some a6 physics questions

quote:

---

Originally posted by ventilator:
is PH_CYLINDER already implemented as collision hull? i saw it in the manual under physics constants but it isn't listed under collision hulls and it didn't seem to work when i tried to use it.

---

PH_CYLINDER can be used for testing purposes- it is not officially implemented. The part that is missing (and I would need to double-check that) is Cylinder to Mesh collisions. So if you create a box or sphere entity and set another one to cylinder, they should collide. Because the level is a mesh though, cylinders will fall straight through it.

quote:

---

i would like to set up a car with realistic physics. i know i have to attach the wheels to the chassis per constraints (probably with the PH_WHEEL constraint?).

---

quote:

---

Originally posted by Fredmaster:
Would it be possible to use the physics engine for a car or other complex/multiple-part object using the one object physics in commercial edition?

---

quote:

---

Because the level is a mesh though, cylinders will fall straight through it.

---

that's exactly what i experienced. so i have to use PH_SPHERE instead of it for the wheels of my car. should work too...

if i want the car to go around a curve, do i just need to turn the front wheels? or do i need to apply torque/motor (front wheel drive) so that they "pull" the car around the corner?

...
the physics engine is really cool! in my experiments i had over 200 active physics entities and didn't have any jerks or any noticeable framerate drop. i didn't expect that!

quote:

---

P.S. @Ventilator: How did you get your update to A6 already?

---

ordered yesterday, got it today!

code:

---

action sphere
{
	my.shadow=on;
	
	phent_settype(my, PH_RIGID, PH_SPHERE);
	phent_setmass(my, 5, PH_SPHERE);
	phent_setfriction(my, 70);
	phent_setelasticity(my, 50, 10);
	phent_setdamping(my, 20, 20 );
	
	temp.x=cos(camera.pan)*cos(camera.tilt)*300000;
	temp.y=sin(camera.pan)*cos(camera.tilt)*300000;
	temp.z=sin(camera.tilt)*300000;
	phent_addforcelocal(my,temp,nullvector);
}
 
string mdl_sphere=<sphere_0.mdl>;
function emit_sphere()
{
	temp.x=camera.x+cos(camera.pan)*cos(camera.tilt)*100;
	temp.y=camera.y+sin(camera.pan)*cos(camera.tilt)*100;
	temp.z=camera.z+sin(camera.tilt)*100;
	ent_create(mdl_sphere,temp,sphere);	
}

on_mouse_left=emit_sphere;

---

it emits spheres from the camera position!

code:

---

action sphere
{
  my.shadow=on;
  phent_settype(my, PH_RIGID, PH_SPHERE);
  phent_setmass(my, 5, PH_SPHERE);
  phent_setfriction(my, 70);
  phent_setelasticity(my, 50, 10);
  phent_setdamping(my, 20, 20 );
  temp.x=cos(camera.pan)*cos(camera.tilt)*300000;
  temp.y=sin(camera.pan)*cos(camera.tilt)*300000;
  temp.z=sin(camera.tilt)*300000;
  phent_addforcelocal(my,temp,nullvector);

  sleep(5);
  ent_remove(my);
}

---

code:

---

action sphere
{
  my.shadow=on;
  phent_settype(my, PH_RIGID, PH_SPHERE);
  phent_setmass(my, 5, PH_SPHERE);
  phent_setfriction(my, 70);
  phent_setelasticity(my, 50, 10);
  phent_setdamping(my, 20, 20 );
  temp.x=cos(camera.pan)*cos(camera.tilt)*300000;
  temp.y=sin(camera.pan)*cos(camera.tilt)*300000;
  temp.z=sin(camera.tilt)*300000;
  phent_addforcelocal(my,temp,nullvector);


  on_mouse_left=null;
  sleep(5);
  on_mouse_left=emit_sphere; 
  ent_remove(my);
}

---

...could work!

quote:

---

An object's hulltype determines its resistance to torque and thus the way it rotates. You can use one of the predefined constants to use an approximation of the object's shape or use PH_POLY to deduce the actual shape.

---

code:

---

function emit_sphere();

action sphere
{
	my.shadow=on;
	
	phent_settype(my, PH_RIGID, PH_SPHERE);
	phent_setmass(my, 5, PH_SPHERE);
	phent_setfriction(my, 70);
	phent_setelasticity(my, 50, 10);
	phent_setdamping(my, 20, 20 );
	
	temp.x=cos(camera.pan)*cos(camera.tilt)*300000;
	temp.y=sin(camera.pan)*cos(camera.tilt)*300000;
	temp.z=sin(camera.tilt)*300000;
	phent_addforcelocal(my,temp,nullvector);

        on_mouse_left=null;
        sleep(5);
        on_mouse_left=emit_sphere;
        ent_remove(my);
}
 
string mdl_sphere=<sphere_0.mdl>;
function emit_sphere()
{
	temp.x=camera.x+cos(camera.pan)*cos(camera.tilt)*100;
	temp.y=camera.y+sin(camera.pan)*cos(camera.tilt)*100;
	temp.z=camera.z+sin(camera.tilt)*100;
	ent_create(mdl_sphere,temp,sphere);	
}

on_mouse_left=emit_sphere;

---

i am also not really good in programming. i just do trial and error all the time!

code:

---

var FR_wheel_ID;
var FL_wheel_ID;
var RL_wheel_ID;
var RR_wheel_ID;

action FR_tyre
{
	my.shadow=on;
	
	phent_settype(my, PH_RIGID, PH_SPHERE);
	phent_setgroup(my,2);
	phent_setmass(my, 5, PH_SPHERE);
	phent_setfriction(my, 70);
	phent_setelasticity(my, 50, 10);
	phent_setdamping(my, 20, 20 );
	
	FR_wheel_ID=phcon_add(PH_WHEEL,my,p_player);
	phcon_setparams1(FR_wheel_ID, my.x, vector(0,0,1),vector(0,1,0));
	phcon_setparams2(FR_wheel_ID, vector(35,-35,0),vector(-360,360,0),vector(0.9,0.001,0));
}

action FL_tyre
{
	my.shadow=on;
	my.pan=180;

	phent_settype(my, PH_RIGID, PH_SPHERE);
	phent_setgroup(my,2);
	phent_setmass(my, 5, PH_SPHERE);
	phent_setfriction(my, 70);
	phent_setelasticity(my, 50, 10);
	phent_setdamping(my, 20, 20 );
	
	FL_wheel_ID=phcon_add(PH_WHEEL,my,p_player);
	phcon_setparams1(FL_wheel_ID, my.x, vector(0,0,1),vector(0,1,0));
	phcon_setparams2(FL_wheel_ID, vector(35,-35,0),vector(-360,360,0),vector(0.9,0.001,0));
}

action RL_tyre
{
	my.shadow=on;
	my.pan=180;
	
	phent_settype(my, PH_RIGID, PH_SPHERE);
	phent_setgroup(my,2);
	phent_setmass(my, 5, PH_SPHERE);
	phent_setfriction(my, 70);
	phent_setelasticity(my, 50, 10);
	phent_setdamping(my, 20, 20 );
	
	RL_wheel_ID=phcon_add(PH_WHEEL,my,p_player);
	phcon_setparams1(RL_wheel_ID, my.x, vector(0,0,1),vector(0,1,0));
	phcon_setparams2(RL_wheel_ID, vector(0,0,0),vector(-360,360,0),vector(0.9,0.001,0));
}

action RR_tyre
{
	my.shadow=on;
	
	phent_settype(my, PH_RIGID, PH_SPHERE);
	phent_setgroup(my,2);
	phent_setmass(my, 5, PH_SPHERE);
	phent_setfriction(my, 70);
	phent_setelasticity(my, 50, 10);
	phent_setdamping(my, 20, 20 );
	
	RR_wheel_ID=phcon_add(PH_WHEEL,my,p_player);
	phcon_setparams1(RR_wheel_ID, my.x, vector(0,0,1),vector(0,1,0));
	phcon_setparams2(RR_wheel_ID, vector(0,0,0),vector(-360,360,0),vector(0.9,0.001,0));
}

action car
{
	p_player=my;
	wait(1);
	my.shadow=on;
	my.material=mat_metal;
	
	phent_settype(my,PH_RIGID,PH_BOX);
	phent_setgroup(my,2);
	phent_setmass(my,40,PH_BOX);
	phent_setfriction(my,50);
	phent_setelasticity(my,50,10);
	phent_setdamping(my,20,20);
	
	vec_for_vertex(temp,my,306);
	ent_create(mdl_tyre,temp,FR_tyre);
	
	vec_for_vertex(temp,my,36);
	ent_create(mdl_tyre,temp,FL_tyre);
	
	vec_for_vertex(temp,my,5);
	ent_create(mdl_tyre,temp,RL_tyre);
	
	vec_for_vertex(temp,my,304);
	ent_create(mdl_tyre,temp,RR_tyre);	
}

---

quote:

---

vector(var x,var y,var z) ... The vector returned has the same restrictions like a local vector: it can't be passed as vector parameter to other functions. It can be used as vector parameter for vector instructions, but only for one of the instruction's parameters.

---

quote:

---

manual v6.00 -> PH_Wheel -> Parameter 4
... Use (-360, 360, 0) to remove all limits.

---

This you do! And perhaps this is the reason for your snapped wheel... because if you don't allow it to change its angle in this direction, the problem shouldn't appears !??
Because I haven't understand the usage with the angles yet (I wait for the german translation ), I don't know what the right valuas are (/should be).

Greets,
Clemens

P.S. Upload your car-demo when it is done [only a suggestion ]... And all free source code is also appreciated!

EDIT:
Two other things about your Code:
1.)
<EDIT> DELETED! reason: stupid question/idea...
2.)
Do you use a motor, yet?
If "yes":
Where is the updated code ?
If "no":
How do you give the car a force? By giving only the p_player a CENTRALFROCE ?

quote:

---

Sorry - nächstes Mal kommt was sinnvolles!

---

Das kann ich dir jetzt doch nicht so ganz garantieren, da mich dein Problem einfach nicht los lassen will - ich aber trotzdem geistig so am Ende bin, das wahrscheinlich wieder nur totaler Mist rauskommt...
Also, der wirklich letzte Versuch!
Ich habe mir dieses Bild (Handbuch -> PH_Wheel) nochmal angeguckt. Und da ich im Moment anscheinend mehr schlafe als wach bin, faszinieren mich schöne bunte Bilder mehr als logische Gedanken... daher:
Wenn man nun also wirklich vom Bild ausgeht und sich so davon faszinieren lässt (wie ich es im Moment tue), so liegt der Anchor-Punkt ja nicht genau auf dem Rad, sondern einzelne Stangen unterteilen das ganze. Vielleicht ist das das Problem deines abgeknickten Rades:
Dein Rad hat den Anchor-Point genau in sich drinne...

Ich seh's schon kommen: Morgen, wenn ich wieder einigermaßen geistig dabei bin - werde ich es bereuen, nicht einfach schon nach dem ersten Posting (am besten gar keins geschrieben zu haben) aufgehört zu haben! Abgesehen mal davon, dass ich morgen früh mir wieder überlegen werde, was mich davon abgehalten hat, etwa so früh ins Bett zu gehen, dass ich nicht aus Müdigkeit den Gedanken habe, so zu tun als hät ich Bauchweh oder so'wat, nur um länger schlafen zu können...

Endgültige gute Nacht,
Clemens

quote:

---

Originally posted by ventilator:
if a physics entity stops movement completely (but is still a physics entity) it does only require very minimal or no cpu performance right?

---

quote:

---

if i stack some boxes then i can notice that they jitter a little sometimes when i look very closely. how can i avoid this?

---

quote:

---

Ich habe mir dieses Bild (Handbuch -> PH_Wheel) nochmal angeguckt. Und da ich im Moment anscheinend mehr schlafe als wach bin, faszinieren mich schöne bunte Bilder

---

Danke, ich habe mir viel Muehe mit den Bildern gegeben
Der Ankerpunkt liegt dort wo sich die beiden Achsen schneiden. Die Raeder koennen einknicken, so wie es im Bild geschieht, wenn a) die Masse des Autos zu gross ist (bzw. die Gravitationskonstante), b) sich die Raeder zu schnell drehen oder c) die Spring/Damper Konstanten falsch gewaehlt sind. Ich muss mir das morgen im Buero genauer anschauen.

Es gibt ein Problem mit den Schatten, dass diese erst projiziert werden, sobald sich das Physikobjekt bewegt. Das wird z.Zt. untersucht.

code:

---

<edit>newer code above!</edit>

---

quote:

---

Originally posted by Seban:
Dude just look few posts down

I would try to define some more stuff
//
action sphere
{

my.shadow=on;

phent_settype(my, PH_RIGID, PH_SPHERE);
phent_setmass(my, 7, PH_SPHERE);
phent_setfriction(my, 70);
phent_setelasticity(my, 50, 5);
phent_setdamping(my, 20, 20 );
}

---

You're missing my point as you did with grey's original post. My fault as I should have posted later replies not the original.

The error is in phent_settype. I actually cut the above ACTION and put it in, attached to an entity and all the PH code returns the same error 'wrong/missing parameter; unexpected comma" for EVERY PH function. I've tried all manner of capitalization variations and experiments and always the same error for phent_settype. The others don't matter since if I can't get the entity registered, all else will naturally fail.

Again, the error is that I can't register an entity since settype won't work.

quote:

---

Originally posted by Marco Grubert:
Just to check the obvious, before we move to more complicated ideas: you do have A6 Commercial or Professional and run the sample with the A6 engine, right?

---

Actually, It is something obvious then. I'm using the trial edition in consideration of upgrading to A6 pro edition comes in. Since trial == extra, not commercial, the physics engine isn't enabled in the trial version; Silly really considering that if I didn't already need the physics engine and I can't test it in the trial version, how do I know what I'm paying for?

Never occured to me that I would have to pay to test out the upgrade is what I'm saying; Not a good way to promote the key features of the upgrade IMO.

Anyways, thanks Marco, I'm sure that's all it is. I'm quite impressed with the promises of A6 and thus I'll get the pro upgrade tommorrow. My guess is that I won't have any problems then.

code:

---

 

entity* myCrate;
action ball
{
myCrate=my;
phent_settype(myCrate,PH_RIGID,PH_SPHERE); 
phent_setmass(crateMass);
} 

---

quote:

---

Originally posted by elsewood:
Make the code like that:

code:

---

 

entity* myCrate;
action ball
{
myCrate=my;
phent_settype(myCrate,PH_RIGID,PH_SPHERE); 
phent_setmass(crateMass);
} 

---

This DOES NOT fix phent_setmass!!, only phent_settype

---

That's because phent_settype is correct above but phent_setmass isn't. _setmass takes 3 parameters.

Given that i'm pretty sure my woes are edition related, sorry I ever brought this up

code:

---

<edit>newer code above!</edit>

---

i haven't continued working on it but this is my current code... the car doesn't jump around but the front wheels don't rotate properly (see screenshot below) and so i can't steer. driving forwards and backwards works!

the wheels are attached stiffly. i though the wheel constraint has dampers? is that what the spring/damper constants should set?

quote:

---

max_ph_entites
Determines the number of entities that can be registered for physics.

Type:
Variable, read/only.

Range:
0 - no physics engine available (S E)
1 - physics for 1 object (C)
999999 - unlimited physics (P)

CAPS_PHYSICS
Defined if the engine is capable of physics (C P).

---

code:

---

FR_wheel_ID=phcon_add(PH_WHEEL,car_frame, wheel);
phcon_setparams1(FR_wheel_ID,wheel.x,vector(0,0,1),vector(0,1,0));
phcon_setparams2(FR_wheel_ID,vector(-40,40,0),nullvector,nullvector);	

---

quote:

---

Originally posted by ventilator:
i still don't know what the spring/damper constant specifies?

---

quote:

---

Sets global correction factors for the physis simulation. ERP stands for Error Reduction Parameter and defines how quickly misaligned constraints get adjusted. Setting ERP to 0 will have constrained objects drift apart after some time because errors get not corrected. With ERP set to its maximum value any constrained objects that get forced into an unacceptable position/orientation will be corrected within a single frame. This however can cause overshooting and it is thus recommended to use smaller values instead. Constraint Force Mixing on the other hand determines how much a constraint is allowed to be violated. Setting CFM to 0 results in very stiff joints (this will lead to errors though- do not do it). Setting CFM to its maximum value makes constrained objects act as though limited by rubber bands

---

quote:

---

if i want to have shocks / suspension, do i have to use an additional slider constraint for each wheel? and how would i set this constrain up so that it works like a spring? just set a motor for the slider which is active all the time?

---

You could soften up the wheel constraint by altering CFM/ERP, or use a slider attached to the wheel. Try setting the slider motor to a desired velocity of 0, with some margin in its max/min limits. If this does not work then use the technique described below: query the slider position and apply a positive/negative motor velocity depending on whether the position is <0 or >0. Make sure to time-correct this value or else you will have a "lowrider" if the framerate drops and the slider overshoots.

quote:

---

how could i force the front wheels to stay parallel?

---

quote:

---

Motor1: Set desired angular velocity around axis1 and the maximum torque allowed to achieve this velocity, e.g. (10, 50,0) will set a desired angular velocity of 10 Hz and a maximum torque of 50,000 Nm
Motor2: Set desired angular velocity around axis2 and the maximum torque allowed to achieve this velocity, e.g. (10, 50,0) will set a desired angular velocity of 10 Hz and a maximum torque of 50,000 Nm

---

quote:

---

would the cylinder hull work better (less slip) for the wheels because it has more ground contact area than a sphere?

---

thanks for the answers!

i tried to adjust the spring/damper parameters now. it works pretty well but the car wouldn't pass the elktest yet!

quote:

---

b) The other way is to query the axis1 orientation of both front wheels (phcon_getposition) and then apply motor values (scaled by time) in order to rotate them to the desired position.

---

code:

---

function flipl()
{
	phent_addtorquelocal(p_player,vector(-100000,0,0));
}

function flipr()
{
	phent_addtorquelocal(p_player,vector(100000,0,0));
}

on_z=flipl;
on_x=flipr;

---

..could be helpful if your car flips over!

...
the q, a, w, s keys i had in my code didn't do anything except changing the values in the debug panel. so i removed them...

quote:

---

Originally posted by ventilator:
is it possible to apply more than one constraint to an entity?

---