Can you drive an F1 car on the ceiling?

LesiU said:
Thanks Johannes!
OK, next issue - it has just too few polys and car goes very unstable. Can you increase polycount? It's just a test track so I don't mind if the tunnel has 500k polys ;-)

But I must say, it looks promising :)
Click to expand...

I can do that, but I don't think that helps any :D
 
LesiU said:
The loop also has wayyyy too small radius. Make it at least 4-5x bigger (and with much more polys, because it is not smooth).
Click to expand...

I don't think those are the issues. Physics starts to fail when car reaches 90 degree angle. Making tubes and tunnels larger doesn't change the fact that at 90 degree nothing works well. But I can still do that so that you see what I mean :)
 
At 90 degrees you would get 1G lateral even while going straight, but F1 car should be able to still do easily 2G and more cornering, so that should not cause the issue, it could of course be that aero force down is in world coordinates instead of car's local coordinates. Also it can be that there are though range for variables and upside down is something that has not been thought when deciding ranges for those variables, such happen all the time and can cause issues in many other things than just driving upside down.
 
I set up my Murc GT1 to generate downforce of it's own weight (1180kg with driver) at 140kph and 130/160mm ride height :)
At 250kph, when I start to go upwards, it already has 3times more downforce than its weight and that is clearly visible. That's why such high ride height - I had to use relatively soft effective spring stiffness (about 150-200N/mm) due to uneven surface even before I hit 90deg angle... but we will see how it goes with much smoother road.
 
Wow this takes ages to export this large meshes...

My best results this far has come with Dodge Monaco (in tunnel), Clio (in bowl of death) and in rTrainer (loop). All of these attempts failed, but other cars did even more worse.

In real world, the loop is actually rather small in diameter:

Team-Hot-Wheels-Land-Double-Loop-Stunt-1-625x436.jpg
 
Johannes Rojola said:
Wow this takes ages to export this large meshes...

My best results this far has come with Dodge Monaco (in tunnel), Clio (in bowl of death) and in rTrainer (loop). All of these attempts failed, but other cars did even more worse.

In real world, the loop is actually rather small in diameter:
Click to expand...

Monaco is not generating downforce, it is creating lift, maybe that is why it is working better?
 
Johannes Rojola said:
In real world, the loop is actually rather small in diameter:
Click to expand...

But it's smooth :)
Also, at 200-250kph angle of ramp is too high and is like I almost crash into it, like a wall, that's how much susp compression I get at the bottom.


I'm at the point where I feel comfortable going straight near 90* angle. What's the most important, is that I can feel that my car is sucked towards surface, NOT down direction. That means, aero probably work like should.
 
LesiU said:
But it's smooth :)
Also, at 200-250kph angle of ramp is too high and is like I almost crash into it, like a wall, that's how much susp compression I get at the bottom.


I'm at the point where I feel comfortable going straight near 90* angle. What's the most important, is that I can feel that my car is sucked towards surface, NOT down direction. That means, aero probably work like should.
Click to expand...

When working with 3d mesh, you really need to decide whether you want something that is small but very smooth, or something that is large and not that smooth :D
 
HAH that was a blast, managed to get the 370z about halfway up Monaco at a steady drive, slight twitch up and I lost it...great fun though! Love these experimental fiddling with the game :)
 
Going at speed limiter (80kph) I was able to go way past 90 deg. (something like 130-140 deg.). at the loop and lost it ONLY because of lack of downforce. No unusual behavior, nothing unexpected. So I'm pretty sure we will eventually get there :)

Johannes Rojola said:
When working with 3d mesh, you really need to decide whether you want something that is small but very smooth, or something that is large and not that smooth :D
Click to expand...
Third option - get both. I'm fine with 2 million polycount :P
 
OK, a couple of observations.

1). Car does not go past 90 degrees in the tunnel.
2). Car can sit on the wall not moving at near 90 degrees.

This tells me that the physics isn't close for doing this sort of thing. FISI car should have started to slide down the wall the slower I got, but I maintained my distance up the wall until full stop.

EDIT: One other thing, when I got to 90 degrees, no matter how much steering input I gave the car towards the top, it wouldn't change the direction or force a spin in the tunnel at 90 degrees.
 
Gearjammer said:
This tells me that the physics isn't close for doing this sort of thing. FISI car should have started to slide down the wall the slower I got, but I maintained my distance up the wall until full stop.

EDIT: One other thing, when I got to 90 degrees, no matter how much steering input I gave the car towards the top, it wouldn't change the direction or force a spin in the tunnel at 90 degrees.
Click to expand...

I had pretty much the same experiences. I am currently making a new version of the track which has much more dense road mesh so it should increase physics accuracy. But the bad thing is that this is so detailed that it is completely impractical for any gaming usage :(
 
@Gearjammer
Yeah... it won't happen, unfortunatelly :(
I set over 40kN of downforce at 80kph and tried to do the loop and it's the same. Just past 90 deg the car simply loose grip.

Johannes, you were right. Those "bumps" are just an effect of strange behavior when we go past 90 deg angle.

EDIT:
It still is just a test track, so we should not be worry about performance issues with 20 cars :)
 
lest we not forget that the physics coding might not actually allow for anything over 90degrees anyway, *maybe* doing that costs more computing power so its rendered pointless to do for a car racing sim, suppose ISI can only answer that of course.
I am no coder though but I cant see ISI coding in physics for upside down travel as it'd be a waste of time, on the other hand the code for normal driving might include 90degrees+ anyway so all I said above is irrelevant... :)
 
To make things interesting, LesiU was partly right. Loop with increased polycount gave much better results! I could have went through it with rTrainer but that car is not fast enough. I am still rendering new version of the tunnel, maybe it works better as well. I believe so.
 
@MarcG
In theory, there is no need for additional coding required for that. If you have defined gravity correctly and other stuff like aero working in the right direction, there shouldn't be anything special to add. At least that's how I see this, but of course I might be wrong.

@Johannes
That's interesting indeed! Can't wait to test the new version of the track :)
 
You must log in or register to reply here.
Back
Top