Driving to EOT Trackmarks in TANE

While developing a new industry in TANE, I ran across this interesting problem. When trying to have the AI drive to a trackmark near the end of a spur, the driver wouldn’t even move.
The loco is on a track headed to a junction that will take it to the end of the track (EOT) trackmark. After reaching the trackmark, the drive is to reverse and travel up the parallel track. Here is a diagram of the set up (with "L" the locomotive, "JUNC" a junction, "T" trackmarks and, "EOT" the end of the track). Yes, poor diagram but you should get the point.

L |
T |
| |
| |
| |
| |
\ |
JUNC
|
|
T
|
EOT

The driver had unloaded a string of cars and driven to a trackmark, uncoupled from the cars and then just sat there not moving. I checked the track to make sure there wasn’t something wrong by manually driving the loco to the trackmark near the EOT and, no, there weren’t any problems with the track. After several frustrating hours working with several (failed) ideas on how to fix the this, I finally decided to build a test route to work on the problem.

After much trial and error (lots of error), the problem turned out to be the distance of the locomotive from the EOT or the distance, D, shown in the diagram below.
LOCO==================TM======EOT
===<- --------------------D---------------------------- ->

After running several tests with the loco at varying distances from the EOT, I found that the loco would start moving toward the trackmark if D was approximately 700 feet (~213 meters for the metric crowd) or more. At that distance, the loco would creep toward the trackmark with a speed of about 2 mph. With D any less than about 700 feet the loco wouldn’t even start to move which was the case with the industry I was working on.

Well, if you have lots of room and time (I can walk at 4-5 mph so I’d beat the loco to the trackmark with hardly any effort) on your route to have 750 feet of track past the last trackmark that you’ll be running your trains to, that will work fine. However, if, like a lot of industrial spurs, that amount of track isn’t possible or practical, you’ll have to get kind of creative to get the AI drivers to work correctly.

Although there are probably many solutions to this problem, this is the solution I came up with:


  1. Extend the track work, using invisible track, a good 700+ feet past where the existing EOT is located.
  2. Place an invisible signal at the new end of the track.
  3. Place some type of scenery asset as a bumper at the visible EOT. It’s important that the bumper be something other than a "track object" because the AI must not take the “visible” bumper as a real EOT signal and start working with it rather than the “invisible” signal at the real EOT.
  4. Place the EOT trackmark at a reasonable distance from the “visible” EOT bumper, say 50 feet or so.
  5. Set speed limits on the spur such that the driver will have plenty of time to stop the locomotive on the EOT trackmark and not over shot onto the “invisible” track. Nothing will happen, of course, but having the locomotive moving "through" the scenery bumper and past the visible track work (or, in my case, into a hillside) will be “interesting” if the loco doesn’t stop in time.

I tested this solution with my industry and everything worked fine.

Again, there are probably many solutions to this problem in TANE but I couldn’t think of any but the above. I don’t remember having this problem in TS12. The drivers would creep toward the EOT trackmarks but I don’t remember the drivers not starting to move to the EOT trackmarks inside any certain distance.

Getting the AI drivers to do what you want can be a challenge (understatement) in Trainz. Developing this solution taught me more things about TANE which, in the end, is a good thing.

Take care,
 
You will always need a signal at the end of a spur. Most buffers are in fact signals and a train will stop at a set distance from the signal. To get closer to the EOT a second invisible signal a short distance beyond the buffer allows the buffer to show caution allowing the train to get nearer, and reduce the radius of the TM will allow for even closer stopping to the EOT.
 
Sorry, stagecoach, but I've have run the test with what you have suggested before and just ran it again with the same result as I stated in the original post.

Yes, I always but a signal, be it a bumper, buffer, or some sort of signal, at a end of a spur.

Yes, a bumper, or buffer, signal will show caution if put in front of a EOT signal. In Driver, you can see the signal state in the Map view. However, the situation I pose in the original post still occurs. The EOT signal has to be better than 700 feet past the EOT visible trackmark to have make the AI work in any reasonable action.

Thanks,
 
If the EOT is the only option for the engine to go to complete another command you shouldn't need a trackmark., unless the engine wants to go back the same way it came in to the industry...Which I have seen before do requiring you to need a TM but I also found that some older or poor engine files were the root of such issues with AI s...Dave
 
In previous versions of Trainz, a longer extension of invisatrack past the bumper, with 3 invisible signals, usually made a reluctant AI train venture to the end of track, also does placing "Signal Thingies" signals, and trackmarkers out beyond the EOT bumper on the long invisatrack
 
Last edited:
I just got done working up a test route in TS12 and the AI Driver had no problems whatsoever driving to the trackmark near the EOT signal regardless of the distance the locomotive started from the EOT signal.

I've run the same test in TANE on several different routes (to make sure the phenomenon isn't route specific) and, no, it is not route specific. Place a loco inside the ~700' distance to the EOT signal and the loco will not move (the "!" icon finally comes on the Drivers picture when the confusion is total). Place the locomotive anywhere outside the ~700' distance to the EOT signal and the loco will travel to about the 700' distance, slow down to a crawl (~2 mph) and take about 4½ minutes to get to the trackmark . . . but it does get there. I've also installed the fix I explained above to some of the problem spurs in my routes and they work like a charm after the fix.

I'm thinking of reporting this as a bug mainly because I can't believe this problem is unique to my version of the game. But stranger things have happened.
 
I have found the same situation (slow speed) when the AI is coupling to a consist and starts at a distance outside the 700' distance. The loco slows to 2 mph when it gets to the ~700' distance until very close to the consist and then speeds up somewhat to do the coupling.

I'll be submitting a bug report on this issue.
 
Here are some oldies but goodies that I use in TANE with spectacular results for doing set-outs and pick-ups. This is a listing of how I use them.

Nav or Drive to TrackMark (short of trailing point junction)

<kuid:401543:1090> Halt Train - waits until train stops

<kuid2:66277:80005:2> UnCouplezFrom - uncouples behind noted car, (start your count with zero)

<kuid2:131986:1018:1> SetDCCThrottle - set speed and direction using numbers from 1.00 to -1.00. Test your locomotive and find what numbers work best.

<kuid2:131986:1017:1> DrivePastJunction - When last vehicle in train passes junction, next command is implimented. If you are pulling a string of 20 cars to set out, or just the locomotive to make a pick-up, your train stops shortly after it clears the junction. Use "exit schedule on "Leave" option.

<kuid:401543:1090> Halt Train

<kuid:131986:1011> ControlJunction - sets junction or multiple junctions to siding

<kuid2:131986:1018:1> SetDCCThrottle - this would be a negative number between -0.00 and -1.00 to back the train onto the siding or spur.

For a Pick-up, I place a trackmark at the junction end of the siding or spur and then drive to that TM at a higher speed than coupling.
<kuid2:131986:210023:1> TCDriveViaTrackmark

<kuid2:131986:1018:1> SetDCCThrottle - this would be a negative number for coupling speed.

<kuid2:131986:1015:2> CoupleAhead - couples to first car found (at speed set in previous command)

<kuid:401543:1090> Halt Train

<kuid2:131986:1018:1> SetDCCThrottle - number between 0.01 and 1.00 to move forward

<kuid2:131986:1017:1> DrivePastJunction

<kuid:401543:1090> Halt Train

<kuid2:131986:1018:1> SetDCCThrottle - negative number between -0.01 and -1.00

<kuid2:131986:210022:1> TCDriveToTrackmark

<kuid2:131986:1018:1> SetDCCThrottle - this would be a negative number for coupling speed.

<kuid2:131986:210023:1> TCDriveViaTrackmark

<kuid2:131986:1015:2> CoupleAhead - couples to first car found (at speed set in previous command)

<kuid:401543:1090> Halt Train






Continue on your way . . .

These controls ignore the signals and allow all kinds of switching maneuvers. And they work in TANE ! !

David
 
Last edited:
After sumbitting the issue to the Help Desk, I got an emai back from them asking for the test route so they could see if they can replicate my results. As I got the route in shape to send them, I replaced the locomotives with new ones and ran the session to insure it functions ok. Lo and behold the locomotive drove to the EOT trackmark no matter where the locomotive started; i.e. the 700' distance didn't matter. What?

I was about ready to respond back the Help Desk, "Never mind !" But I realized that I hadn't really duplicated my origingal testing. I replaced the new locomotives with the orignial ones and, yes, the problem was present as I explain above.

The original locomotive in the test route (there were different ones on the other routes) that has problems with driving to the EOT is a UP GP9 (kuid2:506034:120170:1)'

The locomotive that didn't have any problems is an ATSF SD45 (kuid2:45324:100140:2).

So, in other words, the problem is an asset related one.

We'll see what the testers come back with and whether or not we have to resort to my fix in some cases or use David's commands for certain locomotives.

Thanks,
 
Back
Top