If a pair of sticks had a diagonal connecting bar on only one side of the hub, making a sort of U shape, the other pair parallel with it could have a connecting bar on the other side of the hub. It would still be possible to pull the sticks out in one direction, but there would be a “safe” direction where it collided with the hub, and there would effectively be half as many stick components to get lined up in sync.
that is interesting, it would remove some of the symmetries of the system (that is, it would halve the amount of them), and visualizing it in my head, i cant decide if i find it more aesthetically pleasing or not
@@terdragontra8900 The symmetries of the physical object will always be imperfect. But what we should be contemplating is an idealized version, where the racks are replicated out to infinity, and there's an infinite 3D grid of cores. Or how about multiple interleaved grids?
What comes to mind for easier assembly is: 1. Make a jig which can hold all the racks and the core in the right positions, but doesn't touch the gears. 2. Assemble the core puzzle. 3. Put all the parts in the jig except for the gears. 4. Slide the gears onto their axles. That way, all the racks can be put into their central position without any constraints, rather than needing to start in the “just falling out” position all at once.
You could have temporary gears without teeth, but the correct contact diameter and an extra flange, to work as the jig. You could then push the racks in place one by one and once aligned, then start changing the temporary gears into actual ones one pair at a time.
5:08 Different Idea: you can have the driving rack have a waffle iron pattern and the driven racks can both be driven. Think of the driven racks as tools that carve tracks out of the driving rack, if you carve both patterns out of the driving rack it will only have little parallelogram nubs that don't interface with the whole sliding surface, but it will be able to go both ways
This idea is used in drum winding mechanisms, with the addition that the pattern "merges" at the ends, so that a follower automatically turns the other way.
Here's what worked for me! First, mark the center of each rack on the back side. Then, starting with a bare hub (no pinions), arrange four racks and put a rubber band around each end of each set to hold them to the hub. Then again with another set of 4 and rubber bands. Get them all approximately centered; they don't have to be in exactly the right position or rotation, just close. Then choose a pinion that will contact three of the racks and insert it (with the bolt), careful to center each of the 3 racks it contacts as you do. Repeat for the opposing pinion, then the other two that contact the racks placed so far. Repeat one more time with the remaining four racks and two pinions. I'm not saying it's easy, but it was relatively straightforward. EDIT: Also, thanks for the cool design, and informative video! I'm sure it will be a great conversation piece!
It's a bit hard to explain, but there's another possible configuration: the bottom rack could have a grid of diamond shapes to act like overlapping diagonals going in opposite directions, then each of the top racks could have diagonals that going opposite directions like the last demonstration he showed.
@@minecraftiangenetics Yes I had the same thought. It's functionally the same (or at least very similar to) as the bolt and nut that were machined to have both left and right handed threads that was doing the rounds on RUclips midway through last year.
lovely - i reached for my star burr puzzle and pulled it apart as you did in your video - so thanks for naming that wooden puzzle of mine that i guess me dad bought decades ago - but also - i now have to remember how to put the star burr puzzle back together again - x
You could connect the axial pairs by using concentric rings, just set the rings on the inside or outside of the pair and have mounting tabs, this would allow for a full range of motion. Stemming from this, the concentric rings could also have a guide pin that sits on the outer face of the rings with a pole that sits on the curved surface of the core, where the axle for the gears sits, this would help for assembly. For assembly, the various constraints are what make this easy, the system is constrained to where all centerpoints align, and the gears just slot into place, the guide pins are a temporary constraint whereas the gears are a permanent constraint. For clarification on the concentric rings: Say the square of rods is a base circle. One pair of rods would get a larger concentric circle, the other pair gets a smaller concentric circle, both with mounting tabs for the pair of rods they connect to. For example, axis positive would get a larger circle, axis negative would get a smaller circle; though, personally, I think each pair should get both sizes of rings on the two ends, which could be further used as an endstop against the core. The guide pin would have a flange that seats on both circles at the same time, thus constraining them to the same centerpoint position; and in combination with the guide pin's rod seating on the core, this constrains the axial pair's centerpoints to the core's centerpoint. Let's look at one set of rods, say the X-axis. The core constrains the rods in the Y and Z axes, allowing free movement within the X-axis. The concentric rings will constrain the positive and negative pairs of the four rods, which this later becomes redundant with the installation of the gears. The guide pin will constrain the X-axis position of both of these pairs; at least as a soft constraint with one pin, and a hard constraint with two pins, but really only one is needed. Repeat this for the Y and Z axial pairs. The installation of the gears would now constrain the rods to the core, the two opposed axial pairs together, and an axial rod to its perpendicular neighbor, thus constraining everything together. The original product already had this final constraint, but with a more difficult assembly, but mimicking these constraints during assembly makes the process easier, and as a coincidental byproduct the system that binds two rods together is also usable for setting up the assembly constraints.
5:22 On the lower single stick you could change it to a diamond pattern where it is both diagonals cut out of it. You could also add a connector bar to each end of only one pair of sticks per side and it would also stop the sticks from falling out while allowing the ends to move past each other (add the bar to both ends of the pair moving one direction along the x axis but not the other parallel pair of sticks that moves the opposite direction and same for y and z axis).
Wow, this is absolutely fascinating! I can't think of any practical use cases in my head right now (though there are probably a good amount), but I'm thinking some implementation into a crazy CNC based system for experimental manufacturing!
Hey use giant pendulums as a protection from earthquakes. I suppose it could work in a similar way but in smaller scale. Or instead of a gearbox add a gyroscope and you could dynamically change directions of spacecrafts or small satellites.
I find all of these so mechanically pleasing. To assemble, line up all the racks on the core with toothless gears/wheels so everything slides easy, tape racks to hold the shape, and swap out the wheels for gears 1 by 1 meshing teeth as you go.
To stop the racks from falling out of the core you could make little screw on caps for the ends of the racks that would be too big to get pulled out, similar to how the gears are mounted but with a flange rather than a gear.
what happens when you do the design on 5:40 and you extend/overlap the pattern? they always keep following there straight what so it is possible to have more motion right?
at 5:12, I wonder what would happen if you didn't make all the racks have grooves in only one direction, but formed a kind of a diamond-shaped, bidirectional grooves (a diagonal criss-cross pattern). Something similar to those bidirectional bolt threads: ruclips.net/video/v96LTfmtDPU/видео.html
My first thoughts exactly. Just make use a Boolean intersection operations between the two mirrored angles. But the problem I see with that is that ALL the bars would have to have that design, so I'm not sure if the direction could be maintained between two interfacing diamond patterns. Maybe make half of them diamond? Loses some symmetry, but maybe a working configuration is possible.
My guess is that it might be likely to jam. There are 24 places where two sticks move against each other. If any pair of touching sticks can decide to go whichever direction they want, there could be conflicts. I don't know though, maybe it would work out!
@@henryseg I also thought of the same idea. I think OP meant that only the bottom rack would have the diamonds. The two racks on top can remain unchanged.
@@cliveso Interesting... but in the whole configuration, I think that every rack would need to interact with two other racks that move in opposite directions. Maybe that's still not a problem though? I'm not sure if there's a way to split the racks up into "diamond" and "non-diamond" so that diamond racks only interact with non-diamond racks.
There is a way of having this be a rack-rack interaction: Just have one of the racks have a pattern of a diamond lattice that you get when you do a union of the two diagonal rack patterns.
If you add a size increase at the end of the sticks so they won't fit through the gears it will block them from moving through in that direction. If you add that as a stick-on to the other end as well, it will be forever locked in place.
Interesting if you place a Cartesian coordinate system at the center of the gear mechanism you'll have a mechanism of constant mass centroid location but variable off axis moment of inertia, product of inertia
personally, the star burr puzzle is my favorite. as at first glance it seems impossible to put together or take apart because each piece holds each other like a borromean rings. but once you know the solution, it's a trivial dexterity toy.
You could get all the sticks in the right place easily by having cylinders that are the same as the cogs but without the teeth on the hub, that would restrain the racks' movement to the correct axis without coupling their movement to one another. You could then go around and replace each of these cylinders with a cog to add all that coupling back in. You could also use this as a basis of a way to stop the racks falling out by skipping the first and last notch for the gear teeth to go into on each rack, that way they would slide in and out easily with the smooth cylinders, but once you have the cogs in place they would be blocked from turning any further when they get to the end of each rack. You could even use that as the way to get them lined up correctly when putting the cogs on by having the cog start off meshed with the first or last remaining notch on the racks it meshes with.
About making a rack push two other racks in opposite directions: Could you make worm drive racks such that the linear motion pushes in one direction while the screw spinning motion pushes in the opposite direction? Then the two racks you want to push...i guess one would be sliding and the other rotating in a threaded hole, to go oppositely? I suppose that wouldn't 'count' either.
For version that do not have cogs, if you make bar with only row of pegs they could drive two orthogonal bars that are set in opposite directions. Round pegs will not block move of orthogonal bars as they are exactly same when mirrored.
lovely design and would it not be easy to build clips that grab onto the pairs of racks and while they are bottomed out in the core attache the gears and bolt it down
5:35 There is a hand powered screw driver called the 'Yankee Screwdriver. The core has two sets of paths for one rod, somehow it can be pumped to make a continuous rotation. I've also seen a bidirectional nut, that has two paths cut into it. Would this concept be used to solve this particular problem?
Hello! 5:30 If you cut the two-way slat where it meets the transition of the teeth on the other slat, you get a two-way "X" toothing that runs its entire length. Best regards: A.P. Hungary🙂
Super awesome! You have a beautiful mind. Speaking of moment 6:57, you could connect bars a and d on the right end, and connect bars b and c on the left end.
5:37 instead of having 3 plates all designed with slats, have 2 slatted plates (opposing directions of the slats) interface with a diamond studded plate? That way the slats themselves determine the direction that they move?
To put them together I would take out the gears, put all horizontal stix in the core (stack themfrom lowest to highest), then stick the vertical stix in a fixture or a lump of putty on the table. That way I could (hopefully) line up all the stix in equal llength and have one hand free to reinstall the gears, locking everything together.
I wonder if you could make the racks largish circles (ideally either small enough to still be printed or else arcs that can be connected together) with the core riding on the circle what that would look like, could you connect multiple cores per circle, what about loops of core-circle-core-circle-core-circle-core-circle in say an octahedral arrangement (because of the even number of edges around a vertex)?
It is beautiful, as always ✨. I wondered if there was a way to move two different ways using the principle of the two-way screws ? With the driving stick having a kind of grid pattern ?
On 3:36 you claim that a six axis rack has the tetrahedral symmetry. A tetrahedron has 4 corners, 4 sides and 6 edges (all at different angles). Your six axis rack has six "ends", all rods are parallel to one of 3 orthogonal axis. So your six axis rack has the same symmetry as a borromean rack or a cube (6 faces, 8 corners, 12 edges parallel to 3 orthogonal axis), not a tetrahedron.
If you want 2 racks to interact with a single one moving in opposite directions you could cut the groves all the way in both directions on the stationary rack leaving diamonds behind (similar to the people who have cut both left and right handed threads on a screw).
Instead of having bars going across pairs of sticks, you could just put attachable nubs onto the ends of the sticks (e.g. bolt on) that are bigger than the holes between the gears, which would prevent the sticks from going all the way through.
No need for gears. It's enough if each one from the six axis interacts with two others, not three. You can make a closed chain of the six sticks that way. Now, get a cube, set it on one corner, and look from above, so it will look like a hexagon. Arrange that chain of six sticks onto the sides of this hexagon, alternating between "upward" sides of the cube, and the "downward" sides of the cube. Done, you "just" have to design the way to keep them together. On each corner of the hexagon you have a join between two axes. It will look somewhat similar to the three axis solution, but the sticks that lay on the same axis will be moving in opposite direction to each other.
In 5:36 would it work if you had this alternating pattern on all 3 racks, instead of just on the bottom one. So when one of the top racks is about to stop, then it switches to the right pattern and can continue the sliding?
As far as how to synchronize them... I see that they've got a 1/2 tooth offset on the two sides of each stick. So if you had a worm gear with 2 teeth on it (double-helix) you could insert those one by one, thread them in/out until they're all the same length, probably via some sort of mark at the center of each worm, then temporarily attach the sticks to the end of each, push them all through in sync, then remove the worms from the sticks.
instead of sticks would it be possible to make this with rings so you they can just keep going an every direction? i think it would look really cool although that might put it past what's feasibility printable size-wise
I wonder if it would be possible to turn the sticks into hoops, so instead of going back and forth over the length of a terminated stick, the gears of the core would keep rotating in one direction while the hoops "rolled"
5:31 - what if the bottom rack is "threaded" both ways (i guess it would look like a series of rhombuses instead of teeth) so that when it moves, the two top racks move opposite directions?
There was a video going around a while back where a machinist made a screw that was threaded twice, so that a left hand and right hand nut would move on it, that's what makes me think of this.
2-way hinged crossbar/caps across the pairs of sticks that move in the same direction parallel to each other would allow them to cross through another pair of sticks so movement isn't limited, but perhaps the bulk of the axis mechanism could block the hinged crossbars from allowing us to easily deconstruct the sculpture? I was thinking of how nice the dual-directional swinging doors of a laundromat allow 2 directional flow or with lack of a central bar create one large opening for larger loads...I'm not a mechanically minded person so I may be relying too much on the central mechanism to somehow block the "doors" from allowing the sticks to be pushed all the way out, but feels close.
Since the core holds together without the nuts, wouldn't it be possible to get the sticks lined up more easily by installing the gears *after* the sticks, rather than before?
@henryseg 9-5 Could you make some kind of rack version of Moebius strip like Moebius racks somehow?? Thanks for sharing Henry and really hope you can answer this and my other question or email whenever you can. I would really appreciate it. Thanks very much.
Would love to see this skinned in a silicone blob shape. I think it's movements might confuse people. Ever considered going into animatronic territory?
you can have racks that go both way (mesh in two directions) without the gears. Similar to a nut and bolt that can screw and unscrew in both directions.
Could the core have a zig-zagging groove pattern, and the interfacing sticks only mesh with either the zigs or the zags, going in opposite directions to each other? I don't have quite the engineering mind to imagine this, but if someone could explain or experiment, I would love to hear if it works or why it doesn't
6:52- put the joint on the other side, or in one it's on the edges and the second is on the middle, if you have access to it. actually the middle feels like a bad idea...
The two aligned "sticks" on each axis should be connected with a bar. The two opposing sticks should be connected with a RING around the periphery. This will link the 2 "sticks" without them colliding. YOU'RE WELCOME!
Not 100% sure what your main goal is here, but im curious as to if you dont want two sticks moving parrell in the same direction, if you add in beveled gears sure they will move in same directions but i think you fix the parrell issue maybe.
biblically accurate gearbox
this is the best comment of all time
banger
Gearcube
top teir joke
Ahh yes very correct
this seriously looks like one of those satisfying looping animations
If a pair of sticks had a diagonal connecting bar on only one side of the hub, making a sort of U shape, the other pair parallel with it could have a connecting bar on the other side of the hub. It would still be possible to pull the sticks out in one direction, but there would be a “safe” direction where it collided with the hub, and there would effectively be half as many stick components to get lined up in sync.
I was going to suggest temporary connectors like this. Then remove them once it's together.
I was about to make both of these suggestions, but someone beat me to it!
that is interesting, it would remove some of the symmetries of the system (that is, it would halve the amount of them), and visualizing it in my head, i cant decide if i find it more aesthetically pleasing or not
@@terdragontra8900 The symmetries of the physical object will always be imperfect. But what we should be contemplating is an idealized version, where the racks are replicated out to infinity, and there's an infinite 3D grid of cores. Or how about multiple interleaved grids?
I was going to say that too
What comes to mind for easier assembly is:
1. Make a jig which can hold all the racks and the core in the right positions, but doesn't touch the gears.
2. Assemble the core puzzle.
3. Put all the parts in the jig except for the gears.
4. Slide the gears onto their axles.
That way, all the racks can be put into their central position without any constraints, rather than needing to start in the “just falling out” position all at once.
This is a great idea and honestly I would watch a video of this process.
You could have temporary gears without teeth, but the correct contact diameter and an extra flange, to work as the jig. You could then push the racks in place one by one and once aligned, then start changing the temporary gears into actual ones one pair at a time.
The way the sticks mutually hold each other in place immediately makes me think of the Borromean Rings.
I can't even comprehend the sticks holding each other around the core, forget about the movement! so cool!
5:08
Different Idea: you can have the driving rack have a waffle iron pattern and the driven racks can both be driven.
Think of the driven racks as tools that carve tracks out of the driving rack, if you carve both patterns out of the driving rack it will only have little parallelogram nubs that don't interface with the whole sliding surface, but it will be able to go both ways
Yeah I was thinking the same. Like those 3d printed nut/screw pairs that can translate in either direction for a given rotational direction
Exactly. Just like these 3D-printed counter-threaded screws, where the direction of rotation is defined by how the nut is threaded.
@@philbunting5963yep, same thought here
This idea is used in drum winding mechanisms, with the addition that the pattern "merges" at the ends, so that a follower automatically turns the other way.
Engendering students really need this as part of their core continuum.
Here's what worked for me! First, mark the center of each rack on the back side. Then, starting with a bare hub (no pinions), arrange four racks and put a rubber band around each end of each set to hold them to the hub. Then again with another set of 4 and rubber bands. Get them all approximately centered; they don't have to be in exactly the right position or rotation, just close. Then choose a pinion that will contact three of the racks and insert it (with the bolt), careful to center each of the 3 racks it contacts as you do. Repeat for the opposing pinion, then the other two that contact the racks placed so far. Repeat one more time with the remaining four racks and two pinions. I'm not saying it's easy, but it was relatively straightforward.
EDIT: Also, thanks for the cool design, and informative video! I'm sure it will be a great conversation piece!
7:15 "But I do want to talk about the core on its own, so I guess I have to…"
I see you have not yet discovered the magic of making two of them!
I love your videos. Great videography, wonderful explanations of how things work. It's both eye and brain candy.
5:34 Can't you have diagonal racks on both sides of a stick to achieve direct contact opposite motion?
I thought that too
It's a bit hard to explain, but there's another possible configuration: the bottom rack could have a grid of diamond shapes to act like overlapping diagonals going in opposite directions, then each of the top racks could have diagonals that going opposite directions like the last demonstration he showed.
@@minecraftiangenetics Yes I had the same thought. It's functionally the same (or at least very similar to) as the bolt and nut that were machined to have both left and right handed threads that was doing the rounds on RUclips midway through last year.
lovely - i reached for my star burr puzzle and pulled it apart as you did in your video - so thanks for naming that wooden puzzle of mine that i guess me dad bought decades ago - but also - i now have to remember how to put the star burr puzzle back together again - x
You could connect the axial pairs by using concentric rings, just set the rings on the inside or outside of the pair and have mounting tabs, this would allow for a full range of motion. Stemming from this, the concentric rings could also have a guide pin that sits on the outer face of the rings with a pole that sits on the curved surface of the core, where the axle for the gears sits, this would help for assembly. For assembly, the various constraints are what make this easy, the system is constrained to where all centerpoints align, and the gears just slot into place, the guide pins are a temporary constraint whereas the gears are a permanent constraint.
For clarification on the concentric rings: Say the square of rods is a base circle. One pair of rods would get a larger concentric circle, the other pair gets a smaller concentric circle, both with mounting tabs for the pair of rods they connect to. For example, axis positive would get a larger circle, axis negative would get a smaller circle; though, personally, I think each pair should get both sizes of rings on the two ends, which could be further used as an endstop against the core. The guide pin would have a flange that seats on both circles at the same time, thus constraining them to the same centerpoint position; and in combination with the guide pin's rod seating on the core, this constrains the axial pair's centerpoints to the core's centerpoint.
Let's look at one set of rods, say the X-axis. The core constrains the rods in the Y and Z axes, allowing free movement within the X-axis. The concentric rings will constrain the positive and negative pairs of the four rods, which this later becomes redundant with the installation of the gears. The guide pin will constrain the X-axis position of both of these pairs; at least as a soft constraint with one pin, and a hard constraint with two pins, but really only one is needed. Repeat this for the Y and Z axial pairs. The installation of the gears would now constrain the rods to the core, the two opposed axial pairs together, and an axial rod to its perpendicular neighbor, thus constraining everything together. The original product already had this final constraint, but with a more difficult assembly, but mimicking these constraints during assembly makes the process easier, and as a coincidental byproduct the system that binds two rods together is also usable for setting up the assembly constraints.
Also the rings could face in oposite directions and not concentric, that would look nice!
5:22 On the lower single stick you could change it to a diamond pattern where it is both diagonals cut out of it. You could also add a connector bar to each end of only one pair of sticks per side and it would also stop the sticks from falling out while allowing the ends to move past each other (add the bar to both ends of the pair moving one direction along the x axis but not the other parallel pair of sticks that moves the opposite direction and same for y and z axis).
Wow, this is absolutely fascinating! I can't think of any practical use cases in my head right now (though there are probably a good amount), but I'm thinking some implementation into a crazy CNC based system for experimental manufacturing!
Hey use giant pendulums as a protection from earthquakes. I suppose it could work in a similar way but in smaller scale.
Or instead of a gearbox add a gyroscope and you could dynamically change directions of spacecrafts or small satellites.
I feel like I'm watching secrets of the universe while watching this channel
Great as always! To connect 2 rods you could probably add a brace that does not connect straight across, but in an arch that skips over the other rods
I find all of these so mechanically pleasing.
To assemble,
line up all the racks on the core with toothless gears/wheels so everything slides easy,
tape racks to hold the shape,
and swap out the wheels for gears 1 by 1 meshing teeth as you go.
One of the best RUclips videos I've seen in a while!
To stop the racks from falling out of the core you could make little screw on caps for the ends of the racks that would be too big to get pulled out, similar to how the gears are mounted but with a flange rather than a gear.
Amazing!!!
Your videos on their own are enough of a reason to buy a 3D printer
MC Escher: **heavy breathing**
what happens when you do the design on 5:40 and you extend/overlap the pattern? they always keep following there straight what so it is possible to have more motion right?
Great! Like to see the rods circular!
at 5:12, I wonder what would happen if you didn't make all the racks have grooves in only one direction, but formed a kind of a diamond-shaped, bidirectional grooves (a diagonal criss-cross pattern). Something similar to those bidirectional bolt threads: ruclips.net/video/v96LTfmtDPU/видео.html
My first thoughts exactly. Just make use a Boolean intersection operations between the two mirrored angles. But the problem I see with that is that ALL the bars would have to have that design, so I'm not sure if the direction could be maintained between two interfacing diamond patterns. Maybe make half of them diamond? Loses some symmetry, but maybe a working configuration is possible.
Or make ithe rods double sided as someone else suggested
My guess is that it might be likely to jam. There are 24 places where two sticks move against each other. If any pair of touching sticks can decide to go whichever direction they want, there could be conflicts. I don't know though, maybe it would work out!
@@henryseg I also thought of the same idea. I think OP meant that only the bottom rack would have the diamonds. The two racks on top can remain unchanged.
@@cliveso Interesting... but in the whole configuration, I think that every rack would need to interact with two other racks that move in opposite directions. Maybe that's still not a problem though? I'm not sure if there's a way to split the racks up into "diamond" and "non-diamond" so that diamond racks only interact with non-diamond racks.
I am in love with this 😍
There is a way of having this be a rack-rack interaction: Just have one of the racks have a pattern of a diamond lattice that you get when you do a union of the two diagonal rack patterns.
was thinking the same, surely he thought of it and it didn't work for some reason, I'd like to see it done
is it possible to design the green parts as a circle so the movement is infinite?
If you add a size increase at the end of the sticks so they won't fit through the gears it will block them from moving through in that direction. If you add that as a stick-on to the other end as well, it will be forever locked in place.
oh yeah, this hits all the right spots of my ocd. simply amazing.
Love your work Henry! 👍👍👍
Interesting if you place a Cartesian coordinate system at the center of the gear mechanism you'll have a mechanism of constant mass centroid location but variable off axis moment of inertia, product of inertia
oh, hey, it's me!
personally, the star burr puzzle is my favorite. as at first glance it seems impossible to put together or take apart because each piece holds each other like a borromean rings. but once you know the solution, it's a trivial dexterity toy.
You could get all the sticks in the right place easily by having cylinders that are the same as the cogs but without the teeth on the hub, that would restrain the racks' movement to the correct axis without coupling their movement to one another. You could then go around and replace each of these cylinders with a cog to add all that coupling back in. You could also use this as a basis of a way to stop the racks falling out by skipping the first and last notch for the gear teeth to go into on each rack, that way they would slide in and out easily with the smooth cylinders, but once you have the cogs in place they would be blocked from turning any further when they get to the end of each rack. You could even use that as the way to get them lined up correctly when putting the cogs on by having the cog start off meshed with the first or last remaining notch on the racks it meshes with.
Great work, thanks for the video
About making a rack push two other racks in opposite directions: Could you make worm drive racks such that the linear motion pushes in one direction while the screw spinning motion pushes in the opposite direction?
Then the two racks you want to push...i guess one would be sliding and the other rotating in a threaded hole, to go oppositely?
I suppose that wouldn't 'count' either.
For version that do not have cogs, if you make bar with only row of pegs they could drive two orthogonal bars that are set in opposite directions. Round pegs will not block move of orthogonal bars as they are exactly same when mirrored.
lovely design and would it not be easy to build clips that grab onto the pairs of racks and while they are bottomed out in the core attache the gears and bolt it down
5:35 There is a hand powered screw driver called the 'Yankee Screwdriver. The core has two sets of paths for one rod, somehow it can be pumped to make a continuous rotation. I've also seen a bidirectional nut, that has two paths cut into it. Would this concept be used to solve this particular problem?
Hello! 5:30
If you cut the two-way slat where it meets the transition of the teeth on the other slat, you get a two-way "X" toothing that runs its entire length.
Best regards:
A.P.
Hungary🙂
Excellent! With the angled groove bars could that be made in a way where the top surface had //////// grooves and the bottom had \\\\\\\\ grooves?
Sure, but then could all the sticks be shifted and angled in such a way that they all mesh in the right way? I'm not sure.
Super awesome! You have a beautiful mind. Speaking of moment 6:57, you could connect bars a and d on the right end, and connect bars b and c on the left end.
5:37 instead of having 3 plates all designed with slats, have 2 slatted plates (opposing directions of the slats) interface with a diamond studded plate? That way the slats themselves determine the direction that they move?
To put them together I would take out the gears, put all horizontal stix in the core (stack themfrom lowest to highest), then stick the vertical stix in a fixture or a lump of putty on the table. That way I could (hopefully) line up all the stix in equal llength and have one hand free to reinstall the gears, locking everything together.
0:00 new favorite animated wallpaper
Jesus christ Henry, you've done it again. It's recursive racks.
I wonder if you could make the racks largish circles (ideally either small enough to still be printed or else arcs that can be connected together) with the core riding on the circle what that would look like, could you connect multiple cores per circle, what about loops of core-circle-core-circle-core-circle-core-circle in say an octahedral arrangement (because of the even number of edges around a vertex)?
5:18 could you instead of having solid bars on the lower rack, have diamonds instead that allow both directions of movement?
It is beautiful, as always ✨. I wondered if there was a way to move two different ways using the principle of the two-way screws ? With the driving stick having a kind of grid pattern ?
What about using sliding racks over a diamond grid (so both angle racks intersected) allowing for movement in both directions
On 3:36 you claim that a six axis rack has the tetrahedral symmetry. A tetrahedron has 4 corners, 4 sides and 6 edges (all at different angles). Your six axis rack has six "ends", all rods are parallel to one of 3 orthogonal axis. So your six axis rack has the same symmetry as a borromean rack or a cube (6 faces, 8 corners, 12 edges parallel to 3 orthogonal axis), not a tetrahedron.
Great! This sculpture should be easier to copy in LEGO or fischertechnik than those where the racks mesh directly.
If you want 2 racks to interact with a single one moving in opposite directions you could cut the groves all the way in both directions on the stationary rack leaving diamonds behind (similar to the people who have cut both left and right handed threads on a screw).
Instead of having bars going across pairs of sticks, you could just put attachable nubs onto the ends of the sticks (e.g. bolt on) that are bigger than the holes between the gears, which would prevent the sticks from going all the way through.
No need for gears. It's enough if each one from the six axis interacts with two others, not three. You can make a closed chain of the six sticks that way. Now, get a cube, set it on one corner, and look from above, so it will look like a hexagon. Arrange that chain of six sticks onto the sides of this hexagon, alternating between "upward" sides of the cube, and the "downward" sides of the cube. Done, you "just" have to design the way to keep them together. On each corner of the hexagon you have a join between two axes. It will look somewhat similar to the three axis solution, but the sticks that lay on the same axis will be moving in opposite direction to each other.
In 5:36 would it work if you had this alternating pattern on all 3 racks, instead of just on the bottom one. So when one of the top racks is about to stop, then it switches to the right pattern and can continue the sliding?
As far as how to synchronize them... I see that they've got a 1/2 tooth offset on the two sides of each stick. So if you had a worm gear with 2 teeth on it (double-helix) you could insert those one by one, thread them in/out until they're all the same length, probably via some sort of mark at the center of each worm, then temporarily attach the sticks to the end of each, push them all through in sync, then remove the worms from the sticks.
What applications could this possibly have?
at 5:30 why not go with a double-threaded approach?
5:22 How about overlaying both patterns and substrackt?
There are 3D models of screws that can be screwed both ways I think…
You can do a hatch pattern on the bottom plate and it should let you move the two top plates on opposite directions.
instead of sticks would it be possible to make this with rings so you they can just keep going an every direction? i think it would look really cool although that might put it past what's feasibility printable size-wise
If you make the racks two sided, could you have one side have the pattern required for the reverse motion without the gears?
I wonder if it would be possible to turn the sticks into hoops, so instead of going back and forth over the length of a terminated stick, the gears of the core would keep rotating in one direction while the hoops "rolled"
5:31 - what if the bottom rack is "threaded" both ways (i guess it would look like a series of rhombuses instead of teeth) so that when it moves, the two top racks move opposite directions?
There was a video going around a while back where a machinist made a screw that was threaded twice, so that a left hand and right hand nut would move on it, that's what makes me think of this.
What if you used cross hatching on some of the racks? Would that allow for the same motion in rack to rack configuration?
How does making the sticks extra long, assemble then trim to length sound?
2-way hinged crossbar/caps across the pairs of sticks that move in the same direction parallel to each other would allow them to cross through another pair of sticks so movement isn't limited, but perhaps the bulk of the axis mechanism could block the hinged crossbars from allowing us to easily deconstruct the sculpture? I was thinking of how nice the dual-directional swinging doors of a laundromat allow 2 directional flow or with lack of a central bar create one large opening for larger loads...I'm not a mechanically minded person so I may be relying too much on the central mechanism to somehow block the "doors" from allowing the sticks to be pushed all the way out, but feels close.
What's wrong with the racks moving in the same direction? Do the ones along the third axis jam?
For the non-geared version can you use interrupted teeth?
i have no idea what these could be used for, but they are still cool.
Hmm, couldn’t you replace gears with a part that inverts motion? It’s annoying, but possible, right?
Since the core holds together without the nuts, wouldn't it be possible to get the sticks lined up more easily by installing the gears *after* the sticks, rather than before?
Yes, but the sticks slide around all over the place without the gears...
@henryseg 9-5 Could you make some kind of rack version of Moebius strip like Moebius racks somehow?? Thanks for sharing Henry and really hope you can answer this and my other question or email whenever you can. I would really appreciate it. Thanks very much.
I think the key in lining them up would be in adding the gears on Afterward.
Would love to see this skinned in a silicone blob shape. I think it's movements might confuse people. Ever considered going into animatronic territory?
I cannot wait for the seven axis racks
Does anyone else have the sudden urge to eat crinkle fries.
slotted racks with the correct spacing and phasing should allow you to get the same results without gears similar to a bidirectional bolt
I'm sure you thought of this, but for racks what if the two different angles cut through each other?
What if you somehow combined this with tensegrity structures and magnets?
Also, staggered herringbone gears are the best gears.
you can have racks that go both way (mesh in two directions) without the gears. Similar to a nut and bolt that can screw and unscrew in both directions.
What would be the use case?
Does it need one, other than tickling the brain?
Would you be able to use circles instead of sticks for the racks?
Very large circles presumably, although weight would start to make it unwieldy.
Can you post the end cap stl for the tetrahedral racks please?
Babe wake up, a new Henry Segerman video just dropped.
The video starts at 1:33. Go back to the beginning for explanation later.
Mesmerizing
Can you provide a link for the files to make the Star Burr puzzle?
cool! I would like to learn too. what software do you use to simulate the mechanics?
I use Rhino/grasshopper. I’m not simulating things so much as animating them - it doesn’t really have a physics engine in that sense.
@@henryseg Thank you very much for your reply. It is useful.
Could the core have a zig-zagging groove pattern, and the interfacing sticks only mesh with either the zigs or the zags, going in opposite directions to each other? I don't have quite the engineering mind to imagine this, but if someone could explain or experiment, I would love to hear if it works or why it doesn't
6:52- put the joint on the other side, or in one it's on the edges and the second is on the middle, if you have access to it. actually the middle feels like a bad idea...
1:58 would make good endless gif material
I've got that exact shot as a loop over on my instagram!
The two aligned "sticks" on each axis should be connected with a bar. The two opposing sticks should be connected with a RING around the periphery. This will link the 2 "sticks" without them colliding. YOU'RE WELCOME!
That’s a pretty good idea. THANK YOU!
With eight of those chores and more sticks, you could create a cube center. Something like a 3D #.
So this is the gearbox that every borderlands game starts by mentioning
Not 100% sure what your main goal is here, but im curious as to if you dont want two sticks moving parrell in the same direction, if you add in beveled gears sure they will move in same directions but i think you fix the parrell issue maybe.
AAAAAAH this is so satisfying it's like geometrical crack!
So, next step a lattice made up of six axis racks. And then put it into a universe with "asteroids" bounding conditions for infinitely moving racks.