Tires

[osiris]

it's a tire of Ferrari 308 or 328 according to http://www.miata.net/sport/Physics/

But there is no motion of the moment so the moment comes from vamos http://vamos.sourceforge.net/vamos-docs/Cars.html#Cars
I have just switched and I don't have a lot of time to test them but other tires on the other topic are the one I've tested and they are good(values are alike).
I think that it's good for powerful cars, I've tried new tires on the T73 AX2 and G4 with satisfaction.(but the topic isn't over)

# Lateral force
a0=1.799
a1=-0
a2=1688
a3=4140
a4=6.026
a5=0
a6=-0.359
a7=1.0
a8=0
a9=-0.00611
a10=-0.0322
a111=0
a112=0
a12=0
a13=0
# Longitudinal force
b0=1.65
b1=0
b2=1690
b3=0
b4=229
b5=0
b6=0
b7=0
b8=-10
b9=0
b10=0
b11=0
b12=0
# Aligning moment
c0=2.07
c1=-6.49
c2=-21.9
c3=-0.416
c4=-21.3
c5=0.0294
c6=0.0
c7=-1.2
c8=5.23
c9=-14.8
c10=0.0
c11=0.0
c12=-0.00374
c13=0.0389
c14=0.0
c15=0.0
c16=0.639
c17=1.69


According to http://www.miata.net/sport/Physics , values are simplified.

[osiris]

# Lateral force
a0=1.1
a1=-0
a2=1688
a3=4140
a4=6.026
a5=0
a6=-0.359
a7=1.0
a8=0
a9=-0.00611
a10=-0.0322
a111=0
a112=0
a12=0
a13=0
# Longitudinal force
b0=1.0
b1=0
b2=1690
b3=0
b4=229
b5=0
b6=0
b7=0
b8=-10
b9=0
b10=0
b11=0
b12=0
# Aligning moment
c0=2.07
c1=-6.49
c2=-21.9
c3=-0.416
c4=-21.3
c5=0.0294
c6=0.0
c7=-1.2
c8=5.23
c9=-14.8
c10=0.0
c11=0.0
c12=-0.00374
c13=0.0389
c14=0.0
c15=0.0
c16=0.639
c17=1.69

Is tires I'm using, old ones are just the ones I've taken from somewhere, they aren't good(I don't know if my values are real or not but the result seems real to me so it's te important to me)

[reece146]

I did a little casual reading up on these curves and such.

If I am following this correctly, we need to develop a LUT of these datasets to correlate to specific brands and sizes/aspect ratios, correct?

There is no way to "decompose" the size or aspect ratio directly for the data set you have posted, correct?

So, if the above assumptions are true, we need to start tracking down these curves for tires of interest.

Where to start? I've never seen these curves posted to tire manufacturer or enthusiast sites - at least not in the public domain.

If we had a decent set of curves for various size tires for a specific tire - say S-02s in a spread of 205 through 285 (at least) that could cover a lot of tire options for the cars in the game.

[osiris]

It's just what I'm testing.
I've just changed 2 values to try.
A0 and B0 with aproximately the same factor(to make nothing weird)
Where do you find a large amount of real values?

[reece146]

Where do you find a large amount of real values?

That's the bother... Short of contacting the tire manufacturers I'm not sure if that data is even available. And I'm going to guess that the tire manufacturers probably say it is "proprietary information".

The ax, bx, cx values are curve fits to the Prankja curves (I think). I'm going to guess that same make/model tires will have the same shape curve, sizing is what makes the difference in the ultimate amplitude of the shape (wildly speculating). There must be a way to model/generate these curves - even something "approximate" would help in our case. It's just a game after all.

[thelusiv]

The Racer community has struggled with this problem for quite some time, since they use the same tire model we do. Whenever a Racer car has been imported the tire values have come with it. Look through the Racer sections of the Race Sim Central forums, you'll find quite a bit about Pacejka constants.

There is no easy way to get these values. Many tire companies do measure them (I think) but don't publish them, and purchasing this sort of information can be very expensive, as far as I know. However, I know very little about it altogether, and there may be a way to abstract the formula and estimate the values. There are also Pacejka editors, Racer comes with a simple 2D one and I've found simple 3D ones on the internet before.

[reece146]

The Racer community has struggled with this problem for quite some time, since they use the same tire model we do. Whenever a Racer car has been imported the tire values have come with it. Look through the Racer sections of the Race Sim Central forums, you'll find quite a bit about Pacejka constants.

There is no easy way to get these values. Many tire companies do measure them (I think) but don't publish them, and purchasing this sort of information can be very expensive, as far as I know. However, I know very little about it altogether, and there may be a way to abstract the formula and estimate the values. There are also Pacejka editors, Racer comes with a simple 2D one and I've found simple 3D ones on the internet before.

I guess what needs to be done is to try to find a decent model of a car (Racer or other) and then take the curves to represent the tires that came on that car from the OEM.

If say, a `88 911 came with Dunlop D40 205/55VR16 front tires and the curves seem reasonable we can say in our look up table that D40 - 205/55VR16 is this curve. We'd just need to build this table at that point.

This assumes that the Racer model hasn't been teaked and is a valid set of curves of course.

It's something I want to take on eventually; culling this data out of the various forums and presenting the table.

Maybe an "ultimate" thing to do is to make a "vdrift spec" tire that has our own curves. I'm thinking ahead to the tuner/career stuff in the game.

[thelusiv]

We have plenty of good cars, and the Racer guys for the most part tested the different tire values a lot. However we still can't really know if our values are realistic or not without any real measurements of real tires.

Maybe if we had a good Pacejka editor/curve visualizer, it would be a lot easier to create/edit tire curves.

[joevenzon]

The only real values I've been able to find are these Racer forum posts supposedly of Alpine tire data:

http://forum.racesimcentral.com/showthread.php?t=45296

I think it'd be excellent to have a set of realistic tire parameters, but I think we should also generate some "arcade" type tires that make handling easier for people who don't want unforgiving, realistic handling.

[reece146]

Read that thread and the few threads referenced within.

It looks to me that we are more or less stuck with whatever curves we can find off the web and/or from informed contacts. I'll try shaking down some contacts within my local community.

Probably the most reasonable thing at this point is to come up with a way to draw a desired curve and then decompose it to the coefs. A software tool or "grip" method would be adequate at this point.

[joevenzon]

Anyone know of a good pacejka coefficient visualization/editor tool?

[cotharyus]

No. But I did find Pacejka's original paper on how this is supposed to work....

[thelusiv]

Interesting thread: http://www.lfsforum.net/showthread.php?t=7720
Interesting paper: "On the Dynamics of Automobile Drifting"

This guy wrote a Pacejka editor with a 3D display, in fact I contacted him a while ago and asked for the source and he provided it, but I never got around to porting it to Linux. The source is on this page though and it uses SDL, he provides a binary for Windows. I think we could use this to work on our Pacejka models for our cars:
http://www.i-gear.net/projects/mftk/main.php

If someone has time to port this it would probably help us a terrible lot.

[reece146]

Appears to be a lot intellectual "wanking" in that thread that was posted.

That said, there was one intersting link:

http://www.delft-tyre.com/mf_datasets.htm

Unfortunately you'd need to purchase the models.

[thelusiv]

There are some stupid posts in that thread, but there are some others that are very good.

Which reminds me, I'm glad our forums haven't been inundated with trolls like many other racing game forums...knock on wood...