacker12800 responds "It's something to go "meh" over. We'd have to find an New York Central or LIRR foamer to tell us when the first MUs topped 100. The first M1s were delivered in 1968. The M1/M2/M3s are rated at 100. Metro North's track in NY is 90 MPH track. I assume the expresses go 90 in NY."

    Blogger timz said...

    " MetroNorth NY to Croton-Harmon 32.2 mi 19 stops 68 min"

    There are very comfortable dwell times, so that's about 19 minutes in dwell. Yes, the trains are hitting 90 mph.

    "The express schedules don't pin you in your seat either:
    NY to Croton-Harmon 32.2 mi one-stop 43 min start-to-start"

    And there we go. The trains have to go reasonably slowly in the Park Avenue tunnels from Grand Central to 125th Street, and very very slowly across the Harlem River and along its curves (track straightening would be impossible at the bottom of a cliff next to the river), and don't start hitting high speeds until north of Yonkers, so yes, a 45mph average reflects 90 mph running.

    ReplyDelete
  87. "Blogger Alon Levy said...

    SSFF is the only proposal on the table in the LA Basin, where there's no plan to bring Metrolink into the 21st century as there is for Caltrain."

    I think this is because most of the Metrolink lines are *major* freight lines, and the usual FRA types cannot conceive of a vehicle which could operate both on freight tracks and on high-speed tracks. :-P

    The Antelope Valley line might be converted to all-lightweight operation, as it is probably feasible to kick all freight off it (or replace it with specialized night-only freight trains or something), which would give four shared tracks in that area. It might also be possible to remove FRA-compliant freight from off the San Bernadino line, which is also lightly used. It's not feasible for the others, though, as they include the UP and BNSF mainlines and the only set of freight-usable tracks to San Diego.

    ReplyDelete
  88. I know the FRA is sabotaging any good commuter rail attempt. What I'm complaining about is that Metrolink isn't trying to get waivers the way Caltrain is.

    ReplyDelete