Welcome to European Tribune. It's gone a bit quiet around here these days, but it's still going.
Display:
No problem, I still don't have a clear picture of ETCS versions myself
(compare the Wikipedia entries for LZB and ETCS, and you'll see that the latter is rather unclear)

here I repeat one message from a discussion in LinkedIn, with some additions:

After my reading on signaling systems in Europe a decade ago, I fail to see what improvements would ETCS bring over the already existing LZB (or TVM-430).

It would be much simpler to settle into a selection of either the German (LZB) or the French (TVM-430) system, and be done with it decades ago.

Instead, we saw yet another system that has some added negative aspects:

  • cost (debugging, installing/retrofitting in locomotives etc.) of a new system instead of using one already existing system

  • incompatibilities even between ETCS versions (when you ask a supplier "I want ETCS in my locomotive", he'll ask you "which version do you want?").
And your ETCS locomotive is not going to work over all Europe anyway, since each national network has its own ETCS version.

  • adding yet another system into already entrenched areas (LZB and TVM-land) was never going to work (why rip out an already installed system?), with suppliers asking a 6-digit figure for installing it.

  • AND you have to install LZB anyway, if you want to run in high speed (160+ km/h) routes in Germany anyway (Indusi/PZB is standard equipment in locomotives sold in Europe).
So, you pay two times for the signaling equipment of your locomotive.

Even today, there are new signaling systems created (SCMT in Italian network), despite the existence of ETCS.
Shouldn't this tell you something?

When I visited Innotrans in last September in Berlin, I saw a Siemens Vectron locomotive outfitted with 11(!) different signaling systems in one rack.
But all this added complexity comes at a great cost.
We could reduce that to 10 if we left out ETCS, and everything would work nicely.

And I really fail to see what improvement would ETCS bring over LZB/TVM-430.

Frequency?
There are TGV trains running at 3 minute headways in rush hour.

Higher speed?
RENFE still doesn't permit above 300 km/h runs on ETCS-2

Safety/Reliability?
After the disaster in China (where the signaling is supposedly based on a local version of ETCS-2), I have my doubts
(and the idea of using wireless for sending safety-critical data makes me VERY uncomfortable)

Lower costs?
Compared to already existing and in production systems, I fail to see it

Features?
Note that compared to the LZB video I showed, ETCS Level 2 cannot provide AFB (automatic train-pilot) from what I know, neither it is as flexible regarding braking distances
(I mention the latter one with reservations, since I haven't seen any ETCS/ERTMS documentation yet)

N.F.

by nfotis on Sun Aug 7th, 2011 at 10:27:34 AM EST
[ Parent ]
Some notes in addition to what was covered in my previous comment:

what improvements would ETCS bring over the already existing LZB (or TVM-430).

ETCS L2 means less lineside equipment than LZB, and thus potentially lower material costs (especially in cable thief endangered areas). The real big benefit would be in moving block implementation (ETCS L3, which requires little extra equipment and more extra software), but with ETCS L2 still not mature, it is in the uncertain future.

incompatibilities even between ETCS versions (when you ask a supplier "I want ETCS in my locomotive", he'll ask you "which version do you want?")

ETCS was plagued by two kinds of incompatibilities. One was incompatibility between manufacturers: there were ETCS L2 problems (mainly in the BeNeLux) stemming from manufacturers implementing still evolving standards differently. The other was and is program and parameter setting differences: the effect of physical things like braking distances, failure mode permitted top speeds, allowed regenerative braking power, allowed adhesion can differ between railway networks (partly for climatic reasons) – these all affect the loco's control software itself, too –; and data format. For the locomotive owner, however, this is not an insurmountable problem but a question of the price of extra software and extra storage space.

AND you have to install LZB anyway, if you want to run in high speed (160+ km/h) routes in Germany anyway

This now goes both ways: you have to install ETCS L2 anyway if you want to run on Italian high-speed lines or via Berne and the Lötschberg Base Tunnel in Switzerland, and at least ETCS L1 if you want to run into Spain (once the Barcelona link is completed), and again ETCS L1 if you want to run a freight train on the Betuwe Line.

SCMT in Italian network

SCMT is quite similar to ETCS L1, the trackside part is compatible while the data format can be packaged, thus it was decided to migrate it to ETCS. An extra is compatibility with the older Italian systems.

After the disaster in China (where the signaling is supposedly based on a local version of ETCS-2)

That disaster throws up many questions, but none of the potential accident factors I read of go back to the original ETCS design. According to Railway Gazette, CTCS-2 is based on ETCS L1, and "radio dispatching" (if reports about its use are right) was not ETCS L2 application but a 'fallback' level also used on conventional railways (usually it also means that only one train can travel between two stations). If, instead, the ministry said the truth and the signalling had a software error resulting in a failure to set a signal red, that's probably not even a CTCS-2 error (though the control centre's failure to notice the false signal setting could be).

ETCS Level 2 cannot provide AFB (automatic train-pilot)

From what I read, AFB is rarely used under LZB as locomotive drivers still can execute a smoother (more energy efficient, more passenger comfort) velocity curve. Note that AFB is a separate, overlaid automatic train operation (ATO) system. In theory you can overlay an ATO with ETCS, too (see Czech trials).

*Lunatic*, n.
One whose delusions are out of fashion.

by DoDo on Sun Aug 7th, 2011 at 01:39:08 PM EST
[ Parent ]
AFB is a separate, overlaid automatic train operation (ATO) system. In theory you can overlay an ATO with ETCS, too

Forgot to note the drawback: of course, it remains yet to be developed. For that reason, for example, London's Crossrail will be equipped with standard metro CBTC with ATO overlay on the central tunnel section, while trains will also have ETCS L2 and British TPWS for the connecting sections.

As an overall comment, while I am also generally negative about ETRTMS and especially ETCS L2, now that the rollout progressed in some countries, it would be nice if at least all mainlines with significant through traffic on networks with magnet-based (intermittent) train control systems (like Germany) would be switched to ETCS L1.

*Lunatic*, n.
One whose delusions are out of fashion.

by DoDo on Mon Aug 8th, 2011 at 01:38:33 AM EST
[ Parent ]
Forgot to mention that SCMT was developed after ETCS, violating the supposed standardization and adding yet one more system to install into the fray (now, if you want to run locomotives into Italy, you have to install also SCMT - doesn't that defeat the whole idea of standardization?).

In the video on Youtube I showed above, the driver was using exclusively AFB (you can see that as soon as the target speed changed, the brake/throttle was immediately changed).

A friend who works as a driver in DB Schenker told me that drivers in LZB territory depend much on AFB auto-pilot, especially on high-speed runs like the Taurus in the example, since it can be fast-reacting and smooth-braking at the same time (at least in theory)

N.F.

by nfotis on Tue Aug 9th, 2011 at 10:44:34 AM EST
[ Parent ]
SCMT was developed after ETCS

Do you know that for fact? I couldn't find when the development of SCMT started, only when deployment started (2003). However, that's close enough in time to ETCS (development start: 1996, first line tests: 1999, first commercial ETCS L1 line: 2001) to be parallel developments. At any rate, SCMT deployment started before the directives mandating ETCS on conventional lines (2004 onwards), and simultaneously with ETCS L2 deployment on new high-speed lines in Italy.

*Lunatic*, n.
One whose delusions are out of fashion.

by DoDo on Wed Aug 17th, 2011 at 10:34:33 AM EST
[ Parent ]
I couldn't find when the development of SCMT started

According to RFI, in late 1999, with test sections equipped by 2001.

*Lunatic*, n.
One whose delusions are out of fashion.

by DoDo on Wed Aug 17th, 2011 at 11:54:20 AM EST
[ Parent ]
In the video on Youtube I showed above, the driver was using exclusively AFB

It was a single run for the loco.

*Lunatic*, n.
One whose delusions are out of fashion.

by DoDo on Wed Aug 17th, 2011 at 12:04:45 PM EST
[ Parent ]
There is nothing wrong with using wireless for signalling (plane and shuttles must use wireless!)  The problem is in designing for failure modes.  The same applies to signalling - as DoDo mentioned, cable thieves are a problem just as devastating to signal strength as a failed transmitter.  Indeed, with sensible hardening wireless is probably safer.
by njh on Tue Aug 16th, 2011 at 11:39:41 PM EST
[ Parent ]
Having read of sudden train stoppages for that reason, I used to think that unsteady signal reception is a problem due to a fail-safe requirement of constant signal reception. However, in ETCS L2, the train only stops if the signal loss is at the end of a movement authority (in practice: at a block limit/signal), and there are software patches for both momentary and local wireless signal loss; and these seem to be working as the present ETCS L2 problems seem to be unrelated to signal reception (also see a commenter in one of my diaries who specialises in ETCS and denied that signal reception is the problem).

*Lunatic*, n.
One whose delusions are out of fashion.
by DoDo on Wed Aug 17th, 2011 at 09:45:52 AM EST
[ Parent ]

Display:

Occasional Series