Why are U.S. Railroads resisting standardized signals?

[Paul_D_North_Jr]

Really, a mixture. 

Most of the NEC is elevated at least slightly on embankments, for drainage and a result of grade-crossing elimination when constructed or later projects in the cities and built-up areas. 

The only subterranean stretches are in the big cities - beyond Washington Union Station, Baltimore, Philadelphia, Trenton, New York City, and maybe a couple further north.  Most of that ranges from a few blocks to a mile at most, except for NYC. 

Elevated on bridges/ trestles is pretty much only at the bridge approaches across the various rivers. 

Ground level is not uncommon, but with the above most of that is in the open areas of NE Maryland and Delaware, across NJ, and Rhode ISland and Massachusetts.

- Paul North.

[MP173]

Is most of the NE Corridor sub terrainian, ground level, or elevated...or is it a mixture?

ed

[Paul_D_North_Jr]

blue streak 1

Railway Man

There are also very few grade crossings on the NEC, so that is much less of a problem.

RWM

This brings up a burning question that I haven't asked. How much savings on signal costs are available on ABS, CTC, ATC and PTC if a grade crossing is eliminated. Of course the number of other crossings in range of a crossing, # of tracks, any interlockings in range, etc. would affect the cost. As I remember all grade crossings (auto) have been eliminated NY - WASH sometime after AMTRAK. anyone know a date?  If the cost of grade crossings on PTC installations are as great as has been stated here could that be another reason for wholesale grade crossing eliminations before activating the PTC.

Wikipedia says 1976, under the heading of "Grade Crossings" a little before halfway down the page - but I think that's too early - I think early 1980s: 

http://en.wikipedia.org/wiki/Northeast_Corridor 

Supposedly it was Springfield Road new Bowie, MD - under "Amtrak 2002" about 90 % of the way down the page: 

http://www.trainweb.org/oldmainline/pa1.htm 

1981, per the Brotherhood of Locomotive Engineer's "Today's News" for Friday, Sept. 10, 1999, under the heading "Amtrak will upgrade safety devices in the Northeast", about 2/3 of the way through that item, which says "The last grade crossing south of New Haven was closed 18 years ago in Baltimore.", at:

 https://www.ble-t.org/pr/archive/headline0910a.html

That matches with my recollection - I spent most of 1983 at the nearby Aberdeen Proving Grounds (Maryland), and there was a fairly new bridge in downtown Aberdeen that allowed the closing a a major crossing there.

Hope this is helpful.

- Paul North.

 

[oltmannd]

aut1rml

I have been reading all posts and see logic in all,however I would like to remind those of how the Car makers resisted seat belts and such because of cost. How many lives has Ralph Nader saved over the years? I asume that air trafic and runway control thru out the world is standardized at least I hope.

Actually, it was the car buyers that resisted seat belts.  In the mid 1950's Ford focused on safety offering seat belts in their cars - advertising this heavily.  Chevy focused on performance, pushing their small block V-8.  Chevys sold.  Fords didn't.  Ford gave up trying to sell safety.

Even after the gov't mandated seat belts, starting with front lap belts only in 1964, hardly anybody would wear them. Usage was around 10% through most of the 1970s.  It wasn't until the usage laws came into effect that usage started to climb.

Ralph Nader, as far as I know, had little to do with the push for seat belts.  It was the Chevy Corvair where he made his mark.  (although his book didn't get published until after Chevy had already fixed the car's problems)

Interest in safety didn't really climb until the baby boomers started having kids.

[aut1rml]

I have been reading all posts and see logic in all,however I would like to remind those of how the Car makers resisted seat belts and such because of cost. How many lives has Ralph Nader saved over the years? I asume that air trafic and runway control thru out the world is standardized at least I hope.

[blue streak 1]

Railway Man

There are also very few grade crossings on the NEC, so that is much less of a problem.

RWM

This brings up a burning question that I haven't asked. How much savings on signal costs are available on ABS, CTC, ATC and PTC if a grade crossing is eliminated. Of course the number of other crossings in range of a crossing, # of tracks, any interlockings in range, etc. would affect the cost. As I remember all grade crossings (auto) have been eliminated NY - WASH sometime after AMTRAK. anyone know a date?  If the cost of grade crossings on PTC installations are as great as has been stated here could that be another reason for wholesale grade crossing eliminations before activating the PTC.

[R. T. POTEET]

tree68

How many years did it take for "right turn on red" to make it across the country?

It only took long enough for either Richard Nixon or Gerald Ford--I can't remember just which one was in office at that particular time--to affix his signature at the bottom of a piece of paper called a Bill.

When your Air Force transfered me to Leftover Air Force Patch in Massachusetts in 1964 I had to attend a base briefing. The moderator in this briefing cautioned all us new arrivals that if we were from the west where right turns on red were allowed they were prohibited in Massachusetts. "Massachusetts," he said, "does not allow right turns on red because, when all is said and done, Massachusetts uses rotary intersections to confuse drivers and Massachusetts drivers can only handle one confusion in a lifetime!". My downstairs neighbor who hailed from Rhode Island speculated that Mass drivers were too adle-brained to understand that they were going to have to stop and check for cross traffic before turning into the intersection

Congress mandated "right turns on red after coming to a full stop" in either 1974 or 1975. It was done as an energy conservation measure--I remember a figure bandied about that about 3% of all gasoline was wasted at red lights by drivers waiting to make a right turn--and carried a provision that states were required to post signs at locations where "right turns on red after coming to a full stop" were prohibited. Massachusetts promptly put up prohibition signs at every intersection which drew the flat of the federal government's blade; right turns on red could only be prohibited for compelling reasons of safety. When I read about this in a newspaper I could not help but remember that briefing of a decade earlier. I have not been back to Massachusetts since I left in '65 but I have often wondered just how Massachusetts came to grips with those compelling reasons of safety. I know they didn't but I wish the government had outlawed those rotary intersections at the same stroke!; I made many, many flat-wheel stops trying to navigate my way through those intersections. Guess I might be a little adle-brained myself.

In regards to signal standardization the only way this could occur is if it were mandated by federal law--and that, of course, can only be accomplished with federal specifications. What might these specifications be? and remember the old rejoinder about the horse designed by a committee. (For those uninformed readers this bromide states that a horse designed by a committee is likely to look like a camel.) Any study is likely to look at existing systems to draw conclusions of efficiency . . . . . . . . . . and once this committee has come to its conclusion and ordered compliance to a set of specifications who is going to pay for it. Some roads might even find the expense prohibitive to continued operation while others could find the cost of compliance relatively minimal.

If anything should be standardized it is the automobile industry. Did you ever pick up a rental car at an airport and it takes you ten minutes to figure out how the radio works and another ten minutes to figure out how to manipulate the heat and air conditioning and still another ten minutes to decipher the intricacies of the cruise control.

Standardization should be left to (possible) agreement within the industry. Everytime I hear the word standardization--particularly standardization through some kind of governement fiat--I think of that ditty from the 50s about the:

     Pretty boxes on the hillside;

     Pretty boxes made of ticky-tacky;

     Pretty boxes;

     Pretty boxes;

     Pretty boxes all the same.

 

[oltmannd]

Railway Man

Good analogy Don.

It's a warm, fuzzy solution for those who think that if you can't hear a relay ticking in the instrument house, it couldn't possibly work.

Sounds like you've met "the godfather of ACSES" somewhere along the line.....

[Railway Man]

Good analogy Don.

It's a warm, fuzzy solution for those who think that if you can't hear a relay ticking in the instrument house, it couldn't possibly work.

[oltmannd]

ACSES is kind of like trying to make a diesel locomotive from a steam locomotive by bolting on parts and beating it with a hammer.

It starts with PRR 4 aspect coded cab signal and suppression braking speed control (and later, Conrail's goofy quasi-predictive LSL speed control) and tries to make it do things it never was designed to do. It was a wonderful thing in 1950, but....

First, more signal aspects were needed to accomodate higher speed crossovers, so a second layer of cab signals was added.  But, that doesnt' easily handle curve and civil speed restrictions, to they layered on track trasnponders, which gives you two separate systems, each trying to independently say something about authorized speed..

To make matters worse, they display BOTH of these authorized speeds to the engineer, making the governing one bright and the other one dim.  Not exactly a good man-machine interface.

The only good news is that it's generally backwards compatible with the PRR cab signal/LSL system, so NS and CR can run their trains over the NEC, although I don't know how they'll accomodate a "positive stop".

Can you tell I'm not a fan?

[Railway Man]

blue streak 1

Railway Man

Moreover, ATC can't do a lot of things that PTC will do, except with expensive, complicated, ad hoc, inflexible, inefficient, and capacity-robbing workarounds:

1. It cannot consistently stop the train short of the signal (requires predictive braking).
2. It cannot enforce restricted speed
3. It cannot put two trains into the same block and and keep them from colliding
   
4. It cannot enforce permanent speed limits (curves, bridges, tunnels)
5. It cannot enforce temporary speed restrictions (Form As, General Orders, Track Bulletins)
6. It cannot enforce work zone limit
7. It does not work in dark territory
8. It cannot be installed without creating enormous problems with trade-crossing signaling because of frequency conflicts, which means several billion more to redo thousands of grade-crossing predictors.
   
9. It cannot be low-maintenance
   

I can make an ATC system do all of these things listed above, but no one will like the price tag or the harm it does to railway capacity

RWM: Could not find my copy of TRAINS that had the article on AMTRAK's ACSES signal system. What I remember it could do most of the above items except  #1. Your comments will be appreciated. Do you have any cost comparsions of it to ATC and PTC?  I will post later my concerns about using a GPS based system.

 

ACSES can do all of this 7, 8, and 9, but the cost is high.  On the other hand, the train-control environment and operating regime on the NEC is not the same as on other lines, thus costs are not strictly comparable.  There are also very few grade crossings on the NEC, so that is much less of a problem.

RWM

[blue streak 1]

Railway Man

Moreover, ATC can't do a lot of things that PTC will do, except with expensive, complicated, ad hoc, inflexible, inefficient, and capacity-robbing workarounds:

1. It cannot consistently stop the train short of the signal (requires predictive braking).
2. It cannot enforce restricted speed
3. It cannot put two trains into the same block and and keep them from colliding
   
4. It cannot enforce permanent speed limits (curves, bridges, tunnels)
5. It cannot enforce temporary speed restrictions (Form As, General Orders, Track Bulletins)
6. It cannot enforce work zone limit
7. It does not work in dark territory
8. It cannot be installed without creating enormous problems with trade-crossing signaling because of frequency conflicts, which means several billion more to redo thousands of grade-crossing predictors.
   
9. It cannot be low-maintenance
   

I can make an ATC system do all of these things listed above, but no one will like the price tag or the harm it does to railway capacity

RWM: Could not find my copy of TRAINS that had the article on AMTRAK's ACSES signal system. What I remember it could do most of the above items except  #1. Your comments will be appreciated. Do you have any cost comparsions of it to ATC and PTC?  I will post later my concerns about using a GPS based system.

[Paul_D_North_Jr]

Railway Man

blue streak 1

[snips]

What does the two above paragraphs mean? Any good lawyer is going to present this argument and a jury will buy it. The NTSB report is going to be very damming. This is the reason that restricting needs no ambugity.

 

Bluestreak -- I'm not a lawyer (and did not stay at a Holiday Inn Express last night) but the NTSB recommendations change nothing about the exposure to liability as far as I can see.  Two trains ran into each other.  They're not supposed to do that.  Passengers have a reasonable expectation when they get on a train that they'll arrive the other end sans collision.  Those facts are not doubted.

I think the question you and others are alluding to is that railways have become grossly neligent, that the use of restricted speed, restricting indications, and different signal aspect and indications among railways is so risky and so unsafe that it should never be used.  Anyone is welcome to make that argument in court, and probably will, but that doesn't mean they are inevitably destined to win.  The number of actual casualties where failure to adhere to restricted speed or misreading signal aspects had a role, is a really small number.  Calling that grossly negligent to me seems like saying that a hotel that permits you to step into a bathtub without having a properly secured slip-and-fall protection harness provided by and hooked up by the hotel is grossly negligent.  Level of risk matters.  There are far bigger risks than these.  

There's no natural law that defines a sharp boundary between acceptable risk and unacceptable risk.  Industries, technology, and social expectations evolve, and the boundary moves.  I don't see the recommendations of the NTSB in this case as being even remotely close to the line where the cost of eliminating the risk is equaled by the benefit to society received.  Risk reduction costs money.  My opinion is we should spend money to eliminate the big risks with proven high casualty costs, then as we have money left over, work down toward the little risks.  The number of people killed by train-motor vehicle collisions is more than 100 to 1 as great as the number of people killed in train-train collisions.  Is it smart to pour all our money and attention into removing the splinter from the little finger of a patient who has blood gushing from an severed artery?

RWM

[emphasis added - PDN.]

Well, there's about 20% of a semester of a good law school's course on Torts (which includes, but goes beyond, negligence theory) in what RWM wrote above.  Some supplemental comments and observations:

- NTSB findings and reports are totally inadmissible in any civil or criminal proceeding arising out of the accident.  The NTSB may provide a great "road map" for the claimant's lawyers to get to the same evidence with their own experts by using a sharply focused "discovery" process (subpoenas, depositions, requests for production of documents, examinations of equipment, requests for admissions, etc.) without having to totally "re-invent the wheel", but the lawyers will have to go through that formal process to be able to accumulate, analyze, and present that evidence and findings as their own at a trial or any mediation, arbitration, settlement negotiations, etc.;

- "Res ipsa loquitur" is the Latin phrase meaning "the thing speaks for itself".  That doctrine applies to both of RWM's points - trains aren't supposed to do that, and passengers have a reasonable expectation that they won't.  Hence, the claimant's don't have near as much a burden of proof of 2 of the 4 classical elements of a negligence action - "duty/ standard of care" and "breach of same".  Alternatively, the doctrine - particularly for a common carrier - may enable use of a "strict liability" theory, which pretty much dispenses with the duty and breach parts, and gets right down to the causation and damages (injuries) elements;

- We're probably making a liability mountain out of a safety molehill here (I meant to post this thought before, but probably couldn't due to one of the numerous recurring "Oops - something has gone wrong" error message problems here).  This accident had a remarkably small number of casualties, as I recall.  While something like 75 went to the hospital, I believe that only 6 were admitted - none were seriously injured, and no one was killed.  I don't know how many other similar accidents have occurred - either freight and/ or passenger - but it's not my understanding that this is a common problem or one that invariably causes severe risks - compare with the Chatsworth wreck.  While fixing this seems like a "no-brainer", in the grand scheme of things there are probably other problems that need fixing more urgently, as RWM points out - such as sleep deprivation / circadian rhythym-caused or related accidents, which are a recurring problem;

- The NTSB is "unencumbered by the economic process" [shamelessly adapted from Click & Clack's CarTalk "unencumbered by the thought process" put-down].  It's the NTSB's job to be the watchdog and identify anything that might need to be fixed - note that they usually have from several to many recommendations in each of their formal reports.  It's up to other agencies and organizations - such as the railroads themselves, the FRA, Congress, the courts, etc. to look into those recommendations further to see if they are reasonable = make sense from the standpoints of either cost-benefit, efficiency, safety of passsengers/ crews and other human lives, moral grounds, and/ or security, etc.  That's where the analysis, balancing and trade-offs of fixing problem A or B or C and the relative costs and benefits of same are worked out, argued, and decided, not by the NTSB.  That's not to say that the NTSB isn't right or justified in their conclusions and findings - it's just that in our system that's been set up, it's not the NTSB's role to mandate implementation of its recommendations;

- The railroads, in contradistinction, have direct exposure to the economic consequences of their actions or inactions.  Without getting into a dissertation here, a brief theory of strict liability is that by making the carriers essentially 100% responsible for any and all damages and injuries suffered by their passengers, it forces the railroads to "put their money where their mouth is".  Under the old-style negligence theory, the railroad would be liable only if the claimant could shown that the railroad been careless in some regard, and that the claimant was utterly blameless (for example, any "contributory negligence" - even 1 % - by the claimant was a complete - 100 % - bar to any recovery).  That can be a tough burden of proof, and if that couldn't be accomplished, the result was that the railroad paid and the claimant received nothing, even though the claimant had clearly been injured.  From an economic perspective, in such cases the railroad got off completely free - it had no monetary incentive to change anything or improve safety as long as it met the bare minimum "standard of care" as viewed by the often pro-business interests courts of that day - even though it might have been partially at fault, which I believe is technically referred to as "external dis-economies".  Over the years that wasn't working out too well - for a variety of legal and extra-legal reasons - so society and the courts adapted by creating/ applying the doctrine of strict liability.  The doctrine of strict liability has the effect of imposing and internalizing all of the costs of accidents to the railroad, so it suffered right along with and as much as its victims.  Thus -  in theory - any cost vs. benefits calculation of possible safety improvements should take into account the full reduction of costs resulting from lower damage and injury claims.  Therefore, the railroad - acting in its economic and moral self-interest - will naturally seek to implement those safety changes with the highest return on investment / costs - which tend to minimize the injuries and reduce damages.  Accordingly, what we should see is that the railroads fix the big problems first, and defer the minor ones until "later".  What their actions tell us in this instance is that they view the "Restricting" signal confusion problem as one of the minor problems that doesn't need an immediate fix.  Actual experience - such as this relatively rare accident - seems to support that view. 

Enough philosophizing and theorizing for now.  "Back to the salt mine with ye!"

- Paul North.

[Railway Man]

blue streak 1

Having been the aviation business for a long time and working on safety issues for 15 years I have the utmost respect for the NTSB. They have mostly been very astute in their recommendations. Human factors has been the hallmark of their decisions. Human factors? the accident at Tenneriff of Pan Am and KLM was a litany of human errors resulting in 500+ fatalities. EAL 401 into the swamp near Miami. More human factors with a few mechanical and warning displays (aspects) thrown in 180+ deaths. These were examples that I taught to my classes and the lesson got thru to some of them. The NTSB recommendations have definitely saved many lives in the aviation industry and the other transportation areas as well.

Now of all aspects displayed on RR signals I feel that restricted is the most precarious. Red - you stop. Yellow / flashing yellow  - proceed at some approach speed. Green - proceed at track or interlocking speed. Since either a lunar or flashing red means restricted in some rule books then that should be a standard. When you can look in the same rule book and under rule 286 (R over Y) you have medium approach and rule C-290 (R over Y) you have restricting then human factors are going to happen sooner or later. I've even observed NJ transit use flashing red at Hoboken terminal to signify restricted instead of a R over Y or a R over R over Y (they are on NOROC rules ?)

Now for the accident site in question. Was the siding a non signaled or signaled siding? If  non signaled then the lower Y could easily be changed to a lunar as CSX is doing on their unsignaled sidings. If signaled then an additional target will be needed to give a R over Y over Y to provide for a medium approach.

What does the two above paragraphs mean? Any good lawyer is going to present this argument and a jury will buy it. The NTSB report is going to be very damming. This is the reason that restricting needs no ambugity.

 

Bluestreak -- I'm not a lawyer (and did not stay at a Holiday Inn Express last night) but the NTSB recommendations change nothing about the exposure to liability as far as I can see.  Two trains ran into each other.  They're not supposed to do that.  Passengers have a reasonable expectation when they get on a train that they'll arrive the other end sans collision.  Those facts are not doubted.

I think the question you and others are alluding to is that railways have become grossly neligent, that the use of restricted speed, restricting indications, and different signal aspect and indications among railways is so risky and so unsafe that it should never be used.  Anyone is welcome to make that argument in court, and probably will, but that doesn't mean they are inevitably destined to win.  The number of actual casualties where failure to adhere to restricted speed or misreading signal aspects had a role, is a really small number.  Calling that grossly negligent to me seems like saying that a hotel that permits you to step into a bathtub without having a properly secured slip-and-fall protection harness provided by and hooked up by the hotel is grossly negligent.  Level of risk matters.  There are far bigger risks than these.  

There's no natural law that defines a sharp boundary between acceptable risk and unacceptable risk.  Industries, technology, and social expectations evolve, and the boundary moves.  I don't see the recommendations of the NTSB in this case as being even remotely close to the line where the cost of eliminating the risk is equaled by the benefit to society received.  Risk reduction costs money.  My opinion is we should spend money to eliminate the big risks with proven high casualty costs, then as we have money left over, work down toward the little risks.  The number of people killed by train-motor vehicle collisions is more than 100 to 1 as great as the number of people killed in train-train collisions.  Is it smart to pour all our money and attention into removing the splinter from the little finger of a patient who has blood gushing from an severed artery?

RWM

[CShaveRR]

The accident site in question was on NS, not CSX. And no siding was involved--the train was crossing from one main track to the other.

On the Railway Age news site this morning:

The National Transportation Safety Board makes the following recommendations to the Federal Railroad Administration: establish uniform signal aspects that railroads must use to authorize a train to enter an occupied block, and prohibit the use of these aspects for any other signal indication; study the different signal systems for trains, identify ways to communicate more uniformly the meaning of signal aspects across all railroad territories, and require the railroads to implement as many uniform signal meanings as possible

[blue streak 1]

Having been the aviation business for a long time and working on safety issues for 15 years I have the utmost respect for the NTSB. They have mostly been very astute in their recommendations. Human factors has been the hallmark of their decisions. Human factors? the accident at Tenneriff of Pan Am and KLM was a litany of human errors resulting in 500+ fatalities. EAL 401 into the swamp near Miami. More human factors with a few mechanical and warning displays (aspects) thrown in 180+ deaths. These were examples that I taught to my classes and the lesson got thru to some of them. The NTSB recommendations have definitely saved many lives in the aviation industry and the other transportation areas as well.

Now of all aspects displayed on RR signals I feel that restricted is the most precarious. Red - you stop. Yellow / flashing yellow  - proceed at some approach speed. Green - proceed at track or interlocking speed. Since either a lunar or flashing red means restricted in some rule books then that should be a standard. When you can look in the same rule book and under rule 286 (R over Y) you have medium approach and rule C-290 (R over Y) you have restricting then human factors are going to happen sooner or later. I've even observed NJ transit use flashing red at Hoboken terminal to signify restricted instead of a R over Y or a R over R over Y (they are on NOROC rules ?)

Now for the accident site in question. Was the siding a non signaled or signaled siding? If  non signaled then the lower Y could easily be changed to a lunar as CSX is doing on their unsignaled sidings. If signaled then an additional target will be needed to give a R over Y over Y to provide for a medium approach.

What does the two above paragraphs mean? Any good lawyer is going to present this argument and a jury will buy it. The NTSB report is going to be very damming. This is the reason that restricting needs no ambugity.

 

 

r 

[Paul_D_North_Jr]

ed -

Thanks for the 2nd opinion on the "balance" of the Langewiesche article.  You've summarized my "take-away" understanding of Sharkey's role in it - essentially an informed but uninvolved bystander.  Also, him having a "sour grapes" attitude is pretty plausible in the circumstances.

Yes, I've already printed the pertinent portions of this thread and squirreled it away to remind me to read the space shuttle and Somalian (?) pirates articles.  I appreciate the recommendation and also the estimated reading time for it.  I'm not a "Type A" personality maniac for scheduling life down to the minute, but with web pages of unstandardized lengths it can be hard for me to know quite what I'm getting in to when I start reading something as detailed and fact-specific as these articles.  While reading "The Devil at 37,000 Feet" I remember thinking it was going kind of slow, but I had quite a few interruptions that morning, too.

RWM - I concur with your "tangential sideshow" summary of Sharkey's role in Langewiesche's article.  As far as Sharkey's "street cred"[ibility], it then appears to be - logically enough - more as a "trade press" writer than a researcher or investigative reporter/ journalist (if I'm using any of those "terms of art" correctly, because that's another world that I'm not terribly familiar with), who doesn't have to worry about "biting the hand that feeds him" / conflict of interest kind of thing.  You know more of the background there than was apparent from Sharkey's protest, and that helps to assess and place it into proper perspective.  This morning I was going to write, "Methinks he doth protesteth too much" - well, it seems I should have ! 

- Paul North.

[Railway Man]

MP173

Paul:

I agree with your assessment...didnt quite catch the "pile on Joe Sharkey" portion of the article.  Perhaps it wasnt there.  Sharkey was mentioned a couple of times in the article, but not in a negetive manner.  Perhaps he is a bit upset that an article was written by another writer, instead of the one who should have naturally gotten the byline.

It was very well written and seemingly very fair to everyone.

Paul, if you have a chance, invest an hour and read the one in Atlantic (archived) on the space shuttle crash.  Very well written and documented.  His understanding of aviation gives him a great advantage, plus the ability to recount the actual events with well researched and documented information.

ed

 

Someone who came to the article from the "Ask the Pilot" (Patrick Smith) reference, like I did, knew about this conflict between Langewiesche and Sharkey.  Smith though that Langewiesch was too hard on Sharkey, but that it was a minor, tangential sideshow in the article.  

Sharkey is not the person that could have written that article.  He's a business-travel reporter, not an industry or technical reporter, that's compromised himself with too many boondoggles, too many encomiums, too many paens to the industry.  His market is the high-end, first-class, and private-jet passenger, and his source is the marketers for the airlines and the airframes.  He is compromised from unabashedly sitting down and writing, "hey, guess what, some of this 'magic' doesn't work."  That is not a story his market or his sources want to read.

RWM

[MP173]

Paul:

I agree with your assessment...didnt quite catch the "pile on Joe Sharkey" portion of the article.  Perhaps it wasnt there.  Sharkey was mentioned a couple of times in the article, but not in a negetive manner.  Perhaps he is a bit upset that an article was written by another writer, instead of the one who should have naturally gotten the byline.

It was very well written and seemingly very fair to everyone.

Paul, if you have a chance, invest an hour and read the one in Atlantic (archived) on the space shuttle crash.  Very well written and documented.  His understanding of aviation gives him a great advantage, plus the ability to recount the actual events with well researched and documented information.

ed

[Paul_D_North_Jr]

Railway Man

[snips] Langweische is my favorite author, by far.  No one else gets the technical details as right, explains them as clearly, or turns a phrase so eloquently.  If the last paragraph in the "Devil at 37,000 feet" article doesn't make you shiver, you might be dead from the neck up.  Patrick Smith, the "Ask the Pilot" columnist and a 767 captain, said that Langweische made exactly one technical mistake in the "Devil" article, a very minor one, misnaming a control center, which might have been a spellcheck error in fact.  That's as impressive a recommendation as you can get. 

RWM

Note:   For those who've been following the "Devil at 37,000 Feet" aspect of this thread, you may be interested in the following blog dated March 11, 2009 that I ran across earlier this morning.  It appears to be by Joe Sharkey - the reporter/ writer for Business Jet Traveler magazine who was also aboard the Embraer Legacy private aircraft:

http://joesharkeyat.blogspot.com/2009/03/enough-already-vanity-fair.html 

As you'll see, Sharkey's neither happy with Langewiesche nor very impressed by the article.  It feels to me like Sharkey's straining pretty hard to find a basis to be able to righteously proclaim that he's been insulted by Langeswiesche.  But I'd have to reread both the Vanity Fair article and Sharkey's blog post to make my tentative conclusion more definitive, and I don't have the time for that right now.

- Paul North.

[MP173]

Topic divergingly, slightly...

I have read three of Wm Langewiesche's articles and highly recommend them.  The before referenced article on the collision at 37,000 feet was excellent.

The other two were accounts of Piracy (very appropriate based on the recent activities) and last night was his look at the space shuttle's 2003 mishap.  The look at NASA was extremely descriptive and informative. 

Langewiesche's articles are archived on both Vanity Fair and Atlantic.  Fair warning...these articles are well written and very long, each taking an hour or so.  He sets up each story quite well and leaves you very well informed.

ed

[erikem]

Railway Man

The cost difference I'm seeing for making ATC or some sort of intermittant cab-signal system work vs the Wabtec platform is 3:1.

 

Sounds reasonable and thanks for the data point. My point was that a lot of what PTC does could be done with enhanced ATC, but that's emphatically not the same as saying that an enhanced ATC should be done instead of PTC.

On a similar note, there's a lot of interest in using wireless sensor networks on the factory floor - one tidbit was that the cost of providing a wired connection was about $40/foot - probably something similar for a railroad. 

[jchnhtfd]

Thank you, RWM!  Really what I was looking for, and actually a two part answer.  First part -- people are looking at the problem, and realistically (not to tread heavily on toes, but have you ever tried to convince the folks at EADS that the software in their fly-by-wire does NOT have a fail operational mode for some stupid mechanical failure?  Like wonky integrated starter/generators on GE CF6 engines?  No?  Don't, if you value your sanity) and second that, in general, a failure is a fail safe (fail stop).  Expensive, but at least you can park a train...

And it makes me very happy that the Wabtec system is on-board.  That, I like.

[Railway Man]

erikem

RWM,

Are the 12 frequencies you've mentioned for ATC use or for audio frequency overlay?

I'd be pretty sure that a continuously variable ATC speed signal is possible (or something that has enough steps to be close enough). The tricky part would be ensuring that the ATC receiver is picking up something that can only be an ATC speed control signal - mistaking a stray signal for an ATC speed limit signal is a great recipe for disaster.

 I shouldn't be surprised that audio frequency overlays can have problems in estimating distance - there are some parallels with what I'm doing at work. Another method that could work is using a magnetic sensor to detect the presence of a train within a few feet of the sensor (problem here is either doing the wiring or replacing batteries for a wireless sensor).

Anyway, it is probably cheaper to go straight to PTC than to implement a souped up ATC.

- Erik 

 

About a dozen each for both the ATC and the grade crossing predictors.  That is all the manufacturers are willing to certify as safe.

The active track transponder you reference in your second paragraph is the usual method for knowing distance to a target.  Those things are quite expensive.  The ACSES system in the NEC in fact has gone that route. 

The cost difference I'm seeing for making ATC or some sort of intermittant cab-signal system work vs the Wabtec platform is 3:1.

 RWM

[Railway Man]

jchnhtfd

My previous post on this regarding my one concern with PTC (which hasn't been answered, somewhat to my surprise) having vanished for some reason, I'll briefly restate it:  What happens when something in the system breaks?  Having considerable experience with aircraft accident/incident investigation, I can assure folks that the magic does, in fact break.  Not referring to bad programming (endemic) or operator error (also endemic) but just plain simply breaks.

In the present signalling system, that usually means a dark signal (stop and find out what's going on) (yes, I know, there are ways in which a signalling failure can result in a false clear, but they are pretty rare).

In general, seems to me that some form of PTC is indeed a very good thing and should, in the long run (after amortizing the money) result in better -- perhaps much better -- productivity and lower costs, and I'm delighted with the progress being made.

But I'd still like to know... what happens when something breaks?

 

Ah - sorry, I'd interpreted the earlier post as a rhetorical question rather than a question question.  

Oh dear, failure modes.  There is a long list that lead to the system refusing to do anything, which means the train stops right there, and the outcome is the railway is tied up, but trains shouldn't run into each other.  But I think what you're looking for is failure modes that lead to an authority violation or an overspeed or open switch being run through.  Most of those failure modes have at their root someone failing to update a change in system input data: a new turnout is installed, but the database isn't updated, or a signal is moved, and the database isn't updated. 

The Wabtec platform puts all the processing power and control on-board the locomotive.  There's no requirement for handshakes with the central dispatching server except when an authority is extended or created.  Failure modes there consist of information drops which leads to no valid authority being issued.  The onboard computer uses tandem processors -- both have to agree for something to happen. 

Unlike aircraft there is very much less input into the system.  The inputs consist of a track database (uploaded and stored in memory), an authority (issued by the PTC server), a location (deduced from GPS, track tag, and axle count inputs), and peer-to-peer querying of wayside interface units (on turnouts, fixed signals).  A failure of a WIU is the same as what you get now when a switch-circuit controller doesn't work or a signal gives a false clear -- the WIU consists of two wires to a radio, and if it tests the first time it doesn't change.  But a switch-circuit controller can be out of adjustment (just like now) and call a switch closed when in fact it's open.  The architecture is simple, on purpose.

A common failure mode is the locomotive not mapping itself to its track database.  The database says, for example, "I have according to the logical network and my axle odometer proceeded 5,000 feet timetable west on track #100002 since my last good GPS fix which means my GPS waypoint should be XXX/YYY.  But my GPS antenna says I am at at XX1/YY1.  That's not possible, either the database is wrong or my GPS is wrong, so I quit, and I'm stopping here."

More than this, I'd have to start emailing you pdfs.  It's hard to explain a negative, and that's what your question consists of: "How do I know this is reasonably perfect."  The answer is, you don't.  But there is a lot of test data out there from the Beardstown Sub on BNSF and some other test platforms, and if you read 49.236 Subpart H, you'll know a lot more about how demanding the proof has been.

RWM

[jchnhtfd]

My previous post on this regarding my one concern with PTC (which hasn't been answered, somewhat to my surprise) having vanished for some reason, I'll briefly restate it:  What happens when something in the system breaks?  Having considerable experience with aircraft accident/incident investigation, I can assure folks that the magic does, in fact break.  Not referring to bad programming (endemic) or operator error (also endemic) but just plain simply breaks.

In the present signalling system, that usually means a dark signal (stop and find out what's going on) (yes, I know, there are ways in which a signalling failure can result in a false clear, but they are pretty rare).

In general, seems to me that some form of PTC is indeed a very good thing and should, in the long run (after amortizing the money) result in better -- perhaps much better -- productivity and lower costs, and I'm delighted with the progress being made.

But I'd still like to know... what happens when something breaks?

[erikem]

RWM,

Are the 12 frequencies you've mentioned for ATC use or for audio frequency overlay?

I'd be pretty sure that a continuously variable ATC speed signal is possible (or something that has enough steps to be close enough). The tricky part would be ensuring that the ATC receiver is picking up something that can only be an ATC speed control signal - mistaking a stray signal for an ATC speed limit signal is a great recipe for disaster.

 I shouldn't be surprised that audio frequency overlays can have problems in estimating distance - there are some parallels with what I'm doing at work. Another method that could work is using a magnetic sensor to detect the presence of a train within a few feet of the sensor (problem here is either doing the wiring or replacing batteries for a wireless sensor).

Anyway, it is probably cheaper to go straight to PTC than to implement a souped up ATC.

- Erik 

[jeffhergert]

Thanks RWM.  I was thinking you meant the changing values part, but wasn't sure.  Even a collision when a train is running under restricted speed conditions can be deadly.

Maybe we need another thread on PTC, but I have a couple of questions.

When approaching a Form B (work zone GCOR), how will a train be cleared up?  Will they contact the foreman verbally, who will then clear them verbally and enter into the system that the train can proceed at whatever speed he prescribes thru the limits?  Will the process be completely digital with the engine sending, either manually or automatically, a request for instructions and the foreman responding without actually talking to the train?

The other question is, will on-track equipment and hi-rails that may not shunt the track circuit (where there is signalled track) be equipped with GPS so that the system knows when they are out of their track authority?  Most of the time people think that it is only when a train overruns a Form B that accidents can happen, but we have had instances where MOW forces left their authority.  One extreme case was a track inspector getting a Track and Time on one track in two MT territory, than setting on the wrong track! 

Jeff    

[Railway Man]

erikem

If I'm getting this right, the ATC equivalent of PTC in your siding example would be dividing the siding up into short blocks with decreasing restrictive speed enforcement (e.g. 20, 15, 10, 5, 2.5 ...) - difference being is that the block length would adjust to expected braking performance of the train.

I'd think it would be possible to implement the equivalent function in a modified ATC system - use and audio frequency overlay to judge the distance of the train from the end of the siding, and have the ATC signal give either a continuously variable speed limit or with small speed steps. This would require  replacing all of the ATC equipment on the locomotives and at least some trackside equipment - at which point PTC makes more economic sense.

A note about technology. When continuous ATC systems were first implemented, the technology for resolving the received frequency required a relatively large spacing to reliably distinguish frequencies. It is now ridiculously easy to reliably measure frequency to a fraction of a Hz (cycle per second).
 

The solution you describe is a lot more frequencies than are available.  I think there would be great difficulty in getting it to work with "five nines" reliability without losing a lot of train speed and track capacity.  I am not educated enough to explain why in a way you'll find satisfactory.  The track circuits are imperfect due to varying degrees of resistance every time the weather changes.  As I recall from the last installation I worked on (a couple of months ago), we had only 12 frequencies to work with. That's with the latest, greatest ATC equipment.  It still is having issues with grade-crossing predictors that don't predict due to frequency conflicts and problems with all sorts of real-world misery such as rusty rail when trains don't run for a day, squishy track subgrade that is wet one day and dry the next, and so forth, and a lot of air-brake issues. 

RWM