Thingiverse
Tensegrity Table OpenSCAD by clsn
by Thingiverse
Last crawled date: 5 years, 4 months ago
OK, my first Thingiverse upload is going to be lousy and poorly-organized, because I'm tired of refraining from uploading things because they aren't "right."
There've been a bunch of variants on the "impossible table" lately, the "tensegrity" table that is held together by strings and doesn't look like it should work. I should have links here but I don't. I had already made these "final word" (to me!) scripts on the subject, but it's really something you have to tinker with in the code, not set up for customizing, etc. Also not all that well-written. But people might want to see them anyway.
First attempt is impossibletable.scad. Mostly customize at the top: choose how many sides (triangle, square, etc), the diameter of the circumcircle, height, size of various parts, etc. This version by default creates spokes from the center of the table to the corners but doesn't make a flat surface (you can always put a plate on it or something). Experiment with it a little if you like. (You have to print out two of them, and the same is true below.)
But the better one is newtable.scad. Many of the same parameters, but used a little better and more consistently. You have two modules to choose from at the end: regulartable() and multilegtable(). The first makes a table like impossibletable.scad, with one leg, on a spoke toward the center. The other makes a table with legs along each side of the polygon, so you can make it stronger.
It's pretty flexible and also pretty good mathematically. Both modules take parameters (corners, hi, lng, howmuch). corners determines the polygon, hi is how high the legs should be, lng is how long horizontally (that is, their projection on the horizontal plane), so you use these to control the angle of the leg. howmuch determines how high up the leg the reinforcement goes: this is a solid prism from the point underneath the tip up to some fraction of the leg.
It's reasonably smart, in that if at all possible, it will make the tips of the legs of the multilegtable fall out exactly at the halfway point along each side, which prevents there from being any sideways torque.
For tying the strings, I've been experimenting with all too many possibilities. You can go with the single strand with a knot or the loop (yes, that is something that is missing: there should be a notch in the tabs for the loop). I've even made one using magnets (attracting) for the center string (obviously not a great plan practically: strings pull stronger when you stress them; magnets pull weaker. A little too much pressure and it falls apart.) If you have the Ashley Book of Knots, check out knots #1987, #1988, #1989 for ones you can slide along and then lock in place. #1987 is probably best if you're using a loop.
Still tough to get them Just So. Been working lately with trying to "hang" the top from a framework placed on top of some cups (to hold it level) and doing the outside strands first; seems to have some success.
OK, sorry this is so disorganized; otherwise it would never get out because I'd spend forever getting it "right".
There've been a bunch of variants on the "impossible table" lately, the "tensegrity" table that is held together by strings and doesn't look like it should work. I should have links here but I don't. I had already made these "final word" (to me!) scripts on the subject, but it's really something you have to tinker with in the code, not set up for customizing, etc. Also not all that well-written. But people might want to see them anyway.
First attempt is impossibletable.scad. Mostly customize at the top: choose how many sides (triangle, square, etc), the diameter of the circumcircle, height, size of various parts, etc. This version by default creates spokes from the center of the table to the corners but doesn't make a flat surface (you can always put a plate on it or something). Experiment with it a little if you like. (You have to print out two of them, and the same is true below.)
But the better one is newtable.scad. Many of the same parameters, but used a little better and more consistently. You have two modules to choose from at the end: regulartable() and multilegtable(). The first makes a table like impossibletable.scad, with one leg, on a spoke toward the center. The other makes a table with legs along each side of the polygon, so you can make it stronger.
It's pretty flexible and also pretty good mathematically. Both modules take parameters (corners, hi, lng, howmuch). corners determines the polygon, hi is how high the legs should be, lng is how long horizontally (that is, their projection on the horizontal plane), so you use these to control the angle of the leg. howmuch determines how high up the leg the reinforcement goes: this is a solid prism from the point underneath the tip up to some fraction of the leg.
It's reasonably smart, in that if at all possible, it will make the tips of the legs of the multilegtable fall out exactly at the halfway point along each side, which prevents there from being any sideways torque.
For tying the strings, I've been experimenting with all too many possibilities. You can go with the single strand with a knot or the loop (yes, that is something that is missing: there should be a notch in the tabs for the loop). I've even made one using magnets (attracting) for the center string (obviously not a great plan practically: strings pull stronger when you stress them; magnets pull weaker. A little too much pressure and it falls apart.) If you have the Ashley Book of Knots, check out knots #1987, #1988, #1989 for ones you can slide along and then lock in place. #1987 is probably best if you're using a loop.
Still tough to get them Just So. Been working lately with trying to "hang" the top from a framework placed on top of some cups (to hold it level) and doing the outside strands first; seems to have some success.
OK, sorry this is so disorganized; otherwise it would never get out because I'd spend forever getting it "right".