Guys, come to talk about tires

Little more fun:

Actually made a thread with all settings etc.
https://www.beamng.com/threads/nordschleife-hotlapping-open-challenge.58749/

There is one thing I have been thinking to test, how inertia and speed of object behave, would faster object need strong thruster to steer front end to right for example? That would require zero gravity or means to limit object to level flight as it would need test without tires.

Probably they should steer the same, if aero of sides is not keeping thing straight.

 

Well it's really sad that a big group of peoples have to bash devs about something, so we can finally see how much of work did they put on that subject. In past I was continously asking for another devblog about tires'n'stuff that @Diamondback has promised years ago, we had to push @tdev a bit from time to time to actually see http://blog.beamng.com/something-is-still-on-the-horizon/ and another not so visible stuff, I realy enjoy @estama posts, very complex, exciting and mindblowing stuff, but it's visible for few peoples looking in specified thred where somebody had problem with something. It was the same with "driving on ice problem", 1st, 2nd devblog about tires and finally video with @B25Mitch voiceover that changed a lot of people mind

Spoiler: Video

Such an inteligent, openminded and clever guys working there, but we don't see it untill somebody is bashing on something. I've created BeamNG.Science thread https://www.beamng.com/threads/beamng-science.28426/ and looked like almost nobody knew about all the physics that's working in their "crashing simulator game". You guys (devs) really have something beyond other companies in the world and we know so little about it, yet I obviously don't ask for physics code, but the problems you have to face, the expreiment you do, how much effort it take to improve some stuff. It's really interesting and eye opening to read your blogs about this stuff, what have you acheived what kind of research have you made, comparisions with older iterations etc. I know about all the stuff you published since beginning like tennis racket theorem, pressure change with height and dozens of really neat stuff that's uncommon for most peoples aroud here. Yeah the most amazing stuff about whole BeamNG.drive game is yet the most uncommon too, imo because you hide it too much with "under the hood work". You did only one video that's a good showcase of your physics and it's the one with handling. The gameplay stuff yet exciting it's what other games offer too, so "nothing new", but connected with your physics it's becoming something special, but you didn't recently explained and show to higher volume audience this stuff. Some physics summary vid on youtube, with lot's of stuff that's going on, with nadeox1 skills to create nice looking vids, not only basic footage. Maybe a series showcasing
A) Aero works with comparisions, slow-mo footage,
B) Then another one about driving aids like ABS, ESP, Line-lock, Two-Step,
C) Next overall thermals; brakes, oil, coolant, clutch, exhaust,
D) Physics phenomenons that occur straight from physics in game, withou scripting,
E) New refreshed Handling vid, with comparisons, better camera shoots, suspension work, camber, toe, spring setups wnd what it does
F) ...And so on
^More documentary, learing and promoting style with Mitch's voiceover . Showing it on devblog, youtube, steam, twitter and facebook to change popular opinion of "just crashing sim game with weak handling that just let you drive good enough so you can crash they way you want" because that's a popular opinion about BeamNG.drive, (mostly spreaded by popular youtubers that does't have time to look on forums and blogs, but just take a glimpse about easy materials like videos) but I also see a bit of devs fault about it in not sharing your acheivements, hard work, research, failed attemps, experiments, ideas. Sometimes it looks like you update the game with mostly gameplay stuff and bugfixes once a 3 months (major update) makes game development look slow. While you work on a lot of other stuff too, but if we don't see it then how can we know?

That's just idea, most likely I ask for too much, but maybe then people would get a proper vision about your project, because who have to show them what's going on? You as devs have to promote yourself. But you have to remember that peoples are lazy this times and don't like to read and search for it, but if you give easy for them material within a reach of a hand so they can just sit, listen and look to see what's it all about, that coud be a big game changer, physics was the selling point for a lot of us, so maybe it's worth it to invest this working time on advertising this way also?


Sorry for my weak English, I know too much repetitive words and typos.

 

Automation devs have really interesting videos about how their engine simulation works and how they plan to develop stuff etc.

What I think is that devs might not want to do that quite yet as it ties down resources and things are still rapidly changing, they might do more of that in a future though as tdev mentioned about motivating his people to share more stuff, but I think you have very good points there, exposure I think it was called that is needed and good videos as well as explanations in common man terms, those would really do wonders to challenge traditional sim crowd to see new ways.

Steam store page could have BeamNG vs reality physics test updated to current version etc. but maybe they are going to do those after they have things stable?

 

I'm pretty stupid and should not really do much thinking, but one thing I have been wondering is drag racing tires.

Data I have seen about them was claiming that they get friction coefficient of 3 or something like that on glued track, but when they started slipping friction coefficient did go up to 5 or something lot higher, can't remember if numbers were exactly like that though.

Does glue become cement or something? How this works in BeamNG?

 

A way to create a similar effect is to increase the surface area of the nodes with surfaceCoef. This unfortunately has the side effect of causing the node more difficulty of sinking into the ground which... depending on the tire type... would be unrealistic. If it's a thin tire (like the tractor tire) it would ideally sink into the ground with little resistance as to get the treads deep enough to get in contact with more structured surface.

The best way to get this effect to work correctly would be to figure out a way to make it directional so that you could increase the surfaceCoef in one direction and decrease it in another... I am guessing it would have to be radially directional to the hub as well to work properly.

The most extreme example of this would be in Swamp Buggy Racing.


This is mostly how tires work in snow as well... but I am not going to mess with that for the moment.

Drag tires are specially designed to do nothing but grip asphalt. They are VERY sticky and last a very short amount of time due to how quickly they wear out (granted, doing a burnout at the start of every race doesn't help that fact). Speaking of which, doing that burnout at the beginning of each race essentially soffens the rubber until it is practically nothing short of tacky rubber cement. On top of that, a prepaid drag strip will have a substance laid down on the first stretch of the course (We did 500ft at our local track, but then again we were using an airport) called Track BIte... which is essentially like just putting glue on the ground... so it's like getting glue to stick to glue. This enables them to literally stick to the ground and effectively catapult the car forward when the time comes. They also run really low air pressure as to use the rubber in the tires sidewalls almost like a rubber band. This helps the tires hook up under extreme loads by giving the contact patch that little extra bit of time to stay in contact with the ground while the car launches... They are also shockingly lightweight as I found out this summer helping someone change their car to drag slicks at the local races.

As for having a kinetic coef of friction higher than their static coef of friction... that I have never heard. When the tires start spinning, depending on the power of the car (if your using drag slicks, you have enough power), the tires will effectively grow and for a brief second will be pressed into the ground causing a bump in the natural force... after the briefest of moments though, this effect would be lost and then you would be doing nothing but spinning all your power away.

Just skip to 1:50 to see the effect.


How Beam handles it exactly, I am not sure as I have never used the drag tires yet.

 

I know about jokes with throwing tires to wall and them sticking there on vertical surface.

Even track strip start area is sticky, like walking in a glue, because that is practically how glued track is, covered in glue.

Still it is mind boggling how friction coefficient can increase so much when tire slips, it is like they would work completely wrong way.

However so it is with tractor tires and soil, slipping tire has more friction as it digs rocks and whatnot from the ground, there was a paper about it I read more than 10 years ago, but it was fascinating how that too did work completely opposite of what is usually thought that slipping tire has lower friction coefficient.

Snow and ice are incredibly complex, because there is thin layer of water forming under the tire and it even is not even, but varies for different part of contact patch, then slipping tire on snow/ice changes picture again completely. Complexity of that is really just beyond my brain, add studs/spikes to that and it is something we will not see in any sim, ever probably, it is way too complex to do, surface is so dynamic and changing so wildly which is what makes ice and snow fun to drive on.

I doubt I will even see a sim doing that to believable level, it just is not possible without having huge number of points under the contact patch, just imagine what road would mean in terms of data

 

Yeah... might not actually be a joke. When the guy asked me to hand him one of his tires (after me just taking the stock one from him) I almost accidentally threw the thing across the track. The tires are so big, but they weigh almost nothing. I could literally lift a drag radial and the rim on it with just one finger... it was really an eye opener.

As for sticking to stuff... when he was switching back to his stock tires he handed me one of the drag tires (still with a lot of heat in it) and I set it down on a 2x4 that the guy next to us had been using to keep his jack from sinking into the ground. When I went to grab the tire it actually stuck to the 2x4 enough to lift if off the ground... it was only a foot long board, but still... I wasn't expecting that to happen.

Does the coef of a drag tire increase when slipping? That doesn't make any sense to me as that would then be adding so much heat to the tire that you would basically be liquifying the rubber... then it wouldn't really be coef of friction with the surface, but more like liquide shear of the molten rubber.

This is a quote from an Aerospace engineer from a physics forum

Pretty much that description is for a perfect run with no loss of traction. Another thing that helps is weight transfer and tire width. As the car launches, it presses as much weight as it can onto the rear tires which causes the tire to press into the pavement (all those little divots and what not act like hand holds for a mountain climber) which helps the tires "grab" the pavement and propel themselves at incredible friction levels. The bigger the tire, the more rubber there is on the contact patch to be pressed into the pavement for this grappling effect (another reason to run low pressure)

If it wasn't for this effect and the fact that rubber shears under a certain amount of stress, the size of the tires wouldn't even matter as friction doesn't actually have anything to do with surface area (well... mathematically anyways)

 

Yeah, that is what has been puzzling me over 10 years, how it can be as such should not be.

So I'm thinking glue gets hot enough to become cement, or something insane like that, how it can be so really, but that is what study claimed what I read back then.

When looking lauch of a dragster in super slowmotion, one can see how tire is rolling bit faster than what would be no slipping condition, especially at launch, but is it optimal or less than optimal that way?

Then could it be as those are so soft rubber that slip actually is rubber deforming, then such thing as static would not even exists with soft tires.

Bicycle tires for mountain biking have super tacky compounds and those are pretty much dragster kind of softness I think, because they wear really quick, but grip is also quite wonderful, downhill, hate them when climbing though, so one set was enough for me of that sort.

 

Alright... after doing enough reading to probably allow me to pass a final in college on drag tire physics... I have the answers.

Drag tires don't generate a higher coef of friction while spinning... but "some" are designed to spin.

Some tires are designed to have a "controlled spin" while others are designed to "dead hook". Dead hook tires are designed to be used on low to medium horsepower automatic cars that use a torque converter. Since they use a torque converter, they have slip built right into the driveline so there is no need to have the tires slip to match the engines speed. Manual transmission cars and high horsepower Automatics must use controlled slip drag tires. These are required to allow the tires to properly hook up while slipping to let the engine and drivetrain spin at a speed that doesn't bog down the engine and keep your RPM's at their optimum range... essentially acting like traction limited launch control.

If they don't spin, you could end up bogging the engine, creating tire shake, causing wheel hop, causing excessive tire wadding, or worse... destroying vital parts of the drivetrain like the universal joints, differential, driveshaft... or worse...

So when the wheels are spinning, they are spinning in a controlled manner for first gear with the gear ratios set up just right to enable the exact amount of slip at maximum attack. In setting the car up like this, you can have tire slip while also staying really close to maximum traction (this is difficult to set up and get it right, too little slip and the tire will end up dead hooking too early, too much slip and you fall off the top of the traction curve, reduce your coef of friction, and just spin all your power away), and thus, allowing for maximum acceleration, keeps from breaking stuff, and lay down crazy times again and again and again.

 

To be honest I think the way cars handle is by far superior to other games, and the driving feel is really good. I honestly don't see why people are complaining .

 

Here's my 2 cents,
Tire simulation in .drive is very realistic. Cars handle the way I'd expect them to in most scenarios. There are many aspects that need improving/adding, though.

Tire thermals (as mentioned many times previously) would greatly improve simulation for things such as dynamic grip (race and drag scenarios especially) and pave the path to more advanced simulations.
Flatspotting, though mostly for full-fledged sim racers, I think would be very doable with how the tires are set up in .drive; punctures, I believe these are simulated to a degree but need heavy tuning; tread simulation, very visible on big offroad tires (also mentioned above).

Now here's where I really want to put my attention- FFB. The FFB in .drive is one of the best simulations I've seen, but takes a hell of a long time to get right. I've spent hours just tuning the FFB, going into the real world to get benchmarks, and tuning back on my wheel. I've gotten it very nice, but it's still not right. Response curves are a big issue, they are horribly implemented. I ran wheelcheck 3 times and averaged the values as you're supposed to, ran that sheet through lutgenerator, put that in AC and .drive. AC feels very nice with it, less so without it. .drive is the opposite, feeling pretty good without it but terrible with it. At strength 10 and max strength 2.3, it still clips and I get no resolution. At strength 230ish and max strength 2.3 It doesn't clip unless under extreme loads, but it doesn't give the correct torque because there's no response curve.

I love all the effort going into this and I'd love to help in any way I can, and so far tires look/feel great, but there are some issues.

 

What a great thread. After reading the first few posts i wouldnt have thought it turns out so good with multiple in-depth posts from the devs. Thank you for the insights.

 

yeah, too bad tires were permanently disabled in the new update. :/

 

I thought the same thing when I read that... my tires work fine, so I dunno.

 

Did they get tired and run away?