26 November 2010

Future EMU

The newest automobiles are shown in Paris, Tokyo and Detroit. The latest aircraft are shown at Farnborough and Le Bourget. What about the newest passenger trains? That show, in Berlin, is known as Innotrans and provides a window on the latest evolution in the type of EMU (electric multiple unit) trains that Caltrain is pursuing. One of the future train concepts shown at the recently-concluded Innotrans 2010 is Bombardier's Omneo product platform, ordered in large quantities by French regional transportation authorities and due to enter service in mid-2013. That order is in itself noteworthy, because Bombardier beat French firm Alstom, the incumbent provider of the previous generation of double-deck regional trains, on its home turf.

Tailored slightly for U.S. dimensions, a possible Caltrain Omneo concept is shown in the sketch above. The following key features set it apart from the 1990s-vintage EMUs that feature in Caltrain's electrification plans--the very EMUs that the Omneo is set to replace:
  • Articulated design. The wheels are placed between, rather than under each train car, leading to a smoother ride. Gangways between cars, traditionally cramped, dark, loud, shaky and unwelcoming places, are transformed into full-width, seamless passages that turn the train into one continuous space. Vehicle systems (traction equipment, air conditioning, etc. shown in dark gray in the figure) are moved entirely out of passenger spaces, below the floor and above the ceiling. With inter-car gangways and systems cabinets gone, open sight lines promote a sense of space and safety inside the train.

  • A usable amount of high-floor space. Bi-level EMU cars are always a configuration challenge, since passenger spaces must fit around numerous vehicle systems while still providing full ADA accessibility. Accessibility means that wheelchair spaces, bike spaces and accessible toilets must all be provided on the door entry level. Traditional bi-level cars do not provide sufficient room for all these amenities in the high-floor vestibule areas over the wheels, and thus typically board on the lower level where more contiguous space is available. Articulation gets rid of this design constraint.

  • High-level boarding. While the traditional bi-level EMU architecture tends to allow boarding only on the lower floor, this concept features a 1220 mm (48 inch) boarding height that would be compatible with high-speed rail platforms. The Omneo is offered in 550 mm, 760 mm and 920 mm boarding heights, but additional vertical clearance available in the U.S. makes a 1220 mm floor height feasible. A shared platform interface between Caltrain and HSR presents enormous operational advantages by allowing any train to access any platform, especially at San Francisco Transbay.

  • Dual, High-Low Doors. To allow a gradual transition from today's 8-inch platforms to 48-inch HSR-compatible platforms, each vestibule features two doors, one for each height. While this temporarily restricts door capacity during the transition period, the modular door assemblies provide for easy reconfiguration to 100% high platforms once the transition is complete. Dual height doors are not unprecedented.

  • Extra wide interiors. The articulated design shortens car bodies from the traditional 26 m (85 feet) to just 10 or 15 meters. Shorter cars, for a given loading gauge, can be made wider than longer cars. A U.S. spec Omneo car body could be 3.1 m (122 inches) wide in the 15-meter double deck section, and 3.15 m (124 inches) wide in the 10-meter single deck section, all within AAR Plate F. That's 4 and 6 inches wider, respectively, than today's Bombardier bi-level cars used in Baby Bullet service. Such wide interiors might enable comfortable 5-abreast seating.
Bombardier vs. Bombardier

It's instructive to compare the Omneo EMU concept to another product in Bombardier's portfolio, the bi-level commuter cars that Caltrain operates for its Baby Bullet service. The diagram at right shows a direct comparison of a Caltrain car to a pair of Omneo cars of roughly equivalent length.

At three seats abreast, an 11-unit Omneo train measuring 145 m (475 ft, a bit shorter than a 493 ft Baby Bullet train) accommodates 695 seats and 48 bikes, compared to 674 seats and 48 bikes for a Baby Bullet. With a two-abreast seating plan making use of the extra width for aisles and standees, the Omneo would still accommodate 590 seats (not counting 48 flip-down seats.) Using a metric that Caltrain is fond of, seats per unit length of train, here's how it stacks up:
  • Baby Bullet: 1.37 seats/ft or 4.5 seats/m (including locomotive)
  • Omneo 3+2 seating: 1.46 seats/ft or 4.8 seats/m
  • Omneo 2+2 seating: 1.24 seats/ft or 4.1 seats/m
These figures show that an Omneo-like EMU would achieve rough parity with the Baby Bullet in terms of seating density. Of course, as we've discussed before, it's not all about seats... standees are passengers too, if you give them somewhere comfortable to stand.

Innovative Design Enables Platform Height Compatibility

This article is not intended as an endorsement of Bombardier's products; their Omneo train merely serves as an illustrative example. Other manufacturers (Alstom, Siemens and Stadler) aren't standing still, and are also increasingly focused on articulated EMU designs with innovative and flexible interior packaging. In light of these new technological developments, Caltrain's relentless pursuit of 20-year-old double-deck EMU designs locked into a low-platform architecture seems downright archaic.

Cutting-edge European vehicle designs will not compromise the basis of Caltrain's FRA waiver, and might help achieve one of the key tenets of compatibility: a common platform interface for HSR and Caltrain, without billions of dollars and years of platform reconstruction up front. The trains can come now, as part of the electrification project, and the rest can follow later as time and money allows. What is required is a little bit of forward thinking to future-proof the system and enable any train to use any track to access any platform.


  1. Equipment with doors to handle both high and low platforms? Good grief, I've been seeing such equipment for a couple of decades now in visits to Pittsburgh, Pa., on the light rail line there.


    Another photo, different source:


    Note that this articulated trolley car has three high level double doors, and a single low door. These cars were ordered for a system that was still running PCC cars, and which also still has some sections with street trackage (and street loading, from low platforms).

    It's a pretty wild system, with stations clinging to cliffs, awesome grades, tight clearances, private rights of way that take you between houses as if you were in an alley, bridges that have you looking down into people's chimneys (and some Californians are up in arms about aerial trackage--they ought to look where this thing goes), a town hall (Castle Shannon) surrounded by tracks and wires, some running right past doors in the building, and streetcar curves, including one that makes you think you are going to go right into that store on the other side of the sidewalk as the trolley makes the turn onto a street where it runs in the curb lane.

    Oh, a minor detail--the system is broad gauge, 5 feet, 2 1/2 inches, inherited from horse car days. It used to be a common trolley gauge in Pennsylvania, including several interurbans. It's also the track gauge of the New Orleans system.

    General Pittsburgh photo links:



  2. Hmm, for some reason the links as shown don't work. You will have to copy and paste to use them, at least for now.

  3. A postscript on platform height transition.

    Even if a large height transition were necessary (and I claim it isn't: that a ~650mm high platform is optimal for both HSR and regional trains, both at system start and into the future), the solution that Clem has outlined for doing so may be unduly expensive now and ongoing.

    By building dual low/high platform business into the rolling stock we have created a Gift that Keeps in Giving.

    Now for regional French service -- nation-wide regional services --, with intractable numbers of stations and lines to height-convert (or leave unconverted), the argument can be made that the lifetime of the rolling stock will exceed the length of the station conversion process.

    But that's not Caltrain. With a trivial number of stations, a (comparatively) staggering amount of civil engineer construction coming to the line (grade separations and crazy PB quadrupling madness will mean that every station will be rebuilt from scratch), and a small life-expiring fleet, we are in danger of designing, paying for and operating a fleet of trains that cart around a set of vestigial transitional kludges for 75% of their service life (say 20 of 25 years.)

    The odd sets of high and low doors of the odd single-deck sections of the "Omneo" design don't come for free: they have structural and weight and capacity penalties, as well as awful passenger distribution (ongoing dwell time issues.)

    Note I'm not saying this is bad engineering: in fact I think it is a very interesting one for lower-density RER networks with three or more heights of platforms and decades (or centuries) to come before heights are homogenized.

    But for Caltrain I think there may be better solutions.

    The best solution is to raise all platforms to a ~600-700mm height, which should be the state-wide HSR/Caltrain/Metrolink-NG standard. This can be done with little or no change to existing Caltrain equipment during transition, and new Caltrain equipment can serve both older (via temporary lifts) and newer (level roll-on boarding) platforms unaltered.

    The second best solution, if backwards-looking NEC/Shinkansen-style high-level 1100+mm platforms really are forced on us, involves a coordinated and phased transition of the stations (remember, there are only a handful of them) along with Caltrain line along with introduction of new equipment.

    The phasing plan is a little long to outline, but I've worked it out in detail:

    1. Start with building cross-platform transfer island platform stations at "bullet" stops, with the inside (slow) tracks temporarily raised in order to make the platforms "lower" for the slow trains.

    2. Introduce the new high-platform rolling stock to run bullet-only express stops. Cross-platform transfer (as Clem has outlined here) with low-platform local (legacy) rolling stock.

    3. A miracle occurs.

    No, really what happens is that we start reconstructing platform heights at local stations to the high level, starting from Santa Clara and working north. It's complicated, but it can work, given the cross-platform transfer business.

    As sections (SJ-Mountain View, Mountain View-Palo Alto, Palo Alto-Redwood City, Redwood City-Hillsdale, Hillsdale-Millbrae, Millbrae-Mission Bay) are successively high-level converted, more and more of the local service is assumed by the new rolling stock.

    There are tricky bits. But anyway, it can be done. The nice thing is that phased station reconstruction can accompany phased rolling stock roll-out with no "big bang" and no multi-year system transitional shut-down and no need for "Gift That Keeps On Giving" rolling stock.

    I have a truly wonderful proof but it doesn't quite fit in this text box.

    Summary: I have a lots of doubts about (1) the desirability of high-level (>750mm) platforms and (2) the wisdom of building multi-height docking for a shortish-term transition into the structure of brand new rolling stock.

  4. I'll have to say that given the options, you have to admire the current bombardier car layout. It's very specious, the doors on lower level provide a relatively easy entry, and all while using only 2+2 seating.

    If bombardier could EMU'ify this layout and throw in some articulation, it would be quite a design to beat.

  5. The platform height transition can be aided by using temporary platforms: just build wooden platforms and ramps to whatever height you want. You can even keep half the platform as high and half as low during the transition.

    Oh, and it's the 600-700 mm platform height that's "backwards looking", in the sense of being intended to provide backwards compatibility with legacy low platforms. With a train, unlike a tram, there's no reason to have a floor as close to the ground as possible, and a train design with no stairs inside is much simpler and I would think is noticeably cheaper. Not that any of this implies anything about what Caltrain should do.

  6. Does anyone know the comparative costs of building 550 mm and 1,200 mm platforms, keeping all else equal?

  7. @Richard: maybe I didn't make this sufficiently clear. The low doors would be converted to high doors as soon as the last low platform is rebuilt. The end state would be 100% high-level. It's not crazy to do this if it's designed in from the beginning, things being modular.

    Also, the French trains do not feature dual level doors. That was a flourish that I added, guessing that it would be feasible from a systems packaging standpoint.

    The bigger issue here is this:

    (1) the clock is ticking on Caltrain's fleet, and this solution decouples the schedule for fleet replacement from the schedule for platform reconfiguration. No coordination required.

    (2) An integrated solution would be great, but ask yourself honestly if you believe there is sufficient capacity for advanced planning of a complex logistical exercise such as you suggest, or if there is sufficient inter-agency coordination to pull it off? Not in this world, my friend.

    (3) The monkey wrench is the environmental clearance process of the respective elements of the corridor. That's a schedule that is nearly uncontrollable because it is subject to so many external influences. The HSR EIR will not include electrification of Caltrain, but it would include all the impacts to platforms. Even the most forward-thinking agencies would be hard-pressed to plan around that.

    @Martin: yes, the bilevel car is a marvel of packaging. But there's just no room left for pantographs, transformers, traction inverters, compressors, signal cabinets, etc. There's no way to make it into an EMU, or it would already have been done.

  8. Another constraint: money doesn't grow on trees. Assuming that the peninsula HSR project will be delayed until 2020 or 2025, and assuming Caltrain can scrape together $1.5 billion for electrification (a TALL order), it's going to be really difficult to find another half billion to redo platforms to a different height, only to demolish them later on for HSR.

  9. @Clem: I am a little confused on your reply to Martin, isn't making Omneo design an EMU the samething as making it "multi-unit control" capable, i.e. control multiple linked units from one cab?

    Also, with the two ends floor height above the boogie, it probably shouldn't be too hard to design an ending with gangway instead of sleek ends, allowing passengers to move between units.

    Anyways, I hope someone in Caltrain sees this, especially the PRP guys.

  10. @William: Martin's comment was about the Baby Bullet cars, which are locomotive-hauled. The Omneo already is, of course, an EMU.

  11. ... continued

    So I think that:

    (A) The insistence on "high level" (> 800mm) platforms is "backwards looking" because it only works well with single level train designs and results in sub-optimal double deckers.

    (B) The need for double deck trains, especially for HSR, is readily apparent, and because of horrific inexcusable grossly unprofessional SF terminal misdesign, will be an issue for CHSR(+Caltrain) almost from Day One.

    (C) There is no reason not to build a uniform double-deck CHSR fleet from the outset. If there's one thing I can guarantee you, it is that we will not under any political circumstance be buying an "off the shelf" train: therefore why not go with the best that contemporary engineering can offer, rather than with designs based upon backwards compatibility with 19th century passenger car construction?

    (D) For regional (Caltrain, Metrolink-NG) trains, either single- or double-deck work with low(er) level platforms, and there is no religious reason to prefer double deckers given an appropriate (non "commuter railroading") level of service. That said, double deck regional trains work (station dwell and interior circulation) far better with low(er) level platforms than with backwards-looking high platforms.

    (E) Low(er) level platforms offer a ZERO COST ZERO EFFORT fleet transition plan for the regional trains. New trains can serve old platforms (with some lift kludgery) and old trains can serve new platforms (with almost no modification at all.) THE PROBLEM IS SOLVED, right out of the box. Why borrow trouble by carrying out an incompatible do-over? The result of high platforms works no better for either HS or for regional, so why undergo the massive pain of getting there?

    Please note that for the last 15 or so years I was a huge believer in 760+mm platforms for the Caltrain corridor and HSR state-wide. Looking at and carefully thinking about the Bombardier design for SBB has convinced me that I was wrong. Paying attention to what the guys with the most efficient and intensively used rail network on the planet are up to is always a good idea. I strongly believe that others should at least try to see if the same analysis works for them.

  12. One other issue with the high-low design: what happens when a passenger in a wheelchair gets on at a high level platform and wants to get off at a low level one? These trains either need built-in lifts or else they'll have to keep using the hand-cranked lifts on the platforms. As for the Bombardier bilevels, there were actually plans to make an MU version for the electrification of GO Transit that never happened. I would imagine that it would be possible to stick things in the space occupied by the bathroom and the three seats under the stairs. It would probably be easier to have cars in married pairs with one car having the transformer and the other having the bathroom, for example.

    And, there's nothing wrong with looking backward. If BART did that, they'd have at least had standard gauge tracks on their "elevated duorails".

  13. [... continued due to broken blogspot]

    DWELL TIME is perhaps the most important rolling stock issue. The less time spent at 0kmh the less top speed matters, the less power used to maintain a existing timetable, the fewer tracks and platforms needed, etc; or, the more service can be provided with the infrastructure.

    Designs like the Bombardier "Omneo" are disastrous in this respect as they concentrate doors at limited points along the train. (Don't get me wrong; it's an interesting engineering solution for exurban regional networks decades or more away from train-platform height homogeneity.)

    Moreover all high floor double-deck trains must have bad door spacing because doors in such trains are more or less (yes I know about MI2N) required to be positioned at the ends of each car -- or worse than that, in the Omneo design.

    In contrast, with lower platforms boarding at the quarter points of each car can be readily arranged for both single and for double deckers.

    In fact uniform door spacing combined with excellent promised energy efficiency (and lifecycle price) were what won Bombardier the Swiss contract: those guys really really really care about not gumming up their stations and platforms with stationary trains.

    [... Concluded several comments above with the "So I think that..." that should have followed this exposition. Die, blogspot!]

  14. Adirondacker1280027 November, 2010 16:41

    In contrast, with lower platforms boarding at the quarter points of each car can be readily arranged for both single and for double deckers.

    Doesn't that lead to a lot of stairs on the double deckers?

  15. 1.) It looks like a high-door wheelchair lift would not as much constrain use of low doors?

    2.) How do different platform heights discussed on this blog, anywhere from 490mm up to 1200mm, impact the efficiency or optimization of the HSR bogey and high-speed shock absorbing suspension system? Put another way, does a lower common platform height constrain the design of the suspension? Even if a design exists and operates on xyz system in the world, given a choice a suspension designer may have a preference.

    3.) At some point does a higher platform start to make every tunnel in the system a little taller, adding to tunneling cost?

  16. 2) Generally not at all, as low floors will never extend over the bogies of high-speed trains, and there is sufficient room under high floors (at the level of high platforms) for such equipment.

    3) Similarly, you will never see a second deck above a high floor (except on gallery cars). When you have high platforms, you generally have a split level design where the entry areas are single floor and then that splits into an upper level and a lower level between the bogies. For low-entry traincars, you just eliminate the mid level, so there is no effect on overall height.

  17. Off topic I know... I live directly across the tracks from where someone was killed on the tracks in Burlingame yesterday and CalTrain was slowing down and laying on the horn every time they passed. I wonder if that is the drivers being respectful, a new policy, or just a temporary thing. I listened to it all day and it made me sad that a grade separated system that would prevent this is being held hostage by money centric nimbys. Why isn't this part of the discussion in the media?

  18. @Richard: the door spacing thing is key in Times Square, Chatelet Les Halles, or even Embarcadero BART--places where platform capacity and massive boarding/alighting volumes are constraining factors. On little old Caltrain, it's not a big deal and never will be, and boils down to purism. In practice, dwell time is set by door capacity (i.e. number of doors), but not spacing. It's simply not a design driver for a system that will carry, someday, maybe, 100k weekday passengers.

    As for selecting the best platform height, that's going to be done by divine PBQD intervention. There is absolutely nothing that you or I or even Caltrain can do about it, and all indications are pointing towards high platforms and single-deck HSR. I'm just pointing out that even with those (sub-optimal, to purists) high platforms there is a relatively painless way to go with the flow. Platform height compatibility is far, far more important than platform height itself.

    @James: no, tunnel size is driven primarily by aerodynamic free clearance, not train height. See engineering docs.

    @Burlingame: that's temporary, while crews are still at the scene. A guy holds up a W sign which requires engineers to lay on the horn to warn people on and near the tracks as they pass through. They even do that when emptying garbage cans on station platforms. Safety First.

    I hope for your sake that you didn't see or hear the event.

  19. By the way, and this is kind of off topic, SMART looks to be going with FRA-compliant high floor DMUs, to be made by Nippon Sharyo. But there were a good number of other bidders. And one of the technical requirements was a level floor through the train for wheelchair access, which is important when one car has the bathroom and the other one has the snack bar.

  20. @James: wheelchair lifts should not be used on modern commuter trains, as they increase dwell times and force the trains off-schedule.

    @Arcady: why does a commuter train need a snack bar? Are they

    @Richard: I know you think CAHSR isn't going to use off-the-shelf trainsets, but both agency statements and a look at the politics of it say it will. PBQD et al do not design trains, and gain nothing from using one-of-a-kind trains. The lobbies for trains are going to come from the vendors, whose agenda is to sell their off-the-shelf designs at the highest possible price. Nobody has any intention of designing a new bilevel high-speed train with distributed traction, without which there's no capacity gain from using bilevel trains at all.

    The best industry practice for cramming as many seats as possible into one train is in Japan. There, the high-speed trains are mostly single-level; the bilevels are used only for high-speed commuter trains and are limited to 240 km/h. The JRs prefer to maximize capacity with 3+2 seating, distributed traction, no restaurant car, and 16-car trains instead of two 8-car trains coupled together. That way, a single-level 16-car Shinkansen has more seats than two TGV Duplex trains coupled together, despite having a higher seat pitch.

  21. Adirondacker1280028 November, 2010 09:11

    the door spacing thing is key in Times Square, Chatelet Les Halles, or even Embarcadero BART--places where platform capacity and massive boarding/alighting volumes are constraining factors.

    Grand Central has less platforms. Without looking at turnstile numbers Grand Central is busier. Time Square on the other hand has at least two different kinds of trains serving it, with varying door spacing. It's more or less evenly spaced along all the cars. Nothing in the Bay Area is ever going to be Times Square or Grand Central. It's never going to be one of the busier express stops on single line either.

    Go to town, keep in mind that the MTA only counts people when they enter the system.


  22. @ Alon Levy

    The snack bar in the SMART trains is removable and replaceable with either extra seats or another restroom. The mountings for the extra seats are already built into the floor.

  23. Have to agree with Richard's comment that the Omneo is a disaster for dwell time. And if you read some of the recent Caltrain reports, dwell time is already a huge problem.

    Also, the articulated design would normally be a wonderful solution for evenly distributing load throughout the train -- particularly bicycles. But here, that feature is wasted. For example, how is a cyclist supposed to move to the next car if it means going up and down stairs?

    So (once again) the question has to be asked: what problem is Caltrain attempting to solve by using bi-level rolling stock?

  24. That's 4 and 6 inches wider, respectively, than today's Bombardier bi-level cars used in Baby Bullet service. Such wide interiors might enable comfortable 5-abreast seating.

    4-6" will not make 5-abreast comfortable.

  25. Dwell time can matter in two different ways: the maximum dwell time at a mid-line station is one of the things that determines the line capacity, and dwell time at a terminal determines how many tracks you need there, or how many trains per hour you can turn. But total dwell time also matters, because it affects how long the train takes to get from one end of the line to the other. That affects average speed, which customers care about, and the faster a train gets to the end of the line, the sooner it can turn around for the run back to transport more paying customers.

  26. Did anyone notice in the "electrification plans" that Clem linked, that Caltrain is planning on keeping the newer MP36PH locomotives and Bombardier BiLevel cars for both the Baby Bullet and Gilroy services? And this AFTER electrification?

    They're also planning on keeping their three newer F40PH locomotives after electrification for Gilroy service.

  27. "That's 4 and 6 inches wider, respectively, than today's Bombardier bi-level cars used in Baby Bullet service. Such wide interiors might enable comfortable 5-abreast seating."

    Single-level suburban cars in the NY area are already 126 inches external width at seat level (they pinch down to 120 inches at floor level). People still turn up their noses at 5-abreast.

  28. @Drunk: I share your doubts about Caltrain's fixation with double deck trains. Single deck trains would do just fine.

    From Caltrain's point of view, they are committed to double deck EMUs because that is the fundamental premise of the FRA waiver. Take that away, and suddenly you can buy an FRA-compliant Kawasaki M-8 right off the shelf... apparently not what they're aiming for. This and other justifications for the waiver are found on page 10 of the structural analysis report.

    The key passage:

    An operating scenario using FRA-compliant EMUs to replace the locomotive-hauled service was examined early in the Caltrain 2025 study. However, it did not meet Caltrain’s requirements for the following reasons:
    • Higher train weight limited the energy savings.
    • A Single level EMU with 2x2 seating can only accommodate 90 seated passengers at best. Caltrain needs at least 100 seats to meet ridership and platform length requirements.
    • High-level boarding would not accommodate existing Caltrain rolling stock and would not provide adequate clearance for Union Pacific freight trains.
    • FRA-compliant multi-level EMUs do not exist today.

    As to your comment about bicycles, as a regular Caltrain bike rider I can tell you that moving bikes through the passenger areas inside of the train is a non-starter. The distance between the door and the rack must be minimized or chaos will ensue. Stairs are a red herring in this respect.

    @Peter: you can see the parochial insanity of making the entire corridor architecture, for generations to come, contingent on retaining a small sub-fleet of existing equipment that would fetch excellent prices in the second-hand market. Since the ADA prohibits steps down into trains, the maximum platform height allowable with Bombardier cars is 25 inches.

    They're happy to let poor peak-hour timetable planning drive a bloated fleet size requirement, but oh no, they can't acquire a few more because they must be stingy and retain any rolling stock "investment" that still has life in it. That's a very odd way to squeeze value out of their capital equipment.

    If I had to do it, the diesels would be gone, and you could operate the 6 TPH overtake with as few consists as they have operating today's rush hour (21). And there would be NO FRA trains allowed north of CP Coast (heck, even south of CP Coast... keep 'em on a separate track, as discussed previously.)

  29. @Tim: wow, I had no idea they were that wide. Here I thought the East Coast had generally more constrained dimensions... I agree that 5 abreast is probably not going to work. It's better to provide usable space for standees, in the form of wide aisles.

  30. @ Clem

    Because it would be really hard to purchase a few electric locomotives to pull the Baby Bullets, while purchasing a couple of high-floor DMUs for Gilroy service, which they're even calling a "shuttle" service. Hell, purchase a few of SMART's DMUs, run them through to Salinas. That way they wouldn't even have to worry about FRA compliance.

    But instead we have idiots in charge.

  31. Please, they're not idiots. They're stubborn, maybe, and dealing with what they perceive to be Unique Local Conditions, for sure.

  32. As to your comment about bicycles, as a regular Caltrain bike rider I can tell you that moving bikes through the passenger areas inside of the train is a non-starter.

    Copenhagen knows more about bicycle transport than anyone. Here is how the S-tog looks.

    Of course, having bikes near a door helps, but what does one do if that spot is full, but the next car isn't? Exit, run down the platform, and re-enter the train? With S-tog, I never saw issues with aisle access, and they use wide seating.

  33. Speaking of wider trains and 5 abreast: I'm pretty sure East Coast trains are 10'6" over the door sills, with Amtrak's Amfleets being narrower at door level for no particular reason. Maybe they wanted to make them look more like airplanes? Or thought the tube is a stronger shape? Who knows, but the most restrictive clearance is into Grand Central, and the trains that go there are M-series MUs, Comet-series coaches and Genesis engines. As for 3-abreast seating, I think it has to do partly with the seat pitch and the low back of the seat closest to the center found in many commuter trains. If you have facing pairs of seats rather than all in one direction, I feel like people would find it a bit more convenient. That and Australia and the UK use 2+3 seating on their inner suburban trains, and their clearances are more restricted.

  34. The purpose of double-decker trains is to offer more seats. While a local train making all stops might encourage a few more people to stand, most of Caltrain riders take the baby bullets with average trip length of 28 miles or a little over 30 minutes. Heck, even local trips average 20 miles or probably over 30 minutes in length. So reason #1 for pushing bi-level cars is so people don't have to stand for 30+ minutes. (See page 5 of ridership link for stats: http://www.caltrain.com/Assets/Stats+and+Reports/Ridership/2010_Caltrain_Ridership_Counts.pdf)

    Additionally, when I walk down the isle of a morning baby bullet, over half the passengers are on laptops working. Table seats go pretty fast. Those passengers take Caltrain so they can be productive during their commute. If you force them to stand, they'll probably drive making it another reason why Caltrain prefers seats.

    While you could make up for less seats with more frequent service, many passengers rely on employer shuttles to drive them between the office and the station. Those shuttles are timed to meet the baby bullet trains, and unless the number of the baby bullets goes up, you can probably expect the same loads for those trains due to employer shuttles being timed with them. Most employers only run only 2-3 runs in AM and PM timing them to first baby bullet, last baby bullet, and a random one in the middle. More frequent service might not spread out this load very much.

    While I enjoy reading the block, I'm still not understanding the anti bi-level sentiment. Is it mostly because high-floor bi-level EMU's are rare?

  35. @Clem: all mainline US loading gauges are the same width. The dimensional constraints are on height.

    3+2 actually makes more sense on intercity trains than on commuter trains, because the reduced standing space is not a problem. Commuter trains should plan on having plenty of room for standees, and not just in the vestibules. If it's inconvenient to stand in the corridors between the seats, then the vestibules can be overcrowded while the seating area is undercrowded; this happens on the older single-level trains on the RER.

  36. Why are we still talking about EMUs? I am now more focused on Caltrain/community compatibility than Caltrain/HSR compatibility. I think it's a shame to put money into electrification that could go into improving the safety of our grade crossings. I think it's a shame that we have not insisted that Caltrain rewrite its program and project EIRs with DMUs or DEMUs as alternatives to the EMUS, instead of the no-build alternative that is there now.
    Why are we wasting money on an catenary infrastructure that is (thankfully) unlikely to be needed for a long long time?
    The latest message from PRP is that that PRP will continue to listen to our communities. What more can we say that we haven't already said? The issues in not so much their listening as their speaking.

  37. Caltrain-HSR compatibility IS Caltrain-Community compatibility.

    If Caltrain and HSR are compatible, then the amount of development on the Caltrain corridor, as in the effects on the community, can be reduced.

  38. @ Clem

    Yes, they are not idiots, but exceedingly stubborn. Calling them idiots was just venting.

    If they insist on locomotive-hauled Baby Bullet service, that does not preclude selling their current locomotives and purchasing electric locomotives. They could even purchase non-FRA-compliant equipment (cheaper, less wear and tear) as the Baby Bullet service would not run south of Tamien and they already have their waiver for non-compliant operation. Then just purchase some non-compliant high-floor coaches, I believe Bombardier even makes double-deck ones.

    Presto, you have your electric locomotive-hauled Baby Bullets.

    This of course is all assuming that HSR lets Caltrain use the express tracks at all. Otherwise it's moot, anyway.

  39. @Peter: even electric locomotives are a really suboptimal solution. The problem is the same as for bilevel EMUs: the market for them isn't as developed, leading to high costs. On top of it, their axle load is normally about 20 t, versus about 10-15 for suburban EMUs and 11.5-17 for HSR trains.

  40. @ Alon Levy

    Yes, I agree. Buying some extra EMUs rated for the speed they want the Baby Bullets to run at would be the best way solve the problem.

  41. The market for non-FRA-compliant EMUs built to work on Caltrain currently has exactly one potential customer: Caltrain. The market for electric locomotives is much more developed, and there's already one fairly modern model in active production for use in the US (the ALP-46A) and a Siemens product soon to be in production for Amtrak. And I'm sure NJT or Amtrak would be happy to lease any of their electrics to Caltrain, especially now that there's not as much of a shortage. Amtrak would probably even be willing to sell its HHP-8 fleet, which was built in 2000 but is going to be replaced to ensure fleet uniformity (Amtrak only has 15 of them).

  42. The market for non-FRA-compliant EMUs built to work on Caltrain currently has exactly one potential customer: Caltrain.

    Correct. "non-compliant" is limited to a handful of structural requirements. The other 99% of FRA requirements still apply. There must still be grab irons, ditch lights, number boards, a big 'F' stenciled at the end designated as 'front', door status lights, and hundreds of little details that make American trains, well, American. You can forget buying anything European "off the shelf" at prices remotely comparable to what is seen in Europe. Just wait for the RFP.

  43. Arcady, the modifications required for full FRA compliance are more substantial than those required for Caltrain-FRA compliance. That's why the cost of those electric locos you're mentioning is about $7 million per unit.

    Single-level high-floor EMUs may not be fashionable, but they're basically a commodity, and cost $2-2.5 million per unit, about the same as or less than what Amtrak pays for an unpowered coach.

  44. Wonderful. So CalTrain's waiver is contingent upon (a) bilevel trains (b) low-level platforms and (c) continued freight operations. And I used to think there was just a tiny bit of hope of doing this right...

  45. Oh yes, single-level high floor EMUs are pretty common, even in FRA-compliant versions. There are at least the M-8 and Silverliner-V already in production, and I imagine Nippon Sharyo and Siemens would be happy to make an electric version of their proposals for the SMART DMU. New Jersey Transit and Denver are likely to be buying new MUs soon as well. The NJT order in particular is probably going to be around 250 cars. And of course as NJT gets their replacement MUs, they'll probably be willing to lease their old Arrows for very affordable prices, and probably their ALP-44s locomotives as well. In fact, I imagine that if Caltrain were to magically get electrified tomorrow, NJT would be willing to lease or sell them enough ALP-44s to run the SF-SJ service. Sure they're not shiny and brand new, but they're not old for electric locomotives, and they're still 7000 hp machines, twice the power of the current diesels.

  46. "I'm pretty sure East Coast trains are 10'6" over the door sills, with Amtrak's Amfleets being narrower at door level for no particular reason."

    If you get a chance to look sometime I bet you'll find the gap between the car and a high platform is the same for all of them. No? US passenger cars exceed 120 inches width at high-platform level; it seems bilevel/gallery cars like to be the same width for their full height, so they "can't" be as wide as single-level cars. (Don't ask me to define "can't".)

  47. Nope, I'm pretty sure the Amfleets are narrower at the door sills. They've got to fit in the same width clearance, but instead of being a square cross section, they're round, so there's a noticeably wider gap between the platform and the train, but it's only on the order of a couple of inches.

  48. Joey wrote: "So CalTrain's waiver is contingent upon (a) bilevel trains (b) low-level platforms and (c) continued freight operations."

    Define the "requirements" and you define the outcome. Works every time under every circumstance.

    Nice work for some very, very very special signal and rolling stock consultant pals, that's for sure.

    But honest, we'd just love to be able to do something other than CBOSS and Alstom TER-2 and CPUC platforms ... but the mean old federal government has tied our hands (... because we told them exactly how to tie the knots). Oh woe! We're so very sorry! We'd sure love to consider alternatives but it is so late in the process! Don't worry, next time for sure things will turn out differently, trust us. But for now we need all you little people to hold your noses and give us your support because if we don't electrify which by definition requires CBOSS then PUPPIES WILL DIE.

  49. I get the low platform and TER-2 dig, but in what way does the waiver require Caltrain to develop CBOSS?

  50. Alon, dude, if you bothered to read the waiver, you might answer your own question.

  51. "Nope, I'm pretty sure the Amfleets are narrower at the door sills. They've got to fit in the same width clearance, but instead of being a square cross section, they're round, so there's a noticeably wider gap between the platform and the train"

    Pretty sure that's your mistake. Comets are rectangular cross-section down to just above floor level; the body of the car is 126 inches wide, but the door sills are 120 inches apart. But I never did actually measure them when I had the chance, so that isn't 100% certain.

  52. I did read the waiver, when it was given. It mentions CBOSS, but for the FRA's purposes, ERTMS can achieve the same goals, no?

    Thanks for the anonymous snipe, though. Very productive.

  53. Adirondacker1280002 December, 2010 19:28

    Depends on how you view the FRA's requirements. The trip lever system used by the NYC subway fulfills most of them. Last I heard the LIRR is going for a waiver because the system they use meets all of the PTC requirements the FRA has proposed. .... but then the Peninsula is special and can't possibly find anything similarities between it and the Port Washington branch....after all wouldn't want anyone to think there's some similarities between Millbrae and Woodside...

  54. Do you actually have a reference for the LIRR going for a waiver? I've heard nothing of that sort on the usual channels.

  55. No, the trip lever system is not enough. It has to stop the train BEFORE it passes a red signal, never mind the fact that overlaps and speed control signals mean the train is guaranteed to stop before hitting anything. And Metro North/LIRR style ATC is not good enough either, since it doesn't actually enforce a stop, only a slowing to restricted speed. And yes, LIRR/MNR want a waiver, but a completely different one from the one that Caltrain is getting. They want a waiver from the PTC requirement, on the basis of their existing cab signal systems being "good enough", which, really, they are.

  56. Adirondacker1280003 December, 2010 16:11

    The block ahead of the stop signal is always empty on the subway. Trip the lever and there's nothing to run into. Pass a stop on the LIRR and the train goes into penalty stop, been that way since the train wrecks in the early 50s.

  57. While selecting an appropriate Caltrain EMU a system designer should consider the character of the passenger traffic to be served. Caltrain’s total weekday ridership is in the fairly impressive 40,000 passengers per weekday range. Every weekday 5 trains per peak hour schedules are spread over 5 hours simultaneously in both directions. No regularly scheduled train has more than 700 seats. Upgrading this service by combining modern electronic controls with a 4 track grade separated 125 mph electrified right-of-way would make a 24 tph, 6 local and eighteen 125 mph peak speed expresses each way per hour schedule possible and affordable if no paid employee was required on every train. Of course average service speeds for most passengers could improve immensely with a 75% express verses local schedule. For example a non-stop 125 mph train between Palo Alto and Millbrae with a 50% greater power to weight ratio than BART, a 3 mile per hour per second, until power limited, acceleration and braking rate, a 25 second dwell time and a 3.3% en-route schedule recovery period would leave Millbrae 10 minutes 6 seconds after leaving Palo Alto. Caltrain’s local service schedule between these two points is now 37 minutes.
    The considerable traction power required to maintain this rapid pace could be substantially conserved by tailoring each train length and frontal area closely to actual traffic requirements. Train capacity adjustments can be most precisely accomplished with single deck un-married EMU’s.
    The current peak demand (arriving near 8:00 am in downtown San Francisco) would not likely double no matter how effective the service because of limited need and saturation effects. (Most people will not arrive in downtown SF early in the morning unless they have a job in the area and close to half of those who do appear in the central core area of downtown SF at that time already take commuter trains. See TBT-2002 EIR). Therefore two duplex or three single deck cars per train may be enough for most peak hour trains. Of course long term traffic may grow considerably above this estimate especially if one-seat-rides from the peninsula to Market Street in downtown San Jose, the SJ and SJ airport terminals, the Central Expressway through Sunnyvale and Santa Clara, San Jose’s Golden Triangle, and San Francisco’s Montgomery and Embarcadero stops on Muni’s Market Street Subway. In order to meet possible extraordinary future growth one should consider railway technology’s ability to increase capacity without increasing seat density for a given train length. Adding cars to trains will increase single track station net capacity by approximately 85% of the train length increase. While being responsive to transit industry acceleration limits and safety margins I will provide a mathematical proof supporting this capacity claim upon request.

  58. The arguments for the CHSR-Caltrain system to allow any train on any track are quite persuasive for a host of solid reasons. Since the projected CHSR system will be pushing far closer to practical physical limits at their peak speeds common standards required for interchangeable track use such as loading platform height should usually be governed by CHSR’s rolling stock requirements. For example any inherently expensive system should be designed to minimize its largest cost elements. For CHSR operation power consumption costs during high speed running will be quite significant. Over 93% of the traction power needed to sustain 220 mph on level ground would be consumed by wind resistance. (A proportion implied by Bernard de Fontgalland’s description of passenger train power requirements and the wind losses deduced from the performance of the 300 series Shinkansen.) A careful reading of my fluid dynamics text book (Air’s dynamic properties below 300 mph behave closely to that of most liquids.) shows that air resistance rises in proportion to the boundary layer air flow’s rate of change in velocity between a train and nearby surfaces. For railway systems the shortest distance between a train and a fixed object (and therefore the area of greatest air resistance to forward motion) is usually between a moving train and its track-way. Therefore it should be no surprise today’s HSR level boarding platform heights range between 1069 mm and 1250 mm above the rail; significantly above today’s duplex commuter car level boarding entrances.
    Given that bogy frames are fabricated as steel castings, an EMU manufacturer might agree for a moderate fee to adjust, within a limited range, the exact floor height for a single level car only a few months before production starts.

  59. John Bacon: you've got a few things wrong. First of all, floor height is orthogonal to bogie frame size. The gallery cars and bilevels have the same size bogies, and the same height floor above the bogies. And I bet the Superliner-derived rolling stock has roughly the same sized bogies.

    And I'm also curious about your claim that half of peak period travel to the SF CBD is on commuter rail. Is that just for the peninsula, or does that include BART as well? And the Muni Metro?

  60. I think the Bombardier Omneo would be a perfect fit for CAHSR. Clem, if you aren't already aware, but a group of people, myself included, are strongly advocating that the California High Speed Rail Line only travel on the Peninsula as far as Redwood City, then leave the line, travel over the bay on an existing rail bridge that would be rebuilt to acommodate the line. It would then stop to serve the East Bay and travel over the Pancheo Pass to the Central Valley. A separate service from San Jose to L.A. would travel up the East Bay on conventional tracks and connect to the high speed line. Though you may think that this won't occur, IYO, the likelihood of the alignment in the Bay Area changing before construction begins is increasing every day. If this happens, then everything changes for Caltrain. P.S. Clem Do you think that the Alstom AGV would be a perfect fit for the CAHSR? And another thing, if the FRA Rules were to change, would a product like the Bombardier Innovia be appropriate to replace NJ Transit's Comet Fleet and operate on the NEC?

  61. Sorry, I meant Bombardier Spacium in my earlier post.

  62. Adirondacker1280013 March, 2011 17:58

    Spacium,,, too slow for the NEC. They'd have to redesign it for 10'6" loading gauge and 48" high platforms. Comets are unpowered. If you want to replace Comets the unpowered cars in their portfolio might be a better choice.