Physics experiment

[detbear]

Hey everyone,

 

I need to simulate(If possible).....a weighted but Bouyant object in a tube filled with water.

 

It needs to be able to be flipped kinda like a "Sand Timer" and have the bouyant object rise to surface according to real world physics.

 

Is it possible to use Newton Physics in A:M to simulate Bouyancy like this?

 

Thanks so much,

 

William

[robcat2075]

How many times does it have to turn over?

[robcat2075]

How about.... make the object out of stiff "cloth", put it in a tube that is a cloth deflector and set the gravity of the scene to be upside down. Then it would always rise instead of fall.

[Fuchur]

You may want to have a look at the ocean rig.

It is really a very powerful and cool thing!

 

See you

*Fuchur*

[robcat2075]

Ta-daahhh...

 

float.mov

 

 

Float03Simmed.prj

[detbear]

Hey Rob,

 

Looks really cool. I'll have to study what you have going on there. I'm in a research mode where I need to accurately simulate something virtually. So it needs to accurately work like Archimedes principle. I was hoping newton plugin would be able to do that, but it may not.

 

REGARDLESS, your example is really awesome. Thanks for working that up. It looks really good. I need to study the project to see how you got there.

[robcat2075]

"Air simulated with Static Fluidsurface" looks like it might be a simulation of an object floating to a certain level.

[detbear]

Very hard for me to easily get Newton to work with what I'm trying to do. Crashes every time. I'm sure theres a way to get it to work but too much trial and research time.

[robcat2075]

Very hard for me to easily get Newton to work with what I'm trying to do. Crashes every time. I'm sure theres a way to get it to work but too much trial and research time.

 

Cam you post the PRJ? It might at least make a AMReport.

[robcat2075]

Maybe the "air" project wasn't quite what you are looking for.

 

I just tried the static fluid surface "simplefluid" project and modified it to start the "Brett" (board) below the surface of the water and it did indeed bob to the top. Perhaps there is hope there.

[detbear]

I have a sphere set as a dynamic spheroid and when it drops it goes straight through a Dynamic object below. This Below dynamic object is set up on a base so that it will spin. The base is set up as a static object.

 

However.....when the sphere drops, it passes straight through the dynamic object. I even copied the settings on the objects in Steffen's "Wing" example and it still didn't work.

[robcat2075]

"Wings" seems like a completely different thing than your floating object situation.

 

However, when i run the simulation on the Wings PRJ it does seem to work so I have to presume there is some setting not set in your modification of it.

[detbear]

Robcat.....I agree with ya 100%....... But so far I have investigated all elements in each model and their settings in the choreography....and haven't discovered where I'm going wrong. My setup does not have the same Paddle type geometry though. I did notice that the paddle wheel has a different Gram mass on the individual paddle flaps while the rod/ arms are set to a lighter mass(in the models "newton"groups)

 

I'm looking at this also...because there are more physics in the project than just buoyancy. Trying to get my feet wet in the newton plugin world.

 

I will endeavor on

 

William

[robcat2075]

Well, you've made me start trying Newton.

 

I find that normals have to be facing out to prevent pass thrus.

 

Here's a test i made from scratch where a ball knocks a spoon around in a bowl. These are default settings except for "linear damping" raised to 0.1 on the ball to make it settle down eventually.

 

Setting the spoon's "high velocity object" ON makes the difference between it passing thru the bowl after the ball hits and not passing thru.

 

SpoonHiVelocityOFF_ON.mov

 

SimpleDropTest02BallDamping_unSimmed.prj

[detbear]

Watch how this balance beam gets thrown out of its seating starting from the first frame of simulation.

 

newton_test_A.mov

[robcat2075]

Watch how this balance beam gets thrown out of its seating starting from the first frame of simulation.

 

newton_test_A.mov

 

Maybe the clearance between the parts is too small

[detbear]

After I made the rings a good bit larger, the same results still happen. The drive goes flying away.

[robcat2075]

I'll look at it if you want to send it to me.

[robcat2075]

I've tried creating the rotor from scratch and mine flies off too. But somehow the original works fine.

[robcat2075]

If i just simulate my axle with no arms off it it will behave. If I add one arm it will try to fly off and if i add both arms it does fly off.

[robcat2075]

Any Newton experts who can make this test setup work? All it needs to do for now is to settle on the rings and not fly off.

 

Simple_Rotor01.prj

[agep]

Any Newton experts who can make this test setup work? All it needs to do for now is to settle on the rings and not fly off.

If you enable the setting "create hull for starttime" A:M will create a model that represent how Newton sees it. By doing this on your Rotor model we clearly see why it pops off, and why it worked with just the crossbar

 

 

You can fix this simply by create Newton_0x groups for each part of the model you want Newton to interpret separately. Looking at the new hull, you see that it looks more correctly

 

 

And now its working:

 

newton_000.mov

 

I hope this made any sense

 

Stian

[robcat2075]

I hope this made any sense

 

that makes a lot of sense! thanks!

[detbear]

So what settings do you make for each of the "groups?"

[detbear]

OK..... I think I got it. Kind of like naming bones specific things to work in a rig....Newton requires the "newton 01,02,03....." group naming convention to trigger it's effect on objects.

 

 

Hey Agep.......On an earlier post in this topic I was asking about adding Buoyancy to a simulation. Is it possible to do this somehow?

 

Thanks,

 

William

[detbear]

I'm noticing that my droping ball never actually makes contact with the rotor but influences it with some remaining space between it. What setting can I make to allow an actual "touch?"

[serg2]

Rotor - Example.

 

I understand how works plugin "Newton physics helper" - Creates a bone in the center of "Newton_#N" or "Explode_#N" groups

 

Such a structure model is required to determine the internal spaces, which are common within the bounding volume model.

 

Simple_Rotor09.prj

 

 

ADDON

 

Explode Rotor & Holder

 

Simple_Rotor09_Explode2.prj

[robcat2075]

I'm noticing that my droping ball never actually makes contact with the rotor but influences it with some remaining space between it. What setting can I make to allow an actual "touch?"

 

If there isn't a setting for that... I'd do the simulation, then swap in a slightly larger ball that closes the gap.

[robcat2075]

I just noticed that Newton is creating keyframes at 25fps even though my PRJ is 24fps. Is there a Property somewhere I need to set for that or is that a bug?

[robcat2075]

I'm noticing that my droping ball never actually makes contact with the rotor but influences it with some remaining space between it. What setting can I make to allow an actual "touch?"

I'm starting to think this gap may be due to the mismatch between newton simulating at 25fps and A:M interpolating to 24 or 30 fps. If I simulate something and set my PRJ to 25fps the object contacts seem to be exact.

[detbear]

Maybe the Newton plugin is set to calculate in only the PAL scale.

[Gerry]

Just *slightly* off topic here, but have you heard the one about the neodymium magnet and the copper tube?

 

[robcat2075]

Just *slightly* off topic here, but have you heard the one about the neodymium magnet and the copper tube?

 

 

Wow, I keep waiting for him to get a magnet thru the skull.

[serg2]

I'm noticing that my droping ball never actually makes contact with the rotor but influences it with some remaining space between it. What setting can I make to allow an actual "touch?"

I'm starting to think this gap may be due to the mismatch between newton simulating at 25fps and A:M interpolating to 24 or 30 fps. If I simulate something and set my PRJ to 25fps the object contacts seem to be exact.

 

 

Newton calculates for any fps

 

for 39 fps (a random value)

Simple_Rotor09_39fps.prj

Simple_Rotor09_39fps_SYM.prj

[robcat2075]

Newton calculates for any fps

 

for 39 fps (a random value)

 

How do you set it?

[serg2]

for you example from Post #14

 

Just set Scale World = ON

 

SimpleDropTest02BallDamping_unSimmed_0.prj

SimpleDropTest02BallDamping_unSimmed_SYM.prj

 

 

 

Newton calculates for any fps

 

for 39 fps (a random value)

 

How do you set it?

 

[robcat2075]

for you example from Post #14

 

Just set Scale World = ON

 

I just tried that but I still got 25fps from Newton.

[serg2]

for 30fps

 

SimpleDropTest02BallDamping_unSimmed_30fps_SYM.prj

[robcat2075]

for 30fps

 

SimpleDropTest02BallDamping_unSimmed_30fps_SYM.prj

 

I'm not getting it. that one definitely has keyframes made at 25fps. If I set the PRJ to 30 fps the keyframes are not lining up with the timeline numbers.

 

[serg2]

Yes. I put different fps during the simulation.

Noticed that during simmulyatsii FPS forced to become equal to 25 - and on completion of the simulation was set to any initial value.

This is strange.

[robcat2075]

Well, I've made an AMReport of it. Since Steffen did the Newton plugin he'll probably know if it's really a bug or if I'm just doing it wrong.

[serg2]

Good )

[robcat2075]

Steffen has fixed it so test it out in the next beta

[serg2]

it interesting