New Tire Model

Vehicle physics, suspension, tires etc

New Tire Model

Postby Eugene Cojocar » Fri Oct 19, 2012 7:38 am

The new tire model is quite ready. It has advantages and disadvantages. The main advantage is that it allows defining various behaviors for the long/lateral forces and aligning torques. The disadvantages are that it's too complicated and requires much more computing power.
Previous version is improved also. Added longitudinal forces dependence on camber angle and improved combined slip behavior.
I call them MF1 (Magic Formula 1) and MF2 (Magic Formula 2). Now XMR supports 3 tire models. I think that's quite enough. I prefer to use MF1 because it gives good results and quite fast.

Just a comparison between MF1 and MF2 input data for the lateral forces on the pictures below.

MF1:
Image

MF2:
Image
XMR Developer
User avatar
Eugene Cojocar
 
Posts: 937
Joined: Thu Oct 18, 2012 2:49 pm

Re: New Tire Model

Postby Towers » Sat May 03, 2014 1:41 pm

Hi all,
I played with this last night making some tries with the skyline r34 tires, because I was searching for a solution about "unexpected" forces coming in the ffb while going straight, and I found this car showed me this problem less than others.
However if you want to try it out here is the physic file you have to select in the garage before going on the track.
I packed two file a pha and a phs (don't know if are both required), just unzip in your "Data->Cars->Skyline R34".
It's only a try so please comment and share your opinion about the car handling.

Thanks
Attachments
Skyline R34 tires test.rar
(66.93 KiB) Downloaded 138 times
User avatar
Towers
 
Posts: 82
Joined: Wed Oct 16, 2013 12:13 pm
Location: Italy

Re: New Tire Model

Postby Eugene Cojocar » Sat May 03, 2014 4:29 pm

A phs file is enough. I will try it today. Thanks for the sharing.
A pha file is a copy after tweaking in the game quick physics menu.
XMR Developer
User avatar
Eugene Cojocar
 
Posts: 937
Joined: Thu Oct 18, 2012 2:49 pm

Re: New Tire Model

Postby vit » Sun May 04, 2014 4:13 pm

Interesting setup. I'm not a physics guru and my answer may in fact be wrong.
vit
 
Posts: 6
Joined: Wed Feb 05, 2014 1:25 pm

Re: New Tire Model

Postby Towers » Sun May 04, 2014 8:50 pm

Well I'm not a physic guru too, I try to understand how it works and make tries... those tires loose too much, need a rework...

Thanks for your feedback
User avatar
Towers
 
Posts: 82
Joined: Wed Oct 16, 2013 12:13 pm
Location: Italy

Re: New Tire Model

Postby Eugene Cojocar » Mon May 05, 2014 5:39 am

I had a look at the tires too. The Shape Factor (C) (b0) for the Tractive mismatches real tires. It should go down more at its peak. For example, try to enter the b0 between 1.6 - 1.65. Otherwise the car will be like a car in NFS. You can't do burnout or get lateral sliding of the rear wheels on skidding.

As to the cornering properties (Lateral forces). In this case tires lose many lateral forces at slip angles more than 18 degrees.
Try to play with the a0 (Shape Factor (C)) between 1.3 - 1.5 values.
XMR Developer
User avatar
Eugene Cojocar
 
Posts: 937
Joined: Thu Oct 18, 2012 2:49 pm

Re: New Tire Model

Postby Chronus » Fri May 09, 2014 2:19 pm

Hello, Eugene. First post here.

My 2 cents here:

the TM has evolved a lot. Had the chance to try the Nissan GT-R35 last night and felt good. More organic, complete. (Even though this and other cars I have recently downloaded need a rework, which I'll do soon enough.)

In regards to tires that "lose too much" (whatever it is they lose)...

There's little point in questioning a tire model (or any physics model) until we have explored it deeply (with different physics calibrations) and find it wanting.

To explore the TM deeply, my experience tells me you need 2 things: hard data (and little more than a handful of people are or have been privy to racing teams/car manufacturers/tire companies data) and the proper methods/knowledge/tools. Lack of any of these (usually, ALL) is the cause for so many bad mods -physics wise- for Racer/rF/Race07/GSC.

God really is in the details. From pro sims (based on Deft Automotive's latest models) to ISIMotor2/ISIMotor2.5/iRacing/NetKar Pro or even pCars, good physics calibration can ruin vehicle dynamics, no matter how good the physics engine is.

From my experiences last night, nice work, Eugene.
Chronus
 
Posts: 18
Joined: Fri May 09, 2014 1:38 pm

Re: New Tire Model

Postby Eugene Cojocar » Fri May 09, 2014 3:36 pm

Hello Chronus,

Glad to see you here!
I had some tires data, but, of course it's hard to collect data of many modern tires. A lot of testing can help to improve handling.
Agreed that some mods must be re-worked especially in terms of tires and sounds. Hope they will improved once.
I can't comment pCars and its physics model and tweaking because never tested it. I know there are divergent opinions about it.

Please don't forget to update the XMR via UpdateMaster.
XMR Developer
User avatar
Eugene Cojocar
 
Posts: 937
Joined: Thu Oct 18, 2012 2:49 pm

Re: New Tire Model

Postby Chronus » Fri May 09, 2014 8:17 pm

Hello, Eugene. Glad to visit and exchange opinions with you and the fans of XMR.

I was very pleased to see how XMR's physics engine reacted to a couple of cars. There's a lot of potential here, I wish more and more people know about it (I have spend years modding ISIMotor2 sims and Racer, as well as digging into vDrift code, and XMR already feels very good in comparison).

I know you never comment on other sims (which is a very good sign), so I will just say the following:

pCARS and AC seem to be approaching tire physics from the physical modelling avenue. The risks, for anyone with a bit of knowledge and experience, are obvious and extensive, but hardly anyone dares discuss this. To be honest, rF2's quasi-SWIFT model is much more appealing to me and also more realistic and objective, problem is...modding for it (a nightmare).

This is where XMR begins to shine more and more to me: the TM is becoming better and better and your modding tools allow us to work comprehensively.

----

I have work cut out for me, I will modify and test "a pedibus usque ad caput" a few cars just downloaded and try them at several tracks.

[It's also noteworthy how lighting, gfx in general have reached a high level as well.]

---

Yeah, I'm going to update via UpMaster. Thank you for the heads up.
Chronus
 
Posts: 18
Joined: Fri May 09, 2014 1:38 pm

Re: New Tire Model

Postby Eugene Cojocar » Sat May 10, 2014 2:58 pm

Thanks, I hope to see your physics modifications once.
Seems we talked about it, but I prefer MF1 as more simple and convenient model instead of MF2.
I don't know how TM has been implemented in pCars and AC, but in XMR each car can have unique tire model. It depends on the creator of cars of course.

As to the gfx, it still uses DX9. I hope to port it to DX11. DX11 rendering is ready for 15-20 percent only - all core functions. I has to rewrite several thousand of shaders to complete DX11. I'm not ready for that yet.
XMR Developer
User avatar
Eugene Cojocar
 
Posts: 937
Joined: Thu Oct 18, 2012 2:49 pm

Next

Return to Physics

Who is online

Users browsing this forum: No registered users and 1 guest

cron