Subway Operations at the Wye

The post on George Smitherman’s transit platform has spawned a comment thread dealing with how the TTC manages operations on the subway.  These comments really don’t belong where they were posted, and I have moved them here.

16 thoughts on “Subway Operations at the Wye

  1. Re: M. Briganti’s Wye:

    The way the TTC operates it’s current OOS wye is enough to justify never doing so.

    If there is a slight delay, all of the lines are affected. Anyone who has pie-in-the-sky visions of a University via Lower Bay to (wherever) service, should look at this past weekend’s service levels.

    I dare say that all three lines were in some-way affected by this maintenance work.

    Certainly YUS trains were held up to every-other or every-second through the junction.

    It is absolutely un-necessary to run subway service from where you are to where you want to go. Changing trains by walking/escalating/elevating up/down to another platform of another line is BY FAR less of an inconvenience than holding up the entire city so you can go from York U to Don Mills & do a little shopping.

    If the system were capable of handling these movements inside acceptable service levels, than all would be well. The TTC, and the subway’s current configurations absolutely forbid these kinds of movements.

    Re: George’s plan: Lovely lines on a lovely (expensive?) map.

    Like

  2. To Drew’s comment …

    The whole point of the westward extension of the Sheppard subway is to connect the Finch W. LRT to the Sheppard E. LRT and create a northern crosstown link. We would have to somehow send the Sheppard trains up Spadina to get to Finch West Station — otherwise, passengers would have to take the Spadina line for two stops. So, it’s not that stupid an idea after all — especially since George’s plans do not include an extension of the Finch LRT from Finch W. Stn. to Don Mills Stn. Besides, the TTC always wanted that link for moves to Wilson.

    I rode through the Bloor-Danforth diversion last weekend and I thought everything was working fine. When the schedule is that tight, there are going to be delays. Do you gripe about the same kind of slowdown trains go through on their approach to the terminals? It’s the same thing.

    Steve: The issue at the wye is that small disruptions caused large delays, and this was with only an off-peak service level. Having trains sit for extended periods while staff sort out passengers is not a viable operating scheme on a regular basis. In a “normal” integrated operation, there would be no need for through BD passengers to change trains at Museum, and anyone who boarded a “downtown” train by mistake would be expected to sort themselves out.

    As for terminals, the biggest problem is that trains have too much running time at the end of the peak period, and they tend to back up at terminal approaches. This is related to minimum headways through crossovers and the length of time required for crews to change ends or swap trains at the terminals.

    Like

  3. Just to be clear to Robert W., I entirely agree with your option for the Eglinton Line through the Black Creek valley. I think my wording wasn’t quite right. No matter what happens however, the problem at Weston road will not be solved unless they use stacked tunnels.

    As to the problems with the TTC’s operation of the wye to Museum, the root cause of the delays is human interaction. Yes there is too much service to pass through the wye at once. Yes the confused public causes excess station dwell time. However it is quite clear that the wye cannot be operated efficiently mainly because the interlocking is being operated manually.

    There should have been a system in place to automatically read the trains’ destinations and set the routes accordingly on a first come, first served basis. This is a simple matter with computers – it has been done elsewhere for a long time. Most of the wasted time was from having to verbally communicate destination and location by radio and then have the routing set by hand each and every time. This also meant holding trains further back from the junctions than might have otherwise been necessary to prevent routes being locked out by the interlocking.

    On top of this was the painfully obvious lack of experience by the train drivers from the BD Line who didn’t know how to operate at normal speed through the wye and YUS trackage and timer signals. If they had simply read and followed the signal aspects and the controller markers they would have passed through these sections like any ordinary YUS train. Many of the drivers seemed truly paranoid. Experienced and confident drivers passed through properly and efficiently. It was quite rare for the trains’ passage to actually be slowed by red signals.

    No matter what the level of experience it was also clearly evident that many staff really had no idea what the track arrangement was through the wye. Especially the first time they tried this there were numerous erroneous public announcements on board the trains about the routings which further confused the passengers who already had no idea of their bearings. At least this time around they had signs to hold up on the northbound platform at Museum to more effectively communicate the trains’ destinations. It was a good thing too because some smart guy decided that the BD trains wouldn’t have their rollsigns moved to the right indication if they were pointing the ‘wrong’ way.

    It’s rather unfortunate that by the time they get this method of operation running smoothly they will no longer need to use it.

    Steve: When the wye operated as scheduled service, the signalling was automatic. However, the operation was compromised by the schedule design which integrated the three routes so that a train might be a Keele-Woodbine on one trip and a Woodbine-Eglinton on another. This required absolute sequencing to be maintained, and a late train on any branch held up the other two. Where routes interline on other systems, it’s a first come, first served approach at junctions.

    Like

  4. It’s more painful than funny that the TTC is fixated on schedules and yet cannot keep the vehicles on time.

    Like

  5. I never understood why Sheppard wasn’t done to the standard 500′ length. They could have used platform markers or gates to indicate “ALL TRAINS STOP HERE”.

    To answer Kristian’s comment about the wye, it’s not as simple as first-come first-serve when the overall goal is to have evenly-spaced headways everywhere and a service that alternates back and forth between two destinations.

    Dave Imrie knew how that damn thing used to work like the back of his hand. I never fully understood it because …

    (a) it defied common sense, and
    (b) interlocking/signal theory makes me want to pull my hair out and run into moving traffic

    For example, if a train passes into Lower Bay from Yonge, if another train comes into Yonge just behind it, the interlocking will block the route to Upper Bay for a certain amount of time. Why? Beats me.

    Instead of having the switches simply alternate between routes after trains passed through, photo-electric filmstrip contraptions called ATDs moved the switches (providing the move was safe from an interlocking perspective) according to a schedule that was punched into a piece of film that did one spin every 24 hours — I kid you not.

    Back then, the controller at St. Geroge had to override this schedule so much (in order to keep things moving, and to keep the trains in order on the exits) that I think they stopped using it altogether shortly after the line opened. Even with all this manual intervention, trains did fall out of their proper batting order — that I remember.

    Like

  6. I have a pretty good sense of interlocking theory so I can give you a good guess about that route locking out at Yonge.

    If a train approaches an interlocking signal immediately guarding a switch it sets off a timer that must expire before the switch can be moved in front of it. This is because a train could accelerate fast enough that the tripped emergency brakes could not stop it in time to not foul the switch. Without this protection the switch could move while the train was passing over it. If the previous route dropped as the previous train passed through then the new route could not be set up until the switch timer expired. Therefore the signal would not clear only because the route had to change.

    The problem at the location you cited is that there is no additional interlocking signal (control point) further back down the line to hold the following train sufficiently far back to prevent engaging the safety timer. There are other locations on the system where a double-head interlocking signal can be found much further back down the line, usually at the previous station. Holding a train at one of these signals ensures that the part of the interlocking with the junction doesn’t get locked up because the approaching train is not inside the ‘danger zone’. I gather that this setup is primarily used where a crossover allows a train to cross to the reverse-running side in front of another train. If the control point is not held red in this situation then that routing over the crossover gets locked out and cannot be set up. For my local station Keele with its crossover and yard, High Park eastbound has such a control point signal to prevent locking out of reverse-direction routings at Keele.

    There is no control point signal prior to Yonge westbound because reverse-direction travel is not allowed here by the interlocking. This is unfortunate because this kind of delay could have been avoided if another control point signal had been installed further back. Perhaps they thought the schedule would never put trains this close together. Think of it this way – if the following train was not immediately behind then when the new route was selected it wouldn’t require the safety timer to expire because it would never engage in the first place.

    Let your gears grind on that one for a little while and I’ll write a follow up a bit later.

    Like

  7. I’ve been thinking about a wye for Spadina/Sheppard West for some time.

    My understanding is that once the Spadina extension is operating, the TTC will relocate rush hour short-turn trains from St. Clair West to Downsview. This will reduce the number of trains operating on the Spadina extension by half, and it’s plausbile to build a wye and interline from an extended Sheppard West to fill the gaps in service. It could also potentially improve North York’s vitality with increased York U student residences and foot traffic.

    Furthermore, Downsview short turn trains from Finch via Union could also interline with Sheppard, thereby encouraging downtown commuters to use the Spadina/University line as a single seat ride, alleviating—to some degree—overcrowding on the Yonge Line.

    Steve: There will be track time available for through running, but the structure needed to allow it has not been designed into the Spadina extension. Also, the headway on the YUS side of things may dictate a shorter headway on the Sheppard line to mesh properly, and this bumps the number of trains needed for Sheppard.

    Like

  8. The main problem with the recent use of Museum, from the point of view of someone who intentionally went to use it, was simply that people were not used to the system. People who normally get on a B-D train only need to know which train (or platform) is west, and which is east. For those at Museum, it is either north or south. There is no need to actually look at the sign on the front of the subway to see where the train is going. If interlining came back, people would actually have to look at the sign on the front of the train, otherwise it is a case of “tough luck” if they are not prepared to look. It would be a matter of an exception (like the other weekend) versus a long term change.

    I have ridden streetcars that were clearly mentioned that they would be short turning, and people would frown when the driver told them the car was short turning – all because they could not bother to read the sign on the front of the streetcar that clearly indicated its destination. A lot has to do with people just assuming the the bus/streetcar/subway will go where they want and not bothering to look first.

    Like

  9. I know this is far-fetched (because the Sheppard extension will never be built), but wouldn’t it be easier to connect a westward extension of the Sheppard subway to the future Sheppard West station on the Spadina line, and not Downsview?

    If Sheppard W. were built with two platforms like St. George (with the lower platform roughed in), Sheppard trains could use the lower platform and terminate there, or continue “up” to the Spadina line via ramps just beyond the station. Or, run the Sheppard line in a 2-level 4-track structure between Sheppard W. and Finch W. and terminate the Sheppard route at Finch W. Stn. with two platforms — no interlining. This gives a crosstown link between the Finch W. and Sheppard E. LRTs with only two transfers.

    As for NX/UR interlocking theory … thanks for the explanation Kristian, although I’m still not quite sure I fully understand it … 🙂

    Steve: The problem with Sheppard West Station is that, of course, it is quite a bit north of Sheppard (which already has Downsview Station). The Sheppard W station will be under the GO tracks, and putting in yet another level would be possible, but expensive. Needless to say there is no money in the Spadina project to do this. They are already eating up all of the contingency on things like extra costs at stations where nobody noticed how high the water table was during preliminary design, not to mention a hydro pylon that conflicted with a station structure.

    Like

  10. @TorontoStreetcars:
    It’s also a bit confusing to transfer between a short-turning and standard route vehicle in Toronto. Does one need a transfer? Usually not, but this is entirely non-obvious as everything else on TTC requires a transfer.

    I think short turns would be somewhat less stressful in a time-based POP system like the one YRT uses, because everyone would have a transfer at all times.

    It’s possible that I am just neurotic.

    Steve: No, you’re not neurotic. TTC should move to time based fares as a prelude to a shift to smart cards (or equivalent). It would make calculation of fares and fare validity SO much simpler, and be a good marketing tool.

    Like

  11. There is a slot in the streetcar rollsigns that is red with white letters reading “SHORT TURN.” I find this more effective than the regular black and white with specific destinations as far as getting people to actually notice is concerned.

    Steve: However, there are so many possible places that “Short Turn” could mean, people always have to ask. When we move to all door loading and electronic signs, we may get “501 Queen / Short Turn / Dufferin” or something like that. Mind you, try explaining to some poor tourist how “Kipling” is a short turn, but “Humber” is not.

    Like

  12. Being someonw who has never really spent any real quality transit time outside of T.O. I always wonder how wye work in other cities subway systems and how they are successful. Another question I always had was why did they actually need a lower bay station. Why could not Yonge be the jump off point for the split all it would need would be an enhanced signage and the forward thinking idea at the time to make the platforn larger. I would have rather got off the train and got back on the next train to get to my destination instead of running up and down a flight of escalators or stairs to catch a train.

    Steve: Strictly speaking, lower Bay isn’t needed if you assume that having only every second train serve this stop (while operating an integrated service) is acceptable. It’s probably there as much because there is an upper and lower St. George, although with very different functions (upper St. George was, for a time, a terminus).

    Wyes elsewhere work with signalling that is designed to ease movements and schedules that don’t impose unreasonable constraints on the order trains must pass through the junction.

    Like

  13. Steve, is it possible to use the double platforms at St.George station to make a simpler transfer system? In my opinion, it would be a great help if the North bound YUS trains and the West bound BD trains were on the same level and the South bound YUS trains and the East bound BD trains were on the same level.

    Most of the transfers that happen in the morning rush hours is from Eastbound trains to the Southbound trains and the transfers that happen in the evening rush hours is from Northbound trains to Westbound trains.

    Steve: This is physically impossible.

    Like

  14. Minor quibble: it was the tunnel that was under the hydro tower, not a station.

    Steve: Ah yes, it was the tunnel, but moving the tunnel caused the station to move, and it moved into a more difficult area for construction, etc, etc. I still want to know what bright spark missed so many basic site conditions along the route, and whether we have paid their invoice yet.

    Like

  15. Dwight … Bay and St. George are “inside” the wye and that’s why they have two platforms. Museum is “outside” — that’s why it only has one.

    The grade-separated nature of it requires that the tracks start to change levels gradually starting at Yonge so that they will be completely clear of each other for the turn at Avenue Rd. St. George needed two levels because it was also supposed to act as a terminal on the weekends and between 9am – 4pm, and 7pm+ on weekdays. That was the original plan.

    To my knowledge, at the time it was built, no other transit property had an interchange comparable to ours in the sense that two lines, running at right angles, merged and diverged from every direction to every other direction on a 2 minute headway!

    Like

  16. Karl Junkin wrote, “There is a slot in the streetcar rollsigns that is red with white letters reading ‘SHORT TURN.’ I find this more effective than the regular black and white with specific destinations as far as getting people to actually notice is concerned.”

    The logical side of me agrees with this, but the experience side reminds me of my commuting days from downtown to Birchmount and Sheppard. When an express version of the Birchmount bus was added that stopped only at main roads up to Sheppard, they would place a yellow card with ‘EXPRESS’ in black letters in each and every window on the side of the bus. Despite this, and the presence of people waiting to NOT board the bus, and the driver’s announcement before departing, it seemed there was someone at least half the time who rang the bell as soon as the bus pulled out of Warden station. Nothing wrong with that, since ringing the bell meant that you want off at the next stop, but the bell ringer usually was shocked to discover that Eglinton was the next stop!

    Too many people move around in our world while they are lost in their own world. 😉

    PSC noted, “it was the tunnel that was under the hydro tower, not a station.”

    If I’m not mistaken, there was an issue similar to this during the construction of the SMART tunnel in Kuala Lumpur (Moaz Yusuf Ahmad: help me with my facts!). I seem to recall that a stretch of the tunnel passed right under several towers/pylons and they had to take extra care to make sure there was no shifting in the ground at the surface – and all the while using a 13.2 metre TBM!

    Steve: In this case, the tunnel is close enough to the station that it would be in the cut and cover area. Levitation would be required to hold up the pylon!

    Like

Comments are closed.