The online racing simulator
Tire physics
(88 posts, started )

Poll : Which curve is more realistic?

LFS current curve (in red)
48
My attempted curve (in black)
26
Quote from jtw62074 :Getting it proper might really take more than one curve.

Would in the transitional area of the curve there be a significant amount of randomness (noise) in relation to the percentage of elastic grip verses frictional grip also? This could possibly give real tyres a smoother transition (i.e. more "feel") also

Or just one curve with a certain percentage of random deviation applied only to the frictional component I'm way out of depth here so had better shut up before I make a total dick of myself but can't help myself I love thinking about things like this yes I'm geeky
#52 - JTbo
But couldn't you just modify one curve with formula to get different curves out from it? Then just interpolate everything that is left between this newly created and old curve, maybe something like morphing would be handy to make interpolation smooth, I really don't know what I am talking about, but got just this idea thing again hitting my head like some blunt object
Quote from Glenn67 :Would in the transitional area of the curve there be a significant amount of randomness (noise) in relation to the percentage of elastic grip verses frictional grip also? This could possibly give real tyres a smoother transition (i.e. more "feel") also

Or just one curve with a certain percentage of random deviation applied only to the frictional component

Not really. What you're doing with a graph like this is saying, "ok, right now the slip ratio is 0.2 (20% slip), so what's the force?" Now apply that to the tire or car and off you go. The transitional area is just the part that is rolling off into the peak after the initial climb. That's modelled just fine by measuring the tire forces and duplicating them with such a graph (or more typically, creating or using a math function that reproduces all those force data points on it given slip ratio).

When measuring a real tire there is indeed some high frequency noise, but as covered in an earlier post, you won't really gain anything from including it. It might actually make the car feel a little floaty and detached.

By using more than one curve, I meant that the curve at 10 mph might look very different from the one at 50mph in reality. Really though the issue people seem to bring up most nowadays is the "hey, I can spin the wheels like crazy and don't lose any acceleration" bit. That's good that people are now focusing on this. It means the rest of everything is very good and has come a long way!

Quote :
I'm way out of depth here so had better shut up before I make a total dick of myself

Nah, it's fun to discuss such things

Quote :but can't help myself I love thinking about things like this yes I'm geeky

As a fellow geek, I love it too
I assumed that the curves were being calculated on the fly for the different slip angles etc so there would be the mathematical function for the initial curve itself and another that would alter the curve for different speeds, temps etc and it was just a matter of tweaking the second function now but I guess that would produce unrealistic results as it would be affecting the whole curve rather than whats more likely happening in reality is the elastic and frictional areas of the curve are being affected in different ways So yeah several mathematical curves for different speeds sounds like a good solution
Quote from DaveWS :Ok, but I remember Tristan talking about the steering torque reducing a lot when he turned the wheel too much for the corner in his F3 car. And that this doesn't really happen in LFS.

Surely, assuming Tristan is correct, this would mean the lateral force DOES actually does drop off after the peak?

No, only the caster effect (the tendency to straighten out the wheels) is reduced, not the lateral force (or not by as much).
Quote from Glenn67 :I assumed that the curves were being calculated on the fly for the different slip angles etc so there would be the mathematical function for the initial curve itself and another that would alter the curve for different speeds, temps etc and it was just a matter of tweaking the second function now but I guess that would produce unrealistic results as it would be affecting the whole curve rather than whats more likely happening in reality is the elastic and frictional areas of the curve are being affected in different ways So yeah several mathematical curves for different speeds sounds like a good solution

That may be so. I don't know
Regarding longitudinal grip versus slip, there was a time back in the late 1960's and early 1970's where top fuel dragsters posted about the same ET's whether or not the tires were spinning all the way through a 1/4 mile run. You'd see smoke from the tires for the entire run on one car, virtually no smoke from the tires on the other car and the ET's would be very close.

This isn't true currently with fuel dragsters. Smoke the tires at all in top fuel, and the ET suffers greatly.

However, because a typical stock clutch used for a manual transmission in a lot of sports cars doesn't grip well if the engine is at high rpms while the tires are not, it's a common method for magazine testers to simply find an rpm that they can just drop the clutch and spin the tires a bit for the most consistently fast launches.
Quote from JeffR :No, only the caster effect (the tendency to straighten out the wheels) is reduced, not the lateral force (or not by as much).

Right, Jeff. As you've experienced in your own time on the track in your racing experiences, the torque you feel on the steering wheel is not the same thing as the the corrnering force you get. There's probably a sort of middle ground there where the steering gets just ever so slightly lighter where you actually turn the hardest, no?

For others, this whole fiasco is just called aligning torque. Somebody else posted earlier (sorry, I forgot who ) that it can reverse. Yes, this sure happens.

Anyway, this is all a force feedback thing :P When I'm running a sim I'm not judging anything on that at all. It is important though to differentiate between those who are and aren't... Feel..
Quote from jtw62074 :For others, this whole fiasco is just called aligning torque.

I couldn't remember the term, although I do remember seeing independent graphs for aligning torque and lateral force.

Other than clowing around in go-karts as a kid (where spinning them was a common goal), and in my cars at a nearby large parking lot for informal pro-solo type events, or on fun runs at a track. My only true track experience with side by side racing has been with indoor karts at K1 speed, (www.k1speed.com) in Irvine, where everything happens real quick. The fast way is to keep the kart sliding (drifting) a little, including using some deliberate oversteer to rotate the karts for the tighter turns to line it up for corner exit. The steering does get lighter once the tires are drifting, but it's hard to notice when the karts are pulling well over 1g in the turns, and this side force on your body is the dominating force that you feel, distracting a bit of the feel in the steering. It's not that big a deal, because you can feel the entire kart drifting much easier than trying to sense what going on with the steering wheel. Unlike normal shifter kart racing, since they own the karts, they don't allow all the bumping that goes on with normal kart racing, where it appears the common passing method is to stuff your cart inside another and use the other as a berm.

The Caterham I used to own didn't feel that much different than a go-kart, as it's bascially just an oversized go-kart anyway. Similar to the go-karts, it had some lift throttle oversteer (not as much as the karts, but I suspect they are playing some games with the karts computer setup, as you can't power oversteer them, the motor cuts out if the kart is yawed too much).

Now with the Z06, the power steering provides a lot of assist, so there's not a lot of steering wheel feedback, but you stil get plenty from the car itself. I probably won't be tracking it because the tires for it are so expensive.

Personally I have more fun sliding the cars and karts around more than I should for best lap times, as I don't do it very often, so it's just a fun thing for me.
Quote from JeffR :Regarding longitudinal grip versus slip, there was a time back in the late 1960's and early 1970's where top fuel dragsters posted about the same ET's whether or not the tires were spinning all the way through a 1/4 mile run. You'd see smoke from the tires for the entire run on one car, virtually no smoke from the tires on the other car and the ET's would be very close.

Awesome, I had no idea and never heard this before... Thanks.

Quote :
This isn't true currently with fuel dragsters. Smoke the tires at all in top fuel, and the ET suffers greatly.

However, because a typical stock clutch used for a manual transmission in a lot of sports cars doesn't grip well if the engine is at high rpms while the tires are not, it's a common method for magazine testers to simply find an rpm that they can just drop the clutch and spin the tires a bit for the most consistently fast launches.

Learning Carry on
Quote from jtw62074 :
Quote from jeffr :Regarding longitudinal grip versus slip, there was a time back in the late 1960's and early 1970's where top fuel dragsters posted about the same ET's whether or not the tires were spinning all the way through a 1/4 mile run. You'd see smoke from the tires for the entire run on one car, virtually no smoke from the tires on the other car and the ET's would be very close.

Awesome, I had no idea and never heard this before...

The traction was probably less, but not by much, with those tires, and maybe it was all the force (torque) that the engines and clutches at the time could deliver.

If the tires were spun hard at the launch, it would mess up the run, but once past the initial launch, if the tires started spinning the drivers could apparently stay on the throttle.
Interesting

But how did you get the LFS curves? Are they exact?

Furthermore, curves are one thing.. In general LFS is a lot better than most sims when it comes to behaviour over the limit. But the slightest dab of throttle or feathering of the steeringwheel can induce oversteer with normal road car type car setups. I'm convinced that there is something fairly major still not quite right, even if the curves are 'good'.

Edit: if you try 'realish' curves in ISI games you still find that it doesn't really work very well
Quote from Niels Heusinkveld :But the slightest dab of throttle or feathering of the steeringwheel can induce oversteer with normal road car type car setups. I'm convinced that there is something fairly major still not quite right, even if the curves are 'good'.

Chaisis flex probably will go along way to improving that I'd guess
Quote from Glenn67 :Chaisis flex probably will go along way to improving that I'd guess

What leads you to that guess? The assumption that lfs models everything presently in a perfect fashion? Personally I can't see at all why chassis flex would prevent the present oversteer behaviour on road tyres. It always feels to me like the cornering forces are too evenly distributed between steered and non steered wheels. Rear wheels should have much less work to do because the direction has already been 'set' by the force on the fronts, as far as my rudimentary thinking can see
#65 - w126
Quote from Niels Heusinkveld :But how did you get the LFS curves?

http://www.lfsforum.net/showthread.php?p=212598#post212598 (follow the links in this post)
http://www.lfsforum.net/showthread.php?t=3667

Quote from Niels Heusinkveld : Are they exact?

Now I know that grip values (also longitudinal and lateral tyre forces) calculated by RAFTyreExtract are almost uniformly too small by 1-2 % (the relative error is close to (single wheel mass) / (total car mass)). There may also be additional errors when driving over the kerbs due to the way the program approximates the road surface, needed for coordinate transformation.
But I think it is good enough to distinguish between the curve options shown in the first post.

Quote from Niels Heusinkveld : Furthermore, curves are one thing..

Real (experimental) data curves of the types discussed here show steady state tyre behaviour, i.e. when the parameters do not change or change very slowly. It is more difficult to find real data about dynamic tyre behaviour, which I guess might be the area where LFS and ISI engine outputs differ the most.
Quote from Blowtus :What leads you to that guess? The assumption that lfs models everything presently in a perfect fashion?

No, I was having a particularly hectic day at the factory and in my only 5 mins break I quickly replied without much thought and was thinking on the way home later it was a stupid response DOH!

Quote from Blowtus :It always feels to me like the cornering forces are too evenly distributed between steered and non steered wheels.

I guess that is along the same thought I was having and tried to convey but failed... I was thinking drive train flex and the flex and give throughout the suspension and chasis not to mention clutch slip would soften the wieght and force transfers abit more and enable you to have more feel for the cars.

But actually after stopping and thinking some more, after recently wading through a 30 page or so report I had sitting on my hard drive about truck and car tyres (Todd thats your fault because you got me thinking and then I went off in a great tangent as you do...), one thing jumped out at me, and that was the very dramaticly different lateral grip curves produced at different speeds for truck tyres.
Now I know they don't perform the same as car tyres but the thing that stood out was that at lower speeds if you exceeded the maxiumum lateral grip of the tyre the drop off in lateral grip was very miniumal but as speed increased the drop off became more and more prononced.
So it crossed my mind that the behaviour of tyres at low slip and or low speed is likley to be noticably more grippy beyond the tyres limit than at high slip or speeds (obvious I know )

I don't know how dynamic in relation to speed the grip curves are in LFS and how things interact between lateral and longitudinal grip but I tend to believe that's why LFS is better than most, but also it is still an area that can be further developed

Todd was suggesting two curves for longitudinal at different speeds might better represent longitudinal grip and the way it behaves when accelerating from a stop or low speed to high speed, but I'm thinking why stop there and perhaps it should be applyied to lateral grip also

And then obviously temperature, pressure and load on a tyre which will all also effect the grip curves behaviour, so it is a vary dynamic beast to tame - to get the tyres to feel natural under all conditions is certainly a tough task without shortcuts

Ok too much thoughts... time for this little sick puppy to go to bed
ahh thanks for that..

Curves are probably important but if you look at Racer, LFS and ISI.. With Racer and ISI you can change curves.. with LFS they are 'given'.. But even if you sort of make the curves similar, the handling of each sim is not just slightly different; the differences are HUGE! This indicates that the curves are a relatively small part of the physics model. It is nice to see that LFS is 'closer' to reality than most sims but it doesn't mean the physics are 'close' to reality per se. LFS always has me torn... Its very good in some ways but not so good in other ways..
Quote from Glenn67 :And then obviously temperature, pressure and load on a tyre which will all also effect the grip curves behaviour, so it is a vary dynamic beast to tame - to get the tyres to feel natural under all conditions is certainly a tough task without shortcuts

I wondering if LFS currently uses a complex formula to model the tyre curves. If that is so, then I would have thought that it would not be too much trouble to add extra variables into the formula, such as car speed, tyre temperature etc, which would change the shape of the curves accordingly.

P.S. Isn't it annoying not really having a clue about how things in LFS are actually modelled?
Quote from Niels Heusinkveld :Its very good in some ways but not so good in other ways..

When you say not so good in other ways, do you mean what I mean?

- The way a full-throttle wheelspin start gives as good a start as a controlled start
- The way the car slides around quite easily, and does so almost very gently, with no real "snappiness" between sliding and gripping

Are we all on similar lines here??
Is it just me that under weight transfer to the front when braking or lifting throttle, the rear wheels get too little grip?
Quote from MillerM :Is it just me that under weight transfer to the front when braking or lifting throttle, the rear wheels get too little grip?

Could this be a setup issue, perhaps if you try moving to brake bias more to the front wheels?
#72 - JTbo
Tire curves are certainly just part of thing, sure they are base and other parameters and variables are then used to shape it to get final grip.

There is tire load, pressure, temp, and at least with rfactor there is quite many parameters to play with to get grip to what it should be in all situations so it is not easy like drawing proper curve, not at all like that, there is other parameters that are need to be adjusted so that behaviour of tire is more real like.

For example tire load effects to lateral and longitudal curve is one thing. It is not going to be anything real even curve would be spot on if these parameters are not set well and their effect is HUGE, badly set values do cause feeling like that car has predefined spin angle (99.999% of mods? ) even most perfect tire curves has been used.

So to funny part, how we can discuss about these parameters of LFS if we have no way knowing how those are weighted and what they are using, we can say tire curves are ok, but is parameters, does temp and load modify curve just linearily or are they possible modifying peak?

One picture I saw from this thread did show how when tire load is increased, even tire had rather flat curve without peak on low loads at high loads there was more peak, even more load and even more peak there was even overall grip level was higher. So it can't be modified linearily if we look realistic behaviour.
Now what I see in ISI curves, those look like high load ones, but they should not be as there is parameters to take care of falloff and peak things so I would say that is where they have gone wrong. But also you should keep in mind when looking ISI curves is that they are often 0-100 degrees of slip angle while typical curves we see are 0-20 maybe even 0-40 and it might do some difference to curve too if you plot only first 20 degrees, but surely there is too much peak.

I'm still learning lot of this, but that is now what I have understood, might be wrong might be right, you tell me
Quote from DaveWS :When you say not so good in other ways, do you mean what I mean?

- The way a full-throttle wheelspin start gives as good a start as a controlled start
- The way the car slides around quite easily, and does so almost very gently, with no real "snappiness" between sliding and gripping

Are we all on similar lines here??

Exactly my thoughts. Sometimes it doesn't seem to regain grip as fast as a real car would.

There is still a lot of oversteer even with a open differential (which helps a lot )

I really also would love to see some kind of ESP like the TC which helps against oversteer for beginners, but makes you slower like in real life by taking away throttle and braking single wheels.
#74 - axus
Another big thing is the combination of lateral and longitudinal forces. Ie, if your slip ratio is 0.4 and your slip angle is 20deg, the lateral force isn't the same with slip ratio of 0 and slip angle of 20deg. This is something you can't tweak in ISI sims either (to my knowledge). Its important for how snappy the rear is for instance, with oversteer. Most ISI sims tend to do a pretty piss poor job of it, I find, with the whole grip, grip, grip... no grip at all. There's no progression, its way too snappy. LFS is pretty good there, I think.
#75 - JTbo
Quote from axus :Another big thing is the combination of lateral and longitudinal forces. Ie, if your slip ratio is 0.4 and your slip angle is 20deg, the lateral force isn't the same with slip ratio of 0 and slip angle of 20deg. This is something you can't tweak in ISI sims either (to my knowledge). Its important for how snappy the rear is for instance, with oversteer. Most ISI sims tend to do a pretty piss poor job of it, I find, with the whole grip, grip, grip... no grip at all. There's no progression, its way too snappy. LFS is pretty good there, I think.

I was thinking that too, but now, I'm just not too sure, I have not yet got into all parameters, but already I have tires that do have this small warning that tells you that tires are biting hard and might not be able to grip much longer. Anyway I must look this more before I can say anything to any direction, but one thing I already know, it is really lot better than many thinks, it has not used properly just yet.

Tire physics
(88 posts, started )
FGED GREDG RDFGDR GSFDG