SFMTA Transit Director Predicts Better Muni Metro Service in Coming Months

A problem with signal cables being torn by sander hoses on the trains is one reason for all the recent Muni Metro delays. Photo: ##http://www.flickr.com/photos/timmy/4598573869/##Timmy Denike##
A problem with signal cables being torn by sander hoses on the trains is one reason for all the recent Muni Metro delays. Photo: ##http://www.flickr.com/photos/timmy/4598573869/##Timmy Denike##

The series of frustrating and consistent delays that have plagued the Muni Metro system for the last few months have been narrowed to three causes that SFMTA Transit Director John Haley told the SFMTA Board yesterday have mostly been fixed for now.

Still, according to Haley, it’s “too early to declare victory” on a problem with the sander hoses under the Breda light-rail vehicles that have been tearing the automated signal cables that sit between the rails.

Haley, who updated the board at the request of SFMTA Director Malcolm Heinicke, explained that the vast majority of delays have been caused by signal failures with the Automated Train Control System (ATCS) when the trains have to switch from manual to automatic as they exit the surface streets and enter the underground tunnels at West Portal, the Ferry Building and Church and Duboce streets. Problems with the ATCS have persisted since Muni first acquired the system 11 years ago.

On average, the ATCS failed to get a signal from trains more than 200 times each month in September and October. Every time that happens, the system “fails safely,” and slows down every line.

“The cause of that is either something in the wayside or something in the vehicle,” said Haley. “The impact of that from a service perspective is when the train’s not recognized it puts it in manual. There’s a restrictive set of procedures that must be followed. The safety spacing is increased so the system is slowed down.”

Haley said the agency has been working in concert with the signal manufacturer, Thales, to try and resolve the problem. Last month, the ATCS got a software upgrade and since then the signal failures have decreased by more than 50 percent.

“What that tells us is that the software fix is working, that the trains are accepting and getting automatic train control. In addition to that, what we need to continue looking at is the maintenance of the wayside equipment and other causes on the vehicle that may cause a train not to get the signal system,” he said.

SFMTA spokesperson Paul Rose said they are also working with Thales to do a more thorough upgrade to the ATCS.

Another problem that has proven just as difficult is sander hoses underneath the trains slicing signal cables. It’s happened an average of a dozen times each month in the last few months. The sander hoses are used to drop sand on the tracks in the event of an emergency brake application. For the fourth time since Muni acquired the Breda light-rail vehicles, 130 trains currently in operation have now undergone a sander hose retrofit, but there have been at least three slicings in the past 10 days.

“I’d like a little more time to be able to say that we’ve got the problem cured and this is the right fix for us so we’ll continue to work with vehicle engineering on that,” said Haley. “The service impacts of this have been very, very difficult because in the situation I described delays can extend for long periods of time.”

Eventually, said Haley, the plan is to replace all of the signal cables but it doesn’t make sense to replace them until the sander hose issue is fixed.

Another problem causing the delays is “disturbed blocks.” The system recognizes a train by counting axles, explained Haley, “so if there’s six axles in a particular block the system knows there’s a train there.”

“On occasion, a block will go disturbed and what that means is that there’s something, either a piece of metal, a piece of a track break or some other metal that could be loose or hanging from the undercarriage of the car that tells the intelligence of the system that you’ve got another axle in there. In that particular case what happens again is the system ‘fails safe,’ you go to manual and you take a train and you clear the block,” he said.

Although this problem hasn’t caused as many system-wide delays as the other issues, Haley said it happened on average 50 times each month in September and October.

In earlier interview with Streetsblog, Haley said he had also been challenged by the special events that occurred over the past few months, including the extra service that was added for the Giants games.  On the plus side, since LRV service was fully restored in September after the St. Francis Circle Project was completed, operator availability has been at 98 percent and the agency has been meeting its car availability. He said the line management center has improved the overall management of the system and he’s been careful not to over-schedule trains.

“I think the situation would have been much worse from a service delay perspective had we not been more aggressive in the way we’ve been adjusting service,” he said.

Heinicke thanked Haley for his presentation and for confirming the delays “were not just an odd coincidence for me,” but he pressed him on whether service would get better in the next few months.

“The riders I encounter daily, the people who call or email me, they want to know: is the system crumbling and are we going to have to expect more of this or were September and October an aberration in part because of problems that you had started fixing long ago. A more direct way to ask that question is, will the LRV service in November and December of this year be better than it was in September and October of this year?”

“Okay, I will stretch the crystal ball here but I would say yes,” Haley responded

  • Alex

    Lipstick on a pig… Haley is merely putting lipstick on a pig here. Operator availability has been at 98%? BS. According to the daily service reports[1], the Green (LRV + F-line) division rarely saw absenteeism of less than 23% (Q1 for September = 23.8%, October = 22.9%)[2,3]. Even *if* you only take into account AWOL drivers you’re not going to see availability of 98%.

    Meeting your vehicle availability would be impressive if demand weren’t grossly underestimated. LRVs have been siphoned off the JKLM lines to fill in for the T… we’re back to packed one car trains in the evening and weekend hours on the L. That’s not much to crow about if you ask me.

    FFS, the MTA finished the St. Francis Circle project… and for all of that disruption LRVs still do *not* have signal priority. That means waiting at a light for ~5 minutes depending on how the LRV driver times it.

    Also: Thales vs Thalus (altho their products do seem as if they’re designed by a phalus).

    1: These reports also *used* to contain detailed information about ATCS failures by vehicle, but that information has been scrubbed by Haley so we just have to take his word for it. In 2006 you’d typically see about 1% of all vehicles fail to sync up with the ATCS upon entering the subway. On a bad day, however, you’d see 20-50% fail. How about a bit more transparency Mister Haley?

    2: http://farm2.static.flickr.com/1138/5143528869_88711b2b38_z.jpg
    3: http://farm2.static.flickr.com/1372/5143514003_74c62776e7_z.jpg

  • Jon

    A few suggestions for improving the Metro:

    1) Create exclusive lanes for the above ground portions, which are available for use by LRVs only, even if it means reducing vehicle traffic to one lane. (You could also allow Muni buses to use the lanes if it can be done safely.) If they can’t do that at the very least make the area in front of the stops Muni only, so that trains don’t have to stop twice at intersections because there’s traffic waiting at the lights preventing them from picking up passengers.

    2) Install signal priority at every above ground intersection, so trains rarely have to stop at lights. This works best if the stops are located at the far side of intersections.

    3)Install level boarding platforms with pre-pay ticket machines at every above ground stop. This will probably mean losing some parking spaces. Remove/consolidate some stops as well.

    All of the above changes are likely to be implemented on Geary as part of the BRT project. Maybe once SFMTA planners see it working there, they’ll look at doing it on other sections of the ‘rapid’ network. The F-line on Market could also benefit from these improvements, although level boarding is probably not possible given the different shapes of the historic streetcars.

    4) Run all trains as two-car. It costs the same in labor to run two-car trains as it does to run one-car trains.

    5) Allow double birthing in Muni subway stations. Maybe have the N and J stop in one place and the K,L and M stop in another; that would minimize people running up and down the platform to catch their train.

    Just some thoughts. Ultimately running a train system with a portion in mixed traffic as well as a subway portion is always going to be a nightmare, but I can’t see the city coming up with the money to convert the system into a proper subway any time soon.

  • Alex

    Jon,

    1.) Disagree. With the M (and presumably T), at least, most of the problems are signal related. I’d much rather see boarding platforms (high or low) and separation from traffic than simply moving LRVs into their own ‘protected’ lanes.

    2.) Yes.

    3.) Yes. Even if the MTA won’t put TransClipper machines at every stop (and given the light usage of many metro stops I think this is a waste of money), they should put grade separated platforms of some sort at every stop. Buses can pull curbside, LRVs simply cannot.

    4.) Yes — my guess is that there simply aren’t enough vehicles to run two-car service on the N, L, and M. Take a look at the daily service reports. On the J, K, and T I think (but am not sure) that there are a number of places where two car vehicles can’t safely offboard passengers. Certainly there are a few L stops that are unsafe to use from the second car.

    5.) I’ve seen this batted around a few times by the MTA, but it will take a rewrite of the (defective by design) ATCS software.

    Aside from lack of signal priority, most of the problems I’ve encountered with the metro have been underground or at West Portal proper. Likewise, a proper shuttle bus plan should be drafted and adhered to. As-is chaos ensues whenever shuttle buses are brought out.

  • Jon

    Maybe you’re right about 1). Third St is wide enough for dedicated LRV lanes as well as platforms and several vehicle lanes. I’m not sure if all the streets the LRVs run on are wide enough to do that, which means platforms or a dedicated lane is the choice. In that case I agree platforms are a priority.

    After a trip on the L last night I have a clarification for 2). As well as installing signal priority at all the intersections with traffic lights, change the intersections which don’t have lights (four-way stops) so that LRVs never have to stop at a stop sign. Change them to light controlled intersections with signal priority, maybe reverting to yellow-flashing four-way stops after the LRVs have stopped running, and use sensors in the road to help prevent cars from having to wait unnecessarily during quiet periods.

  • Alex

    As for #1, platforms aren’t a necessity. Some sort of separation from traffic is. Take a look at the 23rd Ave stop on the L. The tiny little strip is meager, and far too small (both length and width) but it’s better than nothing. Offboarding into traffic is absurd and simply unsafe.

    As for signals, no. Stop signs are an effective traffic calming measure, and traffic moves fast enough along Taraval and Judah already. The traffic signals are at busier intersections, thus signal priority would make a much bigger difference than replacing stop signs with sensor signals. Giving the LRVs priority at St. Francis Circle where you measure the wait in *minutes* is going to make a much bigger difference to *two* lines than ripping up some stops on Taraval or Judah. Likewise giving LRVs and buses priority at 19th Ave would make a much bigger difference than ripping up a few stop signs.

    Forcing a stop for LRVs makes it easier to catch a train. Additionally, most of the stop signs on Taraval at least are aligned with LRV stops. There are non-revenue stop signs (and those should be removed/reconfigured like on 15th Ave), but they’re not the majority. That said, the non-revenue stop signs should be removed and the stops reconfigured. I’m sure you can easily find my rant about the 20-24th Ave and Taraval stops.

    Speaking of sensor lights, another modification that IMO should be made is at 9th and Judah / Irving. The LRV cycle should wait until the LRV has cleared the intersection before switching to the auto or pedestrian cycles.

  • Jon

    Sorry, but I can’t agree with you about four-way stops. Maybe it’s because I’m coming from the UK, where four-way stops are non-existent, but every time I see one of those LRVs grind to a halt just so the driver can look both ways and start again I can’t help but think about how nuts it is. We want to calm motor vehicles, we don’t need to calm the LRVs.

    Regarding your point about stop signs at revenue stops, not every revenue stop will have a passenger waiting at it, so there’s no need to stop at them all. Ideally these stops should be on the far side of intersections with LRV priority traffic lights.

    I take your point about platforms though. Level boarding platforms would be nice for disabled people, but the main problem is passengers having to off-board into traffic. Any sort of waiting area is better than none.