According to today's Wall St Journal, the crash occured when a turnout on the main line was left, and locked, lined for the siding upon which the CSX freight was parked. No discussion of who was supposed to keep the turnouts lined properly has come thru the media to me.
I was under the impression that who ever lines a mainline turnout to a siding is responsible for putting it back lined for the main line as soon as possible. Is this right? If true, then the CSX crew that put the freight on the siding was supposed to reline the turnout for the main line after they got the freight on the siding. And that didn't happen, causing the wreck.
What does everyone think?
David Starr www.newsnorthwoods.blogspot.com
I think your premise about the crew needing to reline the switch is correct, but not that the one who lined the switch needed to be the one relining it.
The CSX train was apparently a string of autoracks shoving back through the switch into the siding in question. The person lining the move was also the one riding the point of the shove, and he would have been far back relative to the head end, which was itself the length of 91's consist, about 700', back from the fouling point.
Intensive discussion of this on the Trains forum (the more appropriate place for a discussion of the subject) indicated that the correct action to protect the shove was to lock the switch, as the high-security key can't be removed from the switchlock until it is fully locked and no employee should ever leave one of those keys sitting out of his or her control. This at least implies that the 'proper' action for the crew of the CSX train to take would be for someone still on the head end with a key, probably not the engineer, to walk up and reline the switch as part of tying down the train ... and tied down it appears to have been, as no one was on the engine to be killed during the collision, as they almost certainly would have been.
dstarrWhat does everyone think?
ccording to the news, Amtrak thinks it's CSX's fault too. Human error seems to be difficult somehow to extinguish from the equation.
Rio Grande. The Action Road - Focus 1977-1983
Overmodcontrol. This at least implies that the 'proper' action for the crew of the CSX train to take would be for someone still on the head end with a key, probably not the engineer, to walk up and reline the switch as part of tying down the train ...
The conductor should have got a ride back to the headend by contract cab or maintenance vehicle so,he could lock and line the switch for the main.
However..
The photos look more like a sideswipe then a headon since the CSX units looks to be beyond the fouling point..
Larry
Conductor.
Summerset Ry.
"Stay Alert, Don't get hurt Safety First!"
BRAKIEThe photos look more like a sideswipe then a headon since the CSX units looks to be beyond the fouling point..
The critical switch in question is behind the last cars of 91's consist, not anywhere near the CSX power. You can see it clearly in some of the news coverage photos, including a couple of good views almost straight back over the last three cars.
Someone measured the distance using a tool on Google Maps, and got about 700' from the fouling point back to the freight power. 91's P42 and all the trailing cars apparently made the switch and proceeded down the siding to collision with the standing CSX train -- there was no switch in that area for there to be a 'sideswipe' although one poster thought the locomotive might have been rocking severely on its suspension all the way to the point of collision.
Brakie,Actually, if you look at the overhead drone footage (linked at the Trains Forum discussion), the front truck of the Amtrak engine is sitting in front of the CSX loco, which is roughly 700' (the length of the Amtrak train) from the switch (in the clear, IOW).
https://www.youtube.com/watch?v=Fp36W9L741E
It was a head on, straight on, at roughly 60mph at O-darkhundred. Coming around the curve, the Amtrak crew may not have seen the switch target (if it had one) and not known anything was wrong until they hit the switch. The front of the Amtrak loco folded under behind the cab doors, launching the 130+ ton loco up and over the CSX loco (note that it hit the radiator box at the rear of the CSX unit) and dropping the front truck right at the snow plow of the CSX engine.
Also note that a mid-train Amfleet car got snapped in half. Those aren't two Amtubes jacknifed, that's one car in two pieces. Wow. I haven't seen that before. I'm amazed more people weren't killed.
I did some checking on Google map and what I thought was a switch wasn't probably a optical illusion.
Unless the normal position for a switch is in the reverse position you never leave a switch off the main line in that position.
NTSB sez:
1. Speed limit was 59 mph
2. At no time did train 91 exceed 57 mph since its last stop.
3. 7 seconds before the tape stopped (presumably, impact), the train was going 56 mph and the horn began to sound for about 3 seconds.
4. During the final 7 seconds, the brake pipe pressure dropped, the throttle went to idle, the speed went to 54 mph or less, and emergency braking was initiated.
5. The passenger train traveled 659 feet from point of switch to point of impact, and pushed the freight locomotives back 15 feet.
Here are some of my own thoughts:
1. The lounge car that jackknifed was probably unoccupied because it was closed for the night (this occurred about 2;35 am). In all my years of working at Amtrak (now retired), I never would have imagined that an Amfleet would collapse like that.
2. The crew members in the cab might have survived if they had been able to get out of the cab and into the engineroom. But there was probably only 7 seconds to do this, so it was impossible. Survivability in the engine room is questionable, however, since the damage to the locomotive can only be termed catastrophic.
3. No Amtrak personnel had the ability or opportunity to throw that switch. Its incorrect position was entirely the responsibility of CSX.
4. The law will not hold CSX responsible. The costs of this accident will be borne by Amtrak, in spite of Amtrak's innocence.
5. The news media are treating this as they usually do. They are almost all being very superficial, focusing on Amtrak's problems without paying attention to the plain fact that most of Amtrak's problems are imposed by outside influences, and are not the fault of Amtrak or its dedicated employees.
Tom
If I understood Sumwalt correctly, the train was still at full throttle 7 seconds before EDR cutoff, moving at 56mph. The distance reported from 'the switch' to the point of impact is 659' (which indicates to me that the horn was initiated after the locomotive traversed the switch). The horn was sounded for 3 seconds, and during this time (starting at 5 seconds from impact) the blended brake lever was actuated and the throttle went to idle (which I believe is a feature of the P42 design when the brake lever is moved).
Emergency brake actuated only 3 seconds before cutoff. Train was at 50mph at EDR cutoff. I note that the road bridge may have blocked clear view of the presumably dark CSX power up to the moment the engineer reacted with the emergency brake.
I suspect Tom is right in thinking the engineer and conductor would not have survived this in the engine room; deceleration that can shock-buckle Amfleet shells would likely produce aortic dissection among other kinds of high-speed trauma even in an uncompromised 'safety space'.
Apparently the CSX crew actually released occupancy to the dispatcher before relining the switch. Sumwalt indicates they would have set a derail; if so, this might explain the tipping to a glancing collision while leaving the lead truck where it can be seen, right up against the CSX power even after the impact moved it 15' back.
This is not looking too rosy for the two guys on the CSX move.
OvermodThis is not looking too rosy for the two guys on the CSX move.
First I agree but,the engineer is not allowed to get out of the cab to line a switch due to a FRA regulation that came into being after 8888 got loose after the engineer left the cab to line a switch so,it back to the conductor.
Again he should have caught a contract cab or maintenance truck and went back to the headend to lock and line the switch for the main.This he failed to do.
The final results will be the CSX crew will lose their job and the conductor may be charged.
Had there been a head brakeman on that CSX train those two Amtrak employees would be alive today. His job would have been to stay by the switch until the move was finish and lock and line the switch for the main...
When I worked on the railroad and if we was parking the train it was normal to lock and line the switch after the engine cleared the switch.
The real baddie here is the railroads and their job elimination policies.
Just think the railroads want a one man crew and toying with the idea of using crewless trains..
Thank you to those who posted their insights here.
Does anybody know the unit number of the demolished CSX locomotve ?
GARRY
HEARTLAND DIVISION, CB&Q RR
EVERYWHERE LOST; WE HUSTLE OUR CABOOSE FOR YOU
BRAKIEthe engineer is not allowed to get out of the cab to line a switch due to a FRA regulation that came into being after 8888 got loose after the engineer left the cab to line a switch so,it back to the conductor.
This has a bearing on the one-man-crew issue here, as there is no way the move could have been shoved without RCO of the power, which inherently prevents a Crazy Eights sort of situation. While you're quite correct in noting that a head-end brakeman would have kept this situation from being a 'problem', I would argue that a much better solution than a full and expensive return to head-end brakemen everywhere and often would be a FRA rule that the equivalent of three-step protection with parking-brake application be required before the engineer leaves his cab for any reason (including lining and locking a switch only an eighth of a mile away). That actually solves the problem instead of giving bureaucrats an excuse to impose mandates addressing universally assumed incompetence, one of which has turned out in this case to have contributed to taking lives.
For the record, I have no objection to the idea that it was the 'conductor's' responsibility to make it back to the head end, after completing the shove, and then to reline the switch correctly. We will, I'm sure, find out the details of why that was not done by the end of the initial investigation. On the other hand, one of the elephants in the room here is how the dispatcher was given release. Someone had to say something that resulted in Amtrak being given authority to proceed between control points with that switch between them, and this is the real operating failure and cause of the accident as it developed, as far as I can see at this point.
A proper crewless Amtrak train would either have detected the open switch correctly and applied prompt braking or (in a variety of designs) determined it had not yet been correctly 'released' and stopped in controlled fashion. It would also very likely not have been affected by unanticipated hard lateral motion going through an unanticipated switch, if that turns out to be a factor in response here. On the other hand, I don't see a particularly good answer to how a 'crewless' train could have conducted the crossover and shove moves that the CSX train performed without a great deal of sensor equipment hung and configured with a great deal of, in the absence of trained techies, complex equipment -- so it is likely that it would have been at the very least a RCO shove with the single crew on the ground, watching the point via a pan-and-tilt camera on an enhanced EOT device from near the fouling point of the switch, then relocking the switch and boarding the power for the continuation of the shove (still watching via the EOT) to the finishing point. This wasn't particularly rocket science in the '80s, and is certainly not now.
Heartland Division CB&Q Does anybody know the unit number of the demolished CSX locomotive ?
Does anybody know the unit number of the demolished CSX locomotive ?
CSX 130 (AC4400CW).
BRAKIEFirst I agree but,the engineer is not allowed to get out of the cab to line a switch due to a FRA regulation that came into being after 8888 got loose after the engineer left the cab to line a switch so,it back to the conductor.
Its more of a work rule than an FRA regulation.
If it was a shoving move through that switch, then the engineer would have been operating from the set of power Amtrak struck.
At the time of the collision there was no CSXT engineer on the engine.
That means that at some time the engineer go off the engines and left them unattended. I would maintain that if the engineer is getting off the engine and not going to get back on the engine, there is ZERO safety risk with him lining the switch behind.
Saying he can't do something safety sensitive (especially when it results in injury and death) because he can't leave the engine unattended isn't a good excuse when he turns around and leaves the engine unattended anyway.
Dave H. Painted side goes up. My website : wnbranch.com
Some coverage I've read, (AP, IIRC) has verbiage or implies the Amtrak crew failing to properly react once they knew they went through the switch the wrong way. 7 seconds is not a lot of time and even if they reacted to what happened delayed ony by normal reaction time, it would make no effective difference in what happened -- trains don't stop that fast, even in emergency, at the speed they were traveling just before things went pear-shaped.
Positive Train Control (PTC) could very well have made a difference - maybe. There was some discussion I read that brings up a signal crew working on implementing PTC as involved wth the switch. Got no idea myself, too soon to say that was a factor. Presumably, PTC would have detected a mismatch between switch alignment, routing, permissions for track occupancy, and speed and might have stopped Amtrak before even approaching the switch.
If PTC would have reacted only once the Amtrak took the erroneous route because of a sensor failure or untested maintenance intervention, then would have made little to no difference, again too short time vs distance to matter.
Mike Lehman
Urbana, IL
OvermodOn the other hand, one of the elephants in the room here is how the dispatcher was given release. Someone had to say something that resulted in Amtrak being given authority to proceed between control points with that switch between them, and this is the real operating failure and cause of the accident as it developed, as far as I can see at this point.
And that finger points right back to the conductor unless that call wasn't made and the dispatcher thought the conductor had already informed him the switch was lined and locked for the main. This could easily happen since the DS may be in communication with several crews plus maintenance men.
I would not want to be in that CSX crew's position.
I'm all for a three man crew since that would be far safer then a RCO pack-I've seen some close calls and minor side swipes while railfaning because that man with the RCO pack could not see his engine or the other side of his train 20 plus cars away.
dehusmanSaying he can't do something safety sensitive (especially when it results in injury and death) because he can't leave the engine unattended isn't a good excuse when he turns around and leaves the engine unattended anyway.
The train was parked..The crew was gone.And according to my railroads that is a FRA rule.
If no one from the CSX train ever reported to the dispatcher that the switch was relined for the main, then the dispatcher should not have allowed any train on that track.
If someone from the CSX train reported the switch relined to the main, and knew it was not, that person lied. And should be criminally charged.
If someone from the CSX train reported the switch relined to the main because another crewman told him it had been, then the other crewman lied, and should be criminally charged.
Since the switch was NOT relined to the main, and the crew left the train, it means one or both knew they were leaving an unsafe situation behind them.
Perhaps they just "forgot". Is that a good defense?
Ed
dehusman BRAKIE First I agree but,the engineer is not allowed to get out of the cab to line a switch due to a FRA regulation that came into being after 8888 got loose after the engineer left the cab to line a switch so,it back to the conductor. Its more of a work rule than an FRA regulation. If it was a shoving move through that switch, then the engineer would have been operating from the set of power Amtrak struck. At the time of the collision there was no CSXT engineer on the engine. That means that at some time the engineer go off the engines and left them unattended. I would maintain that if the engineer is getting off the engine and not going to get back on the engine, there is ZERO safety risk with him lining the switch behind. Saying he can't do something safety sensitive (especially when it results in injury and death) because he can't leave the engine unattended isn't a good excuse when he turns around and leaves the engine unattended anyway.
BRAKIE First I agree but,the engineer is not allowed to get out of the cab to line a switch due to a FRA regulation that came into being after 8888 got loose after the engineer left the cab to line a switch so,it back to the conductor.
On the road Dave retired from and the way local managers in my area handle an engineer leaving the cab, engineers can only get about one engine length away from the engines when unattended and the train is unsecured. Once tied down, then the engineer can go farther away. To realign a switch for example.
Jeff
mlehman Positive Train Control (PTC) could very well have made a difference - maybe. There was some discussion I read that brings up a signal crew working on implementing PTC as involved wth the switch. Got no idea myself, too soon to say that was a factor. Presumably, PTC would have detected a mismatch between switch alignment, routing, permissions for track occupancy, and speed and might have stopped Amtrak before even approaching the switch. If PTC would have reacted only once the Amtrak took the erroneous route because of a sensor failure or untested maintenance intervention, then would have made little to no difference, again too short time vs distance to matter.
If the signal system would've been operative, it would've detected an occupancy. This from the switch not lined correctly in this case. The signals approaching the switch would be at their most restrictive. This whether or not PTC was in use.
If PTC was operative, it would've "seen" the signal system occupancy and would allow a train or engine to enter the block (assuming the entering signals aren't absolutes) only at restricted speed. PTC "knows" the switch is there but not how it's lined. Upon getting close, it would prompt the engineer to verify the switch position by using a button key on the PTC display. It does this even if the switch is lined correctly. Failure to verify, the train wpuld be stopped by a PTC initiated brake application.
BRAKIEI'm all for a three man crew since that would be far safer then a RCO pack-I've seen some close calls and minor side swipes while railfaning because that man with the RCO pack could not see his engine or the other side of his train 20 plus cars away.
Unless the person running the remote is on stilts, there's no way they can see as well as being aboard an engine. In my mind, this and other issues suggest that remote control is something that should be sharply limited to locations and duties where it is safest to use it. Yet the RRs are, I think, looking to the technology to achieve a major reduction in crew demand and they seem to intend to use it widely to do so.
jeffhergertIf PTC was operative, it would've "seen" the signal system occupancy and would allow a train or engine to enter the block (assuming the entering signals aren't absolutes) only at restricted speed. PTC "knows" the switch is there but not how it's lined. Upon getting close, it would prompt the engineer to verify the switch position by using a button key on the PTC display. It does this even if the switch is lined correctly. Failure to verify, the train wpuld be stopped by a PTC initiated brake application.
Jeff,
Thanks that clarifies some things. So, does PTC automatically slow the train enough so that points can be seen easily, because track speed, curves, and other obstructions are going to limit the distance at which their position can be verified? That would seem to be OK temporarily, but I would think that such a switch would soon acquire the electronic necessities to avoid being a bottleneck. Or is such a seemingly simple thing as point position still a technical challenge more than a matter of funding?
One of the greatest unfounded-mandate concerns and complaints since 2008 is that PTC demands all manual switches have positive electric confirmation of point position. That requires in practice somewhat more than approximate position of the linkage. To my knowledge no one has suggested programming PTC to chronically approach manual switches within restricted-speed rules for visual point-position observation, although it seems very logical to develop that capability for times just such as the situation at Cayce, where some of the PTC safety systems would be operating but others, specifically the ability to read or confirm switch lining remotely, were not.
Research into even semiautonomous drones provides sufficient answer in almost all cases for offside visibility during a shove; it's long been established that a good camera head and the right machine-vision algorithms in the background are at least as safe as some poor schmo riding the point of a shove without a shoving platform. There is more to effective single-crew operation than handing a guy a RCO pack and telling him to switch cuts of nondegassed Bakken crude for Irving Oil.
The PTC display gives a look ahead of six miles. There is a Distance to Next Target line on the display. This will also display the allowed speed at the next target or 0 mph if a stop is required. There is also a Warning Distance and Stopping distance displayed. The WD and SD also appear as moving lines on the display. This link is to BNSF info, but all the major carriers are using the same PTC system. Amtrak's on it's own lines is different, but Amtrak will have to use this system when operating over the freight RRs track.
http://www.smartlocal202.org/site/assets/files/BNSF/BN_ROAD/PTC_Screen_Elements_9_27_2012.pdf
PTC can only stop a train if it calculates that a train will overrun it's next target speed. That's where the warning distance and stopping distance come into play. For a switch in signalled territory, it doesn't know the position of the points. It will allow entering the block at restricted speed but won't let you by the switch location until the engineer has verified the switch position on the screen. If the switch is lined wrong, the engineer could still verify it as correct and run over it onto a side track, but it would be at restricted speed. Still, people have died in accidents that happened at restricted speed.
PTC can't and won't eliminate all collisions, it's still possible to have slow speed accidents. Even with everything in working order.
mlehmanUnless the person running the remote is on stilts, there's no way they can see as well as being aboard an engine. In my mind, this and other issues suggest that remote control is something that should be sharply limited to locations and duties where it is safest to use it. Yet the RRs are, I think, looking to the technology to achieve a major reduction in crew demand and they seem to intend to use it widely to do so.
Mike,Railroads wants to start using a one man crew on road freights and some local freights uses one man with RCO and railroads is toying with the idea of crewless trains..The scary thought is the technology is there for crewless trains. See January 2018 Trains magazine.
So far the Feds has fought for two men crews sighting safety issues.
jeffhergertFor a switch in signalled territory, it doesn't know the position of the points.
I thought the PTC mandate required to stop a train before it can operate over an improperly lined main track switch?
Your quote shows something different or do I misunderstand something? It seems uneconomical to approach all switches at restricted speed.
It shouldn't be impossible to include the switch orientation into PTC. In other countries it is mandatory.Regards, Volker
VOLKER LANDWEHR jeffhergert For a switch in signalled territory, it doesn't know the position of the points. I thought the PTC mandate required to stop a train before it can operate over an improperly lined main track switch? Your quote shows something different or do I misunderstand something? It seems uneconomical to approach all switches at restricted speed. It shouldn't be impossible to include the switch orientation into PTC. In other countries it is mandatory.Regards, Volker
jeffhergert For a switch in signalled territory, it doesn't know the position of the points.
Since the reversed switch shunts the signal system causing a restrictive indication, you're going to be go through the entire signal block at restricted speed. Even if you stop and line back a switch to it's normal position. Just in case there is something else that is causing the signals to be restrictive. If you enter a block under restrictive conditions and there is a hand throw switch within it, PTC will ask for verification of how such switch is lined when approaching it. Even if it is lined correctly in the normal position.
I imagine in non-signalled territory each individual switch is incorporated into PTC. That's because there is no signal system to otherwise detect an open switch. With signals, PTC is an overlay of the existing signal system, which already detects an open switch. I'm guessing the reason in signalled territory they didn't go for tying in every switch was cost since the signals are already tied into PTC.
In the future if they can ever work out "rolling blocks" and remove the signal system as we know it, then they would have to tie in every switch.
One thing to remember about PTC. It isn't foolproof and low speed collisions can still happen in some circumstances.
Looking at what I saw and heard of this crash i am a bit confused. I thought that as a general rule switches need to be aligned on the main route unless an actual move is occurring in and out of the siding. Thus if the switch was in the incorrect position that would be CSX’s fault since it was not returned to the main after the freight train was parked.
I operate thw SIW this way but am I wrong?
Joe Staten Island West
joe323I thought that as a general rule switches need to be aligned on the main route unless an actual move is occurring in and out of the siding.
Joe,That is correct and the only exception is the normal position for that switch is in the reverse position and that would be rare.
jeffhergertSince the reversed switch shunts the signal system causing a restrictive indication, you're going to be go through the entire signal block at restricted speed.
I hope I understand correctly. If the signal system detects the switch in normal position there is no restrictive signal indication.
jeffhergertIn the future if they can ever work out "rolling blocks" and remove the signal system as we know it, then they would have to tie in every switch.
I fear therefore someone has to re-design PTC to a stand-alone system.
That was the critizism by a number of independent industry experts that railroads had given away the rolling block advantage by letting the signal department develop PTC. The advantage wasn't unknown as the FRA had it in their cost-benefit analysis.Regards, Volker
VOLKER LANDWEHRI hope I understand correctly. If the signal system detects the switch in normal position there is no restrictive signal indication.
In block signal territory, every main track switch has an indicator that tells the signal system whether the switch is open (whether it actually indicates position or just indicates open, I don't know) and causes a restrictive signal in the signal system. All that is there, it just has to be fed to the PTC system.
The problem isn't with PTC its with the wayside signals and the fixed signal system, that's where all the hardware is. All those little cabins along the railroad filled with relays or electronics. They actually detect occupancy and tell the dispatch and PTC systems about the condition of the tracks, then the PTC system, based on their inputs, the train operation and parameters communicate with the train. PTC really doesn't care what's giving it the inputs, but the signal cabins that are hard wired to the track are pretty rigid.