Exploit of physics flaw or what?

On relatively high friction surface, this seemingly appears impossible, yet in game fairly easy to make happen.





Even on ice, such is challenging to make happen IRL, but in BeamNG there seem to be total loss of friction or something along those lines even on pavement, also not sure about engine torque behavior, my experience is that engine torque needs bit more squeezing to be put out to axles and tires, but I don't have numbers on those of course.

What everyone probably agrees is that with 35hp you don't do a burnout with sport tires and with heavy ETKi like vehicle you don't spin tires at 1200rpm with 15% throttle, that just should not happen.

Another interesting experiment was to put 4.5l V8 from pickup to BX200 and see if it can be stalled, seems to be that idle is enough to make tires spin at 1st gear even if car is rolling backwards close to 40kph and you dump the clutch.

Seen at end of the video:


I don't know much or understand much, but what I know is that when curiosity explores what is, something usually is found out, who knows maybe there is something bugging out or something not finished yet.

I have often noticed that engines are stronger at keeping rpm than tires making engine to change rpm, that is compared to IRL experience at least.

Also idle of small engines is pretty hefty, doing this IRL certainly would stall the engine.


I'm just curious of what makes these quite apparent flaws to happen, maybe someone brighter can figure out something smarter out of this, but somehow something making these odd behaviors possible is connected to challenges of making very light vehicles to grip well.

 

I Agree, the only time i've seen a car "gradually" lose grip like that was i think a diesel drag car warming up its tires.
I wouldn't be surprised if this gets fixed when they eventually add tire thermals.

 

I guess the lack of tire temperature/wear simulation is the culprit. When they add those, it will hopefully improve a lot

 

Devs said that current tyres grip is set to "warm and brand new tyres" so wear and thermals only would get tyres less grip. So it would be even easier to do.

 

I wouldn't bet on tires thermal and wear simulation, as now, tires are always at their optimal (best) performance, so thermal and wear sim will just make it worser due to less grip. - Ninja'ed by robert357

I certainly have to agree with OP, stalling a car irl is much much more easy, I wanted to learn my gf some pedals work, how to start in a car etc, but I figured out it doesn't make sense, because it's too hard to stall the engine. I've got hyped for base models of piccolina, as I wanted to experience some difficulties, mad downshifts to keep it rolling, but nah, eazy start on uphill roads of WCUSA with no throttle imput. Feels like every engine have too much torque. About loosing grip, feels like the moment you loose it, you loose too much of it so then tires roll like there's no tommorow, and that may be the case why it's so hard to catch a car back while irl I don't have a problem with that. Could be that due to video game limitations ( lack of G-forces, speed sensation etc.) and hardware delay or other imperfections makes it so difficult compared to real life, but it seems much easier to control car in a drift in other sims, especially in a legendary Live For Speed.

 

Stalling is difficult, because you get 100% of engine's maximum torque without touching the throttle.

Even with relatively modern cars (2000 and later, I'm old) idle control valve did not allow full throttle situation, not sure about modern moderns that have throttle flap actuated by computer.

Another thing with those were delay, it did behave like too much smoothing in throttle code, but tech was just slow.

Older carburetted cars did not have anything, idle was adjusted by setting throttle butterfly open tiny bit and that was all air you got at idle, so stalling indeed was easy.

I made actually little experiment about idle torque earlier, you can see how engine stalls at point where flywheel torque reaches torque that is available in torque curve, that is 100% throttle situation, steep hill was requiring bit more that there was in engine, hence the stall.


In above experiment I had set custom parameter to Jbeam, maxIdleThrottle to 0.035 I believe, which is lot less than default 0.15 or 0.1 I believe?

I did even go as far as trying to hack maxIdleTorque from LUA to something what I would except, but I did lack of skill to do that.

It is likely that such feature has been set to help AI/keyboard/gamepad users or something like that, idk. However developers don't necessarily want to go to full on realism with this as people would complain about stalling cars then, but it is bit excessive way it is now, not sure if it is a bug though.

---

Tires then, well Goosah I think did post quite some time ago about their tire testing rig results and how tires are really good and I think tires are really good for most part.

Problem is when there is really lot of tire slip or burnout situation, where tires are like loosing more and more grip as more slip happens, which indeed appears to be bit excessive by the practical experiments.

How engine torque reaches it max with small throttle opening might be part of that of course as quite big part of engine torque comes available with quite small throttle opening.

While it is true that at lower rpm partially opened throttle butterfly allows all the air in that engine needs, there still is need to push quite bit of throttle pedal to get anywhere near maximum available torque, so this can probably contribute to things.

Then yet more another thing is how laid rubber under the tire is adding to friction even it is smoothing out the surface, sticky rubber surface just is quite sticky and as obviously laid in rubber is not simulated can affect things.

It is interesting how these aspect will change as clever developers figure out new improvements to game, anyone remembers how driving was in 0.8 game version and before that? Lot has changed for sure.

Also noticed there is now engine braking parameter in Jbeam, little bits get added all the time

 

I play on a controller (as I'm not able to afford a good steering wheel (G27 or G29)) & stalling the car seems somewhat easy, especially when I just simply dump the clutch without giving the car any throttle input. Even on a flat surface it will either stall or be very close to stalling & will "jump" forward a few times before getting back up to idle RPMs

 

If anything, brand new tires have less grip than tires that have been broken-in...

 

To be honest, I started to alter the 'standard tire slip coefficient' in the jbeam already a while ago
cause with their standard value the restore of a break out is just not realistic enough and this change
makes a lot of difference. It makes driving a lot more realistic and I think it's kept high by the game
so cars would behave more slippery for crash immersion or so idk ? Perhaps these experiments should
be tried with this altered tire grip setting and see what happens. No offense, but has the game been tested
by pro or equally experienced drivers on request of devs like other racesims have been ? I find that the heavier
vehicles (esp truck and bus) respond far more realistic then the lighter ones without my change. The lighter
ones just go woosh and their off the track, no successful restore to keep car undamaged at all..

@Ai'Torror stalling is quite reasonable imo BUT with use of good pedals. I think the axis of the pedals
responds good to the clutch. It takes a while to get used but once you know the threshold it's like
bicycle riding . Indeed like with truck, which has many gears, without trailer or load you should be
able to take off in higher gear and it works, really. Also remember, the moment you start moving is
really noticeable irl but takes a small calculation on the sim. Once you get the hang (with pedals) all
goes well (imo) .

 

What gearbox settings are you using in game?
I haven't been able to replicate most of what you've shown, from the idle to not stalling that is.
The burnouts I have been able to replicate. I've not tried that in real life, so I'm not sure if it's possible. I've not seen any videos of someone trying to do that either.
I'll try send videos when I can.

 

Full on realism of course, no assists or helps of any kind, realistic gearbox.

D15 with 5spd manual was used for idle test and while other test were done with wheel and 3 pedals, idle hill climb test I use did hold left shift, then shifted to gear and stopped holding left shift.

Key thing here is that don't have parking brake on, that surely makes you stall easy if you have it on, but blip throttle so that parking brake is not on and thing just will go on, that was low range I think, it will go fine with high range too, but it is easier to see flywheel torque reaching maximum possible value before stalling when using low range.

ETKi 3L model is also fun, if you take it to that ramp that leads to top of tall building at gridmap, drive to ramp with engine idling at 1st gear and it will climb that 10 degree ramp without issues, like a diesel tractor.


Maximum idle throttle seems not to matter much, as long as it is set to value that allows engine to start and idle normally, one gets 100% of available torque without touching throttle, while more realistic would be 10% or something quite small like that I guess?

There is lot of variables to idle strength though, carb is different from FI, old tech diesels are different from modern ones.

With old tech diesels like tractors, semi and older cars you don't have a throttle at all, just open intake, you adjust amount of fuel by boost pressure and loud pedal.

It is almost like if fueling would be 100% all the in game time so that even tiny throttle input gives maximum torque at low rpm, while at least with carb cars to get 100% torque near idle one would need to push pedal quite deep because there is not enough fuel to make power.

So when very small throttle opening gives more power than one would except, idle control code opens throttle if idle rpm is less than desired, then that leads to stalling becoming difficult.

Well, lot of speculation without proper examination.

---

For tires and how good / bad they are, this post by developer explains it really well:
https://beamng.com/threads/guys-come-to-talk-about-tires.58720/page-2#post-949560

Tires are really good, but certainly there is something odd when things go to extreme slip, with base model BX200, when you get tail slide with 2nd gear, slide just keeps going, tires keep spinning without opening much of any throttle, while IRL with vehicle of comparable power and less weight to keep poor quality tires spinning full throttle was required and then some.

1st gear of said vehicle was up to 43kph or so, 2nd gear up to 82kph or so, quite low gearing too and getting wheelspin on 1st did require serious effort, not chance of keeping tire spinning with less than 60% of throttle and that was 1050kg vehicle.

Well, I'm sure many of you have did your share of burnouts with various rustbuckets and have some idea how much it takes break traction loose or how much it takes to keep tire spinning at steep uphill as that said rustbucket certainly was not able to burn tire if touching the brake, we all have had our first cars...

Problem is that is it tires or is it torque/throttle application? Or is it chassis configuration and gearing related? Or maybe related to all little things contributing?

All I can do is little experiments and wonder, someone more able need to figure out what makes things work in clearly unrealistic manner here.
Well even that is often pushing with my diminishing ability to form even a sentence these days...

---

Full Tasti Cola load to semi and try to get going to even tiniest incline without slipping the clutch, that is pretty interesting challenge

Thing with big rigs was back in old days at least that you don't slip the clutch, you burn it, so slipping was never really done, enough low gear was chosen that you don't need to slip the clutch, or that was how I was taught to drive the things.

You have to remember gearing though, with tractor at max rpm top speed on 3rd or 4th gear might be quite low, depends from range of course, but gearing is something to keep in mind.
--- Post updated ---

I can't afford good wheel either, so I have Thrustmaster T150, which has freeplay from gears and softness of rubberband, while still somehow managing to be unsmooth so there is like 1cm blocks of movement and lot of noise from rubberband to fan, to base/mounting creaking with every turn.

In case it is still unclear, do not buy this wheel

Do you have clutch mapped to controller axis instead of button? I did try that once, but could not really do much with that, my brain processing power was lacking.

Anyway with digital button or gamepad axis, clutch usage might be quite abrupt compared to how you use it with pedals or with real car and clutch assists is there because of that I guess.

But then current setup of idle so that stalling is difficult could be also because they want it easier to controller or keyboard users to use clutch without assists.

Full on realism might not always work for gameplay reasons and that might be why such choice is made, it is a game after all and should work as such, having perfectly realistic is not always working too well.

 

Did some test with altered wheelSlipCoef and besides the vehicle handling better your point is still valid. It takes some throttle
to get the wheel(s) to slip but very little to stall, etki 3.0 > below 14%

What you say about slides continuously going, totally agree.

On the side of clutching, as I said, good setup and good awareness lead to good effects :

 

We were discussing about dumping the clutch and get going, but if you want to practice clutch control, here is setup you might find satisfactory in challenge level, get it going with that setup on that ramp and you have mastered your clutch foot well

IRL that would burn the clutch though.

With t-series thing is that torque drops to meaningless nothing at/below idle:


This thing can lug below idle as there is so much reserve torque compared to car's weight and at idle it gets all 118Nm without touching the throttle, it is like if T-Series would have 1500Nm at idle or something like that.

I doubt that idling IRL comparable car would be getting 118Nm by whatever idle compensation method it has.

Quite little things though, considering how so much is working in realistic manner in the game, but little curiosities.

 

I was referring to this quote below :

By which I mean the Semi IS fine dropping the clutch up til 4th easily and 3rd without a doubt, give it a pinch of throttle and you'll be taking
off in higher gears too. I drove Semi's professionally (and more then that) and it's looking quite good, honestly. Still doesn't mean you're
not right about the torque weirdness though, never looked into that.

I'll check out that mission, but irl, don't expect to go hauling heavy loads on a hill without using the throttle, only with very light gearing like
the 1st of 16 or 18. The 1st of twelve gears won't make it without throttle I"m guessing, though a little pinch suffices.
--- Post updated ---
Oh wait.. it's not a mission ? This (points to the gridmap) is the hill right ?

Well, I had to use the 617hp version to take off from zero on the hill. There is also this you want to know about heavy hauls :
Irl the vehicle is 'launched' on such hill in combination with a high level pneumatic handbrake which is released gradually while
releasing the clutch gradually as well. So I wonder if I connect the handbrake to an axis if that might lead to better results.

 

Gearing matters a lot.

For example, Dump bed config of T-series is following:
1st gear 14kph @ max rpm
4th gear 37kph @ max rpm

T-series dump truck has 3.55 rear end as default, other option is 2.93 rear end for T-series and nothing else.

Construction / dirt moving trucks probably don't get tall highway rear ends IRL, that is what kind variety I have driven.


I think recently I saw 3.91 to 6.71 rear end ratios listed for dump trucks from US based used trucks seller site.
For highway use, typical loads, terrain and distances would affect what rear end there is and it is pretty hard to say what kind of gearing there eventually is going to be.

60 metric tons and hill start, now there is few things that don't mix too well
--- Post updated ---

See first you put front wheels on that climb, then you stop, now get going without using throttle, it will not climb the whole climb of course, but you can get it started on incline if you don't mind frying the clutch.

Clutch settings were there to add to the challenge, instead of having linear simple clutch, have something with a sharp bite point like a race car. Takes few attempts, but thanks to forgiving clutch thermals it is doable!

My attempt was bit of fail, but you get the idea, too easy with default clutch mapping, much more fun with settings I posted earlier:


Because you get 100% of what is in torque curve without touching the throttle as game's logic is trying to maintain idle rpm, that allows you to get going quite well without using the throttle, add to forgiving clutch thermal/wear and you can ride the clutch enough to pull pretty good stunts that IRL would get you in trouble.
--- Post updated ---
Oh yes, I meant to put this in here but got confused and mixed up.

You can see in this short clip how modulating throttle is not really affecting torque at flywheel as idle maintaining code has already maxed out throttle, same thing as with pickup climbing video, but maybe even more clear to see how idle maintaining code is using 100% throttle at idle, despite maxIdleThrottle variable, which I think should limit idle maintaining code to set maxIdleThrottle value?:


Or then I just don't understand what purpose that variable has, but that would be nothing new.

 

You're going way further in detail then I can currently grasp.
But you must understand that 40.000kgs are not going to roll upon that ramp without throttling the engine on a
12 speed gearbox. It's not gonna happen. Doesn't mean you're not right about the other stuff. Ok, I hooked up
an axis controller on the handbrake, gave the T 16 gears (from my 0.15 config) AND (I forgot I did this before) changed
the friction (transmission jbeam) from 20 to 315 (I formerly changed it to 215). Now at least the 312ps T takes off in 1st
on that steep ramp. Don't mind the initial stalling at start of video, just a little tired..

 

Oh idea is not to get up that ramp with idle only, yeah that is not going to happen, idea is to get truck moving without using clutch while front tires are on that ramp, just little clutch control practice, completely irrelevant to other subjects here.

---

I think that it was at 80's that I saw a video where guy demonstrated how spinning tires on ice did make car easy to push sideways. I think they used summer tires on ice for that demonstration.

Inspired that old test I saw on tv, I wanted to test how much later grip is lost in BeamNG with wheelspin, so I put limited slip diff to Covet and tied it with nodegrabber from towhook so I could spin tires without car running away.

Next I did take nodegrabber and started pulling from front wheel to see how much force it takes to move car sideways, turned out to be 20kg or less, unless I did experiment wrong:


That is quite little compared to how much it takes to get tires sliding when pulling with nodegrabber, which was something like 700kg or so.

Those drifter guys manage to keep car under control while tires spin wildly, which puzzles me as if car weights 1000kg or more and it is thrown sideways, would there need to be more sideways grip left than mere 20kg?

However how much car pulls is not so hugely compromised, with D15 with base model tires I got 20000N with wheelspin and 25000N without wheelspin, but again, I have no idea about accuracy of that.

With all 4 tires spinning D15 is very easy to move sideways though, not much difference to covet in nodegrabber readings, but 40kg moves whole D15 very easily, while 20kg moves front end.

So who has a car and old tires so we can tie down a car and do the burnouts and pushing car to sideways in order to find out what kind of force is required?
--- Post updated ---
Tested D15 on ice, when not spinning tires there is more lateral grip than when spinning tires on pavement, or then measurement method is bugged out, but I doubt:


30kph or more wheelspin speed seems not to have much difference, lateral grip is pretty much same what I could observe.

Would need real world test to see how much there is lateral grip with spinning tires, but real world vehicles don't quite spin tires so easily, especially all fours.

Interesting.

 

I can add a bunch of information here, but I will have to do it later, so I am adding this comment in here so I can come find this again.

 

The problem is that tire lose grip "suddenly" upon wheelspin. The sliding friction is significantly lower than grip.
And 50% throttle is 100% torque at lower RPM. You can test this with acceleration at very high gear. Use the engine info ui app and notice that the flywheel torque is unchanged with 50%~100% throttle at low RPM. This is even worse with eSBR.
In the lua files that I discovered. The powertrain "logic" is to find the maximum power of the engine, then output the minimum value of the throttle X max power or the torquecurve. For example, if an engine is 100kW and the throttle is 20%, the engine would either output 20kW(at higher RPM, minus friction) or capped by the torquecurve.