Physics Center of gravity, inertia of rotating car etc

[dud]

I'm a little baffled about how AC is modeling some of this.

For data/car.ini there is

INERTIA=1.60,1.55,4.70  ; car polar inertia. Calculated from the car dimensions. Just enter the generic width,height,

Does that mean does not have a concept of a given car spinning with a mid-engine layout has less inertia than a car with engine and transmission at the outer ends spinning at the same frequency?


Related, how is the height of the center of gravity coded? The center of gravity in the horizontal pane is suspensions.ini:

[BASIC]
WHEELBASE=2.812          ; Wheelbase distance inmeters
CG_LOCATION=0.52

The height is separate for front and rear, each having one of these lines:

BASEY=-0.125 ; Distance of CG from the center of the wheel in meters. Rear Wheel
 Radius+BASEY=Rear CoG  Actual CG height =(FWR+FBasey)+
(RWR+Rbasey))/CG_LOCATION%

I think that's too voodoo for me. Do they model CG height separately per axle?

Can somebody translate that into, say, what would I do if I replace a steel roof with a carbon fiber roof and decide that the height of the center of gravity should now be 0.2 lower?

 

[Stereo]

The CG position is actually fixed at the 0,0,0 coordinates of the car, you move everything else relative to that. So BASEY says where the level of each axle is compared to the CG (negative means axle is above cg) That's only at the design suspension travel though, depending on springs etc. the wheels will end up sitting higher or lower than the starting point. Likewise the visual car offset in car.ini says where the 3d objects are relative to the CG.

If you want less inertia just decrease the inertia length in that direction (mid-engine cars are 'shorter', so instead of 4.7m, maybe make it 4.2m)

 

[ltcars]

The CG position is actually fixed at the 0,0,0 coordinates of the car, you move everything else relative to that. So BASEY says where the level of each axle is compared to the CG (negative means axle is above cg) That's only at the design suspension travel though, depending on springs etc. the wheels will end up sitting higher or lower than the starting point. Likewise the visual car offset in car.ini says where the 3d objects are relative to the CG.

If you want less inertia just decrease the inertia length in that direction (mid-engine cars are 'shorter', so instead of 4.7m, maybe make it 4.2m)

Can you calculate the inertia length or is it just guessing?
And for the width and height, do you insert the real width and height of the car?

 

[dud]

But BASEY is negative on almost all cars.

So the position is not the "axle" (the wheel center's), but the upper fixed position of the suspension, from where suspension travel range etc is offset toward the ground? ETA: no, it says center of the wheel. I think some mods might be suspicious.

Either way if I want to lower CG by 20cm I just subtract 0.2 from both BASEY.

Thanks!

 

[mclarenf1papa]

The INERTIA comment isn't really correct...it's alright for beginners but otherwise it's not actually true. The moment of inertia of the car is calculated from those "dimensions" and the car's mass. The INERTIA input is actually the dimensions of a box. You can't really calculate it as you'd either need to measure it or do a sum equation for every single part of the car (which typically can be done if you have a full CAD model). The data exists, but you'll have to estimate if you don't have it.

The "BASEY" is essentially just there to set a reference point for where the supension mounts on the chassis are (it's just a vertical offset - like Stereo said).

Also, if your CoG drops 0.2m via a bit of carbon, that must have been one hell of a steel roof.

 

[dud]

Yeah. Messing with BASEY just gets the wheels up, through the fender.

So the only way for me to tell the engine that CG is now higher is to put in a bigger height of the overall car.

I think it might be possible to do spin tests in Assetto Corsa, on a special track with a part of very low friction, to come up with some examples on how to tell cars with similar dimensions that either have low inertia at same mass and spin frequency (mid-engine) or higher inertia (engine and/or transmission off-center).

This could go a long way to e.g. simulate the difference between a 718 and a 991 which everybody can experience in Porsche's Atlanta center (both with the kickplate and on the wet skidpad). The 718 turns+spins faster but also loses spin frequency faster under sideways tire friction.

I am (would be?) surprised if Kunos simply ignored this. In any case it would be great to have a more data-based way to tell the engine that a car has a particular high center of gravity. Think of a 4-seater convertible with a hardtop (400 pounds extra) and the top open or closed. That's 400 pounds that go lower and to the rear.

Or think of a Corvette versus a Camaro. Or current Aston Martins. Front engine, but the engine moved forward so that it can be lowered without interfering with the front axle. I think the MX5 and S2000 also do this.

 

[ltcars]

Is there a way maybe in Content Manager to visualize the CoG?

 

[dud]

Well, I think it would be easy enough to do if it is just a box and the CG is in the center. Would be nice to draw the 3D model in the same coordinate system to see how "off" the CG-dictated dimensions came out.

 

[mclarenf1papa]

Jumping to conclusions isn't going to help you. CoG and inertia are modeled correctly in AC.

The reason the visuals changed is as Stereo said, the CoG is at the origin of the car model (or rather, the origin of the car model is at the CoG). Since the visuals have nothing at all to do with the physics, you need to adjust the graphics offset in the car.ini to make them appear correct again.

 

[JackArse]

So fellas, from what you've said I've had a play and I don't get it......
I relocated the wheel graphics to be correctly under the wheel arches. I changed the suspension rod length and the basey to get the ride height and wheel to cg relationship how I wanted it. how do i now lower the car and also the wheels relative to the car down onto the ground? Is this where the car.ini ride height pick up point comes into play?

 

[Stereo]

Decrease rod length to lower the car "by using coilovers" or make basey more positive to lower it "by cutting out the entire suspension and reattaching it higher on the chassis".

If you mean the tires aren't touching the ground, decrease the tire radius in tires.ini.

 

[dud]

Speaking of this thread, I would still like to know how to change rotational inertia, aka declare how far from the center engine and transaxle are.

 

[mclarenf1papa]

Speaking of this thread, I would still like to know how to change rotational inertia, aka declare how far from the center engine and transaxle are.

car.ini [INERTIA]

 

[dud]

car.ini [INERTIA]

Did you see the first post in this thread?

 

[mclarenf1papa]

Did you see the first post in this thread?

Yes. Not sure where the confusion lies (perhaps in the description of the parameter? It really doesn’t have to do much with the dimensions of the car...that’s just for if you have no idea what you’re doing).

Edit: the actual inertia values are calculated based on a box of uniform density. The dimensions of that box are the INERTIA x,y,z components.

 

[Cote Dazur]

Yes. Not sure where the confusion lies (perhaps in the description of the parameter? It really doesn’t have to do much with the dimensions of the car...that’s just for if you have no idea what you’re doing).

Edit: the actual inertia values are calculated based on a box of uniform density. The dimensions of that box are the INERTIA x,y,z components.

Sorry for digging this old thread from the dead, about the inertia, if it is not the dimension of the car and is a dimension of a box, where can we look to get the dimensions of that box.
like for example a Miata mx-5 2021 or any given car model?

 

[mclarenf1papa]

Sorry for digging this old thread from the dead, about the inertia, if it is not the dimension of the car and is a dimension of a box, where can we look to get the dimensions of that box.
like for example a Miata mx-5 2021 or any given car model?

You can't look anywhere for it. You'd need to acquire MOI measurements for the real car, calculate the sprung mass component, then solve for the box that fits it.

 

[Cote Dazur]

You can't look anywhere for it. You'd need to acquire MOI measurements for the real car, calculate the sprung mass component, then solve for the box that fits it.

From reading more on MOI measurement IRL I perfectly understand that IRL, their is much more to getting MOI measurements than just knowing a car width, length and height.
But as far as AC is concerned, looking at Kunos cars, it looks like what they enter/use is width, length and height. Why are you saying, if I understand you correctly, that it is not accurate for AC
Could it be close enough within the obvious limitation of the calculations in the SIM.

 

[Stereo]

It's just a method for guessing when you have no IRL data to work from, there is no limitation on how AC does inertia, it behaves correctly as long as you enter the numbers for only sprung weight correctly.

 

[mclarenf1papa]

From reading more on MOI measurement IRL I perfectly understand that IRL, their is much more to getting MOI measurements than just knowing a car width, length and height.
But as far as AC is concerned, looking at Kunos cars, it looks like what they enter/use is width, length and height. Why are you saying, if I understand you correctly, that it is not accurate for AC
Could it be close enough within the obvious limitation of the calculations in the SIM.

AC calculates the MOI from those dimensions, so they need to be representative (the dimensions of the car are not).

KS cars shouldn't be the gold standard for doing things... It's nowhere near close enough, and it gets worse the more compact the car is internally. Some of the KS racing cars have up to 90% error per axis in their sprung MOI. That is, their MOI is 90% higher than the correct value.