Discussion related to commuter rail and transit operators in California past and present including Los Angeles Metrolink and Metro Subway and Light Rail, San Diego Coaster, Sprinter and MTS Trolley, Altamont Commuter Express (Stockton), Caltrain and MUNI (San Francisco), Sacramento RTD Light Rail, and others...

Moderator: lensovet

  by Palal12
A friend of mine, who works at planning at NYCMTA recently mentioned on another forum that BART has contacted NYCMTA to study ways, which can be used to keep trains rolling on time. Ironic, isn't it.... MTA can't keep their own trains running on time, but they can give suggestions as to how BART should do it :-) ). What do you guys think should be some of the ways BART should keep trains moving?

This is an easy one in my opinion. Retrofit the Concord line trains and add center doors. This would cut way down on dwell times in the market street subway during rush hour.


  by modorney
3 - door cars:

Kandyman has a good idea. If you dig thru old BART Srtp's, there is mention of the three door cars. BART probably won't retrofit the existing cars, but BART does have a car shortage, and will need to buy more cars for the Warm Springs extension.

At any given time, the Concord line has 14 trains running, and, during peak commutes, can have up to 20, or so. That's 200 cars - a lot of money (like half a billion bucks). Of course, there's no reason for all the cars to be 3-door, if every other train was a 3-door train, it would help. Keep your eye on PATH's new car purchase. They are buying 3 door cars, and BART could piggyback on their order. (PATH cars are shorter, but the door positions are about the same. BART wants to keep the 2 main doors 35 feet apart).

AATC part 1:

Right now, if a train is sitting in Embarcadero (southbound), and another train comes thru the Transbay Tube, it will stop, and wait for the train at Emba to leave. By the time the second train starts to pull into Emba, the first train has already opened at Montgomery. But, AATC http://www.tsd.org/cbtc/projects/#San%20Francisco will allow the second train to pull into Emba sooner.

Most train operators watch the train ahead, and wait to close their doors when the train in the next station closes his doors (yellow lights go out). You might as well collect passengers, instead of sitting in a tunnel, stopped. Unfortunately, one delayed train slows the whole crowd.

This part of AATC - in the tunnels - (which is being tested now), seems to be going OK.

AATC part 2:

AATC is designed for the elevated part, also. From West Oakland to Bay Fair, it gets crowded. This will get the trains moving quicker.


Right now, BART connects with over 200 other transit providers - everything from Amtrak to the Green Tortoise (yes - it still exists!) Part of the schedule challenge is fitting in between these connections. BART will cal other providers (like Caltrain and Tri-Delta Transit) and have them wait, but a more efficient system could be better.

Skipping stops:

Some runs (like reverse commutes) could skip stations. That would make run times quicker and more predictable.

  by modorney
How PL's work

BART uses a central control computer to keep the trains on time. There a few things that can be done, to keep things on time.

Most of BART's mainline track is rated 50 or 70 mph. A few stretches (long, straight tunnels, etc.) are rated 80 mph. Each section of track has a speed code on it. Think of the traffic loop, on a side street. You drive up, and stop in the loop, and the traffic light changes. The rails are connected to a signal generator, just like the traffic loop, and a speed code is sent to the train. Antennas on the train pick up the signal. As long as the train you are on is being controlled by central, you will have these signals - look in the cab and the speed codes are the big red numbers (command speed). This number is the same for that stretch of track (unless there's a train up ahead, or some other speed control, and you will get a smaller number, or zero).

The train, itself takes that speed code and modifies it. This modifier is called a "Performance Level", or a PL for short. There are six performance levels - PL1 to PL6. PL1 is the fastest, PL6 is the slowest, and most of the time you are on PL2 - the normal setting. There are some subtle differences in the PL's - Slower ones have weaker acceleration, and braking, etc., but the biggest difference is top speed. If you look in the cab, there is a small liquid-crystal display, which shows the PL level, as well as the operational speed (which is the big red command speed, modified by the PL). If the light, and the contrast knob of the LCD are just right, you can look in and see the LCD, with the PL and operational speed displayed.

Most trains run at PL2. A while ago, trains ran at PL1 all the time, and the mechanics were replacing brake pads all the time, from the extreme braking. And the trip from Fremont to Embarcadero was 3 minutes quicker than it is today. So, to save money, PL2 became the standard. (Along with some minor software tweaks to all the PL's).

So, when the Train Operator pushes the button to close the doors, and before he leaves the station, the wayside computer figures out how fast he needs to go to get to the next station. Then the wayside computer sends the proper PL code to the train, and that is used to determine the operational speed to the next station.

Typically, a train is pretty close to schedule, so you will usually get PL2. If you are a bit ahead of schedule, you might get PL3, which is a bit slower. If you are really late, you might get PL1. Occasionally, after a long run, followed by stations close together, you'll get a PL4,5, or 6. This happens a lot coming from West Oakland. You'll come into Embarcadero, and then be super early. To slow you down, and put you on schedule, you'll get a PL6 to Montgomery, which is only a third of a mile away. You can notice slow PL's - the train accelerates slowly. Then, by Civic Center, you'll be on PL2 again.

The other way to control the schedule is the door dwell. To fine tune the schedule, the central computer will keep the doors open longer. When the computer says "it's time to leave", an electronic beeper starts beeping, and the train operator looks backk to see if the doors are clear, and he closes them. It's not instant, it takes a couple of seconds, so don't feel like he's out to get you when you jump out from behind a post. He had already pushed the button, before you were even ten feet from the post.

So, if your train is a bit ahead of schedule, you will sit there, with the doors open. So, when you are sitting there, on Saturday morning, on your train, watching the escalator, and seeing nobody, the long door dwell is to keep the train on schedule. Almost all the schedules are the same times, whether it's rush hour, or "oh-dark-thirty" in the morning, so the long dwells seem funny when there's 16 people on the train, but seem too quick when there's 1100 people on the same train.

Commuters are smart and quick. So, train operators rarely do an extra door cycle when it's busy. The extra 20 to 30 seconds may seem like a small amount, but the commute schedule has very little slack, especially between Daly City and MacArthur. Chances are pretty good that a train is already 30 seconds behind schedule, and another 20 seconds will screw up 5 or 10 thousand commuters. Besides, there's another train in 2 minutes (or less)! Even the reverse commutes are time sensitive. Many commuters have bus (and Caltrain) connections, and, if the train is late, central has to call them and hold the buses.

Off peak and weekends is a lot easier. There's more time between trains, and there's fewer people. So, if all 30 of your party take 3 minutes to load through one door at Walnut Creek, then a short dwell at Lafayette, Orinda and Rockridge will take up the slack. And the timed transfer at MacArthur allows some more slack. Many weekend riders are inexperienced riders, so they take longer to get on and get off. And BART does maintenance during off peak, which causes some delays, but they still like to make the timed transfers work.

The operators have some degree of manual control. If they look ahead, and can see a train in the next station, they'll wait, with doors open and collect a few more passengers, rather than sit in a tunnel. If they know certain stations are slow, and others are crowded, they'll get out of the quiet statins quickly, knowing that they'll take longer to get their doors closed at busy ones. Long after central sends the beep to close the doors.

If they are on time, and they know they have a timed transfer ahead, they'll often open the doors for a latecomer. Particularly at night and weekends. A happy casual rider may become a steady daily commuter. But, if you come up the stairs, and the doors are already closed, and you don't get them opened personally, just for you, then take a look at the schedule on the wall. Chances are that train is already late. And, if the scrolling platform sign says "next train in 7 minutes", instead of the usual 15 or 20, then you know that train was super late.

When it works well, or even when it's a bit (3 or 4 minutes) behind, it works well. But, when it's a mess, it takes creative controllers, supervisors and train operators to get it back together.

  by djQFI
modorney wrote: (snip)
Skipping stops:

Some runs (like reverse commutes) could skip stations. That would make run times quicker and more predictable.
IIRC, shortly after the SFO Extension opened, there were some SFO-Richmond trains during the afternoon rush which ran "express" from SFO to Daly City, then reverted back to the regular "local" service past Daly City. It sounded like a pretty cool idea, but I imagine the scheduling was pretty tight, since you also had Dublin-Pleasanton/SFO and Millbrae/Bay Point trains sharing the same tracks.

  by modorney
> ... trains during the afternoon rush which ran "express" from SFO to Daly City, then reverted back to the regular "local" service past Daly City. It sounded like a pretty cool idea, but I imagine the scheduling was pretty tight ...

That's a good idea, during the afternoon rush, a lot of people get on at Millbrae and SFO, while the next three stations are quiet. After Daly City, there are a lot of trains, some only 2 minutes apart, so skipping stations doesn't work out too well.

But, there is talk of rebuilding 24th street station, and adding 30th street, and including a third track, so a passing siding might help. The basic design is like the middle track at Daly City. The Fremont trains would go on to the middle track, and change direction. But, that would tie up the middle track for about 2 minutes, every quarter hour. The rest of the time, it could be a passing siding.

And AATC will allow lots of "creative" solutions.

  by Palal12
Thanks for the insight about PLs.

Modorney: Good post on how the BART computer system works. I spent time in SF back in 2000 and being the rail transit fan I am,I covered the entire BART system. I learned that MUNI passes are valid between the SF stations(Embarcadero thru Balboa Park) on BART. What I also remember hearing about is that BART has a 3-hour time limit on their tickets from entry. I also learned that you can pay the minimum fare if you exit at a different station from the one you boarded. If you entered and exited at the same station the excursion ride fare applied. I covered lines in the system using the BART schedule and watching the time. If I planned to get off outside SF I bought a ticket to that point and took a side trip on it. I would have loved to have seen from the front end the transbay tube-since I did not ask,would a friendly train operator allowed it? Does BART plan to replace the BART-1 cars from the early 70s? in my opinion the BART cars need a super-reflective RED stripe on the front end for visibility. To me the extension to San Jose should be a top priority. Riding from Fremont I almost felt that I was on the Babylon Branch of the LIRR. I enjoyed exploring BART and I hope that I can return to cover the new SFO extension as well as cover Caltrain to and past San Jose. When I was there in 2000 I stayed near the N-Judah LRT line-covering MUNI METRO was for me no problem. Memories and observations frpm MACTRAXX

  by modorney
I'm glad you had a good time, Mactraxx -

Since 9-11, BART, like all other transit agencies is more strict about letting people in the cabs. But, your best bet is to hang out and wait for an A-car, since the view is better. However, there is quite a bit of reflection from the (passenger) lights in the car, so many train operators hang a cloth curtain, or a coat over the door, so that would block your view. (This isn't because they don't like railfans, it's to improve visibility, occasionally, a homeless person will wander into a tunnel, and it's hard to see them, just as your eyes are adjusting.

So, here's the trick. The A-cars belong to Richmond yard, so you have a good chance of riding in an A car if you are on a Richmond (Red) train. Plan on a weekday or Saturday, daytime. Richmond to Daly City trains don't run Sundays, or after the commute rush in the evenings. Start out at MacArthur, and hop on an A-car (that's going to the city). If you can't see out, then get off at 19th, and try the next one. With four stations, you are bound to get one with a good view. Coming the other way, you can start looking at Daly City (these trains fill up - there's a steady stream of Berkeley riders all day long) By Civic Center, you should have a good view train.

If you are railfanning on Sunday, or late at night, there is usually one or two "A-Trains" on the Pittsburg to SFO (Yellow) line. The Richmond line "lends" trains to the other lines, so you could find A cars anywhere on the system. They are not common, so grab the first one you see, it may be the only one.

In the Afternoon, there are extra trains that run from Daly City to Pittsburg, and a couple of these have A-cars. Since these trains have to fit into a tight schedule (2 minutes behind and 2 minutes ahead of another train), they will occasionally skip Daly City and Balboa Park and begin taking passengers at Glen Park. And, if a Dublin or Fremont train gets taken out of service, these extra trains often turn into replacements for the missing train.

As far as new cars, BART plans to buy new cars for the San Jose extension. I've heard lots of options and scenarios, and the SRTP's over the past couple of years have talked about getting new cars. The situation I would like to see is for the SFO-Pittsburg line to get the new (three door) cars. PATH just got new cars, and these cars would be a lot like the new PATH cars. They could piggyback on PATH's order, and hopefully get them for a good price. It takes 140 cars to fill up the Yellow line (14 trains), plus a few more for spares, etc. The Yellow line has the biggest commute load, so the three doors make for quicker loading and unloading. I think all of PATH's cars are cab cars, BART would probably buy a mix of cab and non-cab cars. BART is a specialized, custom system, a BART car should be a little more expensive than a PATH car, but the non-cab cars should save some money.

That would free up cars for the other lines. And would allow BART to rebuild some of the oldest cars. All the A and B cars have been rebuilt, that leaves about 230 (C cars) that could use some work. When they did the A-B rebuild program, the rebuilder (Morrison Knudsen) would take 36 cars at a time, and start rebuilding them. The easy rebuilds were finished in 2 months, and the tougher ones took longer, but, after a few months, they had all 36 done. Then they'd get the next batch of 36, and do the same thing.

They would probably do the same thing with the C-1 cars (about 150) and C-2 cars (about 80) - rebuild them in batches of three dozen, or so. Most likely, they would remain as cab cars, but would probably become the same (there are minor differences between them).

Ridership is slowly picking up, as the Bay Area economy improves, and the engineering of the first part of the San Jose extension (Warm Springs) is ongoing. We should start seeing serious talk about more cars sometime around late this year, or next year.