Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Collision on Washington DC Metro

2 views
Skip to first unread message

Mizter T

unread,
Jun 22, 2009, 6:11:58 PM6/22/09
to
Reuters (via Reuters.com front page):
"Several dead and up to 55 injured in Washington, DC subway collision:
local media - 5:56pm EDT"

Washington Post:
http://voices.washingtonpost.com/getthere/2009/06/commuter_alert_red_line_derail.html
---<quote>---
Two Dead, Many Injured in Red Line Collision

A six-car Red Line train headed in the direction of Shady Grove
derailed between the Takoma and Fort Totten Metrorail stations and
then was involved in a collision with another train at 5 p.m., Two
people are dead, according to preliminary reports from Metro.

D.C. Fire/EMS spokesman Alan Etter said: "We have a number of
injuries. We have at least one fatality. And we have dozens of
injuries. The fire depatment is still in the proecess of extricating
people from the cars. We're using heavy rescue equipment to cut open
the cars to get whoever's trapped in there out."
---</quote>---

Peter Masson

unread,
Jun 23, 2009, 4:08:10 AM6/23/09
to

"Mizter T" <mizt...@gmail.com> wrote in message
news:998bb186-e146-4187...@e21g2000yqb.googlegroups.com...

Now 9 fatalities reported.
BBC report:
http://news.bbc.co.uk/1/hi/world/americas/8114003.stm

Peter

Light of Aria

unread,
Jun 23, 2009, 7:20:47 AM6/23/09
to

"Mizter T" <mizt...@gmail.com> wrote in message
news:998bb186-e146-4187...@e21g2000yqb.googlegroups.com...

Holy smoke. It would appear their construction standards and safety cage
technology appears to be behind us (in Europe), akin to how we were after
the Clapham Junction crash.


Miles Bader

unread,
Jun 23, 2009, 7:27:01 AM6/23/09
to
"Light of Aria" <light...@2009.no.valid.domain.com> writes:
> Holy smoke. It would appear their construction standards and safety cage
> technology appears to be behind us (in Europe), akin to how we were
> after the Clapham Junction crash.

If only they'd build subway cars out of fluffy pillows, we wouldn't have
these problems!

-Miles

--
We live, as we dream -- alone....

Neil Williams

unread,
Jun 23, 2009, 7:43:07 AM6/23/09
to
On Tue, 23 Jun 2009 12:20:47 +0100, "Light of Aria"
<light...@2009.no.valid.domain.com> wrote:

>Holy smoke. It would appear their construction standards and safety cage
>technology appears to be behind us (in Europe), akin to how we were after
>the Clapham Junction crash.

It's a light-rail metro. Would a Metrolink tram, for instance, have
survived better?

I must admit to being surprised at how it's just fallen to bits,
though.

Neil

--
Neil Williams
Put my first name before the at to reply.

David Hansen

unread,
Jun 23, 2009, 8:29:50 AM6/23/09
to
On Tue, 23 Jun 2009 12:20:47 +0100 someone who may be "Light of
Aria" <light...@2009.no.valid.domain.com> wrote this:-

>> http://voices.washingtonpost.com/getthere/2009/06/commuter_alert_red_line_derail.html

>Holy smoke. It would appear their construction standards and safety cage
>technology appears to be behind us (in Europe), akin to how we were after
>the Clapham Junction crash.

The crash at Clapham Junction involved elderly rolling stock. Most
of it stood up remarkably well and protected the occupants better
than one would expect if one listened only to the doom merchants.
The only really severe damage, almost total destruction, was to two
coaches. However, these two coaches bounced off the rear of a
stationary train and then ran in the narrow gap between the
stationary train and a train moving the other way, so destruction
was likely.

In the absence of Kryptonite modern designs would do better, but
people would not step out of them without a scratch. The relatively
modern local train at Ladbroke Grove did remarkably well in a much
higher speed crash, but one coach was almost totally destroyed and
another one was fairly badly damaged.

A modern underground train in London may well have behaved in much
the same way as the trains in Washington did. The important question
is why the two trains came into contact in Washington, having come
into contact at what appears to have been a considerable speed the
results were predictable. The design is fairly old, but there are
severe limits to what can be done, especially in a suburban railway.

--
David Hansen, Edinburgh
I will *always* explain revoked encryption keys, unless RIP prevents me
http://www.opsi.gov.uk/acts/acts2000/00023--e.htm#54

tsha...@aol.com

unread,
Jun 23, 2009, 8:36:55 AM6/23/09
to
On 23 June, 13:29, David Hansen <SENDdavidNOhS...@spidacom.co.uk>
wrote:

> A modern underground train in London may well have behaved in much
> the same way as the trains in Washington did.

"Ye canna change the laws o' physics, captain."

--
gordon

Miles Bader

unread,
Jun 23, 2009, 8:48:35 AM6/23/09
to
David Hansen <SENDdavi...@spidacom.co.uk> writes:
> A modern underground train in London may well have behaved in much
> the same way as the trains in Washington did. The important question
> is why the two trains came into contact in Washington, having come
> into contact at what appears to have been a considerable speed the
> results were predictable.

What seems bizarre is that some of the witness reports I read seemed to
say that the train actually started speeding up, to well above the
normal top speed when it collided (with a stationary train), and there
was no apparent attempt to brake...

-miles.

--
Success, n. The one unpardonable sin against one's fellows.

David Hansen

unread,
Jun 23, 2009, 9:14:20 AM6/23/09
to
On Tue, 23 Jun 2009 21:48:35 +0900 someone who may be Miles Bader
<mi...@gnu.org> wrote this:-

>What seems bizarre is that some of the witness reports I read seemed to
>say that the train actually started speeding up, to well above the
>normal top speed when it collided (with a stationary train), and there
>was no apparent attempt to brake...

While it is too early to speculate on the cause of this crash there
are historical examples which might explain this, if it happened.

I gather this line has Automatic Train Operation, the train normally
drives itself. In the early years of BART in San Francisco a train
crashed through the end of the line and ended up with half a coach
off the end of the viaduct. The train had been slowing down properly
for the station, following a series of commands from the track. It
then came to a command which meant something like, "coast but at not
more than 20mph". Because a component on the train had failed, or
been installed wrongly, the train interpreted this command as,
"accelerate to line speed", which the train did though it did not
get to line speed. By the time it came to the command to stop it was
moving too quickly to stop before reaching the end of the line.

That doesn't mean such a sequence of events was involved in this
crash, but perhaps it is one of what are probably still many
possibilities the investigators may be considering.

John B

unread,
Jun 23, 2009, 9:39:10 AM6/23/09
to
On Jun 23, 1:29 pm, David Hansen <SENDdavidNOhS...@spidacom.co.uk>
wrote:

> In the absence of Kryptonite modern designs would do better, but
> people would not step out of them without a scratch. The relatively
> modern local train at Ladbroke Grove did remarkably well in a much
> higher speed crash, but one coach was almost totally destroyed and
> another one was fairly badly damaged.

Unless the Washington Metro is the world's most epically awesome
subway, it seems unlikely that the closing speed between a stationary
underground train and an underground train pulling into a station
would be in the region of 130mph as at Ladbroke Grove. A closing speed
of 20-30mph should not lead to telescoping and significant fatalities
in modern rolling stock.

This NTSB report from 2006 may be relevant:
http://www.ntsb.gov/Pressrel/2006/060323.htm
...as it recommended the accelerated retirement of the type of rolling
stock used in the rear train in yesterday's accident, due to its
inadequate collision protection and vulnerability to telescoping.

--
John Band
john at johnband dot org
www.johnband.org

David Lesher

unread,
Jun 23, 2009, 9:53:21 AM6/23/09
to
"Light of Aria" <light...@2009.no.valid.domain.com> writes:

>Holy smoke. It would appear their construction standards and safety cage
>technology appears to be behind us (in Europe), akin to how we were after
>the Clapham Junction crash.


The front train is newer 5000 and 3000 series cars; the rear is all
30 yo 1000 series Rohrs. In past incidents, NTSB has recommended their
retirement/retrofiting; but there's a basic issue here:

If you hit a stationary ~450,000 lb [200000 Kg] consist with a
similar one, at 50 MPH [22 mps]; where exactly is the 0.5mv^2
energy supposed to go?


--
A host is a host from coast to coast.................wb8foz@nrk.com
& no one will talk to a host that's close........[v].(301) 56-LINUX
Unless the host (that isn't close).........................pob 1433
is busy, hung or dead....................................20915-1433

Christopher A. Lee

unread,
Jun 23, 2009, 10:08:34 AM6/23/09
to
On Tue, 23 Jun 2009 13:29:50 +0100, David Hansen
<SENDdavi...@spidacom.co.uk> wrote:

>A modern underground train in London may well have behaved in much
>the same way as the trains in Washington did. The important question
>is why the two trains came into contact in Washington, having come
>into contact at what appears to have been a considerable speed the
>results were predictable. The design is fairly old, but there are
>severe limits to what can be done, especially in a suburban railway.

In the 1990s a BART train was for some reason being worked wrong line
through the tunnels at Oakland, and came off at a turnout whose
extremely tightly curved diverging route was normally trailed through.
It was ripped open where the tunnels diverged

Scott M. Kozel

unread,
Jun 23, 2009, 3:31:03 PM6/23/09
to
David Lesher <wb8...@panix.com> wrote:
>
> "Light of Aria" <light...@2009.no.valid.domain.com> writes:
>
>> Holy smoke. It would appear their construction standards and safety cage
>> technology appears to be behind us (in Europe), akin to how we were after
>> the Clapham Junction crash.
>
> The front train is newer 5000 and 3000 series cars; the rear is all
> 30 yo 1000 series Rohrs. In past incidents, NTSB has recommended their
> retirement/retrofiting; but there's a basic issue here:
>
> If you hit a stationary ~450,000 lb [200000 Kg] consist with a
> similar one, at 50 MPH [22 mps]; where exactly is the 0.5mv^2
> energy supposed to go?

Was the impact speed that high in this accident?

David Lesher

unread,
Jun 23, 2009, 4:41:00 PM6/23/09
to
"Scott M. Kozel" <koz...@comcast.net> writes:

>>
>> If you hit a stationary ~450,000 lb [200000 Kg] consist with a
>> similar one, at 50 MPH [22 mps]; where exactly is the 0.5mv^2
>> energy supposed to go?

>Was the impact speed that high in this accident?

Sounds like it -- Various survivors have mentioned several stops with
fast runs between. It's a long, flat section with speed limits of about
that. [?58mph?]

Peter Masson

unread,
Jun 23, 2009, 5:04:32 PM6/23/09
to

"David Hansen" <SENDdavi...@spidacom.co.uk> wrote


>
> I gather this line has Automatic Train Operation, the train normally
> drives itself. In the early years of BART in San Francisco a train
> crashed through the end of the line and ended up with half a coach
> off the end of the viaduct. The train had been slowing down properly
> for the station, following a series of commands from the track. It
> then came to a command which meant something like, "coast but at not
> more than 20mph". Because a component on the train had failed, or
> been installed wrongly, the train interpreted this command as,
> "accelerate to line speed", which the train did though it did not
> get to line speed. By the time it came to the command to stop it was
> moving too quickly to stop before reaching the end of the line.
>

A recent RAIB report into a DLR derailment
http://www.raib.gov.uk/cms_resources/090622_R162009_Deptford%20Bridge.pdf
showed that trains had been exceeding temporary speed restrictions, although
this was not the cause of the derailment. The computer in the control centre
had to be rebooted quite frequently, and temporary speed restrictions had to
be re-entered manually. The control centre staff on occasions failed to do
this.

Peter

David Lesher

unread,
Jun 23, 2009, 6:23:56 PM6/23/09
to


The NSTB press conference reports:

A) It appears the train was in automatic train mode.

B) The speed limit there was 59 mph.

C) The emergency stop [mushroom] was depressed and brake wheels
blued [overheated] indicating it had been deployed pre-crash.

Charles Ellson

unread,
Jun 23, 2009, 7:51:43 PM6/23/09
to
On Tue, 23 Jun 2009 11:43:07 GMT, wensl...@pacersplace.org.uk (Neil
Williams) wrote:

>On Tue, 23 Jun 2009 12:20:47 +0100, "Light of Aria"
><light...@2009.no.valid.domain.com> wrote:
>
>>Holy smoke. It would appear their construction standards and safety cage
>>technology appears to be behind us (in Europe), akin to how we were after
>>the Clapham Junction crash.
>
>It's a light-rail metro. Would a Metrolink tram, for instance, have
>survived better?
>

A rear-ending on IIRC the Bakerloo Line at Neasden in the 1960s didn't
seem to tear things apart so easily. Even the tube train that was
rogered by a 313 at Willesden had less damage. The nearest comparison
damage-wise would seem to be the class 142 [?] that was
semi-demolished a few years back.

Scott M. Kozel

unread,
Jun 23, 2009, 9:33:58 PM6/23/09
to
David Lesher <wb8...@panix.com> wrote:
>
> "Scott M. Kozel" <koz...@comcast.net> writes:
>
>>> If you hit a stationary ~450,000 lb [200000 Kg] consist with a
>>> similar one, at 50 MPH [22 mps]; where exactly is the 0.5mv^2
>>> energy supposed to go?
>
>> Was the impact speed that high in this accident?
>
> Sounds like it -- Various survivors have mentioned several stops with
> fast runs between. It's a long, flat section with speed limits of about
> that. [?58mph?]

Then the fussing about how badly the car was crushed, was misplaced. I
would have thought that at that speed that 4 or 5 cars would have been
completely demolished.

--
Scott M. Kozel Highway and Transportation History Websites
Virginia/Maryland/Washington, D.C. http://www.roadstothefuture.com
Capital Beltway Projects http://www.capital-beltway.com
Philadelphia and Delaware Valley http://www.pennways.com

Miles Bader

unread,
Jun 23, 2009, 10:34:37 PM6/23/09
to
John B <sp...@johnband.org> writes:
> Unless the Washington Metro is the world's most epically awesome
> subway, it seems unlikely that the closing speed between a stationary
> underground train and an underground train pulling into a station
> would be in the region of 130mph as at Ladbroke Grove. A closing speed
> of 20-30mph should not lead to telescoping and significant fatalities
> in modern rolling stock.

haha, the phrase "world's most epically awesome subway" really made me
laugh... :)

The speed seems to have been significantly higher than 20-30mph, though,
in the 60-70mph range according to one report I saw (from a news-story
in the Washingon Post that I can no longer find though).

-Miles

--
The secret to creativity is knowing how to hide your sources.
--Albert Einstein

gl4...@yahoo.com

unread,
Jun 24, 2009, 1:39:41 AM6/24/09
to
In article <4a40bf9e...@news.individual.net>,
wensl...@pacersplace.org.uk (Neil Williams) wrote:

> On Tue, 23 Jun 2009 12:20:47 +0100, "Light of Aria"
> <light...@2009.no.valid.domain.com> wrote:
>
> >Holy smoke. It would appear their construction standards and safety cage
> >technology appears to be behind us (in Europe), akin to how we were after
> >the Clapham Junction crash.
>
> It's a light-rail metro. Would a Metrolink tram, for instance, have
> survived better?
>
> I must admit to being surprised at how it's just fallen to bits,
> though.


They were made by Rohr in the 1970s, when defense contractors were trying
to prove to the government that they could build better trains than the
train builders could. Being off the standard railroad network, they were
not required to meet Federal Railroad Administration safety standards.
Instead, a somewhat more chaotic set of regulations governs "rail tranist"
equipment, and at that time those standards were not very good.

Unfortunately, the same was at one time true of the automatic train
operation system both Washington Metro and BART used, though BART
definitely had the worse reputation of having many computer problems when
it first opened.

--
-Glennl
Please note this e-mail address is a pit of spam, and most e-mail sent to this address are simply lost in the vast mess.

Martin Edwards

unread,
Jun 24, 2009, 2:42:21 AM6/24/09
to

I once met a woman from Oakland who said that there were no railways in
her area. How did she grow up not knowing about BART?

Neil Williams

unread,
Jun 24, 2009, 3:33:11 AM6/24/09
to
On Wed, 24 Jun 2009 00:51:43 +0100, Charles Ellson
<cha...@ellson.demon.co.uk> wrote:

>A rear-ending on IIRC the Bakerloo Line at Neasden in the 1960s didn't
>seem to tear things apart so easily.

Isn't LUL unusual in being (in some ways) a heavy-rail metro, though?
And the smaller cross-section of the stock will make it stronger to
some extent anyway.

Neil Williams

unread,
Jun 24, 2009, 3:35:36 AM6/24/09
to
On Tue, 23 Jun 2009 21:33:58 -0400, "Scott M. Kozel"
<koz...@comcast.net> wrote:

>Then the fussing about how badly the car was crushed, was misplaced. I
>would have thought that at that speed that 4 or 5 cars would have been
>completely demolished.

In the Ladbroke Grove crash in the UK, only one coach of the Turbo
train (hit at a closing speed of around 125mph) was demolished. This
being the case, a closing speed of 50-60mph should certainly not
destroy 5 or 6, and the separation of the floor from the coach in this
instance is quite concerning - most UK stock is designed so that
wouldn't happen in such a collision.

David Hansen

unread,
Jun 24, 2009, 4:22:39 AM6/24/09
to
On Wed, 24 Jun 2009 00:51:43 +0100 someone who may be Charles Ellson
<cha...@ellson.demon.co.uk> wrote this:-

>A rear-ending on IIRC the Bakerloo Line at Neasden in the 1960s didn't
>seem to tear things apart so easily. Even the tube train that was
>rogered by a 313 at Willesden had less damage. The nearest comparison
>damage-wise would seem to be the class 142 [?] that was
>semi-demolished a few years back.

I have in one of my books a photograph of a train on either the
Central or Bakerloo lines, taken in the 1960s or 70s, where the only
thing holding the sides of the train together appears to be a chain
threaded through the windows. Presumably the weight of the roof was
pushing them apart, the floor being largely missing.

David Hansen

unread,
Jun 24, 2009, 4:38:28 AM6/24/09
to
On Wed, 24 Jun 2009 07:35:36 GMT someone who may be
wensl...@pacersplace.org.uk (Neil Williams) wrote this:-

>In the Ladbroke Grove crash in the UK, only one coach of the Turbo
>train (hit at a closing speed of around 125mph) was demolished. This
>being the case, a closing speed of 50-60mph should certainly not
>destroy 5 or 6,

It doesn't look like 5 or 6 coaches were destroyed, from what I saw
of the photographs.

Neil Williams

unread,
Jun 24, 2009, 5:03:23 AM6/24/09
to
On Wed, 24 Jun 2009 09:38:28 +0100, David Hansen
<SENDdavi...@spidacom.co.uk> wrote:

>It doesn't look like 5 or 6 coaches were destroyed, from what I saw
>of the photographs.

I know, but my post was in response to someone who suggested that
might almost be expected performance.

John B

unread,
Jun 24, 2009, 5:48:12 AM6/24/09
to

Or indeed, the various Amtrak lines around Oakland:
http://en.wikipedia.org/wiki/Capitol_Corridor

Mr Thant

unread,
Jun 24, 2009, 6:53:38 AM6/24/09
to
On 24 June, 08:33, wensleyd...@pacersplace.org.uk (Neil Williams)
wrote:

> Isn't LUL unusual in being (in some ways) a heavy-rail metro, though?

Washington Metro cars are 23m long and weigh 33 tonnes each. That puts
them in the same ballpark as most mainline multiple unit stock in the
UK (which may well be considered "light rail" in the US).

London Underground stock is generally well under 20m and 20-25 tonnes
per car, which actually puts it closer to DLR and Manchester Metrolink
stock (if you divide each tram/unit by two).

U

Scott M. Kozel

unread,
Jun 24, 2009, 7:13:22 AM6/24/09
to
wensl...@pacersplace.org.uk (Neil Williams) wrote:
>
> "Scott M. Kozel" <koz...@comcast.net> wrote:
>
>> Then the fussing about how badly the car was crushed, was misplaced. I
>> would have thought that at that speed that 4 or 5 cars would have been
>> completely demolished.
>
> In the Ladbroke Grove crash in the UK, only one coach of the Turbo
> train (hit at a closing speed of around 125mph) was demolished. This
> being the case, a closing speed of 50-60mph should certainly not
> destroy 5 or 6, and the separation of the floor from the coach in this
> instance is quite concerning - most UK stock is designed so that
> wouldn't happen in such a collision.

In the high-speed ICE train crash a few years ago, several cars were
demolished. The scene looked like a plane crash.

With the kinetic energy of two 6-car trains colliding at 50 to 60 mph,
there will be severe damage to at least the two cars that hit.


Christopher A. Lee

unread,
Jun 24, 2009, 9:22:59 AM6/24/09
to
On Tue, 23 Jun 2009 22:39:41 -0700, gl4...@yahoo.com
(gl4...@yahoo.com) wrote:

>In article <4a40bf9e...@news.individual.net>,
>wensl...@pacersplace.org.uk (Neil Williams) wrote:
>
>> On Tue, 23 Jun 2009 12:20:47 +0100, "Light of Aria"
>> <light...@2009.no.valid.domain.com> wrote:
>>
>> >Holy smoke. It would appear their construction standards and safety cage
>> >technology appears to be behind us (in Europe), akin to how we were after
>> >the Clapham Junction crash.
>>
>> It's a light-rail metro. Would a Metrolink tram, for instance, have
>> survived better?
>>
>> I must admit to being surprised at how it's just fallen to bits,
>> though.
>
>They were made by Rohr in the 1970s, when defense contractors were trying
>to prove to the government that they could build better trains than the
>train builders could. Being off the standard railroad network, they were
>not required to meet Federal Railroad Administration safety standards.
>Instead, a somewhat more chaotic set of regulations governs "rail tranist"
>equipment, and at that time those standards were not very good.

Rohr - say no more.

>Unfortunately, the same was at one time true of the automatic train
>operation system both Washington Metro and BART used, though BART
>definitely had the worse reputation of having many computer problems when
>it first opened.

With BART it wasn't just computer problems but also incredibly shoddy
electrical work in the trains which were also late and overweight - by
Rohr who were trying to break into rail. They began in WW2 by
repairing military aircraft, and rather like Boeing Vertol moved into
transit with no knowledge of it and produced (to put it bluntly) crap.

Christopher A. Lee

unread,
Jun 24, 2009, 9:41:12 AM6/24/09
to

Depends what she meant by her area.

San Francisco Bay is like a 100_ mile long thin conurbation bent into
a U shape. It's flat along the waterside, but gets into steep hills as
you go inland.

Some of the communities like Union City where I lived, were almost
entirely in the flat part. But other older ones like Oakland expanded
a long time ago up into the hills. These were a lot more miles wide
than deep.

The railways were built where it is flat.

In Union City we had three. The Southern Pacific closest to the water
and the Western Pacific just below where the hills started. Both these
are now part of the Union Pacific. Then came BART which followed the
Western Pacific route on stilts as it were.

Downtown Oakland is on the flat part near the docks. The main lines
converges at a huge yard also near the docks, and BART obviously went
via downtown.

So your friend could have lived in the hills, several miles East of
any railway. Not just BART but also Amtrak which apart from the daily
Seattle to Los Angeles and the transcontinental to Chicago, also had
trains to Sacramento, Bakersfield and San Jose.
In Oakland the railways converged on the docks, and the

chunkyoldcortina

unread,
Jun 24, 2009, 12:03:53 PM6/24/09
to
Scott M. Kozel wrote:
> David Lesher <wb8...@panix.com> wrote:
>>
>> "Scott M. Kozel" <koz...@comcast.net> writes:
>>
>>>> If you hit a stationary ~450,000 lb [200000 Kg] consist with a
>>>> similar one, at 50 MPH [22 mps]; where exactly is the 0.5mv^2
>>>> energy supposed to go?
>>
>>> Was the impact speed that high in this accident?
>>
>> Sounds like it -- Various survivors have mentioned several stops with
>> fast runs between. It's a long, flat section with speed limits of about
>> that. [?58mph?]
>
> Then the fussing about how badly the car was crushed, was misplaced. I
> would have thought that at that speed that 4 or 5 cars would have been
> completely demolished.
>

But better design including crumple zones and monocoque construction (rather
than separate body and underframe) can still help.

Look at the differences here

http://www.youtube.com/watch?v=L3whFgsNKUQ

Roger T.

unread,
Jun 24, 2009, 1:05:31 PM6/24/09
to
>
> But better design including crumple zones and monocoque construction
> (rather than separate body and underframe) can still help.
>
> Look at the differences here
>
> http://www.youtube.com/watch?v=L3whFgsNKUQ

Very impressive.


--
Cheers.

Roger T.
See the GER at: -
http://www.islandnet.com/~rogertra/


Mark Brader

unread,
Jun 24, 2009, 4:36:26 PM6/24/09
to
Martin Edwards:

>> I once met a woman from Oakland who said that there were no railways in
>> her area. How did she grow up not knowing about BART?

Christopher Lee:

> Depends what she meant by her area.

Also, a "railroad" to many in North America is a separate thing from
a "subway"(*). BART is rather like London's Metropolitan Line, with
long suburban sections above ground but also an important underground
section in downtown San Francisco and Oakland, so it's perfectly
reasonable to classify it as a subway.

(*) Which, of course, brings us right back to the previous point about
crashworthiness standards.
--
Mark Brader, Toronto | "If you wish so, we write your consummations
m...@vex.net | on your bill." --Swiss hotel services handbook

My text in this article is in the public domain.

Mark Brader

unread,
Jun 24, 2009, 5:05:45 PM6/24/09
to
Scott Kozel:

> With the kinetic energy of two 6-car trains colliding at 50 to 60 mph,
> there will be severe damage to at least the two cars that hit.

Note that "closing speed" is not the right measure when you're talking
about kinetic energy -- because it varies as v^2, you need the speeds
of the two individual trains. Energy-wise, a 50 mph crash into a
stationary train (which is the case here) is twice as bad as a head-on
collision of two trains the same size each moving at 25 mph.
--
Mark Brader That would be the opposite of "non idiotic",
Toronto assuming there's some good word for that.
m...@vex.net --Ken Jennings

Christopher A. Lee

unread,
Jun 24, 2009, 5:07:56 PM6/24/09
to
On Wed, 24 Jun 2009 15:36:26 -0500, m...@vex.net (Mark Brader) wrote:

>Martin Edwards:
>>> I once met a woman from Oakland who said that there were no railways in
>>> her area. How did she grow up not knowing about BART?
>
>Christopher Lee:
>> Depends what she meant by her area.
>
>Also, a "railroad" to many in North America is a separate thing from
>a "subway"(*). BART is rather like London's Metropolitan Line, with
>long suburban sections above ground but also an important underground
>section in downtown San Francisco and Oakland, so it's perfectly
>reasonable to classify it as a subway.

She still had Amtrak, both transcontinental, long distance and
regional. Not to mention huge amounts of freight to and from the Port
of Oakland which is one of the major destinations.

>(*) Which, of course, brings us right back to the previous point about
>crashworthiness standards.

General American railway safety standards aren't up to those of the UK
let alone Western Europe.

Prime example:

Most commuter lines share track with heavy freight, and signalling is
for the freight with a few miles between the yellow and the red. So a
few years back a commuter train stopped at what amounted to an
unstaffed halt after the yellow. The engineer forgot there was a red
ahead, set off at speed, passed it at a junction and hit a train on
the main line coming in the opposite direction.

Scott M. Kozel

unread,
Jun 24, 2009, 5:45:40 PM6/24/09
to

Interesting ... I've never seen a railroad crash test before. A major
part of the energy transfer is by pushing the lead train at a
considerable speed.

John B

unread,
Jun 24, 2009, 5:59:04 PM6/24/09
to
On Jun 24, 10:07 pm, Christopher A. Lee <ca...@optonline.net> wrote:
> >(*) Which, of course, brings us right back to the previous point about
> >crashworthiness standards.
>
> General American railway safety standards aren't up to those of the UK
> let alone Western Europe.

I"m sceptical (in a '99% sure you're wrong') sense that the rest of
Western Europe has higher safety standards than the UK (aside from the
point that the UK is a subset of Western Europe, not a separate
entity).

Clark F Morris

unread,
Jun 24, 2009, 7:44:21 PM6/24/09
to
On Wed, 24 Jun 2009 07:13:22 -0400, "Scott M. Kozel"
<koz...@comcast.net> wrote:

In this case, the impact speed was probably no more than 60 mph / 96
kmh. The ICE crash was at 100 - 125 mph / 160 - 200 kmh. The cars on
the ICE were also aluminum bodies on a steel underframe like the Rohr
cars involved in the crash. Assuming a better body shell integrity
and strength would have made noticeable difference given the forces
involved is something that seems intuitive but really should be
analyzed carefully to verify there is a highly probable net
difference.

Scott M. Kozel

unread,
Jun 24, 2009, 9:23:13 PM6/24/09
to

The ICE train also hit bridge piers and abutment, which are fixed and
essentially immovable objects.

Charles Ellson

unread,
Jun 24, 2009, 10:27:39 PM6/24/09
to
On Wed, 24 Jun 2009 09:22:39 +0100, David Hansen
<SENDdavi...@spidacom.co.uk> wrote:

>On Wed, 24 Jun 2009 00:51:43 +0100 someone who may be Charles Ellson
><cha...@ellson.demon.co.uk> wrote this:-
>
>>A rear-ending on IIRC the Bakerloo Line at Neasden in the 1960s didn't
>>seem to tear things apart so easily. Even the tube train that was
>>rogered by a 313 at Willesden had less damage. The nearest comparison
>>damage-wise would seem to be the class 142 [?] that was
>>semi-demolished a few years back.
>
>I have in one of my books a photograph of a train on either the
>Central or Bakerloo lines, taken in the 1960s or 70s, where the only
>thing holding the sides of the train together appears to be a chain
>threaded through the windows. Presumably the weight of the roof was
>pushing them apart, the floor being largely missing.
>

If it has reached the lifting stage then there is a good chance
someone has already chopped things into convenient bits. Even with
tube trains the chassis has a fair amount of strength in it and (as
IIRC applied with the 313 v tube train) will usually catch the bogie
of any larger vehicle although ISTR one rear-ending by a battery
locomotive did cause a degree of over-riding but not the peeling apart
that seems to have occurred in the US accident.

Mark Brader

unread,
Jun 24, 2009, 11:13:26 PM6/24/09
to
Martin Edwards:
>>>> I once met a woman from Oakland who said that there were no railways
>>>> in her area. How did she grow up not knowing about BART?

Christopher Lee:
>>> Depends what she meant by her area.

Mark Brader:


>> Also, a "railroad" to many in North America is a separate thing from a

>> "subway" ... it's perfectly reasonable to classify [BART] as a subway.

Christopher Lee:


> She still had Amtrak, both transcontinental, long distance and

> regional. Not to mention huge amounts of freight ...

Uh-huh. But see your own previous remark. "Area" is a very vague term.
--
Mark Brader "...there are other means of persuasion
m...@vex.net besides killing and threatening to kill."
Toronto --Dashiell Hammett, The Maltese Falcon

Martin Edwards

unread,
Jun 25, 2009, 2:31:30 AM6/25/09
to

Thanks for the detail. :-)

David Hansen

unread,
Jun 25, 2009, 3:50:59 AM6/25/09
to
On Thu, 25 Jun 2009 03:27:39 +0100 someone who may be Charles Ellson
<cha...@ellson.demon.co.uk> wrote this:-

>>I have in one of my books a photograph of a train on either the


>>Central or Bakerloo lines, taken in the 1960s or 70s, where the only
>>thing holding the sides of the train together appears to be a chain
>>threaded through the windows. Presumably the weight of the roof was
>>pushing them apart, the floor being largely missing.
>>
>If it has reached the lifting stage then there is a good chance
>someone has already chopped things into convenient bits.

This was at the run it along the track on a spare bogie stage. The
current fashion in the UK for lifting everything with a road crane
had not yet come into being.

It is possible bits had been removed after the crash, though I doubt
it as the priority then was to get the line open a soon as possible.

Graham Murray

unread,
Jun 25, 2009, 8:30:29 AM6/25/09
to
m...@vex.net (Mark Brader) writes:

> Note that "closing speed" is not the right measure when you're talking
> about kinetic energy -- because it varies as v^2, you need the speeds
> of the two individual trains. Energy-wise, a 50 mph crash into a
> stationary train (which is the case here) is twice as bad as a head-on
> collision of two trains the same size each moving at 25 mph.

What about one where a train travelling at 150mph runs into the back of
one travelling (in the same direction) at 100mph?

Peter Campbell Smith

unread,
Jun 25, 2009, 11:13:57 AM6/25/09
to
> m...@vex.net (Mark Brader) writes:
>
>> Note that "closing speed" is not the right measure when you're talking
>> about kinetic energy -- because it varies as v^2, you need the speeds
>> of the two individual trains. Energy-wise, a 50 mph crash into a
>> stationary train (which is the case here) is twice as bad as a head-on
>> collision of two trains the same size each moving at 25 mph.

Graham Murray <news...@gmurray.org.uk> wrote in
news:87my7wh...@newton.gmurray.org.uk:



> What about one where a train travelling at 150mph runs into the back of
> one travelling (in the same direction) at 100mph?

Mark is correct in saying that the amount of kinetic energy which has to be
dissipated in the '50-0' crash is double what it is in the '25-25' one.
However, the amount which is dissipated in the initial 'bang' is more
nearly equal and roughly proportional to the speed differential.

If you assume that the trains are of equal mass and entirely inelastic, in
the 25-25 case they would hit, crumple and remain where they were. In the
50-0 case they would hit, crumple, and the merged mass would move off in
the direction the moving train was going at 25mph, and then presumably the
brakes or dragging wreckage would stop it, thereby absorbing some of the
energy in a relatively benign fashion.

In Graham's example, they would hit, crumple and continue moving, initially
at 125mph.

Damage is (approximately) done by force, and force is proportional to
deceleration. Stronger trains resist being deformed by larger forces.
Trains with crumple zones decelerate less rapidly in an accident than more
rigid ones. Both of these help to reduce the force applied to the
passengers.

Looked at another way, energy is force times distance, so the longer the
distance you travel while dissipating the energy, the less the force.

Peter

--
Peter Campbell Smith ~ London ~ pjcs00 (a) gmail.com

spsffan

unread,
Jun 25, 2009, 5:22:01 PM6/25/09
to
Martin Edwards wrote:
> Christopher A. Lee wrote:
>> On Tue, 23 Jun 2009 13:29:50 +0100, David Hansen
>> <SENDdavi...@spidacom.co.uk> wrote:
>>
>>> A modern underground train in London may well have behaved in much
>>> the same way as the trains in Washington did. The important question
>>> is why the two trains came into contact in Washington, having come
>>> into contact at what appears to have been a considerable speed the
>>> results were predictable. The design is fairly old, but there are
>>> severe limits to what can be done, especially in a suburban railway.
>>
>> In the 1990s a BART train was for some reason being worked wrong line
>> through the tunnels at Oakland, and came off at a turnout whose
>> extremely tightly curved diverging route was normally trailed through.
>> It was ripped open where the tunnels diverged
>
> I once met a woman from Oakland who said that there were no railways in
> her area. How did she grow up not knowing about BART?


While Oakland has plenty of railways (they run right down the middle of
the street at Jack London Square) according to Gertrude Stein, "There's
no "there" there.

Perhaps she meant streetcars??

Martin Edwards

unread,
Jun 26, 2009, 2:21:19 AM6/26/09
to
Didn't she mean LA?

David Hansen

unread,
Jun 26, 2009, 3:22:22 AM6/26/09
to
On Mon, 22 Jun 2009 15:11:58 -0700 (PDT) someone who may be Mizter T
<mizt...@gmail.com> wrote this:-

>Reuters (via Reuters.com front page):
>"Several dead and up to 55 injured in Washington, DC subway collision:
>local media - 5:56pm EDT"

The Washington Post has reported (what certainly in the UK would be
called) a wrong-side track circuit failure.

I did see a reference to this on a UK newspaper web site last night,
before Michael Jackson coverage edged it out.

Roland Perry

unread,
Jun 26, 2009, 9:56:52 AM6/26/09
to
In message <vft845d3pgnj03u53...@4ax.com>, at 08:22:22 on
Fri, 26 Jun 2009, David Hansen <SENDdavi...@spidacom.co.uk>
remarked:

>The Washington Post has reported (what certainly in the UK would be
>called) a wrong-side track circuit failure.

http://www.bloomberg.com/apps/news?pid=20601087&sid=afA2NXqddOrU
<http://www.latimes.com/news/nationworld/nation/la-na-train26-2009jun26,0
,4470392.story>
--
Roland Perry

Peter Campbell Smith

unread,
Jun 26, 2009, 10:27:59 AM6/26/09
to
> In message <vft845d3pgnj03u53...@4ax.com>, at 08:22:22 on
> Fri, 26 Jun 2009, David Hansen <SENDdavi...@spidacom.co.uk>
> remarked:

>>The Washington Post has reported (what certainly in the UK would be
>>called) a wrong-side track circuit failure.

Track circuit failures (false positives and false negatives) are not
uncommon in any railway, especially ones in the open air. Because of that,
the automated systems I have been involved with track the progress of
individual trains, and if one 'disappears', it is assumed to have stopped
at the last point at which it was detected.

Usually the operating procedure is for the following train to proceed
slowly to confirm that the section is clear -- or not. Some systems will
assume that if a train is seen at A, then disappears, then reappears at C,
that the detection at B is faulty and that B is now clear. This is a safe
assumption if train detection can determine the identity of the train at C
and doesn't just report occupancy of the section.

I don't know the Washington system in detail, but I don't think it's
credible that failure of a single track circuit (or whatever they use to
detect the presence of trains) would cause such a devastating accident.

Perhaps significantly, the Washington Post is reporting that the stopped
train was being driven in manual because the driver saw that the platform
ahead was occupied. That suggests to me a more complex failure.

Roland Perry

unread,
Jun 26, 2009, 11:01:09 AM6/26/09
to
In message <Xns9C369D5...@news.albasani.net>, at 14:27:59 on Fri,
26 Jun 2009, Peter Campbell Smith <see.sig....@ct.details>
remarked:

>I don't know the Washington system in detail, but I don't think it's
>credible that failure of a single track circuit (or whatever they use to
>detect the presence of trains) would cause such a devastating accident.

The press reports give the impression that the extent of the automatic
running is to determine if a train is in the section ahead, and stop if
one is.

>Perhaps significantly, the Washington Post is reporting that the
>stopped train was being driven in manual because the driver saw that
>the platform ahead was occupied. That suggests to me a more complex
>failure.

Maybe the circuit protecting the train in the platform had failed too?

(Or in both cases, the equipment communicating the state of the circuit
malfunctioning).
--
Roland Perry

spsffan

unread,
Jun 26, 2009, 6:24:00 PM6/26/09
to

No. That's "100 suburbs in search of a city." but I don't recall who
said it.

Matthew Geier

unread,
Jun 26, 2009, 6:35:14 PM6/26/09
to
On Fri, 26 Jun 2009 14:27:59 +0000, Peter Campbell Smith wrote:

>> In message <vft845d3pgnj03u53...@4ax.com>, at 08:22:22 on
>> Fri, 26 Jun 2009, David Hansen <SENDdavi...@spidacom.co.uk>
>> remarked:
>
>>>The Washington Post has reported (what certainly in the UK would be
>>>called) a wrong-side track circuit failure.
>
> Track circuit failures (false positives and false negatives) are not
> uncommon in any railway, especially ones in the open air.

> I don't know the Washington system in detail, but I don't think it's
> credible that failure of a single track circuit (or whatever they use to
> detect the presence of trains) would cause such a devastating accident.

One of the articles I scanned said 'coded track circuit', which would be
transmitting a speed code to the train. That could cause an automation
failure. There has to be some system for governing the trains in auto
mode. Speed coded track circuits seems obvious.

> Perhaps significantly, the Washington Post is reporting that the stopped
> train was being driven in manual because the driver saw that the
> platform ahead was occupied. That suggests to me a more complex
> failure.

One would presume that if the train 'lost' the speed code, it would
stop, may not emergency stop, but would stop.
But a fail safe system has just failed - what it had failed in such a
manner that it was transmitting a higher speed code ?.

It would a worst-case for the operator, train in auto mode, probably not
paying all that much attention, the train is driving itself, has done
safely for 100s of shifts before. Suddenly the train speeds up when it
should be slowing. Just how much reaction time are we going to get out of
a human in this situation ?. Probably not much more than 'Ohh @$@# i'm
going to die'. Not a situation i'd like to be in.

Automation is a double edged sword - great increases in capacity and
efficiency can be gained, but the remaining humans in the loop can get de-
skilled and their reaction times dulled.

I doubt the train ahead being in manual is significant, only that that's
train's operator may have passed over the faulty coded track circuit and
not noticed the allowable speed suddenly jumping.

A failed track circuit shouldn't result in a runaway train. A coded
track circuit transmitting the wrong code should be as near as possible
be engineered to be impossible. The investigating engineers have their
work cut out - all possible failure modes are going to have to be
meticulously investigated to ensure it can't happen again.

There are automated metro's all over the world that rely on 'fail safe'
methods to transmit the allowable speed or the 'distance to go' to the
train. I imagine a lot of people are watching this disaster very very
closely and wondering if their own systems could have such a disastrous
wrong side failure.

David Lesher

unread,
Jun 26, 2009, 10:15:47 PM6/26/09
to
Matthew Geier <mat...@no.sleeper.no.apana.no.org.no.au> writes:


>One of the articles I scanned said 'coded track circuit', which would be
>transmitting a speed code to the train. That could cause an automation
>failure. There has to be some system for governing the trains in auto
>mode. Speed coded track circuits seems obvious.

True; read the NTSB report on Shady Grove; I posted the URL's in
m.t.r.a yesterday,


> One would presume that if the train 'lost' the speed code, it would
>stop, may not emergency stop, but would stop.

That it would.

> But a fail safe system has just failed - what it had failed in such a
>manner that it was transmitting a higher speed code ?.

There is no indication of that. 59 MPH was the limiting speed
in that block.

> It would a worst-case for the operator, train in auto mode, probably not
>paying all that much attention, the train is driving itself, has done
>safely for 100s of shifts before. Suddenly the train speeds up when it
>should be slowing.

No indication that 112 lost signal. Rather it appears that 214 did
not trip its track circuit; so the ATP had no reason to stop 112....as
far as it knew..

>a human in this situation ?. Probably not much more than 'Ohh @$@# i'm
>going to die'. Not a situation i'd like to be in.

My SWAG is she needed ~900 ft to stop from 59 MPH [3.2 MPH/s is the spec]
but she had ~300 ft.

> A failed track circuit shouldn't result in a runaway train.

Did someone say it did?


--
A host is a host from coast to coast.................wb8foz@nrk.com
& no one will talk to a host that's close........[v].(301) 56-LINUX
Unless the host (that isn't close).........................pob 1433
is busy, hung or dead....................................20915-1433

David Hansen

unread,
Jun 27, 2009, 12:39:13 PM6/27/09
to
On Fri, 26 Jun 2009 14:27:59 +0000 (UTC) someone who may be Peter
Campbell Smith <see.sig....@ct.details> wrote this:-

>Track circuit failures (false positives and false negatives) are not
>uncommon in any railway, especially ones in the open air. Because of that,
>the automated systems I have been involved with track the progress of
>individual trains, and if one 'disappears', it is assumed to have stopped
>at the last point at which it was detected.

Certainly "modern" computerised systems tend to report if sections
are occupied out of the usual sequence, which may indicate a
detection failure. However, the first of these was in the UK in the
mid 1970s, so the system in Washington is probably not as advanced.
My guess would be something similar to the Victoria Line system,
computer monitored at best, rather than computerised.

David Hansen

unread,
Jun 27, 2009, 12:45:06 PM6/27/09
to
On 26 Jun 2009 22:35:14 GMT someone who may be Matthew Geier
<mat...@no.sleeper.no.apana.no.org.no.au> wrote this:-

> One would presume that if the train 'lost' the speed code, it would
>stop, may not emergency stop, but would stop.

I would guess that there is a code for stop, which is a normal
condition, so that can be differentiated by the train from the
abnormal condition of no code. The train would probably react
differently to both situations.

A great deal of thought was put into the rather old-fashioned art of
coded track circuits [1]. IIRC on the Victoria Line it was even
arranged so that the normal running codes would not start a stopped
train. Rather it had to receive a special code which would allow it
to start, after which it would continue on the normal running codes.

[1] IIRC initial installations of coded track circuits were made in
the 1920s or 1930s, for a cab signalling system on steam locomotives
of the New York Central Railroad.

James Robinson

unread,
Jun 27, 2009, 9:21:37 PM6/27/09
to
David Hansen <SENDdavi...@spidacom.co.uk> wrote:
>
> [1] IIRC initial installations of coded track circuits were made in
> the 1920s or 1930s, for a cab signalling system on steam locomotives
> of the New York Central Railroad.

The Pennsylvania Railroad, actually, developed by the Union Switch and
Signal Company. The first line controlled by cab signals was opened in
1923.

It was considered the first application of industrial electronics to a role
other than person to person communications. It required the development of
an amplifier that could reliably detect the very small currents induced in
detector coils suspended above the rails in front of the leading wheels of
a locomotive.

David Hansen

unread,
Jun 28, 2009, 5:37:21 AM6/28/09
to
On Sun, 28 Jun 2009 01:21:37 +0000 (UTC) someone who may be James
Robinson <was...@212.com> wrote this:-

>The Pennsylvania Railroad, actually, developed by the Union Switch and
>Signal Company. The first line controlled by cab signals was opened in
>1923.
>
>It was considered the first application of industrial electronics to a role
>other than person to person communications. It required the development of
>an amplifier that could reliably detect the very small currents induced in
>detector coils suspended above the rails in front of the leading wheels of
>a locomotive.

Thanks. Presumably detecting such currents is a far more difficult
task than detecting currents induced by say an Indusi magnet.

James Robinson

unread,
Jun 28, 2009, 7:40:37 AM6/28/09
to
David Hansen <SENDdavi...@spidacom.co.uk> wrote:

> On Sun, 28 Jun 2009 01:21:37 +0000 (UTC) someone who may be James
> Robinson <was...@212.com> wrote this:-
>
>>The Pennsylvania Railroad, actually, developed by the Union Switch and
>>Signal Company. The first line controlled by cab signals was opened
>>in 1923.
>>
>>It was considered the first application of industrial electronics to a
>>role other than person to person communications. It required the
>>development of an amplifier that could reliably detect the very small
>>currents induced in detector coils suspended above the rails in front
>>of the leading wheels of a locomotive.
>
> Thanks. Presumably detecting such currents is a far more difficult
> task than detecting currents induced by say an Indusi magnet.

The technology in 1920 was more limited than what is available today. I
don't think the detection of either is a particular problem today. I don't
really know how the relative difficulty of the two compares.

David Hansen

unread,
Jun 28, 2009, 9:48:51 AM6/28/09
to
On Sun, 28 Jun 2009 11:40:37 +0000 (UTC) someone who may be James

Robinson <was...@212.com> wrote this:-

>>>It was considered the first application of industrial electronics to a


>>>role other than person to person communications. It required the
>>>development of an amplifier that could reliably detect the very small
>>>currents induced in detector coils suspended above the rails in front
>>>of the leading wheels of a locomotive.
>>
>> Thanks. Presumably detecting such currents is a far more difficult
>> task than detecting currents induced by say an Indusi magnet.
>
>The technology in 1920 was more limited than what is available today.

I should have perhaps used the phrase, "detecting such currents was
a far more difficult", though perhaps not as the passage of time has
not altered the physics.

>I
>don't think the detection of either is a particular problem today. I don't
>really know how the relative difficulty of the two compares.

In my view a rail is not specifically designed as an aerial; while
Indusi, TPWS or ETCS transponders, to name a few, are designed as
such. So I think that the task of detecting coded track circuits is
more difficult, other things being reasonably equal.

James Robinson

unread,
Jun 28, 2009, 11:04:45 AM6/28/09
to
David Hansen <SENDdavi...@spidacom.co.uk> wrote:

> On Sun, 28 Jun 2009 11:40:37 +0000 (UTC) someone who may be James
> Robinson <was...@212.com> wrote this:-
>
>>>>It was considered the first application of industrial electronics to
>>>>a role other than person to person communications. It required the
>>>>development of an amplifier that could reliably detect the very
>>>>small currents induced in detector coils suspended above the rails
>>>>in front of the leading wheels of a locomotive.
>>>
>>> Thanks. Presumably detecting such currents is a far more difficult
>>> task than detecting currents induced by say an Indusi magnet.
>>
>>The technology in 1920 was more limited than what is available today.
>
> I should have perhaps used the phrase, "detecting such currents was
> a far more difficult", though perhaps not as the passage of time has
> not altered the physics.
>
>>I
>>don't think the detection of either is a particular problem today. I
>>don't really know how the relative difficulty of the two compares.
>
> In my view a rail is not specifically designed as an aerial; while
> Indusi, TPWS or ETCS transponders, to name a few, are designed as
> such. So I think that the task of detecting coded track circuits is
> more difficult, other things being reasonably equal.

I'm not sure it's more difficult, even though rails weren't specifically
designed to be aerials. Coded track circuits have been successfully used
all around the world with very high safety, and they continue to be used
for the application.

It appears that economics rather than physics is more of a determining
factor for the choice of technology.

Stephen Sprunk

unread,
Jun 28, 2009, 12:26:31 PM6/28/09
to
James Robinson wrote:
> David Hansen <SENDdavi...@spidacom.co.uk> wrote:
>> On Sun, 28 Jun 2009 11:40:37 +0000 (UTC) someone who may be James
>> Robinson <was...@212.com> wrote this:-
>>> I don't think the detection of either [currents or magnets] is a

>>> particular problem today. I don't really know how the relative
>>> difficulty of the two compares.
>>
>> In my view a rail is not specifically designed as an aerial; while
>> Indusi, TPWS or ETCS transponders, to name a few, are designed as
>> such. So I think that the task of detecting coded track circuits is
>> more difficult, other things being reasonably equal.
>
> I'm not sure it's more difficult, even though rails weren't specifically
> designed to be aerials. Coded track circuits have been successfully used
> all around the world with very high safety, and they continue to be used
> for the application.
>
> It appears that economics rather than physics is more of a determining
> factor for the choice of technology.

To an engineer, "difficulty" applies only to the intellectual work
needed to figure out how to solve a problem. Once the solution is
known, the problem is then "trivial" and requires only the application
of sufficient resources (time, money, etc.) to implement.

Coded track circuits may have been more difficult to figure out than
magnetic or radio transponders, but the solutions are all known now, so
the only difference is how much each solution costs.

S

--
Stephen Sprunk "Stupid people surround themselves with smart
CCIE #3723 people. Smart people surround themselves with
K5SSS smart people who disagree with them." --Isaac Jaffe

Clive D. W. Feather

unread,
Jun 28, 2009, 3:30:05 PM6/28/09
to
In message <Xns9C369D5...@news.albasani.net>, Peter Campbell
Smith <see.sig....@ct.details> wrote:
>>>The Washington Post has reported (what certainly in the UK would be
>>>called) a wrong-side track circuit failure.
>
>Track circuit failures (false positives and false negatives) are not
>uncommon in any railway, especially ones in the open air.

Please provide cites.

Right-side failures (reporting a train present when there's no train) is
not uncommon, having a variety of causes.

Wrong-side failures (failing to detect a train) are always a significant
concern and I'm skeptical about your "not uncommon". The last
significant incidence I'm aware of was the problem with Pacers on
lightly-used track, requiring actuators to be fitted.

>I don't know the Washington system in detail, but I don't think it's
>credible that failure of a single track circuit (or whatever they use to
>detect the presence of trains) would cause such a devastating accident.

Nor do I.

--
Clive D.W. Feather | Home: <cl...@davros.org>
Mobile: +44 7973 377646 | Web: <http://www.davros.org>
Please reply to the Reply-To address, which is: <cl...@davros.org>

Clive D. W. Feather

unread,
Jun 28, 2009, 3:38:27 PM6/28/09
to
In message <iqic45pm4n9a9dtcd...@4ax.com>, David Hansen
<SENDdavi...@spidacom.co.uk> wrote:
>I would guess that there is a code for stop, which is a normal
>condition, so that can be differentiated by the train from the
>abnormal condition of no code. The train would probably react
>differently to both situations.

Not necessarily.

>A great deal of thought was put into the rather old-fashioned art of
>coded track circuits [1]. IIRC on the Victoria Line it was even
>arranged so that the normal running codes would not start a stopped
>train. Rather it had to receive a special code which would allow it
>to start, after which it would continue on the normal running codes.

Not so, either.

The Victoria has four track circuit codes:
420 - run at full speed, may accelerate, may start from rest
270 - run at up to 25mph, may accelerate, may start from rest
180 - coast at up to 25mph or brake, do not motor
120 - emergency stop
A stopped train will normally be receiving a 180 code until the signal
(or virtual signal) clears, when it will get a 270 or 420 depending on
the state of the line ahead.

Selection of the code is done by relays driven in turn by track circuit
occupation ahead. So the normal arrangement is 420 codes until the
braking point for a red signal, then 180 codes plus a spot command (see
below), causing the train to stop at that signal. There will be a 120
beyond the signal, but the train should never see it. When the signal
clears, the 180 changes to a 420. That track circuit can't transmit 270
at all.

If a train is drawing up behind another train, 270 codes are used
instead of 420s to restrict the train's speed.

There are separate "spot codes" used to tell the train to brake or
coast, or to give it a specific speed profile. But the primary decision
on where to stop is done by measuring the distance gone since the last
station stop.

Clive D. W. Feather

unread,
Jun 28, 2009, 3:50:16 PM6/28/09
to
>Selection of the code is done by relays driven in turn by track circuit
>occupation ahead. So the normal arrangement is 420 codes until the
>braking point for a red signal, then 180 codes plus a spot command (see
>below), causing the train to stop at that signal. There will be a 120
>beyond the signal, but the train should never see it.

Just to clarify, the layout is:

Spot code Signal
v |-O
>---+---420---+---420---+---180---+---120---+---120---+--TRAIN-->

|<-------------------->| Braking distance from line speed
|<------->| Braking distance from 25mph

Note that the *previous* signal won't clear until the train is at least
braking distance at 25mph beyond the signal *and* braking distance from
line speed beyond the start of the 180 code. Thus if a train either
fails to see the "brake now" spot code or stops braking once it reaches
25mph, it will be tripped and stop before hitting the next train.

>When the signal clears, the 180 changes to a 420. That track circuit
>can't transmit 270 at all.

That applies for a starting signal or one in the middle of an
inter-station run, not those approaching a station, where:

>If a train is drawing up behind another train, 270 codes are used
>instead of 420s to restrict the train's speed.

--

Stephen Sprunk

unread,
Jun 28, 2009, 7:53:22 PM6/28/09
to
David Hansen wrote:
> On 26 Jun 2009 22:35:14 GMT someone who may be Matthew Geier
> <mat...@no.sleeper.no.apana.no.org.no.au> wrote this:-
>
>> One would presume that if the train 'lost' the speed code, it would
>> stop, may not emergency stop, but would stop.
>
> I would guess that there is a code for stop, which is a normal
> condition, so that can be differentiated by the train from the
> abnormal condition of no code. The train would probably react
> differently to both situations.

I don't know the WMATA signaling system in particular, but it is general
US practice for "Stop" to be signaled by the absence of any code on the
rails. However, the systems are designed so that trains should never
reach the uncoded areas; the previous block would be coded for
"Approach", usually 10-20mph, which tells the train operator (or
computer) to stop before the end of the current block.

David Hansen

unread,
Jun 29, 2009, 3:31:00 AM6/29/09
to
On Sun, 28 Jun 2009 20:38:27 +0100 someone who may be "Clive D. W.
Feather" <cl...@davros.org> wrote this:-

>The Victoria has four track circuit codes:
> 420 - run at full speed, may accelerate, may start from rest
> 270 - run at up to 25mph, may accelerate, may start from rest
> 180 - coast at up to 25mph or brake, do not motor
> 120 - emergency stop

I have checked and according to G M Kirchenside and Alan Williams we
are both slightly wrong. However, they described the system as it
was in the mid 1970s, so things may have been changed since then.

They say that the train does not resolve the last code, though it is
put onto the rails, the train "sees" it as no code at all.

>A stopped train will normally be receiving a 180 code until the signal
>(or virtual signal) clears, when it will get a 270 or 420 depending on
>the state of the line ahead.

That say that only a 270 code will start a train between stations
and this is only put out as a preceding train pulls away.

They also say that in a station a train can start with a 420 code.
This is because it is not starting automatically but in response to
the start buttons being pressed.

Peter Campbell Smith

unread,
Jun 29, 2009, 4:33:04 AM6/29/09
to
"Clive D. W. Feather" <cl...@davros.org> wrote in
news:y9f1n9H9...@romana.davros.org:

>>Track circuit failures (false positives and false negatives) are not
>>uncommon in any railway, especially ones in the open air.

> Please provide cites.
>
> Right-side failures (reporting a train present when there's no train) is
> not uncommon, having a variety of causes.
>
> Wrong-side failures (failing to detect a train) are always a significant
> concern and I'm skeptical about your "not uncommon". The last
> significant incidence I'm aware of was the problem with Pacers on
> lightly-used track, requiring actuators to be fitted.

My main experience is on BART where I managed the rewrite of the non-vital
control software in the 1980s. Certainly right-side failures are much more
common, but clearly you have to allow for the possibility of wrong-side
failures.

There are also circumstances where the track circuit is correctly reporting
occupancy to the trackside equipment, but a failure of a comms component
prevents the indication reaching the central control system. Conflicting
indications to the train operator and the controller are potentially
unsafe. My observation over the years is that most collisions on automated
systems are where someone has overridden the automatic controls.

David Hansen

unread,
Jun 29, 2009, 5:07:27 AM6/29/09
to
On Sun, 28 Jun 2009 20:30:05 +0100 someone who may be "Clive D. W.
Feather" <cl...@davros.org> wrote this:-

>Wrong-side failures (failing to detect a train) are always a significant
>concern and I'm skeptical about your "not uncommon". The last
>significant incidence I'm aware of was the problem with Pacers on
>lightly-used track, requiring actuators to be fitted.

It also happened with 158s, on regularly used track. One of the
problems was that the suspension was too good, so it was de-tuned
slightly so that the wheels run on more of the rail head to keep
more of it clean. They were also fitted with track circuit
actuators. It wasn't that the track circuits failed to detect them,
it was that track circuits did not reliably detect them.

I have assumed that more modern DMUs are fitted with track circuit
actuators and I have a vague impression of having seen them on 170s.

http://thejunction-dmu.fotopic.net/c583138.html and
http://thejunction-dmu.fotopic.net/c550598.html have photographs of
the trains for those who don't know British trains.

Clive D. W. Feather

unread,
Jun 29, 2009, 6:55:47 AM6/29/09
to
In message <ihqg45lfdpmp910v1...@4ax.com>, David Hansen
<SENDdavi...@spidacom.co.uk> wrote:
>>The Victoria has four track circuit codes:
>> 420 - run at full speed, may accelerate, may start from rest
>> 270 - run at up to 25mph, may accelerate, may start from rest
>> 180 - coast at up to 25mph or brake, do not motor
>> 120 - emergency stop
>
>I have checked and according to G M Kirchenside and Alan Williams we
>are both slightly wrong. However, they described the system as it
>was in the mid 1970s, so things may have been changed since then.

I've dug out the actual instruction manual for the system.

>They say that the train does not resolve the last code, though it is
>put onto the rails, the train "sees" it as no code at all.

That is correct - there is no detector for the 120 code. If none of the
other three are received, the train goes into emergency stop. But that
is the meaning of the 120 code.

>>A stopped train will normally be receiving a 180 code until the signal
>>(or virtual signal) clears, when it will get a 270 or 420 depending on
>>the state of the line ahead.
>
>That say that only a 270 code will start a train between stations
>and this is only put out as a preceding train pulls away.

A 270 will normally appear as the preceding train pulls away, because
there isn't enough distance for a 420 to be safe.

>They also say that in a station a train can start with a 420 code.
>This is because it is not starting automatically but in response to
>the start buttons being pressed.

This is true in a station; the logic can distinguish between station and
signal stops in various ways.

I've been through the manual and it simply isn't clear whether a train
can restart between stations on a 420 code or only a 270. On plain line
only the latter is ever going to happen (absent a track circuit
failure), but at a facing junction not at a station it would, in
principle, be possible to need a 420. Having said that, I can't think of
any locations where this would happen.

Clive D. W. Feather

unread,
Jun 29, 2009, 7:03:42 AM6/29/09
to
In message <Xns9C39612...@news.albasani.net>, Peter Campbell
Smith <see.sig....@ct.details> wrote:
>>>Track circuit failures (false positives and false negatives) are not
>>>uncommon in any railway, especially ones in the open air.
[...]

>My main experience is on BART where I managed the rewrite of the non-vital
>control software in the 1980s. Certainly right-side failures are much more
>common, but clearly you have to allow for the possibility of wrong-side
>failures.

Certainly allow for them, I completely agree. I just disagree with the
"not uncommon" bit.

>There are also circumstances where the track circuit is correctly reporting
>occupancy to the trackside equipment, but a failure of a comms component
>prevents the indication reaching the central control system.

If so, that's incredibly bad design.

The opposite is understandable: TC reports "clear" but the message gets
back to the central system as "occupied". An SSI interlocking can do
that - if the datagram gets corrupted, the interlocking will assume all
the track circuits are occupied.

But the message should be "clear", not "train present".

>Conflicting
>indications to the train operator and the controller are potentially
>unsafe.

True.

>My observation over the years is that most collisions on automated
>systems are where someone has overridden the automatic controls.

That doesn't surprise me at all.

David Hansen

unread,
Jun 29, 2009, 7:11:11 AM6/29/09
to
On Mon, 29 Jun 2009 11:55:47 +0100 someone who may be "Clive D. W.

Feather" <cl...@davros.org> wrote this:-

>I've been through the manual and it simply isn't clear whether a train

>can restart between stations on a 420 code or only a 270.

Messers Kirchenside and Williams said that only a 270 code will
energise the traction motors by itself.

James Robinson

unread,
Jun 29, 2009, 7:37:25 AM6/29/09
to
"Clive D. W. Feather" <cl...@davros.org> wrote:

> Peter Campbell Smith wrote:
>
>> There are also circumstances where the track circuit is correctly
>> reporting occupancy to the trackside equipment, but a failure of a
>> comms component prevents the indication reaching the central control
>> system.
>
> If so, that's incredibly bad design.

The communication link is considered as non-vital, and not part of the
safety circuits. The only time this would be a problem is when stop
signals are being passed on verbal permission. In these cases, operating
procedures, such as only allowing movement at very slow speeds, are
supposed to ensure that safety is protected.

Clive D. W. Feather

unread,
Jun 29, 2009, 5:45:26 PM6/29/09
to
In message <j78h451uonala2hjp...@4ax.com>, David Hansen
<SENDdavi...@spidacom.co.uk> wrote:
>>I've been through the manual and it simply isn't clear whether a train
>>can restart between stations on a 420 code or only a 270.
>
>Messers Kirchenside and Williams said that only a 270 code will
>energise the traction motors by itself.

I'm sure they did. However, what I haven't found is a primary source
saying the same thing.

0 new messages