/robowaifu/ - Actuators For Waifu Movement Part 3

Name
Subject
E-mail
Message	Max message length: 6144
Files	Drag files to upload or click here to select them Maximum 5 files / Maximum size: 20.00 MB

Spoiler images
Password	(used to delete files and postings)
Use bypass

Actuators For Waifu Movement Part 3 Kiwi 12/06/2023 (Wed) 01:18:16 No.27021

(1stl thread >>406 2nd thread >>12810) Kiwi back again with a thread for discussing actuators to move your waifu! Part Three! Let's start with a quick introduction to common actuators! 1. DC motors, these use brushes to switch the ferrous core electromagnets on a rotor to rotate its magnetic field relative to surrounding magnets! They're one of the cheapest options with an average efficiency range of 30 to 90%. Larger DC motors and motors with higher turn counts are more efficient. 1.5 Coreless DC motors, by removing ferrous materials, losses from hysteresis are almost eliminated, dramatically increasing efficiency to nearly 90% even in small motors. Eliminating the ferrous materials reduces flux focusing, resulting in weaker fields and higher speeds. 2. Brushless DC motors (BLDC), these use a controller to switch the electromagnets on a stator to rotate the magnets of a rotor! Without brushes, they have the potential to be more efficient with higher power density compared to DC motors. Their efficiency and behavior vary depending on the algorithm and sensors used to control them. Coreless brushless motors exist but are rare and only used for very niche applications. 3. AC motors, a wide and incredibly varied category. They all rely on AC’s frequency to control them. With single phase AC motors relying on shaded poles, capacitors, or some other method to induce a rotating magnetic field. 3 phase AC motors naturally have a rotating field which usually gives them higher efficiency and power density. Notably, most AC motors are brushless. The most commonly used brushed AC motor is the universal motor, which is 4. Stepper motors, brushless motors with ferrous teeth to focus magnetic flux. This allows for incredible control (stepping) at the cost of greater mass, subsequently giving them higher rotary inertia. Usually 50 to 80% efficient depending on control algorithm/speed/and quality of the stepper. Due to their increasing mass production (& ubiquitous low cost controllers), they have appeal as a lower cost alternative to BLDC motors if one carefully designs around them. 5. Coiled Nylon Actuators! These things have an efficiency rating so low it's best to just say they aren't efficient. (0.01% typical, 2% achieved under extremely specific conditions in a lab.) Though they are exciting due to their incredible low cost of fabrication, they’re far too slow and the energy requirements are nonsensical. https://youtu.be/S4-3_DnKE9E https://youtu.be/wltLEzQnznM 6. Hydraulics! These rely on the distribution of pressure in a working liquid to move things like pistons. Though popular in large scale industry, their ability to be used in waifu's has yet to be proven. (Boston Dynamics Atlas runs on hydraulics but it's a power guzzler and heavy) Efficiency varies wildly depending on implementation. They would work great for a giantess! 7. Pneumatics, hydraulics lighter sister! This time the fluid is air! This has the advantage in weight. They aren't capable of the same power loads hydraulics are but, who wants their waifu to bench press a car? (Too loud and inefficient for mobile robotics.) 8. Wax motors, hydraulic systems where the working fluid is expanding melted (commonly paraffin) wax! Cheap, low power, and produce incredible forces! Too bad they're slow and hard to control. 9. Explosion! Yes, you can move things through explosions! Gas engines work through explosions! Artificial muscles can be made by exploding a hydrogen and oxygen mixture in a piston, then using hydrolysis to turn the water back into hydrogen and oxygen. None of this is efficient or practical but it's vital we keep our minds open! Though there are more actuators, most are derivatives or use these examples to work. Things like pulleys need an actuator to move them. Now, let's share, learn, and get our waifu moving! Servos! These use an actuator and sensor to control motion. DC, coreless DC, and BLDC motors are the most commonly used actuators for servos in mobile robotics. Potentiometers or hall effect sensors are the most commonly used sensors to detect rotary motion in servo systems. >Soft muscles (pneumatic system) with origami-inspired skeletons: https://youtu.be/OJO4FP0DXgQ[ >Printed pneumatics (TSA can also be used instead of pneumatics for actuation.) https://youtu.be/_X0rDW6NQ58 >Using sugar as soluble support material for printing silicone muscles: https://youtu.be/L0Z0-y3qpNk >Cavatappi artificial muscles (hydraulic coiled nylon hybrid): https://youtu.be/yXAJGH5s4cs https://youtu.be/MpCFumHFZvU https://www.designnews.com/automation/cavatappi-robot-muscles-have-5-times-strength-human-muscles >Twisted string actuators (TSA) Be extremely careful when choosing your strings. They will be put under extreme strain and must be durable. https://youtu.be/N4VMoYFrusg https://youtu.be/hFuzQ4ed-t0 https://youtu.be/J26y1nn7JMM https://youtu.be/zYrHGMiqC9A https://youtu.be/PABVsuV7Y1M https://youtu.be/tP9B3aqc4CI https://youtu.be/Y1uceDzhjKY >Continuous ransmission (CVT) / torque converters https://youtu.be/kVPjhmTThPo https://youtu.be/cd2-vsTzd9E https://youtu.be/c9e2y-5DMNc https://youtu.be/PEq5_b4LWNY >=== -edit subj

Edited last time by Chobitsu on 12/06/2023 (Wed) 03:06:55.

Kiwi 01/06/2025 (Mon) 21:51:30 No.35396

>>35303 >>35389 Before I respond, could you please draw out your system? Frankly your description of how it looks and functions confuse me. >>35386 You continue to be one of my favourite posters. Keep doing what you do. Just remember your decimal points. I can edit them in if you'd like. (Unless you specifically declare you want the error to remain, I'm going to fix it. It really bothers me.) >>35394 >Green or blue? Buy heaps of blue's when sales hit :)

Mechnomancer 01/06/2025 (Mon) 23:36:54 No.35404

>>35389 Oops sorry I forgot to carry a zero lmao.

Robophiliac 01/06/2025 (Mon) 23:37:21 No.35405

>>35303 >>35393 >You could also use the sliding mount in the chest to recover speed. >Care to spell this out more clearly for the uninitiate? This comes from old-school model airplane servo installations, before the days of mixing channels or computers in transmitters. Back in the day, if we wanted to have a non-standard control surface arrangement, elevons, tailerons, flaperons, ruddervators, etc., we did it by having two servos mounted on a sliding tray. For example, ruddervators in a v-tail. One servo would have a double-arm with drive-rods going to the actual control surfaces, one to each. The other servo would move the tray. The double arm servo acting alone provided rudder function, moving the surfaces in opposite directions, and the tray-moving servo provided elevator function, moving the surfaces in the same direction. Acting together the two functions were mixed. Sometimes a single servo would perform the operation and slide the tray it was mounted on at the same time. This might be necessary when operating landing gear that traveled fore-and-aft, shifting the center of gravity as big wheels swung up or down. We could have the servo move in the opposite direction as a balance weight. In this case we might want the servo to move itself at the same time as the motion it's causing to augment (or retard-unlikely) that motion. Depending on the linkage, or power train- in this case a chain drive- we need to have a system for adjusting the length of the chain as the servo moves, a tensioner, or possibly a double tensioner. We would want the chain to be "pinched" to resemble a figure eight or infinity symbol when the servo and driven sprockets are closest together, and a simple oval when they are farthest apart. It's basically bicycle and scooter parts applied to 60 year old tech. Most likely the servo will be hacked for continuous rotation, and the feedback pot mounted in the robot's shoulder joint. The system for moving the servo could be another chain and sprocket, or a control arm, it depends on your design. The installation would depend on the available space, so you might not be able to add simple brakes or clutches to maximize the performance of the chain for direction of travel, if you even wanted to. KISS. btw, that chain and sprocket system is still used in real aircraft, so it's very reliable. >>26046 Search the web for online copies of old model aircraft magazines and articles about servo installations, or plans. There was a time when designers had one servo with multiple control rods open the landing gear doors, raise the gear, and then close the doors on a scale P-38 Lightning. There are probably lots of forgotten gems we could use in that old tech. >>35396 Kiwi, is this enough? Right now I'm up to my eyeballs in eyeballs: https://www.thingiverse.com/search?q=inmoov+eyes&page=1&sort=newest

Mechnomancer 01/06/2025 (Mon) 23:41:20 No.35406

>>35396 >Just remember your decimal points. Yeah whoops! lol. Just make an edit in parenthesis stating I forgot to carry a zero. My point still stands tho: mechanical disadvantage sucks >:V >>35389 >servos, sprocket chains and bowden cables Quit the jaw flapping and build it, man! :)

Kiwi 01/07/2025 (Tue) 00:13:44 No.35408

>>35405 I get it.

Robophiliac 01/07/2025 (Tue) 00:26:43 No.35409

>>35405 >>35406 And I made a mistake, corrected it, saw a typo in my post and fixed it, but reposted the mistake. >There was a time when designers had one servo with multiple control rods open the landing gear doors, raise the gear, and then close the doors on a scale P-38 Lightning. Should have been: There was a time when designers had one servo with multiple control rods unlock and raise the gear, and then close the doors, or reverse the process on a scale P-38 Lightning, with it's separate pod and booms. On the P-38 One of my favorite planes. How could I have forgotten? the doors stay open when the gear is down.

Mechnomancer 01/07/2025 (Tue) 00:54:35 No.35410

>>35389 The math ain't mathing. In order to get mechanical advantage you need r2 to be greater than r1, however you sacrifice speed/total movement for greater torque.

Robophiliac 01/07/2025 (Tue) 01:16:55 No.35411

>>35410 Now extend your drawing to the left. Add a "cable" from the bottom of the servo wheel to an anchor point to the left. Assume the servo is not fixed in position. What happens when the servo wheel rotates counterclockwise? This illustrates the theory. NOTE- NOT "bowden cable", but steel, ie; bicycle cable and steel sheath.

Robowaifu Technician 01/07/2025 (Tue) 01:34:58 No.35412

>>35411 so this thing basically

Robophiliac 01/07/2025 (Tue) 01:47:04 No.35414

>>35410 >>35411 >>35412 No. We are increasing the torque on the arm by increasing r2. This slows the arm movement. However, our servo is not fixed, but arranged to pull itself away from the shoulder at the same time. It is reeling in cable from opposite sides, reducing the length of the cable system twice as fast (if the cable pulleys are the same size). Since the shoulder is the "free" end of the system, the arm moves faster.

Mechnomancer 01/07/2025 (Tue) 01:47:34 No.35415

>>35411 Bicycle cables are called bowden cables :D What is the point of adding a second cable that does the opposite of the first? It doesn't multiply your forces or anything. >>35412 Even if one were to use a block and tackle the more pulleys you add the more you need to pull the cable for the load to traverse the same distance.

Robowaifu Technician 01/07/2025 (Tue) 02:00:21 No.35416

>>35414 so you want to reduce the load by increasing the distance but reduce the distance and increase the load? you cant have your cake and eat it

Robophiliac 01/07/2025 (Tue) 02:04:06 No.35417

>>35415 >Even if one were to use a block and tackle the more pulleys you add the more you need to pull the cable for the load to traverse the same distance. So close.... If it was a snake, it would have bit him. Weekley live stream starting so goodnight all.

Robophiliac 01/07/2025 (Tue) 02:09:01 No.35418

>>35416 The load is staying the same, the torque is changing because the load is being applied to the arm somewhere else. I'm out.

Grommet 01/07/2025 (Tue) 03:22:07 No.35419

>>35389 >180 kg/cm divided by 100cm = 1.8 kg Yes. My thinking for the lowest cost, cheapest method so far is you can have a fast lower power motor and use cables that wind up on a pulley, as noted, or you can have an electric actuator with a really long set of smaller forces all acting at once like a Switched reluctance linear motor. https://en.wikipedia.org/wiki/Switched_reluctance_linear_motor "If" the SWLM can be made economically it would seem to me to be the best choice. It's simpler. No big mass of cables, pulleys etc. to route and deal with. I also keep coming back to the notion of scissor jacks. I have not figured this out. I have some ideas, which will likely not work, but the basic idea is a take on an electrical scissor jack type mechanism but all electric. The big idea of this is to take less powerful forces, line up a huge mass of them and use tiny leverage to move large forces. Speed would be reasonable for our task because it's electric. I don't like hydraulics but...I think that you could do them way cheaper if you count the labor to assemble. If you used large cast bladders it would beat everything else on cost and rapidity of assembly. Power output, would not be to difficult to get fairly large forces. I have an idea for valves, which is the cost problem in normal hydraulics. Imagine an oval feed line to a bladder type actuator. On the feed line is a clamp held down by a spring. The clamp is held down but could have either a small solenoid or better, another small control bladder actuator that would take off the spring pressure, opening the valve, which would then feed fluid to the bladder actuator. Since the force is the pressure times the area. The same pressure could be used for the valve opening bladder, but it would have a large area compared to the small area that squishes the line closed. The opening function could be a small bladder that pushes a pivoting arm that releases the spring pressure, leading to fluid filling your main actuator bladder. The small "control" bladder could be actuated by a very small tube since not much fluid would be need to fill it and take pressure off the main actuator bladder valve control spring. Yeah I know hard to garble this out. The point being you have one pressure so that minimizes cost. The originating control could be very small electromagnetic solenoids, simplifying the electronics. So the solenoid opens a smaller feed line to control the bladder that opens the big flow line to the main muscle actuator bladder. Turn one on to turn another, etc. until you have a large flow into the main actuator bladder. The other, smaller, actuating bladders have very little flow so need less fluid, so are faster to act. All of your muscles, no matter how, many could be fed with one large supply line. The supply line would be connected to a bladder tank that is filled and pressurized by an electric motor. I could see you being able to have a large mold. Hang some sort of reinforcement net in it (nylon?polypropylene?)then fill with some sort of flexible resin, cure and then you have the large majority of all actuators in one big ass pour. Hard to beat that for time. Certainly soldiering thousands of connectors would take more time. But...I hate hydraulics. I have a picture, a little off, slightly different actuation and not drawn very well in this comment, >>28447

Grommet 01/07/2025 (Tue) 03:54:14 No.35420

>>35392 >>Why do you think you scale motor power by mass linearly? >Show me where I said that. >You can not. >Yes, I can, easily I don't want to sound like a dick but your fucked up, you talk like a fag and your shits all retarded. One is electric the other is human. I only used these numbers to show the difference in power to weight ratio. It says nothing about scaling. Nothing at all. Stop saying stupid shit that means nothing and is wrong. The more you say the sillier you look. 1. You confuse the power to weight ratio of electric motors and humans. Declare that this mean scaling. No it's a comparison of force. Not scaling. Scaling would be if I said that the power of an electric motor would be the same "per weight" in a smaller size. I did not say that. You are ether trolling me or have extremely poor comprehensions. Like really, really, really bad. 2. >Humanoid robots aren't comparable to humans This is important for us all to understand. We aren't robots, and robots aren't us Don't lecture me when you are so brain dead that you "can not comprehend" how to make an electric motor that is anywhere near 3 orders of magnitude in power to a Tesla motor. I might as well spell it because you may actually not know what this means. It's a 1,000 times difference. You can't possibly be so stupid as to believe that I do not know robots are not humans, or are you??? And since you should know this, why are telling me? You just look foolish and trollish. If that's is what you want people to think about you, fine but you're doing it to yourself. >Ignoring torque for watts >You cannot bypass physics You have no idea how foolish you sound. I suggest you use the torque-watts calculator I linked. I got a big laugh out of it. I got a flash back from some guy arguing about vaccines with me. You sound like him telling me,"vaccines are science", totally missing the point.

Chobitsu 01/07/2025 (Tue) 05:50:57 No.35421

>>35420 >The more you say the sillier you look. Lol. I'm all for the le ebin trole on the occasion, Grommet. But please remember we're all on the same team here. The more ad hominems you use, the more like an Eglin AFB jew you look. <---> IMO Kiwi (and the others here, including yourself) are all acting on good faith. Debate the numbers all you want, but please restrain yourselves from (in effect) degenerating into calling each other niggers, heh. Cheers! :^) >=== -minor edit

Edited last time by Chobitsu on 01/07/2025 (Tue) 05:55:49.

Grommet 01/07/2025 (Tue) 09:04:24 No.35426

>>35421 >IMO Kiwi (and the others here, including yourself) are all acting on good faith Is he??? Notice I explained myself, yet he would not leave well enough alone and kept lecturing me, making up things I did not say all the while getting the whole thing wrong and pretending he was talking down to me. I'm only dismissive and short with people that are that way to me. And Jews act the way they do to demean people who didn't ask for it. I only demean people who do so to me and will not let it well enough go. I asked him to stop giving me advice. I don't need his advice. It's wrong and he doesn't understand what I'm talking about and refuses to even entertain the notion that he could be wrong. "If" he would have asked me questions, like real ones, to clarify what I was talking about I would have readily done so but I have no time for dismissive people who do not know what they are talking about. The way he argues his point also really annoys me. It's Womanish. Passive aggressive. Like, kiwi,"I'm so offended, I was only trying to help". Please, don't piss on my leg and tell me it;s raining. (said in a whiny voice) >oh you need to learn physics >robots aren't humans > maybe some day you can be of use Give me a damn break. Fuck that, I'm a Man and my response is going to straight up tell him he is acting foolish.

Grommet 01/07/2025 (Tue) 09:21:03 No.35427

A good idea of some of the basics I'm striving for is this part I saw in a video looking at solenoids. This part 874F Push/Pull Solenoid Has 2,000 Newtons of force (449.61 lb. force). It has a small stroke (14mm) but can do so at 20 to 40 ms duty cycle. Now they of course are not doing this continuously but it does show what sort of forces "can" be had. So if you are able to spread these out a little, for heat, and have a lot of small ones the forces add up very fast. And because some people are so picky, I'm not saying "use this solenoid, I'm only using it as a benchmark as to what might be achieved. In real life what we would have would be far less but using many small magnetic attraction elements that would not matter. https://www.geeplus.com/push-pull-solenoids/

Chobitsu 01/07/2025 (Tue) 13:27:25 No.35429

>>35426 >And Jews act the way they do to demean people who didn't ask for it. I only demean people who do so to me and will not let it well enough go Fair enough. And please forgive me for the misunderstanding, Grommet. I'm not in any way implying you are one. No one so generous-hearted & giving of yourself towards our common goals here could be presumed to be so by any right-thinking man. One kind gives freely, and the other kind only takes. But rather simply that such behaviour looks similar on the surface. However, you have my full trust, Anon. :^) <---> Personal differences aside please, I'm basically asking everyone here to extend much grace towards one another, and to walk humbly with one another. Christ gave such an admonition to all of us (the so-called "Golden Rule", cf. >>3 ). If each of us fails to do so this year, then the laughter of our common enemies will be our only reward. Let it not be so! Cheers, Anons. :^) Onward! >=== -fmt, minor edit

Edited last time by Chobitsu on 01/07/2025 (Tue) 15:59:30.

Mechnomancer 01/07/2025 (Tue) 20:00:24 No.35449

>>35418 Just because the servo is further away from the joint doesn't mean the force decreases, energy cannot be lost from a system. If the servo reels in a cable from the shoulder and moves itself the sum of the forces on the servo still add up to the original load.

Robophiliac 01/07/2025 (Tue) 20:37:45 No.35450

>>35449 >Just because the servo is further away from the joint doesn't mean the force decreases, energy cannot be lost from a system. If the servo reels in a cable from the shoulder and moves itself the sum of the forces on the servo still add up to the original load. That's what I said I just used fewer words. : >>35418 >"The load is staying the same" Did you mean this to reply to RT at >>35416 ?

Robophiliac 01/07/2025 (Tue) 21:46:18 No.35454

>>35405 I found a video showing a form of the 'sliding servo mixer' that should make the concept clear. It may be a useful idea for some anon's robot. Maybe in the neck or spine, feet or toes? YMMV https://www.youtube.com/watch?v=h6GrhfN52lQ

Robowaifu Technician 01/07/2025 (Tue) 23:20:51 No.35458

>>35454 so its about combining x,y motion i think he already did this with the head

Chobitsu 01/08/2025 (Wed) 07:41:00 No.35469

>>35405 >>35454 Got it! Thanks Robophiliac. I love elegant solutions tbh, whatever the domain. Cheers. :^)

Robowaifu Technician 01/09/2025 (Thu) 11:35:29 No.35498

The whole bldc vs hobby servo conversation is somewhat dumb or maybe rather misguided, in my opinion. You're not going to be able to afford all off the shelf bldc motors in every joint. Even if you can afford it, not everyone can. Not only that, using all bldc motors is going to require much more custom power electronics (3 phase sinusoidal) and still requires some type of gearing, direct drive bldc is difficult because you need lots of power per motor (adding lots of battery weight) and they are very costly. Most will have to use off the shelf bldc sparingly where it makes sense and use hobby servos everywhere else you can. It's more of a using the right tool for the job argument. Also, hobby servos are noisy because of the high-speed rotation of their small internal motor and the gears meshing within the gearbox. High speed motor + friction + impacts between the gear teeth == audible noise. Ultra high torque brushless servo exist, but are just as expensive as a bldc motor and still have friction + gearing impact noise. My point of view is to create custom servos and iron-less axial flux motors for everywhere you can. More work yes, but significantly cheaper than good hobby servos or off the shelf bldc, and far better because you can tune the motor parameters, motor housing shape, and mounting surface to your specific robowaifu design. I personally think this is a much better use of time than trying to shoehorn in existing servos/bldc into a robowaifu design. Using existing servos/bldc seems to cause an extreme 'ulgy-ification' of robowaifu and more generally robots. Look at robo-one stuff to see what I mean in a general sense. These are designed with cost and function over form. A true robowaifu imo, should be prioritizing form, cost, and function in that order. Function's last because robowaifus should be modular and you can always modify your robowaifu for more function over time. Axial Flux > Radial Flux. Don't bother with radial motors for robowaifu if you can. See https://www.emworks.com/blog/electromechanical/axial-and-radial-flux-permanent-magnet-machines-what-is-the-difference for more information on why axial is the best choice. Gearing stages with either custom servos or bldc should be targeting toothed belts to drastically reduce noise over meshed gears. Maybe use planetary or harmonic, or cycloidal gears if you have spacing requirements. A bldc with encoder and gearing is just another phrase for a high torque brushless servo btw. I have developed a shoulder/hip motor that only costs 8.50 usd dollars per motor, cost varies based on filament, bearings, magnets, and hardware used. All you need are a bearing, few stainless nuts/screws, 3d prints, magnets, and double coated enamel wire. This is ignoring the encoder stage but that can be added on later in the gearing or on the motor itself. Also, once you have a well designed motor/servo, you can get the 3d printed parts machined out of metal which generally still works out to be cheaper than buying off the shelf stuff. I got quoted for 300 usd on a set of 8 machined motor's parts, it gets even cheaper if you go the cnc router/cnc mill/lathe route and has an added benefit of being able to sell motors to others or quickly recoup a lost motor due to negligence/accident. Quality bldcs are outrageously expensive, really good ones can cost 200 usd or more a piece, a few of them are in the thousands. If you take the cost of 8 of my motors with machined parts, it works out to be only slightly less then 400 usd. This is a drastic cost reduction. That's also why i went the cnc router/mill/lathe route. If you don't know how to make your own coreless axial flux bldc motor, refer to https://cadenkraft.com/designing-a-coreless-axial-flux-motor-part-1/ for a complete walkthrough. Hope this helps.

Chobitsu 01/09/2025 (Thu) 17:16:36 No.35504

>>35498 >See https://www.emworks.com/blog/electromechanical/axial-and-radial-flux-permanent-magnet-machines-what-is-the-difference for more information on why axial is the best choice. Thanks for the information, Anon! I'm glad there's a wide variety of somewhat differing opinions of robowaifu engineering topics going on here. This type of honest & healthy debate can only lead to a better outcome in the end for everyone involved here. <---> Please keep the good ideas coming, Anons! Cheers. :^)

Kiwi 01/10/2025 (Fri) 22:54:20 No.35549

Do you mean just using this motor? Or, do you have your own design? Need to know more before I can respond with anything of substance on your custom motor. When it comes down to the BLDC vs hobby servo argument, I don't think there is one. Both have their purpose as you noted. As for axial vs radial flux, they both have merits and demerits. It depends on use case. Radial flux are more popular because they are cheaper to mass produce. This is mostly due to material properties of the rotor. rotors go through uneven loading. A cylinder is more mechanically resistant to that uneven loading causing problems. Axial flux rotors are flat disks, they form waves more easily, meaning the material must have far greater rigidity to reach the same power output safely. Their high torque to volume/material at larger scales does lend them an advantage in mobility platforms. This is why when you look for motors, you'll mostly come across radial flux motors. I personally like them both equally, honestly wish that they were easier to mass produce so they would be easier to get at smaller scales.

Kiwi 01/11/2025 (Sat) 00:34:56 No.35551

Great introduction to the physics of counterbalancing. https://www.youtube.com/watch?v=OdmpYBSMWzM

Mechnomancer 01/11/2025 (Sat) 00:53:39 No.35552

>>35498 That certainly is a neat way to get your own motors. Definitely turn that speed down into torque tho. Then you just need to steal a control board off an ASMB-04b and you could control it just like a servo (not sure how much voltage it needs but the ASMB-04b does up to 24V). You can see the control board has separate connectors for the motor, motor power, and signal/board power. But being 3d printed it is probably going to be quite noisy. I might try building one eventually, but I'm focusing on just getting SPUD built fast n cheap as possible (without going mad, of course), and I think I'm doing pretty well at minimizing the uglification :) >>35454 Okay, now the concept is clear. Sorry about that :) Inmoov actually uses something kinda similar for the neck -which I modified for SPUD at one point, see vid.

Robowaifu Technician 01/12/2025 (Sun) 04:32:22 No.35570

>>35504 >This type of honest & healthy debate can only lead to a better outcome in the end for everyone involved here. Indeed! >>35549 >Do you mean just using this motor? Or, do you have your own design? I've designed my own. Though his is pretty good. I have designed three different sized axial motors (shoulder/hip/waist motor, knee/elbow motor, wrist/ankle motor), and two 'micro-ish' radial motors for neck and spine. The two radials are being redesigned as space is a premium and i can layout/fit some stuff better if they were larger diameter with thinner thickness and axial is well suited. I provided a link to that tutorial because it's well done. Largely my argument for using axial is because they can be easily diy'd at home since you can build them as air core, you can wind coils yourself easily with spools, you don't need costly arc magnets and you can pick and chose the kind of magnet you want, circular or bar. The axial flux efficiency/torque is largely just a bonus. The weight savings with axial is not insignificant either. >A cylinder is more mechanically resistant to that uneven loading causing problems. Axial flux rotors are flat disks, they form waves more easily, meaning the material must have far greater rigidity to reach the same power output safely. It largely depends on the construction, materials obviously, but the number of bearings and their placement, the coil/pole numbers, and your capabilities on making the airgap as small as possible. It's possible to remove a large amount of uneven loading from axial, it just requires some forethought. Like i said it's also better to get the parts machined out of some metal and use the 3dprints as a prototype. You can also stack coils and linked-stators in rotationally offset layers to further reduce uneven coil loading. I helped design a stacked and rotated layered axial pcb motor for school which uses large number of layers and it has low uneven loading and it has no bearings at all, just nylon spacers. There's a published paper on it, im sure you all can find it easily, I don't want to link it because I'd be doxing myself. >honestly wish that they were easier to mass produce so they would be easier to get at smaller scales. Personally after creating two different sized 'micro-ish' radials, using http://jystator.com/ I'd just stick with nema size motors if you can fit/afford them. It would be great if it radial were constructible at home, but it's not really feasible for most if anyone. Nema 8/11/17 are quite small and they can be gotten from China extremely cheaply for what they are. Sadly radial motors are quite heavy, for the same output you can see a 2 to 4 times weight reduction with axial, and you can use the weight savings for a better construction or live with uneven loading. The weight savings are extremely beneficial. Lower weight means less rotational mass in limbs and lower weight overall meaning smaller battery capacity can be used or exploit the savings for longer battery life. There is a reason most commercial humanoid robots today are using axial. >>35552 >Definitely turn that speed down into torque tho. Yeah toothed belt gearing on axial is the best way imo. my axials have the toothed belt gearing directly embedded on the outside of the rotor and the housing has spots for belt gears for torque stacking. >Then you just need to steal a control board off an ASMB-04b and you could control it just like a servo Interesting, that's a good idea. Does it support 3 phase? All my axial motors are 3 phase y configuration. i like doing the electronics from scratch as I can select components that will match the motor, I can select higher quality chips/components with the savings from diy-ing the motor, and I can move the high heat components to specific places for better cooling. plus a motor controller is pretty easy to make with a pico/pico2 with pio and the extra components i noted before. But i know not everyone wants to make everything from scratch. Personally everything from scratch is the fun part for me.

Chobitsu 01/12/2025 (Sun) 06:31:49 No.35573

>>35551 THANKS! Achieving reasonably-pleasing humanoid gaits (with our feasible, available, hips/spine design choices available rn) will certainly need a number of these counterbalance design approaches (notably, for the central battery-mass located proximal to the robowaifu's pelvis/belly center-mass volume) to enable the nice, arc-based motions+balances needed. Great find, Kiwi. Cheers. :^) >=== -minor edit

Edited last time by Chobitsu on 01/12/2025 (Sun) 06:37:25.

GreerTech 01/12/2025 (Sun) 07:12:31 No.35576

>>35573 Walking gets exponentially harder when the robot gets bigger, because of things like the square-cube law. It's easy for the Femsapien (the toy robot on OP's post), but for any human sized robot, it's going to be difficult. Note that I'm not saying impossible.

Mechnomancer 01/12/2025 (Sun) 14:53:36 No.35583

>>35570 >Does it support 3 phase? I thought there was something wonky with that DIY motor, unfortunately the board only supports bog standard DC motors. I'm not sure about the official terminology (I am orkpilled after all) but it seems bog standard dc motors are technically 3 phase motors, they got a funny little dingus called a "communicator" which takes the standard dc electricity via brushes that bush against multiple metal contacts along the shaft so it mechanically handles sending current to coils 120 degrees out of synch as opposed to using an external controller. Or something like that. Me ork. Gib motah elecktrisity it zoggin' zoomz. Simpul az. Amusingly, brushes are typically the first things to go on motors /generators, the most notable I've experienced being a car's alternator (at approx 120-140k miles). Replacing them aka "rebuilding" an alternator can cost between $5-$35 (turns out a pontiac 6 shares the same alternator brushes as certain models of Hummer!), as opposed to buying a new one at $100+. A little DIY can save you lots of money and eliminating brushes can prolong the life of your robot.

Kiwi 01/12/2025 (Sun) 22:54:43 No.35584

>>35570 >Axial flux actually makes sense from a DIY perspective I’m in manufacturing, so my biased understanding can obscure some potentiality. I’ll take your word for it that axial flux is generally a superior design for mostly 3D printed motors. I’d like to know more. Read a few papers but, overall I still feel as if a definitive design that is easily reproducible remains elusive. Partially because coil winding requires special jigs to maintain proper tension and alignment. I’d like for you to create a dedicated thread on home fabrication of axial flux motors. You seem uniquely qualified for such a task and I’d like to work with you on the subject in a thread where it’ll be easy for others to find and follow along. >Nylon spacers as bearings You could also use glass or ceramic spheres/beads to make thrust bearings. Works great with PETG. I’d run a track near the center and near the outer extremes of the coils. 3D printing small indents to place them in would be easy and provide great support. Perhaps the nylon spacers are better, I’d say both methods are worth comparison. >Shoulder and hip motors You speak as if these are direct drive implementations. Care to elaborate their dimensions/volume, mass, torque speed curves, and cost/estimated time to fabricate? Deeply curious about the potential. >>35573 >Humanoid gates Walking like a person is sadly far away still. Still, Good to have eyes on the prize. >>35576 >Easy for Femisapien to walk That same walking gate can easily be scaled up well beyond human sizes. It’s just a simple mass shifting statically stable design. She moves some mass to adjust center of gravity from her midpoint to being over a foot. The design has some aspect that allows the rest of her to tilt, this can be done with slight slants or springs in her feet, springs or loose joints in her hips, or a more complex internal mechanism, which causes her other foot to be off the ground, enabling a step. Femisapien uses springs in her feet and hips, so when she twists her trunk with her arms splaying in and out to adjust her center of gravity, her legs naturally tilt due to gravity, with one foot hovering a few mm off the ground. Her thighs then move her lower legs back and forth, causing her to step. What makes Femisapien especially clever is that she distributes her mass via twisting her trunk and arms which puts less load on her motor relative to leaning. When scaled up to human scale, I’d recommend having an eccentric mass, (perhaps her breasts?) to cause her to move her center of gravity. This would reduce how much mass the arms would need to move. Though, the issue would then be giving her gigantic feet to maintain stability. Something like Nana in picrel. I’m a tad obsessed with mass shifting mechanisms, gravity is the most reliable power source. The most reliable actuator is the one that’s not there. No one has ever had a problem with Femisapiens hip and ankle tilt actuators.

GreerTech 01/13/2025 (Mon) 05:47:35 No.35590

>>35584 >Though, the issue would then be giving her gigantic feet to maintain stability. Something like Nana in picrel. That's where the square-cube law will strike HARD. Imagine walking in lead-filled snowshoes.

Chobitsu 01/13/2025 (Mon) 20:26:07 No.35604

This thread is really paying off for it's general conceptual approach rn. >>35549 >honestly wish that they were easier to mass produce so they would be easier to get at smaller scales. This. >>35552 Good luck with your neck actuation assembly, Mechnomancer. It's looking good so far. >but I'm focusing on just getting SPUD built fast n cheap as possible (without going mad, of course) Lol. >and I think I'm doing pretty well at minimizing the uglification :) She's charming! One day you'll look back and enjoy the 'family album/growth chart' of dear SPUD and see just how far you've actually come with her, Anon! Keep moving forward. >>35570 >The two radials are being redesigned as space is a premium and i can layout/fit some stuff better if they were larger diameter with thinner thickness and axial is well suited. Exactly so, and this set of priorities is in play for several actuation sites around the robowaifu. Her hips, for example. >The weight savings with axial is not insignificant either. Low mass is arguably our single-most important design philosophy bullet point at this phase of our history (cf. >>4313, et al). Thanks for all the great information in your post, Anon. >>35576 >Walking gets exponentially harder when the robot gets bigger, because of things like the square-cube law. The basic reason we need to focus on the square-cube law is generally because of the (usually) concommitant increase in mass, as the volume cubes (see link above why this is a critical issue rn). If we use every trick in the book to keep our robowaifu's mass down, then it's simply a matter of the increased mechanical forces required for the longer limb 'levers'. A more-delicate balancing act, if you will. :^) >>35583 >(I am orkpilled after all) Lolwut? :D >>35584 >I’d like for you to create a dedicated thread on home fabrication of axial flux motors. You seem uniquely qualified for such a task and I’d like to work with you on the subject in a thread where it’ll be easy for others to find and follow along. This. All in favor of this idea, personally. >Still, Good to have eyes on the prize. Yup. IIRC, you yourself were the one who posted images here on the board of the prototype that had a mass located inside the torso volume that swung back and forth to provide the needed counterbalancing? My position is simply that we design a suitable system to use the high-mass batteries to swing back and forth instead of deadweight. Make sense? >What makes Femisapien especially clever is that she distributes her mass via twisting her trunk and arms which puts less load on her motor relative to leaning. Her designer is quite brilliant, IMO. >The most reliable actuator is the one that’s not there. This. The same is quite true of software as well, btw. :^) >>35590 >That's where the square-cube law will strike HARD. Imagine walking in lead-filled snowshoes. I think there's some subtleties there, GreerTech. The Square-Cube Law [1][2][3] is formulated regarding volume. For meatspace things like elephants, blue whales, and birbs, this roughly-translates directly into increased mass -- which is our general nemesis. But notice the last example, birbs. God designed them with a very low mass bone structure, and same for their primary lifting mechanism: feathers. As long as special tricks are used to keep the strength up and mass low (as with birbs), then the square-cube isn't quite the problem it would be otherwise. In Kiwi's example case, I'd suggest the concept is kind of like this instead: imagine you're walking on a pair of stilts, with 2'/60cm wide hoops affixed to the bottoms. While the results may be somewhat awkward (perhaps even comical), they wouldn't really require a significant increase in force required, but rather in the finesse involved. With the vase-printing approach (particularly if the 'mesh shell' paradigm [4] is used), her extra wide 'calves/feet' shouldn't really represent a significant increase in mass. Make sense? --- 1. https://worldbuilding.stackexchange.com/questions/215851/consequences-of-an-altered-square-cube-law 2. https://en.wikipedia.org/wiki/Square%E2%80%93cube_law 3. https://www.wolframalpha.com/input?i=cube+calculator 4. (cf. >>16170, >>16171, >>16525, et al) >=== -minor edit -add 'vase-printing' cmnt -add 'mesh shell' cmnt/crosslinks

Edited last time by Chobitsu on 01/13/2025 (Mon) 21:36:17.

Mechnomancer 01/14/2025 (Tue) 00:14:32 No.35618

>>35584 I kinda tried similar to the femisapien walking mechanism least year, not easy as it looks (especially when manually controlling the servos and some of them are the wrong way around!). That's why I'm gonna beef up a smol robodog with some ASMC-04bs and try to mount SPUD on the front of that. ASMC-04bs are pretty good in they have considerable torque (180kg/cm) and they control just like a servo except for an extra set of terminals for actual motor power. >>35604 >Good luck with your neck actuation assembly, Mechnomancer. It's looking good so far. Oh that is from like 6 month ago and I don't advise using that mechanism. The screw-drive takes too much current.

Kiwi 01/14/2025 (Tue) 23:50:35 No.35644

>>35590 >Square cube law More like square cube "guideline." That "law" assumes the part is 100% solid and scaled with complete uniformity. Reality can be whatever you want it to be, I can make the same part scale linearly with respect to volume and mass. Just need to get creative. In our case, we actually need to get creative because we are focusing on a large volume, low mass machine. >>35604 >Shifting internal mass That takes me back. Posted about that in 2017, the year I first got into this endeavor. Back then I was considering a wheel with bolts that'd work similar to old gyro trains. I'd say having a battery pack where the heart is, eccentrically connected to a shaft would be a far more clever, efficient, and practical. I have an itch to test this now. Got the ol' noggin' joggin'. >Femisapiens walking scales in processing "finesse" more than power/force required. Eloquently put, couldn't of elaborated better myself. >>35618 >Ork Femisapien legs You'd need an Ork psychic gestalt field for that to work. To put it seriously, your ankles and hip tilt shouldn't be powered. Your PVC has far too much play, cancelling out any potential height differential the legs could have. Most importantly, you're missing her entire body. The one part that is supposed to impart a height differential via manipulating her center of gravity. Your swing servos and parallel mechanism are correct. Your feet are also well thought out in terms of size and placement. It's a clever design, just needs a body, move her servos around, and find a way to reduce the play in her legs. Currently your design is similar to TecFoot from David Buckley. You may enjoy his work. http://davidbuckley.net/DB/A_Minimalist_Approach_to_Biped_Walking.htm

Edited last time by Kiwi_ on 01/14/2025 (Tue) 23:53:09.

GreerTech 01/15/2025 (Wed) 01:00:47 No.35647

>>35604 >>35644 I see. Ironic that I failed to account for hollow designs, considering my design is mostly hollow. Most of my theorizing was based on the RoboSapien toy I had, there the walking and balancing was achieved by having giant feet with the relatively heavy batteries in the feet, hence the "lead-filled snowshoes", plus my own experimentation with building tall robots.

GreerTech 01/15/2025 (Wed) 01:02:31 No.35648

>>35618 Reminds me of the first Honda robot prototypes

Mechnomancer 01/15/2025 (Wed) 11:12:12 No.35654

>>35644 >You'd need an Ork psychic gestalt field for that to work. You also didn't notice the ankles were tilting the opposite way they should be :D I was only using 3 servo channels to control 6 of 'em. Frankly I'm amazed it even woobled. Springs might work well for shifting the mass of a smol robot, but a large robot's walk cycle could be easily upset if it was carrying a load or accidentally bumped which is why I tried powered. But once again why reinvent the wheel when you can yoink from other products lol

GreerTech 01/15/2025 (Wed) 14:13:34 No.35656

Here's some videos of the Femsapien and Robosapien walking https://youtu.be/IfyB558ItD8?si=0idXfxuLnUVGrjsu https://www.youtube.com/watch?v=4N0pERu9gQ4

Chobitsu 01/15/2025 (Wed) 21:15:54 No.35658

>>35644 >gyro trains Lol, that's very cool-looking Anon! :D >I have an itch to test this now. Got the ol' noggin' joggin'. Please do! I think a single actuator that drives a pushrod in either direction should probably suffice. It's what I'll plan the control interface for. Cheers, Kiwi. Maidcom Mini's coming right along! :^)

Kiwi 01/15/2025 (Wed) 21:29:36 No.35662

>>35647 We're all learning, I've made far worse assumptions. Wish I had someone to correct me in the early days. >>35654 >Powering extra DOF for safety Sensible, I can understand how that makes sense to ensure she doesn't topple over when dealing with large inertial disturbances. Circumferential Flux Motors A new kind of motor that is essentially a rotary solenoid. Coils produce a flux field that imparts kinetic motion within a magnet, see gif for gist. This would potentially allow for a motor that has nearly perfect efficiency with the dirt simple control of brushed DC motors. The only site I found with information on them; https://www.motion-robotics.co.uk/cf-motor How do they interface with magnets inside the coils? Their coils are wound as a C-shape. I'd make one by having magnets with teeth which snap together to form an annulus/ring gear for a planetary gear train. Sets of coils would be wrapped around the annulus with space for planet gear meshing. The planets are held in place on the carrier to prevent clashing with. the coils. (Carrier not in picrel because it looked weird.) Sun gear providing the final output for the system. I put a hex connector because I like them. How would you design or implement a circumferential flux motor?

Chobitsu 01/15/2025 (Wed) 21:44:16 No.35664

>>35662 >gif related Wow that's wild. What's the little 'tube car', some kind of battery or other? >How would you design or implement a circumferential flux motor? Could we co-locate 3 of these together into a tight volume to create a universal direction ball-actuator in the hip joints, ala Sukabu's designs, et al : (cf. >>35663, etc.)?

Kiwi 01/16/2025 (Thu) 00:10:15 No.35670

>>35664 It's an AA battery with neodymium magnets at the front and back. Those magnets function as both brushes to induce and electromagnetic field in the copper coil, and they act as the motive element by "pushing" against that induced field. An ingenious machine, essentially a flexible solenoid with power held within the armature. >Using three for her hip I very much intend to make that happen! Though I want to make it by end of year, it'll likely take at least 2 considering MaidCom 1 is primary focus.

NoidoDev ##pTGTWW 01/16/2025 (Thu) 00:28:50 No.35673

>>35662 That was something I saw long time ago, but I overlooked that the "train" was a battery, and I thought I could use it to pull "muscle" strings to move robot parts. But it turned out that it was too weak: >>4429

Kiwi 01/16/2025 (Thu) 00:56:25 No.35676

>>35673 Just need to calculate based on your needs for your solenoid. https://calculatorultra.com/en/tool/solenoid-force-calculator.html#gsc.tab=0

Grommet 01/16/2025 (Thu) 09:37:34 No.35695

>>35498 Great comment. I've thought about Halbach arrays but didn't think about them in a coreless usage. The paper got me to thinking.

Report/Delete/Moderation Forms

Delete

Password Delete only files (Removes the file reference to the posts) Delete media (Removes the saved files from the server)

Report

Reason Global