The MTA buried its good news on a summer Friday last week as word got out that the L train would enjoy a massive frequency upgrade. Shortly after the initial amNew York story made the rounds, Transit issued a press release touting the service improvements, but since I was out of town for the weekend, I had a chance only to write up the bare bones of the story. So let’s delve in. There’s much to see here.
As we know, the MTA has been working on a communications-based train control system on the L line for years. As the L, at the time the test was first proposed, was not a particularly crowded subway and had the added benefit of sharing no sets of tracks with any other train line, Transit thought it had found an ideal testing ground. Over the past decade though, L ridership has gone through the roof, and the need for greater capacity along the BMT Canarsie line has gone from a luxury to an imperative.
Transit announced the official details on Friday. Beginning this past Saturday, the MTA started adding 98 weekly round trips on the L train: 16 each weekday, 11 on Saturday and seven on Sunday. In terms of wait times, the L trains will now operate every three minutes during the a.m. rush and every six minutes at midday, down from 3.5 and 7.5 minutes, respectively. Saturday evening riders will enjoy service every six minutes, and Sunday evening straphangers will find the same level of frequency, down from one train every 8.5 minutes.
“This is a perfect example of how our commitment of capital dollars to improve our signal system directly impacts the safety and quality of our service,” MTA NYCT President Thomas F. Prendergast said in a statement. “As a result of fully integrating Communications-Based Train Control on the L line, customers will have the added benefit of more trains that will help to ease overcrowding on a line that serves continuously growing populations in Brooklyn.”
Local politicians were excited by the improvements. Daniel Squadron, who has long fought for increased L service, led the charge. “Anyone tired of crushing crowds and overflowing trains will now have an L trip less likely to feel like hell,” Squadron said. “This is a big step toward a subway system that works for its riders every day of the week.”
One quote from Squadron though struck a chord. As he noted that some rush hour trains will likely be below the MTA’s load guidelines, he let slip a key line. “This is not going to be the silver bullet, but this is real good news for L train riders,” he said. “Anyone tired of the crushing crowds and overflowing trains will now have an L train trip less likely to feel like hell.”
On Friday, I noted the strange use of the silver bullet phrase, and as I thought about it this weekend, the stranger it became. What exactly does Squadron expect? What kind of silver bullet does he want? The MTA isn’t about to build a parallel line through Williamsburg or third-track the L train, and running trains every 180 seconds should be at least sufficient to ease some of the crowding concerns.
In a way, the CBTC implementation and the increased capacity shows what happens when the MTA has the money to invest in a systemwide upgrade. If the Queens Boulevard line can be converted to CBTC, if the Lexington Ave. IRT trains were CBTC-ready, the MTA could add significantly more trains to some very crowded areas. Of course, this all costs money and a lot of it. Even as the MTA says the increased service will result in only a $1.7 million jump in operating costs, the capital upgrades are in the hundreds of millions, and that money doesn’t grow on trees. Furthermore, the Canarsie CBTC implementation was, shockingly, years later and way over budget.
Daniel Squadron, one of the more ardent transit supports on the local political scene, knows all this, but he still wants a silver bullet. So what’s that silver bullet? A better run, well capitalized MTA? For the foreseeable future, that’s not a particularly feasible goal.
47 comments
Some recently built rapid transit lines aim for a frequency of 1.5 minutes. Every 90 seconds on the L would indeed be a silver bullet…
A silver bullet that costs a lot in rolling stock acquisition and train storage space. The MTA doesn’t have the rolling stock or physical space to store enough trains to run them every 90 seconds.
Terminal capacity is also an issue.
But CBTC has the ability to accomodate 90 second headways. Since terminal capacity is less of an issue on the Flushing Extension, it would be feasible there.
CBTC certainly cannot accommodate 90 second headways, except perhaps at a purely theoretical level. (But wayside signals can also theoretically accommodate 90 second headways.)
CBTC can certainly handle 40tph. As long as the terminals are built right, it can handle beyond 50tph. The challenge is managing such a system, as well as getting rail car design right. It does no good if the door and seat layout prevent very rapid and highly efficient passenger boarding and alighting. If you have high station dwell time, it greatly negates the signal system capabilities.
P.S. There are fixed block systems in Russia that can operate at 44tph service levels. The Bklyn Bridge once handled 60tph, with a complex high capacity terminal at Park Row and a series of lines at the other end.
Larry wants 40 tph in New York City. CBTC isn’t going to bring him 40 tph in New York City.
I’m not familiar with the Russian fixed block systems, but, like my objection to Stephen’s claims, do they actually achieve 44 tph on a regular basis, or is that merely the goal that they schedule for?
Trolleys operate on line-of-sight. without signal protection. That removes the capacity constraint that we’ve been discussing.
The rated capacity of the latest driverless metro systems in some European cities is a train every 75-80 seconds. None of those lines needs that capacity, but it’s feasible.
It would also be unnecessary. By NYCT rush hour guidelines (145 people per car or 1160 per train), a train every 3 minutes is adequate.
Flip to page 4 of the L report. It shows, in 2010 (the most recent year available), 22912 riders in the peak hour through the peak load point. At 17 tph, that’s 116% of the guideline. But at 20 tph, it would be just shy of 99% of the guideline.
Of course, I don’t know what’s changed since 2010, and the L is only getting 19 tph now, not 20. But a 12% increase is still pretty good.
What if L ridership continues to increase?
Then add service once the next shipment of cars comes in. A 90 second headway (if it were feasible) would only be appropriate if loads through the peak load point doubled.
That’s how guidelines work. Service is adjusted up or down based on train loads through the peak load point.
Peak within peak service demand might warrant a service higher frequency for a short period of time. There is a spike in demand for those who want to make it to work in Midtown and Downtown Manhattan/Bklyn by 8:45-9:00AM
I’m not so sure the spike is as narrow as you suggest (some people work on 14th while others have to transfer to other lines). I assume that 19 tph means a solid 3 minute headway for a bit less than an hour rather than a slightly longer headway for the full hour.
The desired arrival time at destination is probably within that time frame, but as you state the peak demand gets spread out on the L. Some work on 14th st, others go north to Midtown, or even Upper East and West Side or south to Downtown Manhattan/Brooklyn.
Dont forget the flushing cbtc too.
A silver bullet, is by definition, instant, simple, and magical. I believe all the Mr. Squadron was saying is that this “flipping of the switch” won’t instantly solve all of Williamsburg’s transit problems but it’s a step in the right direction. I wouldn’t read much more deeper into it than that.
That’s how I read it also.
The ‘silver band-aid’ for now would be to boost M service, at least during rush hours. That could alleviate some of the L crowding by giving people traveling to/from locations between the Canarsie and Broadway lines in Willamsburg the option of using a more frequent M (though to really get people attuned to the B’way-B’klyn option, extending weekend M service past Myrtle would be better, since it would keep the non-railfan riders from having to remember what days the M does and doesn’t operate to Manhattan).
Long-term, the silver bullet is going to be taking the mistakes made in putting CBTC in on the L and not making those same errors again when CBTC comes to the 7. Combined with the non-proprietary components, that will be the better test to see if the MTA can roll out a system at a lower cost, in less time, and one that isn’t locked in to technology that becomes outdated before it’s fully implemented (which will be another future problem for the L — it doesn’t matter if the 7 is running a different CBTC system from the L. But if the Flushing Line’s system becomes the default for the A and the B divisions, the L will be locked into the 143s and whatever handful of R-160s or any R-2xxs the MTA opts to buy that will be the only trains capable of running on the Canarsie line).
[…] New L Trains to Be Accompanied by Time-Saving Signal System (Transpo Nation, Kabak) […]
Got down to the platform this morning. Hundreds of people standing there. Waited about 10 minutes. Left the station and walked to the J. Cheers to the service increases!
Does anyone know what happened on the L train this morning, anyway? There were announcements that an earlier incident was causing trains to run at lower speeds…
When I got down to the platform this morning I just missed the train by seconds and the platform was clear. 5 RP/MW trains, two Manhattan bound trains that didn’t stop, and 25 minutes later, I was finally able to get on a train. I felt bad, though, for the people standing on the opposite edge of the platform with hundreds of people in front of them.
One option which could be considered a silver bullet, which has been noted here before, is to extend the Eighth Avenue tracks just 100 feet westward so that trains could enter at higher speed, shortening turn around time and allowing for more frequency. But in addition to operating more trains more frequently, perhaps the MTA could think more on the lines of improving access to and from the station platforms. 8th avenue in particular could use another set of stairs between the L platform and the ACE platforms, as the current staircase constantly gets congested by commuters. The new staircase could be on the west side of the 8th avenue line to supplement the current staircase on the east side. Other stations too could benefit from having more entrances.
It would be great to see the L train’s platforms extended to 600′ to accomodate 10 car trains. However, as some have noted here, this would be quite an expensive undertaking as not only would it involve extending every single platform, but also expanding canarsie yard and store these trains. Eminent Domain would likely be necessary in order to expand the Canarsie Yard. Still, if Williamsburg and Bushwick continue to grow in the next 20 years as they have the last 20, then these options might have to be seriously considered.
I was under the impression that with ATO, trains do in fact enter Eighth Avenue at speed.
That’s certainly been my experience.
Mine too. Slight slow down at the very end vs. a normal stop, but barely noticeable.
Prendergast credits CBTC as making 3 minute headways possible. Can someone please explain why other lines can operate at 2 minute headways with the standard block signals. I can see CBTC making 90 second headways possible which is not even being planned, so why does he make it appear that this improvement could not have been accomplished without CBTC?
The low velocity small radius curves right outside of stations, subpar terminal designs, and safety rules make for lower than average train frequency. Those low velocity curves make station dwell time (and rail car layout) a premium.
Another issue is fleet size. If you want greater frequency, you need a larger fleet, or far fewer spares, and maybe more manpower. That gets to yard capacity and work rules.
Which other lines operate at 2 minute headways? The closest you’ll find is the Queens Blvd. express, shared by the E and F, with two north terminals and three south terminals.
The 6th Ave Exp used to run 30tph when B,D,Q all ran down that track. You could stand at Broadway-Lafayette and see trains enter the station before the last one completely left at the other end. The Flushing Line does run at 2min headway unintentionally when trains get bunched up. You can see it at Queensboro Plaza. This also occurs at 60th St-Lexington Ave with the N,R,Q.
The B, D, and Q didn’t run 10 tph until the full Manhattan Bridge reopening in 2004. When they all ran on 6th, service was frequent but it wasn’t 30 tph.
When lines are recovering from problems, trains always bunch up. That’s one of the reason service is never scheduled at the absolute theoretical capacity of the line.
There’s only one pair of tracks with service scheduled at 30 tph, where service actually runs at 30 tph on a regular basis. That’s the Queens Blvd. express.
Equipment acceleration and braking rates determine service level capacity, not the signal system. If on bothers to look back nearly 60 years ago, the TA reported that they were running rush hour trains every 2 1/2 minutes on the 14th Street Line. That’s 20% more trains per hour than the “new, improved, CBTC enriched” rush hour operation being touted by the TA’s press releases.
Had the TA simply kept the old “obsolete” signal system, 14th Street riders could have had service levels adjusted to increasing demand. All the TA had to do is reassign cars on less crowded lines to 14th Street. However, going the CBTC route removed this obvious overcrowding solution. In order for a CBTC system to work, all trains on the line must have working CBTC equipment. The presence of a single non-CBTC train forces the CBTC system to revert to “safe mode” operation of 1 train every 20 minutes.
The TA did not require conversion kits to retrofit older rolling stock for CBTC, when they embarked on this program more than a decade ago. Only new rolling stock would be CBTC compatible. The TA did not purchase enough CBTC equipped rolling stock to meet the requirements for what was already Brooklyn’s most crowded line.
The TA has been starved for funding. However, they squander money when they do receive it. The CBTC implementation for the 14th Street Line has cost several hundred million dollars and has resulted in less service level capacity than what it replaced. Two other examples from this era are the new South Ferry Terminal and the Fulton Street Transit Hub. Over 1 billion dollars will be spent on both these projects without adding a single rush hour train.
Had the TA simply kept the old “obsolete” signal system, the 1920’s signals would have failed at an increasing rate and would have cost more and more to maintain. As I said yesterday, the reason the signals on the Canarsie line were replaced was that they were old and in need of replacement. The reason they were replaced with CBTC rather than with new wayside signals was to serve as a testbed for future CBTC installations. In the long run, CBTC has plenty of benefits over wayside signals – greater capacity, higher speeds and more fine-tuned speed controls, more consistent train operation, less wayside equipment – to the point that sticking with wayside signals forever doesn’t make sense.
As was proven a few years ago, in addition to the R143’s which make up the bulk of the L’s fleet, the R160’s are compatible with CBTC. When it became clear that there weren’t enough R143’s to serve the needs of the L, eight R160 trains had CBTC equipment installed. Retrofitting older rolling stock is pointless – East New York has only a few trains of older rolling stock, and they can continue to run on the J until they’re retired. The R179 order includes some extra trains for East New York.
I’ll add, I have extreme difficulty believing that the stated service levels in that graphic ever operated in practice. Many of the capacities are also much too high to be practical.
What’s the basis for your doubts?
The graphic was included as part of the TA’s First Annual Report in 1954. It shows both theoretical and ACTUAL am rush hour service levels. There are also timetables from that time that corroborate these actual service levels.
I have no doubt that service was scheduled at those levels. I have serious doubts that trains actually ran at those levels. You’ve shown us a graphic based on paper schedules, not records of how the trains actually operated in practice.
The theoretical signal capacity is not directly relevant in real life, since it assumes that trains are spaced perfectly evenly and that all train operators (excuse me, motormen in 1954) always operate at the maximum speed allowed by the signal system (including keying by red automatic signals). That’s not realistic in practice. The practical capacity is lower.
Had trains been lengthened to their current size already?
As development continues, more and more people are being added to the L. And development is continuing. There are large scale projects along the L through Myrtle-Wyckoff, and projects of other sizes too.
The M is not much of a viable alternative. Even if headways decreased, it is much slower, and it doesn’t go through the heart of the development anyway.
So additional time was bought before the L once again is overcapacity.
My impression was that the MTA’s interest in CBTC was always more about the potential manpower savings rather than increasing service capacity. But since the unions have mostly blocked single-operator trains, it doesn’t look like reduced operating costs will be realized any time soon.
There are a number of other relatively easy things that could be done to make L riding more pleasant: construct entry/exit points at Avenue A, for one. The stairwells at Union Square are another major choke point. I’m not saying those would be simple fixes (as I recall Ben had a post a while back looking at the various challenges of adding an Avenue A entrance), but they’re a hell of a lot easier than building an express track.
If we all live along enough, maybe one day we’ll see the L extended west and north to meet up with the extended 7 somewhere in Chelsea.
OPTO is not contingent on CBTC. But CBTC does generate manpower savings in the long term – trains move faster, so fewer trains are required to provide the same frequency; as with ATS, switches are computer-operated, so there’s no need for tower operators; and the signal system, once mature, will probably require less maintenance.
I’d love to see an Avenue A exit. (An express track isn’t happening, and an express track wouldn’t do anything for capacity in any case.)
Just how does CBTC permit trains to move faster?
Current end-to-end operating time is 40 minutes, up from 35 minutes in the 1970’s.
But isn’t dwell time also longer today because of more passengers? But I’m still not convinced that CBTC was responsible for the 3 minute headways. It was just something they wanted to do now and it can be used to justify the expense of CBTC.
As I’ve already said multiple times, the purpose of installing CBTC on the Canarsie line was not to support 3 minute headways.
The old signal system could accommodate 3 minute headways as well. And when the Canarsie line was picked for CBTC, nobody anticipated that it would need 3 minute headways, certainly not so soon. Ridership grew much faster than anticipated. The R143 car order, which was supposed to be large enough to include a few trains for the M, wasn’t even large enough for the L alone!
When the line was picked for CBTC in 1997, its signals were old and in need of replacement. To set the stage for CBTC (with its capacity, speed, and cost benefits long-term) as the general approach for most new signal jobs going forward, the Canarsie line was selected as the testbed, partly due to low ridership.
Why do I feel like I’m repeating myself? Have you read the L report yet?
Have you actually ridden the trains? They move much faster now than before CBTC.
From the long waits to get into the terminals, it’s obvious that the schedule hasn’t been rewritten to adjust the running time since full deployment of CBTC.
And a lot aside from the basic signal system has changed since the 1970’s. Train operators are no longer allowed to key by red signals, car performance was slowed following the 1995 Williamsburg Bridge crash, etc. You’re looking at too wide a time frame – it includes a lot more than the implementation of CBTC.
And what appears in a paper timetable is influenced by what the agency wants to promote. Does it want to promote the impression that the trains are fast? (Then the running times in the timetables may be underestimated.) Or is it more concerned with achieving high OTP? (Then the running times in the timetables may be overestimated.)
Searching for the L train’s ‘silver bullet? All you need to do is order beer from Golden Colorado.
Will they discontinue turning around emptied trains at Myrtle/Wycoff?
Unlikely. It’s for when there is maintenance work being done, and it’s one of the few places to turn a train.
I think Deb Parise is referring to scheduled turns during rush hours.
I have to assume some trains are still being turned at Myrtle, if only because service was just increased, and running all trains to Rockaway Parkway would require even more cars.