PDA

View Full Version : ? Table on 4 pistons: IK solution?



jeric_synergy
11-20-2016, 11:30 AM
I'm not really interested in any Bullet or dynamic solutions here......

Imagine a heavy 'table' supported on 4 hydraulic pistons.

Is there an IK rig, maybe connected to sliders/nulls, that would effectively give the animator separate control over each piston, and would accurately tilt the table according to various extensions of the pistons?

...as I typed that I was thinking something like ...bones in the table targeting the ends of the pistons..... :stumped:..... Can't figure out what the hierarchy would look like...

jeric_synergy
11-20-2016, 02:08 PM
still thinking about this.

Im thinking an X of bones in the table, targeted to nulls representing the tops of the pistons, with the table itself constrained to be at a Y position that is the average of the Y position of the four nulls. That sounds like it might work.

Now to figure out how to "constrained to be at a Y position that is the average of the Y position of the four nulls". Right now, my guess is an Expression.

++++++++
hmmm, maybe a perfectly flat references, HardLinked to the actual table mesh.....

RudySchneider
11-20-2016, 03:16 PM
Rather than four pistons, have you considered a Stewart PLatform? This allows full six-degree motion control: up/down, lateral, and rotational motion:


135136


https://youtu.be/Yilr3kQkNk4

jeric_synergy
11-20-2016, 10:47 PM
Rather than four pistons, have you considered a Stewart PLatform? This allows full six-degree motion control: up/down, lateral, and rotational motion:
135136
https://youtu.be/Yilr3kQkNk4
LOL, to be sure, but still: how would one rig such a thing in LW? :beerchug:

ernpchan
11-20-2016, 11:23 PM
This would be a basic ik targeting setup. The top part of the piston would target the bottom and vice versa.

jeric_synergy
11-20-2016, 11:32 PM
This would be a basic ik targeting setup. The top part of the piston would target the bottom and vice versa.
It's not the pistons that are confusing me: it's the table. Because its attitude and location are dependent on 4 different points.

That is: I want the pistons (actually, 4 nulls) to drive the table, not the table drive the pistons.

pinkmouse
11-21-2016, 03:15 AM
As I said the last time you brought this up, mechanically, that design won't work. Imagine you have a solid table, with fixed legs. Alter the height of only one leg and it will rip the table apart or shear the leg off. To stop that happening, each leg/piston needs to be in a ball joint with three degrees of freedom at the top and bottom. Now you have freedom of movement, but the whole thing will become dynamically unstable, and the slightest sideways movement will cause the whole thing to topple over. That's why the Stewart platform was invented.

As for the animation, to drive the pistons from nulls I see no other option than to go nodal, but it will require lots of complicated maths, (currently beyond me! :) ), as each piston will need to be constrained by the the bottom fixing and the corner of the table it attaches to, so changing the length will not just change the height but also the angle of the table top and the angle of the piston, as the "corner" will move in 3D space.

RudySchneider
11-21-2016, 07:52 AM
...how would one rig such a thing in LW? :beerchug:

I modeled and animated that youtube video in LightWave. As ernpchan mentioned, rather than a "rig," per se, it's target based. I used to work for a company that manufactures motion platforms. The Stewart platform is one example, and yes, the control algorithm is tricky, but it' just a lot of trigonometry.

RebelHill
11-21-2016, 08:35 AM
I want the pistons (actually, 4 nulls) to drive the table, not the table drive the pistons.

Nope... because a change in any one piston changes the WHOLE table, not just that one corner, and thus the other pistons MUST update their positions to fit that table position. Each piston can only be controlled either manually or automatically... not both at the same time.

Pose the table, automate the pistons.

jeric_synergy
11-21-2016, 09:05 AM
Interesting, seemed a simple problem but obviously isn't. Perhaps dynamics is the way to go.

Thanks for everyone's expertise.

RudySchneider
11-21-2016, 09:39 AM
If you don't need full 6-DOF motion, you could have a table with three pistons. This would allow up/down, roll, and pitch motions. In this case, each piston would be locked on one end, with a U-joint at the other. Since three points determine a plane, this would allow you to move each piston independently without worrying about "breaking" either the table or the pistons.

jeric_synergy
11-21-2016, 10:29 AM
In my imagined scenario, the table connected to the pistons via some sort of retaining device, maybe something as simple as a wide sleeve, that the piston top slopped around a bit in as the pistons became slightly mis-matched. In general, the scene will have the platform essentially level, no fancy moves necessary.

RudySchneider
11-21-2016, 10:39 AM
Sketch?

pinkmouse
11-21-2016, 11:21 AM
If you don't need full 6-DOF motion, you could have a table with three pistons. This would allow up/down, roll, and pitch motions. In this case, each piston would be locked on one end, with a U-joint at the other. Since three points determine a plane, this would allow you to move each piston independently without worrying about "breaking" either the table or the pistons.

Still wouldn't work. The three points may define a plane, but move one and the "triangle" changes shape, i.e. the table breaks. ;)

RudySchneider
11-21-2016, 11:26 AM
Still wouldn't work. The three points may define a plane, but move one and the "triangle" changes shape, i.e. the table breaks. ;)

Ah, yes, you're correct. I mis-spoke just a bit. One piston can be fixed, but at least two need a U-joint on both ends: https://www.youtube.com/watch?v=VvBojB_RUBA

To further clarify: One piston is constrained to move strictly up and down, but five of the remaining six piston vertices need a U-Joint.

jeric_synergy
11-21-2016, 07:16 PM
Still wouldn't work. The three points may define a plane, but move one and the "triangle" changes shape, i.e. the table breaks. ;)
Well, considering I just moved, in RL, a 4'x8' half-inch steel table and it did NOT break, including placing it on its side, I'm pretty sure there's room for some slop. :\

And it was bloody heavy.