Skip to content

Anatomy of a Train Wreck

Last updated on August 17, 2018

Taking a quick detour from photography, I saw this on Video Sift the other day and after reading though the commentary (some of which were way off) I thought it was worth taking a few minutes to explain what’s going on, at least as far as I understand it. But first the video.


First, a diagram, signals are numbered at the bottom, signals have their bases pointing in the direction of approach.


As I don’t know which way the trains are actually traveling so for the sake of this discussion I’m going to say the train with the camera (Train A, from now on) is heading east and would be located west the signal labeled 1 on the diagram at the beginning of the video.The first thing I clued in on this was the aspect of the first signal (Signal 1, on the diagram) Train A passes. The yellow over yellow aspect is defined by the General Code Of Operating Rules (GCOR), which the BNSF operates under, and is called Approach Medium and reads:

Proceed prepared to pass the next signal not exceeding 40MPH and be prepared to enter a diverging route at prescribed speed.

This means that the next signal aspect the train can expect is either some aspect of clear (diverging (red-over-green) or otherwise (green or green-over-green)) or approach (diverging (red-over-yellow) or otherwise (yellow or yellow-over-yellow) or possibly restricting but not stop or stop and proceed. In this case it’s clear that the second signal (Signal 2 on the diagram) was initially set to diverging clear (red over green). Which is consistent with Train A being put in the hole for the meet.

We can also infer from the aspect of signal 1 that:

  1. The turnout has already been lined for the siding
  2. The outbound signal on the main (the top Signal 3 on the diagram, the one the second train should have stopped at) is already displaying stop
  3. The preceding signals for the second train would have reflected the track configuration (approach (Signal 5) followed by the stop at the siding (Top 3))

Had the second train run past the signal (Top 3) before Train A had reached the first set of signals, Train A would have been presented with a different aspect at signal 1 depending on one of two conditions:

  1. If the second train had overrun it’s outbound signal (Top 3) but not the inbound signal (Signal 2), the first signal we see would have displayed approach (yellow over red), because the next signal would have been stop (Signal 2).
  2. If the second train had overrun both it’s outbound signal (Top 3) and the inbound signal (Signal 2), the first signal we see would have displayed stop (red over red), because the second train is already occupying the track protected by Signal 1.

The next item that seems to be disputed is, “Is the second train moving at the time of the impact”. I believe it was, and that it can be readily determined by simple analysis of the video.

The signals give us the best evidence that the second train is still moving. The signal (Signal 2) just before the turnout protects all track from that signal to a set of signals similar to the 2 single head signals in the background (3 Top and 3 Bottom) there but at the other end of the siding (4 Top and 4 Bottom). Or more simply signal 2 protects all the track from Signal 2 to the two Signal 4s.The only way for a train to occupy track between those signals without tripping the incoming signal is to be behind the outgoing signal for its track.

In fact we can estimate the speed of the second train using the time between the dropping of inbound signal to the collision and the distance between the turnout frog and the signal. We can get the time readily from the video, which is 6 seconds though there is probably some processing lag in the signal system so lets call it 10 seconds. The distance is a tad harder, probably on the order of 500 or so feet. This gives us an approximate average maximum speed of 35 MPH, which seems high to me but it’s hard to tell from the video. Of course the second train is likely in emergency at that point and slowing down rapidly so that will the estimate is going to be high because of that too.

The second indication of the second train’s movement is observed in the angles of the second train’s plow relative to the left rail of the diverting track. Given that the camera is mounted on the right side of train A, the apparent distance of the plow should seem to increase from the rail as the camera approaches. Instead it covers the rail further, the only way for that to happen is if the second train is still moving.

Some notes:

In determining signal aspect on 3 light signals (such as those in the video) in the US the lowest signal is red, the middle is yellow, and the top is green. This practice dates back to semaphore signals (the wooden bored that moved), in order to have a fail-safe system the stop position was the one that the board would fall to (i.e. the bottom, or horizontal).

I should also point out I don’t work for a railroad; this is all biased on my understanding of operating rules and the copy of the GCOR rulebook I have. Further the actual investigation of this incident should be public record though the NTSB so it would be a simple case of validating this by tracking down the actual investigation. I really only undertook this because I found it to be an interesting exercise, and there seems to be so much confusion here about what actually happened and what the signals mean.

Published inUncategorized