Nine carriage trains – could they work?

Today’s Herald Sun raises an issue that has been pondered for some time: 9 carriage trains to relieve overcrowding.

This would seem like a good opportunity to dig out my Bargearse remake video.

The reasoning behind adding carriages is that it’s close to becoming impractical to add more trains in peak onto the busiest lines, such as the Pakenham line, so they should make the trains longer instead.

Obviously they’d need to adjust signals and sidings, and make platforms longer. The question would be: can they make enough platforms longer that the trains could stop at the right places to take the bulk of the passenger load?

Let’s assume they couldn’t do the City Loop underground stations because of cost, but they could do Flinders Street and Southern Cross, which would connect much (but not all) of the CBD directly. Let’s also assume they would do Richmond (extensive connections) and South Yarra (which is increasingly becoming a major destination, as well as a big source of passengers) but that trains would skip the quieter stations on the line.

If, say, you lengthened platforms at the following: Caulfield (9791 boardings per weekday), Oakleigh (6820), Dandenong (6687), Springvale (5572), Noble Park (4554), Clayton (4323) and Huntingdale (4267). That’s about 42,000 weekday boardings. The remaining stops have about 22,000, so those 7 stations account for 65% of boardings.

Of course, this is making the assumption that the majority of people boarding from those stations that might get the longer platforms would also be travelling to other stations that would get them.

Obviously a lot more detailed analysis is needed. But it does mean that maybe, just maybe, “super-sized” trains stopping only at specific high-traffic stations might actually work. (Though of course the more stations they stop at, the more patronage they’d pick up, and the less the express running would reduce track capacity for other services.)

On the other hand, it might be easier to try double-deck trains again. Okay they’d have longer dwell times, but at least they could use the City Loop, and they don’t necessarily have to have only two doors per carriage side — Paris’s RER line A has trains with three doors per carriage side.

Upgrading signalling, grade separation and/or track expansion, and more off-peak trains (to spread the peak load) are also no doubt under consideration.

By Daniel Bowen

Transport blogger / campaigner and spokesperson for the Public Transport Users Association / professional geek.
Bunurong land, Melbourne, Australia.
Opinions on this blog are all mine.

17 replies on “Nine carriage trains – could they work?”

When I was in Tokyo the trains on the Jōban line were 15 cars long, with 2 of those set aside as Green Cars (swipe your smartcard above the seat for the right to sit in it, for a few hundred extra yen).

The other 13 cars were configured with seating against the walls only, and they were all packed to standing room. Headways were typically somewhere between every 4 minutes in peak and every 12 minutes early on Sunday morning. I won’t pretend it’s fun walking past that many cars at Ueno trying to find a seat (I didn’t bother to buy a smartcard), but meh. 15 car consists worked fine for them.

Instead of lengthening the platforms, couldn’t you have a group of carriages marked for certain platforms and program the doors to only open on those carriages at those stops? So if you want to get out at Parliament, you get on carriages 3, 4, or 5, and if you want to go to Richmond, you can get on carriages 1 to 8.

I don’t understand how it’s close to becoming impractical to add more trains in peak onto the busiest lines – are the trains bumping into each other?

Just because a handful of stations account for most of the boardings doesn’t mean that the passenger’s destination is the same handful of stations. Platforms on the whole network would need lengthening. That kind of investment seems unreasonable when you could do grade separation instead and take away a bunch of problems. More, smaller trains is preferable to the same number of larger trains if only for turn-up and go type frequency.

Does the PTUA still think double-deckers are too much money for not enough gain? Perhaps the same reasoning could apply to 9 carriage trains?
From :

While it is undeniable that double-
decker trains can carry more
passengers than single-deck trains,
we are inclined to agree with the
Department of Infrastructure
position that they are not at present
a practical solution for Melbourne’s
While the City Loop was (sensibly)
built for possible future double-
deck operation, most of the rest of
the system would need extensive
(and very expensive and disruptive)
works to raise the height of
hundreds of bridges and tunnels to
cope, and the concerns over
increased dwell times are valid, as
they would negate much of the
We continue to be of the view that
with upgraded signalling, removal
of single track sections, more
reliable operation, and improved
frequencies (including peak, peak-
shoulder and off-peak times) a great
deal more capacity can be found in
Melbourne’s rail system with the
current single deck fleet, at minimal

Hah! Just got to watch the video. Pity you didn’t cut it so that after the citybound train finally crosses, a outbound train goes (fade to black) – but that would require the kind of synchronization that you only get at Essendon station just before closing time at the Post office (two suburbans then a Vline).

Haven’t there been calls for signalling upgrades on the rail network going back for a couple of decades now? Is it ever likely to happen?

I’m just looking at the timetable for Caulfield to the city for 7am to 9am right now, and there are times when there are 6-10 minute gaps between trains. I’m sure this could be squeezed down to three minutes or so (like it was on the Lilydale/Belgrave line in peak, back in the 80s/90s, before Kennett gave us the permanent summer timetable)

Yes, you could run trains at closer headways with the existing signalling. The main problem is the level crossings. More trains means they are closed to road traffic for longer, and there is already a lot of political pain between Caulfield – Oakleigh (for example) from the existing timetable. One problem with closer headways and signalling will be congestion points (e.g. Flinders St, Richmond, Caulfield) where trains close up. This may need resignalling.

9 car trains will need platforms to be lengthened by 50% – that’s quite a lot. At Richmond, for example, the subway will prevent lengthening at the Melbourne end, and the Swan St bridge will limit lengthening at the other. Fixable, of course, but at a cost. Extending platforms may also need the signals to be moved (so that signals are not on platforms). This may need signals in the rear to be moved to give sufficent braking distance.

The lines to Pakenham/Frankston/Belgrave/Lilydale have already been adjusted to run double deck trains.

Well Yes and No. It would be expensive to extend every platform. How about 6 set double deck. That would be more space friendly.

My opinion on the matter isn’t made up, but my initial thoughts are pretty much this.

When people like Paul Mees point out that there isn’t a capacity crises and that the loop simply isn’t being used to its potential (and that it used to run more trains), the most common response is that it’s running more express services which eat into the total number of trains that can run.

If its impractical to extend all platforms for at least a given line, then there are two choices:

1) As you’ve said, trains would have to run as expresses between stations that can accomodate them; or

2) Trains would have to run with some kind of system where you either choose your carriage and make sure you get on one where you can get off at your destination, which I imagine would be confusing as all hell for irregular passengers, but I know they do on V-Line and in places like Tokyo. The big difference with Melbourne and Tokyo is that even if you don’t speak Japanese the communication with passengers is pretty clear and its easy to tell whats going on. In Melbourne speaking English doesn’t seem to be much of an advantage in understanding our absurdly poorly managed system.

With choice 1) I can’t see much advantage coming from running 9 carriage trains. Sure they can carry 50% more passengers, but you’d have to run less trains as a result, which means longer wait times and possibly only small gains in passenger capacity.

With choice 2) I can see the advantage, but it’d have to be coupled with some serious addressing of passenger information, which currently is average at the best of times, and outright miserable at the worst. With this sort of communication, there would be problems abound if passengers had to work out what carriages to board for certain stations.

Some good comments here, as always.

@Josh, since new designs in double deckers have emerged (Paris RER line A), I suspect it’s worth looking at again. Happily, the Pakenham line is one of those that was adjusted for the 4D double deck trial train in the 1990s, though there might be clearance problems with the Dandenong platforms at Flinders St.

@Paul, check the gaps for V/Line services. And note the level crossing issue.

@Andrew is right. Level crossings are the issue on the Dandenong line. Already there is immense political pressure to fix spots like Murrumbeena, and Carnegie and Hughesdale aren’t much better. Clayton, Springvale and Grange Road also have big problems. You can squeeze more trains on, but until a solution is found to the level crossings, I doubt we’ll see many more. (Note most of these roads have bus services, and pedestrians, including train passengers, also get delayed by the boom gates, since most of these stations don’t have pedestrian underpasses/overpasses.)

@Phill, no, 6-car double deck trains won’t work. The shape of them can’t clear platforms etc at curves, and the small number of doors would mean long dwell times. 8-cars might work.

@Julian, yeah I suspect there’d be problems with short platforms and long trains. It works on long distance trains when almost all the passengers board at big stations with long platforms, and there’s no issues making your way along the train to the appropriate door. Different story on metropolitan services with a lot more origin/destination combinations where in peak you can barely move.

Paul Mees had a really good point about CBD rail capacity, though I’d note that in the last few years a lot of the things he called for (such as running more trains direct to FSS) have happened. I suspect that we’re not far off reaching the old record for trains per hour through central Melbourne (if it hasn’t been reached already).

Paul Mees had a really good point about CBD rail capacity, though I’d note that in the last few years a lot of the things he called for (such as running more trains direct to FSS) have happened

There’s still also a lot that haven’t: having station staff assist wheel chair passengers instead of drivers; investigating different locations for driver change overs instead of Flinders Street and having standby drivers waiting at the appropriate end of the platform for trains that must ‘reverse’ out of Flinders St; and terminating the La Trobe Valley V/Lines at Flinders Street instead of Spencer Street (with standby drivers waiting!).

@Julian, yep, that’s right.

I’m hearing on the grapevine that a couple of stations (Box Hill and Parliament???) now have staff standing by for wheelchairs, but I haven’t seen it myself. They do seem to have started having drivers ready to go when trains need to reverse at FSS in peak.

I’m not sure the termination of Latrobe Valley V/Lines will ever happen at FSS. It’s only 1-2 trains per hour, and it’d be political suicide to be seen to be presiding over it.

I’ve recently seen staff at Parliament standing by for wheelchair passengers. A good move, I think, as are the ‘helpers’ at peak hour providing additional information, asking the doorways be kept clear, etc.
Ditto for Richmond – having the information clearly broadcast (as opposed to garbled, or on screens which may/may not be that visible or informative) helps tremendously.

Capital works to extend platforms by 50%, works to signalling and siding lengths. It doesn’t really sound like the best use of the finite budget we have to begin with…

If they were to run express between the chosen super length stations, then you lose further train paths for possible future frequency improvements – especially having spent the kind of cash we’d be looking at on said upgrades.

On the other hand, V/Line’s annual report mentioned that they are trialling door sensors that prevent doors opening if there is no platform outside them with the intention of introducing 9 car Vlocity sets…
I see no reason why this couldn’t be applied to the electric trains… this could allow a more regular stopping pattern, but would obviously require good signage at stations and within the trains to limit confusion!

I have given a bit of thought to double deckers since the last time this came up here… Sydney’s 8 car trains for example, only have 2 doors less per side than our 6 car trains, and while (from my perception) the choke point seems to be the internal stairs rather than the vestibule doors, any cost to dwell time (however long or short) probably doesn’t matter all that much with regards to our system as we probably have too many level crossings (and will for the foreseeable future) to increase train frequency to the point where this will become an issue? As far as I can tell, dwell times would really only be of concern at the ‘confluence’ points closer to (and in) the core… Paul Mees suggested the core is capable of 181 trains per hour if used to its design, how many would this reduce by with an extra say 30 seconds dwell time?

As has been said, the loading gauge is a consideration for the introduction of double deckers as they wouldn’t fit everywhere – although this seems to be the case with our current fleet anyway! It is also another type of train (obviously) in terms of maintenance skill-sets, crew training etc, but I have no idea how big an issue that might be…

I wonder about the possibility of running 7 car trains as M-T-T-T-T-T-M instead of 6 car M-T-M-M-T-M… If we are in agreement that we still require more trains, we could have Nth Ballarat pump out only trailer cars for a while… 7 new T cars would turn one 6 car train into two 7 car trains of 163 metres (according to vicsig) – yes longer than our platforms, but Sydney’s trains are longer than their platforms according to that report enno referenced, but they are not so long that the passenger doors aren’t accessible from unmodified 160 metre long platforms… You also lose the dead space in the middle, so passengers can traverse the full length of the train internally which seems like a plus…

“I’m not sure the termination of Latrobe Valley V/Lines will ever happen at FSS. It’s only 1-2 trains per hour, and it’d be political suicide to be seen to be presiding over it.”

I don’t see how it would be any more an act of political suicide to deciding to stop half the suburban trains from travelling through the loop.

There’s some opposition to moving suburban trains out of the loop, but much of the change has already happened. In contrast, petitions have been raised in the past regarding Gippsland trains, by those who are now in power.

One of the ideas floating around is to terminate one of the northern loop lines at Southern Cross (again, in the name of CBD capacity); it’d be interesting to see what response that would get from passengers.

Comments are closed.