The online racing simulator
Quote from jtw62074 :Very interesting data there. Do you have a link to his thesis? I haven't seen it yet.

Todd, it's available -- for a fee -- from the link at the bottom of this page: http://www.mae.buffalo.edu/research/phd-kasprzak.shtml If you want to avoid the fee, you could probably ask him if he has a copy he can forward to you. It's an interesting read and well worth a look for those interested in the subject matter.

I too would be very interested to hear Doug Milliken's take on the data shown in those figures and, more generally, on the various approaches used to model tyres in these consumer level simulations.
totally agree with this thread starter.

My XFG has a sport setup - lowered suspension etc, but it acts completely like a fiat seicento ! ( i had that doubtful pleasure to drive this car)
Quote from Dupson :totally agree with this thread starter.

Did you proceed to read the rest of the thread after reading the first post?
Quote from George Kuyumji :Thats right.

Big part besides of the sense of speed is how the car is setted up. Setups in LFS are often very loose compared to normal Road cars. I'm just wondering what type of wheel and pedals are you using? There are big differences in Pedals, there are plastic ones which measure the amount of braking applied by the way the pedal traveled, or more expensive pedals that feel alot better and work by the amount of force applied to it.

Using M/soft FFB2 wheel and pedals, which I realise isn't exactly state of the art by today's standards but I've become used to them over a couple of years.
Quote from choxaway :Using M/soft FFB2 wheel and pedals, which I realise isn't exactly state of the art by today's standards but I've become used to them over a couple of years.

Don't worry, plenty of us still using those beauties
I don't think FF technology has moved on significantly, if at all, but they are a bit lacking in the rotation department. Fine pedals though, even if there are only 2.

Car setup is so important to the way a car, and therefore a sim, can feel, that it's hard to make accurate comparisons with anything else. For instance, I had been using XFGti setups with the locked differential until recently (on the basis that it was faster), and my previously favourite car had become one of my most disliked. Then I got sent one with a very open clutch pack with apparently no performance deficit, and it really feels like a pointy little hot hatch again.
Quote from BuddhaBing :Todd, it's available -- for a fee -- from the link at the bottom of this page: http://www.mae.buffalo.edu/research/phd-kasprzak.shtml If you want to avoid the fee, you could probably ask him if he has a copy he can forward to you. It's an interesting read and well worth a look for those interested in the subject matter.

Thanks for the link.

Quote :
I too would be very interested to hear Doug Milliken's take on the data shown in those figures and, more generally, on the various approaches used to model tyres in these consumer level simulations.

Doug and I have gone back and forth a bit on this via email the past couple of days. He went through this thread and read some of the posts (ours in particular) and looked at the data. While he's too busy to get involved in a debate right now, he did clear up a couple of things with me.

Essentially, there can indeed be a peak and a slight drop off when enough negative camber is present. When we had discussed peaks in the past, what he had in mind were the book illustrations found all over the place that show lateral force plummeting after the peak over a very small slip angle range, which is wrong. So it seems when he was reinforcing that there wasn't a drop off, it was meant more relative to those sorts of illustrations and wasn't to mean that there never was any measureable drop off after the peak in any situation at all. When there isn't any camber or camber is positive, the curves are indeed flat after the peaks as I've been preaching.

When there's a fair amount of negative camber there can be (but not necessarily always is) somewhat of a drop off in the direction where the tire is leaning into the turn, which normally flattens at a slightly lower value (similar to previous LFS versions). This isn't to mean it drops 10 or 20% or anything like that, just that there can indeed be some measurable drop. Away from the turn or at 0 camber the curves are just flat after the peak.

The picture you showed from Kasprzak's thesis with the really big drop off was at -9 degree inclination, which is some very serious camber and probably not a normal operating condition for that tire and would likely explain what's happening (this is the highest drop I've seen in lateral to date aside from drum test data that often looks similar to that). Note that some of the FSAE tire data showed a slight drop while other data was pancake flat. This can most likely be attributed to camber differences in the tests. I suspect pretty strongly that the 4 pics you showed were all indeed flat belt data since it came from Calspan, so my suggestion that it might have been showing a drop due to being measured on a drum is very likely wrong.

However, the point is, as you said, that there is indeed data out there that shows a drop off. At 0 camber, positive camber, and mild negative camber the curves are quite flat as I've been preaching, but when you start dialing in the negative camber in fairly significant amounts the situation can indeed change. I wasn't aware of this, so I've learned something here and thank you for challenging me over the past few days on this

I must now agree with your earlier statement about the GTR2 curves not being all that bad since they only show a few % drop once you've tripled the slip angle beyond the peak as long as this is in a situation where there's a fair amount of negative camber involved (granted, the inside tires which are now in positive inclination should then be operating quite differently, but since the loads there are much lower in comparison it's not as critical).

The original rFactor curves I posted though are a still way overboard, but perhaps GTR2 curves are not all too bad up to and a fair ways past the peak. I suspect it's still excessive though as the countersteering problems were still there (at least in the demo I tried it was), but the point is that some drop off is allowed as long as it's in negative inclination (essentially camber), which is contrary to the data I've seen up until now and what I've been telling everyone as a result.

Can someone pass the salt? My foot tastes a little bland

Another thing to watch out for when looking at curves, especially if there's any drop off at all, is the axis system it was measured in. If the data is in the tire plane then normally things will be flat after the peak as long as the negative camber isn't up there too high. If it's measured in the vehicle plane then of course there will be a drop off right down to 0 force @ 90 degrees. Just another thing to watch out for...

Something else that was pointed out was that in their business (racing) they aren't interested in what happens after the peak, so when they fit their model to the data they don't care if the modelled force plummets into the abyss afterwards or varies considerably from the measured data. They are only interested in getting a good match below and up to the peak. If the mathmatical formulation that gives the best fit to that winds up causing the force to drop, rise, drop, rise, then shoot off to the moon in either positive or negative direction after the peak, that's perfectly fine with them.

We can't have that luxury in the simulator business though of course since it needs to operate well everywhere
Quote from jtw62074 :Essentially, there can indeed be a peak and a slight drop off when enough negative camber is present.

have you got a intuitive explanation for why this happens ?
and also a rough rule of thumb number of how much it will fall at worst ?

Quote :When there's a fair amount of negative camber there can be (but not necessarily always is) somewhat of a drop off in the direction where the tire is leaning into the turn, which normally flattens at a slightly lower value (similar to previous LFS versions).

I must now agree with your earlier statement about the GTR2 curves not being all that bad since they only show a few % drop once you've tripled the slip angle beyond the peak

ah but gtr2 curves keep dropping without ever flattening so if you dont countersteer faster than the curve drops or run out of lock all youll achieve is making the situation worse ... or on the other hand steering into the slide will be your best option which it really should not

thing is if youre not biggie bawbag niels or peri chances are you wont feel the spin comming until its too late with strictly monotonic decreasing tyre curves
Quote from Shotglass :have you got a intuitive explanation for why this happens ?

Since this is new to me, I really don't know. Friction coefficient in reality is different everywhere in the contact patch. It varies with several things, all of which are not known yet by anybody as far as I know.

One factor is vertical pressure (load sensitivity on a local scale), which again is different at any given spot in the contact patch. Adding camber or braking/accelerative torque alter the vertical pressure distribution throughout the patch (in the case of braking/acceleration it's very significant). Changing slip angle at constant load ought to do this too. The FSAE tire test videos showed the tire moving up and down significantly with slip angle in an attempt to keep the load constant. (Increasing slip angle caused the tire to squat considerably). Add in some camber in combination with that and who knows how the load distribution is effected? If the area under the curve of the entire distribution is changed, then the total friction will too, as will local surface temperatures, which would then modify friction again on top of this.

Again, I don't know exactly what's going on in the patch or why it happens, but it does. This is the realm of FEM modelling by tire companies on supercomputers.

Quote :
and also a rough rule of thumb number of how much it will fall at worst ?

I don't know.

This is the biggest drop I've ever seen in a car tire (thanks BuddhaBing):

http://i11.tinypic.com/4l96wyh.jpg

That's at -9 camber and appears to be an extreme case with a tire that's very stiff in the lateral direction. As I've said before, I haven't seen anything measured in the tire plane until now that showed a drop that was worth noting and was outside of the usual tire noise. I'll have to see more data at high slip angles with a couple of degrees of negative camber to give a useful answer. I'll definitely be keeping an eye out for it

Quote :
ah but gtr2 curves keep dropping without ever flattening so if you dont countersteer faster than the curve drops or run out of lock all youll achieve is making the situation worse ... or on the other hand steering into the slide will be your best option which it really should not

thing is if youre not biggie bawbag niels or peri chances are you wont feel the spin comming until its too late with strictly monotonic decreasing tyre curves

Yes, I agree. I'm beginning to wonder now if perhaps the combined slip modelling might be more at play than I thought previously. I'll have to try some curves that drop off a bit in mine when negative inclination is present and see what it does.
Just to say thanks for the debate.
That is a big thing; you're not 'riding on the curves', because the result on the road is often a combination of longitudinal and lateral slip. If you have real curves, measured perfectly, and you enter those curves in a sim, the output will only be perfect if the force combining is perfect as well. Since, aparantly (Todd said... ) force combining is tricky and easy to do slightly wrong, likely seen in LFS and ISI, the whole point of realistic curves in sims sort of stops being all that important! The target is not 'real curves', the target is 'exact final force vector'. So I am saying, even if you would find that curves drop a lot after the peak, copying that curve to a sim won't necissarily make it more realistic if the force combining isn't perfect!

I've been spending too much time with rFactor and only when you get to the nitty gritty of it you tend to get some appreciation how changes in the curves (god knows how they are combined), lead to changes in car handling. With the released corvette, sustaining a high angle powerslide is a little nervous; seemingly lateral grip isn't as dependable as I'd like. I've since only slightly changed the shape of the curve. At a slip angle of 26 degrees, the new curve has 1% more lateral grip. Hardly spectacular, but this made a noticable change in the handling; more confidence in keeping the back stepped out at a big angle. Given the likely inaccuracies in force combining, a completely flat curve after it reaches its maximum would probably be a simplification one could defend in court..

Longitudinally I've never really been sure what is going on. Its common knowledge that you can screw up your 1/4 mile time if you don't launch smoothly. The conclusion is that excess wheelspin must mean a reduction in grip, which could VERY well be the case. Thinking on the matter though, say you're doing a drag run, with ''knowledge'' in the back of your mind that wheelspin is bad. When you hear / see the wheelspin, what do you do? Get off the gas to try and stop them from spinning... So where does the lost time come from, the spinning tyres or the gas pedal not being pressed for a few tenths of a second! Not saying this is true, but its worth considering at the very least!

Another longitudinal ''thing'' that I can imagine is the excess heat generated by noticable wheelspin. When smoke comes from tyres, I can imagine the composition of the rubber is sort of fluid and mushy right at the surface of the tyre. This might reduce available grip somewhat, though the extend is open to debate. If you would find data up to high slip ratio, and it would show a drop off to, say, 80% grip at 1.5x slip, this would already include the effect of the rubber heating up. It seems that most sims use base curves and, on top of that, add temperature effects, reducing grip further. Here also, force combining is what makes the driving experience, so some amount of simplification might well be justified.

Attached the current grip curves for the Corvette, plus a comparison (that 1% difference at 26 degrees slip) that you can barely see, yet do notice when driving, of the new curve and the one that is in the Version 1 release of that car.

So in semi conclusion, speaking about tyre curves and how much they drop off tends to assume the force combining is 'perfect' which it probably is not. Secondly, you can't see how the handling will be affected by different amounts of lateral dropoff; you can't predict how a drop to 80% feels just by looking at the graph. From my limited but practical (rFactor that is) experience, 1% difference in grip at a big slide angle of 26 degrees makes a noticable difference. My slightly premature conclusion would be that any tyre that drops noticably towards, say, 30 degrees slip angle, will be VERY hard to drive beyond the limit, unlike what you see so many real drivers do in real (road) cars.


PS: once again, a really informative thread about cars / physics; LFSforum - respect!
Attached images
c6_curves_current.jpg
c6_curves_release1.jpg
Quote from jtw62074 :
However, the point is, as you said, that there is indeed data out there that shows a drop off.

There is but as you say there is also a lot of data which shows no or very little drop off after the peak. For instance, here's a paper by some engineers in the R&D department at Yokohama which shows measured data for two tyres: a 195/65R15 and a 205/55R16. Neither shows a significant drop off after the peak. Note: the paper is in two parts; the second part shows the slip angle v lateral force curves.

http://scitation.aip.org/journ ... l_34/iss_2/84_1.html#div8
http://scitation.aip.org/journ ... ol_34/iss_2/100_1.html#F6

I've also seen a paper which included some data from Bridgestone which showed little to no drop off after peak but I can't lay my hands on that paper now.

At some point, I'm sure the materials and physics will be sufficiently well understood that physical models can be developed which accurately predict tyre behaviour throughout the full range of operating conditions.
Concerning dropoff with camber, load sensitivity is the first thing that comes to mind which Todd already touched on. Could it be partially due to much more rapid heat build up in the portion of the tire in contact with the road; or specifically the portion receiving the bulk of the weight transfer - since the surface area handling that weight is reduced... It actually seems logical that this would happen, but I certainly never thought of it before. Obviously that's a simplistic view, but it does make some sense I think. Note that with less load, the dropoff is not as severe, even at -9 deg.

Nice work on the debate BhuddaBing!

Man I love these threads!
Tire is not most stable object as it is made of rubber so grip also changes quite a lot by different stress etc.

What grip tire provides could be modeled as certain quite linear curve, but what is final grip is complicated and of course many things affect to grip and also we could make curve to be different in different situations too.

But question is, do we like to make curve from 0 load tire with 0 camber and modify that or something else and work it out from there.

Anyway it is not much of all indeed if curve is perfectly right, force combining + other modifiers are so big part that those are things to concentrate when doing tires for example to rfactor mod. (imagine if you have 40% drop off in curve and 60% dropoff from other data, would that be slippery or what?)
Quote from Ball Bearing Turbo :Concerning dropoff with camber, load sensitivity is the first thing that comes to mind which Todd already touched on. Could it be partially due to much more rapid heat build up in the portion of the tire in contact with the road; or specifically the portion receiving the bulk of the weight transfer - since the surface area handling that weight is reduced... It actually seems logical that this would happen, but I certainly never thought of it before. Obviously that's a simplistic view, but it does make some sense I think. Note that with less load, the dropoff is not as severe, even at -9 deg.

Nice work on the debate BhuddaBing!

Man I love these threads!

Yes, that could very well be part of the picture. Load distribution throughout the contact patch is quite bizarre frequently. The ones I've seen for tires operating even at 0 slip angle, 0 camber, and free rolling, show a somewhat unusual pattern across the width of the tire. It's often a fairly unpredictable looking wave. Local load sensitivity (imagine the contact patch made up of thousands of little chunks of rubber, each with its own friction coefficient that's varying with load) could very well be having an effect here.

A tire engineer in an engineering thread somewhere said he didn't really believe that rubber's load sensitivity was responsible for this, more or less denying that it exists (lots of sliding rubber research indicates otherwise, however), and that the load sensitivity of the tire as a whole can be attributed to carcass effects. In my mind at least, carcass effects vary the local load distribution and would explain much of what's going on provided the rubber itself is indeed sensitive to load (all rubber data I've seen shows this to be the case). So I'm not quite sure what the tire engineer was on about.

Anyway, when you start dialing in camber, slip angle, and especially slip ratio (braking/acceleration torque), the pressure profile can change dramatically. I'm having a hard time visualizing what might happen to this pressure distribution at slip angles past the peak with a lot of camber, though.

I've played quite a bit with FTire (the simple FEM type of tire model) and it never predicted much of a peak at all with the default tire that comes with it, even with quite a bit of camber. That's using local friction coefficients that vary with load and quite a lot of other stuff. There is probably a mishmash of all sorts of things happening in the rubber.

I hesitate to speculate on a dominate effect, however. A tire engineer with the really serious FEM stuff they're using these days could probably answer that.
Quote :A tire engineer with the really serious FEM stuff they're using these days could probably answer that.

Surely someone can bend one's ear to get him to give Scawen some useful info?
Mike85
Well, don't you think it's a tad presumptuous to speak this way? We haven't a clue what information Scawen has, or who he's befriended during the development of LFS... He's not likely to be just pulling things out of his rear to get LFS this far.
-
SEO Services Company India (rosesmith) DELETED by Scawen : spam

FGED GREDG RDFGDR GSFDG