Union Pacific collision between two trains in Hoxie, Arkansas.

[NP Eddie]

ALL:

I saw a report of a UP collision of two trains in Hoxie, Arkansas. In checking with an old OG, I found this is former Missouri Pacific trackage. Two Little Rock TV stations had aerial pictures of the wreck. It looks like one CSX and at least one UP locomotive were destroyed.

Can anyone shed more light on this wreck? How near to the former SLSF diamond was this?  Was this CTC or single track ABC/TWC?

I will check Monday if the NTSB will investigate the wreck.

 

Ed Burns

Retired NP-BN-BNSF from Minneapolis, MN

[CShaveRR]

UP's line through here is CTC.

[MidlandMike]

It was somewhat south of town and the the BNSF diamond.  This is also on the Texas Eagle Amtrak route.

http://www.arktimes.com/ArkansasBlog/archives/2014/08/17/train-derailment-in-hoxie-kills-2-homes-evacuated

[oezbob]

It is hard to tell from aerial photos -- cars flopped all around and tracks misaligned -- so, I wonder if this section was double or single track? It appears that two parallel tracks leading to a switch were damaged, but I cannot tell if the switch was for a passing siding or if it is part of a crossover, which would make a big difference. And, I am not sure there was a switch at all as I could not see a frog or points ... just two tracks merged together, perhaps from the impact.

Also, I could make out only two locos in the aerial pix -- one, a UP on its side with nasty cab damage, and another that appears vertical, probably on rails and burned gray (perhaps that is the CSX motor mentioned, but I could not see that ID). No other locos are in sight on those 5 pix, but I am sure they are there somewhere. Perhaps they caromed onward out of view.

Bad news! ... Bob

[samfp1943]

MidlandMike

It was somewhat south of town and the the BNSF diamond.  This is also on the Texas Eagle Amtrak route.

http://www.arktimes.com/ArkansasBlog/archives/2014/08/17/train-derailment-in-hoxie-kills-2-homes-evacuated

 Hoxie,Arkansas is an at grade junction. The lines lay for the UPRR on a roughly slanted course acroos this junction. The Union Pacific is the former MoPac line between Walnut Ridge,Ark and Little Rock,Ar. on a cant from Southwest to North-North Easterly direction ( US Hwy 67 is parallel to it through Hoxie-on Business US 67).  The UPRR line is double tracked with sidings, as I recall right through the Hoxi area it has some triple tracking(?)

The current BNSF ( Thayer Sub) is on a sort of North-North Westerly line and goes in a southeastery line towards Jonesboro,Ark and Memphis)   The BNSF is a single tracked line with some passing sidings scattered along its route.  

 I believe both lines are CTC controlled.  The UPR line is virtually table-top flat through Hoxie. The BNSF might have some line of site issues as it runs through the small community and an area of houses and buildings. BOth lines have much activity.  It used to be hard to travel through that area in an automobile. USHwy 63 is a major route from the Memphis area into the Ozarks and on to Springfield,Mo. and US 67 was a major route from the St Louis area and on towards Little Rock.  Many of those problems were aleviated in the erly part of 2000 by building an 'Interstate quality' by-pass for US Hwy 63 and removing the T- Junction at Hwy 67, and construction of Ark Hwy 290 towards Walnut Ridge,Ar.  

The accident took place within the City limits of Hoxie according to the linked article which probably put it sout of the By-pass bridge on the South edge of town. That is the bridge that would carry USHwy 63 over Business 67 and the UPRR.  An area of fairly straight tacks(?).

I'd leave cause as to the experts to decide.  

Area of HOxie Ark from Google Map: @ https://www.google.com/maps/@36.0521076,-90.9744953,1661m/data=!3m1!1e3

 

[CShaveRR]

The collision took place at or near a spot where two CTC main tracks converged (becoming a single track to the southwest).  That, right there, gives you an important truth about the wreck:  it was caused by the southwest-bound train exceeding the limits of its authority, for whatever reason (I have ideas about that, but those would be speculation).  

Sam, do you know where the crew-change point for UP north (northeast) of Hoxie would be?  I'm assuming that (North) Little Rock is the southern end of the trip. 

As for the Eagle, reports I've seen suggest that it was behind the southwest-bound train, and had to be brought back to Poplar Bluff to detour around the wreck. 

[dehusman]

The crew district is Dexter/Poplar Bluff to N Little Rock.  There is no regular intermediate crew change.

[CShaveRR]

No intermediate crew changes, but it now appears that the southbound crew boarded the train at East St. Louis. That's quite a long, and I daresay tiring, trip. 

[dehusman]

CShaveRR

No intermediate crew changes, but it now appears that the southbound crew boarded the train at East St. Louis. That's quite a long, and I daresay tiring, trip. 

 

You have received bad information.

 

The train originated at E St. Louis, the crews changed at Dexter.

[CShaveRR]

Okay, thanks.  Both trains were still well into their trips by then.

[narig01]

To view the location on Google maps type in the following location.:

SE Front St & SW Texas St, Hoxie, Ar

The wreck is at the south end if the siding. About a mile south of the US-63 overpass.

It looks like the northbound Texas Eagle went thru this morning. The southbound went down the other way.

Rgds IGN

[narig01]

One other comment question.
Is UP running at the limits again? The last few weeks I've several intermodal trains with only one unit on the train. Down my way in SW Arkansas it's flat so not a lot of power is needed. Still I have to wonder.
Thx IGN

[samfp1943]

narig01

To view the location on Google maps type in the following location.:

SE Front St & SW Texas St, Hoxie, Ar

The wreck is at the south end if the siding. About a mile south of the US-63 overpass.

It looks like the northbound Texas Eagle went thru this morning. The southbound went down the other way.

Rgds IGN

Seems that TRAINSNewswire of this date August 20,2014 is carying the following headline:  FTA:

NTSB: Southbound train missed stop signal in Arkansas collision

Published: August 20, 2014

Investigator Mike Flanigon is seen with the event recorder from the lead locomotive of the southbound train involved in the Hoxie collision. Photo by National Transportation Safety Board JONESBORO, Ark. – In their continuing investigation into a Sunday collision and derailment in northern Arkansas that killed two crewmembers, federal investigators now say the southbound train involved missed multiple signals, including a final stop signal, WMC-TV reports. Investigators are now collecting documentation and taking measurements at the scene...." (WMCTV is a Memphis Television Station)

Seems like we are turning over all sorts of investigations to the Media to be the source for information that would normally be released after the data had been studied and filed by the appropriate entity... Just my 

[NorthWest]

Looks like the NTSB is about to release a report on this collision:

http://www.msn.com/en-us/news/us/automatic-horn-may-be-behind-fatal-arkansas-train-crash/ar-AAlbKXM?li=BBnbfcL&ocid=UE13DHP

I can't even wrap my head around what this article is trying to say. What is an 'automatic horn'? Another PTC ad?

[zugmann]

NorthWest

I can't even wrap my head around what this article is trying to say. What is an 'automatic horn'? Another PTC ad?

Sounds like they are very confused on the function of an alerter.

[BaltACD]

zugmann

 

NorthWest

I can't even wrap my head around what this article is trying to say. What is an 'automatic horn'? Another PTC ad? 

Sounds like they are very confused on the function of an alerter.

 

Kelly P. Kissel has absolutely no knowledge about what he is trying to write about.  He is trying to write about the operation of an Alerter, but knows nothing about it's operation in the real world.

[Euclid]

The "automatic horn" is the automatic horn sequencer that blows grade crossing horn signals.  Apparently it has been miswired in conjuctiontion with the alerter whereby the grade crossing horn signal is taken as a human input that satisfies tdhe alerter to not sound for that one time.  That is only my approximate understdanding.  More information may be linked to the following which is the explanation given by the NTSB:

NTSB Urgent Rail Recommendations: 'Alerters' Watching for Human Engagement Should Not Be Misled by Automated Functions

2/6/2015

Page Content ​ WASHINGTON - The National Transportation Safety Board has issued urgent safety recommendations to the Federal Railroad Administration, the Association of American Railroads, the American Short Line and Regional Railroad Association, and the American Public Transportation Association to help ensure that electronic alertness devices or “alerters” work as intended on trains. The alerter helps crew members maintain vigilance in the locomotive cab by monitoring locomotive engineer activity.   If it has been too long since the locomotive engineer performed an input or action to reset the alerter, the system issues visual and audible alerts, and applies train brakes if there are still no inputs from the crew.  However, the agency found, an alerter’s reckoning of “idle time” can be reset to zero by inputs that do not necessarily demonstrate a crew member’s continuing engagement.   “The alerter is an automated system to make sure the human is engaged and, if necessary, to take action,” said NTSB Acting Chairman Christopher A. Hart.  “We found that the alerters were acting from automated events as if they had been human inputs.” The NTSB discovered this safety issue and issued urgent recommendations as a result of the ongoing investigation of the 2014 collision of two Union Pacific freight trains, one southbound and one northbound, in Hoxie, Arkansas.  The accident resulted in the deaths of two crewmembers, the derailment of 55 cars, a release of diesel fuel, a fire, and the evacuation of about 500 nearby residents. The southbound train was equipped with a horn sequencer. The horn sequencer is activated with a single push of the sequencer foot pedal under the engineer’s control console. Once activated it continually sounds the cadence for approaching a highway-rail grade crossing – a series of two long sounds, one short sound, and another long sound -- until the sequencer foot pedal is again depressed. The NTSB’s examination of the southbound train’s event recorder noted that the horn sequencer reset the electronic alertness device each time the horn blew, as if the locomotive engineer were commanding each sound manually. This prevented the device from providing an alarm to the train crew or activating the brakes.  While the NTSB has not yet reported the accident’s probable cause, the agency issued these urgent recommendations to address the safety issue that came to light during the investigation. “The Union Pacific railroad has moved to fix this problem. The FRA needs to require that other railroads understand the problem and fix it where it is necessary,” said Hart.

 

 

 

[Euclid]

Here is the detailed explanation:

http://www.ntsb.gov/safety/safety-recs/recletters/R-15-006.pdf

 

 

[RME]

Oh.  My.  God.

The law of unintended consequences combines with the "automotive engineer's" love of cheap "equivalent" functionality.

What's the right way for an alerter's processor to accept an "input" that the crew is moving locomotive controls -- here, specifically, the locomotive horn?  That's pretty obvious: put a switch of some kind on the horn valve, or tap into the input from the physical button or cord in the cab, that registers that someone has actually done something in the cab - a human moving an actual control.

Automotive engineer:  Let's make it wireless!  We'll just listen for the horn and let the system understand the crew has to be blowing it.  Why else would you be hearing a locomotive horn in the cab?

Government:  We want quiet zones, but they don't seem to work at reducing accidents as much as we hoped, so we'll install directional horns that are only loud for people approaching the crossings.  (A good idea.)  Problem: they are also loud for vehicles approaching the crossings at right angles to the traffic lanes...

Now, remember the point Randy Stahl just made about alerters at Megantic: they are NOT meant to be depended on as safety devices -- they are backstopping crew alertness in specific respects.  And are annoying in some of the ways they do so.  Here, in an effort to make them less annoying, poor engineering choices and poor operating choices appear to have combined to open the door to disaster.

Next, I suspect, we will find out that microphone-based alerter systems aren't designed to allow hardwiring of additional switches for the "microphone" functions without a lot of expensive modifications on a unit-by-unit basis.  Plus a bunch of programming that has to overwrite object code in PLCs, probably in some subset of an arcane language or toolkit.  Next we'll learn that the alerters don't have flashable nonvolatile ROM.

The take-home message, though, is exactly as Hart said.  I'm hoping that as part of the 'understanding' the railroads look at any other assumptions that have been coded into the alerter programming, and perhaps proactively 'fix' them too.