Stopping Runaways - The Sandhill Derailer
Updated: Mar 7
As many of you know, Mill Creek Central Railroad is not a level track but more of a mountain railroad, especially on the Mountain Division. Over the past several years the track has had occurrences of railcars that would break loose from a train and come careening down from Tower, through Varian Tunnel, down Sandhill and into Barney Yard. If you are familiar with the track you'd know that by the time the cars were at the bottom of Sandhill Block they would be going a very high rate of speed. This past year we had at least three breakaways, one traveling all the way from APEX! See the video of one of the breakaways below.
In the past we have strongly recommended that all railcars operating on the railroad utilize safety chains to help minimize these occurrences, but we have never required their use. I have seen breakaways even at those tracks that require safety chains, and if the track is not extremely diligent breakaways will occur. One of our primary locations of concern is Barney Yard. Should a breakaway occur on the Mountain Division there is a very good chance the railcars will end up in the railyard and, in fact, could end up all the way down in the Creek Block, a distance of three-quarters of a mile from a possible breakaway.
With the blind spot due to the track rolling around the shop into Barney Yard the crew at Mill Creek Central Railroad began to discuss how we might be able to protect the railyard, the track leading to the Valley Division below, and provide some additional equipment and personnel protection for those operating in Barney Yard. Sandhill and Tunnel blocks have a grade mostly around 2.5% but there are areas, especially in and around the tunnel, that exceed 3%. The combination of this grade and the one on Keim Block is where we see the greatest acceleration on the railroad and where it would be the most difficult to safely derail a car or group of cars.
The first idea was to place a manual derailer on Sandhill Block where if a breakaway was recognized someone could make a mad dash to the manual derailer in an attempt to set the derailer. The crew added this safety device to the track but after some additional discussions it became clear that getting someone to safely run to the derailer was not a very good idea. Also, in many cases, there is a good chance that the runaway cars may not be noticed until it was too late anyway.
Another idea that came up was to add a switch or two to these steeper grades. These switches would be placed in the divert position so that the cars would move off the mainline and into a siding and just maybe onto a ramp that would slow the cars or drop them into a sand pile. This too would work but would require the building of additional switches and sidings. But the biggest issue was how would we know when a runaway was occurring and how could we make a safety device that was foolproof.
Our crack brain trust, Nelson Riedel and Dick McCloy, got together and came up with a plan. Nelson, who is in charge of our signaling system (at least until he finishes pawning it off on Kelley Henry 😉) volunteered, as he so often does, to come up with a derailer design, similar to the one that was installed for manual use. The thought was to utilize Nelson's newly designed switch motors for the track and try to utilize them to "flip" the derailer into place during a runaway. He drew up some cad plans and soon had a rough prototype manufactured from his shop. The plan was to test the new design and determine if the mechanism would function properly on the track. After some initial testing along with some minor modifications Nelson built the remainder of the device and brought it out to the track.
After installation it was time to give the device a try. Some of the questions were, would it damage the railcars, would it be damaged after use and would the track be damaged by the forced derailment. There were other questions as well like, would it flip fast enough, where should the derailer be placed and still the big question, how would it be triggered. So, I guess you don't know until you try so that's what we did! Want to see how it did, play the video below.
As you can see, one of the benefits of the derailer was that it was able to remove the car from the track without flipping it over. With the successful testing behind us, the next issue to tackle was how would a derailing device be triggered. So first we had to think through the logic of the process. Clearly, we didn't want the derailer on the track as trains were passing by, nor did we want it to activate while there was a train in the block. Another option was the use of a switch. Clearly a thrown switch could be used and would not be an issue for trains moving uphill (westbound) as the train and cars would throw the sprung points out of the way as the train moved up the hill, just like all of our other switches do. But what if we had a roll back. Also, it should be noted that Sandhill Block is a fairly short block, only about 250 feet. Clearly, we didn't want to have a switch thrown or a derailer on the track with a train coming downhill (eastbound) either. We could have traincrews throw the switch points or flip the derailer through the use of a switch box but then the traincrews would have to remember to throw them, not once but in the eastbound case, twice, to allow them to be properly positioned. Clearly this methodology was not "foolproof!"
For purposes of moving the project forward we choose, at least for now, to utilize the derailer device. So, when was it safe to have the derailer engaged? Clearly if no one was using the block the derailer could be engaged. But how would the "track" know that no one was on the Sandhill Block? After some additional discussion we determined that we could use the signaling system. There was still an element of human error should the traincrew not trigger the block but the traincrews are supposed to trigger the blocks before they enter so if they don't then they'd get quite a surprise.
So, the crew decided to use the signals, but how? Nelson to the rescue again. With the recognition that the only time the derailer should be engaged was when the block was clear and the signals green, Nelson developed a relay circuit in the signal head at the bottom of Sandhill Block. The new relay circuit would recognize the flow of power through the green light circuit and that relay would then trigger the derailer to disengage when the block was "taken," and no power was flowing to the green LEDs. When the block is released the green lights would again trigger the relay and reverse the polarity allowing the derailer to again be engaged. After Nelson completed that circuit, he and Kelley installed it and began testing. The signal team also incorporated red and green warning lights on the derailer to let the engineers know the status of the derailier, another safety precaution.
So, after a couple of weeks of testing the derailer was working as planned and the system was done, but not quite. After some additional discussion about how the track was used someone brought up another issue, the length of Sandhill Block. As many of you know who visit the track, many of our visitors love to run double headed on the railroad and these trains can get rather long, sometimes nearing 200 feet in length. That may create a problem, right? As a double header enters Sandhill Block, they take the signal, turning the green signal head to yellow, which deenergizes the green light circuit and thus triggers the relay to disengage the derailer, perfect. But now as the double header gets to the end of the block the engineer releases the block, now we have a problem! As most of the train is still in the block the signal has returned to green and the derailer is now engaging, DERAILMENT TIME! So, if we were to use a sprung switch instead (which we had determined that we weren't) this would resolve the problem going up the grade but it certainly doesn't resolve the issue going down grade. Maybe we should require that all double headers have a conductor on the back and make them release the signals, but that doesn't seem very foolproof either.
So how do we solve this issue? After some additional discussion the crew came up with another idea. What if there was a double trigger, what if the derailer would engage only if the signal light was green and if something was detected on that track. For that matter only if something was detected on the uphill side of the Sandhill Block. But where would the detection go? The answer came in the Wilson passing siding. As we know the passing siding tracks are not bidirectional and therefore if a group of breakaway cars are coming down the hill on Wilson, the track they would utilize would be the eastbound track, not the westbound. As trains are going uphill on (westbound) Sandhill they pass the downhill trains traveling eastbound on Wilson on Wilson's westbound track. So if we were to place an optical sensor, similar to a garage door sensor, on the eastbound track just before the switch we could accomplish what we were looking for! Also, doing so would reduce the number of cycles that the derailer would be required to make as it would only "engage" when the signal light in the Sandhill Block was green AND if the optical sensor detected movement along the track just before the lower switch in Wilson (see attached diagram). This location also works well as a train moving down the hill (eastbound) from the Wilson passing siding into Sandhill Block will "take" ("take" shown in yellow at the top of the diagram) Sandhill Block before reaching the optical detection (located on the eastbound track in Wilson at the top of the diagram next to the Sandhill EB Signal shown below in Red) and therefore avoiding the triggering of the derailer.
So this spring Nelson and Kelley will be installing and testing the derailer with the added optical sensor. We may even move the derailer further down Sandhill to ensure there is enough time to flip the derailer into place before the breakaway cars get to it. Is the system foolproof? Not quite, but it is about as close as we think we can get. Should we be using a switch instead of the derailer? That debate continues. Are there other ways of doing this? I'm sure there are but this was the simplest and lowest cost method we could come up with. Do we still recommend the use of safety chains, ABSOLUTELY! If further use and testing proves beneficial there is the possibility of additional installations along the railroad. Why don't you come out and try the new derailer (well... not exactly derail... you know what we mean 😂) and let us know what you think. Do you have a different/better idea? We'd love to hear it!