rec.autos.simulators

Hi !

If we simplify a force damper is

F = k * v

v = damper piston velocity
k = damper oil viscosity coefficient  ( k<0 )

Do somebody have an idea of k value for a stock car damper ????

In reality k = F( v ), because on the piston there are some valves that
opens when v incerase

Do somebody have this function ?

Can't find those infos at Ohlins web site or anywhere :o(

regards,

--
Seb -
Game Developer
http://www.racesimcentral.net/
GPLRank -41.68
http://www.racesimcentral.net/

Hi Sebastien,

Starting with the formula you presented (F = k v ) is a very good
approximation, and, depending on how far you want to go with it, should
perhaps be just sufficient (I have a feeling GPL and N4 don't use
anything else).

Generally, however, you would have

F=F(v),

and a good approximation for many a damper is to have a function

F = k1 v                     for |v|< v0

and

F = k2 * (v - v0) + k1 * v0  for  v > v0

and

F = k2 * (v + v0) - k1 * v0  for  v <-v0

where |k2| < |k1|, since at a certain speed, as you mention, additional
valves may open (in GP3, for example, the coefficients k1 and k2 are low
and high speed dampers, respectviely). The function F is therefore a
line with kinks into shallower lines at |v|=v0

To get an estimate for the dampening required for any kind of a car,
assume that each part carries about an equal weight (the total mass of
the car being M), and four springs, all with the coefficient K. If you
look at the differential equation that governs the total up/down car
movement,

M x'' + 4 k1 x' + 4 K = 0 ,

the optimal damping will be achieved when

k = sqrt( M * K ) .

Start with this and then tweak to reach a good setup.

-Gregor

Well, to not answer your question, I would have thought it would be
something like:

if (v<v1) then F=k1*v else F=k2*v

or words to that effect, ie basic low/high speed circuit

BUT THEN

I'm just filling in time and haven't ever tried to model this kind of thing,
and have just introduced new variables k2 and v1 which I also don't know the
value of!

Ben...who isn't sure why he responded

mb.si>:

I've have an SIA document in front of my eyes
( SIA = Automotive Ingenior Society )

and the damper formula is quite complicated, it's involved

Su = Section of damper piston minus damper shaft ( PI * r^2, r = 2cm )
rho = oil mass ( 876 kg/m3 SHELL)
mu = oil viscosity ( 0.7 )
d1 = valve diameter  ( 8mm )
V  = damper piston speed ( variable )
sig = piston leak section ( PI * r^2, r = 4mm )
j1 = valve opening ( 0 at low speed, and j1 is max when V > V0 )

and the formula is .. glups...

F = rho / ( 2 * mu^2 ) * ( Su^3 ) / ( sig + PI * d1 * j1 )^2 * V^2

Well, except of j1 = F( V ), all other componant are not variable in time.

The important things is that it's not k.V but k.V^2 !!!

I'll try this formula and yours with V^2 instead of V, because k.V didn't
give me satisfaction.

Seb

--
Seb
Game Developer
http://www.eden-studios.fr
GPLRank -41.68
http://magicfr.multimania.com

Hi Sebastien,

what you seem to have is a paper that tries to get the damping from
basic principles of physics. This may not be exactly what you're after
since going from basic physics in a complex fluid dynamics problem will
necessarily neglect certain important effects while exposing others. It
seems to me that it's the high speed behaviour which the paper you have
deals the most with, while low speed behaviour is what you would need
for the basic handling.

I am surprised, though, that the linear damping model didn't work for
you, as it should be very, very close to what you want. Can you tell
more about what is wrong with it?

-Gregor

mb.si>:

Well, the high speed behaviour is a big problem for me because there is lot
of high speed behaviour on jumps . moreove the tracks is very bumpy... you
know, like a rally track ;o)

It's working well, but i feel that it could be better ;o)

In fact, i try to have a different model than k.V that is very dependant
with integration problems ! if k is to high then dampers generate force
greater than the gravity instead of just resisting. So the car is hoping a
lot :o(
If i decrase k there is no more hoping but the car has too much rolling, and
diving because of too smooth dampers :o(

I'll hope that a V^2 model + high/low speed rate can resolve this.

BTW, i can't integrate my model faster than 100hz .

I let you know if it's works better with this model.

--
Seb -
Game Developer
http://www.eden-studios.fr
GPLRank -41.68
http://magicfr.multimania.com

Hi again, Sebastien,

The pure V^2 model for small velocities will cause the car to not dampen
the small oscillations, so the car will always keep on rolling a bit,
just as is the case with too smooth dampers.

Another thing you should probably take into account is to have a
different damper rate for bump and rebound. Therefore,

F= k_a * v   for v>0

and

F= k_b * v   for v<0,

where k_a and k_b are different, and the coefficient that corresponds to
the bump (compression) is smaller than the one for the rebound
(extension). I believe this is what would work best for you, as this is
what real cars use as well.

-Gregor

hehehe, it was obvious for me, it's already implemented in my model :o)
I also have other "expert" parameters in springs, as spring preload and
stroke length .

I also have front and rear anti roll bars.

Yes,this is what lot of car companies recommend, except Citro?n who
recommend equal coef for compression and extension, but they have hydro-
pneumatic dampers that are very "special".

Anyway, i'll try V^2 model and it's very, very bad ;o) ... maybe the SIA
formula din't calcul a Force but another thing ... i didn't understand
everything even if i made physics studies ;o)

Your "double" linear formula pleased me ;o)

--
Seb - Le nain, des sensations biens.
Game Developer
http://www.eden-studios.fr
GPLRank -41.68
http://magicfr.multimania.com

Hi Sebastien,

I'm glad to hear this :). I'd try first with setting V0 to some
reasonably small value, and then k2=0, which effectively limits the
maximum force the dampeners can exert. That should take care of the
jumping cars.

-Gregor

mb.si>:

hi,gregor,

Here is the setting i used , and it's works fine:

front

k1c = 300000
k1r = 500000

k2 = k1/3;

v0 = 5cm/s

kspring = 75kN/m

rear

k1c = 300000
k1r = 400000

k2 = k1/3;

v0 = 5cm/s

kspring = 45kN/m

The car is a mitsubishi lancer 1230 kg, weight distribution is 60% front.

the kspring values are taken from Ohlins web site, Pro suspensions for the
Mitsubishi Lancer.

Tkanks for your helps.

ps : i haven't modeled tire/ground viscosity yet ;o)

--
Seb
Game Developer
http://www.eden-studios.fr
GPLRank -41.68
http://magicfr.multimania.com

Hi Sebastien,

that v0 setting seems to be a bit low, but what do I know ;). I'll check
the Millikens' RCVD later to see if I can find some data on a good
value.

-Gregor

mb.si>:

That will be great !

well, i'm going to try higher value ;o)

regards,
--
Seb
Game Developer
http://www.eden-studios.fr
GPLRank -41.68
http://magicfr.multimania.com

IIRC, RCVD has some shock force vs shaft speed graphs for a couple of racing
dampers at various knob settings, and you could just do pretty normal
curve-fitting to reproduce those results--while fluid dynamics are
inherently non-linear, many of the shocks actually strive for a "boring"
linear function by artful washers and holes and such.  I bet you could go to
Koni's web site and download graphs for their aftermarket dampers--I'm
assuming by "stock car" you mean "production car", not "tube-frame race car
with vaguely sedan-like sheet metal?"
You could probably take the softest Koni setting and scale it down by some
factor to approximate street car values...

Another thing to keep in mind is that street cars are typically set up to be
critically damped, while race cars are usually overdamped, so you could just
do some basic quarter-car analysis on your model to find those settings.
-Dave P.

Hi Sebastien,

from what I could see in that book (no strong data, just some examples),
the V0 parameter should be of the order of magnitude 0.1-1 m/s, probably
leaning to the lower side for the off-road application.

A bit of racing sim philosophy; what's amusing is that, once you create
a good physics model, you are then given the role of a car engineer in
that you need to find the best suited springs, dampers, tyres etc. so
that your model is at least approximately driveable on the limit. For
example, that Champ car in my sim is still not quite driveable at speeds
when aerodynamics steps in, and I'm really having a tough time setting
it up! I guess you're in the same game at this point ;).

-Gregor

mb.si>:

Hi Gregor,

Waow, i'm far from this ( 0.075 m/s ) ... Well, the problem in rally is that
we need to have good setting for off/road ( dirt/snow ) and for tarmac (
corsica/spain )... So we need to setting up the car for a wide range of
application !!

Yes, that's true, it's quite hard to find correct setting for tire/surface
grip, springs , dampers, etc...

I'm lucky that aerodynamics is not very important in rally because there is
no high speed and no big spoiler ... anyway we have simple aerodynamics
force too .

That champ car you working on is for a personal project or for a game
company?

--
Seb
Game Developer
http://www.eden-studios.fr
GPLRank -41.68
http://magicfr.multimania.com