Hi everyone, The tire model is obviously a hot topic in BeamNG, and I realize that the devs are hard at work improving it. However, as I am an alpha tester, I feel it's my duty to report my findings. It would seem that BeamNG doesn't simulate a difference between the coefficient of static friction and the coefficient of kinetic friction. In reality a car should take almost 30% longer to come to a stop when locking the wheels or spinning out, relative to controlled braking at the limits of grip. Let's use the Civetta Bolide as an example. Test 1: Civetta Bolide braking test (100km/h - 0km/h), brakes applied, tires at limit of static grip. (imported from here) Test 2: Civetta Bolide braking test (100km/h - 0km/h), spin-out with brakes and handbrake applied. (imported from here) In Test 1, the car comes to a stop in 2.64 seconds, a similar time to modern high-performance vehicles such as the Ford GT and Chevrolet Corvette C6. However, in Test 2, the car comes to a sideways stop in a ridiculous 2.25 seconds, which would make it the world's fastest braking car. This is the inverse of what I expected: the 'spin out' test should yield a result of approximately 3.7 seconds, given that the kinetic coefficient of friction for rubber is roughly 70% of the static CoF. Cars should take that much longer to stop when traction is lost. My findings seem to indicate that the tire model needs to better differentiate static vs kinetic friction, as it is not currently simulating reality.
Well, we have made substantial changes to the friction and physics code, as well as the vehicles, so maybe you can re-test with the race update and see if something is still off
This is actually really awesome for you to bring this up and conduct a comparison. Not complaining, but yeah, the friction differences between skidding and slowing down at grip limit always seemed a little strange to me, and like gabe said, definitely compare these results to the race update results. Im sure (I have immense faith in the devs for improving the game) the race update will be much better. awesome thread!
Thank you Gabe for your swift reply! It's great to have developers in such close contact with the community.
The tires were always way off, im glad these threads aren't getting shut down with "Shut up, the game is realistic, have you ever driven a vehicle" like they used to be. Looking forward to the changes in the next update
There is a difference between static and kinetic friction. Have a look at lua/groundmodel.lua to see and change the values. The problem here is the low tire node weight. It seems the static friction doesn't work properly with very light nodes, so the tire friction coef must set higher than normal to reach realistic cornering or braking grip. When a node slides fast, the friction model works more like it should. But then the friction coef is too high, so a car stops faster with locked brakes, accelerates too quickly with spinning wheels etc. Heavier tire nodes would fix it, but that causes bigger problems for cars that should have light wheels. Vehicles with big, heavy wheels don't need nearly as high friction coef for realistic grip.
Nice to see someone actually doing scientific testing for once instead of just signing on and saying Time to add this to my list of things to hope for from the race update. Every day it seems to get better and better.
Now that the pre-race update is out, would anyone care to do a similar test? Initial reports suggest this has been improved vastly, but since I'm away from my PC I won't be able to try it for another week.
ditto ^^...and I cant believe that this thread only got a small handful of replies, also since Gabe himself commented....so yea, now that the tire model has been improved, lets soo the difference. conduct the same test, with the same car, on the same map, and see whats going on. very exciting time we live in!!
the groundmodel.lua now has values for both static and kinetic friction. i would say this issue is fixed.