From the Red Line
Published in

From the Red Line

The quest for doors

A recurring theme in vehicle purchases is to get more doors out of them.

It has to be noted that all new buses delivered in 2021 have had three doors — including the Linkkers and the two types of three door double deckers. More TEL trains were also launched with five doors per side.

This appears to be an attempt to do something about long boarding times. But do we really need to do that? And how do these initiatives work?

From the last time I examined the usefulness of all-door bus boarding, I still retain the view that greater policy changes are needed before we can even consider all-door bus boarding. While some have achieved benefits, it is a mixed bag. Apparently, London has had to cease all door boarding on the New Routemaster due to high fare evasion rates.

But first of all, why not bendy buses? SBS Transit asked themselves the same questions at the turn of the millennium, and here are the alleged results:

SBS998Y and SBS999U’s supposed problem was with the near 18m length taking up space, negotiating difficulties and the depot facilities/ driver training were catered for double deckers — the mainstay high capacity vehicle type for SBS since the 80s.

Now, these are excusable with a sufficient critical mass — and in fact TIBS did have that critical mass with 315 bendy buses bought. But unlike SBS, TIBS operated mainly in the new towns of the north (CCK, Woodlands, Sembawang, Yishun), where wide stroads are more forgiving of unwieldy trailers. TIBS also would have sufficient motivation to design their depot and interchange facilities to handle bendy buses, again due to the critical mass.

However, the argument on being unwieldy still sort of remains, with the MAN A24 and its supposed greater maneuverability needed before bendy buses could be deployed on Changi Airport routes, delivering much needed capacity relief to 858. Bendy buses may take up more road space and incur higher infrastructural adaptation costs — apart from the terminals, bus stops also need to be lengthened for them, and perhaps lesser buses may be able to stop at the same time if there are more bendies out there.

But what are the cases for bendy buses? They often fall down to one or more of the following arguments:

  1. The lack of an upper deck means people can just keep pushing towards the back of the bus, instead of having to choose whether they want to proceed to the upper deck or remain below
  2. The third door allows for faster entrance and exit
  3. And perhaps bendies may have more roof space for more batteries, compared to double deckers where the batteries take away space from the passenger compartment

Proponents of bendy buses made a video to prove their point, albeit with some very faulty science. A better comparison might have been to use Service 302 or some other route seeing both bendy buses and double deckers, at the same stop, around the same time.

But the authorities don’t think so, for reasons known only to them. Perhaps it might be that a policy movement away from stroads and towards narrowing lanes to reclaim road space for other uses may limit the amount of slack that a bendy bus can take advantage of. “Bendy buses take up more road space” isn’t something that applies only to the length of the vehicle, after all. It might say something that the first SMRT double deckers were deployed on route 972, which loops at Bencoolen Street, and which had never seen a bendy bus even in their heyday.

Maybe the point about passengers just pushing to the back of the bus may not work too, looking at the amount of publicity material that had to be made to get people not to crowd around the standee/wheelchair area and its respective door. And perhaps a policy movement to shorter bus routes may also render any increase in battery capacity for range unnecessary too.

They may think the mad science approach is worth pursuing, which led to the construction of the three door double deckers, one of the first for left hand drive markets after the New Routemaster. The idea is that such vehicles deliver the best of both worlds as some form of middle ground, with passengers exiting from the upper deck no longer having to cross paths with those trying to board on the lower deck.

It’s understandable why they think that way, since Berlin too has bendy buses coexisting with three door double deckers, cancelling initial plans to phase out the double deckers. They’ve even ordered more, and Alexander Dennis even offers three door vehicles with two staircases as standard in Europe, much like what we have.

The three-door MAN we have, well, isn’t that great. The adaptations required to provide a similar design to Singapore’s right hand drive market are awkward at best, and perhaps make using the rear staircase somewhat akin to climbing out of a kids’ bouncy castle. Alexander Dennis, perhaps due to their experience, has a significantly improved design, so that might be a better thing to examine.

Back down (source: LTG)

Conceptually, though, I think what might be the cause of the preference for three door double deckers would be to increase the throughput of the front door, which appears to be the bottleneck especially at interchanges, terminals and major stops. After all, we still board from the front with everyone having to push through the one-person wide aisle, so having one or two alighting doors is never the issue. The idea behind three door double deckers might be done by giving people the choice to go upstairs much earlier, and the rear staircase also avoids the issue of them having to wait for alighting passengers from the upper deck to clear out.

Considering bus travel times and the supposed preference for comfort over speed, the second staircase and third door also makes the upper deck more lucrative by providing a direct exit. If you’re going decently far on a feeder, the upper deck now not only provides more seats compared to a bendy, where you might have to stand the entire distance — you don’t have to worry about getting out either.

Of course, if we stop taking cash on buses, we could probably swap around the entry and exit doors like they do in Japan. Enter by the middle, exit from the front or rear — but without moving to three doors on a single decker bus too, this can get complicated real quick. But that may not be an issue especially with adoption of electric vehicles — the simplicity of electric drivetrains may have allowed both Linkker-STE and CRRC to offer three-door single deckers without the complicated mechanical redesign as seen on the double deckers.

Like buses, trains face the same problem too. A key lesson learnt in the design of the TEL train was to find a way to add an additional door supposedly to speed up passenger flow. This means:

  • all other lines have four doors per side
  • TEL and perhaps CRL train cars have five doors per side
  • JRL and RTS train cars have three doors per side

Four doors per side is something that we have had since the Initial System and something that wasn’t really questioned until TEL. Well, looking at the origins of the DTL, you couldn’t question things at that time anyway, considering how it started out as an extension of the Circle Line and thus would have been designed around the Circle Line train.

Well, what’s built is built. You could find a way to change out all the NSEWL trains to have five doors per side like the TEL ones, maybe if you want to replace the platform doors at the same time. But for the NEL onward, platform edge columns at several stations limit the reconfiguration works that can be done — a phenomenon known as the Changi Problem on this blog — and if keeping some doors closed is unacceptable as a solution for longer trains, it’s not going to be acceptable here.

But why is it better, and is it a retrograde step for JRL trains to only have three doors? A simple table explains why — capacities are from Wikipedia. The lower the passengers per metre of doorway, the faster people can board and alight since less people have to share one doorway. The JRL train in fact has the best metric here, at 44.4 passengers per metre — the smaller vehicles do offset the lesser doors.

Perhaps, though, the situation can be helped. While it may not be possible to change the physical position of the doors, they can be made wider, which is something that can be considered for future traincar orders. Think something like the LRT doors, or maybe those as wide as London’s NTfL or the Tokyo Metro 15000 series. Of course, it has to be noted that wider doors will take longer to close. It’s also important to remember that the ATO systems also need to be tweaked in order to achieve the more delicate stopping margins needed for larger doors to be effective within the current PSD layouts.

What would it look like now? The results are similar to that of 5 door cars, assuming 1.7m wide doors are deployed — not that much different from the NTfL, but slightly less wide than the Tokyo Metro one.

To me, wider doors, or more doors, will only pay off if the increased passenger flow rate allows for station stopping times to be reduced. This is all the more so in the case of wider doors, since the wider doors will take a longer time to actually open and close. Albeit this may only be in terms of seconds or even milliseconds.

Saving only just three seconds every station stop builds up, especially on the 3 car lines with denser station spacing. On a line with 30 stations, that three seconds builds up to roughly 90 second faster trips in one direction, and a saving of three minutes when both directions are added together. This can mean one or two lesser trains are required to operate a given service level and results in some savings for the operators. In fact, I’d wager that this may be a reason why train trips are slightly slower than they were 10 years ago, since more people are using the trains and that means they stop longer.

Of course, quieter stations can have lower stop times than present already, and if it can be managed, might be low hanging fruit. The data is out there. The LTA and the train operators would even have better data than peons and thus would be able to find out on average how many people board and alight trains at a given station, and perhaps they might be able to micromanage the stopping times instead of just applying the 28–45–60 rule.

But doing this would require travel patterns to first settle down before the analysis can be done. The downside of running “normal” timetables while not everyone has yet returned to the office is that station stops meant to cater to large passenger volumes feel awkwardly long if not so much people have to use it. The other problem is that you have to find a way to convince aunties not to rush for the train doors.

On buses, though, the savings can easily amount to a significant portion of the bus fleet and also speed up passenger journeys in a more obvious way. Since the bottleneck they’re solving appears to be at the front door, while three-door buses may pose the best possible techno-fix, I’m still of the intention that all-door boarding should be pursued in order to deliver a comprehensive solution that can be done here and now across the entire network. But, as I’ve said, that requires a significant reinvention of the whole bus network, starting from reconsidering its purpose.



A blog on transport issues in the Garden City of Singapore. You can say that I love controversy. Posts can get technical! Abuse of comments may be blocked. Subscribe to Telegram for updates:

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store

Sometimes I am who I am, but sometimes I am not who I am not.