Silver Line frequency during SafeTrack

Over three-and-a-half years ago Sound Transit launched a study to solve the myriad problems blocking proper real-time arrival information at Link Stations. Since that time, the returns have been meager: ST took responsibility for OneBusAway  from the University of Washington. OneBusAway still reports only scheduled data for Link. Only the two U-Link stations have onsite displays of next train times, which have problems of their own. These issues will soon be resolved, while onsite info at the other stations is nowhere in sight.

First, the good news: beginning sometime in January 2017, OneBusAway (and any other application using the data feed) will display actual next train times for OneBusAway. Not only will mobile devices have accurate information, ST’s IT department says that a software upgrade last September “improved the accuracy of the ETA data” for Capitol Hill and UW Stations.

Unfortunately, there is no clear path to upgrading the other existing stations. The existing “dynamic” screens are severely limited and cannot accept the heavily processed data feed that OBA uses. The study estimated the total cost of replacing the screens at $475,000, which ST deemed “prohibitive.” Blocked there, ST is looking at a “feasibility study” that will figure out what can be done with the existing system, how much it will cost, and funding sources for the work.

ST could not provide a timeline for this project, but they “remain very committed to finding ways to provide Link arrival and departure information to our riders, and to have this information available on station platforms and in other convenient ways.” It sounds like it’ll be quite a while before we have universal RTA signs, much less ones that list the length of trains (as pictured in Washington, DC, above).

75 Replies to “Real Time Info Coming Closer, Farther Away Than Ever”

  1. How exactly is “remain[ing] very committed [emphasis added] to finding ways to provide Link arrival and departure information to our riders, and to have this information available on station platforms and in other convenient ways” expressed in finding the expenditure of $475,000 to replace the existing screens in Link stations “prohibitive”?

    Sure, one doesn’t want them chasing the latest “hot new thing” like Donald Trump frothing in a tweetstorm, but since bus-train kerfuffles are to continue for another year and a half, giving people information about delays makes sense.

    1. Wait, I thought that was $475,000 per screen or per station. This is for all 13 existing stations? $475,000 for all of them doesn’t seem in any way “prohibitive”.

      Maybe put a fancy design on them and call them art? lol

      1. Yep:

        The study estimated the total cost of replacing the screens at $475,000


        Seems a bit like loose change among $20 billion, no?

    2. Most of the time, schedule information, alone, is sufficient, but not always. When service is disrupted due to a car-train collision in the Ranier Valley, real-time arrival information becomes very important to the rest of the system.

      The information screens in the older downtown tunnel stations are ridiculous. As far as I can tell, they seem hard-wired to provide two types of information: train arriving in two minutes, and pre-recorded announcements at pre-determined intervals apart. Sometimes, the announcements can actually interrupt each other (e.g. “The next train, northbound, is arriving in…proof of payment is required to ride Link Light Rail. Tickets may be purchased at the ticket vending machines on the upper level.”). And, of course, ST has the volume turned up too loud, so that every announcement sounds like someone shouting in your ear. (Visit the departure lounge of SeaTac airport for a more reasonable volume level).

      1. asdf2, you forgot the default message of the signs, which indicates that you are in the downtown transit tunnel. Just in case you aren’t sure where you are after descending 100′ and seeing the trains and buses.

  2. This is a reoccurring theme on almost everything.

    We are committed to providing services. But it’s easy to just not do anything. Transportation in general.

    Stupid train cars with driver seats on both ends.
    No real time tracking as mentioned for something on a track.
    Rapid stations with out pre pay boarding
    Orca contract locked until 2019 and can’t live in the 21 st century
    Building a new downtown ferry terminal but need to queue the cars out on the surface street highway.

    1. Richard, your first paragraph explains why the second one is happening. Have seen psychiatric term “depression” defined as “learned helplessness.” Number of easy things Donald’s opponents can’t fix shows value of an education.

      Subject needn’t be on the WASL. It is the WASL. So I like your fighting spirit Ben, I like your fighting spirit. But with all the deserving candidates for the title, don’t award Stupidity lightly.

      We’ve had to build our whole regional system step by step. So there was a chance we could have had to run single car trains for many more years.

      A car with only one cab is hard to reverse. So the required loop at every terminal costs either surface parkland or very expensive elevated or subway structure. Portland’s single cab SR70’s could run single-car if the wonderful park system included a loop.

      Most of the great PCC cars had only one driver’s station. By the 1930’s, the car’s main selling points imitated its dangerous new rubber tired relative. Car’s greatest selling point was probably replacing the conductor with a farebox.

      Several Toronto lines loop gracefully in parks. Glenn, are all the great loop-positive park donors of the last street rail era all really dead?

      As is, every SR70 runs coupled to a car with a cab facing opposite direction, either a matching SR70, fastened tail to tail, or an older Bombardier, which has cabs at both ends. Our future walk-through trains could have many segments between the cabs.

      Meantime, Ben don’t lose heart. Shouldn’t be hard to find a replacement item. But Richard, given twitterbox contents to date (would somebody empty the damn thing) worst could be yet to come. What if approaching President loves light rail?

      Luckily for rail division, Vladimir Putin has probably already got his invitation to smash a bottle of wine across the front of the first Route 7 trolleybus over the mountains to Ellensburg. What better way to honor his hero Stalin. Real reason for seizing the Crimea.


  3. Living in a time when new flat screen TVs are literally being given away why do the “new” train arrival boards look like they are coming from a 1960s system in DC? The only thing worse may be the RapidRide real time information boxes. I would comment on the display but I have never actually seen one working.

    I think the NW region in general has some issues with signs, maybe we are looking down at our phones to much to pay attention. From highway toll signs that are so cluttered with information most drivers have no clue what they are paying to non existing and non functioning bus/rail departure times that almost never work correctly.

    With the exception of Sea-Tac arrival/departure boards I’m not sure we have anything that actually works.

    1. Because the dot matrix displays are a lot easier to read than a flat screen LCD from way down the platform. Metro Transit in Minnesota is replacing some of their dot matrix displays with flat screens; those stations have effectively lost useful information. Let WMATA do its thing with displays for Metrorail, they are by far the best information signs on a rapid transit system I’ve seen. Don’t encourage adoption of flat screens, they’re too invisible and too harsh for rapid transit information.

      1. Flat screens can be “dot matrix” also. Whoever owned it has taken it down, but there used to be a huge plasma screen just south of the King/Pierce county line on I-5 so bright that it was mesmerizing enough to be a traffic hazard. (Maybe that’s why it’s gone).

        There are probably thousands of “fifty inch class” plasma TV’s moldering in a warehouse somewhere. Buy a couple of hundred (for backup) and turn the programmers loose to produce meaningful displays.

        It’s not rocket science. But is it “train science”, RS’s predecessor?

    2. Right? Big screen TVs are dirt cheap. If it breaks or gets vandalized, it’s still dirt cheap to replace. And as far as getting some basic train prediction, as long as there’s a system in place for providing train location, this shouldn’t be a issue. Actually, it sounds like a fun problem. I bet I could get acceptable accuracy using basic statistics. Maybe we could go really crazy and use some machine learning and factor in things like time of day, driver, or the lights on MLK.

      1. General purpose TV sets are not suitable for outdoor use, they are not suitable for 24/7/365 use and will fail pretty quickly under such conditions, indoor 48″ LCDs that are of a commercial grade are around $1200 each, and if you want one that can withstand the rigors of being outdoors…

        Also ADA may be an issue as well. while LCD screens are becoming more commonplace they may not be able to provide enough contrast vs. a regular LED matrix display. the LED matrix displays also have a known long service life and are used throughout the world.

  4. If they would open the interface to the displays at ULink stations, I could get real-time arrival from OneBusAway on them using Raspberry Pi for about $30 of supplies each. Anyone at Sound Transit want to hire me for a weekend project?

    1. Was thinking the same thing.
      Yeah, Pi for $30, or just buy $100 24″ LCD screens and voila, real-time.
      I imagine a GoFundMe could easily come up with the $5-$10k needed for all supplies + some cash for basic labor.

      The only question would be just getting some space to hang the screens on ST/SDOT property.
      So, they could save $450k+ and have real-time in just months. Crazy.

      1. AlexKven, that was almost exactly my reaction.

        I was once told that the most important skill a technical manager needs is the ability to tell, when given an unexpectedly high estimate, is whether it’s because of some unforeseen issue, or just because the engineer doesn’t want to do the job. So why exactly don’t ST want passengers to have this information?

      2. I swore there was a RFI that they put out for transit screens. I thought I tweeted it, but apparently I’m imagining things.

    2. Thinking the same thing here, even $1000 per sign would be too much. Just two screens, a small computer, minor electrical wiring, and labor cost for one worker in one day, altogether should run about $500 per sign. Two signs per station makes this project $16,000 total.

    3. Exactly. If it cost double or quadruple because the government has to have records and procedures and bidding processes to avoid graft, that’s OK.

      But 400k is just infuriating.

      1. I respect your hack-it attitude, but keep in mind that whatever piece of technology gets installed will need to withstand moisture and be tamper/theft proof. If these things are not secured, they’ll be on Craiglist in no time.

      2. The new TriMet screens cost about $8,000. It’s $5,000 for the hardware and $3,000 to install. Keep in mind many bus stops may not have power and data where you need it, do you may need to get the electric jackhammer out and install conduit in the concrete.

        See TriMet propaganda at:

        Now, TriMet has only one tunnel and the system shown in the propaganda sheet relies on a cell network data feed, so it works well for surface stations. It may take some special jury-rigging to get it to work in tunnels. However, it seems like there should be some solution to that. It’s not like ST is the only transit agency operating trains in a tunnel.

      3. 400k seems cheap for such a project. Remember these are not normal LCD screens they would need. These are commercial grade, long life, outdoor all weather rated displays. Low and High voltage electrical work would need to be done by a licensed and bonded contractor, with all the electrical in conduits and everything mounted in weatherproof NEMA type enclosures. and that just the start.

    4. Right… “The existing “dynamic” screens are severely limited and cannot accept the heavily processed data feed that OBA uses” – does anyone actually believe this?

      Either ST is incapable of using technology or they are incapable of buying the right technology to begin with. I can’t decide which is worse.

      1. One thing everyone forgets, is that this system was designed in-between 1997-2003. All the technology on the line is representative of design decisions made during that era. The displays are exactly that. Sound Transit of the era also specified the current system as giving you a two minute warning prior to train arrival. We got exactly what was specified and available for the time.

      2. Real Time Arrival is a lot older than even 1997. Even then of course they weren’t designing these systems at the very beginning of the project.

      3. Heh…in the artist’s renderings used for the DSTT EIS and promotional information in the early ’80s they showed what appeared to be a real-time arrival sign. At least somebody back then thought that might be kind of cool….

      4. Metro buses had real time information since 1995 with technology developed at UW!

        Train stations and airports long have had electric/mechanical departure boards, since the 60s. It’s nothing new.

  5. What kind of requirements gathering was done when procuring the original screens?

    In other words, why the hell did ST buy something that they weren’t sure would ever have worked?!

    1. It works for what it does. The goal was to keep costs low to ensure ST1’s passage, back when there was no rail on the ground to experience . It’s the same mindset that led to a surface al8ignment in Rainier Valley because the flat terrain did not “justify” the cost of grade preparation. Never mind that a tunnel would have eliminated collisions with cars and peds and their associated delays and expenses and allowed the trains to go 55 mph (which should have been a minimum requirement, like real-time signs).

    2. Don’t believe for a second those LED screens can’t display real time arrival. The problem can only be in the signaling and communications. There’s nothing preventing those screens from displaying anything ST feeds into them.

  6. I’m still waiting for a fix to all the signage about trains going To Seattle — at stations that are already IN Seattle. Apparently created by an out-of-town consultant who thought that Seattle is only Downtown Seattle.

    1. I don’t really mind that. It’s short, and it gets the point across. At Rainier Beach, one direction says “Seattle” and the other says “Seatac,” and it’s clear which is which. Just like at Judkins Park, one way is going to say “Seattle” and the other “Bellevue.”

      1. Sorry, but a sign at the Capitol Hill Station that reads To Seattle just doesn’t make sense.

        I’ve also noticed confusion at Westlake. The fixed signage says trains go to Airport and Angle Lake. But when the train shows up, it just says Angle Lake. Riders unfamiliar with the line can wait for the Airport train to show up…until they give up or ask someone for information.

      2. I’m convinced Westlake with always be full of very confused out of towners who would be confused by anything train related. We can’t even get half the people in this city to use escalators correctly.

        Having worked customer service and hospitality for many years I can assure you there are people walking around in a constant state of confusion. And they won’t read any sign placed directly in front of their faces.

      3. Barman, I don’t accept the notion that because many people are easily confused we should therefore give up trying to provide good signage and wayfinding. We should at least make a credible effort, and ST hasn’t.

      4. It’s a disagreement on what good signage is. Your position that “to Seattle” is inappropriate anywhere inside the city limits is one position, not the universal or authoritative position. On the one hand are thousands of people wanting to go to Seattle by which they mean between Westlake and Intl Dist. Many of those are visitors who are especially prone to getting lost and want easy-to-understand words like “Seattle” and “Airport. On the other hand is the theoretical problem of people thinking Othello is outside the city limits, being confused about that, or thinking ST is an idiot because it’s not pedantic. Visitors don’t go to Othello unless they’re staying with somebody there, and whether it’s inside Seattle is not a priority concern. Residents know whether Othello is inside the city, or if they don’t it doesn’t matter much unless they want to know who to send their building application to or complain to their councilmember.

        The San Diego Trolley has the city name in small letters below the station name in the suburbs. That would be one way to do it.

      5. There’s a sign on I-205 just north of SR14 that says something along the lines of “Washougal / City Center”

        Which city center? Vancouver? Camas? Washoigal? Downtown Portland?

        So, simply saying “Downtown” might not be ideal either. At some point, there will be several downtowns in play.

      6. I agree with the point that To Seattle is generally understood, but I’m not sure why using Seattle-Downtown would not be more clear. It’s a reasonable suggestion that it should be changed to a more specific location. The direction reference could change within the city from Seattle to Seattle Downtown, much like the way the Metro and Sound Transit do on the their bus Marquees.

    2. The signs are mainly for visitors. When a highway sign says “Seattle 12 miles”, it means 12 miles to downtown, not 12 miles to the city limits. If you noticed the other way they say “Airport & Angle Lake” for the same reason. Washington is the most trade-dependent state in the country, and Seattle’s economy is based on a lot of visitors (not all tourists), so it’s not surprising they emphasized it a bit. The #1 question people ask me at Westlake is, “Is this where to get the train to the airport?”

      1. CBD is a planning term that few visitors understand. Downtown or City Center (like Portland) is more common.

      2. Here again “City Center” is the international term. “Downtown” isn’t used much outside North America. When I was thinking of downtowns once I had a fun time trying to decide where London’s downtown was, and concluded it doesn’t have one. It could be the City or Canary Wharf or Leicester Square or Charing Cross or Oxford Circus or Victpoia Station or all of inner London, but none of those are really satisfactory.

      3. Wouldn’t work in Seattle. Too close to “Central District” which isn’t core downtown.

  7. Thanks for the WMATA photo! Nothing shows better where ST should be than to show another system that puts continuous line, end station and train car length as well as real-time arrival on a lower-grade sign. Also notice that the sign doesn’t say “WMATA Metro Heavy Rail” as a restatement of the obvious on the bottom of every sign – like we have!

    Why do board members agree to spend billions yet can’t push for a modest amount to improve the most basic aspect of the system: information on signage?

    We should not accept a ‘we can’t do that’ attitude from a transit agency when it’s common to find it almost anywhere else. These systems are designed and installed by the very same firms that ST hires or would like their business.

  8. I find it laughable that $475,000 is “prohibitive” when we just passed a referendum to have Sound Transit spend $54,000,000,000.

  9. Someone at ST should talk to someone at TriMet.

    The 1980s era MAX displays on the Banfield line were updated to show real time information.

    1. It seems often that transit agencies feel that they have to go it alone, and don’t reach out to their peer organizations when trying to solve a problem: RTA system; general station signage; contactless fare media/system; station/platform design; train seating layout…
      It’s either they think they are the first ones to ever to try to solve a certain problem, or that if they simply copy other known best-practices, that it will be seen as not trying hard enough. I think it’s the fact that at the implementation level, it’s the engineers that set the methodology, and their bosses, being politicians, don’t know how to correctly manage the situation.

      1. That’s the kind of thing that a non-elected ST board is supposed to prevent. Yet they continue to make stupid political mistakes like this all the time. I bet the average highschool computer nerd could figure out a way to get those signs to work.

  10. Considering that U-Link finished $200,000,000 under budget you’d think Sound Transit could find $500k in the couch cushions to provide useful information to their customers…

    I hope they’re still planning on accuracy improvements for the signs at CHS and UW. Even since the update in September they are still sometimes unusable. I’ve seen people walk into the station, see that the next train isn’t coming for 20+ minutes (wrong, it just didn’t display the next couple of trips) and walk back out.

    1. “I’ve seen people walk into the station, see that the next train isn’t coming for 20+ minutes (wrong, it just didn’t display the next couple of trips) and walk back out”

      Hilarious. Especially because they probably spent 20 minutes descending and then exiting the station.

  11. Everything that’s been said already. These are dot matrix LED displays. There’s a controller somewhere that spits letters onto them. Put in a different controller. If it’s costs half a million dollars at least we have functioning signs for the next generation. Can we let some grad students have a crack at it?

    1. Hell, turn the LED displays into expensive clocks, vampire their power/data, and roll out something with commodity TVs run by Raspberry Pi or similar computers like everybody else has been suggesting. It’s not like anything that requires a ladder to access it is particularly vulnerable to theft and vandalism when there’s on site security. This is just another example of what happens when ST staff doesnn’t use the product they are maintaining.

      1. “This is just another example of what happens when ST staff doesn’t use the product they are maintaining.”

        A claim without evidence.

  12. Besides fixing the real-time data issues, I don’t get why they can’t use actual t.v. screens instead of this dot matrixy crap. Toronto uses t.v. screens, so they are able to display news, weather, videos, etc. in addition to the arrival info. In fact, the tvs are tuned into a local station (CP24).

    1. Please god no do not let there be constant interruption and flashing images while we are trying to wait for a train.

      1. Sadly, that seems to be the direction all this stuff is going. All the sign makers seem to be advertising the ability of their systems to bring revenue and provide “infotainment” rather than accurately and legibly display what passengers actually need to know.

        In the Mass Transit Magazine advertisements, I have yet to see a single sign or station display maker advertise “accurately and legibly display arrival times” as a system feature.

        I want to know when the train / bus is coming, not who is rubbing hips with who in Hollywood.

      2. You realize you can just not look at it if you so choose, right? I’d rather be able to read the news, weather, sports scores, etc. than have nothing to look at. I don’t see what you’re whining about.

      3. I can get my news and entertainment elsewhere, like a phone or a book. No need to clutter the one space where the transit agency can communicate to all customers with anything else.

        Visual pollution is a real thing. Otherwise cities wouldn’t ban billboards and signs from public areas. Even if you look away, it’s still there, in your peripheral vision. You can’t turn it off.

      4. Yeah, Santiago (Chile) does this as well. They are actually TV monitors (or something very similar) suspended above the platform, and the information is very small compared to the news/ads/whatever is on there. That said, they look as though they would be relatively inexpensive–I prefer LED, but whatever–and could likely be programmed to show actual useful information just as easily as ads.

      5. The small size of the information would violate ADA signage requirements for minimum text size and contrast so the Santiago/Toronto style is a non-starter. You could do separate slides for other content, but the train information has to be readable from a distance.

  13. I really don’t understand the difficulty or cost of adding real-time info to existing signs. It’s already partway there with the “next train arriving in 2 minutes” announcements, so we know they can receive remote updates. And I’ve worked with the OBA data feed for my own projects; it’s trivial to pull the necessary arrival data and format it a way that’s suitable for a screen. To me, it just seems like a software update on their server, no new hardware required.

    1. My suspicion is that the signs themselves are hardwired into the SCADA that controls the signaling on the line. The dot matrix displays, if broken free and controlled independently, which would probably require rewiring them, installing a computer to control them, and getting a new network connection to the controlling computer that would cost some money. Further, you may have to re-certify the signaling system because you have removed these devices from it which also costs more money. Should it be done? Hell yes. Its just a lot of money for an agency who is still not very familiar with transit operations and working with capital projects outside of a voted and approved on “plan” which in my opinion has been a great shortcoming of the agency for many years, even before LINK opened. It seems to me that if a major project is not in their “plan” it just wont happen. I honestly was surprised when they bought the Bonney Lake P&R from Pierce Transit as it wasn’t in their “plan” and I did not think they could deviate much from it.

      1. This is where I got confused. The article was a bit vague, and understandably so, since things get technical pretty quickly.

        But if I get what you are saying MrZ, is that there there is data being fed through the lines and it is being used for two purposes. One is to feed the monitors, the other to actually operate the vehicles. If you messed with the data to display better (or just different) information, you could screw up the vehicles. Is that right?

        If so, it explains everything. Basically, they need to rewire things and set up a brand new network, and that is expensive. This is very unfortunate, as it probably would have been trivially cheap to set this up properly in the first place, but like a lot of what ST has done, they really didn’t consider the future.

      2. So I assume the signs as dumb terminals. As far as the data feed is concerned, it is my understanding that we already have that (in One Bus Away data). So it may simply be a matter of converting one data type to another and sending it downstream (to the screens/dumb terminals) which shouldn’t be that expensive. You could even make it an open source project and people might do it for free. But even if it was a private project, it seems like it shouldn’t take months and months. A three person team should be able to figure this out in a few weeks. The whole things sounds like it should be done for less than 100 grand.

        But my guess is that there is something else going on here, as MrZ suggested. Maybe the data feed is very complicated and uses an old, obscure protocol. If so they might figure that there aren’t many people around who know it, or can pick it up easily, and contracting that out would be expensive (sort of like the Y2K thing when companies struggled to find folks who knew COBOL). Even so, it doesn’t add up to me, as it doesn’t sound like a huge task. I would like to know more details, because at this point I don’t understand the problem.

      3. Real time data feeds don’t come out of thin air. Link’s One Bus Away data is only schedule data, not real time. They have to do the same thing they did to the buses: take vehicle location information and feed it into an algorithm to generate arrival time predictions. Problem is train location using track circuits isn’t as accurate as GPS and GPS doesn’t work underground, which means you need a lot of clever math to estimate it.

        There’s already an open-source project that takes bus GPS data, couples it with schedule data, and produces a real-time feed.

      4. Ohhh! SCADA technical and legal requirements are plausible, justifiable* and really hard to work around.

        * Long-standing safety practices. Not what I generally want to see disrupted, even when it’s inconveniencing me.

  14. Embarassing. Sound Transit. This is Seattle 2017. How can this not be solved within weeks of CH and UW stations opening? Ridiculous. I use Link about eight times per week and always subtract one to two two whole minutes from posted departures. This is for the trains.

    Maybe a hackathon is the solution if Sound Transit can’t program this on their own.

    They are flush. They have to provide this service to the ridership.

Comments are closed.