rec.autos.simulators

Also, I havehandling forclicking on a face in the car model and
highlighting it.  When this method is called, the car suddenly starts to
whirl and spin and eventually a parameter gets out of bounds.  I can
only imagine that the method is taking too long and the time in between
physics updates is getting too big.

I've tried varying time between physics updates (the base is at
100,000/second (is this too little/too much/just right?), but this seems to
make ZERO difference (except that if I set updates to 1/millisecond
(.001 seconds, it crashes and burns pretty quickly).

Any ideas?  The jitters definitely seem to come from settling in the
Pacejka tire model.

Thanks!

Aaron

--
"You're the worst thing to happen to musicals since Andrew Lloyd Webber!"
  - Stewie from The Family Guy

SR = (Rotational_Velocity * Tire_radius - Tire_longitudinal_velocity) /
Abs(Tire_longitudinal_veloctity)

(Thanks again to Brian Beckman for simplifying a big mess of stuff there for at
least one of us in one little formula that can take a slow, hairy IF block to
do another way.)

What happens to the tire_longitudinal_velocity when the car almost stops?  What
happens to the slip ratio when the car is supposed to be sitting in one spot
(velocity =~ 0)?

This is only one of the dimensions of the problem (besides sideways/lateral
force), but it's the most influential one and might illuminate the cause of
other one..

Todd Wasson
---
Performance Simulations
Drag Racing and Top Speed Prediction
Software
http://PerformanceSimulations.Com

My little car sim screenshots:
http://performancesimulations.com/scnshot4.htm

Depends on the method, but by the sounds of it, I imagine that your imagining
the cause of this one exactly right.

This super-super small time ste[, yet still having a stability problem thing,
is related to the "approaching 0 veloctiy" bit in the other post.  You'll
figure that one out, I think.  That problem will blow the time step that
appears necessary out of proportion, even more so the closer you get to being
completely motionless.  This depends a lot on how you're doing the tires
(wheelhop mode?  5 rigid bodies or not? ) and what integration methods you're
using, but regardless, you should be able to reel this one back to 200/second -
300/second (100hz-200hz).  If things are running ok at even 1000hz, you'll be
on the right track, although that's still an extremely, extremely, small time
setp.

You're observations are right on this one too, I bet.  You've got a good eye
and feel for what's going on in your model :-D

From before, take a look at the slip ratio calculation and see what happens
when the tire's velocity gets really close to 0.  A similar thing happens in
the slip angle calculation, but it's generally not visibily noticable....  Let
us know what you discover.

Todd Wasson
---
Performance Simulations
Drag Racing and Top Speed Prediction
Software
http://PerformanceSimulations.Com

My little car sim screenshots:
http://performancesimulations.com/scnshot4.htm

Except for Pacejka, also check for stiff suspension springs. Can also
do a lot of harm.

Ruud van Gaal
Free car sim: http://www.racer.nl/
Pencil art  : http://www.marketgraph.nl/gallery/

It was a problem with my suspension.  I had my roll center defined, but
had dumply left it relative to absolute car position...I hadn't taken
the position of the suspension in to account!

Needless to say, this led to all the forces going through the suspension
being off!

Thanks again to Ruud for making his source public as I doubt I would
have seem this dumb mistake otherwise.

My car now sits quite happily still with only one oddity: Immediately
upon starting, the car rotates about 5 degrees clockwise.  I'm sure it
will be something similaryl stupid, though.  Hopefully it won't take me
as long to figure it out, though!

Aaron