In another related thread, member, greyhounds made this comment:
“We don’t know that the crossing was timed for the exact minimum. And that’s the fictional flaw the shysters are using. Is there a crossing activation system that will provide the exact programed time for every train under every circumstance? I’d be surprised if such a perfect system existed.”
This is an interesting point. The news reports the NTSB as referring to the specific 20 second activation warning. They refer to this interval being specified by law as the minimum; and they say that is exactly what the Midland, TX crossing had.
But how in the world can you precisely control this interval? At some point, the train approach has to be sensed, and then the signals activated in response. But the system would have to know how long it will take for the train to get to the crossing before it could know when to activate the signals. How can the system know how fast or slow the train will go once the activation has been made?
These systems provide the same time interval regardless of train speed. In the past, they used to provide a distance interval; however, it was hard to know what distance should be built into the installation. If it were too short, a fast train might not provide a long enough warning. But they also had to worry about a slow train creating an excessively long warning.
Excessively long warnings are like the little boy crying wolf. They frustrate drivers with a seemingly pointless extra delay, which causes drivers to take risks by trying to beat trains in the future
The systems do sense the speed of the approaching train, it is not a matter of tripping a switch at some distance. It meausures a current flow between the rails and the rate of change of that current as the wheels of the train basically short the rails together electrically.
The closer the train, the higher the current and the faster the train is moving the quicker the current raises. Thus if the current changes slowly, the train is determined to be moving slow and the circuit can delay the activation to when it predicts the train is 20 seconds from the crossing; likewise if the train is moving fast, then the current will change quickly and the circuit can predict when the train will be 20 seconds from the crossing and activate the signals appropriately.
There are limits where the signals are activated no matter whether the train is moving or not.
I think it can even measure a rate of change to the rate of change (acceleration) and handle that prediction, too.
On another thread, Bucyrus posted: “I read that the float driver was being cleared and protected by a police escort, and therefore he lowered his guard and relied on the protection promised by the police escort.”
If that is true, it would seem to diminish the float driver’s negligence. So where has the system broken down here? 1. It really is silly to keep proclaiming the engineer’s innocence when no one is stating he was at fault. 2. Gates and crossing signal time? Perhaps. 3. The crossing system, where we rely on an old technology geared towards much shorter, lighter and slower trains, to provide safety in a congested zone with long, heavy 70 mph trains that take a long distance to stop. And when it fails, with a consequent tragedy, should anyone be surprised?
I would think that was the main motivation for developing the predictive crossing signal. By having a consistent and reasonable time from when the lights start flashing to the time the train reaches the crossing, drivers will more likely obey the signals as they will consistently see the train crossing seconds after the signals activate. The relatively short and consistent timing is particularly effective where high track speeds are permitted - drivers coming up the to grade crossings on the ex-AT&SF Surf Line are pretty good about not stopping on the tracks.
Another poster mentioned that some of the predictive circuits also measure the rate of change in the rate of change in distance from the crossing (2nd derivative). This is particularly important in cases where the grade crossing is adjacent to a station - the circuit would stop the flashing and gates lowering when a train stops at the station.
I would imagine that the minimum time delay is more based on the speed limits for the road that crosses the tracks and note that yellow times for most traffic lights are only 3 to 5 seconds. On that basis, twenty seconds is plenty of time for any urban road, might be a bit short for a rural highway in Nevada.