I've been planning a smallish (3'x8') double-track twice-around n-scale layout for the basement, and there's a section that I would really like to automate to prevent trains from running into the wrong end of the points. I have read through the Wabbit instructions a couple of times, and while I get the concepts, I'm just not sure how I can, or if I can, use the capabilities of it to solve things. Here is a schematic of the section in question. I've labeled the two lines that make up the mainline (A and B) and the two sidings that lead off of the section (X's). The space between turnout #3 and turnout #5 is about 2 feet, and 1,2 and 3 and 4,5,6,7 and 8 are really stacked like that, turnout to turnout. I have read all about how you can use a small section of isolated track as a trigger for the Wabbit, causing it to basically throw the points towards the section of track that it receives the trigger from. I'm just wondering how I can set this up so that say, if a train were approaching from the left on line A (turnout #1 clear), and hit the trigger (green blob) right before turnout #5 turnout #7 would throw, but only if turnout #5 were thrown. The other side of this is that if it were to hit the other trigger on the top track, #7 would go to clear, but only if #4 was set to clear. All of the logical for each turnout can be boiled down to something as simple as "[Throw/clear] if trigger at location [x] and turnout [y] is [thrown/clear]", but I'm just wondering if something like this can be implemented using some sort of block detection, triggering or some sort of other new-fangled technology. I haven't purchased anything yet, so I'm open to hearing about any and all ideas. Or if I should just do it by hand and enjoy that aspect of things, that's cool too I just think it would just add a little excitement to watch turnouts auto-correct themselves on the fly.
I have not thought all the way through this...so at this point this is just food for thought... It seems to me that the pairs (2,3), (5,7), and (6,8) are just three single crossovers, so each pair should be thrown as a unit. Also, it seems to be complete there should be a trigger on track A before turnout 1 to clear 3 when 1 is clear. All of this assumes that due to spacing there is no way you can go part way across a crossover and then wait for another train to clear, which seems a reasonable assumption, but if we are going to throw the crossovers automatically it means that we can't even approach the crossover while someone else is on it. Just making sure that's understood. So, we have multiple ways of controlling each crossover, with each only valid under certain conditions. Just a digital logic problem. Let be get some food, and then think a bit. (Though that may come later, I might should work some, too.....) Jeff
Yeah, certainly no reason to not have 2-3, 5-7, and 6-8 set up as pairs. Not sure why I originally thought that this wouldn't work, but yeah... if one is thrown, it's not going to do much good unless the other one is thrown, and if one is clear, the other one might as well be clear, since it's not going to do any good throwing off to a dead-end. Interested to hear what you come up with. I liked Logic Design in college so much I took it twice! (yeah... that's the reason why) edit: ooh, and if I combine those crossovers, I basically get two more DPDT switches, which should make implementing the logic part of this a little bit easier.
The more I think about this, the more I think some kind of route selection might be more suitable than automatically trying to 'fix' misaligned switches. For one thing, to fix them there are a lot more cases to consider, and it could get very messy, especially if there are multiple trains around. I'm back to thinking....
I think you're right. Like Route 100 Clear could set the A line to use the top path (1T,4C,5/7C) and Route 100 Thrown could send set 1C,3C, 5/7T), and something similar with Route 101 for B. I can think of another few routes I'd want to program (everything clear, 2/3 & 5/7 thrown, etc).
Do you want to control the turnouts via DCC or switches or buttons? I think the Wabbit will work, though it might be necessary to add some contacts. I need to fugure out it one can use more than one trigger section for the same action. I can see ten cases where a switch could be approached 'wrong'. Of course if a second train comes along at the wrong time all heck could break loose! Jeff
The 'problem' with using the routes is that it doesn't eliminate the possibility of hitting a switch wrong, if you come in the wrong way. And the routes might be a bit complex, since you can have two trains in the section at the same time. Even so, I think there are only about ten unique, non-interferring routes, so it might not be too bad. Might be able to have routes, and then automatic protection. The Wabbits appear to be preyy smart. Too bad they can't write a manual (in my opinion). Jeff
I'd like to be able to use both, which is what initially drew me to the Wabbit. Yeah, I think that avoiding problems because of a second train entering the area would be unavoidable without some sort of block detection and braking or automatic path selection black magic. What do you mean by contacts? Yeah, that's pretty much the solution I'm looking for. -Jeff (also)
Just more of what the Tortoise already has, the switches inside that you can use to power route to frogs, drive lights, or whatever, based on turnout position. Jeff
If 2/3, 5/7 and 6/8 are wired so that they always mirror one another, that would add another pair of DPDT switches. I'm sure someone somewhere has found more ways to add more contacts off of a Tortoise.
That's why I'm not worried about that. You could use one Tortoise for each of the crossovers, of course that loses the switches, but saves Torti. the parts to do it don't cost quite as much as a Tortoise, though they try.
I see 'just' ten non-interfering setups. There are 4 routes for each A and B, so there would be 16, but there are some that interfere, so those are out. I think the Wabbits could be set up to get anytrain entering against the switches through, where it ends up would be dependent on where the switches were when he came in, but at least he would not be on the ground.
I am not vewy familiar with the Wabbit, but in looking at your original question, two thoughts came to mind: 1) Given the (slow) speed of the tortoise, can it react fast enough between the time a loco hits the sensor and the points complete the desired action? 2) What happens if a train is coming the other direction and hits a sensor that controls a turnout that part of the train is still on? All that said, this looks like a primo application for the DS64 stationary decoder and its programmable routes, particularly if you are running Digitrax DCC (but also doable with other brands, or even if you don’t run DCC at all). Try this: Re-label turnouts 2&3(ganged) to 23 Re-label turnouts 5&7(ganged) to 57, and 6&8 (ganged) to 68 Re-label separate turnouts 1 and 4 to 11 and 44 respectively. Now label the three tracks in the center from top to bottom as 1, 2 and 3 respectively. You will need 2 DS64s connected via a LocoNet cable. Each can drive four separate tortii or tortoise pairs, although you will only use a total of 5 outputs (Perhaps the first set up to drive 11, 23, and the other to drive 44, 57 and 68). The LocoNet connection allows commands generated by the route table in the first DS64 to trigger turnouts controlled by the second DS64. In one of the DS64s, set up the following routes identified by the left, center and right track, with “dummy” 3 digit turnout addresses used as route addresses: Route/Route Address : Turnout commands A1X/110t: 11t, 44t A1A/111t: 11t, 44c, 57c A2A/121t: 11c, 23c, 57t A2B/122t: 11c, 23c, 57c, 68c B2A/221t: 23t, 57t B2B/222t: 23t, 57c, 68c B3B/232t: 23c, 68t B3Y/233t: 23c, 68c Thus by sending a DCC switch command to “throw” turnout 122 (“122t”), the route will be setup from track A via center track 2 to track B (A2B). You can still address individual turnouts directly (e.g., “57t”) if you’d like, and you can wire the sensor inputs on the DS64s to panel mounted pushbuttons to trigger any of these routes. You can also use block detectors such as the BD1 or BD4 to trigger routes via the sensor inputs, much as I think you originally proposed to do (Digitrax has an app note for the DS64 that gives an example). The DS64 has a lot of flexibility. It can get a little intimidating, but then so is your application <G>. Look into it.
That's a cool way of thinking about the routes, based on the way it comes in, the path it takes through the middle, and the way it exits, dstuard. Thanks for putting the time into thinking about this stuff for me, guys. It's helped a lot.
I count 8 paths through the complex. So one could control these with 6 toggles.Note that if the path A via turnout 1 and turnout 4 to X is selected, then both xovers 57 and (23 or 68) must be thrown.But allowing B68R would mean a train in reverse direction A75 would encounter 1 in thrown condition. So the only other route allowed when A14X is selected is B2357A. Additional points on the toggle that selects A14X will throw 23 and 57 to assure this. (or it can be done with a diode matrix) The next possible path is A14nA (where 4n means 4 not thrown). Since 4n, power will not be routed to the 4X path (either by the points on 4 or additional points on the route A14nA toggle) and a train approaching 4 from X will stop. Also, 57n is required to get from 4 to A. and of course 23 must be thrown. So path B23(57n)26nR is the only allowable path when A14nA is selected. The next possible path is A1n57A. Again power must be interupted on the 1n to 4 leg to prevent a train fron X fouling 1. Since 1n, (23n) is required and 68 is required. Thus the only other path available when A1n57A is selected is B(23n)68R. The next path is A1n(57n)R. Following similar logic, (23n) and (68n) are required and the only other path allowed is B(23n)(68n)Y. The next path is B2357A and it is allowed if A14nA is not selected. It also requires 68 and power interupted between 2 and 6. So the only other route is A14X which is in agreement with the control of that route. Next is B23(57n)(68n)R and only A14X is possible again. The B(23n)68R and finally B(23n)(68n)Y can be worked out by the same logic. Armed with this set of logic, the circuit can be constructe using toggles and neither block detection or extra points on the tortices is required.
Hello, I think Doug has it the closest. You just need a "logic" circuit and some sort of detectors to control the area. This can be done with some basic TTL gates (descrete circuit) or what Doug was saying, the Digitrax DS4. Hope this helps,
You don't need TTL. An old fashioned diode matrix and toggle switches will do the job nicely. TTL logic requires a regulated power supply, sockets and thus a pc board, etc. All a pain in the neck when diodes soldered directly on the toggles will do the job simply. I can draw a circuit for you, but you'll learn more by doing it yourself. I'd be happy to kabitz if that would be helpful.
