One of the things they did after 9/11 was require all aircraft to monitor a common frequency. This actually saved an A320 plane from a mountain terrain disaster when the previous radar center noticed a blip out of position and transmitted on that frequency.
TheLoadedDog, The same is true here. The conductor is less senior to the engineer but is in charge of the train. It’s a little weird for the reasons you cited plus the conductor isn’t skilled in train handling. I’ve only been a qualified conductor for just over two years but I’m fully prepared to take matters into my own hands. I know what to look for. In fact, if you ask our engineers, one of their biggest gripes is that they feel like their conductors aren’t responsible enough inside the cab.
If any crew member messes up they usually will hold the entire crew accountable these days.
How does the “Two trains cant have a form D for the same piece of railroad” system work? Is it an administrative procedure or is there some physical fail-safe that prevents two being issued?
One of the railway experts will probably be able to say more but as I remember in the UK a “token” system is used where electrical interlocking ensures only one token for each block can be issued at a time. I suspect this is done electronically now rather than the driver being handed a physical token by the signalman but I think the principle still applies.
They call this phenomenon the “authority gradient”. We heard a lot about it after the crash Loaded mentioned. Most of the examples they gave us were from the airline industry. Airline X’s chief pilot makes a glaring error, dooming the plane, while the junior co-pilot sits and watches, too scared to question the senior man. Our resident plane geeks will probably know all about this stuff. Things are a little better than they were on our railway now, at least we’re aware of the problem. Cultural change is a slow thing though.
There seem to be conflicting stories about the driver - conductor roles at Metrolink. Someone said that the driver and the conductor are both responsible for observing and reacting to signal indications. Then someone else said that the conductor wasn’t on the locomotive at all but was back on the train somewhere (luckily for him, in this case). How does the conductor observe signal indications, confer on same with the driver, and take appropriate action if necessary, from back in the train somewhere?
You don’t need necessarily a token system for single lines to prevent conflicting and/or following movements. Electrical relay, and later on, computer based interlocking has been protecting single line sections for going on a century. As long as you can detect the presence of a train in the section with track circuits or axle counters, you can run trains on signals, no token required.
Even in so called “dark territory”, a system of forms often referred to as “train order” working can control trains over single lines, with varying degrees of safety, depending on the way it is implemented. The more primitive versions are definitely less safe than a “token” based system. You save the messing around with tokens but open up the prospect of death by administrative error. The wags here call them “suicide notes”. We have a pretty decent version of train order working here in New South Wales, on some of the country lines which have not been deemed worthy of CTC signalling. Our system uses a computer for the train controller (dispatcher) and codes on the paper forms used by the trains. I’m not fully conversant with the system but I believe that the computer keeps track of how far the controller has authorised each train to travel and won’t let him set up a conflicting move. The driver gives the controller a code number from the form he fills out when he gets his order. This code number goes in the computer. When the train has carried out the instructions on the order, the driver gives the controller a second code number off the form, which lets the computer release the specified section of track for another train on another order. As far as train order working goes, it’s not a bad system. It still relies on drivers pulling up where they’re told to though.
I can’t vouch for train order working in other parts of Australia or the world. There seem to have been more “corn field meets” on train order systems than on token system, hence I’m dubious of train orders. None of this is really relevant to the Metrolink crash, which was in signalled territory.
Yes, but is all this worthwhile? Railway systems in the western world have been operating for 100+ years using little more than lights and warning bells with an excellent safety record. The trouble with rail is that on the incredibly rare occcasions when you do get a smash, it’s all very telegenic with lots of casualties and everyone gets their knickers in a twist for no good reason.
After the Ladbroke Grove crash in the UK, there was a big hoo-hah about how inefficient and unsafe the victorian-style lights system was, and a decision was made to immediately install a fully automated system for stopping trains, despite an estimated cost of £1-2 billion, approx £14million per estimated life saved. Then the cost estimates started going up and up and up, the immediate hoo-hah died down, and everyone quietly agreed that spending billions to save a few dozen lives a decade on the railways was just silly when people were being smushed by the thousand on roads without anyone caring. And so they are still using lights & bells, and over the next decade or two will gradually roll out a (much cheaper) partially automated train stopping system.
It’s remotely conceivable that something broke, but given that pretty much everything to do with railways is very carefully designed to fail safe it’s extremly unlikely equipment failure caused this unless a human who was supposed to spot it failed to do so. Operator error? Not necessarily, but I’ll bet a dollar that a human somewhere messed up.
But a foolproof system doesn’t need to cost billions. Hell a bluetooth transmitter connected to the red light, and a bluetooth receiver in the cab hooked to a loud obnoxious horn would work just fine. Just don’t let the deaf guy drive.
Like I said before it isn’t rocket surgery.
Two things about the LA crash-
One- the MTA spokeswoman who stated that it was human error and took responsibility on the MTA’s behalf is now out of a job.
Two- the NTSB has already said point-blank that if the MTA had implemented the suggestions the NTSB made after the 2005 crash in LA, this would not have happened.
:eek:
WHAM! (That’s the sound of me hitting the floor after I fell out of my chair.)
Her statement did seem odd when I heard it. Now that I think about it, an admission like that will open them up to even bigger lawsuits than they would be facing.
The NTSB opinion does seem on point. Way too much riding on the vigilance of one person.
I think dispatchers have some sort of blocking device they use to prevent them from letting two trains into one track but I’m not sure. I do know of a recent case where a dispatcher mistakingly gave one train a form D with another train in the block.
When someone has a form D they can give back pieces of railroad as they pass station signs. Their rear car must be two miles beyond the last station sign.
We have a new system where I work and no one really seems to understand it all that much, which is scary. We did away with form D’s and now have track authorities. It seems that our company made policies for the bulk of the system but not for those of us at the end of the line who have to work on several railroads. We have to have with us all the rule books for all the railroads we travel over, which gets unwieldy to carry let alone memorize.
Rick & Slaphead…like I said upthread, we already use LSL’s on amtrak, which makes the engineer conform to cab signals within a given time frame or the train stops. The software is already on the trains. It’s just that all railroads don’t use it.
I read that the only fatality among the crew of either train was the engineer of the Metrolink train. I don’t know if the crew of the freight train were injured or not.
It was pretty stupid to say. They may “know” that the train went through a red signal, but they don’t know that it was the driver’s fault. That’s why they do an NTSB investigation.
I was reading through some NTSB accident reports, because they are fascinating, and one collision between freight trains was caused because during a one hour period in the afternoon, the sun caught the (unlit) green bulb in a signal and made it as bright as the red. So the engineer got an impossible signal, and didn’t handle it right.
It’s clearly a case of diminishing returns once you go beyond conventional signalling and start engineering out human factors. You’re right, Metrolink don’t necessarily have to go all out with an advanced ATP setup. It looks lile they need to do something though, two head on collisions in six years is abysmal. They’ve killed 27 and injured 157 people in those two accidents. The lawyers are going to take a big bite out of whatever money they saved by not doing more for prevention.
There are plenty lower tech defences to be considered. The old GWR in the UK had an electro-mechanical system for stopping the train if the driver ignored the caution signal before the one with the red light. That was early 1900’s, on steam locomotives. Trainstops, mechanical devices that engages a valve on the passing train, applying the brakes, have been used all over the world, mostly on multiple unit passenger trains, for a hundred years or so. Trainstops generally only activate the train’s brakes after it has passed a stop signal, so you need a decent “overlap” distance engineered into your system to allow the train to pull up. There are many “off the shelf” train protection systems, of varying complexity and expense, from reputable signalling equipment suppliers.
To avoid head-on collisions, having points (track switches) at the entry to the single line which will divert an approaching train towards a dead end track when the protecting signal is at stop, is a simple, effective measure. It’s better to derail a train, in the place of your choosing, than to keep it on the rails and let the train coming the other way stop it. This is a good option because it would work with the current signalling system, would cover all trains, and no additional equipment would need to be fitted to locomotives.
Metrolink’s problem is that it’s not the sole user of it’s tracks. They could spend a fortune making their own trains and signalling system safer, but they share the route with freight trains. Every locomotive leading a train needs to be equipped with whatever system you use. The pool of potential freight locomotives would be huge. The freight operators aren’t going to be keen to pay to fit all their locos or restrict the pool of locos than can lead over Metrolink’s lines. The inter-operation of locomotives and rollingstock in the US is one of the great strengths of the system, but it’s a bastard for fitting non-standard equipment.
Ideally, the US railroads would all agree on a standard train protection system. Even if they didn’t all adopt such a system, those that wanted or needed better safety would know what to buy.
This is a nitpick, but if you’re talking about the accident I think you’re talking about (the one that killed two), that wasn’t a Metrolink error from what I recall. I thought a freight train ran a signal and hit the Metrolink train, which was stopped at the time.
ETA - from Wikipedia, and I was half-right:
But the 2005 crash wasn’t really between two trains–a guy left his car on the tracks at a crossing in Glendale. The southbound Metrolink train crashed into the car and derailed, crashing into a stationary freight train and a moving northbound passenger train on other tracks. There were three trains and three tracks.
A signal system isn’t going to detect and alert about a car left on the tracks. But investigators did say that when a train is being pushed (rather than pulled) it’s more likely to jackknife and derail in a head-on collision, generally leading to more deaths if people are in the first car. In the 2005 crash, if the engine had been pulling the train it probably would have just totaled the car. Friday’s wreck did not involve a pushed train.
At least, that’s what I remember.
Yes indeed. I think push-pull is an utterly stupid, stupid thing, and should be outlawed.
Here is an article that explains the dangers inherent in this system.
There has been a lot of focus in this thread on active (before an accident) rail safety, but not so much on passive (during/after an accident) safety. This latter can be summed up as follows:
- Try to keep your train on the rails
- If it must derail, try to stop the couplers breaking. Then you have a heavy locomotive acting as a battering ram against trees and other obstacles, with the lighter and more fragile (and loaded with human beings) carriages following relatively safely in its path.
- Try to keep the derailed carriages upright even if they are freed from the rails (following the loco will help in this).
Better to have the much heavier locomotive acting as a ram removing obstacles than as a piston shoving the passenger carriages into those obstacles.
How is this any better than the ancient (as in, Victorian) system of rigging a plate between the rails to ring a bill in different patterns depending on whether the signal just passed was at one amber, two ambers, or on red? Your suggestion is not foolproof, and I’d be astonished if they weren’t using something that is functionally equivalent on the stretch of rail in question - I can certainly remember the regular signal bells coming from the cab of the crappy little diesel ‘skipper’ train I used to ride to school back in the eighties.
I believe most SPAD incidents are because the driver for some reason missed the visual signal and overrode/ignored the audio alert from force of habit. Like I said, human error is the primary cause of these horror shows, and most of the simple straightforward solutions to reduce it were put in place decades ago.
chicken wire? - I’m not a train geek, I only know how things work here in the UK from decades of riding the things around and soliciting explanations as to why they have failed to get me to where I want to get to on time. Does the US not use cab signals and so on?
Way oversimplified. One, the Bluetooth transmitter puts out too little power for this to be even considered. Second, just because an alarm sounds does not necessarily ensure it will elicit a reaction. Third, if the engineer does have a stroke/heart attack/otherwise loses consciousness, it would do no good. Fourth, even if the engineer does brake immediately, it may be too late, depending on the location and speed of the oncoming train, whether the engineer can reverse his train back before the signal he passed in time, and so on.
I will assert that no system will be 100% foolproof, especially one that is software based or requires a circuit board to function.
Traveling via rail is still inherently safer than travel by highway by several orders of magnitude. Cite: http://www.bts.gov/publications/transportation_statistics_annual_report/2005/html/chapter_02/transportation_fatality_rates.html
We have no evidence that the engineer was incapacitated. A much more likely scenario is that he failed to notice the one and only one red light that was between the spur for Chatsworth Station and the main line. (LA Times had an illustration of the track arrangement and the location of the lights. Two out of the three lights are located south of the station, and would not be visible once the train was in the station.*) If there was a low power radio transmitter that triggered a reciever in the cab that had a latching relay hooked to a Mother Fucking loud horn** that required the brakes to be applied to shut off, it would fix any and all inattention issues.
*Maybe it’s just me, but this strikes me as a really shitty arrangement of lights. I would have put the second light where it could be observed from the engine while the train was parked in the station; but hey what do I know?
*Or if you want to go high tech go to a voice synthesizer that says Brake or die, hey idiot, brake or die!
Black Train Song, is this what you are describing? Or this kind of thing? I see it mentions the cost, and incompatibility:
From twincities.com.
I’ve been on the railways for almost 20 years now, first in signalling system installation and maintenance, and more recently as a driver. This is in Australia though, so anything I’ve picked up about the US is second hand from books and other sources.
The US rail network is huge and there would be quite a lot of variety in the level of sophistication of the signalling used. They certainly do have cab signalling in some places, but not everywhere. I don’t know if Metrolink have anything in the way of in cab signalling, but if they did, it seems not to be backed up with anything to enforce correct actions by the driver, or we wouldn’t be having this thread.
I really don’t think that poking the driver with a stick on passing a red light is as good as automatically applying the brakes. Throwing the emergency brake on will get his attention every bit as well as your MotherFuckingest loudest horn, with the added advantage of at least starting to pull the train up before the even louder bang if it hits something.
The railroads like your thinking though, and have been blasting drivers drivers with nasty noises for years, in the vain hope that it’ll keep them out of trouble. The US term is, I believe, an Alerter. We call the same thing a vigilance control system. Same idea.
Basically, when the train is set up to move (ie the locomotive brakes aren’t fully applied) a timer starts running. The timer can be variable (to keep the driver guessing) or fixed. It can be speed linked to cycle faster at higher track speeds, or not. It can also be linked to the locomotive controls, so that if the driver is seen to be awake and doing stuff, it resets. This timer thing will typically be set for between 30 seconds and a minute (maybe varying randomly or with the speed of the train, depending on the system used). At the end of that time, a light will start flashing in the driver’s cab. Five or ten seconds later, an audible warning device will start up. In the old days, this was a Fuckoff Loud whistle in the cab, with air from the locomotive’s main reservoir stuck up it’s arse at 120 lb’s per square inch. More recently bells and electronic noise makers are favoured (because you could muffle the old whistle with a bit of rag and deafness claims are expensive). If our poor driver can sleep through that, we figure he might be dead or incapacitated, and open a magnet valve to let the air out of the train’s brake pipe, stopping the train.
OK, so we’ve got this device that will reliably scare the bejesus out of the driver once a minute. We admit though that this isn’t exactly sporting and conducive to a happy driver. So we give him a button to press, to prove that he is infact alive and awake, and pressing the button restarts the timing cycle. There are two options with this button. We could let him press it whenever he wants. The trouble with this option is that the driver will learn to press the thing all the time, without thinking. I’ve heard stories about old drivers who could sleep like a log, one hand twitching at that button every ten seconds or so. The second and probably better option is to lock out the button until the system demands a response at the flashing light stage. We moved to this second option here in NSW after a two man crew on a massive coal train somehow dozed past several red signals and snotted the back of another massive coal train. Luckily, the sight of the rear of the second train looming up revived them from their stupor and they jumped for their lives. No, we don’t have a proper train protection system in a lot of places either.
The thing with this alerter/ VC thing is, you can’t make a tired, intoxicated or comatose person perform better by yelling at them. You can’t improve their cab discipline or take their mind off their wife running off with the next door neighbour, with an alerter. You’ve got to help the drivers manage their health and fatigue (which is often 180˚ to the interests of the employer, who wants them to work long, ungodly hours), and you’ve got to accept that the driver will fuck up at the worst possible moment, and try and put in more and better defences than a bell or whistle.