>>37672 >I still believe using this concept is objectively correct for us to pursue. Excellent! I do too, Kiwi. IMO, we're simply not likely to successfully manage all the tradeoffs needed for effective & pleasing robowaifus without incorporating design elements relying on tensegrity principles (whether they be 'canon', or no) into her frame & support systems. I'm personally excited by a specific thread on this very-cool topic. Thanks, OP! :^) https://en.wikipedia.org/wiki/Tensegrity https://www.youtube.com/watch?v=jnpshtyvWr0

>>37673 >related This straws+strings (jumbo drinking straws+ultra-high molecular weight fishing line, actually) approach is right up my alley, concerning building ultra-lightweight, very low-cost robowaifu skellingtons. Jumbo drinking straws are pretty much near the limits of physics in providing strong, lightweight (and cheap!) support structures in space.

>>37674 >ultra-high molecular weight fishing line BTW, slightly off-topic (but only slightly): this stuff can be woven into fabric. * Combined with good sewn textile patterns+vacuum forms+a little hardening resin (interior/concave sides only -- for smooth exteriors), this presents us all with interesting possibilities concerning our robowaifu's shells (very lightweight, resilient to impacts/resistent to punctures, reasonably inexpensive, easily-repaired). Properly-designed, these shells can then also directly provide 'support frames' for underlying internal tensegrity structures (perhaps with the addition of some simple interior 3D-printed hoops/'ribbing'). Lots of potential benefits, AFAICT. :^) --- * One example vendor http://uhmwpefiber.com/1-4-industrial-textile-fabric-fiber/

>>37675 Uhmwpe is a wonderful material but it creeps(lengthen over time with mininal stress) thus ruins tension over time. Collapse is easier when the critical length is not constrained properly. It is attractive to not create ball and socket joints that requires geometric accuracy and lubrication, and subsequent shielding away from liquids and dust but a ball joint is still dead simple and would hold such accuracy much better under compression load. Walking would be hard if joints are wobbling. But the idea that actuator/cable double as extra width against bending stuck with me. I sketched up a capstan actuated full motion joint with 2d extention and contraction as well as axial rotation hoping to creat a superior design over individual sets of wire and linear actuators. My afterthoughts later reminds me that individual sets of cable and actuator is going to be more resiliant and less likely to fail completely than one continuous cable running on capstan being the single point of failure. The number of motors is still 4, at least 3, and since I haven't ran the math, there may not be an attachment point solution that keeps the wire length constant. Without that exact solution, some range of motion would result in popping out the joint or snapping the wire unless I make one of the capstan ride on suspension, adds wire tensioner, or pneumatric cousion inside the joint as a hack job.

>>37676 Artbyrobot, is that you? :D >Uhmwpe is a wonderful material but it creeps(lengthen over time with mininal stress) thus ruins tension over time. Collapse is easier when the critical length is not constrained properly. Fair point. But I don't see much else that could easily fit into the same 'readily-accessible, relatively-inexpensive, lightweight, strong, durable, super-conformable, highly-adaptable, tensioning system' category apart from perhaps extremely-fine (even dangerously-so) stainless steel wire. And even that will sag with time & usage. Ideas? Carbon wires are exorbitantly costly and will be reserved only for the most topflight (read: super expensive!) robowaifus produced several years down the line from now. >It is attractive to not create ball and socket joints that requires geometric accuracy and lubrication, and subsequent shielding away from liquids and dust but a ball joint is still dead simple and would hold such accuracy much better under compression load. Beautiful. My own pottering about with this years ago turned into a 'cylinder & spindle' design, tensioned with big rubber bands, situated orthogonally up along the (also tensegrity-esque) 'spinal' column elements. You might think of them as lightweight, swivel-able 'ribs'. This approach allowed for very flexible (resiliently-bendy) internal support structures, for otherwise completely-inflexible components (eg, the so-called 'breadbox' electronics housing, et al) within the robowaifu's torso-volume. >My afterthoughts later reminds me that individual sets of cable and actuator is going to be more resilient and less likely to fail completely than one continuous cable running on capstan being the single point of failure. An important design consideration!

>>37676 >>37677 Slackening tension of support cables is a problem going back for many centuries (at least!) in the civil engineering & nautical worlds. The aviation industry solved this over 100 years ago with lightweight, inline (but manually-operated) turnbuckles, as a regular maintenance item. [1] Can these be a) designed to be very smol, and b) automated to re-tension periodically on their own (and without much additional cost+mass)? --- 1. [2] https://en.wikipedia.org/wiki/Turnbuckle 2. Also confer lockwire used to tie-down the turnbuckle after re-tensioning (a safety-critical item): https://en.wikipedia.org/wiki/Safety_wire

>>37678 >Slackening tension of support cables is a problem going back for many centuries (at least!) in the civil engineering & nautical worlds. Now that I think of this, surely that 'Old Italian Dude', the inimitable Leonardo di ser Piero da Vinci must have a solution for this. Anyone know if he had a re-tensioner system for his Mechanical Knight Bro? (Surely a precursor to Robowaifus if ever was. :^)

Various resources from the blog: http://jirolfing.blog.fc2.com/?date=201501 https://www.youtube.com/watch?v=k6FaULbOmnE https://www.youtube.com/watch?v=uMug6XzP1R4

>via: http://kennethsnelson.net/tensegrity/ Also, numerous animations (helpful to visualize the structures involved): http://kennethsnelson.net/animations/

>>37677 Hi chobitsu I m sorry that I don't know who that is. I am the newone posting uduki tsukasa on the other alog site and haven't used a user name yet. Some offers zero creep polymer ropes but tbh i don't trust their claim. Tension is the necessary evil for all thing except masonry buildings. Maintenance on exposed soft sides is what bond building is all about. Ideally the actuator self adapts to lengthened wires, and if not the robot has its own hand to tighten it and docking station should make it easy. Steel is probably the best wire. Piano tuning is seasonal only and probably due to temperature change. But don't get body parts caught in one and tempered carbon steel isn't the best against oxidation. (Just boil rust away) The cylinder and spindle idea cpunds interesting. Can the cylinder be blow molded like soda bottles? As for carbon, i think the best way to utilize them is making few layers graphene oxide by mixing graphite in milk whey protein in blender or magnetic stirbar for many hours, dry and adds that to any plastic/resin of at like .5% weight. Or be used as ink to print soft pcb https://www.markoneill.com/ACS-AMI-2020.pdf

>>37682 >Hi chobitsu I m sorry that I don't know who that is. Just a Nihongo pronunciation of "Chobits", the mango/animu that created the board's mascot Chii (cf. >>495 ), who inspired me personally regarding the possibilities of robowaifus. That's all. <---> >and haven't used a user name yet. We tend to namefag here b/c as a (mostly-)consistent group of engineers, artists, etc., working towards common goal(s), we can 'cut to the chase' (get to the main point) easier by getting to know one another as a community. >tl;dr Imagine working at a big engineering firm together where you yourself, and every.other.anon. in the building all have 50 First Dates [1] syndrome. It would be pretty hard trying to get anything practical done together in such a scenario, for a highly-complex project (such as conceptualizing/researching/designing/prototyping/testing/manufacturing/distributing effective & appealing opensource robowaifus [& kits] -- all from scratch). And stylometry only goes so far! Therefore by consensus our board culture uses namefagging, typically. OTOH -- we are Anonymous; through & through. So yeah, it's a bit of a weird (and fun!) mix having many of us with a moniker, yet still absolutely maintaining the OG Anon mindset toward one another. Don't think I know of many other boards like us in that regard (and a few other regards, as well). >ttl;dr Use a name, or don't. Perfectly fine either way, Anon. <---> >and if not the robot has its own hand to tighten it and docking station should make it easy. Interesting idea. OTOH, structural lines for tensegrity tension are going to (at least conceptually) spread all throughout her internal skellington/support-structures. Hundreds of them, I'd imagine. It'd certainly be hard for a human to do maintenance on such a collection! :D >Steel is probably the best wire. It is better than fishing line in many ways, that's certain. But cost & mass are two of the highest priorities in the long litany of concerns we all have to balance here on /robowaifu/ . If we can find a set of reliable sources for extremely-fine, high tensile-strength stainless steel wire, then that might be a real consideration. Otherwise the mass of hundreds of such steel wires throws the "equation" right out balance. UHMW fishing line is currently much, much better for us in these (availability/cost/mass) [2] regards, IMO. >Can the cylinder be blow molded like soda bottles? Great question! A few of us here constantly keep manufacturability concerns in mind at all times. And personally, I'm always on the lookout for """alternative sourcing""" (read: waste recycling) materials ideas, so that men around the world can have some chance at building their own robowaifus out of whatever's at hand. >carbon/graphene >flexi-PCBs >information linked-to, for everyone's benefit Thanks! Nice ideas like that are always welcome here, Anon. Cheers. :^) --- 1. https://www.imdb.com/title/tt0343660/

>>37678 > Can these be a) designed to be very smol, and b) automated to re-tension periodically on their own (and without much additional cost+mass)? I had an idea on this topic: Since my designs are using jumbo drinking straws (thin plastic tubes) for the floating tensegrity 'rods' (cf. idea-related : >>37674 ), and the fishing lines travel down inside the straws...why not insert smol, motor-driven line tensioners inside the straw tube as well? Not only could they be used as (re)-tensioning devices, but they could actually assist with initially assembling the tensegrity structures (within jigs) as well. The proverbial "Two birds, one stone." <---> They'd need to be: a) very compact (fit down snug inside straws/tubes) b) lightweight (in total, at least; probably have only a few strategically-placed around the tensegrity structures -- drawing everything tight together) c) inexpensive (perhaps 3D-printed frames around cheap, tiny motors) d) effective (the motors would turn some kind of friction feed [possibly a tiny triple-rubber-wheel arrangement, each offset 120' around the central 'feed zone' core]) e) lockable, when not in use (perhaps drawing the 3-wheel feeder head back up into a 'compression funnel' to squeeze & lock the ensconced line in-place). Ideas?

>>37683 I'd stay Anon for now, maybe for another 2 weeks or when it got too confusing. Distilled ideas and works can and do live without authorship. Ultimately both steel and UHMWPE are used to anchor things in the sea and some plant fiber are resistant to salt too. I just think less tweeking is better. https://www.mcmaster.com/product/9667K25 Looks like they are selling 3 feet of pianowire in a pack of 6 for 10 dollar. I don't think it would break the bank. As for weight, tensile elements are usually not heavy when compression element has to add additional width to hold bending stress to prevent buckling. The cable is half of the weight to the tower and not counting the concrete floor slabs. https://www.goldengate.org/bridge/history-research/statistics-data/design-construction-stats/ >>896 Straws will be weak in the other two directions. Id thicken it like water pipes in profile. Specific strength are good when only the self weight is concerned like running and jumping. But in case of user falling over, stepping on or a bookself falling all over the robot, foreign bodies weight justifies absolute strength to save her from total rebuild. Besides, air weight stuffs are afraid of air blowing. A 70kg companion can easily have 20 to 30kg dedicated to structural strength while 1 steel bike is like 10 kg. This isn't a nasa mission where a large tower of rocket fuel are spent to send a tiny peak to the orbit and repeat to send a smaller fraction to mars. In civilized world there are power sockets everywhere to power the battery and in rural area plenty of hydrocarbon to burn. One more thing is that uhmwpe are one of the best at sliding, used as joint replacement, and many glue cannot stick to it. I think gluing requires resin and solvent like in ballistic plates. I don't think friction drive would work on slipery rope but either a pulley, wench or capstan with many loops would work.

>>37685 Thanks Anon! I'll look into that information. I appreciate your valuable input here. Cheers. :^)

A /comfy/ Fishing Bro Anon made an interesting post about fishing lines: https://trashchan.xyz/comfy/thread/13171.html#13313

>>37676 Please share more illustrations and post updates as you build your joint mechanism. >>37679 >Da Vinci's Knight Excellent example of a design that could be adapted to tensegrity principles. Floating shell panels using variable length cords to move about is the future coming from the past. In this case, the ropes creep wasn't factored. Da Vinci likely just used thick enough rope that the creep wasn't an issue for the machines expected life. >>37682 >Zero creep rope Not physically possible. You are wise for realizing their lie. Creep can be incredibly small given the right material is used for expected tension. Notably, many elastronomers (such as silicone and rubber), tend to have favorable creep given their cost.. >>37683 >Steel is heavy It also has a catastrophic failure mode that would be devastating for her master hugging her. I've been whipped by a steel cable that snapped suddenly. >>37684 >Dynamic line tensioners embedded in her tubes Definitely potential in this concept. Not too far off from the research presented here; https://www.frontiersin.org/journals/robotics-and-ai/articles/10.3389/frobt.2019.00111/full >>37685 >70kg My personal goal is 10kg or less for MaidCom. Every gram adds cost to building, maintaining, and powering her. Tensegrity as a concept is primarily used because it can provide heaps of durability relative to mass compared other construction methods. >Capstan A great mechanism for adjusting tension and/or actuating her parts.

>>37689 >It [steel cabling/wire] also has a catastrophic failure mode that would be devastating for her master hugging her. I've been whipped by a steel cable that snapped suddenly. Great point!! Safety is certainly a very-high priority concern for us here. Thanks for pointing that one out. > https://www.frontiersin.org/journals/robotics-and-ai/articles/10.3389/frobt.2019.00111/full Wow! That paper is loaded with great information, thanks Kiwi. Here's a copy, just so it doesn't get memory-hole'd on us. > >Tensegrity as a concept is primarily used because it can provide heaps of durability relative to mass compared other construction methods. This. Its literally the whole point behind this research! (Well, that and the fact it also offers us at least some improved potential for highly-lifelike motions/deformations/etc. Important points for hugging & dancing with our waifus! :D Cheers, Anon. :^)

>>37691 >>37687 I doubt I was actually helpful. Good luck working things out.

Do we even have any materials that have better tensile strength compare to flexural and compressive strength? Besides steel rope suspension bridges I can't really think of any common uses for this building method. But on the notion of elastic elements, perhaps nitinol might be a good idea. Nitinol doesn't have as much tensile strength, but is more elastic.

>>37689 I did the joint calculation. Good news is during axial rotation, from neutral ( V ) to maximum rotation ( [X] ), assuming zero radius capstan and zero radius ball socket joint, and assume the span and height is 1 to 1, the length change is only 1.5%, less if span is less than height. And one can easily get 180 degree of travel axially like how we turn our hands/arms/head. An intuition on how the geometry works is that the trace of 2 fixed point sharing a wire is an elipse. This will be all the possible position of the capstan in 2d. Rotate around the long axis gives elipsoid which is the surface which the opposing capstan can be in 3d. The top of the elipsoid is flat enough that a circular path is not far from the true surface. >bad news However planar movement like leaning forward/backward/left and right as a hinge would slack the tension. Turning 90 degree make make 20% of the cable length into slack. Not good for joints but gives a good insight that one can place the wire by bending the joint first, wrap it in and then straighten it out. >new design Hinge movement is needed. To modify the existing architecture for that task, the assumption that the center ball joints are too small to collide with wires are taken back. Tendons in real animal joints are locked in place by other tendons than to freely slide past and loosen on the contracted side. Similarly an enlarged ball joint with wire guides can be placed to lock wire in place. More math is needed to calculate the right geometry given the segments are more complex and the axial rotation is negativelly affacted the bigger the ball is. >wire whipped Ouch!

>>37691 No! It was really very helpful, Fishing-Anon. I hope you'll interject with more good advice here on /robowaifu/ whenever you'd like to (though you should visit over on our main board at Alogs then). Thanks & cheers, Anon. :^) >>37693 That's a good question Anon. Honestly, I haven't had much time to delve into Materials Science sufficiently to be able to answer these types of questions very well. But Fishing-Anon seemed to know something about this, though (cf. https://trashchan.xyz/comfy/thread/13171.html#13371, et al). Maybe Kiwi or our other legit engineers here can provide more insights as well? Cheers. :^)

>>37694 >based hross limb-analyzer Biomimicry is king! :D

>>37695 Will a version of this thread be made on the main board? >But Fishing-Anon seemed to know something about this, though (cf. https://trashchan.xyz/comfy/thread/13171.html#13371, et al). Sorry as this is only barely tangentialy relatted to tensgrity. As to that particular post, Curv and Tegris are fused thermoplastics made of layers of woven polypropylene fabric. These products are commonly used as stiffeners in duty belts, plate carrier cumberbunds, etc.. Picrel is a Curv holster adapter very much so capable of surviving a person falling on it. There is a vidio from the maker of a guy doing pull ups on one without ripping the anchoring screws out of the material. Anyway, I am no engineer, but I believe it would make an ideal semi-regid anchor bed for mounting just about anything. It would give to absorb impact and flex to aid in body softness. I believe it would probably also work very well in thin bands for reinforcing/support a flexible shell exterior. A bit of elasticity could be added by using shock cord anchors on the bands (look up the cumberbund adjustment for a Crye JPC 2.0 for an idea). The video below is of a simple Curv belt. There is no reason to watch all of it, but it gives an idea of how the material will flex. These styles of belts remain very stiff and resist sagging downward as weight is added to them. https://youtube.com/watch?v=frMwLYlL6A4[Embed]

>>37700 >Will a version of this thread be made on the main board? Yes! It will be this very thread in fact (once it's been merged over from here by Robi & myself Kiwi did it because he's impatient :^) ). Other threads that fit properly will simply be merged into already-standing threads there on Alogs ( /meta , R&D, etc.) A few like this one will remain as new threads there. Once we return & are fully re-merged over there, we'll put this (our bunker) board into a relative 'deep freeze' (a few threads here will still be unlocked, however: like backups, filepile, & comfy bunker). This board will then be unlisted from the overboard & webring (rn we have both listed on the WR, lol). Make sense, Anon? update: The move has now begun! (cf. https://trashchan.xyz/robowaifu/thread/55.html#970 ) <---> Excellent post, Anon. This sounds like a great materials resource that could be used in many ways for the structural & internal support aspect of a good robowaifu (whether she's a tensegrity-based design or no). Thanks! I trust your instincts are well-qualified. BTW, most of us are amateurs (or at least relatively-so, considering the vast swath of disciplines involved here! :D So please make yourself at home, Anon. Cheers. :^)

>>37690 Thanks for saving the PDF! >>37693 >Materials with greater tensile than compressive strength Stainless steel buckles under lower pressure than it'll undergo plastic elongation. >Common uses Though "true" tensegrity is rarely used outside of bridges, stadiums, and other large architectural works, the core of the concept is commonly used. Tents, umbrellas, folding chairs, etc... use a mix of compressive and tensile loading to min-max durability per material use. >Nitinol elastic elements Clever idea, makes fantastic springs. It is expensive so, using it in series with another material may be needed depending on scale and budget. >>37694 Looks like you're making a mechanism functionally similar to a stewart platform. I've been contemplating similar concepts. >Tendons in animals All animals use tensegrity in their musculoskeletal systems. Studying how tendons and muscles interact with skeletons provides great examples to learn from. I appreciate you bringing up how faschia and tendons can overlap to limit motion. Frankly, I overlooked that other animals also have mechanisms we can learn from. Thanks for broadening my perspective. >>37695 >Kiwi can provide more insights I'm passionate about tensegrity and want to help others. >>37700 Interesting material, similar to cloth in a resin binder. Depending on geometry, it could easily work for either compressive or tensile elements. >>37702 Any estimates for when this thread will be moved? Turns out, now

>>37704 >Thanks for saving the PDF! Y/w! Remember 1984 bro... >I'm passionate about tensegrity and want to help others. Same. I'm just too ignorant about how to assemble these things ATM though -- seems complicated! How the heck do you build something like that with just two hands!? :D >Any estimates for when this thread will be moved? As I related (https://trashchan.xyz/robowaifu/thread/55.html#970 ) the plan is about a week from now. a) I need to do a modest rewrite on BUMP, but circumstances are stalling that rn. b) Robi will need to complete his BUMP-to-Lynxchan importer scripting first as well. Since these are both indeterminate-timeframe imperatives, it's tough to pin down rn. Hopefully my estimate will come off... we'll see! :^) <---> That being said, there's nothing to prevent you -- Mr. Moderator -- from getting started early with a new Tensegrity bread there today! You don't say? :^) (Thanks for the reminder.) -Kiwi Also, please be giving some thought now about how you think we should merge all these threads here, together on Alogs after the 'move' of these trash/robowaifu/ threads over there. Some (like this one) are clearly new breads. However, most of these Bunkerfest threads definitely should be merged back into aleady-standing threads there on Alogs. Please remember we have a limited 'thread budget' (and we've never liked having dupe threads about on /robowaifu/ , regardless). I'll like to leave that task fully in your hands+Greentext anon's to decide, if possible Kiwi. Cheers. :^)

>>37693 >>37704 I forgot to mention, I know nitinol can change shape with heat, and elastic like rubber bands can emit heat when stretched, so I was thinking that nitinol and elastic could be used together to boost the strength somehow.

>>37716 That's really an interesting idea, Anon. Sort of a 'smart material' type of behavior at that point, I'd say. Cheers. :^)

>>37702 >>37704 >>37705 Heh. :DD <---> You did a great job migrating this thread, Kiwi. Crosslinking, etc., all look good. Nice to have this cool new bread back home here! :^) Forward!!

Another link from a Doctor studying the skeleton and how it's actually a Tensegrity-Truss. https://web.archive.org/web/20151007045949/http://www.biotensegrity.com/tensegrity_truss.php

>>37735 Excellent research material, Grommet. Thank you, Anon. :^) <---> Tensegrity is how we are going to win this race as amateurs, against all the FAGMAN (((groups))) and even against the baste Chinese!! :D Remember the Wright Brothers!!