Discussion in 'Car Modding' started by vilivili, May 3, 2012.
Hey KeiKei and thanks for the translation, really helps
Are you comparing the Lukkojarrutus to ABS-jarrutus brake distances or Optomi (non ABS) jarrutus to ABS-jarrutus?
Looking back at the table, the non-studded tyre information has missing date for the braking distance of "Optimi (ABS) - Jarrutus". I'm wondering if "Optimi.." is referring to the effective performance you can expect in real-world situations (with a human behind the brake pedal and who would probably be reluctant to push down 100% on the brake pedal despite having ABS enabled), in which case we should only compare the equivalent "Optimi.." for ABS and non-ABS braking distance. If that is true, then I think the data agrees with your impressions that a skillful non-ABS driver yields a shorter distance than full ABS for the studded tyre but there is missing data for the non-studded tyre to make the same conclusions.
Perhaps this next part is making too big an assumption, but the numbers coming from "ABS-jarrutus" sound like the 100% brake pedal with ABS scenario where the traction is jumping between 100% grip braking and down to 71% in momentary wheel locks (but for only a nominal fraction of the total braking phase). This is brilliant on asphalt but I remember watching a video on ABS some time ago that showed how ABS for all it's benefits to the normal road user is in fact not as good on snow/gravel. The reason being that if you keep you tires locked, the snow/gravel will grow into a bulge at the front of each tyre (progressively sinking the tyre into the snow/gravel more) which acts to retard the vehicle more rapidly.
Ice i think is a different matter once again (but not certain), where the friction coefficient of ice is so low that the ABS detects constant locking and must releases the brake pressure more frequently and by a larger percentage in order to try and get the wheel rolling again.
Despite ABS being a great piece of kit, i don't think absolute 100% ABS can beat a skilled drivers braking performance with near perfect modulation based on lots of practice that yields skillful anticipation and interpretation of the cars feedback/handling on the fly. Since ABS relies on the wheels having to lock, the tires will always loose significant grip (drops down to 71% according to this table) even if it's just for a moment. The ABS then has to modulate the brake pressure off and back on to full again till the next wheel lock, rinse and repeat. I think a human at an average 98% braking potential is going to be better than this.
Optimi is optimal braking by experienced good driver, something at least 80% of drivers will never reach, fastest way to stop without abs brakes or locking wheels. I believe treshold braking or something like such is proper english term for that?
ABS stops faster on tarmac, that is where it works well, test car in that test was 2005 Volvo V50 from my memory, despite being build for Scandinavian conditions where snow and ice are typical for any bit smaller road (big roads are constantly wet because of insane salting), their ABS still could not handle well ice and tires without studs.
But that test really is giving just some vague idea of street car performance, I have run several years with non studded winter tires, Continental brand and they are not the same as summer street tires for example, they don't have quite the same grip as summer tires, on wet there is larger difference, it is not huge difference, but it is a difference, when wet ABS starts to activate something around 2/3 of force to brake pedal compared on dry and on dry it is quite well possible to get abs activate. With proper summer tires it is quite hard to get abs to activate on dry, not impossible, but you really have to push the pedal trough the floor and at least with my ancient old french crap there is something to be said about brakes as I can indeed floor the pedal with summer tires and abs activating resulting no longer abs activating or stopping quite as well....
But if we compare this to racing slicks then, something like Meganes, Clios or GTR has, what I think is that standard brakes of car probably will not even get them to lock or abs to activate at all, so at least hugely bigger brakes are needed, even GTR brakes have issues locking all four reliably so one needs to mod it a bit to even actually be able to do some proper testing, even more so as ABS with default implementation increases tire's grip which is not wanted.
With ABS then there is 4 wheels all individually doing pressure adjustments 100 times in a second (or more, I remember reading 400hz, but that is probably sensory side) with modern ABS, which makes it quite good for getting tires to be at their peak when braking, but on slippery conditions system gets bit confused as tire is actually stopped, comparing it to other tire that also has stopped it might get bit laggy, in rFactor, if it is done properly, it perhaps compares to ground speed and adjusts based on that, making it too good, but what I found when testing is that efficiency of abs is actually causing brake pressure level to be held at different level depending from setting, that is how it looks from logs.
With 1.0 it allows whole pressure range of adjustment.
When testing tires, one just have to first know his testing tools and methods being reliable, or one will end up with results that don't really say much else than placebo.
Maybe bigger difference in sliding friction is because racing slicks are so stiff, but so soft in compound, I know that there are bias racing slicks sold even today, they are probably bit softer in construction and I know those are lot more forgiving and probably have different kind of locked wheel braking performance compared to radials when comparing locked wheel vs abs braking. 60's cars have something of such, maybe taking rTrainer and putting different tires to it, then making brakes to be insanely good, might be best testing platform because weight would be quite rear so with weight transfer it would not be only test for front brakes, maximizing tires potential or something like that.
Oh yes, about that table I posted, ice in there, I believe was clear ice but studded tires were used to make it bit rough.
My experience is that tires without studs and clear ice do not mix, ABS is making noise, but speed is not really slowing, handbrake helps only so little. However we got mostly white ice, which is packed snow smoothed by tires, this is somewhat softer and tires get that interlocked effect as rubber bites into stuff, so braking distance is faster than on loose snow. Modern studless tires are incredible as they truly cut ice, I had to test this and spinned tires a bit on ice, I got similar cuts to ice as with studs.
With modern studs that clear ice is not much less dangerous as studs are so small today, mostly I get different sound, but stopping is nearly as poor.
So in that table results are very good ones and my experience is same.
Just waiting to see proper snow/ice physics in sim, bit more challenge than making that 'easy' gravel, which with I'm sure ISI devs will be pulling hair or two until it is properly working
I think that I saw Free model of Volvo V50 at one of those free 3d model sites, might be interesting to create correct spec test vehicle of such and see how it performs in rFactor. Shame that it would take months to gather all data and make proper physics. Even I have some from that car, suspension frequencies for example. Creating proper tires might be still a pain even I have tires they had in that test and one even cut open, but it is just so little and I have already two mods (or actually 4 + tracks if counting ones on hold) I'm working with so I doubt there being chance of making of such, maybe hdv body inertia, mass, brakes and springs + dampers, but that is most I could do now.
What is the difference between "Optimi (ABS) - jarrutus" and "ABS - jarrutus" then?
My understanding is that ABS jarrutus is flooring the brake with all might you got and optimi being treshold braking with abs enabled, which is not really proper way to brake with abs when on tarmac, but on ice it works better than flooring the brake and letting abs do the whole thing as ABS bit of sucks on ice.
any idea why they would leave out braking distance data with the "Optimi (ABS) - jarrutus" on non-studded tyres?
I can only guess, my guess is that perhaps they could not make difference between the two, ABS started cut too early?
Still does not explain why only with non-studded would have that behavior, probably there is some other explanation.
I'm comparing all values within same tyre type.
I think this also applies to ESP (Electronical Stability Control). In theory and in most circumstances it controls car lot better than driver because it has the ability to brake different wheels independently. However in extreme circumstances it can cause problems. I've experienced this once when had old non-studded tires and road was extremely slippery (hard ice with thin layer of water on top). Well I was in a hurry (poor explanation) and turned in into 90-degree corner. Front tires started to loose grip and hence caused some understeering. Then I turned steering wheel some more which caused ESP to react. It basically applied some braking force to each wheel and I instantly became a passenger (we have a saying for this: going into the direction of a tangent ). Road was so slippery that wheels didn't start to roll again and there was nothing I could do to control the car. Well actually pulled handbrake as a last resort to avoid direct hit with steel light pole. The light pole survived but the side of our car didn't look so good anymore.
Of course, cannot tell for sure without being there, but in similar situations I was able to unlock the front wheels by straightening them, or rather pointing them in the direction of travel. ESP and traction control can be dangerous at the track as well - I once had ESP turned on in an M3 without me realizing it. So I approach a turn where I normally use aggressive throttle just before the apex to rotate the car, start adding throttle as I normally would, but the car just keeps going on a wider arch, leading into a concrete wall, and there is no way to tighten the arch with the steering wheel. Good thing I had time and presence of mind to change the strategy and lift to cause off-throttle oversteer, which is easy to correct in M3.
Anyway, back on topic - even though I think rfactor is the most accurate sim I've ever tried, there is definitely something wrong with the way it handles sliding of tires. I can only guess what is wrong with the model, but from the drivers seat it works like that - when the tires are sliding, they basically have 70% or so of normal grip, and this grip slows them down a lot, pulling them back. So as a driver, you feel rear coming around, you point front wheels in the direction of travel right away, before the car gets any rotational momentum, and you end up with front tires rolling in the direction of travel with little to no resistance (rolling resistance is negligible), while rear tires scrub across their thread in the same direction, with friction comarable to half of full-on braking, if not more. Anyway, no matter how much friction tires lose from sliding, sliding is still much less efficient then rolling, so the front goes faster than rear (assuming no throttle impact and no brakes). That essentially corrects the rear VERY quickly, and you need to be very fast on countersteering not to end up with a tank-slapper or spinning the other way around - so powerful is the friction of the sliding tires.
In RF2, I often end up with these endless drifts where front wheels roll in the direction of travel (so practically no resistance) while rears slide with the same speed and deceleration rate (so no resistance there too?). THis should not be possible. In real life I would need to apply a lot of throttle to rob rear wheels of traction to maintain such a drift, but in RF2 it just happens by itself, so something is wrong.
Another thing is that in real cars sliding and drifting slows you down a lot - all that kinetic energy just turning into heat, like towing a car that has a parking brake on. In RF2 it seems to be nearly free, which is impossible - ater all, sliding is essentially dragging the car across pavement - it requires a ton of force and energy, and this energy is essentially removed from the kinetic energy of the car, which should slow it down dramatically.
I don't drive irl but are you sure it's not a case of the difference in the car you've experienced this irl to whichever one or ones you are comparing to in rf2?
It may be the case - I am not completely sure. Although, the 350Z GT is very close to a car I had a chance to drive irl. But the difference is pretty drastic - IRL I had to practice to make correction and recovery a one motion, because making it two separate motions was too slow (probably just .5 second delay - still too slow). In RF2 recovery should come with a delay because it takes a while for correction to take effect, or in some situations recovery is not even necessary because you just drift sideways for some time and then rear slowly comes back into place. It's close to driving in the rain irl.
What tells me it may not be the car difference but indeed a physics issue is that in Skip Barber's video in the skidpad episode you clearly see how quickly the rear snaps back into place on correction, and how quickly he needs to recover. I tried that in RF2 and the car goes sideways for some time before the rear comes back. In the video, he needed to apply a lot of throttle to achieve a comparable drift.
There's big difference catching the slide with rear wheel driven cars when one is brave enough to keep throttle pedal down (or fairly down) - then it's possible to catch the slide nicely. Worst mistake is to lift off completely which will suddenly cause huge counterslide. Could it be that in real life when your life is at stake one would tend to lift off the throttle quite easily?
I'm not sure the OP really understands kinetic vs static friction but just so you guys know there is a readme of known issues in the rF2 core folder.
-Contact patch load distribution model is not final, which means tire pressure does not have the full effect that it should.
-Aquaplaning is not implemented, so wet weather driving is only properly modelled when the track is damp but with no standing water.
-Wear (meaning both abrasion and degradation) is incomplete in several aspects.
-There are other relatively minor loose ends.
-Curbs, white lines have almost no grip in rain.
Separate names with a comma.