Message Areas - "PO&N signals, the Christmas tree project."

Train-Sim.Com Message Areas

"PO&N signals, the Christmas tree project."

Email this topic to a friend
Printer-friendly version of this topic
Bookmark this topic (Registered users only)
Previous Topic | Next Topic Conferences MSTS Route Design (Protected) Original message
sniper297 (1027 posts)
Dec-10-01, 00:00 AM (EST) "PO&N signals, the Christmas tree project." LAST EDITED ON Dec-10-01 AT 00:05 AM (EST)It occured to me belatedly that I hadn't shared the results of all my mad scientist experimenting, so here goes.
************************************************************
DISCLAIMER! I'm an amateur hacker like everyone else around here, don't take this garbage for gospel, these are my observations and opinions based on trial and error, not the gospel according to Saint KUJU. If I have some of this wrong, anyone from MS/KUJU is free to tell you how it really works as soon as their vows of silence expire.
*************************************************************
How was that, Rich? Think that'll keep the dummies from formatting their hard drives?1. Interval signals are always active, if you drive thru one that's off of your path it will make the same indication that it would on the path. There are four states, not three, green, flashing yellow, steady yellow, red. As near as I can tell, if you want a green light approaching a yard at the end of the line, you need at least 3 signals beyond that one to get a green indication. A signal at a dead-end will show red, not because the block is occupied, but because there is no signals beyond it.2. Interlocking signals are inactive unless there is a currently active path thru the signal. Linked or unlinked, in an activity an interlocking signal defaults to red in a "sleep state", and if you approach one off the path, it will stay red. An interlocking only "wakes up" and gives the proper indications when the active path comes along in advance of the train, AI or player, to trigger it. After passage of the train, the path is erased behind the train, and the interlocking goes back to it's at-rest default red indication, regardless of whether the track is clear or not.So here's what I came up with, the AI traffic is routed around all switching districts on tracks made just for that purpose, and an AI train always takes the curve at every automatic switch to diverge ("take other exit" in the activity editor), and never goes thru a manual switch. The player is given a short path that does not cross an automatic switch onto a single track main section, and has to read the work order, consult the map, and throw all the manual switches for his path as he comes to them. This is done because if the player path goes thru a switch that has an AI path thru it, the player path takes priority, and the AI train will wait 3 minutes or 3 hours, refusing to move until the player passes.
The first pic shows the signals on a real short single track section, since the player always takes the straight track (automatic switches also have a default "at rest" state when there's no path thru them - they reset to straight!), he never goes thru an interlocking, only the intervals. The doublehead interlockings for the AI path are linked backwards, since the AI train always takes the curve, the top link is to the next curve (diverging path) and the bottom link is for the straight track, where the AI train ain't supposed to go. The problem I had with this initially, it worked OK as long as the player train was already past the base of the switch on the single track section, the AI dispatcher ("HAL 9000") would detect the presence of the player train and not throw the switch.
Player train is in between switches, advancing AI path stops at switch, HAL delays throwing switch for AI path until player train clears.The problem was approaching a switch at the converging point, the signals is clear, then suddenly it's red, both switches are thrown for the curve, by the time you stop you're on the single track section with both switches set for the AI path, can't go forward or back. Replacing the auto switch with a manual is no good, that gives you an escape, but then the AI train is permanently stuck. Interlocks are no good unless you have your path going thru them, with all the complexities and limitations that come with that. But I did find that placing a single face interval signal where the double head interlock is normally located makes it workable, since the path interrupts at the back side of the interval, both switches won't be thrown at the same time, always giving the player an out.
In this last pic, the AI train will be coming from the left, with the path advancing ahead of it (so many signals? so much distance according to speed? I still dunno.). When the path reaches the switch, the player train is not on the singletrack, so HAL throws the left switch. The interval by PL1 turns red, if PL1 happens to be in that few-second window where it's too late to stop, he can just keep going to the next switch (on the right) which will not throw once his train is detected past the base of the right hand switch. Notice I'm talking switches, not signals, the overtaking AI train will not pay any attention to the fact that the block is now occupied, you have to clear the next switch and get out of his way before he catches up and rear ends you. Once you clear the next switch, that will throw for the AI train.
In the case of PL2, coming from the right for an opposing meet, he'll have more warning and an easier time of it, for some reason if he blows thru a yellow onto the single track and finds the next switch is set for the curve and a head on with the AI train, the AI train will stop. I guess they don't mind rear-ending you, but they're too chicken to play chicken face-to-face. However long he sits there on the single track, the AI train won't move, the left switch stays set for the curve, the right for the straight. So the player breaks the deadlock by backing up onto the straight track he just came from, and the right switch throws for the curve, and the AI train goes past.So that's where it stands at the moment, the only two problems are the flashing yellow or yellow for the converge actually meaning stop, and the fact that 'ol HAL will only hold AI traffic for the player train - if you uncouple a cut of cars on the singletrack mainline section, that's a loose consist now, not part of your train any longer, and HAL will clear a path for Crazy Jones to come along and happily plow into it. Nothing like artificial intelligence, hey?
Alert Edit | Reply | Reply With Quote | Top
Table of Contents

Subject Author Message Date ID RE: PO&N signals, the Christmas tre... rgarber Dec-10-01 1 RE: PO&N signals, the Christmas tre... rgarber Dec-10-01 2 RE: PO&N signals, the Christmas tre... sniper297 Dec-10-01 3 RE: PO&N signals, the Christmas tre... rgarber Dec-10-01 4 RE: PO&N signals, the Christmas tre... sniper297 Dec-10-01 5 RE: PO&N signals, the Christmas tre... lwhelan Dec-10-01 6 RE: PO&N signals, the Christmas tre... lwhelan Dec-10-01 7 Sorry Jim, But you are wrong.......... lwhelan Dec-10-01 8 RE: Sorry Jim, But you are wrong...... rgarber Dec-10-01 9 The test track.... lwhelan Dec-10-01 10 RE: Sorry Jim, But you are wrong...... lwhelan Dec-10-01 11 RE: Sorry Jim, But you are wrong...... rgarber Dec-10-01 12 RE: Sorry Jim, But you are wrong...... sniper297 Dec-12-01 13 RE: Sorry Jim, But you are wrong...... rgarber Dec-12-01 14 RE: Sorry Jim, But you are wrong...... sniper297 Dec-12-01 15 RE: Sorry Jim, But you are wrong...... rgarber Dec-12-01 16 Yes Jim, How do you do that........... lwhelan Dec-12-01 17 RE: Yes Jim, How do you do that....... rgarber Dec-12-01 18 RE: Yes Jim, How do you do that....... lwhelan Dec-12-01 19 RE: Yes Jim, How do you do that....... sniper297 Dec-12-01 20 RE: Yes Jim, How do you do that....... sniper297 Dec-13-01 21 RE: Yes Jim, How do you do that....... rgarber Dec-13-01 22

Lobby | Topics | Previous Topic | Next Topic
Messages in this topic
rgarber (792 posts)
Dec-10-01, 00:42 AM (EST) 1. "RE: PO&N signals, the Christmas tree project." Jim,Okay bud.. I'm gonna make you walk me through this... not because I'm disagreeing because I'm having a hard time following ya.First 1 and 2, make sense. Sounds good.The third paragraph I followed you. Then you start in on the explanation. Okay... why did you bring up this track arrangement in particular?And in this track arrangement, and I'm just asking you to explain this... why is there a two head interlocking on one side, and a single interval on the other track coming from the left side of the picture?It looks like you have the same arrangement on the right hand side of the picture with those set of tracks too, right? So this would be my second question but I figure when you answer the first question in this sequence you'll answer that too.Then, why do you have a single interval in the middle? I would've thought if anything, that track would be cleared of a signal.So if you could answer these questions first, I'll get to my next set of questions.Oh one more... is this setup such that an AI train can only come from the single interval branch from the left side? I know that's what happens in the activity. What happens if I would set up a AI train to start from the interlocking position on the left side. Would that still work?Thanks!Rich Garber Remove | Alert Edit | Reply | Reply With Quote | Top
rgarber (792 posts)
Dec-10-01, 11:13 AM (EST) 2. "RE: PO&N signals, the Christmas tree project." On second thought Jim, I remembered this morning that you covered the why's in the previous thread on this. Very interesting though. Rich Garber Remove | Alert Edit | Reply | Reply With Quote | Top
sniper297 (1027 posts)
Dec-10-01, 01:43 PM (EST) 3. "RE: PO&N signals, the Christmas tree project." LAST EDITED ON Dec-10-01 AT 02:07 PM (EST)Too late, I already typed all the answers out, it's copied onto the clipboard, here comes the paste....aaaargh, I couldn't stop it.You know, it's amazing how many new scientific discoveries are made purely by accident. I'm looking for a solution to the strange behavior of signals and AI traffic, and it appears I've stumbled across something denser than uranium.
This may clear it up, altho if you're looking at it in the wee small hours of the morning I can't guarantee it's gonna penetrate the U-235 fog in your head. Don't you ever sleep? In the following diagrams, interlockings are shown in red, intervals are green, the colors have nothing to do with how they respond, I only colored them to avoid confusion as to which type is which. Here's the traditional way to set up signals;
Each double head interlocking guards the approach to the base of the switch (single track going to double track), is linked with the top button to the straight track, and the bottom button to the curve, and is supposed to indicate (1) which way the switch is thrown, and (2) occupation state of the track beyond the signal. The reason this does not work unless the player path goes thru the switch AND IS CURRENTLY ACTIVE is because of the nature of the interlockings themselves, defaulting to an inactive state. ("What was the strange thing that the interlocking did in the activity, Holmes?" "The interlocking did nothing, Watson, that's what was strange!") The "currently active path" part of the equation is important, again, as the train proceeds along the path, the path is erased behind it. This is why reversing points in activities are so difficult, and why their behavior is so ambiguous. First time thru the interlocking works fine, second time depends on whether you have appeased the gods of the reversing point, third time forget it. Take a look at the few original activities that have any switching involved, the switching movements are limited and very basic, AI traffic is sparse, and usually on another track. (Very few opposing meets in the KUJU activities, most KUJU routes are double, triple, quadruple tracked over the majority of the mainline.) When there is any potential traffic conflict, the timing of the meet is rigidly controlled, and any variation from the script by the player screws it up. This may be fine for most people, but it's unacceptable to me, and doesn't feel even remotely realistic. Hunky-dory for high speed passenger trains, hopeless for peddler freights.
There I go off on a rant again, where was I? Oh, yeah, how it works. Well, after trying various combinations and placements of interlocking and interval signals, I found the following;1. At the moment the AI train is spawned, the AI dispatcher (HAL) sets all the switches ahead of the AI train for it's path, until it comes to a previously existing path or a signal. The path triggers and activates interlocking signals (X number of signals ahead, I still don't know how many, and I'm fairly sure that time, speed, and distance are not factored in), and either some kind of temporary "end of path" or wait point is generated at a red signal.
2. If the signal goes yellow or green before the arrival of the AI train, HAL continues laying out the path and setting switches, and the AI keeps going. If the signal hasn't changed, the AI train stops at the temporary end of it's path. This is why the AI trains usually go past red single unlinked interlocking signals until they are just clear of the converging switch, it's actually stopping for the end of path point, which is at the base of the switch regardless of where the signal is placed.
3. The AI dispatcher directly checks the occupation state of each block as he comes to it, if the player TRAIN is in the block, it will stop the path. Note I say "TRAIN", in this case a loose consist does not count as an obstacle. If you uncouple one or more cars in the block, they are no longer part of your train, so once the end of your remaining train clears the next signal, the AI dispatcher considers the block "clear" in spite of the red signals at both ends caused by the loose consist sitting in the block. The AI path continues thru the loose consist, the AI train runs the red signal, bang-shang-a-lang.
4. The AI train does not see signals at all. If a block is clear when the path generation (or advancing movement of the path end point ahead of the AI train, however you prefer to think of it) arrives at it, the path is continued thru the block to the next signal. If another train enters the block after the path has already been extended thru it, the AI train will run the red signal regardless of whether it's an interval or an interlocking, permissive or absolute. Again, it's not stopping for anything but that end of path point. This is why sometimes two AI trains will appear to stop for red signals, and other times go thru and collide with each other, it depends on when the opposing paths were extended ahead of the two trains. Paths meet and stop, both trains stop, paths overlap, bang-shang-a-lang.So. With all that in mind, the player needs to go thru only interval signals, no interlockings, and only straight thru automatic switches. The AI trains have to have a path, since there's no actual "intelligence" in artificial intelligence, so they have no problem with interlocking signals, which are activated before they get to them, and they aren't bothered by red lights anyhow. The AI train always goes thru the curve, always "take other exit" at every automatic switch. So the way I set it up initially;
worked only partially, the single track section was a single block, if the player was on the single track section, HAL would delay throwing the switches for the AI path until the player was in the next double tracked section. Problem was, if the player happened to be going on to the single track section at the moment the leading edge of the projected AI path arrived, both switches were thrown at the same time, it's too late to stop, player train is now stuck on the single track, unable to go forward or back up. Placing a double facing interval in the middle of the singletrack was no good, the AI path advanced to the interval, throwing the left hand switch, the player coming from the right still sees a green, goes thru the interval, now the block behind him is clear, AI path extends past him thru the block behind him, throwing the right switch, stuck again. So I came up with this;
Again, the AI train always takes the curve. Taking a left-to-right movement for example, when the advancing AI path from the curve hits the back side of signal 3, that's a pause point. The path temporarily stops there, the interlocking wakes up, the left switch throws curved. Signal 1 goes from green to red because the switch is now set against it, signal 3 goes yellow, signal 2 flashing yellow. If the player is also approaching from the left and don't have time to stop, he can continue on to the right hand switch, which is still set for the straight track. If the player is approaching from the right and don't have time to stop, he can stop anytime before entering the left switch and back up. No stalemates, once the player is clear of whichever switch is still set for the straight track, it will set for the curve, the AI train continues on, the switches go back to straight and the interval signals clear after he passes. It works in either direction, altho not perfectly, and the player needs to be aware that approaching single track from double track, if the track next to you has a doublehead signal, slow down and use caution, that flashing yellow ahead really means stop. And the player needs to keep the switch window open to check the state of the next switch when he passes a yellow, familiarity with the route helps a lot. Clear as mud yet?"Okay... why did you bring up this track arrangement in particular?"
Because it's the one I'm talking about. The whole convergence/divergence interaction between the player and the traffic is based on it. The example was convenient cuz it's a short one and fits in the picture, they all work the same. "... why is there a two head interlocking on one side, and a single interval on the other track coming from the left side of the picture?"
Some of my insane babblings are actually semi-coherent, you oughtta stay awake thru at least part of them. I thought that was Bo snoring in all the previous threads on this subject, I wondered why it sounded like stereo. Interlocking = player no pathey, no green lightey. You savvey?"Then, why do you have a single interval in the middle?"
I don't. It's two single intervals facing each other, taking the places of the double head interlockings that normally guard the throats of the switches. On a longer single track section they would be much further apart, since they need to be close to the base of the switch.

"Oh one more... is this setup such that an AI train can only come from the single interval branch from the left side? I know that's what happens in the activity. What happens if I would set up a AI train to start from the interlocking position on the left side. Would that still work?"
You have it backwards. Player goes thru interval, AI traffic goes thru interlockings. As far as "what happens if", you're the activity expert, you tell me. I suspect that if you swapped paths or actually made a player path going thru all that, you'd be right back to the same thing, unbelievably complicated programming to set up the simplest meets, rigidly controlled movements that won't tolerate the slightest deviation from the path or timetable without going spastic. Well, that's enough babbling for now, I'm wearing out my keyboard, gotta go get a new one.Uh, one more thing, in case anybody hasn't had enough of this yet, here's some of the previous stuff;http://www.train-sim.com/dcforum/DCForumID6/2595.html
http://www.train-sim.com/dcforum/DCForumID6/2444.html
http://www.train-sim.com/dcforum/DCForumID6/2386.html
http://www.train-sim.com/dcforum/DCForumID7/585.html
I'd put in "The signals that time forgot" thread that Bo started, but I can't remember where it is.
Remove | Alert Edit | Reply | Reply With Quote | Top
rgarber (792 posts)
Dec-10-01, 02:41 PM (EST) 4. "RE: PO&N signals, the Christmas tree project." There's a couple of presumptions introduced here that make me wonder.First, 'all AI trains go through the curve.' Part of me wants to say, this is incorrect because I have seen otherwise. I wrote in my tutorials that the AI trains must have the Begin Passing Path because I kept seeing the opposite. This is not to say you are wrong but I have seen otherwise, alot. I wish I had time to investigate this. One reason I may not be reaching the same conclusions you are is because in places on my route, I used the same swith for both ends of the siding. Flarfann asked me why I did this. I told him.. maybe.. I screwed up??? Now this isn't in all cases. Anyway, typically my AI trains will go straight if I didn't use the Begin Passing Path feature. Trust me, I've met many a face to face stalemated until I included the Begin Passing Path option.Second, you say AI trains cannot see signals. Again part of me is thinking... 'but remember the time...' and once again I'm in semi-disagreement with this conclusion. This again is not to say you are wrong. But when I have seen otherwise my instincts tell me you are of course seeing what you are seeing, but are you conclusions the way it is. I'm not trying to argue to say I'm right, I just want to do it, the way it works.By no means do I claim to be an activity expert. What I did I based on my own observations of doing a number of activities and a thorough reading of the Activity forum. You know as well as I that none of this is chronicled by Kuju so we're all guessing on this one. I've done 19 activities for the EMRR, 4 for the Reto Metro and for the most part these work fine for most people. This doesn't mean I could get the signals to work for your route. But I would like to think that somewhere in all of this that signals do have some kinda of consistent behavior. Other than badly I mean....In another segement of your post, do I read you right that you are setting up unlinked signals, or do you mean the single interlockings where you don't have to set up a link? Just trying to get a clarification is all.What I think we need is a test track, setup from the start for signals. I really wish I had the time to try to understand the PO&N but with so many other projects it just isn't happening. I'm just not sure the PO&N was setup to handle signals and one of things I learned doing the Metro was I had to go back and redo whole segments of the line to accomodate the shortcomings of the things. It maybe that you are getting results from your configurations to work on the PO&N, but would those results be repeatable elsewhere? Would they work in a more complicated activity? Would folks experience the same amount of activity failure using this method?I'm asking these questions because I'm genuinely curious, not to make trouble. My hope is through your exploratory efforts, some new light will be shed on this. Right now I can create just about any activity I want using player trains and AI trains on the Metro. There are some I won't because I didn't do things properly and to try is to fall back on old techniques to make things work. And then if I do get them to work, is it worth uploading a whole new Metro version and ask people to download it again? I dunno. With this next route I'm taking greater care with what I have learned to not repeat the previous mistakes. But then I won't be doing any signal research until I reach that point.What I'm thinking is we create a test route we share amongst each other and try observing train behavior that way. Then maybe once and for all we can reach a consensus on what is happening so at least our suppostions will be the same. But with our experiences being different even at the lowest levels, I think the best we'll come up doing is scratching our heads.Rich Garber Remove | Alert Edit | Reply | Reply With Quote | Top
sniper297 (1027 posts)
Dec-10-01, 06:21 PM (EST) 5. "RE: PO&N signals, the Christmas tree project." You know, I think you may have something there. You musta finally gone along with the Scarecrow to see the Wizard. I'm reluctant to stop work on 5 now, but a test route wouldn't take that long if I kept it small and simple, and the track layout could be set up with positional vantage points where the player engine could stop and observe the actions of signals and behavior of AI trains. In the long run it would save everybody time and aggravation.
Just to clear up one point you're still confused about - WITH THIS PARTICULAR CONFIGURATION OF SIGNALS the AI traffic ALWAYS takes the curve, ALWAYS take other exit. That's the only way to make it work without a player path, the AI dispatcher will set the switch for the curve when an active path goes thru it, automatic switches will return to the default straight ahead position once there is no longer an active path thru the curve. If you're setting up an activity on a route that don't use this type of signal configuration and dedicated AI bypass sections, you can make the AI train always go straight if that's what you want. "Always takes the curve" only applies to this setup, nothing else. And one of these days I'm gonna prove my theory that the AI train don't pay attention to signals, only some kind of temporary stop point the AI dispatcher sets for it. And like it or not, you ARE an activity expert by virtue of knowing more about it than the rest of us. So take a shot at it, write an activity for the PO&N beta test route, see if it's either more or less complicated with a player path, wait points, passing paths, see if you can do 2 way AI traffic, you know, experiment. The whole point of all this is to improve future routes, and I personally won't be put out if MS finally releases a massive patch that fixes all this nonsense and makes all our work a waste of time.
I gotta go make a signal route now, "Catchyas". (Doggone Aussie has me doing it now!) Remove | Alert Edit | Reply | Reply With Quote | Top
lwhelan (37 posts)
Dec-10-01, 06:50 PM (EST) 6. "RE: PO&N signals, the Christmas tree project." Rich & Jim,As I said to you gentlemen in my lengthy email :-f. Made up a test route - wye at each end - long straight bit in between - lots of signals,
switches - 2 tracks - 4 tracks. To test passing paths - stuff like that.
It also has an observer track that the player can scoot up and down to monitor the AI trains,
signals etc - while the AI trains are running. Was spawning AI's at 3 minute intervals.
But it was in vain.................Catchya's
Laurie Whelan Remove | Alert Edit | Reply | Reply With Quote | Top
lwhelan (37 posts)
Dec-10-01, 06:57 PM (EST) 7. "RE: PO&N signals, the Christmas tree project." ............ and as a further development.
I am currently re-writing the scripts and configurations for the default BNSF signals.Whilst very carefully following Jim's thread(s) - I have the off-path excursions working to a degree etcCatchya's again
Laurie Whelan
Remove | Alert Edit | Reply | Reply With Quote | Top
lwhelan (37 posts)
Dec-10-01, 07:21 PM (EST) 8. "Sorry Jim, But you are wrong........"
Your paragraph....."4. The AI train does not see signals at all. If a block is clear when the path generation (or advancing movement of the path end point ahead of the AI train, however you prefer to think of it) arrives at it, the path is continued thru the block to the next signal. If another train enters the block after the path has already been extended thru it, the AI train will run the red signal regardless of whether it's an interval or an interlocking, permissive or absolute. Again, it's not stopping for anything but that end of path point. This is why sometimes two AI trains will appear to stop for red signals, and other times go thru and collide with each other, it depends on when the opposing paths were extended ahead of the two trains. Paths meet and stop, both trains stop, paths overlap, bang-shang-a-lang."
I am looking at my test setup where in the middle of nowhere I have set up a signal that is programmed to be permanentally (??)
red - there is no other traffic in front of it - the player (me) is sitting on the observation track watching (no interference to the AI path(s) at all.The first AI spawned and has stopped at the signal as it is supposed to - what is happening now is that AI's are spawning behind it and each is stopping at successive signals back down the track. Any minute now the activity will crash because an AI will spawn on top of one that has nowhere to go.BTW What I need is an engine (to scoot up and down the observation path) that accelerates quickly, stops VERY quickly and is bl**dy quiet - because sitting watching AI's for ages at a time listening to something like the little Speedster with the sick VW engine is very annoying..........Cheers,
Laurie Whelan
Remove | Alert Edit | Reply | Reply With Quote | Top
rgarber (792 posts)
Dec-10-01, 08:27 PM (EST) 9. "RE: Sorry Jim, But you are wrong........" Oh yeah? I didn't realize you meant you already made up the test track. Can you do a picture of the activity editor. I'd like to see that test track. Maybe too if you could package it up and send it. Sure would be great if we're all using the same test track.One thing... I believe the mystery of the signals maybe in the reason for the two sets of interlockings. I haven't been able to test this. But I noticed in Marias Pass that they use the single head (but you can make it into 3 heads) when it's not a passing siding. When it is a passing siding, then they use what we know as the double head interlocking.For instance, when there's a switching over from one track to the other, they use the single interlocking (w/ multiple heads), but on the passing siding, on each end is a double head interlocking.There's gotta be a reason for that. Can you set up your test track to include a couple of stretches of this?Major thanks Laurie!Rich Garber Remove | Alert Edit | Reply | Reply With Quote | Top
lwhelan (37 posts)
Dec-10-01, 08:59 PM (EST) 10. "The test track...." ... I hope, Rich. Haven't tried this before.
Not what I wanted - how does Jim put those pictures in the middle of his 'ramblings'?Never mind - this will do, I thinkCheers,
Laurie Whelan Remove | Alert Edit | Reply | Reply With Quote | Top
lwhelan (37 posts)
Dec-10-01, 09:18 PM (EST) 11. "RE: Sorry Jim, But you are wrong........" Rich,Do you know roughly where on Maria's Pass these signals are?If not I can see if I can see if I can find them later and set up the test track to suit.Picture is a bit hard to see what is what but basically it is two track all the way with a four lane section in the middle to test passing paths - still trying to fathom those....AI's spawn bottom right hand end travel up right hand route to the top right hand end, reversing point, across top of wye to reversing point left hand side then back down to bottom to terminate. Spawned at three minute intervals they back up down the track waiting to use the top wye. Switch the return track over to the other side to test passing path etc.And one more for Jim - are you listening Jim?We BOTH just learnt something - in your paragraph:-"3. The AI dispatcher directly checks the occupation state of each block as he comes to it, if the player TRAIN is in the block, it will stop the path. Note I say "TRAIN", in this case a loose consist does not count as an obstacle. If you uncouple one or more cars in the block, they are no longer part of your train, so once the end of your remaining train clears the next signal, the AI dispatcher considers the block "clear" in spite of the red signals at both ends caused by the loose consist sitting in the block. The AI path continues thru the loose consist, the AI train runs the red signal, bang-shang-a-lang."I thought this was true too - but alas......
Just did two more tests - placed a loose consist in the AI's path and it stopped at the signal before the block.
Tried again, this time - dropped last two cars of player train then high-tailed it out of there to clear block with player train and next AI stopped at light.***** BUT PLEASE NOTE *****
With the standard default signals only the Interlocking ones will completely stop an AI. This is because the Interval ones, as standard, NEVER, NEVER completely stop a train. Player or AI because the code is commented out by KUJU or whoever.
I have changed this............Enjoying this,
Laurie Whelan
Remove | Alert Edit | Reply | Reply With Quote | Top
rgarber (792 posts)
Dec-10-01, 10:55 PM (EST) 12. "RE: Sorry Jim, But you are wrong........" Ahhh... very good Laurie, you are good. I didn't spot 3 in the 'ramblings'. 3 is definetely incorrect. Leaving any loose cars on the track and a path goes through will stop trains for miles. Turned switches or loose consists will give the player or the AI train a red signal. Permission will be denied. In fact I just finished testing an activity where I believe this is why I couldn't finish the activity. I got a red signal and I couldn't pass through. Miles ahead is a load of cars sitting on the path.Okay, now... in regards to Marias Pass. What I did was fly along the tracks starting at the originating point and look at the signals along the way. In the large town area at the very beginning I didn't pay too much attention. But as you keep on going further on you will notice switch setups where each track has a set of switches to cross to the other track for both directions.These signals use the 1st BNSF single head interlocking, with the check mark for two heads. Every time.Alot farther up the track, I believe near the big.. you will come to the regular two headed interlocking we typically use. I've only seen these used in passing siding situations on the Marias Pass. Every time. You'll find one at each end of the siding with two opposing single interlockings.It may be just something simple. I dunno. I noticed the single interlocking when after having chosen a multiple head, will allow to route itself past a path point, where the double interlocking we typically use, only allows up to the 1st path point. I don't understand the ramifications of this as this was something I noticed after I finished the Metro.Rich Garber Remove | Alert Edit | Reply | Reply With Quote | Top
sniper297 (1027 posts)
Dec-12-01, 02:00 PM (EST) 13. "RE: Sorry Jim, But you are wrong........" Rolaids, quick. Ahh, much better. So much food for thought gives me indigestion.
Well, I hadn't had a collision with an AI train in quite a while so I began to suspect you were right. Sometimes. Maybe. It depends. Then I was adding interval signals on the AI passing routes to reduce the amount of time the player has to wait for an AI train (if there are still six passing sidings between me and him I don't want to be stopped at number seven and have to wait 15 minutes for him), and suddenly, bang-shang-a-lang. Not only were they blowing redlights and coming at me like Kamikazes, I lost all that ability I had before to back out of a deadlock, because now the switches at both ends of the single track were once again throwing at the same time, trapping me on the single track section with that old no-way-out-game-over garbage.
So. Being a Cancer with Libra rising, Venus in Taurus and Moon in Saggitarius, gives me sensitivity that sometimes these things are affected by astrological transits. Figuring that Saturn must be in opposition or something, I gave a call to the Psychic Friends Hotline, and my Personal Psychic said, "Hey, bonehead, did it ever occur to you that it might make a difference which way the little red triangles on those double-faced interval signals were facing?" Wow. She really is clairvoyant. Sure enough, I was adding the second face to make it a two-way interval signal without even thinking about the fact that the triangle only faced one way. So the trick is to face the triangle for the interval just beyond the switch away from the switch, everything works again, AI trains stop when it's red, switches throw one at a time, thank you Psychic Friends. (I asked her about the front coupler bug but it clouded up her crystal ball, and when I asked about a KUJU patch for the game it went dark.)
The following pics have the interval in question slid over a lot closer to the interlocking so it fits in the picture, but no matter how far away it is the principle is the same.
This does not work.
This does work.
Anyhow, that's where I'm at, and I'm still working on the test route. Laurie, you gonna upload yours too? Probably wouldn't hurt to have a few different ones so we can look at this from all the angles. Remove | Alert Edit | Reply | Reply With Quote | Top
rgarber (792 posts)
Dec-12-01, 02:40 PM (EST) 14. "RE: Sorry Jim, But you are wrong........" LAST EDITED ON Dec-12-01 AT 03:07 PM (EST)GooOOLLLY!!! No fair calling in the psychic hotline friends! That's like using a cheat. Well on this one Jim.. I gotta take a stand. *turning to face the audience* What we have here is a christmas tree decoration fetish. For goodness sakes Jim, you got more signals at the one switch than a dog has fleas! No wonder everything is stopping. They're trying to figure out what to do next! Signals in the middle of a switch? FOOUUULLL!!!! Penalty box 3 minutes, cross switching. This madness has just gotta stop! This is a compulisive passive aggressive behavior pattern. Think of your dad! Think of your grandpap! Come back to railroading!!! This signal populating fetish has gone to far! Perhaps a signal exorcism is in order. Also my toes are curdling cause you got them red sensors right on the switch... everybody in tinseltown knows that's a no-no. Rich Garber Remove | Alert Edit | Reply | Reply With Quote | Top
sniper297 (1027 posts)
Dec-12-01, 06:32 PM (EST) 15. "RE: Sorry Jim, But you are wrong........" LAST EDITED ON Dec-12-01 AT 06:40 PM (EST)You been taking sleeping lessons from Bo? Where is Bo, by the way? No offense, Laurie, I like you and all, but Boweavel always livened these things up...at least when he was paying attention. he did such wierd things with the route editor he actually made me look normal in comparison!
Rich. Sit up straight. Drink coffee, shake cobwebs out of brain. No slouching."The following pics have the interval in question slid over a lot closer to the interlocking so it fits in the picture, but no matter how far away it is the principle is the same."Yo comprende now? All that stuff is normally spaced out a lot further, I just crammed it together for illustration purposes. Marc was making fun of our communication problems in his Super Chief thread, and altho I usually can translate my Chicago "English" (if that's actually what we use here) into language that even Australians can understand, sometimes I wonder what you actually habla in the foriegn country of North Cakkelakky. I had a girlfriend about 30 years ago from Edonton, and she sounded like she spoke a sort of English, anyway. Maybe you've invented your own language now.
You do have a valid point, tho, Johnny Reb, one consideration that might need to be addressed is to bury most of the signals on the AI path underground, they would still serve the purpose of controlling the AI trains without making it look like a dragstrip. The other consideration is the speed of the AI trains, I noticed in all my Frankenstien laboratory fiddling with this that an AI train will slow way down for a bunch of succesive signals set close together, and when you're mixing different kinds of AI trains it's hard to space them out (3 minutes, Laurie? Ouch, I had such a hard time bucking the traffic at 5 minutes I generally use 8 or 10!), because the faster ones catch up with the slower ones. Start and end speeds seem to have no effect on the cruising speed of the AI train, you ever found a way to control the speeds in the activity editor? Maybe placing wait points for the faster train would work, even if the seconds always reset to zero it still stops momentarily. And how come I can't get an RS-2 to run as an AI train at all? It don't move!
Remove | Alert Edit | Reply | Reply With Quote | Top
rgarber (792 posts)
Dec-12-01, 07:11 PM (EST) 16. "RE: Sorry Jim, But you are wrong........" Jim... I know this is a little off the beaten path here.. but .... how do you do that? Reading one paragraph of yours is like listening to somebody who doesn't stop to breath. I find myself needing a bit of oxygen when reading one of your ramblings. It's like there's no periods or nothing. One sentence right into the other no stop for Go, no collect $200. I'm all wore out. My tongue is hanging out, I'm gasping for air, my eyes are darting back and forth... it's a gift you got there fella. I mean a gift! Okay, well.. back onto the beaten path. Uh... can't think now... what was we talking about.. Oh yeah, signals. Well.. uh.. IT STILL AIN'T RIGHT!! NO ... NO... NO... and... NO! And besides.. how do we know you squeezed them all in for the picture.. maybe you're just saying that to please your girlfriend, eh?????Anyway, it'll take a rocket scientist to figure out what you've done. You might even get a patent on your technique. Well.. I guess you need to get back to work.. Rich Garber Remove | Alert Edit | Reply | Reply With Quote | Top
lwhelan (37 posts)
Dec-12-01, 08:14 PM (EST) 17. "Yes Jim, How do you do that................?" Hi Jim and Rich,Here I was trying to digest which way the red triangles point etc, and suddenly something about an RS2. What's an RS2 - I think I got ones of those somewhere. Turns out I got two of them but one of them is in there twice but in a different folder. Anyway I thought I'll try and put one in as an AI (you see? Completely lost the plot now! Wandered off on some dark and dismal byway..........)But wait!! This is very interesting.......I put a consist in as an AI (two consists already set up: "RS-2 BN 4002 two units" and "RS-2 SP&S 61 two units" and started a test activity I was using yesterday and all started well the RS2's materialised and started to move down the track as I was also. Then it appeared to stop - I thought at signals to wait to let me through but my light was red. What's going on here? Maybe I stuffed up the script programming somewhere - two reds into the same junction - but his / her signal was green I discovered and guess what? The pair of RS2's are still moving but ever so slowly. Six minutes so far just to get round a 250 20d curve.Well, I will leave that to more knowledgeable people than me about engines and stuff.......... But it is very weird.About the test track - have added a section alongside existing trackwork to simulate Jim's layout of one track to two back to one then back to two etc - so had to rip up all the signals and start afresh after laying the new part. Any change to trackwork and I find I have to first rip up all the signals else problems appear.OK the RS2's just took 13 minutes to negotiate a 90 degree curve, scooted along the straight bit and has now vitually stopped again at the next curve.Enough for now........Cheers,
Laurie
Remove | Alert Edit | Reply | Reply With Quote | Top
rgarber (792 posts)
Dec-12-01, 09:05 PM (EST) 18. "RE: Yes Jim, How do you do that................?" Oh wait a second.. I know what's going on with the RS2's.. I'll bet you guys got an rS2 with the engine file messed up.. oh gosh.. what caused that?? It's something in the file spelled wrong I think.. let me go check and see if I still got it... *pause* I can't find it right now. But there's something in the RS2 engine file that's not right and that's why it's messing up.. so could you guys.. QUIT SCREWING AROUND and get back to SIGNALS!!!! Rich Garber Remove | Alert Edit | Reply | Reply With Quote | Top
lwhelan (37 posts)
Dec-12-01, 09:16 PM (EST) 19. "RE: Yes Jim, How do you do that................?" Sorry Boss.
Remove | Alert Edit | Reply | Reply With Quote | Top
sniper297 (1027 posts)
Dec-12-01, 11:24 PM (EST) 20. "RE: Yes Jim, How do you do that................?" LAST EDITED ON Dec-12-01 AT 11:26 PM (EST)You need a vacation, RAGs to RICHes. Screwing around is how you maintain your sanity. And I haven't got much to maintain, so I need all of it. I just spent an hour and a half screwing around, driving up against the traffic, drop off pick up, run back down with the traffic, pick up, do some switching, all this with a caboose on the end to complicate everything (that's something else I think there was a fix for, losing the cupola view after uncoupling), run back up the hill again, and after an hour and 40 minutes I got caught in a catch-22 on the South pass high bridge just before the tunnel into 77. I had to relay the track there because I found with the overhang on the auto cars it DOES matter if the edges of the ballast overlap a little (you ever try laying doubletrack curve sections on a grade?). I checked after the game over stalemate, but the next interval signal on both sides was facing the correct direction. Admittedly I was daydreaming, sightseeing, and taking screenshots, so it's entirely possible the first switch was already set against me when I went thru it and I ran a red. Touristing when I shoulda been engineering. I'm gonna spend a little time testing that section, but I think it was more likely inattention to duty. Hmm. This don't look good, bad signal layout or bad driving, either way guess who? Can I plead to a lesser offense and get probation? Remove | Alert Edit | Reply | Reply With Quote | Top
sniper297 (1027 posts)
Dec-13-01, 00:38 AM (EST) 21. "RE: Yes Jim, How do you do that................?" AHA! I was looking for some other stuff, and stumbled across this, so just for fun;
http://www.train-sim.com/dcforum/DCForumID6/1740.html Remove | Alert Edit | Reply | Reply With Quote | Top
rgarber (792 posts)
Dec-13-01, 00:49 AM (EST) 22. "RE: Yes Jim, How do you do that................?" That was in the days when I knew something... heh heh or thought I did....Rich Garber Remove | Alert Edit | Reply | Reply With Quote | Top

Lock | Archive | Remove

Lobby | Topics | Previous Topic | Next Topic