/robowaifu/ - DIY Robot Wives

Advancing robotics to a point where anime catgrill meidos in tiny miniskirts are a reality.

Site was down because of hosting-related issues. Figuring out why it happened now.

Build Back Better

Sorry for the delays in the BBB plan. An update will be issued in the thread soon in late August. -r

Max message length: 6144

Drag files to upload or
click here to select them

Maximum 5 files / Maximum size: 20.00 MB

More

(used to delete files and postings)


“I am not judged by the number of times I fail, but by the number of times I succeed: and the number of times I succeed is in direct proportion to the number of times I fail and keep trying.” -t. Tom Hopkins


Prototypes and Failures #2 Robowaifu Technician 01/17/2023 (Tue) 02:10:46 No.18800
Post your prototypes and failures. We fail until we win. From now on with even more madness. Don't forget looking through the old thread >>418 to understand how we got here. >=== -minor subj edit
Edited last time by Chobitsu on 01/17/2023 (Tue) 03:01:26.
>>18790 >with the downside that these parts might not come with the license you like to have ("created by Thingiverse user SorrowBJD, and is licensed under cc") No, the CC licenses are fine Anon. They're compatible with commercialization, unlike, say, GPL. BTW, your progress is an encouragement to us all. Keep going!
>>18770 I'll be checking with MeshLab moving forward. >>18780 Dolls are a myrical that I regret overlooking for so long. >>18782 Going for 1DOF shoulders? I was planning on implementing a cable system so her steppers actuate everything from the floor up. If you're looking for a new chest, could you explain what you'd change about MaidComs chest design? >>18800 Nice get dub dubs
>>18800 This is actually kind of amazing looking over these collages OP. It's been an interesting ride so far!
Open file (100.37 KB 648x715 FaceWIP.png)
Working on a 3D model for MaidCom's head. Advice would be appreciated. Files: .stl https://files.catbox.moe/kw3kcq.stl .3mf https://files.catbox.moe/p8078z.3mf .amf https://files.catbox.moe/idu5vg.amf
>>18810 Nice AYYLMAO character appeal typical of mango & animu. >Advice would be appreciated. I'd suggest you consider a) separating the jaw from the rest of the head so she can move it around while talking, etc. b) make cutouts suitable for driving things like eyelids & eyebrows. (cf. dear Sophie >>15236; also >>18686) c) Where are the cat ears?!!.:^) >=== -add Elfdroid Sophie crosslink
Edited last time by Chobitsu on 01/17/2023 (Tue) 04:59:55.
Open file (86.50 KB 479x547 nya.png)
>>18811 Added cat ears and eye cutouts I hope it'll be cute when printed
Open file (294.93 KB 1920x1080 just_howitlooks.jpg)
Btw, does anyone want to get some of my files as STL? Not sure if they're final, though. I could upload some of the lower part of the body. That said, I didn't print them myself in full body size, so that should make clear that I wouldn't recommend doing so. >>18806 >I'll be checking with MeshLab moving forward. Hmm, okay. I will have to check my parts as well. I'm getting errors sometimes at export. It's probably dependent on the program if it warns or not. Then it can be ignored if the slicer doesn't care, but importing it into another program and using it there, is where it fails. Just FYI, one part from therobotstudio's arm also failed. >Going for 1DOF shoulders? Certainly not, this was just a demonstration. > I was planning on implementing a cable system so her steppers actuate Similar. I was thinking of using small servos in the shoulder areas with strings, to move the shoulder around. But haven't decided how yet. That said, in a later design if you want more human-likeness, I think the shoulders in your upper chest design would need to be able to move. Or we add the moving parts outside. >a cable system so her steppers actuate everything from the floor up. You mean from the belly area or deeper down? >If you're looking for a new chest, could you explain what you'd change about MaidComs chest design? I was talking about my upper body design, chest and belly area. I think I would rather use one that has been sculpted and is being shared instead of sinking to much time into trying to model something alike in OpenSCAD. Yours is kinda limited to the shoulder area so far. I didn't get to (re)read your MaidCom thread completely yet, so I might have knowledge gaps here. I think you wanted to keep her lightweight for mobility, but the "Joystick-waifu" isn't mobile. I wonder what will be there in the belly. Also, I was thinking about the outer shell, the skin so to speak. I mean, boobies and smoothness, for example. Generally, I want my approach to have different parts to switch out. So I'm not saying your part can't be used. I'm not sure if I will end up in my actual build with it, though. >>>18800 >Nice get dub dubs Don't know what this means, but thanks. Sadly I forgot to include the first few designs from the old thread, but it's in the folder now. So next time it's in there. I used just PhotoCollage on Linux.
Open file (162.77 KB 1550x870 elf_elfenlied.jpg)
>>18810 Yeah, cool. But if you don't have much time, focus on the things we don't have and can't get from somewhere else, at least as a placeholder. It's really time that we stitch something together that can move and looks like some waifu. >>18811 >a) separating the jaw from the rest of the head so she can move it around while talking, etc. You're conflating skull and face a bit here, I think. We need parts for creating a mold or printing it in a TPU or other flexible material, and a then also a complex skull which moves inside. Going for a relatively soft head which can't move much is going to be faster and easier. We can switch things out later. There's a reason why I didn't even really get started trying to build the kind of skull I want. It's going to take veeery long, and have many iterations. >b) make cutouts suitable for driving things like eyelids & eyebrows >cat ears Simple things could be added in OpenSCAD and then shown or hidden by changing one value. If it's in a sculpted STL model it needs to be cut out or filled if modification is necessary later. Also, aren't cat ears supposed to be very soft? You're close to building the elfes from Elfenlied, their "ears" are part of the skull. (Spoiler: They're not just cute).
>>18825 >Btw, does anyone want to get some of my files as STL? Yes.
>>17086 >Paradox Drive What happened to it? Does it work? >>18810 > AMF and 3MF exports Btw, dependent on your program and the preferences there, AMF and 3MF files might include the full name to your files and this might be a privacy problem, dependent on your username on your PC and if it is supposed to be a secret. >>18832 Okay, I'll take care of that soon. That aside I thought 3mf would be better to share such models, till I remembered that problem described above. Though, I could turn it off before doing my exports for posting and also it won't be a big deal if it gets leaked. You know the older files are mostly here?: https://mega.nz/folder/nv42EbzT#1MUblavva2UG6DukgD0moA
>>18811 >separating the jaw from the rest of the head so she can move it around while talking, etc. Fixed it (picrels). Better?
>>18814 Heh, niacu. It'll be cute enough before all is said and done, Kiwi. :^) >>18826 >You're conflating skull and face a bit here, I think. Fair enough. My assumption is that ATM, Kiwi is going for a single assembly that's intially hard-shell. I may be mistaken ofc. >>18837 Lol. Now you're talking Anon! :^)
Open file (349.13 KB 1626x1855 MaidHead.jpg)
>>18825 How many DOF do you intend for her arms to have? >I think the shoulders in your upper chest design would need to be able to move I look forward to the day we can afford such mechanical complexity. To provider greater context, I think very similarly to you. I have faith we both are working towards a full replica of the human form. >You mean from the belly area or deeper down? As far down as possible to help with balance and minimize inertial loading on her actuators. > I think I would rather use one that has been sculpted and is being shared instead of sinking to much time into trying to model something I spent years searching, then years learning once I realized that what I wanted did not exist. I want to help you so that you do not have to go through the same. You clearly have great potential. To be clear, I want our projects to be completely compatible and develop standards. For instance, I overlook breast development and having someone to actively develop such features with would be beneficial. >Dubs 4chan slang for a post having repeating digits; double digits.8800 being a case with two sets of dubs, thus dub dubs. >Focus on other things Her head is at a state where I'm fine with it. Picrel is the final head unless someone else volunteers to improve her. >Cat ears Never a permanent fixture. Definitely a planned add-on. >Paradox Drive It functions as a proof of concept. I stopped development once I realized that 3D printed gearboxes are inherently inefficient. Pulleys, belts, and cable based mechanisms can be printed and used efficiently. >>18839 >Kiwi is going for a single assembly that's intially hard-shell. Correct currently, MaidCom is developed as a series of intercompatible parts with hard shells to make her cute.
Open file (1.02 MB 1920x2560 MaidHead2.jpg)
>>18847 Increased eye size 140%
>>18847 >Correct currently, MaidCom is developed as a series of intercompatible parts with hard shells to make her cute. Ahh, thanks for clarification Kiwi. >>18849 That's kind of neat comparing the two. Mango uses varying iris sizes all the time, but your pics made that more noticeable somehow. I also have a vague sense that it affects our perceptions of neoteny, somehow (head & eye ports are the same size obvs, but bigger irises seem younger somehow).
>>18847 Okay, cool. Looks certainly nice. My current waifu face test certainly needs some work. I got lead into a rabbit hole by the whole snout or face depth thing, but it's getting better. The eyeballs would even have to be way bigger in my design, compared to what they currently are. A big problem the current version has is most likely that I didn't make the nose separate. I try to work on some kind of face generator with this, but I also considered to just pay someone on Fiverr some money to make a sculpted model for my own personal waifu. I want to have gynoids I share with the public, but also my personal one which won't act in certain ways on social media. I imported some different faces I tried to work with in the past into my testground file (picrels), just for testing reasons. One is from Project Eve. The other one I found as "head-female", I think. They are female but not necessarily pretty enough anyways, and the Sophie head seems to be low resolution. > For instance, I overlook breast development and having someone to actively develop such features with would be beneficial. Yeah, I can look into that a bit more. Not sure if we really need different breast to choose from at this point, though. Then again, I think it's fairly easy to create some boobie framework file in OpenSCAD. I was planning to look into the neck next >>18862. Aside from finally finishing the joystick waifu body parts and print it. I also need to buy some screws and the elastic strings for moving on with that build, and I have to do something else around the house before I can. >reposted this above, since the filename of picrel 3 was wrong. Two more applied for the role as our first synthetic girlfriends, faces shown in picrel 4. Picrel 5 is the full body of the Anime_style one of those: - Anime_style_1_3_BJD_55cm_Doll_2996049 (left from viewers perspective) - Alternate_Polaris_BJD__4279007 (right from viewers perspective)) The Anime_style one is 700ish MB big and the models are so high resolution, that I had to simplify the head to even be able to import it into OpenSCAD.
>>18847 >>18868 I downloaded an eye from here, only number one for now: https://www.thingiverse.com/thing:4397291/files (male doll, picrel2, but doesn't matter, imo). These eyes have a internal pattern in the iris, which will still be visible if the cavity is filled with colored resin. There are videos on YouTube how to work with these. Again, we could someone here replicate something like this model, but why? Anyone who wants to do this, fell free to do so. Till then we need to work with what we have. I cut a little hole in my downloaded file, for the camera. Just one more test if this works, it did, it rendered fine. For putting in the resin, the hole would need to be closed temporarily of course. No idea if this is going to work or if we would need a glass cylinder in the middle to avoid resin there. One more thing to find out. The back side should maybe also get a camera holder. Anyways, here's the file with the hole, also I made it smaller (3MB instead 10, but 3mf only): https://files.catbox.moe/h2y8cb.3mf
Open file (898.58 KB 1920x2560 CyberEyes.jpg)
>>18868 Here's the latest file for her head. https://files.catbox.moe/oux1jk.stl >Breasts One of the last things to work on honestly. >>18869 I'm going to use follow me eyes (little piece of paper behind glass hemisphere) https://www.youtube.com/watch?v=3BwAM_V2Jhg Moving physical eyes have a charm but, passive eyes that use optics to appear as if they are looking at you are easier and faster. Picrel is MaidCom's head with paper eyes I made. To be clear, I do think your idea for eyes is worthwhile, just would take too long for me as I'm in a rush.
>>18870 Okay, looks nice, thanks. I'll look into the video as well. I have an idea for an water proof eye mechanism but not sure how I will prioritize it after looking into the other approach. Either way, the regular animatronics approach is not acceptable to me. After finding all these doll faces and seeing your design I'm also not going to work on my OpenSCAD based design anytime soon. I'll rather try to figure out other things. Going through Thingiverse was totally worth it, and it's even not the only platform for this anymore. Too bad that none of the lurkers here was looking through it before and posting it into the doll, meta or general engineering thread.
>>18870 Looking good, Kiwi! I assume the glass refracts the printed image, giving some illusion of depth and angular displacement?
>>18871 I started on a neck mechanism, like indicated here >>18862. It's just an early concept and I got the top wrong, since I did it out of memory. Though I have to think about it, it might either not matter or be better anyways, especially with printability in mind. The head would sit on top of this, the pole moving it is still missing, assuming the motor wouldn't be in the head (it shouldn't). >- The head falls automatically forward if the neck string is loose enough. This will be done by how the weight of the head is balanced. >- Can be moved by outer force in all directions at any time. Because springs and elastic strings. >- Can tilt forward fast through some weak but fast motor in the front, and because the string holding it back up would be elastic and just strong enough to hold it in place without any additional force involved. On a string below, or bit higher doesn't matter. It will snap back automatically and fast as soon as the front motor stops pulling, because of that neck string which is elastic or there's a spring involved. The sideway part isn't in the first draft yet. I was still contemplating how to make exactly sure to make it more able to tilt sideways while facing to the front, but not while rotated. >- Sideways movement only by overcoming some relatively weak springs holding the head in the middle, so it will snap back automatically. What is in the sketch already is, that the top where the head will be connected is oval so the head can't spin around. One of the somewhat tricky parts is the same as with my joystick waifu body: Having a pole in the middle going to the top and then also be able to tilt and elongate itself while doing so. This is a mechanism a have to work on. Strictly speaking the neck is also something we do not really need to model and recreate ourselves, if we use one of the other mechanisms here: >>18844 - That said, the ones which would work would probably require to use the motors for which they where made. Maybe SophieDev's design could be used with strings, I guess. Dara's neck design also looks like something that would work fine, at least with the motors for which is was intended. So, I'll look at this as an option and will work on it as I'm motivated, but maybe not give it the highest priority. Then again it's rather something very simple so far and I guess it's almost finished, and so I might feel like go on with it soon. We'll see.
>>18887 Nice work Anon, keep going! Also, you might examine the torso-pivot joint for the project in link-related (>>18890). His use of a simple design, plus accompanying spring-loaded tensioning might be useful.
>>18887 >Neck Went on working a bit on it, but I'll switch to the telescope tilting mechanism now. Without that, neither of my "joystick" mechanisms is going to work. >>18892 > might examine the torso-pivot joint for the project in link-related (>>18890). His use of a simple design, plus accompanying spring-loaded tensioning might be useful. Looked at it, but so far I'm not into it. I might contemplate on it a bit more, though. Thanks.
>>18907 That looks nice & compact, Anon. Question: does the topmost element spin around it's own axis?
>>18880 Somehow works without glass. Works better with glass though. Figuring out how to mount the glass right now. >>18907 >Telescope tilting mechanism Please elaborate?
>>18912 >>Telescope tilting mechanism >Please elaborate? If something can tilt and rotate the pole for rotation might need to have the ability to get longer while doing so. If I'm not mistaken.
>>18919 >>Telescope tilting mechanism Piston tilter might be the better way to name it. Though, the tilting part is still missing, and so are the second horizontal bolts of course. The point of it is, that it can elongate. I believe this to be necessary.
>>18923 Ahh, that's a little clearer to understand. So, there would be another couple of components needed to complete the assembly Anon?
>>18989 >there would be another couple of components needed to complete the assembly Anon? Which one? The whole simple desktop waifu, the neck or only the piston tilter? I'm working on the whole thing from time to time. Every day a bit, sometimes more, sometimes less. Currently trying to make the neck design printable. Also, I realized the piston tilter should rather be in another tube to make it smaller. Also, I had to work on a actual tilting mechanism. I should at least print the piston tilter today to test it, probably both designs.
>>19004 I wonder myself where the time goes, but working on the details of a part is often less rewarding than for example working on parts which look like a female body, and so I'm less motivated to even get started. Thinking about and working out how parts need to be cut to be printable and later can be screwed or glued together, for example. This plays hugely into how the code has been written, maybe reorganizing it and label section better, and so on. Aside from that, I always have a list of other things I should be doing at some point and other also things to which I'm drawn instead. Anyways, I'm still on it. I think much of my work just need some polish before I can go on with printing and testing it. My trick with listening to podcasts also works wonders, after I get started with it.
>>19008 This https://files.catbox.moe/d77e5r.zip is the 3mf version of the files from >>18299, I think. They should come with some support indicators for Prusa Slicer. The part in black will certainly need some rework, also the whole thing is just an incomplete model.
>>19004 I was referring to the piston tilter in that case. >>19008 >I wonder myself where the time goes Ahh, the life of a tinkerer! Try to enjoy the ride along the way Anon. Going back to the basics is always a good idea, and reminding yourself why you're doing all this can help you see the forest again despite all the trees currently in front of you. Keep moving forward Anon!
>>19008 I was just printing and assembling a small model of my neck design. Just for testing it, though there was no doubt that it would work at this point. The springs on the sides are still missing, and many screwholes were too small. It's rather just tinkered together right now, but was just a test anyways. Also, it broke when I used force. I could reinforce some parts but it was smaller than it is intended to be, so it might not break when printed bigger.
>>19078 >so it might not break when printed bigger. Welp, one way to find out, right Anon?
>>19004 My last design was most likely dumb. Unnecessarily over-complicated. I redesigned it, and also cleaned up my code. This time powered by Nightcore waifu music and some dystopian podcast. The new design will just use a screw instead of this whole sliding shell abomination. Still have to work out the best way to print and assembly it. Then printing and testing. >>19180 Yeah, no problem. It both can be done easily. Making it stronger and bigger. Scaling it up is even much easier. The screw holes were to small anyways. It was just printed very quick with a few shells and with moist filament, but this wasn't relevant anyways. I'm quite sure it will also need to be a bit higher to have enough space for that piston tilter pole to have space to work. Reminder, anyone else with a 3D printer could try out one of the other neck mechanisms and show that it works with strings or different servos: >>18887 >Strictly speaking the neck is also something we do not really need to model and recreate ourselves, if we use one of the other mechanisms here: >>18844
>>19191 Those look very clean anon.
Open file (6.54 KB 462x636 pole_example.png)
Open file (24.40 KB 474x474 angle_joint.jpg)
>>19191 I might need another small angle joint or something on top of the whole thing, to get the movement translated as I want it. I also had the idea that at least for the neck, some kind of chain mechanism, like the spine in some print-in-place mechanisms (picrel 3), would be better for the whole thing anyways.
>>19212 >some kind of chain mechanism, like the spine in some print-in-place mechanisms (picrel 3), would be better for the whole thing anyways. Neat idea. One of the cool things about resin printing (AFAIUI) is that you can print complex mechanisms like this in place (ie, already joined together, etc.) Someone here please correct me if I'm mistaken about it.
>>19214 You don't need a resin printer for that, dependent on the exact design, though it might work better. But they are more messy to use and I don't have one. Print in place works with FDM printers as well, if the design is made for it and the printer is tuned very well.
>>19214 https://www.thingiverse.com/thing:4642307 >Printer brand: MK3 >Printer: Prusa i3
>>19216 Hmm. Looks kinda promising. I might at least combine these two. The piston for elongation if necessary, the snake design for a better way of transmission of the rotation. Idk. Will have to play around with it. >>19214 I could make a chain myself with a slightly modified version of those angle joints (tilters) I made myself, as part of that piston tilter design. Then I could use metal screws as bolts or as reinforcement, which would require assembly but also make it way stronger. Using this design would also in general work better for modifications to make it stronger.
>>19216 Thank you Anon. >>19221 That's really cool looking. Have you considered the idea of running a stiff braided wire down the central core longitudinal axis of that 'chained' element. For improved ruggedness I mean.
>>19253 >stiff braided wire No, but I was considering replacing the whole thing with a metal spring that would be strong enough to transmit the rotary movement. Anyways, I'm test printing parts right now, and I'll go back working on others soon.
>>19264 Looking forward to it Anon.
>>19221 I want to help you but, your project still feels nebulous. What is her height? What are her DOF for her body, neck, and arms? What is your desired payload? What will be her primary purpose? So far, I just know she's a motorized joystick that may be similar in height as a human without legs. You've provided some information but, nothing hard. Please provide a hard spec sheet and more drawings of what you're working towards. I can work with the crudest of drawings as long as dimensions are known.
Open file (414.88 KB 800x450 directions_(800p).png)
>>19290 >I want to help you but, What would you do if you knew these things? I can take advice based on different ideas and try to make my design compatible with it. >your project still feels nebulous. Because it is. I'm flexible and make things up as I go. As good as possible. >What is her height? All my parts are supposed to be seen part of a framework which can be used by as many designs and builds as possible. Having that said, the Joystick design can orient itself on a 130-150 cm waifu. This is what I mostly had in mind. But I also thought of her sometimes as a desktop waifu on the kitchen table, so her body size would probably the size of a smaller human. Obviously, she won't have legs. That said, since she also won't be sex-enabled this is one more reason why she also could be smaller. Also, I want my design to cover as much ground as possible. Okay.. looking into this now ... I guess, if she was like a 150cm body than her body without legs would be around 60cm without neck and head. The desktop version of a 40cm upper body plus part of the thighs (without neck and head) would relate to a 120cm full body. They could both fit onto a desk, but the taller one would be quite high up there. >What are her DOF for her body You mean the main part? Forward, backward, sideways tilting, rotate like a human. I thought this was clear from my design. I simpler version could cut out all the tilting and just go for rotation of the upper body, or tilt forward a bit, for some work. >neck as human-like as possible >arms? as human-like as possible >What is your desired payload? What? I have no idea. Of the arms? There can't be a plan for that. This is meant as a platform to test stuff, I will not plan everything ahead. If I would think like this I would never have gotten started. >What will be her primary purpose? - Having some kind of simple waifu design anyone here can actually build and improve. - Separating the development and testing of anything above the waist from the rest, so it can be part of different prototypes. - Improving anything above the waist in a waifu that I actually have and is equipped with AI. Instead of working another five or more years on everything from time to time till I can assemble the whole body. - Having motivation to work on the AI, by having a waifu to interact in daily life. - Maybe starting teaching her working with her hands as soon as she has good enough arms. Like stirring a pot or cutting some ingredients while posted on the kitchen desk. Maybe playing a xylophone or so. - Making videos like Masiro Project and Kibo-Chan. >So far, I just know she's a motorized joystick that may be similar in height as a human without legs. As short term goal, yes. The parts being added and tested over time are of course meant for a later design with full body, but those who don't want or need that, or can't afford it, can stick with the simpler version. >Please provide a hard spec sheet Not sure what this exactly is and most likely I have to say no. This is simply not how I think and work. I also don't see how this would help. >and more drawings of what you're working towards. This here shows it quite well: >>18433 - I might make some kind of collage soon. Maybe I think of reasonable drawings. >I can work with the crudest of drawings as long as dimensions are known. Is this about the program you use? By help you mean designing parts, not general advice? I scaled your other design, and my designs are supposed to work at least in certain ranges. Well, if you need something I might just pick something, but what if I later run into problems and have to change it? Are the values above good enough for a start? Equivalent to a 120-150cm body size...
>>19299 I made a crude preview of the MaidCom head with other parts and the current simple neck mechanism, without a flexible neck casing added. The value how these parts relate in this picrel right now, are: - WC shoulders 0.8 - WC head: 0.9 - simple neck: 0.75 I sized it down for the comparison with the other heads. Of course we could take the MaidCom head as reference and then scale everything based on that. I think I had my problems with the size of the chest. I'm really looking into keeping the shoulders not too wide. The purple part is still the Nema 17 without any added gears, which fits in, but I don't know how well it would work for the arms. It's just an example.
>>19300 I briefly sketched out an idea of how to make floating ribs, which could be held by strings (tensegrity). The holes for that are still missing. It's just to show of the basic idea. Maybe some other way of attachment would be better, since ribs are not supposed to move. But the current chest might require a ribcage below it and fake ribs attached on top of it as well.
These are looking really interesting Anon, keep it up!
>>19221 This here failed, btw. I'm doing some remodeling of the neck tilter, then print it and test it with a long spring, or some combination of the piston-tilter parts, maybe with a long spring. I think the one I have is strong enough to transmit rotary movement.
Open file (44.91 KB 640x480 maze.jpg)
>>19299 >What are you going to do? Come up with viable solutions and help you implement them. >It is nebulous Which is a problem that will impede you the same way it impeded me. >Height 150cm scale, just no legs. That's perfect, I can use this. >As humanlike as possible Can work with this. >I just want an open source Kibo-Chan Heh, understandable, she's really cute. >>19300 Is there a reason for her chest being backwards? Her chest is small on purpose. It's the minimum viable size for a 150cm waifu. Based on measurements of lanky anime girls like Aigis from Persona 3. Though embarrassing, my original design was supposed to work with a NEMA17 stepper but, forgot to include the shaft. As is, her chest would need to be scaled up for a NEMA17 based mechanism to be incorporated fully internally. Externally is also a problem as she starts to have broad shoulders. >>19320 >ribs y tho >>19330 Just use the mechanism in picrel.
>>19336 >Is there a reason for her chest being backwards? Confusion. >Her chest is small on purpose. It's the minimum viable size >...as she starts to have broad shoulders. I wondered which of your parts is the most important reference for size? I scaled the chest down, because too me it was looking to big. Especially around the shoulders it are already to big, since the shoulders are rather in the arm, as you can see in any picrel. >need to be scaled up for a NEMA17 based mechanism It still fits, without additional gears. Well, dependent on other things that have to fit in there. >ribs >Just use the mechanism in picrel. I see. You really want to focus highly on functionality. Ribs were meant for making her body more rigid and feeling more human. Even looking more human if the rib bulges are visible through the skin. I have at least two approaches to ribs right now, ignoring the failed ones. I'll keep a variant of this one here as a option in my mind. I was generally wondering how to build a light body based on tensegrity and came up with these layers of TPU floating and guiding the strings. Though as I wrote above, for the ribs it might not be the right way, I would rather add some rods there. In regards to you mechanism, this looks more like a alternative to a spine and my tilting mechanism.
>>19337 I separated the options into different files. MaidCom is now the one with the orange chest. I used 100% your sizes, kept the servo for now for comparison, and the same arm.
>>19338 I had the idea of importing the heads into Prusa Slicer and there I can see the size and percentage compared to original. Then I can scale the heads more reasonable and compare it to other parts, while also scaling them if necessary. The MaidCom chest size looks way more reasonable with a 16cm head then the smaller one before. Without scaling the MaidCom head I downloaded would be 12.5cm in height (Z), which is to small for the chest.
>>19339 My own opinion is that it's perfectly fine to overscale the head somewhat. This is pretty typical for females in mango/animu (our basic character design guiding-light). We are definitely not going for absolute-realism here.
>>19346 I just compared the head above to an old water bowl for a cat, since I didn't find any better reference. Well, I guess 16cm feels kind of illegal. I think for I one meter waifu it might be fitting, but a 1.50m one should at least have 18cm head height. Without taking the overscaled anime heads into account. 18cm is the diameter of these cork coasters I have. I guess head sizes don't change that much compared to height. I generally think more and more, that part designs should have holes that can be changed dependent on the scaling of the part. Alternatively it needs to be drilled out a bit in case if it's scaled down. Basically, we can't easily guarantee the size of the screw holes. 3D parts need to be designed in way that allows some scaling later. Btw, I really like that near perspective of >>19339 every time I look at it.
>>19359 >Btw, I really like that near perspective of >>19339 every time I look at it. Yes, it's a good camera setup. Keep going Anon, you're doing great! :^)
>>19361 Thanks. Yeah whenever I work with the parts that look human-like I feel the same, but in reality I'm not doing that much currently. Have some AFK stuff to attend as well. That said, some nice preview from time to time should help to keep the morale up for everyone.
>>19368 >some nice preview from time to time should help to keep the morale up for everyone. Yup, true.
Open file (293.47 KB 419x1070 AigisTan.png)
Open file (613.05 KB 1920x2560 Cute.jpg)
>>19337 >Most important size They're all the proper size with no scaling needed unless you want to. Admittedly, they were designed for a cute meter tall robot but, now, they're being used for a 1.5 meter robot with everything else being stretched. It should be similar to Aigis in terms of form factor. (picrel) >Steppers fit She can have up to 3 standard NEMA17 steppers sideways from my real world testing. 4 work if they're all pointing straight up. Which could work for certain arm designs. >Focus on functionality Correct, though I was being facetious rather than curt when I said "y tho". Fully support your use of ribs and I honestly believe tensegrity is the correct >>19339 My designs look better IRL. Her head was designed for a meter tall waifu but, I'm sure it'll be fine at 1.5 meters. Her feeling robotic helps to make up for it being so small. For me, anyways, everyone can use the head they want. >>19346 This guy gets it >>19359 >More holes I agree, there should be more holes. I'll release the final files with and without holes so users can customize.
>>19382 Thanks for the clarification. Your picrel is certainly a more advanced preview than mine. >They're all the proper size Hmm, idk what I did wrong then. I used scaleval=1; which is 100% or just normal. Whatever, not important, I can change it if I need to. Just wanted to report back about my experiences.
>>19330 Still working on it. Just put it inhere as an update. My photos are quite bad, though. Used several lights, and that's the best I get. Still need some connector to test it with my waifu model. Then I still have to figure out the thing from the drawing in >>18430. Not really sure yet.
>>19544 I somewhat tested this above and think my idea is going to work. I really need to get started on servos now. I could need some opinions on shoulders, like of those of Eve from Halodi (picrel 1). I don't plan to make them look very similar, but I need to get an idea about the shape of that part around shoulders. I also think her chest is too thick (picrel 2), aside from the back curve not being cut out and these ugly folds on the side and back. I think at some point I want these shoulders to move, while being shelled into some TPU. But this might require bigger parts of the chest being printed in TPU or with moveable parts. I might be going down another rabbit hole too early here. On the other hand, early discussions and showing of sketches won't harm. The tilter (picrel 4) is also ready, assuming this idea is going to work. But I'm more or less inclined to just print in in full size soon and go with it. Probably one other model size test, during this week.
>>19544 >My photos are quite bad, though. Just at first glance, it appears you simply need to clean the lens? These are coming along Anon. I'm curious if you mean for the screws to be part of the final construction? >>19547 >On the other hand, early discussions and showing of sketches won't harm. Definitely not, that's always helpful. I'd suggest you plan on making the shoulder designs well bigger than you think you'll need to, then back that down once you have your servos & the shoulder articulation joints worked out. Keep up the good work Anon!
>>19548 >simply need to clean the lens? Yeah, the thought crossed my mind but I just didn't bother enough. That and adjusting the light worked to somewhat improve it. So, now you can see and enjoy the awful print quality way better. > I'm curious if you mean for the screws to be part of the final construction? No, these were just ones which fitted. This is just a small model and these are wood screws without nuts.
>>19550 >So, now you can see and enjoy the awful print quality way better. Lol. These are just prototypes Anon, don't worry about it. Indeed don't waste time overmuch on things you know will simply be discarded as the overall design and capabilities improve. It's more efficient if you don't, particularly for time. Once you begin putting these under load, you'll understand then where they need to be beefed up more, to deal with stress better. >=== -minor sp edit
Edited last time by Chobitsu on 02/07/2023 (Tue) 05:04:36.
Open file (203.26 KB 1920x800 mpv-shot0001.jpg)
>>19548 >suggest you plan on making the shoulder designs well bigger than you think you'll need to, then back that down once you have your servos I tend to work the other way around. Thinking of looks first, then trying to make it functional, finding a compromise. Looks and texture are the primary functionality, unlike in any other field of robotics. >>19552 >where they need to be beefed up This is why I'm inclined to integrate some standard metal parts like screws and iron sheets here and there.
Open file (731.89 KB 2371x3533 51324125213.jpg)
highly suggest using foam for mockups before the real thing, can save a lot of time and trouble. I would´ve wasted hundreds on materials had I not done a foam model to see my measurements were wrong >>19232
>>19557 Thanks. Not a bad idea. Maybe some guys want to see some kind of shape outside of a 3D model, before printing and molding her parts.
I'm stuck on the how to go about making the inverter that outputs over 1000 volts. I tried watching some videos but they didn't click.
>>19585 Good to see that someone is working on that. > how to go about making the inverter Ideally, try to find out and share it here. This board isn't very good at answering general expert questions beyond adapting technology to our use case. Can you describe your problem better? There are quite many videos on YouTube about HV inverters: https://www.youtube.com/watch?v=iHkKTGYc6aA Why baby oil, btw? related: >>13966
>>19587 I might go with this one seems like the most straightforward. https://youtu.be/0dtYJJxVh_E Some sci anon told me HASEL muscles are made that way. I looked up what they're made of and it said some silicon compound but I think that's meant for the shell.
>>19588 I suggest looking into different tutorials, so you find the best way and avoid any problems. Godspeed.
>>19556 >I tend to work the other way around. Thinking of looks first, then trying to make it functional, finding a compromise. That's good. I'm glad there are several of us here with differing views on the right approaches to use. This will lead to good synergies I predict, as long as we all stay open-minded and focused on the goals we're all striving for Anon. >>19557 >highly suggest using foam for mockups before the real thing, can save a lot of time and trouble. Excellent point Anon. I've done numerous skellington mockups using cheap resources (and then torn them back down again) and it's a great way to prototype rapidly and inexpensively. >>19593 >I suggest looking into different tutorials, so you find the best way and avoid any problems. Godspeed. Agreed. I hope we can all together find literally several different means by which to accomplish actuation & locomotion. Once we have our first robowaifu models working, then I predict it will become more & more clear which are the most productive routes in the end. But try them all out as we can! >=== -minor sp, prose edit
Edited last time by Chobitsu on 02/07/2023 (Tue) 23:24:22.
Open file (37.73 KB 474x428 HASEL.jpg)
>>19588 Silicone is not required. HASEL actuators function via charge attraction creating hydraulic pressure. Two conductors with opposite charges will attract each other via electrostatic dynamics. By having a dielectric (insulating) fluid between the plates, the fluid will be moved as the plates press together. This fluid pressure can be used to apply pressure on the container, causing it to exert a force as a portion becomes rounder. The container must also be a dielectric (insulator) so that charge is not shorted. Your plastic bag with foil and oil would work. High voltages are required to achieve the electrostatic charge pressure. Most electronics cannot work with these high voltages. Be extremely careful. It also must be DC as AC will cancel out charges and produce erratic behaviour. I cannot in good faith suggest you chase this actuator unless you're comfortable with extensive trial and error. This is one of the most difficult actuators to use, and you will need to dedicate substantial time and money in order to get anything useful.
>>19593 thank you >>19599 Thank you for the comfirmation. It might require extensive trial and error. But it seems like the perfect solution for a robot waifu since its soft. Why would it require money though? The materials seem cheap.
Open file (1.06 MB 2560x1920 GearTest.jpg)
>>19547 >Halodi Decent machine to reverse engineer as a learning exercise. Sorry for my lack of help, I've been busy trying to get practical results IRL before moving forward. Picrel has gears which should be perfectly meshed together and generally does. They have spots where meshing doesn't happen due to the tolerances of 3D prints. The timing belt idea would work perfectly if the teeth were longer to provide for higher allowable tolerances. I prefer pulley's and strings but, need a way to achieve concentric continuous rotation for a mechanism that should enable waifus to become far more accessible.
>>19600 >Cost You'll need to try several times with different materials and high voltages aren't easy or cheap to work with. High voltage will destroy most microcontrollers and most electronics in general. Unfortunately, There's not enough good information on these actuators. This should help. https://onlinelibrary.wiley.com/doi/full/10.1002/advs.201900178
>>19603 Hmm, okay. Interesting, thanks. What I wonder the whole time about: Would it make sense to print these parts with filament that contains metal or paint these parts with something like it, and then galvanize them? >>19607 >High voltage will destroy most microcontrollers That's obvious on one hand, but how? It should be controlled by the controller not going through it.
>>19607 Thank you. I just saw this here. 80,000 volts. That anon just told me over 1000 volts and then ran off without saying anything else. 80,000 volts is quiet a bit more than just over 1000 volts...
>>19612 Missed the pic
>>19608 >Filament gears They are loud, inefficient, and easily wear down/heat up. This is why I'm experimenting with belts, as they are almost as good as metal gears in terms of efficiency. They are silent due to being made of materials that dampen vibrations. I'm currently considering using the belts normally with printed pulley's. This would be a middle ground in terms of efficiency and should still be quiet. >Electroformed gears This is time-consuming and though would help somewhat, they are still inefficient. It's a good thought, electroforming has many benefits and deserves further investigation. >>19612 >Voltage The higher the better for the most part. 8K to 20K seems to be the best from what I've seen. Feel free to test though, just post your findings. >>19613 Glad to see you found out about opto isolation. This is essential for the system to function correctly by isolating the actuators from the controller.
>>19078 Bought some small springs today and found some more lying around. Working on the the spring holders, before I print a bigger model. The orange part is just a cylinder how the spring could look like. I better make a list to remind myself what I need to work on: - Still working out ways the print the neck better and add some screws and nuts. Also, some ways to connect the strings to move the thing. These are rather small tasks. - Upper body tilting tests, probably only with the belly part in the model first. - Upper body improvements. Maybe later, since I cut it into belly and chest parts. - Need to work on the servos - Later, I need to do another part of tilting mechanism on the top of the neck design to let the head tilt without the whole neck bending. It's close, since most of these things are optional before I start printing the first full size parts, but it always feels that way until I find more problems or something gets in the way. After that, head connector, making the head design more slender and print it (molding comes later), eyes, shoulders, arms, ...
>>19673 Improved the back of the upper body a little bit.
>>19616 >Glad to see you found out about opto isolation. This is essential for the system to function correctly by isolating the actuators from the controller. I hope we can find really tiny & inexpensive 'buses' of them Kiwi. They would be great for many uses inside a robowaifu, and would assist in hardening her from external insults like voltage or EM. >>19673 >>19674 Nice work Anon. Yes that back is an improvement IMO.
>>19603 That's quite a nice 'double-power' design Anon. Looking forward to your belt-toothed version! :^)
>>19616 So I'm going to give it a try. The bug zapper seems like a good hack but it'd have to be for each muscle. So I'd need a bug zapper for each muscle... How is this going to work out?
>>19683 Please be careful Anon. HiV capacitance can definitely set you back on your ass.
>>19683 Okay is my shopping list Is there anything I'm missing? https://drive.google.com/file/d/1--BCeBBF57-2lb_KZ6gUF2pMRFdIj1gg/view?usp=sharing I'm not a thai local but I'm living in thailand. Anyways that doesn't matter.
>>19685 Whatever I'm just going to go ahead and buy that...
>>19685 I don't know enough about it, you would have to wait till Kiwi is here. Or ask around in other places as well. This place here is mostly about the integration of things into the process of building robowaifus, not about expertise in very specific areas. Also, I can't read the description. >>19693 Yeah, I buy stuff all the time to test something. No problem, if it isn't too expensive. Make sure you play it safe with high voltages, limit the amount of electricity you could be exposed to (e.g. battery instead power supply in the outlet)
Open file (4.38 KB 198x254 images (2).jpeg)
>>19694 Guess that's how it starts huh. With ziploc bag, some baby lotion, some tinfoil, some cables, a few components and a bug zapper.
>>19695 Meant to say baby oil
>>19685 I sure hope you're going under the assumption this >>19588 wasn't fake. Because it's clearly someone who doesn't understand much and is just trying to get views. Replacing a few capacitors won't multiply voltage like that, it'll just smooth out the voltage. They're an entertainer, not a teacher. You need a voltage multiplier circuit (https://en.wikipedia.org/wiki/Voltage_multiplier) Mini rant over, your shopping list is fine, aside from the fact you need some adhesive to keep your electrodes connected to the bag. Going to assume you're going to use tape or some glue you already have. Where are your diodes? Capacitors are important but, they aren't going to do anything without diodes to control your voltage multiplier. Remember, you need over 8kV of DC. Picrel is a simple circuit that will provide that for you. >>19695 Keep that positivity up, you're not alone. I will help you on your journey.
Open file (45.16 KB 1024x541 111.png)
Open file (845.10 KB 966x935 Newchest.png)
Open file (595.45 KB 996x747 Newhips.png)
Higher poly and more anatomically correct chest and hips.
>>19674 I printed a variant of the tilter and like how it fits, but my more slender upper body is now a bit small compared to the hips of the model. Still things to figure out. I also started using SophieDevs torso model as a reference. I wanted to do something like this for a while. Adjustments and modelling parts based on imported files. But it also looks just nice. Pictures like this keep me going, at least a bit every day. >>19724 Jade waifu
>>19722 Oh... I figured the gator clips would do for keeping the foil in place and transferring the charge. What kind of glue should I use? Also which kind of voltage multiplier should I use?
Open file (26.61 KB 1074x1382 Pep.jpg)
>>19725 Still have a lot of faith in your growth. Slow and steady. A hint, extension is not necessary. Just clear your path of rotation. Pepper (picrel) is a great example because she's very simple. You can see the distinct spherical joints which clearly communicate their range of motion. Look at her neck. You can see how she has a large dip in the front and back to allow her neck to tilt forward and back. Notice how the dips are also swept into a round feature, if she tilted her head side to side, this would accommodate that. Her waist functions similarly. It's helpful to visualize your waifu as a series of spheres that have one to three axis of rotation each.
>>19727 >Gator clips No, they would put too much sharp pressure on the bag and cause it to rip. You can use super glue. More helpful reading https://www.pnas.org/doi/10.1073/pnas.2006596117
Open file (45.16 KB 1024x541 111.png)
>>19727 Used the stacked voltage multiplier. https://en.wikipedia.org/wiki/Voltage_multiplier
>>19732 huh... I think I'll try the bug zapper first. But thank you.
>>19733 You use the circuit with the big zapper to reach the 8kV+ you need. 250V is low for a HASEL actuator.
>>19734 If it wasn't obvious by now that I don't have an electronics or engineering background well I don't... I'd appreciate it if you could tell me how to build it... Still wouldn't the breadboard and wires catch on fire though?
>>19729 >A hint, extension is not necessary. Okay, idk and it doesn't matter, since I already designed the part doing it. It wouldn't be necessary if I switch to another design, after solving the problem? >Pepper (picrel) is a great example Yes, I could make my design simpler at any point if it doesn't work. Or someone else goes with it, if it can be copied so easily. Pepper isn't quite new. >visualize your waifu as a series of spheres It's also a constraint. I already went with a flat tilter. Pepper might also need a motor holding her upright the whole time. She won't be easily backdriveable. Also, my thinking is still about going to a more human-like model, the current design is just the stepping stone. I won't commit to a completely shell based design. It's a completely different design and I'm not giving up my approach. You come up with these ideas right after the fact that a part or area is finished.
>>19725 That's looking interesting Anon. Is that the central pivoting mechanism I see in there now?
>>19722 >>19723 >Keep that positivity up, you're not alone. I will help you on your journey. That attitude is extremely-appreciated by me, Kiwi. Thanks for your engineering expertise as well. Cheers. >=== -minor edit
Edited last time by Chobitsu on 02/10/2023 (Fri) 17:40:03.
>>19753 > the central pivoting mechanism I see in there now? Yes, I'm using the other parts for comparison. Need to figure out the best way how to go about the relations. Also that abdomen shell could be seen as a preview if I would add something like it out of silicone rubber on the outside.
>>19756 >Need to figure out the best way how to go about the relations. If by relations you mean proportions, then I suggest simply using an orthographic view of a character with similar features to what you're going for. Pardon me if I'm misunderstanding things though, Anon. Regardless, she's coming along well. Keep going! :^)
>>19757 >by relations you mean proportions Yes. >simply using an orthographic view of a character I import parts into that model here and then I can decide about the size in relationship to other parts. >she's coming along well. Keep going! Thanks.
>>19759 Idk why, but I'm currently not as motivated as I should be. Fixing errors and making small adaptions is just less rewarding, and I tend to look into other things which are more thrilling. Anyways, for the future I should using such patterns earlier, to make the models closer to a human female. I also need to make my files into what they're intended to be: Frameworks which allow for fast adaptions. It's still to much trial and error. Current picrels show that her front is a bit to wide, but I think I got over it by reminding myself of the fact that this was meant as a dress which could be wider, and also it was never meant to be perfect. That she's not as round on the side and back compared to the pattern also doesn't bother me too much, I think. Though, I still have to figure out how to fit tilter and ribcage (lower chest) together the best way. However, this means my full-body OpenSCAD models aren't correct, since I'm changing the scale of parts later in this assembly file here, which is kind of frustrating to think of.
Open file (46.43 KB 736x414 example.jpg)
>>19736 >How to build it See this video: https://www.youtube.com/watch?v=o6eCkS77rsk You have 250V AC coming out of your bug zapper. You want 8kV DC, you'll want to use 5 voltage multiplier stages. >Breadboard and wires catching fire? Liking your eye for safety. No worries though, it's amps (A) that bring heat. Your bug zapper will output a few miliamps (mA). Your breadboard and wires have a plastic coating with high dielectric strength (resistance to current penetrating through them) your circuit will be fine. It certainly helps that your bug zapper is limited to the battery output wattage, which should be very low, thus limiting the maximum current output accordingly. >>19742 >Finished You've made a concept. The finish line is much farther than it appears. Don't worry, it'll be worth it once you get there. I've been where you are before and it's really tempting to think things will be easy but, they won't be. You don't have actuators or sensors figured out in relation to your concepts. >>19761 >Fixing things is less fun than making other things Very true, it's exciting to start, frustration will always await us though. We are imperfect but that's ok. Your ideas can become something wonderful if you stay with it. Honestly, I would restart if I were you. Work with newfound knowledge, figure out your degrees of freedom, figure out how to drive and sense those degrees of freedom, then apply that to a design which will work with them. The path of least frustration if function into form. Protip, expect your project to take many months, likely a year at least to truly get it right. You can get her done right though, I believe in you.
>>19763 >Don't worry, it'll be worth it once you get there. This. For every anon here, BTW. Eyes on the prize Anon! Keep.moving.forward.
>>19761 >Frameworks which allow for fast adaptions. It's still to much trial and error. Excellent point Anon. Finding commonalities is really important to all design work, and explicitly taking them into account can often greatly speed the processes involved. For a surprising, yet sublimely on-topic take on this, please take a break from your labors and read this book (>>1273). I think you'll come away both enlightened and refreshed. Cheers Anon, you're making good progress. Keep it up! :^)
>>19761 The strange thing about the MaidCom chest here is that it looks reasonable, but it is shorter (not as high) than my own approach. I will have to think about the conclusions I might draw from that. In regards to my other problem. Changing my file was easier than anticipated. I'm currently printing a adjusted tilter. It's still a bit difficult or tedious to transfer my conclusions from the assembly like in picrel to my other files or into scaling of the printed parts. Though, with the small models it's not such a big deal. I just should make sure to not print out two bigger parts which then don't fit together. Still on my list here >>19673
Nice very nice. I did not quiet understand but nice nonetheless... https://youtu.be/4alV5LzHLE4
>>19783 I think the ratio/proportions with the MaidCom chest looks OK for prototyping Anon. >>20058 Nice. Maybe you can link to this in our Electronics general, Anon? (>>95). Good luck with your project! >=== -minor edit
Edited last time by Chobitsu on 02/13/2023 (Mon) 22:11:59.
>>19783 Still working on trial and error. Working out the dimensions is a bit confusing between the different files and programs, but I'm getting ahead.
>>20091 That's looking really nice and clean now. Good job! BTW, can you kind of consolidate all your file & formats together into a unified system? I'm sure that will help with your prototyping & engineering Anon.
Open file (876.29 KB 1920x1080 embarrassed_yes.png)
>>20091 Well, exported for printing we want separate files, since prints can fail and also because of the size. But yes, the assembly file should be importing all the files and then allow for scaling and for more skilled anons also replacing parts and looking at the result. Generally this here is about getting something done as fast as possible, though. Don't expect some easy to use framework for all kinds of changes without any knowledge of OpenSCAD. I'm glad if I can make it soon at this point.
>>20098 >Generally this here is about getting something done as fast as possible, though. Excellent choice. Doing fine Anon, please proceed.
Open file (600.57 KB 1652x2560 Vday.jpg)
Happy Valentines day! Took a small break from engineering to design and print a cute and fun little thing. Torso and hips of a ~80 cm class waifu that's MaidCom compatible. Just something fast.
>>20147 Thanks Kiwi. Hey she's looking pretty good! I like the chest redesign tbh. Much more suited to maintenance now AFAICT. >a ~80 cm class waifu that's MaidCom compatible Don't stop what you're doing, b/c this task is still too early yet, but at some point we all should get together and define interfaces to things. Like, how would NoidoDev's arms connect into a MaidCom shoulder? Interlocking? Push-to-click? Something entirely different? These are the questions enquiring minds want to know! :^) Keep up the great work Kiwi. Looking forward to the brand-new MaidCom thread soon! Cheers.
>>20147 Cute. >>20149 Don't worry, these parts can consist out of different parts and so there will probably be some kind of adapters.
>>20153 OK, understood. Looking forward to a new task to define these adapters for future use, Anon.
>>20149 It was a heart, then I realized the point on top would have nothing supporting it. >NoidoDev compatibility >>20153 We should make a thread to start defining standards for waifu development. It'll help others start and work together.
My electronics kit arrived now all that's missing is the baby oil to make the hasel actuators... Anyways how do I make the voltage multiplier...
>>20156 >define these adapters for future use >>20161 >defining standards for waifu development Or we design something and then we'll see.
Open file (5.85 MB 1920x1080 VID_20230215_131053.webm)
It sparks muahaha
>>20161 >>20165 You guys are the hardware wizards, I'm just a software guy haha! :^) You two figure the new thread out between you, but at the least we should have an entry for it in Tasks (>>20037). Actually, software is absolutely rife with the issues surrounding the general idea of interfaces, so I can at least offer abstract suggestions to you two when you begin this task.
>>20164 Gratz, Anon. Please be careful. BTW, just to give you a heads-up, I'll be merging this thread into our current Prototypes thread (>>18800) soon anon. Please keep us up to date on your progress!
>>20170 Come on somebody draw what I'm supposed to do in ms paint or something...
Open file (20.87 KB 554x554 images (4).jpeg)
>>20173 How am I supposed to arrange the diodes and capacitors on that thing?! Tell me right fucking now!
>>20177 Lol, chillax, m8. :^) Didn't Kiwi already give you suggestions, Zapper-anon? (>>19722, ...)
>>20178 So I guess I'm going to have to look it up and try to do that myself huh
>>20179 Yep. Way of the builder, bro!
>>20181 I'll try to figure this thing out tomorrow but this is kind of bullshit ngl
Open file (3.62 MB 1080x1920 VID_20230215_234353.webm)
Okay I think I got some kind of idea of how to go about doing this. It's kind of scary though. I got shocked a few times... I can't imagine getting shocked with like +100x the the amount of volts lol Good thing I got those gloves but they're kind of cumbersome. But it'll have to wait until tomorrow.
>>20195 Maybe also don't touch it while it is powered? Can't you have a switch or plug for your battery? I don't want your wife posting here that you got a heart attack.
>>20165 If something like this here >>14312 is coupled with a servo or some rotational mechanism on the side where the hole is, then this coupler could be a regular connection to a servo pole. It's a non issue.
Open file (28.17 KB 271x394 1stage.png)
Open file (49.87 KB 603x400 vmulti.png)
>>20173 Here's more resources to help. https://hackaday.com/2017/03/22/how-does-a-voltage-multiplier-work/ https://everycircuit.com/circuit/5239629936852992/voltage-multiplier https://www.youtube.com/watch?v=ep3D_LC2UzU We're going to start with the simplest multiplier, the Cockcroft-Walton Multiplier. This multiplier isn't the smoothest but, will provide you with a workable voltage to use with HASEL actuators. The video uses low quality capacitors and diodes. Your diodes and capacitors should work better. I made a mistake when I previously stated that voltage multipliers doubled the voltage with each stage, I've never used the circuit before and my AC knowledge is based on household electrical work. Turns out, each stage adds the initial voltage to the output. The first picrel is a breadboard circuit example of a single stage multiplier which will bring you up to 500V max, though it will be somewhat less in practice. Second one is 4 stages and have a max of 2kV output, though it will be less in practice.
>>20202 Thank you sorry for being rude... Should have bought more capacitors... I have 4 of the blue capacitors and 5 diodes
>>20205 Well I wanted to get back to a regular sleeping schedule but I can't so might as well try that some time today... One last question though... when I put the wires of the bug zapper on the in and I turn it one they start sparking on the breadboard.
>>20206 Could you provide a video? It may be a crack in the breadboard, allowing for the high voltage to arc across.
>>20207 So these wires that came with the bug zapper are a problem. They fit on the breadboard but they keep coming off. I think I'll try again tomorrow and desolder them and solder the pointy breadboard wires maybe... I think i got it to run once and it didn't work. But I'll try again tomorrow. Thank you
>>20195 >Good thing I got those gloves but they're kind of cumbersome. If you can find authentic rubber (not rubberized-plastic) dishwashing gloves, then I think you'd find them effective in more ways than one, Anon. Maybe double them up since hi-voltage. >>20212 >They fit on the breadboard but they keep coming off. I think I'll try again tomorrow and desolder them and solder the pointy breadboard wires maybe... Maybe you could just solder one of the solid wire leads (the ones actually made to fit into the breadboard) onto each of the zapper wires, and just use that instead? Good luck, Anon. >=== -prose edit
Edited last time by Chobitsu on 02/16/2023 (Thu) 06:02:43.
Open file (2.75 MB 3000x4000 IMG_20230216_121552.jpg)
Well the baby oil arrived. Took a while... Is this really going to work?
>>20243 Your foil will work better if it is smoothed out, and both foils must be across from each other and far away enough to prevent arcs.
>>20244 Well i did not get any sleep so it'll have to wait sorry but thank you for the advice. But soon >muahaha
Open file (54.39 KB 2048x1251 canvas.png)
I can't go to sleep still... I am going to try to solder the pointy breadboard wire end to the end of the mostquito net wire. I tried soldering to pieces of cables in the past and i sucked at it but here's to a second try... If I can't go to sleep still like an hour from now. Boy do I feel like crap. I mean this is not really necessary but the cables keep coming off so.
>>20251 I recommend that once you have a good solid solder join, that you wrap the the connection together tightly in some quality electrical tape, Anon.
>>20252 I stripped the wire and soldered. It works. Now I'm going to try that circuit again but this time with no wire that comes off nonsense.B)
>>20254 I'd suggest you put electrical tape across those joints, and then put the whole thing inside some kind of protective casing. Can you saw off the swatter handle, and then put it back inside the remaining piece it came in? >=== -minor edit
Edited last time by Chobitsu on 02/16/2023 (Thu) 08:59:15.
It did not work however I don't know if I touched a capacitor or what but I got shocked and it felt like a body shock for a second... I might try later with the gloves but I'm kind of scared now...
>>20256 Yeah, be careful Anon. Be sure to wear some rubber-soled shoes when you're working on the project is my advice. You can bet Dr. Frankenstein took electrical precautions! :^) >=== -add funpost spoiler
Edited last time by Chobitsu on 02/16/2023 (Thu) 09:54:37.
>>20257 This is frustrating. If I can't get this working I can't reset this. I have a plan b and that is to just fuse the electronic components and not use the breadboard though.
Related from meta: >>20239 >But first you must design it so that together its a complete system that: >• can be easily removed intact (at least the maintenance item can) >• can be cleaned properly w/o damage >• can be easily re-inserted properly in it's place This seems to be more about the integration into the body. The first question would be if and how often the oil inside would need to be replaced, and same for the bag. If the oil would need replacement it could be flushed out by a internal cleaning system with pumps in a more advance robowaifu, or the part would need replacement by disassembling or opening the joint. This has not much to do with the current development with the muscle. We can not plan everything ahead. Optimization and the integration into a specific body design happens later. Related from meta: >>20258 >If the waifu was made with motors it'd make a lot of noise and be stiff True, there would be some noise, but servos with a small amount of reduction can be driveable by outside force. >if it was made with hydraulics We might use a combination of different systems. Either way, having a way to build a hasel actuator would be amazing. >If anybody wants to give it a try as well please do. Others will, at some point. I'm interested in it, but I don't always have enough time or energy to work on these things. I will add the parts necessary to my next order on AliExpress, but don't hold your breath.
>>20283 If you're too busy to help, too busy to buy components and shoot people down then you're just a negative influence. A net negative if you will.
>>20282 Thanks for posting the webm, it appears that your breadboard cannot hold back the high voltage. Try connecting everything with solder or, twisting together then wrapping the conductive parts with electrical tape to prevent arcing. I'm learning about AC circuitry to help, we're both learning as you go. Really thought the breadboard would handle it. As a general precaution, dipping the final circuit in wax to completely seal all high voltage parts would make it safer. Though, there are countless ways of applying a dielectric outer layer, like putting the circuit in a plastic bag. >>20283 >Oil replacement Shouldn't be an issue. No one has any reason to use degradable oils as the dielectric working fluid. Leaks would be a potential issue that frankly would need to be solved on a case by case basis. Chobitsu's advice on designing actuators that are easily integrated into designs and upgraded over time is objectively correct.
>>20286 Well I can't say I'm surprised that the breadboard that came with the electronics kit is a piece of junk. I don't want to solder the components together but I don't want to wait for another breadboard and I don't know if it'll work. So I guess I'll be doing that...
>>20285 I didn't "shoot anyone down", I'm just pushing back against this notion of planning everything in detail before doing something. I also didn't say I'm not working on anything. Just don't want to make promises about participating in that project here very soon, but I wanted to signal that I'm interested doing so in some time. >>20286 >use degradable oils as the dielectric working fluid Then that has been clarified, good. >actuators that are easily integrated into designs True, but that part is rather obvious. If and how it can be replaced by reinserting it would depend on the design of the body. With a seamless skin it would be rather difficult. How do you want to make sure anything in >>20239 can be archived at such an early stage?
Hmm what now...
I got some sparks out of it but it doesn't seem like it's stronger. I was hoping for a nice arch... I can't verify though cause I can't find my volt meter. Anyways I don't think I can do anything with this. I didn't solder it though I don't know if that makes it weaker.
>>20285 >If you're too busy to help, too busy to buy components and shoot people down then you're just a negative influence. A net negative if you will. Chillax bro. Anons are free to follow their own schedules towards their robowaifus goals here as they see fit -- just like you or I can too! :^) And I can personally assure you from experience, that no regular here today is 'shooting people down'. At all. We're all here to help each other out; in fact right now, we're all going out of our way to try to both encourage you and to assist you as a newcomer on the board. Don't just assume the worst when someone gives you advice here please -- this board is a little different. And if that advice was wrong, you can bet someone else beside yourself will likely call that out before very long. So be patient, would be my strong advice to you Anon. Good luck with your prototyping research BTW, we all anticipate your good success with it soon! :^) >=== -prose edit
Edited last time by Chobitsu on 02/17/2023 (Fri) 06:09:13.
Okay sorry I deleted it. Anyways, this is my job...
Never mind the other stuff I said. Anyways I'd like to keep learning about electronics and what not but I'm kind of in a rush though. Good thing I found this: https://th.aliexpress.com/item/33035175448.html Lets gooooo >=== -rm hotlink tracks
Edited last time by Chobitsu on 02/17/2023 (Fri) 11:13:49.
>>20318 Wait that'd arrive on april the 25th ugh
>>20319 Either way i think its supposed to be 8kv not 80kv I've already spent like $60 I think a lot of it has gone towards shipping though... How much is the waifu going to end up costing... I'm mostly worried about the outer shell of the waifu. Er the skin. I'm also sure there is a cleverer way to generate 8kv but I just kind of see that bag move. Still I don't think the waifus muscles ought to be made out of ziplok bags though.
>>20324 >I'm mostly worried about the outer shell of the waifu. Er the skin. Silicon rubber of medical or food safe quality is expensive. It might be possible to limit it to impregnate textiles for the skin: >>20327
>>20328 Well if we all chip in on this one an get like 20 kilos each we could make it a bit cheaper https://www.alibaba.com/product-detail/Moulding-Silicone-Rubber-TBL-8160-Food_1600262715534.html >=== -rm hotlink tracks
Edited last time by Chobitsu on 02/18/2023 (Sat) 08:34:47.
>>20329 I'd buy the 100 kilos even but what do I do with that liquid afterwards... Do we need the body to one whole hollowed out waifu or what...
>>20330 200 kilos I mean
>>20331 So um how many kilos of that stuff is a thick waifu anyways?
>>20299 3 out of 4 diodes were twisted in upside down. You're testing the positive ouput to the input that isn't ground. You should be seeing intermittent arcs based on leakages which your webm verifies. I added advice to an edited version of your pic. This should help guide you on how to quickly fix your circuit. Also, please cover the twisted leads with electrical tape or something to prevent premature arcing. >>20318 That should work. >>20324 There are far better ways of achieving kV power supplies. This is a learning and building experience. Was surprised to see cheap high kV power supplies being readily available, those make more sense. >>20328 Correct, medical grade silicone impregnated cloth is the ideal skin for a robowaifu. I'm going inside out so, her outer skin will be one of the last things I work on. If someone wants to start researching it now, I'll help.
>>20324 >spent like $60 I think a lot of it has gone towards shipping though That's why it was so fast? We often wait for weeks in Europe till something arrives, sometimes 10 days if it something small.
>>20335 Thanks for this wonderful labelling Kiwi. Very helpful!
>>20335 Thank you i soldered them though -_-' I bought more of those capacitors and 1n4007 diodes So I guess I have to watch which side the diodes are on >That should work. I think the paper said its supposed to be 8kv either way not 80kv and I can't wait until april the 25th
The people who created the hasel actuators could have built a robot... so why didn't they? There is something mentally wrong with some people I think. I'm still trying to figure out what it is exactly but on the mean time I'll try to make one myself...
>>20352 >so why didn't they? Building a fully-autonomous real world humanoid robot is no light task, Anon. And researchers are often fascinated/focused on just a specific area, or (even more often) just a smol sub-area. It takes good general systems design/systems engineering skills to bring all the moving parts into place together effectively. Actuators (of whatever form) are simply a piece of the much-bigger, overall, robowaifu puzzle. Does all that makes sense, Anon? >=== -prose edit
Edited last time by Chobitsu on 02/18/2023 (Sat) 12:55:24.
Open file (6.00 KB 224x225 download (40).jpeg)
>>20353 In the demo they had a mechanical arm that moves... They had a hand that can grab an egg... They could have made a cooking bot atleast...
>>20355 >They could have made a cooking bot atleast... It's a good point. But as you're probably aware, there are cooks, and then there are cooks! :^) Control software, recipe software, power & control electronics, sensors of several different kinds, and some sort of overhead attachment point/rail above, would all be needed too. And probably many other things unforseen until you actually undertake the design, prototyping, manufacturing, delivery, installation, and testing. But yes, just like robowaifus, it's all coming. And soon! :^)
>>20359 We don't have much time... This waiting in between for parts sucks ass... Kind of wish radioshack hadn't gone out of business.
Open file (1.96 MB 498x498 straitjacket-anime.gif)
If a group of anonymous posters can make a robot waifu that must mean a group of anonymous posters can make anything except maybe a cpu or a tv from scratch. A car, an electronic watch, a motorcycle, etc... But we're going to build a waifu robot. HAHAHA pee pee poo poo PEE PEE POO POO.
>>20369 >anonymous posters can make anything ... A car, an electronic watch, a motorcycle, etc... Look on YouTube and Hackaday and you will find examples for all of this. Done by some maker on his own or with his friends. That said, every challenge is different. Also, "from the scratch" is debatable, we don't need make the electronics and electric motors on our own. Also, we're getting OT here: This thread is about posting prototypes.
Can someone make a new off topic thread please? I've been ranged banned from 4chan for some reason. Maybe the glowies don't like me or something.
>>20382 'Glowies' huh? I thought they don't real? The CNN told me so. Check the catalog, Anon. :^)
How are going to make the spine? I think a sex bot doesn't even need balance really. They just need to be flexible. A sexbot doesn't need to do a whole lot of walking.
>>20398 This is off-topic once again, Anon. Please respect all the thread's OP subjects. In this particular case: >Prototypes and Failures #2 >Post your prototypes and failures. you're not discussing one of your prototyping efforts, but rather posting yet-another rando, off-topic question. Please either move it the ecchi-containment thread (Vagoo), or to the off-topic conversation currently-ongoing (/meta). The skellingtons thread might also be suitable. TIA. >=== -prose edit
Edited last time by Chobitsu on 02/19/2023 (Sun) 07:34:31.
>>20399 It is essential that the robot have a spine... Which thread should I post it on?
>>20402 >It is essential that the robot have a spine... You mean your current project. Others might not use a spine, but have a hard shell. >Which thread should I post it on? 1. Look into the catalog 2. You would've found this >>200 3. I you're not sure post it into the current meta next time (not in this instance) 4. You could also compile and install Waifusearch which is linked in the first posting in meta. Then next time, you can search for it. Links could also be in the Archive thread.
I made some progress with my neck mechanism. Got a bit annying figuring out how to print it. The general idea works but I'll need to change some things to make it really work. Also didn't have the right screw length, so it's still rather improvised. The hole in the middle is also quilte small for my piston tilter. Well it's something at least.
>>20409 >Well it's something at least. Yep. Good job Anon. I think it's really good progress for this specific component.
>>20409 *annoying
Open file (981.36 KB 2472x2260 0suk.jpg)
Posting to inspire prototypes, if it belongs elsewhere, please move it to a more suitable thread.
>>20425 Sukabu is a real design talent. The hip joints are a very complicated set of challenges to solve. He clearly at the least understand the geometric range-of-motion part of those. I'll just spolier it in-place Anon, thanks!
>>20409 On my ongoing mission to make my designs more complex...
>>20431 >On my ongoing mission to make my designs more complex... Haha wait, aren't we all trying to simplify things r/n Anon? :^) BTW, can you explain the new additions please?
Open file (2.05 MB 1280x720 VID_20230220_131920.webm)
I was trying to measure how many volts the fly swatter circuit is... So how many volts is it. I'm supposed to get it to 8kv looks like it's 1700 volts. I got 100 1n4007 diodes, those will be good for the voltage multiplier right...
>>20432 These side arms in the roller should make it possible to make the springs on the sides moving forward and backwards if the neck tilts that way. Otherwise one end would be static and get loaded when not supposed to be. Also, the distance is now a bit higher, so that I have more space for the springs. The other thing is, that these arms have a little holes (blue color in picrel one) in the middle for adding strings (rather a wire or a chain) to the upper part of the neck mechanism, for pulling it sideways. Then I made some kind of champfer into the big roller (purple in picrel one), since my long screw is to short to fit the whole way through. That said, it should be two screws instead a long one anyways.
>>20438 I see makes sense now. Are you close to printing your new prototype using this improved design, or are there other changes needed first?
Open file (20.90 KB 181x220 Exa.jpg)
>>20437 Assuming you have the diodes correctly oriented now, is your multimeter meant for the kV range? It's entirely probable that it's meant to measure up to 600V with spikes to the kV range. Which makes sense as most won't see voltages above 480V max. The simplest (and most fun) way to measure kV is to test how far away sparks form. Every mm of travel is 3 kV. Try setting up two conductors 2 mm apart and connect your circuit to them. you should be able to measure an arc across 2 to represent a minimum of 6kV. Picrel is an example of a simple testing setup using two paperclips in foam 2 mm apart.
Open file (2.25 MB 4000x3000 IMG_20230221_081422.jpg)
>>20467 Some anon told me to make a voltage divider and I tried two different ways and neither worked...
>>20467 >Every mm of travel is 3 kV Neat. Is that b/c the dielectric constant of standard air or something, Kiwi?
>>20444 There were some delays, but I'm already printing it.
>>20469 Yes, though it varies, air generally has break down voltage of 3kV a mm. It's always good to think of how you can use nature and what's readily around you to achieve your goal. back to MaidCom dev work
>>20468 >>20472 Well, difficulties are to be expected. But we can't let them stop us. Somebody once said >"You improvise, you adapt, you overcome!" You'll finally figure out a way if you just don't stop Anon. So, I'm curious why you're working on a voltage divider, Zapper-anon? BTW, have you managed your 8KV target yet? Good luck Anon.
>>20478 Good, hope it's a fine print job. >>20515 >t's always good to think of how you can use nature and what's readily around you to achieve your goal. Indeed. In fact, I suspect we'll need to find many different ways to economize things with clever ideas. After all, we don't want our MaidComs to cost US$20K, right? :^)
>>20438 I did some redesigns and printing. The channels for the nuts are still not wide enough. On another part I also have to change something to mitigate for my short screws. For now I'm trying to figure out how to add the springs the best way.
Open file (25.28 KB 464x636 simple_neck_T1v01d.png)
>>20540 The last print is more of a hybrid, since I'm always adding some changes. Generally I want to try to get my piston tilter design in there somehow. I also added more holes: Forward/backward string holes, and another pair to add screws or hooks to the sides of the upper part.
>>20541 Looking forward to seeing your strings installed. I think it will be easier to understand the mechanism that way.
Open file (874.57 KB 2219x2318 pantilt.jpg)
Pantilt mechanisms are fun.
>pic related Experimental "hydrophilic" wet silicone work has started. You need access to an institution to order PDMS-PEG block copolymer but it's only ~25$ for a lifetime supply, since mixing in a few drops into uncured silicone is all it needs. I have no idea how body safe this is except that the research paper says "highly biocompatible" (doesn't cause rat liver cell death). The "smart polymer" silicone becomes hydrophilic when soaked in water for a few minutes. This should allow for a cost-effective, cleaner (anti-fouling), wetter silicone. If it's what I am imagining, then it will be able to self-clean in a closed system. >source: >https://www.nature.com/articles/s41598-019-43625-5 >results: >https://files.catbox.moe/oxg9wh.mp4 The silicone pucks form a thin layer of water instead of a bead like normal. The audible squeaking is only from my glove. The silicone is soaking in IPA for 24H like it says to do in the paper. I tried 80-85%, 60-70%, and 50-55% PDMS-PEO from Gelest.com (the only ones easily accessible to me) at 0.013 parts to 1 part silicone and they all perform the same way, but the 50-55% might get wet a little sooner. All extremely wettable. After the soak the properties may change, will update.
>>20569 Excellent research work, thanks Anon. This has obvious robowaifus snu-snu implications for vagoo & oral. I'm wondering how surfactants such as Dawn detergent affect the hydrophilicity long-term? Also, I just wonder if this can somehow be used to stand-in as a sort of 'highly-viscous synovial fluid' that can be used to cushion/lubricate articulation joints such as spinal vertebrae, shoulders, hips, & knees for instance?
>>20568 That's a good point, Anon. The motorized camera heads are fairly in-abundance out there (if expensive). Surely some prototyping efforts at the least could be shortened by utilizing some of them?
>>20547 >Looking forward to seeing your strings installed. I think it will be easier to understand the mechanism that way. I doubt it, it's just so that the upper part can be pulled sideways and on the lower end it can be pulled downward back and forth tilting. Not very complicated. The left and right tilting snaps back after the force stops being applied, while the same is not true for the back and forth tilting. The head is supposed to fall forward if the string at the back is loosened. Though, this might change. >>20568 It would be more useful to tell everyone here how to find the best solution for some kind of problem. Or you go through all joint movements yourself and put the info out, maybe into the R&D general, and ideally two years ago. Not after someone designed something. This infringes on sabotaging or demoralizing, if it's not just boasting. I'm not going to design a new part after I'm deeply into it, even if in theory it would be the better approach. Too late for now. That said, mostly these ideas are not working anyways, or at least it's not obvious how to solve other problems and since it's too late I don't care. In the specific case, if you wanted to rotate a head on top of such a mechanism, you would most likely have to rotate the whole construction. I don't see how to put a pole through it starting from the bottom.
>>20569 >Experimental "hydrophilic" wet silicone work has started. Why? >then it will be able to self-clean in a closed system. Cleaning doesn't mean to take water in? Rather the opposite.
I think I did it right yet I still can't get a clear reading...
>>20579 Some guy on 4chan again said I need an oscilloscope... but then what do I do with the oscilloscope... I'm starting to get a bit discouraged.
>>20579 What are you trying to make Anon? The voltage divider? >>20596 A) You don't need any expensive gear like an oscilliscope anytime soon, and B) don't get discouraged so easily. You have to learn to pace your expectations Anon. This is going to be a years-long effort. Marathon, not a sprint, alright? Cheers. :^)
>>20578 >>20572 The hydrophilic silicone may have less tendency to keep dirt, hair, ect stuck to it when wet because water runs over it better. Just salty soapy water might clean it well after use with Jlube. >>20572 Ordinary silicone is lipophilic, might work with silicone oil out of the box. Maybe a platinum cure silicone could form a cup around a bone joint with silicone oil inside of it like real joints. Silicone oil doesn't leech inside of EcoFlex. Dragonskin might function as the cartilage/sliding material. I'd rather just use ball bearings anyway.
Open file (2.04 MB 3000x4000 IMG_20230223_201414.jpg)
>>20202 I copied that exactly and it doesn't work...
>>20657 I'd suggest you also post an extreme-closeup of the board, with all the components adjusted apart slightly to clearly show their interconnections onto the breadboard, so that Kiwi can analyze that and give you recommendations Anon. Nice effort Anon, keep going! :^) >=== -minor edit
Edited last time by Chobitsu on 02/23/2023 (Thu) 13:54:13.
>>20602 Thanks for the response Anon. My reasoning behind such a question is that my own design work tends towards the biotensegrity-inspired concepts discussed already for MaidCom. Since we're already working towards such a tension/compression suspension arrangement (just like in a real human), then we may as well work towards shock-absorption/'lubrication' at the pertinent joints also (just like in a real human). Also, thanks for the reminder about the oil. What do you think of combining the two (hydrophilic + silicone oil) together in the same enclosing capsule BTW? >=== -prose edit
Edited last time by Chobitsu on 02/23/2023 (Thu) 14:54:50.
Open file (2.57 MB 3000x4000 IMG_20230224_080219.jpg)
>>20662 I made it a bit neater... It still doesn't work
>>20669 Take it all apart and try again, Anon. Space each set of caps/diodes out down the breadboard more if you want easier working clearance, and easier closeup-photography so anyone else can examine your work better. Good luck. Mind the diode flips between the matching sets. >
>>20673 Okay I tried to make it as clear as possible this time...
>>20675 While that's a big improvement, until you learn to keep your work much neater than that, we'd still need a professional-tier 360 orbit video (or a good stop-motion video of the entire process), to have much hope of understanding what you've plugged into where exactly. Stress on the word exactly, b/c nothing else will do for this stuff. Please clear the entire board and then neatly do just one matching set, as in pic-related: (try hard to make it look just like this diagram, when seen from above) > Also, I'd recommend you use the short wires that are typically provided in a breadboard kit to do your cross-bar interconnections with, and not to use your long jumpers for them (since that significantly confuses tracing leads after the fact). A small needle-nose pliers can definitely help with the process of inserting these small wires (and other parts) into the breadboard. >=== -prose edit
Edited last time by Chobitsu on 02/24/2023 (Fri) 12:59:03.
>>20675 Looks like it should work, what woltage are you seeing at each stage? It's possible you have something faulty, each stage should increase voltage by 200V to 250V assuming the inital V is the specified 250V AC. Have you tested the initial voltage to ensure it's 250V AC? >>20677 Testing one stage at a time is actually an excellent idea.
SoaringMoon's 3D waifu models (Full Archive .zip 5.01 GB) https://mega.nz/file/uwQEQZKD#U5UnaHBeBg4t4JvAafv8GWDnrsLFccXJYLmwi33iJc0
>>20755 I get some odd 'out of disk space in chrome' error for this and I'm not going to install the desktop app. Any chance someone can break this into smaller segments for us plebians? :^)
The stack exchange said that the voltage multiplier is weak and collapses when any kind of load is placed on it. This stuff is useless for what I'm trying to do either way... You know the hasel actuator... I mean if you guys didn't know its whatever, if you did know I was just sent a wild goose chase :^)
>>20757 Can fetch successfully from terminal using megadl (megatools package): https://megatools.megous.com/ Linux and Windows builds available.
>>20788 >hasel actuator >voltage multiplier is weak and collapses when any kind of load is placed on it Can you link to the discussion on stack exchange? >if you did know I was just sent a wild goose chase Only one of us knew remotely how this might work. Don't just rely on this board. This place here is good for talking about how to put tech together into a robowaifu and how to go about it in the first place, also providing motivation and a social network for that, but for technical details and executing you'll need other additional sources. We aren't enough people here, for starters. No enough to know everything. Anyways: Don't give up.
>>20793 >megadl Nice, something called megatools is actually in my repo.
>>20793 Thanks Anon.
Open file (17.13 KB 1519x939 snapshot00.png)
I made a clamp based on this pattern from Kiwi >>19413 (thanks, btw) - Not as precise as intended, and the OpenSCAD code is rather bad, but it works. I would need to make the code more mathematical to scale it in the code, but I just used the export and scaled it down to 26 percent and that's good enough. It fits a 5mm Nema17 shaft perfectly. I added it to the string holder for the motor tower >>18672 (picrel) - which is just a prototype for now. I already printed it and it fits very well. shaft_clamp 3MF: https://files.catbox.moe/x27ojh.3mf shaft_clamp STL: https://files.catbox.moe/dksey1.stl string holder STL: https://files.catbox.moe/neen95.stl OpenSCAD code: https://files.catbox.moe/misy5c.scad
>>20809 I like that you were able to reuse another design on the board. Good job!
>>20798 Okay I'm going to stop beating around the bush. I ordered a 15kv generator and another that claims to be able to generate 1000kv...
>>20818 >I ordered a 15kv generator and another that claims to be able to generate 1000kv... There you go Anon, good job! 'There's moar than one way to skin a cat' or so the old adage goes. :^)
Open file (279.33 KB 564x798 ClipboardImage.png)
WIP of maidcom leg. Shin and foot are blocked out. When thinking about how long it would take to print a whole leg (and reprinting it if the dimensions are off), the tensegrity / dowel rod spaceframe idea makes a lot of sense. You only have to print the joints of each "bone"
>>20846 Nice! Thanks Anon. >When thinking about how long it would take to print a whole leg (and reprinting it if the dimensions are off), the tensegrity / dowel rod spaceframe idea makes a lot of sense. You only have to print the joints of each "bone" Now you're getting it! Biotensegrity-inspired designs bring a metric boatload of goods to the table, economy of material requirements not the least of which (as you touched on). But I would say that, arguably, it's the relatively low mass (often combined with even greater strength & resilience) that is it's primary benefit for us here. Cheers Anon, keep up the good work! >=== -prose edit
Edited last time by Chobitsu on 02/28/2023 (Tue) 05:50:37.
Open file (307.08 KB 1920x2560 tripod.jpg)
>>20846 Using dowels makes everything so quick and easy. Built picrel in under an hour (from concept to assembly). (Inspire by MeidoDev's work.) 2 ft tall and weighs barely anything. Even has built in passive springs in the knees to compliment the natural springiness of the dowels for fun. On a serious note, what size dowels can you attain? I use 6.5 mm thick dowels. Is this commonly accessible where you live? This is an open question for everyone as accessibility is extremely important. >>20859 Bio-inspired-tensegrity is a surprisingly tough nut to crack. Still working on it. The concept has the potential to bring uparralled durability per mass.
>>20890 >Bio-inspired-tensegrity is a surprisingly tough nut to crack. Still working on it. The concept has the potential to bring uparralled durability per mass. I'd suggest just don't even worry about it at this stage. After all this is just prototyping still. We'll figure it out well enough after we've gotten other design work off the table first. Remember we're working on many different things at once here Kiwi. Probably smarter to use more traditional approaches for most things, innovating for just one or two thing at a time. Fall back on your traditional training, yeh? Cheers m8.
>>20890 >I use 6.5 mm thick dowels. Is this commonly accessible where you live? In burgerland, we have 1/4 dowel rods, which are a hair (0.00591mm) under 6.5mm
I think I made it work but it does not feel like 8kv. Either way I'm waiting for other parts...
>>20987 Congrats Anon, proud of you!
>>20809 Finally got into the whole motor controlling business today. Though, only with a L293D shield that doesn't allow for enough current and get hot very fast. I have so much of this stuff lying around, for years. Time to get into it, and it's as annoying as I thought it would be. With wires being wrong in cheap motors and such. Tutorials being wrong or confusing. Anyways, at least I got started. A little servo moved nicely. I also have at least one DRV8825 driver around, which should work better for my steppers. I hope it's more than one, my workplace is a mess. I'll just might order some more.
>>20993 >my workplace is a mess. <I'll just might order some more. Lol, what could possibly go wrong? :^) Congrats NoidoDev, it's encouraging to us all to hear you've gotten back into the electronics again. Godspeed.
Tinkercad is not perfect. I got a circuit that works irl but not on tinkercad...
>>20987 Glad to see it working! You have five stages which when multiplied by the input of .25 kV, should yield 2.25kV maximum. Your spark looks to be about a mm, everything seems to be correct. >>20993 It's frustrating that so many tutorials still recommend the ancient L293D when it's the worst option for anything that requires over a few watts due to its ineffeciency. DRV8825 is a great driver, just remember it requires higher voltages, I'd recommend 9V minimum. Once you get used to driving steppers, I recommend moving onto the TMC2208 or better. Those have helpful features to reduce current consumption and noise.
>>20994 Yeah, I thought I have some of those already but can't find them and there are these CNC shields. But I already have something similar. >>21001 >L293D when it's the worst option for anything that requires over a few watts due to its ineffeciency I might not matter in many cases, especially for getting the feet wet. >DRV8825 is a great driver, just remember it requires higher voltages My steppers need 12V anyways. Yeah I know about these other drivers from 3D printing. Though some might be smooth but slow, I think. I realized that steppers are very likely too slow for what I want to do, so it's more of a test. Which was the initial idea anyways. I'll look into controlling my small BLDC motors as well, just some simple speed control via software at least, also getting into simpleFOC but only with one board. I will probably be in another place for a few month soon, and then order some bigger BLDCs.
>>21001 >Glad to see it working! You have five stages which when multiplied by the input of .25 kV, should yield 2.25kV maximum. Your spark looks to be about a mm, everything seems to be correct. What's the projected voltage he'll need to have for the Hasel effect given the bags he's shown? Is there a 'best' oil (or other apparatus materials) to use Kiwi?
>>21007 >But I already have something similar. Always cheapest if you can make do with what's to hand Anon. :^)
>>20992 >>21001 Thank you. I think i did it right cause the sparks are more continous. That must mean there i tiny arc in there... If I had known i would have gotten 20 capacitors, I thought I wouldn't need the 10 actually. But the guy at stack exchange says the voltage multiplier collapses if its power its drawn so its not much good for much of anything other than making electric arcs I think either way. Good thing I got some other stuff on the way. >>21011 I think the paper said 8kv but it feels like I'd want a good centimeter arch tbh. Something like 30kv. I don't think that bag will move with just 8kv. But I don't know.
>>21013 Well, you're clearly dedicated to this prototyping experiment Anon. I'm sure you'll figure this all out! Cheers.
>>21019 You can make a robot without a hassel actuator but hassel actuators completely change how you'd go about it. If you use say motors you're worried about the limitations and the torque and what not. With the hassel actuators if one is not enough just get more and one alone is enough to pull a robot arm according to the demo.
Open file (725.12 KB 3836x3472 IMG_20230303_092656.jpg)
27Nm peak torque at 100 amps direct drive.
>>21007 I started cleaning up my mess and it turned out I was right about having bought a bunch of drivers, but these are just DRV8833. Which are good up to 10.8V and 1.2 Amps - not really enough. >>21028 Nice. Glad you're going forward with this.
>>21013 Make sure you don't electrocute yourself, especially through important body parts.
>>21031 I told my dad about what I was doing and he said he did not like me playing with high voltage and read a book on electronics basically lol. I really should though...
>>21028 Pretty gorgeous work Anon. Gratz. What are your next plans for the prototype? >>21030 >I started cleaning up my mess Haha, the first step is always the hardest! :^) >and it turned out I was right about having bought a bunch of drivers Nice. See? Already worth it, Anon. >>21031 This. (Although I'm breddy sure Anon meant 'shock') >=== -combine posts
Edited last time by Chobitsu on 03/03/2023 (Fri) 18:59:34.
Open file (1.99 MB 3000x4000 IMG_20230305_011606.jpg)
Open file (1.97 MB 3000x4000 IMG_20230305_011555.jpg)
Open file (1.97 MB 4000x3000 IMG_20230305_011535.jpg)
My 15kv voltage circuit arrived. I followed the YouTube tutorials(cause it didn't come with instructions. But for some reason it doesn't work. I can't verify anything cause I think my volt meter broke when I was trying to measure the sparks last time
>>21065 Looks like there's some problems with your circuits connections. See picrel for advice.
>>21068 Yes I clipped them it still doesn't work... Either the circuit board or the components are nto working. I'm thinking the circuit board tbh. I'll desolder them and try on the breadboard I guess... I know the batteries have charge because they lit up an led.
Open file (21.21 KB 1532x466 spaceframe sketch 1.png)
Leg dowel spaceframe ideas: 1. Cut a dowel hole at the desired angle into the interface plate 2. Make a universal plate with rotatable dowel holes (so you can use dowels of any length and mix/match plates of different sizes) 3. Just put a ball socket and hinge on a hardware store 1" dowel and call it a day :^) The dowel can be any length and you can fit contour plates to it to give it an appealing shape. (Or a similar design with multiple smaller dowels) Thoughts?
Open file (170.81 KB 1920x1924 test.jpg)
>>21074 I like your ideas 1. These can be frustrating because the rods and print need to flex to fit together. It is effective. 2. This is a really interesting idea that I'm going to try. Will post about it tomorrow. 3. What I was planning on doing tonight before option 2 was presented. Still going to try it. Picrel is a 100 gram hip that's surprisingly durable inspite of being a failed print with obvious layer lines from where it failed and needed intervention. PLA can be an effective elastic element when used correctly. Another step closer to full sized ultralight waifus that can take a pounding.
>>21074 Probably will be easier to tell once you have prototypes assembled Anon. Good luck. >>21079 That's looking very interesting Kiwi. Godspeed.
Open file (958.95 KB 3472x3138 IMG_20230305_105315.jpg)
>>21037 The prototype wasn't complete, I printed the frames without a bolt flange for test fitting the internal components. Here is a completed frame with a place for a random number of bolts I had lying around. I have to drill out every hole with a drillpress and print a bulkhead for the rotor, which is the magnet thing on top of the assembly. The thing with all the coils is only the stator. This is getting very heavy...
>>21086 Nice! Glad to see all the components nearly complete now. Very clean assembly. I'm curious how you wound your stator elements? >This is getting very heavy... Still this is quite impressive for homespun engineering ATM. Maybe it can serve in some locomotive base ala the ideas tossed around in MaidCom thread? Once you complete this prototype any plans on scaling it all down somewhat? Regardless, looking forward to your good work to come Anon.
Open file (869.32 KB 1920x2560 round.jpg)
Open file (449.12 KB 1493x2204 quare.jpg)
>>21074 Tested method 2 using flexures for rotation. Round dowels will rotate too much in there sockets. square dowels are perfect.
>>21098 >Round dowels will rotate too much in there sockets. square dowels are perfect. Excellent work Anon. As you're probably aware, circular cross-sections are more resilient agains strain & stress. However in this structural context, the square notch will clearly resist in-socket spar rotations.
>>21087 I didn't wind the stator, I bought it the motor off aliexpress for ~100$. This is the highest capacity frameless motor available on the internet that I've found, 27 N*m peak torque @100+ amps. This big, round, strong motor is going straaaaight into the robot's hips. Every other motor size from this seller might be useful for joints with scaled down torque requirements based on the computer model that I didn't make yet. >https://aliexpress.us/item/3256804320185350.html Motor is around 300 grams, bearings weigh around 200 for 2, printed parts weigh around 200 grams total, and the motor controller + nuts and bolts weigh a little more, so maybe 800g for a fully integrated (plastic) servo. I'm solely building this prototype to characterize the motor weight and real performance for a computer model of the robot. If this works, anyone with 2-300$ (not counting the position encoder $$$?) and a 3d printer can make one ridiculously powerful motor that might be enough for bipedal walking with a low reduction gearbox. I now have to decide which rotary encoder to slap on for it to work. I'm using a moteus r4.11 bldc controller (open source firmware, apache 2.0, publicly available design files). The next iteration of parts will be printed in PLA this week so I can install the encoder later and then perform the torque test. >=== -minor typo patch
Edited last time by Chobitsu on 03/06/2023 (Mon) 18:17:51.
>>21115 >This big, round, strong motor is going straaaaight into the robot's hips. Oh nice! that's just what I was hoping from your first pic. Somehow the later one made me think it was somehow too big. Looking forward with eager anticipation to your completed testbed Anon. Can you please make your design files available to the group?
>>20091 I'm thinking about how to make sure the arms on my servo can even pull the upper body sideways, without pulling the other string working in the other direction. Maybe I should've thought about this a bit more in a earlier stage, but it was simply too much to plan out to every detail. I just worked under the assumption it's gonna work somehow. I tend to figure some things out as I go. Anyways, I'm realizing, that I will probably need my piston extension for that. Picrel One is supossed to show how one string would be pulled down while the servo moves left, and the other string is not being pulled because the arm can bend in one direction and then extend, so the opposing string is not being pulled down. If that doesn't work, I'd need to use more servos, but I'd like to avoid that. >>20993 Some of my problems testing my servos were because I used a power supply that put out AC. I forgot that I had such ones, just looked at the voltage output. *lol* and *facepalm*. Anyways, there seem to be other problems with the wiring not fitting the manual. I have to look into that later, I'm working on something else now. Setting up the driver wasn't that bad, besides that it doesn't work yet. I only needed to confirm that it will most likely work. I also ordered some other drivers and a CNC shield like Picrel Two. Should've done that earlier. I guess I will look through some files now and start printing them in full body size. I'm running out of excuses for not doing that, except my plan of likely not being in that place right now I am currently for a few months. Since I'm not 100% sure about it, this probably isn't a reason to wait, since there's always something.
>>21117 The design files have modified dimensions than what I'd order from a machine shop to work on my 3d printer. Other than that, I will post them as long as I can sterilize my identity from it.
>>21125 I can't say I can exactly picture your goal here, but I definitely agree that a cable-based system (your strings) can very likely give us good actuation dynamics and economy. We're looking forward to seeing what you come up with NoidoDev. Godspeed.
>>21145 >picture your goal here Should've explained it better: The left arm moves down and pulls the string. These three ones are just meant to be one. The three on the right are also the same, in different states. They should bend in one direction, but not the other and only extend the arm if they're the not-pulling arm. Not completely sure if I made a mistake there, but I think I got it right. If the string pulls from the back it won't bend, but if it pulls from the front it does and from the top it extends. Should've worked on this today instead of wasting my time by getting annoyed with other topics.
>>21149 Ah, OK. Looking forward to seeing the prototype moving soon Anon!
>>21098 >Using the natural flexibility of PLA instead of a gopro mount-esque hinge Didn't think about that, very creative! Is there a cookbook or formula for shortcuts like that?
>>21152 The key with PLA is to have the flexure either be large or have minimal strain. PLA unfurtunately has a tendency to permanetly deform, it has high plasticity whereas high elasticity would be ideal for flexures. in my leg design, the flexure element is 2 mm thick and only deforms a few degrees. Having the tension mirrored in the design also helps. 1 mm would have also worked, it would have been less durable. The large lump in the center helps to focus strain onto pivot points. The lump feature should be smaller, it's only large in the picture to help illustrate a sense of scale, this leg segment is similar in length and width as a 150 cm tall human woman's thigh. If I were making it for an actual robot, the lump would be a few cm so that the strain from elastric deformation is more spread out.
>>21154 >The large lump in the center helps to focus strain onto pivot points. Neat. Do the inset elements of the hips also serve kind of similar function, Kiwi?
Open file (59.08 KB 1054x1237 Spaceframe leg 2.png)
>>21098 Hmm, so the implementation would look something like picrel. Next step is to get the balljoint and hinge worked out.
>>21158 Those are nice sketches Anon, good job!
>>21155 Correct, I noticed squeezing her felt wrong with her being hard. I do want to create an outer shell based on silicone cloth to fix that but, I know next to nothing about silicone manufacturing. So, I'm experimenting on ways to make plastic squeezable. The hips were an extreme example, the long flexures move too easily. They also readily make annoying scratching noises when messing with them. More research is needed to get it right, Also, switching to PETG going forward. >>21158 Nice design, will test it tomorrow.
>>21162 >So, I'm experimenting on ways to make plastic squeezable. Excellent. The more consolidation we can make around a single fabrication technique alone will reduce both the complexity and costs of manufacturing.
Open file (680.86 KB 1920x2560 Normal.jpg)
Open file (590.42 KB 1920x1975 Bouncy.jpg)
>>21158 Made a leg with your design. It's incredibly bouncy with the upper section being 160 cm across and the bottom being 73 cm across. PLA still just because I'm almost done with the roll and just want to empty it before switching to PETG. PLA at 50 cm + lengths and between .6 and 1.6 mm can be very elastic. The top stick can move in all directions very freely. It's honestly fun to play with. Will post as actual leg in a bit.
Open file (474.67 KB 1705x2560 Leg1.jpg)
Open file (607.22 KB 1920x2560 leg2.jpg)
>>21177 Here's a simple full scale leg that's reliant on flexing. Knee is the problem, I'm getting really close to a good one though. Would need some panels/discs to fill out woman's clothing. Based on Aigis measurements naturally. Mostly made for testing. >>21165 Ease of manufacturing is my passion. Flexures make everything so much easier. Designed a print in place mechanism only to realize a line of plastic with the correct dimensions did the same job. We have a bright future in simplicity.
>>21177 >>21179 These look really cool Kiwi! If we can somehow solve the joint, actuation, and other design requirements for limbs after such a fashion as this then we'll be much closer to our goals overall IMO. >We have a bright future in simplicity. This. Everything is better when it can be done simply.
Open file (709.16 KB 3472x4624 IMG_20230307_213249.jpg)
Open file (700.90 KB 3472x4624 IMG_20230307_223052.jpg)
Open file (698.61 KB 4624x3472 IMG_20230307_223612.jpg)
Open file (709.76 KB 4624x3472 IMG_20230307_225316.jpg)
>>21115 Motor v1 complete. It doesn't always turn since the frame that holds the stator isn't stiff or accurate. If I made this with a lathe and glued it in, it would be fine? Plastic motors are a terrible idea regardless. The stator collapses in on the rotor sometimes because of 0 frame support holding it into shape and away from the rotor magnets. The rotor itself is warped by 1 mm out of round to begin with. Next is aluminum and an encoder bracket mount, but I'm going to work on chatbots for now since I cut my finger a bit on the stator.
>>21194 Congratulations! I'd consider that a nice milestone if I were you Anon. Also don't let setbacks or misjudgements either discourage you or stall your forward progress. We all of us are going to have to deal with many of these before we reach the lofty goals in the end. Just keep moving forward Anon! :^) >since I cut my finger a bit on the stator. Ouch! Please sanitize and bandage it properly Anon to not get infected. Cheers. --- Please be careful everyone. Remember this is protoyping design work, not yet fully-engineered products, foolproof enough for the masses. Some of this stuff is hazardous in it's incomplete form.
Open file (366.80 KB 2601x1272 hinge 1.jpg)
Ok guys, I made baby's first hinge. With a 6mm hole and no cleanup you can throw in a gopro bolt (M5) and- >>21177 >Kiwi took my rigid plate idea and already has a flexible spaceframe leg out of it DAAAAAYUM
>>21162 >I'm experimenting on ways to make plastic squeezable It's a bit of a kludge and takes some thought but a way to make rigid objects squeezable is tensegrity. Seach for "tensegrity tables" and look at the images. These are rigid objects but they have movement if the cables are flexible. Imagine a set of much smaller elements suspended from a bone with something like nylon fishing line that is flexible. Now you might say assembling such a thing would take forever. Yes by hand and to set it up would be really mind numbing but...say each part was in a jig(set frame), each connection was a wedge that snapped shut on the nylon line like zip ties or model car kits. All the little surface parts laid out flat in the jig. The nylon thread weaved in place, wedge snapped tight, then the whole is gather at the ends and snapped closed making an arm, leg...etc. If you look at tensegrity structures the solid parts and the tension members, nylon, automatically would form the shape as it is stressed closed. The whole object, no matter what shape, would be formed from the interrelated distances from the rigid pieces and the length of the nylon tension members. Not saying this is the best way to make a limb but it is one way you can have hard structures that are flexible. It would most likely take a compouter progra t figure the string lengths and to design the jig. Which with modern 3D printers, if you had the program, would not be so bad to print out. Another way to look at it is bicycle spokes. Except in this case the bicycle spokes would go out to thousands of little small pieces(exterior)and the spokes would be a streachy rubber.
>>21177 I got the idea from looking this comment you made's pictures. These are sort of like tensegrity structures. You're half way there to making something tensegrity.
>>21199 Neat! Great job Anon. How did you get the print to look so smooth? Looking forward to more good work from you. Cheers.
>>21201 >It would most likely take a compouter progra t figure the string lengths and to design the jig. Which with modern 3D printers, if you had the program, would not be so bad to print out. This is a great idea Grommet!
>>21204 > How did you get the print to look so smooth? Low light and Image compression :^)
Open file (676.88 KB 1920x2560 bent.jpg)
Open file (906.05 KB 1920x2560 straight.jpg)
Open file (4.32 MB 1920x2560 CompundBend.jpg)
Open file (981.49 KB 1920x2560 extremebend.jpg)
>>21193 Getting there, mein gott, PETG is incredibly elastic. Currently working on a dual helical flat joint with spherical rotation. A line of PETG does the job but, the point where it connects to the middle piece was too easy to rip off when slamming it against things. The knee is nearly perfect now, bends in an arc that approximates real human knee motion. Needs a knee cap though. Once things are ironed out a tad more, will start designing in anchors for servos to manipulate motion. >>21194 Great design! You're correct, machining an aluminum frame for the motor is going to work much better. I do worry about that mm of warp. May cause vibration and efficiency losses at higher RPM's. >>21199 Nice work anon, keep it up. >>21201 Great to see another fan of tensegrity. I actually started at tensegrity and developed it into flexures as string and rubber bands were frustrating to deal with. Thin strips of plastic are much easier to deal with. I'm just too lazy for tensegrity :^) (All my tensegrity projects broke in testing so I didn't post about them.) (I'm very rough with prototypes.)
>>21199 That is very nice. What did you do that with?
>>21208 >I'm just too lazy for tensegrity The mental strain to set something up like this would be immense. Once done and all the parts set to print it would then be low energy.
>>21207 Haha very tricky! :^) >>21208 >The knee is nearly perfect now, bends in an arc that approximates real human knee motion Yes that does look like a nice curving form. Will you be ready to do a destructive test for mass-carry at some point, do you think Anon? I once managed a smol strut that handled over 90lbs steady-state, vertical compressive force before failure (and that simply b/c tangential force buckling from a shoddy test rig -- it could've held more). ~8.5x0.75" in size, that tubular strut prototype cost ~$0.32 in materials & weighed ~1oz (a ratio of 1440:1). In quantity, that material cost would drop to less than a nickel I'd estimate. :^) >=== -prose edit
Edited last time by Chobitsu on 03/09/2023 (Thu) 18:48:31.
>>21221 Please post your strut, could help anons
>>21222 As I said, its destroyed. squashed flat by 90lbs of iron barbell weights heh :^) But its easy enough to make one however. --- Materials: -6 jumbo drinking straws -1 11x8.5" sheet card stock paper -2 waste milk jug caps -High-quality tape, like Kapton -(Optionally) school glue, & Scotch-tape Tools: -1 oil funnel, trimmed to ~0.70in inner orifice -1 dowel, sized to fit fairly snugly through 1 straw -1 broom handle, ~0.75in dia. Manufacturing technique: -Very tightly roll, then Kapton tape the paper into a tube around the broom handle (you don't need a full sheet; cutting it in half along the 11" side will leave you with 2x 5.5x8.5" subsheets -- enough for 2 tubes). -Remove & slide this paper tube onto the (outside tip) end of the funnel, with a solid friction fit. (Optionally, you could temporarily Scotch-tape it there.) -Putting the dowel through one straw, compress the other five into a ring around this central one (gripping around it all together firmly in one hand, then temporarily securing a piece of Scotch-tape around the end of the straw bundle (away from the funnel) with the other hand worked for me). -Force this collection together down into the funnel, and then on down into the attached paper tube (it should be a very tight fit; that is, you want all the straws to be under stiff compression together inside the paper tube). -Withdraw the dowel (alone) back out of the funnel. -Remove this radial-core strut from the funnel's end, snug all the straw/paper end alignments, and securely fasten the two milk jug endcaps on (if your cap-inner/tube-outer diameter sizes are proper, they should match & twist right on; otherwise adjust sizes accordingly). Glue helps. You're done. --- Extra Protips: -You can even inject standard expanding construction foam down into the straws (once they're inside the paper tube, & before the second endcap goes on; use a very fine injection tip to fit into each straw individually), for extra strength against lateral forces. Please wear eye-protection! -Generously-coating the innermost straw with school glue first (before compressing the other five down around it) will improve this resistance further, once cured. (Be sure to force them all together down into the paper tube -- as mentioned above -- while the glue is still wet.) -3D-printing 2 custom solid endcap fittings, to hold 3x of these tubes tightly together into a single assembly (and wrapping it around with tape) would provide for a ~$1.50, ~6oz, short strut that I estimate should hold 300+lbs up quite handily (and be much more resistant to lateral forces than a single one alone). --- That's it Kiwi. Cheap, lightweight, and readily-assembled from simple components; using no electricity to do so (not counting the materials manufacture, ofc). It has a hollow central core which could be used to provide ordered-passage access for actuation cables or wiring. And it's quite strong for linear compressions too! I was thinking of bird bones when I tinkered it all together like this, while casting around for a set of everyman robowaifu manufacturing approaches; both in cheap materials and simple, easy construction techniques. Good luck Anon. --- >addendum One other thing Anon. This lightweight strut's strong game is standing up to compressive forces. Can you think of any other design approach mentioned recently ITT that utilizes compression as one of it's two key structural forces? :^) >=== -prose, fmt edit -add addendum
Edited last time by Chobitsu on 03/10/2023 (Fri) 08:28:03.
>>21215 I sinned against Richard Stallmann and used Fusion360 for the sake of convenience. It's a very simple piece. I sketched out one profile for both pieces of the hinge, then drew a cutout in the larger piece.
>>21237 > Fusion360 I meant what kind and type of printer. I assume it's one that's liquid based with ultraviolet light hardener??? What does something like that cost and what brand? I really want a 3D printer. The ones that seem to work, or so they say, are expensive. The cheap ones don't seem to work. My thinking is I need some of the stronger plastics to prototype. I think I could build something that would print, the structure, but I think the software angle would defeat me. So much complication. So many details.
>>21255 It was just a plain Ender 3 Pro (~$250) with the stock Cura print settings. AKA a basic, budget printer with very few tweaks done. It only looks ultra-smooth because of the camera, but it's not a bad print by any means.
>>21256 >>21255 I should mention it's also a few years old at this point. You may want to look at the newer models if you want the most bang 4 buck
>>21255 >I think I could build something that would print, the structure, but I think the software angle would defeat me. So much complication. So many details. That's why we're dividing up the load Anon. One of us solves one thing, one of us solves another. We share everything here on the board, and together we all win! :^)
>>21255 What is your budget? There's many printers that work very well for ~200 USD. If you're interested in building one, I'd recommend buying a cheap printer and fixing it up. It's cheaper to fix something than to start from scratch. As long as the printer is using NEMA17's, it is possible to get great prints if you give your machine that care it needs.
>>21259 I'd pick a cnc machine over a 3d printer.
Thanks.That print does look really good. My budget is...non existent. I have a little money put back but I'm afraid to spend it because of the precariousness of my position and the really constrained future that seems to be coming. I could buy one but might find later that I wished I hadn't spent the money. I'll probably not buy one until I really have to have it. I looked at options to build thinking you could save money but I'm really not seeing that. Maybe you could get bigger build platform but with a lot of time and expense. I know things it would be useful for now. but...Sigh... The macro-economic signs are rally bad. It amazing the US economy has been able to stay afloat. The FED has recently raised rates such that they have never done before except...before the great depression in the 1930's. It's like they are trying to destroy us. Lower rates to nothing then when everyone loads up on debt raise them to the roof.
>>21262 If you can't buy, there may be a "makerspace" or public library with a printer you can use. I think there are also places where you can have prints outsourced for a low-ish price. As for software, there are a handful of free CAD softwares out there (both proprietary and FOSS) if you want to try your hand at designing without buying extra hardware.
>>21267 You can have custom stuff printed and then shipped. I don't know what I'd use a 3d printer for as that is plastic and I rather work with metal. Which is why I think a cnc machine would be better.
>>21262 Then I'd recommend you focus on the 'soft' skills portions of our domain. Things like CAD, 3D Animation, Programming, Propaganda, Writing, Arts. As you're well-aware we have plenty else to do here that doesn't involve manufacturing! Good luck Anon, work hard. :^)
>>21262 time to up the grindset. Live with your parents, live in your car, using the gym showers, use starbucks wifi. Maximize those profits and walk out rich after 6-12 months on the grind
>>21271 it is good to focus on all the small things, aside from crafting and art I have my nandroid animated show on the way this year
>>21273 Investing right now is not good and even if it was good you should not expect more than 20% a year... If times are good... Which their not...
>>21274 >aside from crafting and art I have my nandroid animated show on the way this year Neat! Looking forward to it Emmy-Pilled.
Open file (1.12 MB 672x574 boop.png)
>>21276 thanks, so am I
for people looking for ideas to joints, here is the original King Kong skeleton. I figured there were some good ideas in there
>>21267 >there may be a "makerspace" or public library with a printer you can use >>21269 >stuff printed and then shipped Good ideas I didn't think of that.
>>21149 I'm working on printing some parts, and also optimizing them for shorter print times and other things. I was concerned the poles might get into the way of the arms pulling the strings from above, so I made them a bit kinked. Also added some reinforcements on the bottom, to prevent the poles from breaking off. Hope that helps. Printing time per side is around five hours with high speed but high strength (4 perimeters).
>>21365 Looks interesting, how big is it? Seems like it would be huge if it takes 5 hours per side.
>>21371 No, not really huge. Both sides together are 25 cm, so one side is just 12.5 and the poles have 1.5 cm in diameter. I guess it takes a while because of the four perimeters. But FDM printing just takes some time, tbh. Some of my shells went up to more than two days with some options, or at least 19h with more reasonable ones, but then I cut it down to nine.
>>21365 Looking forward to seeing this all assembled NoidoDev!
>>21371 Can you give us all some tips on speeding up our prints Kiwi?
>>21389 Unfortunately, my printer just stopped working, and since I'm in a transition period, I probably won't fix it right now. I looked into it if it could be fixed easily, but it might be the bed. I'll go on working on the 3D files as good as I can and start sorting the ideas about my approach for the AI during the next two month, then I should have another printer (Anycubic Kobra) anyways.
I'm thinking if getting a dc power supply. Should I get the 30 volt 10 amps one or the 60 volt 5 amps one? The latter is a bit more pricey There's also a 120 volts 3 amps one but I don't think that'd be necessary right?
>>21411 I hope things all clear up for you soon, NoidoDev. >Anycubic Kobra What features made you choose this particular printer Anon?
>>21417 Why not the 30V one Anon? After all, there isn't much general need for anything beyond 24V inside a robowaifu, AFAICT. Hasel actuators notwithstanding. :^)
>>21418 >>Anycubic Kobra >What features made you choose this particular printer Anon? We're getting OT here. I just planned to buy a second cheap printer anyways, and my current one is a direct drive while the other one is a bowden style.
>>21419 I already have 3.3 volt batteries and the thing to put 1 and 2 batteries. I could just get the missing 9 volt battery one. I wonder if I need a dc power supply at all...
I would like to chime in on the subject of supplies. If you wish to contribute you'll have to spend some money. Here are some supplies that I think will be needed. A dc power unit, although not mandatory it is quiet nice A 2 in 1 multimeter oscilloscope is better value than them being separate I think. Sure the oscilloscope on those is kind of crappy but why do we need a super accurate oscilloscope anyways. You can also test faulty transistors with the diodes function. A 3d printer. You can get the cheapest one and most reviews say they're good enough but most people recommend the ender 3. A starter electronics kit, you might want to get extra jumper wires, you know the flat one, just get 580 of those. Maybe an extra big breadboard and a smaller one, etc... You'll also need arduinos sooner rather than later That should be good enough to start.
>>21442 Please do not buy an Ender 3 unless you're willing to learn, tinker, and commit to regular maintenance. They aren't build and use printers, they're build, calibrate and trouble shoot then use. There's a reason why there's endless tutorials related to them. For awhile they were best price to performance printer but, that was years ago. Now, you're much better of with a modern Kingroon, Anycubic, or Voxelab printer. You get more for the money with those. They work with very little setup and maintenance. Kingroon and Anycubic are my recommendation for noobs, they are very easy to setup quickly and print if you're getting started and produce high quality prints with little effort.
>>21450 Thanks for the recommendations Anon, cheers. :^)
>>21417 >Should I get the 30 volt 10 amps one or the 60 volt 5 amps one? Get the 30 volt 10 amp. There's a reason behind this. If I remember correctly there's regulations over 48V that are much more stringent. Under 48V much less. And I may be wrong about this gut it;s a factoid that I remember from somewhere. Ok I looked it up and there are codes above "60V" but most people use 48V because of voltage spikes. So at 48V you don't have to worry about that. Much over that and you can be killed by the voltage, under 60V not likely. SELV or Safety Extra Low Voltage are keywords to search for with this. Just my opinion but I think, or at least I am planning, to use every bit of 48V. Wire size between 48V and 24V is 1/2 for the same power. You also get heating effects from the wire. It turns out you can get close to 200 pounds of lifting power at 48V with a 22 gauge wire (high quality wire). I think we'll need this sort of force to lift people. Also semi-conductors go up in price per amperage which is double at lower voltage. Yet we do not want to go over 48V because of regulations relating to the susceptibility of a shock killing someone. I've been figuring power, wiring size and stuff like that and made some notes. I found some wiring on Amazon Bojack 20AWG Flexible Silicone Stranded Electric Wires $17 (5 Colors 19.69Ft Each Color) total 98.45ft This should be good for 20AWG 11A 528 Watts and the silicone covering gives you high temperature and long lasting wiring protection. http://www.kylesconverter.com/power/watts-to-foot--pounds--force-per-second That 1 W = 1 J/s and 1 Foot-Pound-Force per Second: Approximately 1.3558179483314004 Watts (SI). So 100 Watts to Foot-pounds-force Per Second = 73.7562 (73.7562ft.-lbs. /100W) * W = ft. lbs. or watts * 0.7375626651 = ft. lbs. So 528 Watts gives 389ft.-pounds force but remember we will likely share the power among several muscles so it needs to have some headroom.
>>21450 Thanks for the printer recommendations.
For power for prototyping you could get 4 lead acid batteries and hook in series. Used would be fine and have all the power you need. If you want new ones motorcycle batteries are lower cost if you look around. Here's a great site for new and used lithium batteries. Here's a link to one set I bought. They were for back-up power supplies and pack a huge peak current value. https://batteryhookup.com/products/super-beast-module-with-48-headway-38120-hp-3-2v-8ah-24v They're used but they mostly sat around charged all the time for backup.
Using the above battery wired to 48V and, "...Over an 8-hour work shift, an average, healthy, well-fed and motivated manual laborer may sustain an output of around 75 watts of power...." So you have basically enough power for 16 hours at that rate.
>>21489 Yeah I decided not to buy the DC power unit and just use the lithium 3.7 volt batteries I bought tbh. They're even rechargeable.
Okay so I just found out hasel actuators use dc not ac power... Which is bad since transformers output ac power. However, I also found out hasel actuators require low current which is good cause transformers also have lower current... If it does work I just think it'll be alternating between compressing the bag and decompressing very rapidly I think. But anyways I'll test that when I get back home. Surprised nobody else is testing this... :^)
>>21487 >Ok I looked it up and there are codes above "60V" but most people use 48V because of voltage spikes. So at 48V you don't have to worry about that. Much over that and you can be killed by the voltage, under 60V not likely. I once read up on a longer discussion on this. It's way more complicated than just looking at the voltage. How much body surface is exposed matters, same of how much someone was sweating, and the amount of energy that can be drawn from a source.
Open file (104.33 KB 1044x1392 TPUbody.jpg)
Open file (133.59 KB 1044x1392 TPUcrushed.jpg)
Open file (131.62 KB 1199x897 Brokebackwaifu.jpg)
95A TPU printed with PLA settings. 1 wall at .8mm with 9% infil. Hollow, only 4mm thick skin. Incredibly durable and the infill between two thin walls feels somewhat skin like. TPU makes unpleasant noise when rubbing against itself. This is my best version of this concept so far. Soft waifu's are possible but, further research must be one. I now believe that TPU skin over PLA or PETG bones is the correct option. This is being done because shortly after building a mannequin waifu, I accidentally destroyed her through careless snuggling. A waifu must survive aggression in bed to be worth making.
>>21507 Okay so I just thought of something. If I combine two alternating square waves it ends up being like one dc wave right?
>>21556 Tpu? I thought we were going for food grade silicon.
>>21556 Great. Seem to work surprisingly well, though it depends on what can of wrinkles one is okay with. Maybe my whole experimentation with teardrops and such maybe wasn't necessary? >>17630 - Though this was meant to a huge extend for the tissue deeper below the skin and therefore also for thicker parts, but also for a generally heavier build. >>21558 >we were going for food grade silicon. For prototyping this might be better, also the silicon could be added on top of it. If we want layers, maybe with sensors inside, then the body can't just be molded in one step.
>>21557 Well I just asked chatgpt and it told me that two alternating square waves is not like dc, it it is dc. How about that. It also said that if you input a square wave a transformer will output a square wave. So all I have to do is get a couple of non bulky high voltage transformer circuits, make a circuit that outputs two square alternating square waves and voila. The square waves can be generated with 555 timers although it's not perfect it'll do I think.
>>21563 I also asked asked chatgpt if I can use acetone+styrofoam for silicon casting and this is what it said. >Yes, it is possible to use styrofoam and acetone to make a mold for silicone. This process is known as "lost-PLA casting" or "burnout casting". I don't want to get hooked on that chatgpt though lol I know they'll start charging money for it this year...
>>21566 I would highly suggest for making "mother-molds" for silicone to use cheap casting-filler like boat-repair or liquid polyester as seen here >>19234 you can get entire kilo-buckets of it for cheap and it´s dry within 20-30 min (just make sure to be in a well-ventilated or outdoor area when using it)
>>21567 What could be cheaper than acetone + styfoam though
By the way chatgpt kept saying sex bots objectified women and was bad for marginalized groups. I'm like lol. First off what do "marginalized groups" have to do with any of this. Secondly women have no issue objectifying themselves either way. Ridiculous.
>>21569 the things I just listed, you can get them at fairly low prices and very high quantities
>>21563 Oh no I agree the body can't be solid, it needs to be hollow. Which is why the silicon needs to be poured and spread throughout the mold. There is a term for it but I forgot the name. It also ought to be separate the head, the torso, the 2 arm/leg sections, etc and later bonded together. I'm also wondering if silicon is flexible enough for the elbows. But I think it is specially if it's hollow. I thought of making the boobs and butt out of latex but I think we ought to keep it as simple as possible for robot waifu version 1.
>>21557 This is what a bridge rectifier does. >>21558 Others are, I'm still on the fence. Silicone has plenty of problems. >>21563 Just using infill can approximate the benefits of the teardrops. >>21565 You're making an AC generator. 555 timers can be used for such but are not ideal for such. Please look into this further, beyond ChatGPT. Going to a library is recommended. >>21566 This is dangerous and unecessary, please listen to Emmy-Pilled. They are objectively correct and have real world evidence backing up their advice in the Nandroid thread. >>21572 You are referring to slush casting. https://support.polytek.com/hc/en-us/articles/360046061134-What-is-slush-casting-and-what-products-can-I-use-to-do-this- Silicone is a generic term for a polymer which incorporates silloxane into its makeup. There are potentially endless variations depending on which copolymers are used. A polymers ability to bend is called plasticity. There are several copolymers that will enhance silloxanes plasticity, these are generally called plasticizers. Though I you provided no measurements related to this hypothetical elbow, I feel safe in assuming that the majority of silicones will work for your aplication as most have extreme plasticity and the elasticity to stretch over a similar arc that a human elbows skin is subjected to.
>>21573 I like the alternating square wave idea better though. I think the bridge rectifier is kind of wavy and there's a lot of power loss by doing it that way. I get the impression that alternating square waves is more efficient.
>>21572 >boobs and butt out of latex Why? Silicone is the most flexible and the medical one is probably the most skin friendly and probably the least dangerous in terms of micro dust coming from it (if there's any).
>>21576 Well seems like the more conventional way of generating dc from ac is a full wave rectifier I guess. Either way I think I'm getting ahead of myself for the time being. First I ought to see if the thing moves at all.
>>21578 Sorry I mean the bridge rectifier
Okay I'm back home and I just got done extra stuff(a lot which I didn't need I think but I don't like waiting for parts) Anyways here was one of my attempts... I think the current is just going through the aluminum foil and back into the end of the gator clips.
>>21613 Exciting times Anon! I think you'll solve things if you just don't quit. Keep Moving Forward! and get yourself a double-pair of rubber dishwashing gloves to use during your experiments! :^) --- BTW, seems we could already use a Prototypes and Failures #3 thread, OP >=== -combine posts
Edited last time by Chobitsu on 03/31/2023 (Fri) 09:27:59.
>>21616 Okay, that's a surprise. I'll look into it. I'd like to do something similar than before with the pics, but maybe a different background color and of course with the newer pics. Maybe one with only the new pics.
>>21616 Suggestions for the thread. If I don't react to the approval, you'll have to wait or make it yourself, since I won't be able to for at least a few days. Also, I can't change the collages after tonight (EU time).
>>21641 These are all fine, just use your own discretion NoidoDev.
>>21645 Thank you.
NEW THREAD NEW THREAD NEW THREAD >>21647 >>21647 >>21647 >>21647 >>21647 NEW THREAD NEW THREAD NEW THREAD

Report/Delete/Moderation Forms
Delete
Report