Design choice (and if so, why?) or bug: Decoupled car ignores gravity when pushed

[oknotsen]

This is the situation:
A train is pushing a few cars slowly up a hill (say 5km/h or 5m/h). The one on the front gets uncoupled. The train reaches the top of the hill and the first cars go over it.
Now one would expect gravity to get a hold on that first car, making it to get more speed and getting away from the rest of the train, but in Trainz12 it does not. The car just stays connected with the rest of the train. Even when you hit the break, the whole train stops, including the decoupled car; gravity totally being ignored.
When you start pulling the cars back up the hill, the decoupled car stays there; one could make the assumption that someone has put the breaks on. This would also explain why it does not start running down the hill; the breaks keep it pinned against the pushing train. ... but when one brutally reverses the train, the decoupled car sort of gets "shaken off" and then starts speeding down the hill, which would mean there is not any break applyed.

My question(s) is (/are):
Is this a design choice? If so, can someone please explain why so I can figure out if there is a way to by-pass this with custom scripting. Aim is to eventually make a working hump yard.
Or has this been a long open bug? If so, when are you planning to finally fix it and is there a good work-around (not being the "shaken off" way) till then?

 

[fran1]

Have a look at matlin hump yard by Leefer, no idea whether it will suit.
Just found demo hump (dcc or cab) by llj.
Driver command uncouple all by brummfondel, might also help.

 

[Bill69]

Try uncoupling just over the top of the hill while the train is still moving. When you stop the train the brakes are applied to all coupled cars. It works in TS2009 and should still work in TS12.

Cheers,
Bill69

 

[danny5]

That is the way it is. you need to reverse your train in order to completey uncouple from the car,and then it will roll down the hill.
When you make your hump yard use the hump tower by llj it will work inTS12.
Some advice is to put two to tree letters only in the junction names.(like HY1,HY2)
Unless you use both ends of the yard you will need to set just the hill side junctions.( no retarders no signals)
Need more help let me know.

 

[oknotsen]

Thanks for the replies so far and for the work-around / related items created by others.

I know it is weekend in Australia, so they probably dont give a reply yet to this topic, but I hope someone else knows the answers to the main questions of this topic:
- Is this a design choice? If so, can someone please explain why so I can figure out if there is a way to by-pass this with custom scripting.
- Or has this been a long open bug? If so, when are you planning to finally fix it?

 

[maruffijd]

From what I have seen doing this, I notice that the hand brake on the decoupled car(s) will be set when you stop first. So I believe this is a built in feature.

 

[Jananton]

Hmm, afaik the train brake will set the car brakes on in normal situations so the brakes are on when decoupling the car so it doesn't run down a slope. Obviously it seems a decoupled pushed car isn't seen as decoupled by the trainz game engine at this point. Have you tried the same movement using the engine brake (E key toggle) only to see if this will free the uncoupled car as you expect, since this in theory wouldn't apply the brakes to the pre decoupled car?

Greetings from nighttime Amsterdam,

Jan

 

[bobhahn]

This is from a session I did called "Feed the Feedlot" for MSGSapper's route "Watseka & Kankakee Railroad" (both are on the DLS):

In the picture you'll see the hump tower, a switcher with a few cars in the background, and two track markers named "Humptop" and "Humpbottom".

When triggered by the driver's loco passing by (this hump yard shtick is merely eye candy) the switcher's AI driver executes the following commands:

Navigate via Humptop
Uncouple (first car)
Navigate to Humpbottom
Wait 5 seconds
Navigate via Humptop
uncouple (next car)
Navigate to Humpbottom

The AI takes care of stopping and reversing the train at Humptop without setting the brakes on the released car. It rolls merrily down the hill into the yard.
Note the position of the Humptop track marker a little past the "summit" of the hump.

 

[Bill69]

You can also uncouple a car at the hump bottom trackmark and push it up the hill without it recoupling so long as it does not move away from the train when uncoupled.

Cheers,
Bill69

 

[oknotsen]

So far, a bunch of sollutions / quick fixes / work-arounds have been posted. Thanks for that. All nice, yet nobody answered my main questions:
- Is this a design choice? If so, can someone please explain why?
- Or has this been a long open bug? If so, when are you planning to finally fix it?

I am trying to understand this, not looking for more tricks.
Once the understanding and right qualification is there, then (and only then) I can try to break my head on seeing if I can find another sollution or pick one of those already offered.

 

[Bill69]

Yes it is a design choice, the program is set up to apply brakes to all coupled wagons when they come to a complete halt. In addition you will find you need two clicks or more of the throttle to relese the brakes. In earlier versions of Trainz you had to apply the hand brake 'A' key to leave a consist on a slope when stopped. Now the hand brake applies automatically provided you come to a complete stop. That is why I said in my first post "decouple while still moving".

cheers,
Bill69

 

[oknotsen]

Okay, thanks Bill. Then I probably didnt totally understand you the first time, but now I do.

This is probably the moment this turns into a topic for Content Creation Support.
Because of this handbreak, I did fool around a bit with custom code build into a trigger that I placed over the top of a hump so that it does not trigger untill a car is a full length past the top on a 4+ gradient. Here is a code snippit (I wont bother you with the whole code):

vehicles[i].Uncouple(1);
vehicles[i].Uncouple(-1);
vehicles[i].SetHandBrake(false);

The result of the code is an uncoupled car (tested with success) but since it doesnt speed up, I fear the handbreak is somehow still applyed.
This basicly means that either my code is still wrong (which at this moment I consider perfectly possible) or we do seem to have a bug... or at least some unexpected behaviour.

 

[Bill69]

The speed of the uncoupled car will depend on the weight + gradient. By a 4+ gradient do you mean 1 in 4 or 4%? Trainz gradients are usually set by percentage.


Cheers,
Bill69

 

[oknotsen]

By a 4+ gradient do you mean 1 in 4 or 4%? Trainz gradients are usually set by percentage.

To be honest: I was not sure which of the 2 it is. I mean the one you can set / read in trainz, so I assume it is a percentage.
I also tested this with 11% and still the cars stayed attached to the pushing train.
It has been a few years since I was in school, but my guess is the car must be pretty light with some decent resistance of the wheels for any car to still not start rolling without breaks.

 

[Jananton]

Hmm, if set / read in surveyor the gradient value is in percent.

As for the code, I'm certainly no scripting expert and I don't know what you've omitted, but as soon as car(s) are decoupled they become a consist on their own and then it depends on the default behaviour when on creating a new consist the handbreakes are on or off. Personally I would look for an event message that's sent when a consist is created, if such a beast exists and set the breakes to false there.

In the TS12 section of the forum is a same kind of thread about kicking cars, maye of interest for you too.

Greetings from sunny Amsterdam,

Jan

 

[wreeder]

There was a very long thread either here or at the now gone trainzdev where a user did many tests of trainz physics vs real world physics. As I recall his conclusions was that Trainz physics didn't work like real world physics as far as non-powered rolling stock is concerned. As to whether you can script it to behave differently, that is a question for the programmers.

William

 

[Bill69]

Hi oknosten,

I have just hashed up a small hump yard in TS2010 that works. It takes about 15min. to run through, splitting up a consist that supposedly has just arrived in the main siding. I don't know what rollingstock is available in TS12 so it might be usless sending it to you, but I can give you a list of the commands and the order they are in if you would like me to. The other option is give me a list of the rollingstock you are using and I may be able to set it up using the same rollingstock and loco. The maximum incline in my hump section is 2% and I use retarders near the end of the classification sidings to avoid having cars run off the end.

Cheers,
Bill69