Amtrak 91 - CSX Collision Cayce, SC - 2/4/18

I realize there may not be as such is a personnel matter and accordingly sealed from public disclosure, but is there any confirmation that the CSX T&E crew and the Train Dispatcher, and for that matter the surviving Amtrak Assistant Conductor, been Held From Service?

Regarding PTC and switches:

(a)PTC system requirements. Each PTC system required to be installed under this subpart shall:

(1) Reliably and functionally prevent:

(iv) The movement of a train through a main line switch in the improper position as further described in paragraph (e) of this section.

(e)Switch position. The following requirements apply with respect to determining proper switch position under this section. When a main line switch position is unknown or improperly aligned for a train's route in advance of the train's movement, the PTC system will provide warning of the condition associated with the following enforcement:

(1) A PTC system shall enforce restricted speed over any switch:

(i) Where train movements are made with the benefit of the indications of a wayside or cab signal system or other similar appliance, method, device, or system of equivalent safety proposed to FRA and approved by the Associate Administrator in accordance with this part; and

(ii) Where wayside or cab signal system or other similar appliance, method, device, or system of equivalent safety, requires the train to be operated at restricted speed.

(2) A PTC system shall enforce a positive stop short of any main line switch, and any switch on a siding where the allowable speed is in excess of 20 miles per hour, if movement of the train over the switch:

(i) Is made without the benefit of the indications of a wayside or cab signal system or other similar appliance, method, device, or system of equivalent safety proposed to FRA and approved by the Associate Administrator in accordance with this part; or

(ii) Would create an unacceptable risk. Unacceptable risk includes conditions when traversing the switch, even at low speeds, could result in direct conflict with the movement of another train (including a hand-operated crossover between main tracks, a hand-operated crossover between a main track and an adjoining siding or auxiliary track, or a hand-operated switch providing access to another subdivision or branch line, etc.).

(3) A PTC system required by this subpart shall be designed, installed, and maintained to perform the switch position detection and enforcement described in paragraphs (e)(1) and (e)(2) of this section, except as provided for and justified in the applicable, FRA approved PTCDP or PTCSP.

(4) The control circuit or electronic equivalent for all movement authorities over any switches, movable-point frogs, or derails shall be selected through circuit controller or functionally equivalent device operated directly by the switch points, derail, or by switch locking mechanism, or through relay or electronic device controlled by such circuit controller or functionally equivalent device, for each switch, movable-point frog, or derail in the route governed. Circuits or electronic equivalent shall be arranged so that any movement authorities less restrictive than those prescribed in paragraphs (e)(1) and (e)(2) of this section can only be provided when each switch, movable-point frog, or derail in the route governed is in proper position, and shall be in accordance with subparts A through G of this part, unless it is otherwise provided in a PTCSP approved under this subpart.

litz wrote:

justalurker66 wrote: Apparently a switch left locked lining a train into an occupied siding?
A signal suspension in place that removed any alert from the signalling system (restricting signals into that block) that would have warned the engineer?

One rule that was apparently not broken was the speed (unless MAS was slower) ---
300.5 The following speeds must not be exceeded:
b. 59 MPH for passenger trains operating within the limits of a signal suspension or against the current of traffic

(And as a supporter of PTC I'll agree that PTC would not have helped if it also would be suspended along with the signal system.)

So here's a question ... is PTC implemented on manually thrown switches?

In other words, if the switch isn't under CTC control, is PTC aware of it (and its position)?

Read up on I-ETMS and you will find your answers and some

https://www.wabtec.com/products/1454/i-etms" onclick="window.open(this.href);return false;™

ACSES which is considered a form of PTC would have not prevented this crash.

litz wrote:

justalurker66 wrote: Apparently a switch left locked lining a train into an occupied siding?
A signal suspension in place that removed any alert from the signalling system (restricting signals into that block) that would have warned the engineer?

One rule that was apparently not broken was the speed (unless MAS was slower) ---
300.5 The following speeds must not be exceeded:
b. 59 MPH for passenger trains operating within the limits of a signal suspension or against the current of traffic

(And as a supporter of PTC I'll agree that PTC would not have helped if it also would be suspended along with the signal system.)

So here's a question ... is PTC implemented on manually thrown switches?

In other words, if the switch isn't under CTC control, is PTC aware of it (and its position)?

The short answer is Yes. Now here's the full explanation.

Without PTC but in signal territory, every switch on the mainline is tied into the signal system. A switch lets the system know, at the simplest level, that it is either closed or open in relation to the mainline. There are contacts that made in that case, and broken when the switch is not lined for the main (meaning lined for the siding or not lined at all). All this shows up as is a track occupancy to the train dispatcher - commonly called a track circuit or track light. The dispatcher has no idea what it is if it happens randomly, and has to send a field employee (signal maintainer) to investigate. It could be the switch, a broken rail, something shorting the track, whatever. What it does, however, is not allow the signal system to line a permissive signal better than "Restricting" toward the "open" switch. So an engineer would either need permission by a red signal and operate at Restricted Speed prepared to stop short of a misaligned switch.

So a signal system protects against an open switch, or in this case a switch lined for a siding. What it doesn't protect against is an engineer who misses the signals somehow and still manages to operate at track speed into the switch - a "perfect storm" scenario of multiple failures. THAT is what PTC will prevent. It will force the train to stop at the prior interlocking signal, and it will enforce the maximum end of Restricted Speed (either 15 or 20mph depending on the railroad) once the train is authorized past that signal through a Stop Override. The train can still make impact at that speed, unfortunately that hasn't been solved yet, but the physics involved at 15 or 20mph are far different than at 50mph. It also affords an engineer more time to react.

And yes, ACSES II would have prevented this crash.

So in this territory, Amtrak follows the csx rulebook? And the csx rules do not require approaching switches prepared to stop when signals are inoperative (dcs)?

Presuming that is the case, I do not see how this anything but a 100% CSX screw up.

Trinnau wrote:

litz wrote:

justalurker66 wrote: Apparently a switch left locked lining a train into an occupied siding?
A signal suspension in place that removed any alert from the signalling system (restricting signals into that block) that would have warned the engineer?

One rule that was apparently not broken was the speed (unless MAS was slower) ---
300.5 The following speeds must not be exceeded:
b. 59 MPH for passenger trains operating within the limits of a signal suspension or against the current of traffic

(And as a supporter of PTC I'll agree that PTC would not have helped if it also would be suspended along with the signal system.)

So here's a question ... is PTC implemented on manually thrown switches?

In other words, if the switch isn't under CTC control, is PTC aware of it (and its position)?

The short answer is Yes. Now here's the full explanation.

Without PTC but in signal territory, every switch on the mainline is tied into the signal system. A switch lets the system know, at the simplest level, that it is either closed or open in relation to the mainline. There are contacts that made in that case, and broken when the switch is not lined for the main (meaning lined for the siding or not lined at all). All this shows up as is a track occupancy to the train dispatcher - commonly called a track circuit or track light. The dispatcher has no idea what it is if it happens randomly, and has to send a field employee (signal maintainer) to investigate. It could be the switch, a broken rail, something shorting the track, whatever. What it does, however, is not allow the signal system to line a permissive signal better than "Restricting" toward the "open" switch. So an engineer would either need permission by a red signal and operate at Restricted Speed prepared to stop short of a misaligned switch.

So a signal system protects against an open switch, or in this case a switch lined for a siding. What it doesn't protect against is an engineer who misses the signals somehow and still manages to operate at track speed into the switch - a "perfect storm" scenario of multiple failures. THAT is what PTC will prevent. It will force the train to stop at the prior interlocking signal, and it will enforce the maximum end of Restricted Speed (either 15 or 20mph depending on the railroad) once the train is authorized past that signal through a Stop Override. The train can still make impact at that speed, unfortunately that hasn't been solved yet, but the physics involved at 15 or 20mph are far different than at 50mph. It also affords an engineer more time to react.

And yes, ACSES II would have prevented this crash.

If the signal system is suspended and ACSES remains in service, how would it have prevented this crash? ACSES would only enforce civil speeds, temporary speeds, interlocking route speeds and PTS at interlockings.

8th Notch wrote:If the signal system is suspended and ACSES remains in service, how would it have prevented this crash? ACSES would only enforce civil speeds, temporary speeds, interlocking route speeds and PTS at interlockings.

Basically, even though the signals are suspended, you have to suspend ACSES in order for it to not enforce the infrastructure. So the signals to the North would have been enforced if ACSES was left in service, and without input from the signal system is designed to "fail safe" and enforce a stop that has to be overridden.

EDIT: As a follow-up, the new NORAC book only allows 40mph max under ACSES and ABS/CSS both OOS (non-signaled DCS territory). 585(A)(3) and 588.

MCL1981 wrote:So in this territory, Amtrak follows the csx rulebook? And the csx rules do not require approaching switches prepared to stop when signals are inoperative (dcs)?

Presuming that is the case, I do not see how this anything but a 100% CSX screw up.

Correct on both your questions, but there are still 3 possible scenarios based on the information we have now - 2 on CSX and 1 on Amtrak. I'm not CSX rules qualified, but on NORAC these scenarios would exist.

1 - CSX crew reported clear but failed to line the switch
CSX crew reported clear and did not line the switch but notified the CSX dispatcher of a switch in reverse position (might be the case if the crew was close to being on the law and did not have time to normal the switch). This could result in either:
2 - The CSX dispatcher failed to give the switch reverse to the Amtrak crew
3 - The CSX dispatcher gave the switch reverse to the Amtrak crew and they failed to operate accordingly.

We will, hopefully, know more soon. The CSX dispatcher and crew were interviewed today and the surviving Amtrak crew will be interviewed tomorrow. The NTSB will release pertinent details revealed by the interviews once all interviews are complete.

Trinnau wrote:

8th Notch wrote:If the signal system is suspended and ACSES remains in service, how would it have prevented this crash? ACSES would only enforce civil speeds, temporary speeds, interlocking route speeds and PTS at interlockings.

Basically, even though the signals are suspended, you have to suspend ACSES in order for it to not enforce the infrastructure. So the signals to the North would have been enforced if ACSES was left in service, and without input from the signal system is designed to "fail safe" and enforce a stop that has to be overridden.

EDIT: As a follow-up, the new NORAC book only allows 40mph max under ACSES and ABS/CSS both OOS (non-signaled DCS territory). 585(A)(3) and 588.

ACSES would only enforce a positive stop at an interlocking, if this occurred near ABS signals which were OOS then ACSES would have done nothing! Do we have any clarification on the territory? Do we know if this hand throw switch was even protected by an interlocking or just ABS....

You know, this makes me wonder... How often does this mistake happen and the signal system catches it so we never hear about it?

8th Notch wrote:ACSES would only enforce a positive stop at an interlocking, if this occurred near ABS signals which were OOS then ACSES would have done nothing! Do we have any clarification on the territory? Do we know if this hand throw switch was even protected by an interlocking or just ABS....

The switch was hand-throw, NTSB briefings have said as much. So it was protected by ABS and not interlocked. What I'm saying is that ACSES will not let you get past the nearest interlocking without pushing the stop override button, and that only releases you to Restricted Speed. You won't be released any faster than that until you get a good ACSES read along with a good wayside signal.

A look at Google shows 2 tracks to 1 about 2000ft north/east of the hand-throw switch with signals in such a location to indicate this is an interlocking (under the I-26 bridge). ACSES would have enforced a south/west bound train to a stop here if only the signals were OOS (ACSES fails safe in this case because it is not reading a permissive indication from the signal system), and the only way by the signals is to push the button and go no more than 15mph. They would proceed at that speed until either passing a permissive indication or stopping at the next interlocking and repeating the process. The only way to go faster than 15mph is to disable the ACSES system in addition to the signal system.

IT should be interesting reading of what method has used for this signal outage,what orders was given to the CSX freight crew and the Amtrak crew from
the CSX dispatcher.
It will be awhile before that material will be part of the docket for this wreck.

Okay, just to condense what we do know:

• CSX had the switch and signals down. We don't know if it was on the EC-1 or if a Form D was issued.
• CSX crew had locked the switch to go to the siding track, and gave control back to the dispatcher. We don't know if the dispatcher was informed or not.
• Amtrak 91 went in as normal not expecting any issues beyond the known (probably on the EC-1/Form D), max 59 MPH...
• CRASH!

Ugh. Yeah, this is more likely looking like CSX's fault. I'd love to see those EC-1/Form D papers. The only thing that would make it Amtrak's fault if the papers say "STOP AT THIS SWITCH, UNLOCK IT, AND HAND-THROW IT." The dispatcher *may* have some more info.

Trinnau wrote:

8th Notch wrote:If the signal system is suspended and ACSES remains in service, how would it have prevented this crash? ACSES would only enforce civil speeds, temporary speeds, interlocking route speeds and PTS at interlockings.

Basically, even though the signals are suspended, you have to suspend ACSES in order for it to not enforce the infrastructure. So the signals to the North would have been enforced if ACSES was left in service, and without input from the signal system is designed to "fail safe" and enforce a stop that has to be overridden.

EDIT: As a follow-up, the new NORAC book only allows 40mph max under ACSES and ABS/CSS both OOS (non-signaled DCS territory). 585(A)(3) and 588.

You are looking under the PTC section in the new Norac, special instructions supersede Norac which is what Amtrak has in place for ACSES rules in the ETT. ACSES rules are not in NORAC and when we have ACSES questions we refer to Amtrak’s special instructions (all which are irrevelant to this accident.)