I’ve been trying to find a good robotics / cad / 3d printing community. I hope this is it.
So I’ve had this idea of building a sci-fi robot hand. I want it to look somewhat heavy, bulky, strong … but not necessarily very strong in reality.
I want it to look modular and serviceable and I want to use heavy hardware such as bolts and ball bearings instead of glue etc…
Been learning CAD (Fusion 360) designing and printing parts, going through iterations and I’ve finally arrived at … drummroll … at a somewhat acceptable design for a finger. Yay!
Ok so first I tried using 200lbs braided fishing line between the motor and the finger. It was sort of finicky and I couldn’t figure out how to tension it properly.
Also this design wasn’t that great so I did another iteration with a belt one one side and springs on the other. This seems to be working ok.
Also I needed to use a continuous rotation servo and add a pot that measures the servo because my 180 degree servos didn’t move enough.
My test rig is:
]desktop PSU for 12v/:m] ]adjustable DC-DC converter for 7.2v (optimal for my servos)/:m] ]arduino with a display and a pot for testing/:m] ]test “socket” for the finger with a servo and a 10 turn pot/:m]
It can lift a 0.5L water bottle with the fingertip but that’s about it even though the servo can move 20kg at 1cm distance supposedly. I guess that’s just physics unfortunately.
But I think it can hold much more … haven’t properly tested yet.
Also the springs subtract perhaps most of the power and very little is left for actually lifting things.
Btw you can tell it’s a finger because it has 3 joints. Other than that you need yo use a lot of imagination I guess.
I have more pictures and videos on instagram if anyone’s interested: instagram.com/lumikaarel/
Anyway next step is to design the main hand part with 5 sockets for fingers etc … even just trying to figure out where to put the thumb is sort of not obvious at all. To make it look good but functional at the same time.
I have some big 90mm bearings I may want to incorporate into the wrist etc…
I have limited time and I’m learning everything so progress is slow.
Update: playing around with some design ideas for the hand part that connects fingers to the wrist I quickly discovered that it’s going to be larger than my 20x20 printer bed so I’m now thinking about creative ideas how to split it up. Also I’m playing around with the idea of creating a 4th joint for the thumb that allows it to sort of rotate / move horizontally a bit. Otherwise I’m planning to reuse the same design as the fingers so there are going to be 5 identical fingers, 4 of them maybe staggered a bit compared to each other so they don’t look like a square spade and the fifth one positioned like a thumb with some extra freedom of movement.
I’m not into gluing so I need to split the hand in a way that complements the design and attaches with bolts and designing things in cad or even just 3D is very new to me. So this is a major challenge.
Welcome to the community! Visually, that’s a great looking hand, and from what you wrote, that’s the main objective. As you already discovered, the “physics” of the joints don’t allow for much lifting at the end. You are aware you may be able to 3D print the joints as well (SLA, DLP, SLS perhaps). Although not posted here, your “other” hand (the flat one with additional joints in the palm) looked quite simple and effective - what happened to that approach?
FYI - consider joining letsmakerobots.com as well.
Hi!
Thanks!
I believe you’re referring to a picture on my instagram. There are a couple of pictures of the early stages of an InMoov robot build. I actually got as far as building the hand, wrist, arm, bicep, shoulder and part of the torso on that one. But it’s not my design. Gael Langevin and his community have spent years designing that one inmoov.fr/gallery-v2/ and it’s awesome. But it’s not what I want to do. I’m all about the journey with this project and the end result is almost irrelevant to me compared to the learning, designing and building process. Instead it has inspired me to try learning CAD and designing a new robotic hand of my own that doesn’t compete with the great work that others have already done but is something different and maybe a little crazier instead.
To answer your question on what bacame of it: I had just assembled the entire thing but I got a little ambitious with testing. One of the threaded rods spun out of it’s socket, the arm with all it’s weight impacted the torso, the shoulder broke, the hand fell to the floor and broke further into 3-4 pieces. I found the weak link in the design. And I haven’t gotten around to fixing it yet. At this point I’m not sure I will.
Direct drive and it’s going to have to support considerable weight so fingers crossed.
It feels like serious business. At least I don’t want to get my finger stuck in there but will have to do more testing to see if it will move a 400-500g thumb … I may be asking a lot from it.
I just weighed it. The finger is 432g. The hand is going to be quite massive, literally. Which is not completely unexpected so I’m planning to use hardware like this for the wrist and elbow.
The other (rectangular) one will house the servo that swings the thumb about a bit. Made a bigger one that goes all around the servo and has some mounting holes at the back.
Although you’re 3D printing, there’s metallic paint which would go over it well if you want that end effect.
You can also use an acetone “bath” to smooth the 3D prints.
Acetone doesn’t work on PLA which I’m using though. It works on ABS only. I don’t want to rule out painting because that could be really cool and I already have some spray paint I might use but I’d probably need to sand everything a bit, fill, prime, paint, mask, paint again, weather etc and probably I’d need to take everything apart for that and reassemble. Sanding PLA is a real pain too. I’m tempted but I don’t know if it’s realistic for me.
True - didn’t catch that you were using PLA. Sanding… yeah, I hear you.
We actually use a hot air rework station to gently melt the PLA parts - it heats up in seconds, and the parts look quite nice. The more you practice, the more you know about the nuances / limitations.
I’m not sure I dare try but from your experience is it possible to heat treat small mechanical PLA parts like these without warping them?
Also what does your hot air rework station look like? The kind that is used for surface mount soldering? I’ve never used one. How would you go about finding a good one?
I’ve experimented with a hot air gun but it doesn’t feel like the right tool for this. Not enough control.
It tends to burn off very thin strands, and smooth layers. If you keep it pointed at a specific spot for long, it would certainly melt. It also makes great work of areas where support material have been removed.
The idea would be to smooth out the layers for a better finish, nothing more.
Unfortunately don’t have a before / after image, so take a look at: youtube.com/watch?v=cjujx7dlyOk
It’s certainly not the “best” tool for the job, and a quick search reveals there’s no “ideal” way yet.
First attempt at printing something for the palm. It’s kind of like a construction kit part or something and I’m not completely satisfied with the design but my hope is that since I’ve left some extra mounting holes I can design a cool cover for it which will rectify some of it. I sure hope the print doesn’t fail.
I haven’t yet designed the wrist but I left plenty of mounting points for it which I hope will suffice eventually.
So yeah. “hope” is the keyword here but I couldn’t find a way forward otherwise.
