The online racing simulator
Kunos Simulazioni - Assetto Corsa
(8882 posts, started )
Well the box method is not supposed to be the exterior dimensions of the car exactly. But you also need to take into consideration the location of the cog (box becomes longer, wider or taller if the cog is further away from the geometric center). And also like phil said how much is the mass centered near the cog or more towards the edges. So the jag has more weight on the outer edges than the v8sc. Thicker steel (maybe), more heavier bodyshell, heavier windows, bonnets, doors etc.
And that steering of yours is antiackerman Smile. Basically the outside wheels turns more than the inside wheel. Why it feels more stable and wonders less is because you have several degrees of toe and even more in corners.

eI found the car dimensions in various ways. First I just took the stock xj12c. Then in blender after I had finished the model so the exterior dimensions should be fairly close I then measured it in blender. I did lots of cross referencing to get the blender model correct so I now I have measured most of the stuff from that. Frontal areas for aero, dimensions of the car...
Well, this is odd. Yesterday I did steam integrity check for ac and now for me the car seems to have a lot more bouncyness. Before the bouncyness was not an issue for me but now I'm getting oscillations which causes the car to jump around. In the nordschleife the parts which used to not be problems (the early part of the track where you use those concrete/tiled runoffs to cut corners a bit. Or the super fast left hander into schwedenkreutz. Before I noticed there was alot of bodyroll and some bouncing but never so much I'd miss my lines. Now I'm bouncing off the track...).

Also I think I have solved the numb steering issue in the center. Scrub radius. I never assumed it would make such huge difference! First I moved the whole suspension sideways 10millimeters. A lot heavier in the center. Reduced to 6 millimeters and now it feels better I think. There is more weight and resistance in the steering. Which of course makes sense considering what the scrub radius does but I never assumed even 6mm can make such big change. I also moved the steering arms a tad.

If you want I can share the files.

Next I need to go back and double (quad-something-check actually...) those inertias. This time I'll go all out and decide that box system so I can get as accurate as the positioning of the individual parts on my calculator.
Lol I calculated the box dimensions.
What I have now in ac: INERTIA=2.4,1.5,5.2
What I should use according to my calculations: INERTIA=1.89,0.56,3.75
I think I've made a mistake... The actual inertias for the car were 2130 yaw, 469 roll and 1738 pitch. With mass of 1450kg.

You can't explain that.
if you want send me your data I can run the calcs on my side to check for errors
I calculated in on paper Big grin

I can get you photo!
I guess this is too late now, but I fixed the steering issue in a simpler way. Consider a vanilla Jag:

Original
WBTYRE_STEER=0.108, -0.193, -0.149

Changed
WBTYRE_STEER=0.108, -0.163, -0.149

Optional: -0.01 static toe.
I get the same results.

This makes no sense...

That Dynamic Index in AC...is it supposed to be for yaw inertia only? Maybe it's an average of yaw + pitch + roll indexes ??
I just spent the time I could be calculating those dynamic indexes to post a rant on the ac forum about the topic. Big grin
I'd assume it is yaw. So if you calculate one you can probably use the same multiplier for the other inertias as well. So if the yaw dynamic index is too low then raise it and raise the other 2 as well using that same multiplier?
I'd go with the tiny box you calculated and I'd bring the tires DX DY closer to each other. See how that feels

I get a yaw DI of 0.72 and the tiny box gets a DI close to that in AC.
The formula_k uses
INERTIA=1.9,0.8,3.4
And the box:
INERTIA=1.89,0.56,3.75

I have some doubts...
Aris : - The car inertia in the car.ini is calculated as a box. The calculations inside AC are made in a way that if you put the rough dimensions of the car, you get a pretty good inertia values.

So it isn't as simple as take the known inertias and size a box to match that. There is a multiplier or another calculation built inside AC.

Still he is saying that Dynamic Index is a very good tool and the only way to get close to 0.72 yaw DI on your car is to have a small box
Maybe the weight of the car is the multiplier.
No it has to be something on top of that.

What I mean is the way you calculated the box dimensions cannot work in AC. For AC the box dimensions are not the dimensions of a box that has the same inertias and mass as the car. This is what you calculated, a box(x,y,z) that weighs 1450 kg and has 2130,469,1738 inertias. AC does not want that as an input.

About the Dynamic Index, changing box width does not affect its value. Changing box length, box height and mass(wtf? it really shouldn't) affect its value. That means the DI we see in AC is based on mass and pitch inertia alone. Not yaw, not roll.
Quote from PhilS13 :Changing box length, box height and mass(wtf? it really shouldn't) affect its value. That means the DI we see in AC is based on mass and pitch inertia alone. Not yaw, not roll.

Mass does affect the yaw dynamic index...

e.g. Yaw DI = k^2 / ab, given that k^2 = Iz / m

As far as I can tell AC wants the input for car.ini as INERTIA=width,height,length (whole vehicle values i.e. not track,height,wheelbase).

At least that's what every good mod seems to use (except in cases like the AC Cobra which uses a lower height value in car.ini. I assume that this is due to the fact that it doesn't have a roof and the windowscreen would only count for 1% of the cars overall mass or something).
In light of this inertia discussion, I made some small changes to my mod car and I think it feels better: https://www.mediafire.com/?1lmlkymy7lk6acq,ukf7dvrf38fzqar

Things I noticed:
- handles like shit with DI 0.72. Just, no.
- increasing the width value to just 0.01 beyond the actual width of the car caused it to want to rotate constantly, but 0.01 less feels great.
- values significantly lower than the physical dimensions of the car cause the car to drive as if it was on rails and handle bumps far too well.
Quote from JackDaMaster :
Quote from PhilS13 :Changing box length, box height and mass(wtf? it really shouldn't) affect its value. That means the DI we see in AC is based on mass and pitch inertia alone. Not yaw, not roll.

Mass does affect the yaw dynamic index...

e.g. Yaw DI = k^2 / ab, given that k^2 = Iz / m

As far as I can tell AC wants the input for car.ini as INERTIA=width,height,length (whole vehicle values i.e. not track,height,wheelbase).

At least that's what every good mod seems to use (except in cases like the AC Cobra which uses a lower height value in car.ini. I assume that this is due to the fact that it doesn't have a roof and the windowscreen would only count for 1% of the cars overall mass or something).

Mass is already in the Iz. You divide by m to get k. Both masses cancel each other out.
Quote from PhilS13 :
Quote from JackDaMaster :
Quote from PhilS13 :Changing box length, box height and mass(wtf? it really shouldn't) affect its value. That means the DI we see in AC is based on mass and pitch inertia alone. Not yaw, not roll.

Mass does affect the yaw dynamic index...

e.g. Yaw DI = k^2 / ab, given that k^2 = Iz / m

As far as I can tell AC wants the input for car.ini as INERTIA=width,height,length (whole vehicle values i.e. not track,height,wheelbase).

At least that's what every good mod seems to use (except in cases like the AC Cobra which uses a lower height value in car.ini. I assume that this is due to the fact that it doesn't have a roof and the windowscreen would only count for 1% of the cars overall mass or something).

Mass is already in the Iz. You divide by m to get k. Both masses cancel each other out.

Thanks. I thought I was missing something in regards to that.
fml. For some reason yesterday I have decided to delete tyres.ini, suspension.ini and car.ini. So now I need to make all those changes again I made yesterday.

But who cares. All the important stuff is right here in this image:
So the M3 E92 has DI of 1,119...
1.079 here

I tried your car with E92 specs

0.52 Weight dist
1580 kg
2.76 wheelbase
1.8,1.43,4.62 box

Gave me a DI 1.275

Then I found out that changing HUB_MASS changes DI...nice. As long as we don't how DI in assetto is calculated trying to match a real life DI with this obscure value is worthless
The updates for Sierra Cossworth and Ibiza mods from Bobskype, too good now, especially the stock Sierra.
Don't forget the NSX

Kunos Simulazioni - Assetto Corsa
(8882 posts, started )
FGED GREDG RDFGDR GSFDG