On the scanner, one can hear the talking detectors giving clearance or instructions to inspect the train.
The CSX detector nearby gives considerable information, such as:
Track number, direction of movement, train length, speed, and number of axles.
How does the detector compute the train length and speed of the train? Is the speed the average speed or the speed at the end of the transmission? If a train were accelerating or decellerating, would this impact the length of train and speed info?
Now that I don’t know. But it was funny to tell the new hires that there was someone in the box that calculated all the information. One guy even tried to talk back to the detector and even wanted to toss him a bottle of water as we went by.
It wouldn’t surprise me if the same speed for the entire length is assumed when calculating train length, but it’d conceivable be possible to calculate an entry speed or a departure speed just be using the distance between the first two (or final two) axles. Or, more elaborately, have the detector tied in with an AEI scanner that gets car lengths directly from UMLER.
Now, as to how it’s really done, I haven’t the foggiest.
As for temperature requirements, there are certain threshholds at which heat restrictions or cold restrictions cut in.
The temperature is important because is some spots, if the ambient temperature is above a set point for a given time, slow orders kick in automatically due to possible sun kinks.
This is especially true in an area where the rail had to be replaced in a particularly colder temperature, due to wreck repair and other such unusual events…
You can find temperature restriction in a lot of timetables…down here, over 100 degrees F for an hour or more restricts movement over two of our drawbridges, due to joint expansion…you have to stop and physically inspect the bridge before proceeding.
The detectors measure the average speed of the entire train.
They use variations of the physics formula D=vt. If it were me, I’d use the assumed average distance between axles on a freight car truck and measure the time between the rapid axle inspections as a train passes to calculate average speed. Six-axle power will mess this up, but on a long train, the larger number of 2-axle trucks will prevail when calculating an average. Then using that average speed, I’d take the time from the first axle to the last and calculate the train length.
Maybe someone who works on them can explain it better/in more detail.
The typical system used today uses two magnetic transducers bolted to the web of the rail to provide timing signals for train presence, speed, length, and direction. The tranducers synchronize bearing and wheel scan timing algorithms that capture the heat signals from the bearing and wheel scanners and match them to the correct wheelset. The timing signals break out cars and locomotives into individual units within the train consist. This information is transmitted by radio both to the train and to a central office server. Software is available to match the signals to the car numbers using the trainlist, in order to build a service history for each locomotive, car, wheelset, and bearing. Integrating that software and culling out bad data from improper trainlists is the Achilles’ heel.
Average speed is used as Mr. Blysard points out. You could set the software to pick up last speed of the last car, but there’s no value in knowing that.
Dragging-equipment detectors are simpler as they do not need to know which car or axle generated the signal. An island circuit or audio-frequency overlay is used to turn the detector on and off. The track circuit can detect speed and direction of the train by the change in frequency.
LMAO…Or more new Hire fun includes when deadheading back to your terminal, to tell them to promptly report detectors not giving you an entrance and exit message to the dispatcher when going by one in the van[:D]. 15 points if you tell them to get out and roll your van by…[8D].
Thats a first for me,I never heard UP or BNSF detectors give train length or direction, only wheel count , temp , mph and milepost…sometime its a female voice but mostly a male voice…
I’m used to hearing (in a suitably mechanical voice) “Syracuse, New York, Track One, No Defects, Train Speed four zero, total axle count four zero eight…” or something very similar. I don’t remember ever catching an actual defect announcement.
Two transducers just a few feet apart can easily be used to compute train speed. After that it’s just a matter of how it’s applied. Defects are reported by way of number of axles from the rear of the train, as I recall. Given articulated cars and the occasional three axle truck, it’s the easiest way to measure.
Before radar, cops would sometimes use a device with two rubber hoses set about a foot apart to clock speeds.
Anyone who has rode Amtrak’s Empire Builder betweem Milwaukee and Minneapolis on a hot summer day knows that rule all too well. Normally, it’s a pretty quick run for such distance but I took it once when it was only in the high-80s (not at all abnormal for a July day in Wisconsin) and the speed reductions added at least an hour to our trip.
I’ve seen what the wild temperature flux does to track here in Milwaukee though, there are old sections of jointed rail on many of the less-used runs. In the Harborview district for example there’s a short run of track that used to be part of the CNW mainline, but is now used maybe twice a week for switching industries. In winter, when the temperature dips as far as -30 degrees F, there can be as much as an inch between joints. In summer, when the mercury is somewhere around 100 degrees F, the rails but right up against eachother and start to kink very slightly.
There’s one spot that is particularly frightening to me, a really disturbing looking old switchback that ends in street running, I’ve seen them pull heavy gondolas loaded with scrap down this with a GP38-2, the rails protesting the whole way, not the kind of thing a crew takes lightly for sure!
What I find interesting is that each one seems to be distinct to itself in either the voice heard, information given or ending message spoken.
One of the busiest detectors I’ve heard is the one at mp 78.8 on the Geneva sub on the west end of Global III. This one is real easy for me to recognize and is etched in my brain. It goes something like this:
UP detector, mile post 78.8, track one (or two).
UP detector, mile post 78.8, track one, no defects. Total axles 544, temperature seven one degrees, detector out.
Still haven’t heard a detector ever announce a defect condition. Wonder what that one goes like.[%-)]
They get pretty excited. “You have a defect!” It will then tell you the nature of said defect. I can’t be more specific, because it has been a while since I’ve heard a detector catch a defect.
Funniest thing I heard was when UP 3985 went over the detector outside West Chicago. “You have defects. Hot bearing, axle one. Hot bearing, axle two, Hot Bearing, axle three. Hot bearing, axle four. Additional defects. Repeat. You have defects…”
It’s pretty long, but if you just watch the first 3 minutes you will see the train pass and then clearly hear the detector report the hot box, as well as conversation between the train crew and dispatcher. Very interesting for non-railroaders [:D] Jamie
I was listening in on an internet feed of NS scanners out of Harrisburg, PA the other day. One detector kept reporting hot boxes. It went something like this:
“NS detector milepost xyz, track one, defect detected, hot box axle one-three; hot box axle 2-four; hot box…”
It just went on and on; apparently the defect was in the detector, as the dispatcher and yard master mentioned “It’s a ghost train.”
I figured you would have the answer. Thanks. So, if I have this correctly, the detector does not scan each car, but simply reads the axle “hits” and then instantly uploads the information to the main source, which turns these “hits” into expected data.
So, if a train set out a car (or cars) between the detectors, that information would have to be loaded to the source for the detector information to be correct. Otherwise the trainlist would show 100 cars, when in actuality there would only be 99 cars. How would it account for the updating/reporting of the cars? The axle counts would matchup to the wheel report, but when it would cross check the “hits” vs the data, wouldnt it be one car long? In otherwords after the setout, the source trainlist would show 100 cars, unless it was updated instantly, perhaps that is done. With the trainlist showing 100 cars, and the axle count showing 99, the data for the length of train and speed would be slightly off.,…wouldnt it?
With UPS/FedX and others there is instant uploading of delivery/pickup information so that there is very close to real time reporting. Is that done in the rail industry also?
ed
[quote user=“Railway Man”]
The typical system used today uses two magnetic transducers bolted to the web of the rail to provide timing signals for train presence, s
The primary job of the detector is to alert the crew to hot bearings or wheels. It counts axles and tells the crew which axle(s) are above threshold temperature, counting from the head end, and the more sophisticated ones, which axle on which car or locomotive, counting from the head end. The operating rules require the train crew to inspect that axle plus a certain number on either side. The detector doesn’t have to know anything about the trainlist, number of cars that are supposed to be in the train – it develops its own information.
The secondary job of the detector is to report the information to a database, where the axles can be matched up with the trainlist – assuming the trainlist at that moment is accurate. As you point out, if the train has set out or picked up cars since it last went by an AEI reader, and the crew has not updated the trainlist in the computer, then the data is mismatched. However, the crew reports the car number of a car that sets off the scanner, and that is manually input into the database, so the car is flagged.
The value of a database that shows several thousand no-defect reports per year for 99.99% of the cars and one bad-actor that is already manually input into the system, is not obvious. The infrastructure needed to support it (an AEI reader on every scanner) is not cheap. Several railroads are however putting in combined AEI readers/HBDs and seeing if the value of the information generated is worth the cost. If it is, then combined scanners will become the norm.
The only information the detectors on the U P Mainline here in Iowa give is the track either track one or two whether or not you have a defect and the temp, sometimes Larry