PTC is a tough one.  I certainly understand the railroads’ distaste for such a huge unfunded mandate, and one that must be implemented in so short a timeframe.   I’ll need to get smarter on the alternative approaches offered by the AAR and other industry stakeholders.  Thinking strategically about how rail technology must evolve in the 21st Century, I believe PTC is a matter of “when” and not “if.”  I cannot envision large North American railroads operating without PTC (or other technology that achieves the same objectives) on high-traffic lines in 2050.  I’m not sure whether or not PTC lends itself to more gradual implementation, or if, once begun in earnest, it must be carried out system-wide in short order.  (I’m thinking about cab signaling, which is used on only part of the CSXT system; it does present added resource constraints, such as managing appropriately equipped locomotives.)  Can PTC be rolled out in phases?  If so, perhaps the implementation mandate can allow for a much longer period in which to achieve (and pay for) compliance.  If so, the railroads could submit a 20- or 25-year plan to implement PTC, and the FRA could hold the railroads accountable for timely execution of these plans.

Mr. Smith - virtually all of the PTC writings I've seen since Chatsworth have emphasized high cost and short schedule. As both a former signal company employee and a former RR ops manager, I wonder why there hasn't been much said about what I view as the two biggest practicalities of PTC - how it will impact freight lines' capacities, and how to implement it given the almost infinite number of different freight train 'handling' characteristics there are.

Regarding lines' capacities, as I understand it, the intent of PTC is to virtually eliminate the possibility that any train could get close enough to a danger point to hit another train. Let's assume for a moment that four major operating characteristics of acceleration rate, top speed, deceleration rate and weight of all the trains using a given line were EQUAL, which in the real world they couldn't be (the only way this could come close to happening is if a brand new passenger line using equal train consists all the time was the culprit). In such an 'equal characteristics' case, MORE DISTANCE BETWEEN TRAINS THAN IS TODAY'S STANDARD WOULD BE REQUIRED in order to achieve the PTC goal of eliminating train collisions, just for following trains. Here's how several of the USA's older mechanical-train-stop-equipped transit properties, like NYCT, try to eliminate the chance that a following train could rear-end the one ahead of it  - they add an extra signal 'block' between trains, WHICH DECREASES LINE CAPACITY OR THROUGHPUT. Instead of just having a red signal immediately behind a train preceded by a yellow signal and then a green one, (basic red-yellow-green scheme that most freights use on their signaled lines), NYCT inserts en extra block between trains in most places. Thus NYCT trains almost always have TWO red blocks, then a yellow & then a green, behind them (red-red-yellow-green scheme). Mechanical train stops are located adjacent to most NYCT signals at block entrances, and when a signal is red its corresponding train stop is in the up, or tripping, position and will 'trip' a train that proceeds past the red signal, causing that train's brakes to apply at the maximum 'emergency' rate (unless the train stops first and 'keys' by, which is possible at many signals). Even with NYCT's basic train-stop-equipped method there still can be, and have been (Williamsburg Bridge accident in the 1990's comes to mind) rear-end collisions if when tripped the following train is going to fast or its brakes don't operate as they're supposed to.

Now let's assume that all freight trains using a given line have equal operating characteristics (which of course THEY DON'T as I imply above). Frequently such lines are already 'signaled' for maximum throughput. If a second 'red' block is added behind every train to achieve PTC's prime objective, THE MAXIMUM CAPACITY OF THE GIVEN LINE WILL GO DOWN, which as the freights will tell you is contrary to the objective of all the capacity improvements they have made in recent years, and are continuing to make as of this writing.

Regarding implementation of PTC on freight roads given the huge number of different operating characteristics of each freight train (train length/weight/braking capability, weather, temperature, grade, time of day, track curvature, engineer train handling knowledge and fatigue level), WHERE SHOULD YOU PUT EACH ADDITIONAL SIGNAL, EITHER PHYSICAL OR VIRTUAL, TO INSURE A FOLLOWING TRAIN WILL NEVER HIT THE ONE AHEAD OF IT?

Just waving a wand and saying "implement PTC" seems to me to be the same sort of statement as "ONE MAN TRAIN CREWS are all we need" (the people who make this statement have obviously never been on the crew when their train's air is pissed away on a long downgrade in the winter, but I digress).

I'm not saying PTC can't be deployed - but I can't see how given the two points discussed above and the current level of train control technology.

I'm sorry to be so long-winded - others feel free to comment  (however, nit-picking what I've written isn't necessary).

bradyrs raises some interesting points (no nit-picking here).  I suspect that one reason the freight railroads do not discuss increased capacity potential of PTC implementation is that the Congress ordered its installation and deployment strictly on safety ground - and did nothing to contribute to paying for it.  Based strictly on safety ground, PTC has a dreadfully poor benefit-cost ratio, according to the FRA's studies.  Happily, the annual number of collisions between freight trains, either as a result of an engineer blowing through a signal or a mis-aligned switch is really quite small and that lowers the benefit-cost ratio.

I claim no expertise on PTC, although having written about it on numerous occasions, I have discussed it with some of the leading experts in the field.  There should be no need to insert a second block, nor lose the existing capacity, as bradyrs postulates.  As I understand the technology, the system requires that the location of the lead locomotive of a train be known to within about six feet.  The pertinent data is programmed into the dispatch computer system and the system has the ability to apply the brakes if the train exceeds its authority either inadvertently by a less-than-observant engineer or intentionally by a suicidal one.  This applies to speed and protection of MOW personnel, movement through a switch that may be misaligned.  GPS technology allows for the precise knowledge of unit location, although there also are PTC advocates of wayside sensors.

Larry - thanks for not nit-picking (just noticed a couple of spelling errors myself!). You're correct that one of the PTC implementation strategies as I understand it is to remotely cause a 'PTC-controlled' train's brakes to apply if it travels past its authority limit. There's no problem, using GPS, with knowing where each end of such a train is, or its weight, length and makeup (loads vs. empties, hazmat car types & locations in the train, etc.). But in order to establish/calculate the train's authority limit, or boundary, stopping distance must be calculated as well. The problem is that stopping distance is very changeable because of the train's operating characteristics that I discussed in my previous post (forgot to mention train speed - old age is telling, isn't it?). The other factor that I didn't discuss previously is the minimal brake adjustment that is done on most freight cars.

I'm old enough to remember the self-adjusters on automobile drum brakes that when the auto was backed up, tightened its brakes as the shoes wore. Most freight cars' brake rigging only can be manually adjusted to maintain constant shoe force against the wheel as shoes wear. However, NO ROAD MANUALLY ADJUSTS CAR BRAKES to my knowledge. One has to crawl under the car, remove 1"-diameter cotter-key-contained pin(s) and then put the pin(s) in different hole(S) in the linkage, all while prying on parts of the linkage and holding up other pretty heavy pieces of steel included in the linkage. This process is time-consuming, physically demanding and significantly more hazardous to 'car knockers' because of the linkage's location under the cars.

Instead, all that's typically done is brake shoe replacement. Most railcars' break rigging linkage (rods & levers between a car's brake cylinder piston and its brake shoes) is left in a position so that when a car's brake shoes wear down and need replacing, the new shoes will just fit WITHOUT HAVING TO ADJUST THE BRAKE RIGGING. Shoe replacement is performed relatively easily from the side of the car - seldom must one crawl underneath, and a new shoe might weigh 10 pounds. There is a 'key' that holds each shoe in place. By pulling this key (8" long or so) up after the shoe holder is pried away from the wheel, the worn shoe is removed and a new one put in the 'holder'. Then the key is replaced and the car goes on its way (after all below-limits worn shoes are replaced, of course).

There are a few cars that have mechanical 'automatic' adjusters in their brake rigging, but such adjustment mechanisms are complex, need periodic maintenance and inspection and add cost to a railcar. Also, some cars have load sensors to take varying car weight into account and somehow reduce braking force when these cars are not heavily loaded. However, these sensors are relatively unsophisticated, as they are attached to truck springs and typically have only 2 positions - 'loaded' and 'empty' - based on how far the car's truck springs are compressed. As many RR folks know, a 'loaded' car's weight can vary significantly by type of lading - this is especially true with general-purpose godolas, flat cars and smaller boxcars. An empty gon might weigh 30 tons, with its fully-loaded gross weight  for compact heavy steel scrap being 130 tons, but its fully-loaded gross weight containing pipe being only 90 tons. Unfortunately the load sensors can't tell the difference between the 90-ton and the 130-ton weight, so the brakes may apply more tightly than needed in many cases (such continued heavier-than-needed brake applications cause flat spots, of course).

This wide variation in brake system maintenance & condition must be taken into account in calculating stopping distance for trains. The problem is, today's scheduled 100-car coal train #487 is made up of 50 hoppers with new brakes and 50 with worn brakes - tomorrow's #487 has worn brakes on all its 100 hoppers. The upshot of all this variation is that braking distance calculations AND AUTHORITY LIMITS will be hard to accurately calculate. So, safety margins will have to be added. Thus distances between trains will have to be greater, won't they (and capacity will decrease)? For instance, each line's signal locations (block layout) were originally determined based on common train types, weights & performance factors, maximum authorized train speed, terrain, etc. Theoretically when higher speeds are permitted (HSR on freight lines, anyone?), distance between trains must be increased to account for the longer stopping distances required by these speeds.

So - how to implement PTC without significantly impacting capacity? I'm afraid I don't know........

As with my previous posting, comments/clarifications are welcome!

I don't question a thing you've said here, bradyrs, mostly because I'm not competent to do so.  I do know that PTC has been tested as long ago as the mid-1980s on at least one Class 1, and the tests were considered successful and positive.  So positive, in fact, that after more years of R&D, one or more Class 1s were in the process of deciding to go ahead and deploy PTC for the capacity benefits they saw.  Ironically, this was about the same time as Chatfield.  Knowing what they had been told by the railroads, mandating PTC for safety became a really cheap vote for Congress, the kind it likes because it thought it was mandating what the rails were going to do anyway and at their own expense.  I'm quite certain the unanimous condemnation of an "unfunded federal mandate" has caught the Congress by surprise.  Will Congress do anything about it?  When does Congress ever behave rationally and pass up the opportunity to pander and posture?

More specific to the discussion, as I said, I am not competent when it comes to understanding the fine points of brake operation and adjustment.  Accepting what you say, I don't see how brake adjustments, etc., affect the working of PTC, which activates under just one circumstance basically, the locomotive violates its track authority and the dispatch computer actifvates the brake system.  If I've missed something significant, I imagine you and/or others will educate me.

Movement authority is calculated based on train performance, terrain, track layout, etc.  What Bradyrs is saying is that if you don't know the condition of the brakes, you don't know the performance of the train and you have to calculate based on a worst case condition.  Whether this results in more or less capacity than a fixed block system, I don't know.

I wonder if NS CEO is simply trying to throw "2010 is going to be bad" verbiage out there so that they don't get hammered by Wall Street this year.  Ironic now that Wall St judges us on financial stability.....

I think the concept of PTC is that it will allow you to add more density to the lines because trains can get closer but yet be safer.  I'm not agreeing or disagreeing.

The cost benefit doesn't exist though so I'm along the lines of Larry on this.  Chatsworth was the rail event of 2008.  When was the last doozy?  The chemical one for NS a few years before that?

Just more regulation to make everyone that knows nothing about the industry to feel safer at night.

Anyone remember the UP cars in Southern Cal that got away from a crew and ran for 20-30 miles after the air was bottled?  That was a rules violation - human factor.  

NS incident with chemicals - Rules Violation - human factor.

Chatsworth - Rules Violation - human factor.

So unless we go to Zero Man crews there will always be a chance for accidents to occur.  PTC is the warm fuzzy blanket that FRA/NTSB, etc are giving the constituents to fall asleep at night a little "safer" than last night.

I do like the PTC concept though.

Well said, BacktotheFuture.  Cynical, but well said.

It is obvious that we all know less than all there is to know about the functioning of signal systems.  However, a better study for what PTC might do would be the old continuous cab signal system with speed control that the Pennsylvania Railroad had installed on its lines east of Pittsburgh.  (Not all lines were equipped, and not all locomotives had all of the speed control features installed, either.)  Later the system was improved by the Locomotive Speed Limiter for freight train locomotives operating into Amtrak territory.

Basically, these systems used fixed blocks on the main track with a signal generator at the leaving end of the block (where there was usually a track-side fixed signal).  When the train entered the block, an inductor below the leading end of the locomotive would detect the signal.  Lack of a coded signal through the track resulted in a "restricting" aspect showing in the cab.  The next better code produced an "approach" aspect.  One better was an "approach-medium" and finally the best was "clear" in the cab.  With the speed limiting feature working, if a train was exceeding 20 mph without its brakes already applied when entering a block on a restricting aspect, the system would give a few seconds for the crew to apply the brakes and then apply the brakes itself.   Speeds above approximately 30 mph for "approach" and 45 mph for "approach-medium" would produce similar results.  Conrail's Locomotive Speed Limiter would respond similarly, except that the system measured decelleration rather than the having the brakes applied with a "full-service" brake pipe reduction and that the locomotives were restricted to 50 mph on a clear aspect.  (I am not sure how exact this limit was nor if the throttle were closed, if a train drifting down hill might exceed the limit a bit more without penalty.)

The shortcomings of these systems was that the signal system only allowed for the four speeds (Amtrak enhancements allowed for additional speeds, though none were slower than 50 mph, so they only applied to passenger trains.) and they did not cover restrictions such as curves and track conditions.  

Also, the speed controls needed to be set-up for the worst case.  This would be the effective shortest signal spacing, actual distance adjusted for grade, on the Northeast Corridor.  It would also cover the freight train with the most weight per operative brake that could pass a brake inspection.

The system did improve through-put a bit.  It allowed trains to respond to the signal at the end of the block before it was visible.  That is, if the train entered a block on an "approach" signal (diagonal from the right at the top to the left at the bottom with the old PRR position light signals, the same as a yellow on most railroads) when the train entered the block, but the next signal ahead cleared from stop to approach, the cab signal would allow the train to avoid the braking and proceed at full speed after acknowledging the initial signal and also the improved aspect.

Finally, there were a few things that the system could not prevent.  First, because a train might move back and forth between main tracks equipped with cab signals and main tracks that were not equipped, a suicidal or impaired engineman could turn the system off (or never turn it on) and after receiving a restricting cab signal, then accelerate again into an accident.  Actually, a train can do a lot of damage at 20 mph without accellerating (think Chase, Maryland).  Or, like happened to the Federal Limited shortly before the Eisenhower inauguration, the brakes could simply fail.  (For those of us who have read about it, the angle *** on the brake line a couple of cars behind the locomotive was closed on the train so it could not stop when going slightly downhill into Washington Union Station.  Best guess was that it vibrated closed somewhere after leaving Baltimore, but there were conspiracy theories even then.)  PTC should help prevent willful acts of the engineman, but I would doubt that it would have solved the angle *** problem.

The biggest problems with PTC is that the system has been mandated before we even know if we can get it to work and it reverses the old adage.  It is a pound of prevention to effect an ounce of cure.

One problem presented by some commenting here is the view that PTC is another signal system.  It is not.  It is a system to prevent a train from exceeding its track authority, and it relies on the dispatch computer to send the proper signals when the train is detected exceeding its authority.  As for the claim that PTC has been mandated without having been adequately test, that simply is not so.  Interoperability is a legitimate issue, but the BN at least has been testing it back to 1984.  How long does it have to tested.  By the standards some of you would apply, we still would not have adopted the Westinghouse Air Brake.  As for trains moving back and forth between main tracks that might not be equipped with the appropriate signals, this is not a signal system.  The train that moves back and forth has track authority to do so and the dispatch computer has that authority nice and neatly on its hard drive or wherever.  Open minds, please.  A closed mind cannot learn anything new.

CAD software that dispatchers use allow for the movement of trains.  Dispatcher gives a train a lineup (tells the signal system to give proceed indications or some sort, followed by restricting signals dependant on traffic, etc.)

If dispatcher lines up a train to go eastward from point A to B and adds no further instruction to CAD system, if operating correctly the train BEFORE hitting point B will get signals resticting its movement past point B and at point B there should be a red signal or an aspect that is the most restrictive.

Should that train go past point B - GPS or other technology would take steps to shut down the locomotive or send train into emergency brake application, etc.  So the train might not necessarily stop on a dime at point B, but none the less would be stopped and the dispatcher would know that PTC was implemented and shut the train down, now the dispatcher can tell other trains in the area of the situation.  Such as a train going westward approaching point B.

So in effect using Chatsworth, once the train got past the red signal, the train would have gone into emergency.  The likelhood that the dispatcher could have contacted the east/southbound freight train to alert them of the situation the outcome may have been avoided.

I appreciate the commentary in this thread…some excellent info and food for thought on PTC.  My apologies to Jeff Stagl for hijacking and diverting his original blog topic.

Well when a recent railway age mag. feb 09 survey showed that 55 % of railroad custumers did not support railroad regulation...But that also says that 45% of railroad custumers are not happy with there service?

Some interesting commentary about PTC here.

First, to BacktotheFuture's comments about Chatsworth. I'll avoid nitpicking on technical detail, but your description of PTC is a bit offbase. Stopping a train AFTER it passes a red signal is not an improvement over technology that's been in use for the past 75 years, and would likely not have prevented the loss of life at Chatsworth.

The NTSB's investigation shows that the Metrolink train ran through the switch a mere 6 seconds after passing the absolute stop signal, and collided head on with the UP freight 21 seconds later. What this tells me is that any system triggered at the stop signal would not have been capable of preventing this train from penetrating into the route of the freight train. Even if the dispatcher had grasped instantly what was happening, an aggregate of 27 seconds is woefully insufficient to convey any warning to the freight train and for them to take any meaningful action to stop.

The key element that PTC is intended to provide is NOT to take action when the train passes a stop signal or passes a signal requiring reduced speed, but to calculate a point, given the train's speed and braking performance, where the system has to intervene to stop or slow the train BEFORE the violation occurs.

The trick is that with highly heterogeneous train consists and without a way of PRECISELY measuring the weight and braking capacity of each car in a train, calculating a meaningfully accurate braking curve is difficult. Taking a pessimistic view of a train's braking performance means that in most cases, a train will be needlessly slowed down or stopped before it needs to be, leading to slower throughput and reduced capacity. And if you take an overly optimistic view of braking performance, you wind up failing to stop or slow trains where they need to be slowed, defeating the whole "positive" component of PTC!

The industry's been criticized for having tinkered with PTC for 20 years without deploying it as widely as its proponents would like. Yet problems like the one cited above haven't been licked, and remain a serious issue.

There is no question that PTC has some fantastic potential.

However, I have serious misgivings. First, whatever one thinks of Congress at any point in history, legislation passed in the wake of a tragedy when emotion wins out over rational thought seldom goes down in history as wise legislation. Second, the legislation that came out of Congress paints the railroad network with an broad brush. Is it not possible that conventional automatic train stop or automatic train control would be more than adequate for many parts of the network? Instead any main line with passenger service or toxic by inhalation hazard freight has to have a gold-plated PTC system, whether it be Nashville's Music City Star with its six daily round trips or the Long Island Rail Road with trains on 2-minute headways.

I wonder how much money will be diverted from other worthy projects by the 22:1 cost/benefit ratio of PTC implementation, and I wonder how much will be diverted from work (like grade separation between road and rail) that would save more lives over time.

railroadpostoffice wrote: Well when a recent railway age mag. feb 09 survey showed that 55 % of railroad custumers did not support railroad regulation...But that also says that 45% of railroad custumers are not happy with there service?

It means that as many as 45% of rail customers would love to see railroads regulated again so they could have transportation on demand and not have to worry that a competitor just might have negotiated a better rate.  If you were to break down surveys such as this one, you would find that it is not the percentage of shippers that counts; it is the type of freight they are shipping.  Bulk shippers of coal, grain, chemicals, etc., made out like bandits until railroads were deregulated and they had to pay a larger share of system fixed costs than before.  They've been lobbying for more  than 25 years to do away with differential pricing, although they think it's a fine idea that they use differential pricing themselves.  Merchandise shippers, including intermodal customers, are in the 55% that don't want re-regulation.  They are getting better service than ever before and at lower average rates.  Now, all we need, is for the railroads to do a better job of telling their story.

JRR:  I'm not a signalman (or woman), so I'm not going to try to nit-pick your very interesting and helpful comments.  But, as I understand it, PTC would intervene if the engineer fails to slow after going past the previous signal point, and could then bring the train to a stop at the red/stop signal.  

I think you're absolutely right about Congress.  This is what one might call the "do something before the sky falls again" syndrome - a bunch of vote-seekers acting on something about which they know nothing.  EPB implementation, because it can stop a train in less distance and time, might save more lives than PTC.  Some of us are convinced that the real value of PTC is in track capacity.  If you can stop a train in less distance and more quickly, you don't need current standards of block separation, allowing more trains to occupy the line.  The railroads were very close at the time of Chatsworth to making the PTC decision on their own and for business reasons, not safety.  That made the legislation a no-brainer for Congress, including the unfunded mandate; the august legislators were ordering railroads to do something they thought they were going to do anyway.  The 22-1 cost benefit ratio is what FRA came up with, and is a pretty good number, I'm told.  By playing up 22-1 and the fact that Congress ordered railroads to do something they wouldn't do for safety, the carriers are building a case for getting federal financial support for the PTC deployment.  

A minor quibble: grade crossing separation is mostly funded by Sec. 130 of the highway program.  I doubt many, if any, separation will be deferred because of the money being diverted to PTC.  If you could get the states to put up their share of the money, you would have a lot more separations completed.  You're right; that would save lives.

JRR

My apology for some generalities.

Metrolink Ventura Sub is 79 mph track, doubt that the train was doing 79 since there are some nice curves and a good ascending grade heading west (north).

GCOR state that trains must stop 400' short of signals.  I'll stab from signal to switch point is 200' more.  Based on 600' Metrolink was clipping along at 68 mph.  

So maybe I should have more clearly stated based on speed and tonnage of train that PTC would make emergency application at a certain point before major problem.  

So where would Metrolink have gone into emergency via PTC?  400' before signal like rules state?  Or factor in length and tonnage and add that to the 400'?

Based on what I said above point B is the signal.  If the engineer doesn't take some step to slow the train prior to 400' BEFORE the signal then brakes apply.  Train will still get through point B though.

Point is that dispatcher will know a train went into emergency and that gives needed time to alert other trains in the area.

27 seconds is certainly enough time to have gotten Metrolink stopped and dispatcher is calling UP freight to advise them of the emergency.

Don't forget I'm cynical.  :)    

BackToTheFuture:

Metrolink was surely not going 79 because the train in question was accelerating from a station stop. The collision occurred on a portion of railroad with a 40 mph restriction for curves.

There's plenty of factual data (like the track profile, operating timetable, event recorder data, etc.) on the NTSB site, so there's no point in guessing on what transpired. It's all laid out in excruciating detail at www.ntsb.gov/.../default.htm .

I don't doubt that a PTC system would set the target point for the stop at the approach side of the signal with that 400' margin. But I don't know of any ATS or ATC system (other than transit trip-stops) that dumps the train's air; a penalty application is typically 30 psi or a full service reduction. A PTC system capable of calculating a braking curve to a particular target point should damn well be able to find a less abrupt but effective way of stopping the train before the target without an emergency application.

Thanks for the elaboration on your previous comment; your intent is more clear as a result.

Hi, Larry.

PTC is (in theory, since very little of it is in practice today) is supposed to intervene any time it calculates that the human operator is not complying with a signal indication (be it a reduced speed or stop). One system I've examined calculates two braking curves, one warning the human operator that he or she must place the train under control, and the second identifying the point in time at which it is clear that he or she can not comply and where PTC computer must intervene.

Conventional cab signals and train control date back a long time, with one of the first being on the Pennsylvania Railroad in 1928. I won't get into history, but PTC sets out to do several things the tried-and-true systems did not.

First, traditional systems have speed control but the speed is not very granular and the lowest enforced speed is typically 20 mph. This is not acceptable for a high-speed operation; a freight engine that intrudes into the Acela's path is just as much a risk at 10 mph as it is at 25 mph. With today's technology, PTC seeks to enforce an absolute stop where appropriate, but also provides a way to get a release from the dispatcher in the event of a signal failure.

Second, conventional train control systems protect overspeeds at some civil speed restrictions but not others. An example is that most contemporary systems can enforce proper speed through, say, a 30 mph crossover, but do NOT enforce a permanent 30 mph restriction due to grade, curvature, bridges, tunnels, etc. PTC sets out to do this.

Third, conventional systems do not enforce temporary speed restrictions ("slow orders"). PTC sets out to do this. In addition to protecting the train, this feature protects on-track workers; a temporary speed restriction of 0 mph can be enforced to keep errant trains out of a work zone.

I do have to take issue with one statement you made about stopping a train in less distance and more quickly. PTC does not set out to change the laws of physics as they apply to train brakes -- perhaps you are thinking of ECP braking.

The concept of "moving block" is where a technologically rich system could improve capacity in certain scenarios. With conventional signal blocks, the distance between trains changes as the first train passes block signal locations and moves farther into the block and leaving more distance to the following train that is dictated by braking requirements.

The idea of moving block is that the safety cushion behind a trains is always the optimal amount and is not constrained by discrete signal locations or block boundaries.

Yet I haven't heard much about moving block lately, and the issue of calculating braking performance for a wildly heterogeneous consist may mean PTC actually hurts capacity more than it could potentially help it were moving block to pan out.

You hit the nail on the head about Congress. All it takes is a tragedy like Chatsworth to occur, and the halls of Congress are full of hand-wringing about the sky falling and anger about technology that the railroads were allegedly sitting on. Congress doesn't know thing one about the technology, nor do they understand risk analysis.

You're of course right that grade separation funds come mostly from highway funds, though there of course is a railroad component (10% or something like that). The 25 who died at Chatsworth is 25 too many, of course, but I'd rather see the estimated $20 billion to be spent on PTC spent instead on reducing the 35,000 deaths that take place on our highways every year.

Of course it's easier for Congress to legislate with someone else's money than with taxpayer dollars.

JRR thanks for the link.  The switch that was run through was on the straight portion before the climb and curves.  Definitely not going 79, but I was using your 6 seconds from signal to switch and 600' as the factor to back into 68.2 mph.

Heck if this guy wasn't even paying attention to the signals the speed of his train wasn't in his thoughts either.  :O

JRR, you know your stuff - well said! I think there's only one way to fix any capacity decrease PTC implementation will cause, and to Larry's point, that's EPB (or ECP - electrically controlled pneumatic), brake system implementation. From what I've read, ECP braking will reduce stopping distances sigificantly - as much as 50% for heavy trains moving at 50 MPH (I'm not sure about faster-moving freights - maybe only 30% reduction??). This happens because brake control is done electrically by wire or radio, rather than merely by the engineer bleeding air from the train's brake pipe/system thru a .125-inch orifice in the loco cab. Thus the 1 minute-plus it takes for all the cars in a 100-car train to sense a head-end pressure reduction and apply all brakes goes down to no more than several seconds.

As usual, $$$ is the problem. $$$ to install EPC equipment on all rolling stock, $$$ to maintain it, $$$ to re-space roads' signal blocks to allow trains to follow each other more closely, $$$ to revise control office visual displays and operating algorithms to take advantage of more trains on the same trackage.....$$$ - $$$ - $$$ - where will these funds come from?

bradyrs: Thanks.

It's not just the orifice; don't forget the propagation delay of air brakes is limited by the speed of sound, while ECP brakes can propagate at the [much higher] speed of light.

A question.  Let me start by saying I realize what I'm asking will not apply to a most trains as they are made up, at least not initially.

Why not make a segmented EPB/ECB system?  Instead of every car braking immediately, every tench car would control the nine behind it.

Obviously, this'd work a lot easier in trains that didn't get busted up every trip, but there's a fair amount of freight that applies to, often on restricted sections of track.

anmccaff:

A segmented system doesn't seem practical.

Remember that using traditional Westinghouse brake systems, the pressurized air in the trainline is used not only as the source of the energy that presses the brake shoes against the wheels, but the trainline pressure also serves as the means to signal the control valves (triple valves) one each car to open or close certain passageways for air on the car based on the pressure difference between the trainline and the car's reservoir.

Thus, while the engineer is drawing down the trainline pressure (to apply the brakes), it can't simultaneously recharge the reservoirs.

ECP brakes separates the energy source from the signaling function; electrical signals instruct the control valves on the train what to do, and the compressed air exists in the trainline solely as the energy that provides stopping power when needed. This provides a number of benefits such as instantaneous propagation of the signals (speed of light vs. speed of sound), which speeds brake response time. Graduated release is possible, as compared with the all or none release possible with conventional freight car control valves. Finally, recharge time is negligible since the trainline is continuously replenishing the cars' reservoirs.

With ECP, the trainline pressure isn't varying as the brakes are being set and released from the engine – remember, this is being signaled electrically. So non-ECP equipped cars on the train would still see full trainline pressure regardless of what braking commands are being transmitted electrically. Breaking the train into segments that are pneumatically separate doesn't seem workable either; how would the segments on the rear of the train get electrical braking control signals transmitted through non-equipped cars farther forward in the train, and where would they obtain the air needed to recharge their reservoirs without being pneumatically connected to the locomotives' compressors?

And even if it were technically possible, I suspect the necessity to precisely position equipped vs. non-equipped cars would produce little if any net operating benefit.

One of the reasons the railroads have done most of their ECP deployment on captive freight cars (like coal hoppers) is that they know that the cars will stay together and be ECP-usable, meaning the initial investment to convert the cars will yield immediate operating savings, rather than get broken up and scattered among the fleet, seeing actual ECP usage only sporadically.

Yeah, I'd thought of the the practical difficulties, but making a system that had a series of distributed braking control cars would get you 90% of the response speed for....well, I'd swag it at about 20% of the price; it'd obviously cost more to arrange the piping and valves such that it'd fail safe into conventional air brake mode, and store enough -or make enough- air in the specialized cars that you'd get fairly rapid brake release,  I don't think you could fine-tune the release, though, it'd still be close to off-or-on, but you'd only have to modify one car in 10.

I assume you'd have to use RC with repeaters; if you had to wire through each car you'd might as well go the whole hog.

Obviously, as you say, any additional money spent has to be recaptured through frequent use; coal, stacktrains, etc.

bradyrs asks (rhetorically?) where the money is going to come from.  How about customers?  They ultimately pay for everything involved in moving their goods from where they are to where they are wanted.

Fascinating discussion, gentleman.  I'll just sit back and read.  Keep it up.

Good writings, folks!! I suggested to my former signal employer several years ago (they didn't even know what I was talking about, as most of their emphasis is on the transit signal world, unfortunately) that we get into bed with Wabtec to jointly come up with a cheaper way to approximate PTC as you suggested, anmccaff. My thought was to position Wabtec's end-of-train devices every 20 cars in each train, thus for a 100-car consist reducing propogation delay time from 60 seconds to maybe 6. This would be possible because, unlike other vendor's EOT devices which can only be commanded to initiate an emergency brake application, Wabtec's can also be commanded to make normal service brake applications as I understand it. Why 20 cars apart? - because these devices would work like distributed power placed 2/3ds of the way back in a train - the cars both ahead & behind such power feel train line pressure change, reducing brake application time to 1/3 of "normal" (it took me 20 years to realize this after watching unit coal trains with locos 2/3 of the way back in the train rather than at the midpoint  - I'm slow sometimes!)

Of course, Wabtec is interested in the business profit of having to equip every car for ECP, as opposed to selling more EOT units. I also assume these "mid-train" EOTS would have to be strengthened? Are the forces from slack action mid-train better or worse than at the end of a train? And would the radio antennas for mid-train units have to be longer, or the units placed higher on the car, etc. And where would these units be added to a train - in the classification bowl track where the car knockers are coupling hoses preparatory to airing the track & testing its brakes??

Comments?

Regardless of where they are located, would they still be "Freds"?

Larry - as far as I can see, yes. Note that all current FREDs small air hoses only have a single glad hand to attach these units to the conventional air hose at train's end. So, modification of FREDs mid-train attachment to the air line would have to be done - some sort of "T" fitting with glad hands on each end would have to come with each mid-train FRED. You'd apply this T fitting like air hose extensions that are used on extremely long cars because of their coupler lengths & large overhang on sharper curves.

Other issue on segmented ECP is how do you know where/what the tenth car will be when humping/flat switching the train together at origin?

On coal and other unit trains where a trainset typically stays together coal is the only one that stays together nearly all the time other than some bad orders here and there.

How about implementing PTC on Amtrak first since the government then can get a taste of its own medicine?

RSB>"Comments?"

I think there are any number of ways of doing things better, but so much of railroading's past success has depended on uniformity that suppliers and railroaders are (justifiably) very cautious about adoption of new methods.  Railroading also depends on (relatively) cheap money that is heavily based on the fact that equipment is easily moved from one application to another.

I think this is the usual reason why new ideas don't get adopted fast.  You have people in an industry that still remembers 4, 5 decades of cash-strapped times, that depends on extremely long amortization rates for capital funding of much of its equipment, and that depends on reliable, predictable interchange of equipment, if needed, across state and national borders.  Of course they are going to be conservative.

anmccaff: I think you really defined the problem in your second paragraph above.  Being cash-strapped makes adoption of new technology difficult.  You really want to be sure it works before you sink a lot of money into it.  The railroad culture is conservative to begin with; each railroad always believes its way of doing something is the best way.  If you screw up on something like PTC , you're going to be paying for it for 30 years or so.  On PTC, I think we are seeing a lot of propaganda right now.  I know a couple of roads were all set to adopt it on their own, justifying the investment on business grounds, not safety.  Chatsworth just gave the idiots in Congress a chance to pander and posture for free.  If they knew it was going to be adopted anyway, as a few had been advised, then ordering it would cost nothing.  They were wrong, as it turns out, but I'm seeing too many of the same words coming out of CEO mouths to buy the "woe is us" story we are hearing.  There is a bit more conservatism at this blog than I think is justified; no shortage of reasons why something can't or shouldn't be done, not enough being willing to take a risk.  I always have known the railroad and rail supply industries to be very risk averse.

BBTF>"Other issue on segmented ECP is how do you know where/what the tenth car will be when humping/flat switching the train together at origin?"

Yupper.  It'd be a pain in the butt on consists that were regularly broken up.

BTTF>"How about implementing PTC on Amtrak first since the government then can get a taste of its own medicine?"

Joking aside, this is what I'd fear most.  A mistake at the early stages - or a system suited to passenger rail's needs, but not freight's - gets locked in early, and then the dog gets wagged by the tail.

anmccaff wrote:

BTTF>"How about implementing PTC on Amtrak first since the government then can get a taste of its own medicine?"

Joking aside, this is what I'd fear most.  A mistake at the early stages - or a system suited to passenger rail's needs, but not freight's - gets locked in early, and then the dog gets wagged by the tail.

--

Of course Amtrak already *has* PTC in the form of ACSES on the Northeast Corridor and the ITCS pilot project on its line in Michigan. But given how little of the rest of its network Amtrak actually owns, how is it able to install a PTC on property it doesn't own.

I don't think there's much danger of a passenger-centric system getting rammed down the throats of the freight carriers, and I think most everyone in the industry understands that there are differing needs there.

Further, the ACSES system, which has the most mileage of any PTC system in the mainland states today, was built on top an already-existing cab signal installation that was put in on what's now the Northeast Corridor by the PRR in the Thirties. From what I know of the system, ACSES complements the conventional cab installation and thus wouldn't be suitable standalone.

Does a cab signal system protect MOW workers?  PTC is not another signal system.  It is active and it takes control of the train when it exceeds its track authority.  It does't just stop the train if it blows through a signal.  It handles movement from track-to-track in multi-track routes, and if the engineer is doing his job properly, he'll not even be aware of PTC being available.  My point is: this is not a new-fangled signal overlay that someone just dreamed up.  It's been around for more than 20 years in various stages of development, and it has worked where tested.  No one was willing to invest the money to implement it because, happily, there are not enough collisions to justify the expense of preventing them.  The Class 1s working with GE, Rockwell, WABCO, and others, have known all along that the value of PTC is in added track capacity, which is cheaper than building additional track.  All of the negative comments here aren't going to change the reality that Congress spoke and FRA has drafted the rules.  Deal with it; it's going to be installed.

PTC's only involvement in track-to-track moves is again, to avert collisions. Conventional signal equipment handles & protects the routing of a train from one track to another. Thus, while we'd like to believe that the existing signal systems can be replaced by PTC, THEY CAN'T. As I said previously, how will PTC correctly calculate movement authorities based on the perfomance characteristics & stopping distances of following trains given the huge variation of these trains' speeds, weights, lengths & locomotive power? THESE AUTHORITIES WILL HAVE TO CONTINUOUSLY CHANGE! Thus safety factors that affect distance between trains will have to be included, which will DECREASE CAPACITY of a given track.

I believe true moving block with ECP braking is the only way to maximize capacity of "through" tracks. If PTC acts as moving block, and ECP braking comes to pass, through track capacity WILL be maximized. However - we must remember that railroads are intra- & interconnected systems. Maximizing through tracks' capacity will put more pressure on already constrained terminals, hump & flat switch yards, intermodal loading & unloading facilities and interchange ops with other roads.

Just speeding up main lines and/or increasing their capacity will also put more pressure on what's at their "ends", requiring more of those $$$$$ to increase these ends "throughput".....

Larry, no one here is saying that PTC is just another cab signal system, and in fact a number of people including me have SPECIFICALLY mentioned the protection that PTC gives to on-track workers, which no previous system has done.

Please reread what I wrote previously. Amtrak's ACSES is indeed an overlay and DEPENDS on the underlying conventional cab signal on the Northeast Corridor on which Amtrak's deployed it. This was Amtrak's decision based on its analysis of how best to leverage their sunk investment into conventional cab signal equipment while taking the Corridor to the next level of protection. My point in the last reply was that the technical and financial equations that led Amtrak's to their choice of systems are not likely to apply to freight railroads, and for that reason as well as some technical ones, I don't think the tail will be wagging the dog as one other commenter wrote.

Regarding PTC (in a broad sense) being an overlay, I made NO such statement about any PTC installation other than ACSES, and some of the systems being tested in fact do not even depend on the presence of track circuits or a wayside signal signal system, one example being BNSF's ETMS prototype on dark railroad in Illinois.

The fact that PTC has been around for more than 20 years and has worked where tested does not mean that it is not without issues and is ready to be blindly rolled out everywhere. Listen to what signal engineers and operations planning people have to say: there are some real, unresolved issues with PTC that some brilliant people in the industry still grappling with. You tout PTC's benefits in improving capacity, yet there is evidence that PTC produces negligible capacity gain in some situations and even reduces capacity in others. Don't take my word for it; talk to the engineers, planners, and analysts and see whether THEY see a panacea that the industry's been sitting on.

To quote from www.fra.dot.gov/.../PTCRSIAfinalRIA120809.pdf

"At present, the PTC systems contemplated by the railroads, with the possible exception of one,

would not increase capacity, at least not for some time.  If the locomotive braking algorithms

need to be made more conservative in order to ensure that trains do not exceed the limits of their

authority, PTC may actually decrease rail capacity where applied in the early years.  Further

investment would be required to bring about the synergy that would result in capacity gains."

The financial benefit or cost of any change in capacity resulting from the PTC mandate is presumably calculated in the FRA's 15:1 (later revised to 22:1) figures.

About your closing statement that Congress has spoken and that we should just deal with it. Sorry, I don't work that way. As a citizen and taxpayer who prefers to be engaged in public policy that concerns me, the acts and omissions of Congress are important to me -- particularly when Congress formulates a fatuous "solution" whose costs are wildly out of proportion to the benefit that will accrue. As an investor, I should be concerned about the cost:benefit ratio and what this type of expenditure means for the railroad industry in terms of both the capital that could be used elsewhere, and the effects PTC may have on the railroads' capacity to perform their duties.

Finally, for those saying "deal with it" because Congress has acted, railroaders like Wick Moorman are not willing to suffer the PTC unfunded mandate in silence. Why should the rest of us?

JRR - again, well said!!!

JRR:  And, please read my previous post here.  I asked whether cab signals can protect MOW people.  It was neither sarcastic nor argumentative.  I simply don't know the answer and figured someone here would tell me.  Nor have I suggested that anyone just roll over on PTC, although I do believe that once it is a law and I see no one campaigning to repeal the law, it is something you are going to have to deal with.  Wick Moorman, for whom I have great respect, is not suffering in silence, nor is he ignoring Congress.  Speaking out?  Absolutely.  Putting company lobbyists to work on the issue?  Absolutely.  As for the FRA 22:1 cost:benefit ratio, that is focused on the safety issue and not on productivity.  Congress didn't give a rip about railroad productivity when it postured and pandered following Chatsworth and passed the Rail Safety Improvement Act.  I don't disagtree with you that it may be a bad piece of legislation, but legislation it is.  Decrying it, especially at a blogsite populated by railroaders and rail afficionados, is not going to cause Congress to change a thing withe the law.  And, yes, BNSF's ETMS works well in dark territories where they are working on it, which is because it is not a signal overlay, but a new technology.  As for Amtrak and the NEC, I have just two words: Ricky Gates.

Larry:

The red herring about Ricky Gates puzzled me.

Anyone who knows about the catastrophic crash that Ricky Gates caused at Chase, MD in 1987 is surely aware that the Chase wreck led to random drug and alcohol testing for railroaders, federal certification of locomotive engineers, and requirements about all trains on the Northeast Corridor having operable train control device.

Most pertinent to this discussion is that Chase led to FRA policymaking that was a progenitor of positive train control in the NEC, particularly as the New England electrification moved forward, and as Amtrak sought to bump up the NEC's speed to 150 mph in spots.

You surely know that with Amtrak eternally starved for capital, scrapping the existing tried-and-true cab signal system on the Corridor was NOT an option for Amtrak and the commuter carriers operating on the NEC. Amtrak engineered a hybrid solution (ACSES) that made use of the perfectly functional existing cab signal system, overlaying it with transponders, data radios, and track databases to provide PTC functions like positive enforcement of stop indications, on-track worker protection, and enforcement of civil speed restrictions that are not provided by the earlier systems. It doesn't use GPS, but it sure meets nearly every definition of PTC I've ever seen.

I'm trying to understand what your point was to bring up Ricky Gates when I mentioned ACSES on the NEC. I'm sure you're not forgetting that Ricky Gates spawned the push for ACSES. Also, I certainly hope you are not suggesting that the PTC Amtrak engineered for the NEC as a result of Chase is an ineffective system. So I'm scratching my head as to what your "just two words" (Ricky Gates) was intended to imply.

One thing this discussion reminds me is how confusing signaling and train control systems can be to those who aren't familiar with their many mysteries.

You are right that Congress doesn't give a rip about railroad productivity, but we should.

And one can't really compute a cost/benefit ratio without tabulating all the costs (implementation and operating costs of PTC, plus any negative effects on productivity/capacity) balanced against the benefits (lives saved, injuries prevented, cleanup expenses prevented, productivity/capacity improvements). There are obviously arbitrary monetary values involved.

Wick Moorman is certainly astute enough and gentleman enough not to stand before Congress and demand the mandate be repealed, though he may well be doing his best to help them try to come to their senses. And failing that, I'm sure he'd like at least some funding for Congress's mandate.

JRR: I can agree with almost everything you have said.  I do not consider reminding readers of Ricky Gates to be a red-herring, however.  Everything you say about him is so, particularly about the various improvements it triggered in the NEC and in drug and alcohol testing.  Had there been a PTC system in service, I don't believe the disaster would have occurred because obviously Ricky Gates' unit had exceeded its track authority.  And the Colonial, which was moving faster and coming up behind Gates would have been slowed or stopped as I understand PTC.  I do plead guilty to not being a signaling expert, as I've stated on several occasion.  Perhaps some of you experts can explain things in the English language for those of us who do not have degrees in electrical engineering any other kind of engineering.

You're right that Wick Moorman is a gentleman.  I see no reason, though, why he or other rail senior executives should not seek repeal of a law that is badly conceived and flawed in its mandates.  That's perfectly acceptable in a representative democracy.  He and his colleagues have every right to ask, and ask publicly, that the unfunded mandate be funded.  Any funding sought, though, would have to come from the same Cngress that ordered PTC implementation with no funding.  You think many in Congress will admit to having been wrong?  Meanwhile, the Administration, including Ray LaHood and Joe Szabo have been mute on this subject and virtually any subject involving freight railroads.

The FRA cost/benefit calculations support the railroads' position.  Of course we don't know whether the industry is trying to negotiate an ITC trade-off for the unfunded mandate.  I'm reasonably confident that they are not just fulminating, but are doing what they believe the should.  As for the FRA cost/benefit analysis, as I understand the process, because PTC was mandated solely for safety reasons, that restricts what the FRA can/could do.  It attempted to add up all of the capital and operating osts of PTC and balanced that against the identifiable benefits (collisions net experienced), and came up with 22:1, not really justifiable by any rational analysis.  If there are productivity gains, as I believe, that did not factor into FRA's analysis.  I don't know that FRA included any effects on productivity, but it sure had plenty of costs in its analysis.

Larry:

I for one would be happy to help educate folks about some of the principles behind signaling and train control, hopefully without delving into much electrical engineering. I don't know where to begin, because these are tricky subjects, but I perceive that there is a knowledge gap even within the industry, as I hear about capacity planners who don't grasp how signal systems has an immense effect on capacity.

I don't think discussion of Chase is a red herring, but I didn't understand the conclusion I thought you were trying to draw between Ricky Gates and the ACSES overlay to conventional cab that Amtrak used to successfully deploy PTC on the NEC (to the FRA's satisfaction, I might add). Perhaps we were talking at cross-purposes.

Obviously the technology and knowhow weren't available in 1987 to have prevented Chase, but a simplified version of what happened at Chase and what would have happened if ACSES had been in service on that fateful day is roughly as follows:

The Conrail engines passed a wayside signal displaying "Approach" along with a corresponding cab signal indication accompanied by an audible warning (which in fact was taped over in the locomotive in question). These wayside and cab signal indications required a reduction to 30 mph. With conventional automatic train control that was installed on passenger equipment operating in the NEC but not on freight engines, the speed restriction would have been enforced by the conventional systems without PTC.

With ACSES or a similar PTC system, the train would have passed a wayside transponder located between the rails. The data transmitted to the train includes the distance to the next absolute signal (in this case the absolute signal at Gunpow) to allow calculating a warning curve and enforcement curve for the train control system.

The Conrail engines passed a track circuit boundary 4,450 feet from the "Stop" signal. The cab signal would change to its most restrictive indication ("Restricting"), which requires reduction to 20 mph. With conventional train control, this speed would have been enforced. One might conclude that the penalty brake applications would have drawn Gates' attention to his situation even where he ignored the visual and audible signals, but in any event conventional train control systems don't know the difference between an absolute "Stop" signal (such as at Gunpow interlocking) vs. a "Stop and Proceed" elsewhere, so he could well have continued into the Amtrak train's path as long as he didn't exceed 20 mph. One could still have a catastrophe in this case.

Now let's step back to that approach signal and to that transponder I mentioned. The transponder tells the on-board computer how far to the next absolute signal and the input from the conventional cab signal system tells the computer whether a stop needs to be anticipated at that signal. It then uses track profile data to determine the average grade and the train's braking profile (programmed at point of origin) and velocity to determine the braking curves. The warning curve defines the point at which the engineer should be warned that it's time to brake, and the enforcement curve defines the point at which the on-board computer takes over. As long as the engineer keeps within the enforcement curve, the computer does nothing but monitor. But this is being continuously monitored, second by second, and recalculated based on the train's speed and distance to the target.

So what would have happened at Chase is that this system would have intervened at the point at which ACSES determined, based on the train's braking profile, velocity, and distance to target, that the train had to be braked in order to not pass that target.

Even the most conventional of cab signal systems would have warned Amtrak 94 if someone had intruded into its path (or if I had gone out there with battery cables to short-circuit the track circuit). The wayside signal DID go to Stop and the cab signal DID go to its most restrictive indication, but the Amtrak train was traveling around 120 and could only have slowed insignificantly in the short compass of time available before the inevitable occurred. Had Amtrak been a few minutes later, it would have seen restrictive wayside and cab signal indications as the freight intruded into its route, but could well have stopped. Timing was not in the victims' favor that day.

The beauty of a system like ACSES or the other PTC implementations is that the freight train would never have been able to pass its target point behind that "Stop" signal.

Thanks a lot.  That discussion is very helpful.