rec.autos.simulators

Etc. This has been a while ago, but I recently did some figuring out
of effective inertia's for geared components and finally put up a
small article on it on my page:

http://www.racesimcentral.net/

You can get there from the normal Racer page as well through
'Programmer info' and selecting the article from there, but ofcourse a
direct link is easier.

Feel free to shoot at it! But I hope to clarify where all the squared
terms come from (taking a different perspective) in Gillespie's
equations to calculate one big effective inertia for geared
components, and provide a method to check whether your formula is
correct for any gearing combination by using an energy approach.

I hope it's correct, but if you find errors, please let me know so I
can correct them (or remove the file altogether if it turns out to be
all a bunch of ***;-)).

Ruud van Gaal, GPL Rank +53.25
Pencil art    : http://www.racesimcentral.net/
Free car sim  : http://www.racesimcentral.net/

Hi Ruud :)
There are already some Sims with effective inertia build into physical model
code,
specifically at the centripedal forces.
Best
Miguel Romao

I hope so. :) But a lot of us here are implementing their own physics
code, either because of library code licensing issues, or just we all
want to do it our own way. ;-)

Cheers,

Ruud van Gaal, GPL Rank +53.25
Pencil art    : http://www.marketgraph.nl/gallery/
Free car sim  : http://www.marketgraph.nl/gallery/racer/

  Funny coincendence, I just downloaded that yesterday from your site to see if
you'd updated it yet :-)  Will check now.
Todd Wasson
---
Performance Simulations
Drag Racing and Top Speed Prediction
Software
http://PerformanceSimulations.Com

Great timing. :)
There's also a new little chat about track splines, which I put up a
few hours after the inertia document. The track splines are now
working, and I'm trying to get a renewed look at the Oval, with
splines and with a horizon now...

Ruud van Gaal, GPL Rank +53.25
Pencil art    : http://www.marketgraph.nl/gallery/
Free car sim  : http://www.marketgraph.nl/gallery/racer/

  I read about the splines too.  Very interesting.  I'm currently modelling
bumps using a similar method.  I'd like to learn more.
Todd Wasson
---
Performance Simulations
Drag Racing and Top Speed Prediction
Software
http://PerformanceSimulations.Com

I'll have to visualize the splines; I drove an oval with splines
yesterday and it's better, but in the corners it has some modulation
at the wheel, some jittering, which may be due to the difference in
the 2 triangles. Still, a lot better than WITHOUT splines. :)
And somehow the shadow intersections go bad; perhaps the two problems
are interrelated.

I'll just go for locking diffs and some other physics now, the splines
are ok for the moment, and the car needs working, since it spins out

Now where was that printout from Brain Beckman's 'Combined Pacjeka
Longitudinal and Lateral' again. Still haven't read it. Might give
some more clues as to why the car's too touchy.

Ruud van Gaal, GPL Rank +53.25
Pencil art    : http://www.marketgraph.nl/gallery/
Free car sim  : http://www.marketgraph.nl/gallery/racer/

Ruud,

I have managed to get combined slip to work (at least theoretically).  I got
a bit confused with Brian's stuff and ended up using G.Genta's
"intermediate" solution.

Have a look at my lovely graphs (under physics) -
http://sim.twistedtech.co.uk

All the best
Ash

| Now where was that printout from Brain Beckman's 'Combined Pacjeka
| Longitudinal and Lateral' again. Still haven't read it. Might give
| some more clues as to why the car's too touchy.
|
|
| Ruud van Gaal, GPL Rank +53.25
| Pencil art    : http://www.marketgraph.nl/gallery/
| Free car sim  : http://www.marketgraph.nl/gallery/racer/

Hi Ruud,

I just read the inertia stuff; great work, I wish I had the energy to
write stuff down in such an easy to understand manner. :)

There might not be much wrong with your physics model regarding the
ability of the car to spin. With my model, I've found that I need to
very carefully ballance the tyre specifications to create a well
handling car; I mentioned this in another thread a while ago, noting
that once you have a good physics engine, you then need to become an
engineer to sort a proper handling car from it. The first thing to do is
to readjust the front to rear tyre grip ratio (that is, if the tyre
specs are different rear to front in the first place). The next place to
look is softening the rear suspension, as this can create miracles (or
disasters ;) ) as well. I'm still struggling to create a car that will
point nice and straight at low speeds, but not create too much
understeer at high speeds without putting in too much front wing.

What I want to say is, maybe your tyre model is just fine, but the car
itself needs tuning.

-Gregor

On Wed, 20 Jun 2001 11:04:36 +0100, "Ashley McConnell"

The graphs look good; I must say it looks the same as my graphs here
(using the Ferrari Pacejka constants; I have a Pacejka Player thing by
now but it's so crappy that I won't release it yet hehe). With the
intermediate solution you mean taking the square root and making the
lateral force less important than the longitudinal force?
I use the extract below, but I have a very slippery car once it is
thrown sideways. Slides on much too long. I'll have to check if I
don't have a jittering longitudinal force which takes away too much of
the lateral force that can be generated.
I'll see if I can understand Brian's thoughts. He often takes a rather
different approach. :)

Ruud

--- rwheel.cpp extract
  // From G. Genta's 'Motor Vehicle Dynamics', page 83, eq. 2.24
  // (Fy/Fy0)^2+(Fx/Fx0)^2=1
  // Where Fy0 is the lateral force calculated using Pacejka, but
  // without taking into account any longitudinal force,
  // and Fx0 is the MAXIMUM longitudinal force. Fx is calculated
  // using Pacejka, and Fy is the unknown variable
  // Rewriting results in: Fy=Fy0*sqrt(1-(Fx/Fx0)^2)
  maxForce=pacejka.GetMaxLongForce();
  // Adjust maximum force for current friction coefficient
  // (based on slipvel)
  maxForce*=frictionCoeff;
  if(maxForce>D3_EPSILON)
  {
    dfloat Fy,Fy0,Fx,Fx0,ratioSquared;
    Fy0=pacejka.GetFy();
    Fx0=pacejka.GetMaxLongForce();
    Fx=pacejka.GetFx();
    ratioSquared=Fx/Fx0;
    ratioSquared*=ratioSquared;
    Fy=Fy0*sqrtf(1.0f-ratioSquared);
//qdbg("Flat FC'ed from %f to %f\n",Fy0,Fy);
    forceRoadTC.x=Fy;
  }

Ruud van Gaal, GPL Rank +53.25
Pencil art    : http://www.marketgraph.nl/gallery/
Free car sim  : http://www.marketgraph.nl/gallery/racer/

On Wed, 20 Jun 2001 11:04:36 +0100, "Ashley McConnell"

By the way, the rolling resistance seems to go up quite a bit, this
might be ok but for future enhancements, I read in John Dixon's book
that rolling resistance keeps quite low but at a certain turning point
you get waves in your tire that go out of hand and make for a suddenly
much higher rolling resistance. That's why some tires are not allowed
to be driven above x km/h. Like the fast/slow bump/rebound of dampers
a bit.
Ah well, future ideas, we all have virtual tires good up to 400 km/h
ofcourse. :)

Ruud van Gaal, GPL Rank +53.25
Pencil art    : http://www.marketgraph.nl/gallery/
Free car sim  : http://www.marketgraph.nl/gallery/racer/

Hi Gregor,

Thanks. :)

Yes, yikes! ;-)

I will try this. After driving at the oval yesterday I also realise I
will probably have to do that steering ratio thing; steer less when
speed is high.

How would you rate the influence of a (partially) locking differential
at the rear? With the open version that you had before, it seems the
wheel rpm differences can go up quite a bit quickly, so locking might
make things a lot more stable.

Yes, some wings, I haven't got those yet, but will attempt to get a
simple version in. Todd recently pointed my to FoilSim:
http://www.racesimcentral.net/
Which allows you to play around with wings.

Yes, definitely. :) But I'm a little hesitant on doing too much
setting up, as I'm still lacking the differential locking. I think
wings will only make me go faster without losing control, and rolling
resistance, well. And ofcourse the anti-roll bar, also seems to be
important. Enough with the splines, on to the physics again! :)

PS That BridgeBuilder game is really fabulously ***ive! :) Sad too
see it ate all the attention away from the free sim though, lol.

Cheers,

Ruud van Gaal, GPL Rank +53.25
Pencil art    : http://www.racesimcentral.net/
Free car sim  : http://www.racesimcentral.net/

Hi Ruud!

Just be sure that the higher steering ratio should apply only to very
low speeds (i.e. parking maneuvers), and then turns itself off at higher
speeds, as otherwise this can ruin the feel for the car.

Definitely. I am using a viscous type of a differential and indeed it
has a profound stability effect. In principle, the more locked the diff,
the more understeering the car will be in the normal operating regime,
but the power oversteer becomes much more snappy that way.

It's great to see how wings work, but one should be aware of the
limitations. Since I would imagine that the wings on a car rarely
operate close to stall angles, the whole applet could be quite accurate,
but the effects of finite size wings and/or endplates should also be
taken into account, as these can be dramatic. At this point, I
personally would go with empirical estimates.

Yes, wings shouldn't be a priority at this stage. If anything, they make
a car harder to control close to the limit. Antiroll bars, well, there's
something I still need to implement :) .

Now, a question for the more knowledgeable; I heard a few times that
highly aero dependant cars may not use anti roll bars, but a mechanism
quite opposite to it (softer in differential action than collective,
which is achieved by a special arrangement of springs). Is this really
so?

It sure is :).

Cheers,

-Gregor

If we are talking about the same thing then these are often called "third
spring" systems, fitted iirc on F1, CART and probably other cars.  The
third spring pushes equally on both wheels on one end of the car (through
some kind of balance-beam mechanism); another configuration is the Z-bar
(or pivoted leaf-spring "camber compensator") which does the same kind of
thing.

Since there is also an anti-roll bar fitted on many of these cars, there
are really 4 springs for 2 wheels!  The two original ride springs could be
removed and things would still work with just the "third spring" for ride
and the anti-roll bar for roll.

On race cars, a bump stop with "packers" (adjusting shims) is always fitted
on the third spring and this will control the _average_ minimum ride height
for the two wheels.  Note that F1 cars have a plank down the center of the
car that must be kept off the road, and the original purpose of the "third
spring" becomes clear -- you can set packers (a bump stop) that controls
the height of the plank, regardless of the roll angle of the body.

-- Doug

        Milliken Research Associates Inc.

Hello Doug!

Springy things, those cars, aren't they ;). Thanks for the reply, this
is most likely what I had in mind. Now, give me a couple of days to
study this to really get a good picture of what the system looks like
:).

-Gregor