|
04-21-2010, 07:28 AM,
|
|
|
Quote:It would be cool if there was a formula to put tire vertical deformation in and get normal, friction force out.
Will read up on it, it should be possible to strip down the pacejka formula to get just the deformation at the current loads and feed it back in.
BTW, just tried the pacejka editor, nice one, that'll be really useful 
|
|
|
|
04-22-2010, 07:51 AM,
|
|
|
|
carsuspension.h #166
360 on second last turn of ruudskogen (flat out banked turn)
When suspension is at its travel limit displacement_ext is set to travel, then later saved as displacement..... which is somehow different to travel.... anyway, the value for distance gets smaller and smaller until the car floats away, wakes up and realises its not flying and craps its self.
|
|
|
|
04-22-2010, 09:58 AM,
|
|
|
NaN
Posting Freak
|
Posts: 2,024
Threads: 120
Joined: Jan 2010
|
|
The new displacement(wheel displacement) is clamped to [0, travel]. The difference between the old displacement end the new displacement is used to calculate the velocity. Then the old displacement is updated to the new displacement.
Quote:car floats away, wakes up and realizes its not flying and craps its self.
Can you make a screenshot?
|
|
|
|
04-22-2010, 10:13 AM,
|
|
|
Could try for a screenshot but would have to figure out replays The car goes into the bend, squats down properly then gives a small jump straight up.
I'm putting the values out in the terminal and when the suspension is at full travel the spring_force is slowly falling (by no fixed amount, generally around 50, occasionally rises) and the final value for displacement is falling.
I got around it by giving travel a high value in the .car file and letting it bottom out on the spring load (put bump stops in the spring curve on 350Z but stiffer springs would do the same).
Will try setting displacement to travel when maxed in carsuspension.h and see what happens.
|
|
|
|
04-22-2010, 10:33 AM,
|
|
|
NaN
Posting Freak
|
Posts: 2,024
Threads: 120
Joined: Jan 2010
|
|
Quote:when the suspension is at full travel the spring_force is slowly falling
Interesting. Will look into it.
Quote:setting displacement to travel when maxed
The displacement is set to travel if displacement > travel.
|
|
|
|
04-22-2010, 10:39 AM,
|
|
|
Quote:The displacement is set to travel if displacement > travel.
Thats what I thought, just tried with a huge travel figure and the displacement value keeps falling.
|
|
|
|
04-22-2010, 10:53 AM,
|
|
|
|
360 on ruudskogen, second last turn, the flat out banked one.
I committed my version but suspension values aren't changed, 350Z does it too on the same turn only not so badly. The only other cars I tried where lightweights and had no problem, not sure how much difference that makes though. Will try FF with high downforce.
|
|
|
|
04-22-2010, 10:57 AM,
|
|
|
|
Can the body slide along the ground frictionless?
|
|
|
|
04-22-2010, 11:01 AM,
|
|
|
NaN
Posting Freak
|
Posts: 2,024
Threads: 120
Joined: Jan 2010
|
|
You can change the friction value in cardynamics.cpp line 673. It's not recommended to set it to zero(bullet manual).
Code: info.m_friction = 0.75;
|
|
|
|
04-22-2010, 11:14 AM,
|
|
|
Looking at the terminal there are about 300 lines (stdout in carsuspension.h) from fairly balanced displacements to one or more wheels taking off.
Not some sort of self correction gravity in bullet is there?
|
|
|
|
04-22-2010, 11:36 AM,
|
|
|
|
Only had it happen about quarter of way around that turn. Can sometimes get the car back into shape and continue but not often.
Seems like the car is going from squat down to the top of suspension travel in an instant.
|
|
|
|
04-22-2010, 11:57 AM,
|
|
|
|
Just had it happen a lot more gently and it could be its only the suspensions reaction to the surface etc. Other times it felt like the car was being shot off the ground.
|
|
|
|
