Over at Portland Transport, EngineerScotty (also author of the Dead Horse Times) posted on improving transit speed downtown, particularly for MAX. It’s an interesting post and a lot of different ideas have come up in the comments. I was going to respond there but it got long, so I’m taking it here and breaking up my thoughts on the different suggestions that have been made.
The first of these…
Train Length
One of the constraints of MAX brought up in the post is train length – Portland city blocks are about 200 feet, and a two-car consist is about 184 feet (191 feet if it’s a Type 4). All lines run through downtown, so the system is designed around that 200′ maximum length for trains. Early on in the thread, one commenter asked why we couldn’t run a train that’s twice as long (a four-car consist rather than the two-car consists run now) – even if it blocked a street while it serviced a stop, it wouldn’t be there long and this would double the capacity of service.
The exception, not the rule
Mechanically speaking and not taking anything like platforms into consideration, the cars are capable of being coupled together in consists longer than a two-car train. I haven’t really posted about how cars are coupled aside from answering questions in comments, but the trains are coupled in two ways: a mechanical couple and an electrical couple. The mechanical couple is what physically holds the cars together; the electrical couple is what lets the cars talk to each other. For example, this allows the operator to hit the door open button and have all the doors in the train open, not just the doors in the car that the operator is sitting in (this is called “trainlined” and yes, that’s where the safety communication gets its name). That works if there are two cars coupled together, or three, or four. I don’t remember if more than four cars can be electronically trainlined. This does not work for Type 4s. The coupler head located under the cabs of those is there to be used for a dead car tow or push and is capable of being mechanically coupled to any car in the fleet, but there will be no electrical communication between them.
Screen shot of Bob R’s video of the A-cab coupler head
So aside from the 4s, more than two cars could be coupled together and still function. However, there are a number of reasons why it would take so much money in construction costs to run 3-car or 4-car consists to the point where it’s just not worth it.
For one, the previously-mentioned trainline opens all the doors of the train. Assuming you have a four car consist downtown, if the operator stops to service a platform (we’ll use Pioneer Square North as an example), the rear two cars are going to be blocking SW 6th and going back up the block between 5th and 6th. When that operator opens the doors, all of the doors in the train are going to open, and remember that even on the low-floor cars, there’s a drop to the ground below when not at a platform:
Climbing into a Type 2 from the ground
So that would be opening the train doors onto the street, and even for people not using mobility devices, that’s not a comfortable way to get on or off the train. And to lengthen all of the platforms in the system to accommodate longer trains would be prohibitively expensive (just the Washington Park stop alone would be a logistical and financial nightmare)
There’s also the matter of what to do when the train gets to the end of the line.
In the Jackson turnaround
Here at Jackson St, which is currently the end of the line for Yellow and Green trains, the first and third tracks are big enough to accommodate a two-car train, but nothing larger. The circuits in the turnaround are only big enough for one two-car train. I took this picture from the leading car looking back toward the trailing car, and the last wheel axle of the trailing car is just past the insulated joint on the eastern entrance to the turnaround. And the center track can only accommodate a single car train, such as the mall shuttle. In short (pun not really intended), there’s no room for a train longer than two cars here.
So that means no four-car consists on the Yellow-Greens, which is good because that would make things much more difficult for buses driving on the transit mall. What about on the Blue line? Cleveland has a tail track, so there actually is room at the east end of the line in Gresham. Heading out to the west side though, there’s a lack of space. Here’s a view of the platforms at Hatfield Gov Center, the western terminus of the Blue line:
Western end of the Blue Line
As Hatfield is now, there’s no room for a train longer than two cars – to lengthen the platforms would mean shutting down Main Street which runs behind the building there.
It’s not just a lack of space and platforms big enough to accommodate them that that make it impractical to run longer consists.
Paradoxically, longer trains would actually mean slower running speeds in many sections of the alignments. At Goose Hollow (above), for example, the speed limit around that curve for eastbound trains is 10mph, and a train can’t accelerate until the entire consist is clear of the curve. You get thrown around quite a bit if you’re near the back of a trailing car going around a curve and the operator accelerates before you’re out of the curve. If the trains were twice as long as they are now, that’s waiting until another 200′ of train has gotten through a curve before the train can accelerate.
In other places, gravity would work against longer trains. For example, heading into the tunnel westbound, the speed limit is 55mph past the first cross passage. As things are now, if your train is a two car consist with a crush load of people, it’s hard to get to 55mph since you’re climbing a hill with all that weight. If you’ve got twice as many cars and people, it’ll run even slower. Longer trains might mean more capacity, but ultimately they’d mean slower running speeds.
So it’s an interesting idea to run longer trains, but it would involve so much construction to existing platforms, major modifications to city blocks in the CBD, to say nothing of the work involved in changing the circuits in the rails to accommodate longer trains that it’s not feasible to do.
More to come.
Camelopardalis
MAX FAQs 
Great post… a few comments:
I think it would be worthwhile to at least make a formal inventory of stations and create ballpark estimates on what’s needed to lengthen platforms, move tracks, move substations (if needed), etc., so that when the time comes (and it will come) to build a subway or other means of rapidly crossing downtown, we can decide if it’s worth it to expand some or all of the system to accommodate longer trains. (And perhaps compare prices for a 3-car vs. a 4-car system.)
As for the Hatfield station you point out, I think with a bit of creativity 3-car trains could be handled here without changes to Main St. There is space in front of the current platforms, if you move the location of the pedestrian crossing to where the track bends, for 3-car trains on two of the tracks. The problem is that the interlock gets covered, rendering the 3rd track useless. But you could move that over and take out the trees and curb extension at the right side of the picture, and for relatively little effort have a station that can handle 3-car trains.
The Zoo station may indeed be a problem, simply because the excavation work may require a shutdown of MAX operations through the tunnel, not a good thing for such a busy facility.
But if you imagine a future in which there is an east-west subway for “express” service across downtown, while maintaining today’s surface tracks for “local” trains, then “express” trains could simply skip over the Zoo station and go all the way to Sunset TC. Schedules would have to be carefully planned so that “express” trains don’t get caught up waiting for a “local” to leave the Zoo platform, of course.
Finally, you wrote that “in other places, gravity would work against longer trains” and based this statement on the difference in acceleration between light and heavy passenger loads.
But I don’t think your conclusion works for longer trains, because I assume the additional cars would be powered, correct? In which case, each car is responsible for carrying it’s passenger load. Its the heavier load per-car that’s slowing you down, not the number of cars. (Or is there a factor I’m not considering where a 2-car train, for example, doesn’t have as much power as two separate 1-car trains, and therefore a 4-car train has less power than two 2-car trains?)
Interesting thoughts on Hatfield – of course, there too the signals, call loops, etc would all have to be moved.
In a two car consist, both cars are powered, but a two car train does handle differently from a single car, in that it feels as though more power is needed to climb a hill. That’s more noticeable in some sets of cars than others, but it does feel different. And actually that makes me think of something else I didn’t put in the post – some cars play better together than others (like Type 1s and Type 3s coupled together can feel kind of rough sometimes since the 3s were kind of “dumbed down” to trainline with the 1s), so those problems would likely be compounded with longer consists.
Which then makes me think that from a service perspective, longer consists (at least with the existing fleet) could be problematic for people who can’t climb the steps of a Type 1 and need a low-floor car. With current service, it’s not advertised as much as it should be, but it’s generally best when people who need a low-floor car wait in the middle of the platform – that way even if one car of an incoming train is a Type 1, they don’t have to go far to get to the low-floor. How would the Type 1s be doled out in this 4-car consist idea? One per train? More? That could make it more difficult for people who need ADA accessibility. Even if it’s just one Type 1 per train, it wouldn’t be as clear where people who need a low floor should wait.
Thanks for the info re: multi-car consists feeling different, needing more power.
Regarding what combinations would work out of the current fleet, I’m operating from the assumption that increasing train lengths would require ordering more cars/consists, and that 2-car trains would continue to serve in some capacity (either as “locals” as described above, or relegated to more lightly-used routes.) So at the time of procurement of more cars, a model could be ordered/customized which offers maximum compatibility with existing rolling stock.
After all, at peak time, most of our existing cars are already in service. Just bunching/reorganizing the existing fleet into longer units may buy some operating efficiency (fewer operators), but it does very little for overall system capacity.
The way I view the issue of longer trains is that eventually we will reach minimum operating headways with 2-car trains and those trains will be full (at least at peak). Therefore, any hope for increases in transit ridership will have to either come from new parallel lines which absorb some of that service (MAX on Powell might take some of the load off the I-84 segment, but not much), longer trains, or bypasses to the street grid (downtown subway) and enhancements (new river crossing to bypass Steel bridge) which facilitate operating at tighter headways than we do now.
Ah, okay, I was looking at this from the perspective of Bjorn’s first comment on the Portland Transport post, which I read as meaning that we could double the consists now and the only issue would be a temporary obstruction of streets in the CBD. And as things are now, there’d be more issues than just that one. Future planning, absolutely there’s a lot more flexibility in designing things differently. But this post was just more of an answer to “why can’t we do this now?”
PS… I really do think that in future expansions of the system (southwest corridor, for example), a future with 3-car or 4-car trains should be engineered into the design. Even if platforms are only built-out to 2-car trains, and terminus trackage is only installed in 2-car lengths, the ROW, excavations, location of outbuildings, etc., should allow for easy upgrades in length by future generations.
If we really expect light rail to be the spine of a future densified, urbanized Portland metro area with the corresponding increase in transit mode share compared to autos (not to mention our greenhouse/transportation energy goals), then MAX must eventually offer the capacity to carry those passengers.
The idea I want explored is skip stop operation with passing tracks. We have a High Capacity transit study and now we need a High Speed transit study from Metro.
Bob R.: Therefore, any hope for increases in transit ridership will have to either come from new parallel lines which absorb some of that service (MAX on Powell might take some of the load off the I-84 segment, but not much), longer trains, or bypasses to the street grid (downtown subway) and enhancements…If we really expect light rail to be the spine of a future densified, urbanized Portland metro area with the corresponding increase in transit mode share compared to autos (not to mention our greenhouse/transportation energy goals), then MAX must eventually offer the capacity to carry those passengers.
Or…the realization that MAX was improperly engineered and the cost to correct it will be too great; and that the best, most cost-effective solution is to start planning for bus services TODAY that will accommodate growth in ridership; while transitioning MAX away from the hybrid-Streetcar/Interurban role and into more of a true interurban role of rail transport, eliminating the close-spaced stops especially in downtown Portland, Beaverton and Hillsboro. Let buses – in particular, high-capacity, hybrid-electric articulated buses, buses that are designed for both high passenger loads AND obtain their greatest efficiencies in stop-and-go traffic – serve the local riders. An articulated bus carries 40% more riders at virtually no additional operations cost, but a two-car (or three-car, or four-car) MAX train incurs significant additional operations costs.
The under-engineering problem also exists on WES, where to rebuild WES to accommodate “true” commuter rail trains would require rebuilding every single station, and completely relocating the Beaverton station (or building on top of a protected wetland and having to demolish and completely rebuild the Beaverton Creek bridge.) And the Streetcar vehicles can’t be coupled together. But there is virtually little or no cost involved in deploying articulated buses (especially given that so many TriMet bus stops are currently not meeting even the most minimum of basic requirements.)
If you look at the Portland Transport post, you’ll see that we discuss stop elimination/consolidation in great detail, as well as the role of MAX as a local vs. regional service.
Streetcar vehicles can be coupled together (perhaps with some modification, but the original design used is indeed capable of multi-car consists), and platform lengthening is not as complex as MAX, although it does remove parking.
As for the assertion regarding MAX that “the cost to correct it will be too great”, I don’t believe that platform extension and utility relocation for 4-car trains (which would result in a doubling of capacity) would cost nearly as much in the long run as providing that much regional capacity with buses or entirely new true BRT or light rail lines.
So, would you say that given the power issues and speed restrictions (because of curves and things), a single-car train can be faster?
In a way. It’s true that you can accelerate sooner after curves or coming off the Steel Bridge span as a single car (also known as the “sports car” train) since you don’t have to wait for a trailing car to clear which saves a few seconds. But platform dwell time can be longer if people aren’t expecting a single car (and people are never expecting a single car) so there can be a lag in giving people who were waiting to board the trailing car enough time to walk 100′ up the platform and try to fit themselves in. So overall I don’t think there’s a significant amount of time gained or lost running a single versus a two-car consist.
Unrelated comment, but why are the front windows covered on the type 4 trains (or really any of the trains?)
Mostly for glare on the windshield from the lights inside the train (and even though this isn’t Bus FAQs, that’s also why the bus interiors are dark in the front with the pink light). People are also less likely to needlessly tap on the windows if they can’t see through them – knocking to ask a question or if you need help is fine, but knocking the same way some people tap on fishbowls is not!
Can I ask why longer trains would climb slower uphill? After all, these are multiple units, so twice the weight would come also with twice the power, wouldn’t it?
See above in the comments, Bob R & I were tossing this one back and forth. Yes, both cars in a 2-car consist are powered, but operationally it feels different, as in it feels like more power is needed. I’m not sure if that’s a function of how the cars electronically communicate with each other, which might explain some of the variability in that (at least in my opinion) it feels more noticeable with some sets of cars than others.
Car weights may also be a factor – Type 1s are lighter than Type 2s and 3s, which is very noticeable if you need to do a dead car push/tow up a hill and the Type 1 is the only car that works.
Thanks for the explanation and sorry that I didn’t read the comments without commenting.
Interestingly enough, there can be situations when multiple units coupled together mean a slightly better acceleration: in fact, all forces are doubled, except for air resistance at the front of the train, of which there is only one. This drag effect is just like in case of e.g. a peloton of bicycles. Maybe this effect is only apparent at the high speeds of mainline trains.
Is there a reason TriMet isn’t ordering married car sets? I feel like more than 80% of the trains I board are 2 car sets.
I would imagine even ten more seats and 5ish feet of crush space would help and it’s works with most of the infrastructure(sans the shops, right?). They could have 5 or 6 articulation points and 5 or 6 trucks. I think the s70 has the option for a “permanently married” or a “uncouple-able married” set.
Also could the shop handle a permanently married set or would they have to uncouple the set?
Flexibility, predominantly… even though the 4s can’t be run as singles, if one car of a consist goes bad the other could be coupled to a different one and sent out again. And no, I don’t think the shop could be easily modified for a 2-car consist given the way the maintenance bays are set up. Aside from the shop, the Elmo wash rack is for single-cars only. Offhand I’m not sure else where a married consist would be an issue, but I imagine the changes that would need to be made to accommodate that are more work than it would be worth.
How do they wash type 4 cars if the wash is only built for single cars?
As singles – they are uncoupled in the yard and can be operated via a panel in the parlor cab.