Time-delay Turnout

[Fender]

I'm encouraged to learn about the setup at Ridge and PLS. Will have to find out more about how they work. Will probably be at Ridge this coming Feb.

[kcjones611]

Where exactly were you planning on this turnout being?

[Steggy]

REALLY. BAD. IDEA.

When I asked this question, I expected to get some helpful, useful suggestions, as well as some negative, unsolicited opinions. I wasn't disappointed. Thanks.

Always glad to be of service.

It is not a good idea, and I wasn't the only one to point that out. I'm all for automation (I'm well-known among my acquaintances for automating things), but a turnout that could malfunction at the wrong time and put a train on the ground is not a good place for it. In fact, I'm not real keen on powered turnouts in general, since most of the examples I've seen are junky. This is a hobby where oftentimes low price is vastly more important than superior quality, and powered turnouts tend to fall into that category.

But, hey, it's only an opinion, and you know what "they" say about opinions...

[Builder01]

REALLY. BAD. IDEA.

Hi Dan,

I apologize for the guy that posted this response. Now we all know why he won't tell anyone his name or where he lives.

David

[ccvstmr]

Dave...don't discount BDD. I know the guy personally. He worked in the RR industry more than a few years. He places reliability above cost when it comes to electro/mechanical designs. This is something too many people DON'T do. And when working in the railroad industry...the priorities need to set in that order...as train vibration is brutal on anything electro-mechanical. Furthermore, reliability is sometimes improved when the human-element is removed from the equation. Still, controls needs to be in place in an attempt to make any system foolproof.

BDD designed a signal system at my club. The club threw every possible problem regarding signal systems at him based on prior experience. He designed a system that satisfied all possible problems that could come up. The signal system is pretty "solid"...except if it takes a lightning hit (not much can survive that). So whether a contributor name or location is offered...has nothing to do with the man's technical abilities. If that's how you're going to judge the content of material posted on this forum, then you're missing the point. BDD's comment may have been over the top, but the underlying concept he tried to portray might be spot on. Just say'in...Carl B.

[chooch]

David,

I suggest you read the "BUILD" section of an F-7 IN SCALE. I would think if "bigdumbdinosaur" can figure, design, show and Explain everything so clearly as he does post of his building an F-7 IN SCALE.
I would take his word for--"REALLY, BAD, IDEA.

Personally, I know squat about "time delay" turnouts but, I`ve never experienced any track I`ve been on with Time Delay turnouts and don`t think it`s a good idea either.
Thank you.
chooch

REALLY. BAD. IDEA.

Hi Dan,

I apologize for the guy that posted this response. Now we all know why he won't tell anyone his name or where he lives.

David

[Builder01]

Hi Guys,

Yes, I understand BDD is a talented fellow and is capable of explaining things in great detail. I respect that. But, posting a terse response in giant letters is the equivalent of yelling at someone at the top of your lungs for asking a simple sincere question. It really does not matter if BDD's response is right or wrong. It just seems completely out of line with how most folks here deal with the sincere questions posted by others.

If someone had asked this question face to face at your local track, you would not start yelling at them. The response seemed pretty impolite.

David

[chooch]

Capitals might be "Yelling" in speaking but, maybe it is to make a Point, not (here we go) Politically Correct" or whatever. Do I know how to "underline" , no. I`m a Dinosaur on this machine with two finger press a key.
So, I guess I`m NOT like most folks. SO THERE!!!.
Have a nice day.
Thank you.
chooch East Central Florida

[Steggy]

I apologize for the guy that posted this response. Now we all know why he won't tell anyone his name or where he lives.

Please don't apologize on my behalf. We already have an elected official who does that for us.

[chooch]

For what`s it worth, I have just seen on another thread subject, a couple posts with NO locations mentioned. In addition, several other posters here and there also DO NOT (I`m yelling, yes)DO NOT give any locations of their whereabouts.
At times, not very often, I see someone in this group suggest a poster give their General location for various reasons.
Maybe it`s about time to get crackin` and mention to ALL what they are supposed to do. So whose in charge here?? What`s good for one is good for all.
chooch
(edited 1 time)

[Fender]

My thanks to everyone who posted helpful, constructive responses to my request.

[Steggy]

I'm looking for ideas on how to set up a time-delay mechanism to allow the points of a turnout to be set to one position, then automatically return to the other position after (let's say) one minute. This would allow (for example) a switch that is normally aligned to the main track, to be aligned to a siding, and a train to pull into the siding, without the engineer having to walk back the length of the train to throw the points back to the main track. Since no power is available out on the track, a purely mechanical system would be preferable.
Anyone have some suggestions or ideas?

Okay, now that I've caused a bit of an uproar around here, please allow me offer some thoughts on how you might accomplish what you want. My opinion remains that fully automating turnout operation is a bad idea (machines can't think, whereas most human beings can), but being somewhat of a "control freak," I can give you some useful direction on how to develop something. Caveat emptor on what follows, as it's all out-loud thinking that hasn't been subjected to failure mode analysis. Identifying and preventing automation screwups is a bit like attempting to rid a greasy spoon restaurant of cockroaches.

As I understand it, you have a descending track from which branches a runaway stub and you want to normally keep the turnout leading into that stub in the diverge position until a train that is or appears to be under control approaches it from the facing side, at which time the turnout would automagically align to the mainline, allowing the train to continue down the mainline. After the train has passed the turnout would automagically return to the diverge position. If a train is approaching the turnout from the trailing side, no response would be required, as the turnout could be a spring switch that can be trailed.

As I see it, the automation has to be able to do the following:

1. Determine the train's direction of travel (DOT). As I noted above, the turnout doesn't have to do anything if the train is ascending the grade, as it would be making a trailing approach. If DOT is a facing approach then the automation has to...
2. Differentiate between a train that is under control and one that is not. If the automation "knows" that the approaching train is traveling at a speed that is commensurate with controlled operation it can align the turnout to the mainline. Otherwise, it would be assumed that the train is out of control and the turnout would remain in the diverge position, forcing the runaway off the mainline—or derailing it.
3. "Know" when to return the turnout to the diverge position. It is patent that the automation must not realign the turnout as a train is passing over it. So the automation has to know when the last wheelset of the train has cleared the turnout.

The key to achieving this operation would be in the use of one directional block occupancy detector (DBOD) and one non-directional block occupancy detector (BOD), along with a timing device and some simple relay logic.

The DBOD uses a split block arrangement to both detect the presence of a train and establish DOT. In the context of your application, the DBOD's split blocks would be "leading" and "trailing," those adjectives referring to the side of the turnout that they govern. Blocks, as you (should) know, are electrical circuits established in the track by use of insulated joints and bonding wires at non-insulated joints.

The leading block starts at some point up-line from the facing side of the turnout—at a location I'll call the "point of no return" or PNR—and ends a few feet beyond the trailing side of the turnout. The trailing block commences where the leading block ends and extends a short distant farther down-line—the exact distance is not critical.

The BOD would monitor a timed block whose end is the PNR and whose start is a measured distance up-line from the PNR. So there would be three consecutive blocks in all: the timed block established by the BOD and the DOT blocks established by the DBOD.

Operation when a train is descending (facing point movement) would be as follows:

1. A train enters the timed block and starts the timing device, whose period of operation (T) is equal to the minimum amount of time that would elapse before the train reaches the PNR if it is under control. For example, if it has been determined that an in-control train would never exceed 10 MPH (actual, not scale) and if the length of the timed block is 100 feet T would be 6.8 seconds, which is computed by the equation T = D ÷ (V × 1.46), where T is seconds, D is feet and V is miles-per-hour. D in this case represents the length of the timed block.
2. If the train reaches the PNR—detected by entry into the leading block—before T has elapsed the automation will deem the train to be a runaway and the turnout will remain in the diverge position.
   
   Otherwise, the train will have arrived at the PNR when or after T has elapsed and the next steps will occur.
3. Upon passing the PNR, the entry of the train into the leading block will cause the automation to align the turnout to the mainline. An indicator of some type would let the engineer know that this has happened—the engineer should be prepared to stop if he fails to get the expected indication.
4. Upon entering the trailing block, the automation would prepare to return the turnout to the diverge position, but would not do so until the last wheelset of the train has left the leading block. Since the end of the leading block is on the trailing side of the turnout nothing will happen until after the last wheelset is clear.

In the event the train is ascending, which means it is trailing the turnout, it would enter the trailing block first, which would tell the automation to do nothing.

As I early said, I haven't given any significant thought to failure modes, but can think of one possibility.

It's well-known that engineers tend to follow each other too closely, with occasionally predictable results. Suppose a train has passed the PNR after T has elapsed, and right behind it is another train that is a runaway. Having reached the PNR after T has elapsed, the first train is able to continue down the mainline because the automation has aligned the turnout accordingly. If the second train enters the timed block before the first train has left the end of the leading block the automation will not know that another train is in the timed block. The runaway will be able to continue down the mainline as well, and smash into the first train.

Other considerations are:

• What if a train enters the timed block, stops and reverses out of the block?
• What if a train passes the PNR after T has elapsed, stops and reverses completely out of the timed block?
• What if the engineer wants to take the runaway stub instead of the mainline? How would he do it?

As I said, greasy spoons and cockroaches...