How do physics affect the performance of cars in racing games?

[Aaron_Mason]

Hey,

I'm trying to make a racing game that will run completely within a browser. It will be similar to BATracer, except that rather than using carsets - a model employed by games such as TOCA, NASCAR and V8 Supercars - I will follow the model employed by Gran Turismo, where you buy a car, improve it, race it and get glory.

While I know that > realism = < fun, but I'd like for a true play-by-play of races, like BATracer already is. To do that, I'd like to understand how cars work so I can truly simulate the cars.

I've got the gearbox worked out, what I need to know is how various "variables" such as bore and stroke, the weight of the car and the weight of the engine can make an impact on how much engine rotation is sent to the gearbox.

Hope I haven't left anything out, and thanks in advance.

 

[chroot]

Cars are very complex machines. Some variables like total vehicle weight can be used in simple equations to determine changes in vehicle performance. Other variables, like bore and stroke, intake manifold shape, valve timing, etc. are virtually impossible to model with simple equations. Car manufacturers and tuners typically use a dynamometer to empirically measure changes in a car's performance.

- Warren

 

[Aaron_Mason]

Hi,

It doesn't have to be simple. As long as I can implement it in logic for a computer program.

I can understand how much speed a car will go at if 3000RPM is thrown into the gearbox in first gear... what I need to understand is how (and if) I can work out how quickly the car gets to 3000RPM.

Sorry if I sound like a complete n00b, there's a lot of things I don't quite understand and that I wish I could remember from high school physics.

 

[chroot]

It's not high-school physics. Vehicle dynamics is typically an upper-division course in university mechanical engineering curricula. There are textbooks on the topic.

- Warren

 

[Danger]

Aaron, there's a way that you could simplify the programming and still maintain a reasonable sense of reality. At least I think so, but I don't know anything about computers.
You could offer a list of upgrade options that simply give a percentage increase in horsepower and torque, much as you estimate the power of a real vehicle. Bore and stroke increases both simply increase the cubic inches of 'swept volume' in the engine. Bigger is stronger. Adding headers and low-restriction mufflers can add around 20% or so to the power that you already have. A windage tray can give you 5%, and most people don't even know what it is. Porting and polishing the heads adds a few more percent. As with a real vehicle, the more power you have in the first place, the larger that '5%' is. If you have 100 hp, then you gain 5. At 400 initial, you gain 20. Other things such as suspension tuning, different tires, etc. can increase your performance by certain percentages as well. Hypatia can probably help you with particulars, since she's a hard-core racer who builds her own cars. I've been away from the game too long.
Anyhow, all that your programme would have to do then is combine the hp/torque curve with the traction and aerodynamics in discrete chunks rather than trying to calculate every single variable.
That might not be what you're after, but it's a thought.

 

[Aaron_Mason]

Hey Danger,

That helps a lot for logic. I'd still like to know how weight affects this, though. At least offer a slightly higher sense of reality for the hard core racers.

 

[Danger]

Right. Weight, generally, is a detriment. The only exceptions to that are when a vehicle becomes so light that it can't maintain traction, or lacks the structural strength to stay in one piece. The solutions to that are to use aerodynamic features such as inverted wings and spoilers to create a downforce that supplements the gravitational attraction to the ground, and to use composite materials for strength.
In any circumstance, you have only to adjust your performance level upward as your power-to-weight ratio increases.

 

[C12_805]

Aaron,
What you've discribed sounds a lot like Forza Racing [for xbox]. First you need a game plan such how many variables you'd like to include. In Forza you alter just about everything which is interesting to a wrench-head like me. However the level of complexity may limit how many will actually play the game.

Weight is a key performance variable. In regard to the engine it depends on the material that's used, [aluminum head and block vs. cast iron]. I wouldn't put too much emphasis on bore & stroke - it depends on how it's built, [N/A vs. turbo vs. supercharged]. Power [torque] is determined [as mentioned] using a dyno. Typically there's a 15-20% loss through the driveline. Of course other variables are the tire size and gear ratio of the differential. I'd suggest checking out one of the many sites deadicated to automotive science & math.

 

[Danger]

Hi, C12. Maybe it's just because of my age, but I still say that there's no substitute for cubic inches. I'll take my 440 over anything from Mitsubishi any day. (And a 6-71 Roots over a turbo, but that's just personal preference.)
Good advice about the gearing (including tire size).
A couple of points about NOS that I neglected to mention. One is that it should never be used for more than 10 seconds or so at a time, and only at full throttle. Elsewise there's a good probability of blowing the motor. The second is that its effectiveness is inversely proportional to the efficiency of the engine. An off-the-showroom car could see a 50% increase in hp from it; a perfectly tuned race motor will barely notice it.

 

[C12_805]

Hi Dan,
I wasn't really endorsing any particular configuration, but rather referring to the physics...
One reason the muscle cars have such big cube engine is becasue they're heavy, most around 3700lbs. Ironically my muscle car has a better power to weight ratio, 10.5:1 vs. my sports car which is 12.8:1. The Z06 Vette is ~ 5.5:1 [woohoo - fun!], in contrast is the average car which is ~19:1.

 

[brewnog]

Hi Dan,
I wasn't really endorsing any particular configuration, but rather referring to the physics...
One reason the muscle cars have such big cube engine is becasue they're heavy, most around 3700lbs. Ironically my muscle car has a better power to weight ratio, 10.5:1 vs. my sports car which is 12.8:1. The Z06 Vette is ~ 5.5:1 [woohoo - fun!], in contrast is the average car which is ~19:1.

Eh?

 

[leright]

In case you're interested, here is the description of the vehicle dynamics course offered at my uni.

Vehicle dynamics 1:
Fundamentals of vehicle dynamics with focus on acceleration, braking, aerodynamics, axle loading, ride and steady state handling principles, steering and instability (e.g., roll over)

Advanced vehicle dynamics:
Must have a B.S.M.E. or graduate standing with approval of MAE program director. Fundamentals of vehicle dynamics with focus on acceleration, braking, ride and handling, steering, tire dynamics, and vehicle instability (e.g. rollover). Lecture 3 hrs.

 

[Danger]

Well, now... that sounds pretty cool. If you're anywhere near Calgary, I'd love to audit the lectures. (As a high-school non-graduate, I couldn't attend officially. :grumpy: )

 

[dephys]

well its fine to recognise the limitation of the software use
to understand the basics of the car stop playing the games and gotot books

 

[rcgldr]

Some compromises have to be made in order to be able to do the computations in real time on a typical PC. A guy named Todd Watson is currently working on his own crude simulator. You can search for his posts at http://www.lfsforum.net, where he uses the forum name jtw or jtw62074.

Here's a link to a typical thread, look for the post by jtw62074. He's met with some experts in the field to build up good knowledge of what it takes to make a car simulator on a PC type system.

http://www.lfsforum.net/showthread.php?t=7720