Any ideas for how making a flexible border

[asasinuxp]

Any ideas for how making a flexible border aligned to road?

 

[Emery]

Do you mean a string object and the "align to track" feature? Or do you mean making an object that is flexible when struck?

 

[Erwin Greven]

A tire wall that works like a safer barier would be great.

 

[asasinuxp]

Do you mean a string object and the "align to track" feature? Or do you mean making an object that is flexible when struck?

No.The track is not straight,and I found a border(to use in a town),but i cannot align to track because is not flexible(it's a object).

 

[R Soul]

You might have more success with a wall. You can change the shape and materials to make it look like the object.

 

[asasinuxp]

Finally,it work's with a wall(after several attempts).

 

[Emery]

Excellent! Glad to hear you worked it out.

 

[Emery]

A tire wall that works like a safer barier would be great.

I've given a lot of thought to this. The best I've been able to come up with is a tire barrier that is not marked as a collision object and then an invisible wall stuck part way inside, so you'd drive into the barrier a short distance before actually getting a collision.

Wait! I think I've got a great idea on how to make it best work!

In the properties of objects (or is it just sign objects?), you can hinge it at the bottom, with a spring rate, damper, and mass. So you use two invisible walls, the first one being hinged with a modest mass and high spring rate. The second wall is fixed and constrains the first wall from bending too far and acting like a ramp. Or you could use three walls with the middle one having a heavier mass and spring rate.

Anyone want to experiment?

 

[Erwin Greven]

I would

 

[Emery]

Looking up tire sidewall spring rates, around 100 lb/in for an uninflated tire. You'd have several springs working in parallel, say 6, that make it 600 lb/in spring rate during a head-on impact as a starting point. A starting point for the mass should be on the order of 800 lbs, since the impacting car is affecting several stacks of tires at once?

Now you have to convert into SI units... and come up with a damper rate.

And then start experimenting... probably create several test walls with different spring rates and masses and dampers. Keep them visible so you can watch how they react in the replays. Try some without the second wall and a significantly higher spring rate.

I took a quick look and there are two likely approaches to testing methodically:
1) create fresh objects for Xpacker, perhaps labeling them with signs indicating the properties of springrate & mass
2) just twiddle the values in 3dSimEd

Questions needing answers:
a) What differences are seen when struck a glancing blow vs. head-on (glancing blows on real tire barriers 'suck you in' so that it spins the car around)?
b) What difference does speed have on the impacts?
c) What differences does mass (NASCAR vs. Lotus 7) & construction of the car (fenders vs. no fenders)?
d) How critical are the variable values?
e) When using two walls, how far apart should they be?
f) How badly are framerates affected?
g) For a long tire wall, is it better to have several spring walls or one long wall?

 

[Emery]

Aw, nuts... the physical properties of signs, posts, etc. aren't in the object other than specifying what the object should be (sign or post). Instead, they're in the GDB:

Made properties of loose objects (cones, signs, etc.) configurable through track GDB file. Defaults can be set for each type, and can be overridden for specific instances. Values can be specified directly (with positive values) or relatively (by using a negative number which will be multiplied negatively by the original default value). Entering 0.0 for a property is the same as -1.0: neither input will change the property's value. Note that the physics sampling rate for loose objects is somewhat low (to keep CPU usage down), so there are limitations to what range of parameters can be used. If you simply want to increase the mass of an object, we'd recommend increasing all the parameters except friction by the same proportion (e.g. if you want to double the mass, also double the inertia values and the spring and damper). Format is "<type or instance name> = ( <mass>, <inertia x>, <inertia y>, <inertia z>, <spring>, <damper>, <friction> )". Hard-coded defaults for each type are given below:

• CONE = ( 2.5, 0.25, 0.20, 0.25, 1425.6, 72.0, 0.80 )
• POST = ( 1.2, 0.5, 0.1, 0.5, 400.0, 40.0, 0.80 )
• SIGN = ( 5.0, 4.0, 2.0, 6.0, 1024.0, 44.8, 0.70 ) // default inertia parameters for signs are actually calculated based on geometry
• WHEEL = ( 15.0, 0.7, 0.5, 0.5, 5625.0, 45.0, 1.0 )
• WING = ( 10.0, 1.0, 0.5, 1.5, 3600.0, 60.0, 0.60 )
• PART = ( 10.0, 1.0, 0.5, 1.5, 3600.0, 60.0, 0.60 )

So just set the two walls in place, mark the springy one as a post or sign (whichever makes sense based on whether you have other sign or post objects which you don't want behaving like tire walls) and the other is a normal wall. Then edit the track GDB to try different values.

 

[mianiak]

Interesting find about the gdb file there, it's something that has had me baffled for a while now

Also what you can do for this sort of thing is,,
You can change the values of the tirewall in the .tdf file
or
You can add a new entry into the .tdf file, which then corresponds to a material name, all you have to do then is give that material this certain name and it will work however you want it to work.

The only problem is getting that object to have that material name, from memory I have had problems doing this in the past with XPacker.


ie,

( i haven't changed any values here, you will need to calculate them, this is just an example of what could be done)

// DEFLATED TIRES
[FEEDBACK]
Legal=false Spring=20000.0 Damper=10000.0 CollFrict=0.90 Sparks=0 Scraping=0 WallSkids=1 Sound=tirewall
Materials=dftwal

 

[Emery]

Here's some real world info about barrier design: http://www.grandprix.com/ft/ftpw016.html