Suspension damping and wheel modelling.

See the highlights in the quote

 

I honestly think you may be onto something with the heat thing too to be honest. It makes sense. Every single bump and force in these tires need to be absorbed somehow by the beams... Might be a way to make some forces just... disappear as heat to force the tires into behaving a bit more accurately...

Also, I'm not saying the slip angle thing actually happens in code because it's coded that way. It's just that it seems that's what happens. To me, it feels like there's an area between "sliding" and "not sliding" where you have no grip whatsoever. When you're going at a certain high speed most of the cars drive perfectly straight due to there not being resonance. But at certain speeds, with the gas on, the rear ends of a lot of cars just start boating around like they're on ice, and there's barely any way to counteract that by steering... Yet when you do actually start leaving tiremarks, it seems a bit grippier than when it's just "floating"... If that makes any sense at all

As for the contact patch thing they do in other games: Problem here is that these tires have to behave during rockcrawling and such too, so there wouldn't be a proper way to do that I think...

 

Indeed, rockcrawling is where this model really shine. I LOVE it!

And the heat thing.. That is what a damper do. Convert kinetic energy into heat. How else could damping occur? I have no idea how the nodes and rods work, but there has to be SOME damping, and many components or materials or whatever define the properties seem underdamped. But then, looking at stuff in superslomo reveal alot of flex in even the seemingly rigid components, so...

You made cars, how are properties of a material defined exactly?

 

Well that's a pretty heavy content you bring here guys, I admire that really.

At the same time I hope your exchanges will bring something to improve that behavior, as I can really feel something is not right here when driving the cars.

I don't know much like you, but I can't stop me to think that when looking at the videos you have done I see the wheels spinning like they are not spinning at the center of the axle, why does it look like that ?

 

I guess they do convert energy into heat... but I kinda meant, perhaps we need some different kind of damping for tires, more or less progressive for example.

you can simply set the spring and damping amount on any beam you create, which it then transfers into any node. Nodes have weight. If you have too low weight with too high damping or spring, the whole thing explodes or starts deforming on its own...

So basically you can make a material any way you want. If you make something rubber, you're probably gonna have a lot of damping but a lot less spring. With metal you're gonna have more spring but less damping.
Other than that, you can set the material on a node, but that only changes what type of particles it makes when it touches another material. Ie rubber on asphalt will create smoke, metal on asphalt sparks... That sort of thing. Other than particles, it's completely up to weight, spring, damping, deformation and breaking values you put onto your nodes and beams... That's the versatility of this physics engine

 

One of the main reasons the tires\wheels bounce like they do under heavy loads is the flex in the suspension arms\steering rods. If somehow they can be made stiffer i am pretty sure the problem will be lessened greatly. The only negative side effect i can see to doing this is that suspension durability\strength will be over the top and will be difficult to bend\break realistically.

 

I'm curious to see how a Go kart would behave in that simulation with that issue, I'll have to wait someone knowledgeable enough add one

 

I think the tire model is actually really realistic. First of all, let's consider the fact that BeamNG is one of the few games that actually has a tire model, and not a positional physics model (ie accelerator = true, therefore car moves forward). Also it's still in alpha. Think of minecraft when it was in alpha.

The issue here is that the tire is simulated with points, not a surface. This really improves performance and is fairly accurate. However, when the tire is spinning but the car isn't moving, since the energy goes to the path of least resistance, there is less resistance in shock compression then burning the tires, SO, the tire 'skips', as it forces the shock up.

So it would look like the solution would be to increase shock force necessary to compress it, however, as gabester said, the stiffness can't be increased.

The real issue is simply the 'play' or 'slack' in the tire beams which prevent it from spinning perfectly.

And finally - if anyone has ever driven and old RWD on asphalt and floored it - especially if it's light in the back - you'll see that the tires bounce and skip as the tires aren't perfect.

 

Randomshortguy, you are referring to wheel hop, which is caused by resonance from too little damping and too much play in bushings in the drivetrain. Same thing we have here, in reality, but caused by different things. In real life it's caused by imperfections in the road, in the game from the segmented tires exciting the resonance.

 

Actually, no. That's caused more than anything by soft bushings in the motor mounts and suspension.

 

After the recent tire changes, i think its the suspension and wheels are pretty much perfect. Gabe did an amazing job with the new soft tire modifications.
The only vehicle I have only noticed doing the bounce is the truck when going into a handbreak turn (which actually is realistic).
The Truck has old school leaf springs in the rear and doesnt have a sway bar to compress both sides of the truck's suspension which results in the bounce as the spring couldn't handle the pressure asserted on it, for those of you who are curious.
For the tires bending out of shape, thats realistic as well. http://www.youtube.com/watch?v=XttOe4rerLo
Go to the 30 second mark.

In the videos where your wheels are going in and out of alignment while moving at 2mph, I have tried to reproduce but i couldn't.
I took the little hatchback and let it roll from a stand still and I didnt see the wheels doing the twist at all.

 

Tires bend out of shape at ANY load, it's how they excert force. The real issue is segmented wheels that impose a HUGE load on the suspension, much more so than a smooth wheel will do. I think the actual grip and deflection of the tire carcass behave fine, to be honest. The real issue is segmentation, which the devs have stated is a known issue with no workaround at the moment.

And for the guys saying that the cars drive fine; No, they don't.
To the devs saying we don't know anything because we haven't dropped cars from 20 feet; We talk about normal, everyday driving conditions. Even on the grid map with mathematically perfect surface the problems arise. Sorry, but you can't blame that on potholes.

 

For the wheels jumping in and out of alighnment, I would like to bring this video to light...

http://m.youtube.com/watch?v=91dmTS-UvDc

When the front end lands, you can clearly see just how much flex there is in these suspension systems. Granted, this is a bit of a drastic situation, but my point is that the whole suspension system on these vehicals are behaving realistically.

As for the wheel floating around, Mythbuster is well-versed in the physics system of Rigs of Rods which is basically the same underlying system that BeamNG uses. Why the floating is happening in BeamNG more than RoR, I am not entirely sure. I do know that occasionally in RoR whenever I did feel my car was a bit floaty, I usually just combatted it with more downforce to aid in traction. Also as far as the bouncy, jittery feeling at low speed goes, I urge you to get on RoR and take the Lamborghini Countach for a spin. It had tires with only 7 segments in them (I think) and still was not floaty at all, granted Lifter did work some serious black magic with that thing to make it handle like it did.

Now I know in RoR, tire rolling resistance was a huge problem. This was suposedly solved in BeamNG with some new tire physics, perhaps at a cost to controllability and traction? I am not really sure, but I am only speculating.

As for the turn in of the vehicals being really slow, I may be able to offer an explanation to that.

As we are well aware the tires in the beam system are assembled by 2 parallel rings of nodes interconected with beams. This gives us a rather accurate and yet beautifully simple means of simulating a tire. In BeamNG the system was improved drastically by also symulating the rim in the tire in much the same manner. This of course leaves us with a slight draw back. The amount of play that each node can move with reffrance to each other.

What I am getting at is that tires are made of soft, malleable rubber that alows them to flex and contort to the contours of the road. Using that mentality, in RoR anyway, the tire was constructed of soft, malleable beams that would squish and flex in much the same way that the rubber of a tire would. However, as was pointed out earlier, rubber is made up of trilions of molecules which are interconected with bonds. These molecules act as nodes, and the bonds act as beams. As is noteworthy, they can't move very far and as a result, that rubber can only flex so much.

In the Node/Beam tires on the other hand, the nodes can move a great deal. This gives us the ability to have the tires flex and contort in ways that would otherwise seem somewhat unrealistic. Folding up the tire like a pie pice for example. Also, I believe this gives to much flex of the tread nodes when you go in for a turn. For instance, skip to 2:03 in this video, then pause it and look at the tire tread (only vide I could find that had it)

http://m.youtube.com/watch?v=JCC1_WtoZxk

See how the tread flexes away from where the tire is pointing while going through a turn? In real life, a tire only does this under high load situations, like the high speed turn in that track. Also a tire has steel belts to help counteract this problem by keeping the tread nice and stiff.

Our Node/Beam tires in the other hand can flex much more easily. Many rock crawlers (without the wheels2 section) in RoR this is very eisaly visible. Just find a low tire preasure rock crawler or truck (I think the Unimog does it) to test this. Just park the vehical, step on the break, and steer. Any real life vehical the contact patch will just slide leaving very little deformation for the tire to have to do. In our Node/Beam tires however, you will see that the patch on the ground stays completely stationary and the whole tire will twist to keep that patch on the same spot. Start to move and it will straight out fine but never the less, at speed you can see how this would be detrimental to the handling.

So... I wonder, if it is not already in place, if the generated beams around the outer band were stiffer to simulat that steel belt, would that improve vehical response? Could have the potential to I supose, but there is no way of knowing until it is either confirmed or denied through testing.

That's just what I was thinking anyways.

Also I apoligize if there are random words in there that make no sense. I typed this on my iPhone and you all know how autocorrect can be

 

That could be it. Modern tires are alot stiffer in the sidewall than the running surface.

 

Well that's great to hear!!!

That does make sence though if you think about it... Say the wheel is spinning and a force moves a node slightly closer or further from the axle... I can see how that would escalate out of control quite quickly... So this is great to hear!!!

 

Are these updates to the physics something hardcoded, or to the files for the actual parts on the cars?

I'm trying to figure out if these changes affect the separate part files myself and others make for mods, or if we have to remake those parts after these updates.

Thanks.