tyre/tire selection

[zimluura]

What if we use a fixed offset for all wheel meshes, like 0.15 or something. The real offset would be known as 0.15*rim_width for brake positioning.

if we use a fixed offset for all wheel meshes...how about making it zero? that way it gets pushed out as far as you define the offset. and i just redo all the meshes so all vertexes are on the negative side of the x axis (some verts just barely so). we might need, eventually, a definition file for the wheels so we can check what the thinnest wheel_edge/tire they can be used on, but we will probably want one for weight if nothing else. and really so long as you don't try to text-edit a wheel too thin you'd be good. after that, i think, they should respond to offset displacement properly. you see any design flaws, or does it sound good?

i've got the rotor mesh generation almost done. still need to calculate normals, which should be super easy. also need to get some basic textures together.

btw - i'm only a so-so texture artist, and frequently find myself taking shortcuts to get things checked in. so if anyone finds them to be an eye-sore they should go right ahead and fix/tweak them.

[NaN]

I'd like to make sure we are talking about the same offset. With offset I mean the wheel hub offset from wheel center (see wheel cross section images posted). If all vertices are on the negative site how do we make sure that the outer spoke vertices fit the rim edge? The artist would have to figure out the correct offset by himself then.

[zimluura]

yeah, that is totally the offset i was thinking of.

and that *is* the problem i see with it. at first i was thinking it would just be up to the artist to make sure the wheel and wheel_edge play nice together. after all, they sorta have to do that anyway if they want a flange displacement to look good.

maybe we could use something measured from the wheel.joe itself? so it determines the width through a bounding box calculation, and then you know the smallest tire you could possibly fit there (baring any ben-hur spikes ) if we want to have ben-hur spikes i think we'll need some kind of wheel.ini to be ok anyway.

i think maybe we have an issue because a regular wheel you get (irl) will have a specific offset. and a car will want wheels with specific offsets. and in a race-car sim we want to be able to freely interchange without regard to the multiple $1000 at the machine-shop having to modify the hub assembly to accept wheels with a different offset or vice-versa.

so, maybe the artist can figure out a range of offsets that his wheel will work for??

[zimluura]

thinking back to the wheels i already converted.
i think i set them up so all vertexes would be on the + side of the x axis, in a hope that it would be easiest for an artist to meld it with the wheel_edge. and, given the option for a flange offset those will always have to match up perfectly.

and, i guess, since that's on the outside it is more noticable than the wheel mesh intersecting the rotor. still can't quite see a way to get the wheel mesh to be sandwiched (effectively) between the wheel_edge and the rotor, without resorting to a text config file.

btw - rotor is all built (checked in a few days ago).

[NaN]

Where should I place the rotor, at wheel center or within the offset(right to the center)? Should the offset be fixed? What about the thickness of the rotor, fixed or relative to diameter? I've also been thinking about adding a brake caliper mesh. How to figure out the placement of it? I've been thinking about creating a wheel template in blender(tire, rim, brake rotor) and fix their relative positions. Do you think it is a good idea?

[zimluura]

placement of rotor: with most cars i think the rotor goes outside the hub assembly, but i think all cars have it outside of the pivot point (knuckle, i think it's called) on the steering.

fixed offset: i'm not 100% sure of the best approach. i think the 2 most important visual things, at least for the moment, are:
1) for the artist's overall wheel style, the wheel mesh needs to line up just-right on the wheel edge.
2) that we want to be able to make wheel styles that will work on any car (at least sorta fluidly). to have the same wheel style be able to fit on more than one tire thickness will mean having it be variable (i think).

another thing about the offset: i assume that the front-wheel pivot point would be pretty important to the handling characteristics of the vehicle. some of the vehicles we have might not have this point even close to correct. and therein lies another issue: that we could do a wheel offset, and that it shouldn't be the same as the offset from the pivot to the center of the wheel. i think right now the steering rotates around the center of the wheel


i have a caliper mesh already put together. wanted to go give a good look at my car's before uploading it. not sure how simple it should be modeled. i guess it does need to have both inside and outside modeled for open-wheel cars. but the part in between could probably be really sparse on the details.
for placement i think the same x-spot as the rotor. then perhaps an angle of rotation around the rotor (definable in the .car file)? might make it go upside down on some cars (which, for all i know, might be how it works irl). until there is a robust system for buying and storing car mods in-game, we probably don't really have to worry about caliper clearance.

i'm less in favor of a wheel template in blender. but it's just sortof a principle thing, i'm really starting to think there's no real alternative :|

is there anyone here who knows ALLOT about cars? i've never taken my suspension apart so i'm going off of models and a basic understanding.

[NaN]

I've implemented new car config code (converted TC6, 360 already). I'd like to have some feedback before checking it in.

The number of wheels/suspensions is fixed to four(will change in the future). There has to be at least one coilover, tire and brake(names can be arbitary). Tire types will be in carparts/tire and tire textures in carparts/tire/textures.

One issue is wheel mass and inertia. I'd tend to put them into tire definition.

TC6:

Code:

[coilover-front]
spring-constant = 72000
bounce = 7500
rebound = 5000
travel = 0.32
anti-roll = 3700

[coilover-rear]
spring-constant = 68000
bounce = 7500
rebound = 5000
travel = 0.32
anti-roll = 8000

[tire]
size = 215/50r16
type = touring
texture = touring

[brake-front]
friction = 0.73
max-pressure = 3.2e6
bias = 0.61
radius = 0.17
area = 0.015
rotor = rotor_worn

[brake-rear]
friction = 0.73
max-pressure = 3.2e6
bias = 0.39
radius = 0.14
area = 0.015
handbrake = 1.1
rotor = rotor_worn

[wheel-0]
orientation = right
position = 0.75, 1.26, -0.43
mass = 20
rim = 5_spoke_m
tire = tire
brake = brake-front

[suspension-0]
coilover = coilover-front
position = 0.66, 1.26, -0.07
hinge = -3.70, 1.1, -0.43
camber = -1.33
caster = 6.12
toe = 0.0

[wheel-1]
orientation = left
position = -0.75, 1.26, -0.43
mass = 20
rim = 5_spoke_m
tire = tire
brake = brake-front

[suspension-1]
coilover = coilover-front
position = -0.66, 1.26, -0.07
hinge = 3.70, 1.1, -0.43
camber = -1.33
caster = 6.12
toe = 0.0

[wheel-2]
orientation = right
position = 0.75, -1.28, -0.43
mass = 20
rim = 5_spoke_m
tire = tire
brake = brake-rear

[suspension-2]
coilover = coilover-rear
position = 0.66, -1.28, -0.07
hinge = -3.70, -1.0, -0.43
camber = -0.45
caster = 0.28
toe = 0.0

[wheel-3]
orientation = left
position = -0.75, -1.28, -0.43
mass = 20
rim = 5_spoke_m
tire = tire
brake = brake-rear

[suspension-3]
coilover = coilover-rear
position = -0.66, -1.28, -0.07
hinge = 3.70, -1.0, -0.43
camber = -0.45
caster = 0.28
toe = 0.0

[zimluura]

looks allot cleaner than the current one. all those tire parameters i don't understand, mercilessly copied from another car.

on wheel mass and inertia, i think it would be good to have them defined somewhere else, hopefully in a way that can be figured mathematically from the tire code, wheel type, and rotor weight. but i think it might need to have a mass and/or inertia parameter to compensate for the hub and the axle so if it's kept...have it default to zero maybe, and call it something else: additional_mass, infrastructure_mass?
in the tc6 i think i took up the inertia on the rear wheels to simulate the load of the drive-shaft, still that was mostly a work-around for the high powered awd.

[NaN]

be figured mathematically from the tire code, wheel type, and rotor weight

I prefer this way too(will add code for this). But it makes it difficult/impossible to fine-tune the settings. One could still influence them by tire/rim(wheel) selection. Do we want to be able to set wheel weight/inertia directly?

[zimluura]

i was originally thinking that most cars would have its own oem_wheel.joe file in its own directory. 'cause some wheels are really tied to the car that they came with.
for instance:
http://www.cartype.com/pics/3934/full/el..._wheel.jpg

got all of the car wheels imported and scaled in blender a while back (put the exporting on hold while we figure out how to position things in the .blend).

but in the case of cars that have an oem_wheel.joe, then those should have a way we can set the mass and inertia for the wheel itself. but it's pretty rough because on most of the powerful rwd sports cars, the rear wheel will be much wider than the front, and so at least some of it has to come from the tire code, and assumptions about the wheel mass increase based on the tire's size.

but perhaps a mass that gets added to the computed wheel mass. maybe inertia as well though i don't think it matters all that much in the case of hubs and axles where the radius is comparatively low. but i'm not really sure, it's been years and years since i did any rotational inertia equations.

so if making say the 10-speed bicycle we'd definitely need a custom oem_wheel.joe with a really thin tire code, which with the custom wheel mesh could be fine-tuned (oem_wheel.cfg maybe, or just in the .car file). but if they put some 22" spinners from a cadillac pimp-wagon there would be a huge performance decrease.

also i think practical tuning will just get better. the # Lateral force, Longitudinal force, and Aligning moment are certainly things i've never touched before, not really sure if other car builders do or not. frankly, for all i know the settings i copied for the tc6's wheels are equal to a monster truck, or perhaps a baby carriage.

btw - been working on the LE all week. the headlights are evil evil evil. went to a dealer yesterday to take a bunch of photos of them (at strange angles). hopefully i'll get it up in the next week or two. man, swoopy cars are so pretty, but i'm starting to hate modeling them.

[NaN]

zimluura could you please scale and position the wheel meshes. I'd like to position them relative to the outer rim edge.

unscaled:


The outer radius of the rim in the image is 1.08/2. The outer width is 0.75. The inner radius is 1.0/2, inner width is 0.9 * outer width. The center of the rim vertices is 0.75/2 on the x-axis => rim outer edge vertices have x=0.

scaled 225r16 (wheelmesh*0.406, rim*(0.225,0.406,0.406):


The inner radius and width will vary a bit in the generated wheels. We take care of this by a forced intersection of the wheel mesh(spokes) and rim. The wheel mesh radius should be close or equal to outer rim radius to avoid holes in the wheel+rim mesh. The wheel mesh center/origin should 0.0. It will be placed relative to the rim outer edge(position = 0.375 * tire width).

The blend file: http://www.pasteall.org/blend/2910

[NaN]

About the wheel mesh. It's outer vertices should be on the negative side(x <= 0) as rim outer edge is x=0; The inner vertices of the spokes can protrude into positive side if needed(go beyond rim edge).

[zimluura]

The inner radius and width will vary a bit in the generated wheels.

just so you're aware: the outer ring of the wheel edge should vary more than the inner. when i got the wheel edge function uploaded i reduced the tire's sidewall to compensate for the overlap that is typical for a wheel flange.

The wheel mesh radius should be close or equal to outer rim radius to avoid holes in the wheel+rim mesh. The wheel mesh center/origin should 0.0.

you mean y=0.0, and z=0.0, correct? the .blend file you posted has a negative x value for the hub. or, am i completely misunderstanding you?

also, it looks like the .blend file has a smaller wheel mesh than you suggest. it looks like its outer vertexes along the mesh radius are closer to the inner radius on the edge template. in this blend file, there are small holes at the end of each spoke (you gotta be at extreme angles to see them). this can be fixed by up-scaling the wheel mesh in the center as you suggest. is that the correct thing to do on 5_spoke_m?


sorry for lots of questions. but, this is some precise stuff so i want to make sure i understand completely.

[NaN]

outer ring of the wheel edge should vary more than the inner

From the code in mg_rim() the outer edge is:

Code:

innerRadius = rimDiameter_in * 0.0254f * 0.5f;
flangeOuterRadius = innerRadius * 1.08f;
float sectionWidth_m = sectionWidth_mm / 1000.0f;
flangeOutsideWidth = sectionWidth_m - (sectionWidth_m / 4.0f);

This means I always know the outer rim radius and width. The mesh will be placed relative to the outer rim edge. The inner rim edge has some constant values applied independent of the scale, thus unusable as reference point for mesh placement.

.blend file you posted has a negative x value for the hub. or, am i completely misunderstanding you?

Sorry, I've got the coordinates wrong. The wheel mesh is facing in negative direction(x-axis). Its vertices should be on the positive side then. Just import the 5_spoke_m.joe and scale the object (0.406, 0.406, 0.406). You only have to move the vertices (in edit mode) into positive direction(x-axis) and scale them by 1.05 or so. Then it should fit the rim pretty well. Undo the object scale and export the object.

[zimluura]

i think i get it.

i'll get all the shared ones updated, and then you can tell me if i got it correct or not