Irrlicht Engine

Why is it that every time I run a program which uses IrrNewt, the physics are inconsistent? For example, if I apply a force to an object, it will push it so far on the first run. On the next run, it will either push it further or not as far. The next run, the results will be different again. What could cause this?

In anticipation of the inevitable "No one can help me?" BUMP, IrrNewt appears to be abandonware, and Newton is bordering on it. At the time of writing, the download links for both and the Networn forum are all 404.

So rather than spend more time trying to figure out a problem in an unsupported black box, why not just switch to (e.g.) Bullet? At a minimum, you'll have the source to fiddle with.

wolfman wrote:Why is it that every time I run a program which uses IrrNewt, the physics are inconsistent? For example, if I apply a force to an object, it will push it so far on the first run. On the next run, it will either push it further or not as far. The next run, the results will be different again. What could cause this?

I think this is because you have to run the physics at a fix time step.

cassini wrote:

wolfman wrote:Why is it that every time I run a program which uses IrrNewt, the physics are inconsistent? For example, if I apply a force to an object, it will push it so far on the first run. On the next run, it will either push it further or not as far. The next run, the results will be different again. What could cause this?

I think this is because you have to run the physics at a fix time step.

Exactly. Also, use the exact solver (not linear or adaptive).

(and no, Newton is not a "black box" or dead, Julio is switching hosting providers because powweb fuxxed him over)

Well, I set the solver to "precise" and set a fixed time step, but the results still vary from run to run. I must be missing something somewhere or doing something wrong.

I am using irrnewt and I am shooting a tank bullet that hit the ground after 1 kilometer at high speed. An d i always hit at the same point.
What is exactly your problem?

By the way, I am even not using exact solver, and anyway work just fine.

Are you using exact or adaptive friction?

Now I set the friction model to PRECISE, but earlier I was using default (I don't know which was). But it work almost just fine (if I remember correctly).
Why? Are you having some problem like this?

I am having some problem with friction between tires and terrain (obviusly is my problem ), but about the rest is kinda precise.
About the solver model I am using the default too because I need to mantain a good fps and my approach to simulte tank physicly actually is not very optimized .
By the way, you are the one which is making a tank too, what approach about track do you take , covervor velt? wheels? other?
My approach is kinda realistic but very poor about physic performance.

wich version of IrrNewt are u using?

in the latest version the problem was solved by using a timer to fix the timestep