> What I would like to know is what Doug meant by "his only flaw in
> the model is he doesn't do dynamic camber and roll center change with
> roll".Could someone(hopefully Doug)help me out as to what this means.
Actually, I should have left the word "roll" (and "flaw" for that matter) off the
end of that sentance it's just plain dynamic, and is constantly changing in bump,
rebound and roll. Camber change is controlled by whats called an "instant center"
or instant swing-arm lenght, this arm swings in an arc about its center and has a
certain amount of camber gain based on that arc. GPL uses this arc as formed at
ride height and each car in the sim is different. This is formed by the
suspension point placements on the tub and the lengths of the upper and lower
a-arms (see Caroll Smith's books). This arm, on a well designed suspension, is
much longer than the actual a-arm lengths. The reason they are called "instant"
is because they only occur at the instant you are looking at them. As soon as
things start to move, that swing-arm lenght starts to change, and so does the
amount of camber gain. Further complicating it is that the front and rear
swing-arm lenghts are usually different, because the rear has different
priorities than the front.
The same thing goes for roll center heights. They constantly evolve up and down,
and left and right as the car rolls squats and lifts. This affects the roll-axis
(a line drawn between the front and back roll centers) and the torque-arms
between that axis and the CG heights (there's more than one!) that the sway bar
and springs have to resist when cornering. It's an extremely complicated
interaction between all these factors that makes it so calculation intensive. The
formulas are all there, and readily available and Keammer is well aware of them -
it's just not feesable CPU-wise ATM, to do it all. Keammer is a genius - and what
he's done here with GPL - with the tools and restrictions he's been given - is
incredibly impressive.
Once these things become possible, then a true challenge to car setup and
individual car behaviour will come about.
Also, once this is done, the possiblilty of designing our own cars from scratch
can be very realistic - this I_want_to_happen :)........................
In the end, a roll bar is for driver feel and under/oversteer adjustment. The
risidual effect is that it controlls the amount or roll and resultant camber
"loss" in roll. Your job is to match the static camber settings to the amount of
roll you imposed through the srings/rollbar settings by using tire temps. BTW, I
use the word camber "loss" because most race cars suspension systems will go
*positive* camber as the car rolls, so you need to match the final loaded state
with an intial negative setting to make up for that. Complicating that issue is
the track geometry, ie banking which requires more negative camber, the driver -
the faster you go the more G's you pull the more the car rolls the more negative
camber needed or more swaybar..............and on and on........... Confused? You
should be <g>.
No, they *will* effect camber change in GPL because there is still a camber curve
as the wheel travels up and down. As you stiffen the roll bar the loaded wheel
compresses less. However, the roll of the chassis, and the resultant roll of the
suspension points, dosen't change that curve at all - this is the dynamic part.
No, see above :)
The current model is still *extremely* realistic feeling, you just don't have the
possible bad things that the car can throw at you that would effect your setup
decision and car choice.
Once we get all these things and complicated realistic track "surfaces" , we will
have the next generation :)
--Doug
PS: I corrected the spelling of my last name in the header..........don't worry,
it happens all the time I'm used to it :)
Hi Dan,
> Thanks doug for the reply but I still dont know what you meant by no dynamic
> camber change in the GPL physics model.Would it be possible for you to
> further explain what you mean by no dynamic camber change?
What I mean is dynamic camber *curve*change. That is a camber curve that is
constantly changing as the car squats, lifts and rolls. Right now there's *one*
curve that GPL uses for for this and its formed by the suspension geometry at the
*designed* ride height. As the wheels travel, they follow this curve and gain or
loose camber dependent on this curve. In real life, this curve changes shape as soon
as the car starts to change position from the designed level ride height...........
Correct,.:)
No, your Correct......and there's the rub, you see. Where to compromise. If the
track has a fair amount of left and right hand corners then you must set the camber
equal on both sides. Each side of the car takes it's fair share of loaded turns so
you must have the appropriate amount of negative camber on both sides of the car.
If, OTOH, its almost entirely in one direction, Like Monza, then you can start to
think about asymmetrical camber to optimize all four contact patches in 5 out of the
6 corners....a worth while compromise there. But this has the side effect of
squirrely acceleration and braking.
When you stop feeling lost, you're about half way there <g>.
--Doug
[fascinating and slightly confusing (for me at least) stuff]
I didn't ask the original question, but now *I* am a bit confused...
Please tell me if I got this right or not:
1. The camber we dial in in the setup corresponds to a car that is
sitting on a plain surface and not moving.
2. Assuming we have chosen 0 camber: once the car gets "pressed"
evenly to the road, the wheels will gain negative camber due to the
geometry of the suspension.
3. Under the same assumption, the wheels of a car that gets airborne
will have positive camber during the flight.
GPL models all this, but
4. What is *not* modelled is that, say in a high speed turn, the inner
and the outer wheels will have different camber due to the leaning of
the chassis itself. Only the camber change coming from the moving
suspension is modelled, as if the mounting points of the suspension on
the chassis didn't roll with the chassis.
Did I get this right?
And so is Mr. Kaemmer, I should think. But as long as you keep calling
him a genius, I'm sure he won't mind. <g> SCNR
--
Wolfgang Preiss \ E-mail copies of replies to this posting are welcome.
Sorry for butting in Doug and correct me if I'm wrong, but from what I
gather, GPL models camber change due to roll. It uses a fixed camber curve
based on the static suspension geometry. However in the real world this
camber curve is constantly changing, ie dynamic, and this change in the
camber curve is not modelled.
The inside tyres, when cornering hard, do *** all for grip (although I
guess with the low grip levels of 67, weight transfer in corners would be
less). The optimisation of the outside tyre's contact patch is more
important than the inside one. Try putting your inside tyre onto the grass
while cornering hard, it affects grip but not much. Now do the same with
the outside tyre ;-)
Ben
