Second Degree Hand

by orgemd, published

Second Degree Hand by orgemd Apr 16, 2014
1 Share
Download All Files

Thing Apps Enabled

Order This Printed View All Apps



Liked By

View All

Give a Shout Out

If you print this Thing and display it in public proudly give attribution by printing and displaying this tag.

Print Thing Tag


Thing Statistics

18181Views 2651Downloads


This is a mechanical hand made for users with a functional thumb. It includes an opening that allows the user to utilize their thumb along with mechanical fingers.
This hand includes a couple of unusual features:
First, this hand allows for two degrees of freedom in the motion. Most similar hands only allow the user to raise or lower the fingers. This one also allows the user to move the fingers side-to-side. A demonstration of the two degrees of freedom can be seen at:
Second, this hand includes springs on the tuners. This resolves some issues so that the hand has dramatically improved grip. A demonstration of this feature can be seen at:
This design is also fully parametric OpenSCAD files. Although there are STL files attached, they are just for example purposes. The intention is that each individual will create their own unique STLs from the SCAD files.
There is currently no option for a mechanical thumb in this design. There are a couple of other hand designs that include a mechanical thumb. If someone really wants a thumb added to this design (maybe because of the two degrees of freedom) I can maybe add it.
I am not marking this one a work in progress. The one I was making for my sister is pretty much complete and working well. I am working on an updated version, but that will likely be a new version 3.
As always... This is provided in the hope that it will be beneficial. It is provided, however, without any warranty of any sort. Use at your own risk, and seek appropriate medical attention. If you do use it, and you experience any problems, immediately discontinue use.
For more information on where to connect with people who can help someone in need with sizing, printing and/or assembly of a mechanical hand please join the e-NABLE Google+ Community entitled, e-NABLE:
For more information and stories:

UPDATE Version 3 7/23/2014
I have preserved all of the old files for anyone who was in progress or prefers the old version. Those are all in the V2 zip file. The new files are all for version 3. Version 3 was mostly tweaks to the original design. Some notable changes are:

  • I reduced the number of parts to create the main hand from 7 printed parts to 4 printed parts while improving strength across the board.
  • I created an option in the finger files to insert small bearings that make the motion a bit smoother.
  • I changed from printed pins or various screw mechanisms to 3mm PLA.

I find the updated version to be much stronger than the previous version and to function more smoothly.

Most of the pictures are still of the previous version. I did, however, include a bunch of pictures that show the new finger joints. I'll try to get some pictures of the overall hand. I know, however, that some are waiting on this version, so I wanted to get it posted.


Save all of the .scad files to the same directory. Customize the OpenSCAD files. I generally increase the number of elements under the Advanced Preferences to about 64000. Most of the parameters related to the hand are in the Include file. The hand and finger files allow you to select which components to print. The gauntlet file has a bunch of options. Printing the parts from the fingers file, the hand file, and the gauntlet file will give you all of the needed printed parts.

All Apps

Auto-magically prepare your 3D models for 3D printing. A cloud based 3D models Preparing and Healing solution for 3D Printing, MakePrintable provides features for model repairing, wall thickness...

App Info Launch App

Kiri:Moto is an integrated cloud-based slicer and tool-path generator for 3D Printing, CAM / CNC and Laser cutting. *** 3D printing mode provides model slicing and GCode output using built-in...

App Info Launch App
KiriMoto Thing App

With 3D Slash, you can edit 3d models like a stonecutter. A unique interface: as fun as a building game! The perfect tool for non-designers and children to create in 3D.

App Info Launch App

Print through a distributed network of 3D printing enthusiasts from across the US, at a fraction of the cost of the competitors. We want to change the world for the better through technology, an...

App Info Launch App

Quickly Scale, Mirror or Cut your 3D Models

App Info Launch App

Treatstock is an online platform that offers decentralized manufacturing services such as 3D printing and CNC machining for business-to-business and business-to-consumer sales all over the world. W...

App Info Launch App

3D print your favourite design with NinjaPrototype, a professional 3D manufacture with consistent quality and speed.

App Info Launch App

I'm having an issue with scad when I input different value a part is losing its shape!!! Its the MAIN CUFF part which have 2 parts, the one that sits on top of the main cuff where the strings are fed to the guanlet. Any help/ why this possibly can be happening would be great. Pictures attached. Thank you

Also parts come out with errors(EASY LIMITSTATE fix)which is not as big of an error like the part in the pictures.

Hello again,
I was wondering if you could make a parts list or direct me to one if you have it? I'm in the process of printing this and need to find the hardware or similar parts. Thank you!

Hi skeletaldragon. I am afraid I do not have an actual bill of materials I can point you to. I did not create one when I made it because a lot of it was tweaking as I went, and then I gave the end product to my sister without thinking to record everything. If you look lower in the comments, however, you will find a comment from me where I spell out most of what I used along with info like where I got it. I created that much closer to when I made the hand, and I think it is pretty thorough. The biggest weakness is that I do not succeed in listing exact screw lengths, but I think it will still give you a pretty good starting point. You will also find information like how I did the finger rivets. Read through all of that and see if it helps.

One of the big strengths of this design is that you can customize it to fit your needs. Because it is OpenSCAD, you can adjust the sizing and configuration in a great many ways. When you do so, the holes will still be M3 size (if you so choose)More c but the lengths needed will change. Because of that, it is hard to specify exactly what size hardware any builder will need - it depends on how you customize the design. I have tried to comment the OpenSCAD files a lot to provide information on what the various parameters do if you need to customize.

The strengths of this design are:

  • It is more readily customizable than pretty much any similar design. It uses open source software that anyone can get and use. It provides a great many out of the box customizable parameters, along with comments to explain them. I is easily changed outside of the built in parameters by someone who knows OpenSCAD.
  • It provides a second degree of motion at the wrist allowing for a more natural and unrestricted motion. I have not seen any other design even to this date that includes such a capability. My original design was closer to the ones you mostly see today where the wrist can only move in one way. I once had frozen shoulder that developed from imobilization after an injury. That made me concerned about the long term impact of limiting the range of motion of the wrist. My design tries to avoid any such limitations.
  • It is made to be more like a normal hand with different sized fingers. When I started making a hand for my sister, they all used fingers that were one uniform size, It worked, but looked very unnatural. I didn't like that and didn't see any reason why the fingers couldn't be sized to roughly match the users other hand.
  • The finger joints are smoother and stronger than most other designs. The bearings and filament rivets are almost unbeatable in this regard.
  • This design includes an adjustable adaptive grasp mechanism. My original designs would have problems with grasping. When the fingers were closed, the first finger to contact the object would stop the others from moving. The other fingers would not conform to the shape of the object - producing a weak contact. I think the design I have on this hand is probably better than most at solving this problem. Some other designs include an adaptive mechanism. I have not tested to see how effective they are. They do not seem to be adjustable.
  • This design allows a separate adjustable and replaceable elastic tension for each finger. Most similar hands use one elastic for multiple fingers (usually two fingers). The problem with such designs is that often one finger will return properly, and another on the same elastic will not. If the fingers are different lengths, it is that much more important that each one be independent. Also, the elastics are a weak point and fail. When they do, ease of replacement is important. I once replaced the elastic bands in my sister's hand in less than 5 minutes.

Where this design is weak:

  • I do not provide an option that includes a mechanical thumb. My sister had a thumb - so my design only includes the four fingers. I expect it would be possible for someone to add a thumb, but I have never done so.
  • Many of the components are not standardized. For example, elastic is not standardized on pretty much any factor. The amount of stretch, the resistance to stretch, the size, etc. It is very much a trial and error to find a good component for some of those elements - elastic cords, nylon strings, springs. That is probably the biggest issue.
  • Although much of the design is parametric, the screw length is not included as controlled. That means that you often need to cut the screws to the proper length. Because I recommend using stainless steel screws, cutting them is not a small task. You can do it with a dremel tool, but you probably need a diamond cutting disk and need to exercise safety.
  • Although the finger joints are much smoother and stronger than any others I have seen, they are also much more of a pain to create, and involve more expense. You need a handful of small bearings - so it increases the hardware. Creating the rivets can be a bit hard. It is doable, but it takes some effort - and you could mess some attempts up and have to reprint the parts. By comparison, other designs just use printed pins. That is simple to make and simple to replace, but doesn't work as well. My files do provide that option - but I don't like the results as well.
  • The adaptive mechanism adds considerable complexity to the design. It provides more areas where things could fail. It also makes a bit of noise as the springs move over the screws.
  • It is not easy to replace a failed nylon string. The one end is knotted in the tips of the fingers that are covered with rubber. I have replaced them, but it was not easy. I used an exacto knife to cut the rubber out of the tip just over the knot, and then extracted and replaced.
  • Although I tried to make the finger grip good by dipping finger tips in rubber and using 3M gripping material on other surfaces, it is not like a human hand. What it really needs is soft, deformable pads on the fingers. Your fingers squish around an object and increase surface contact. This design doesn't do that.

I will add that my sister does not really use this much - it ended up being more of a novelty. The reason why is fairly simple. By the time my sister was an adult, she had developed adaptations to do the things she wanted to do without having a full left hand. For example, she played softball. She wore a glove on her right hand to catch, then would flip off the glove and grab the ball to throw. She could do it very fast so that there was little to no difference in what someone with both hands could do. She wanted a hand, but found it didn't really let her do anything that she hadn't already found a way to do.

Anyway... Take a look at the comments below. I think you will find most of what you need there.


Anyone else having issues with OpenSCAD? It keeps telling me it has 1000+ warnings and can't generate the SCAD files.

Hmmm... I haven't worked on this for a while, but I looked at the files and it seems that I changed the names at some point and didn't change the include statements. I will re-post the files so that they expect the correct names in the includes. Give me like 15minutes to do that, then download again and give it another try. With that change, they all work for me still in OpenSCAD. If you are still having problems, let me know which file you are trying, and what sort of errors.

Thank you kindly for the changes! Everything works much better now, this is the only issue. When editing the Hand file, one warning comes up: "DEPRECATED: The assign() module will be removed in future releases. Use a regular assignment instead" Other than that it works beautifully. Much appreciated.


I started printing this and once I reached the 3rd stage realised that its a left hand , do you have the designs for a right hand version?


I do not have STL files for a right hand, but you will find the OpenSCAD files posted. (OpenSCAD is free, opensource, solid 3D modeling software). The SCAD files will allow you to make a large number of customizations. The file V3-Include-HandParameters.scad has, for example, the line "Left_Hand=true;" If you change that to "Left_Hand=false;" and generate the parts, you should get a right hand. Most of the key settings are in the single HandParameters file, but there are also some parameters you can set within the individual files - Fingers, Gauntlet, and Hand. I have tried to include a lot of comments within the files to describe what each parameter does. Hopefully they are sufficiently clear. You will want to download all six of the scad files to create your own STLs.

When I was making this for my sister, I made every version out of PLA. That worked just fine. I am not aware of her ever running into a situation where the PLA was a problem. If I was printing today, however, I might choose PETG or Nylon instead - just because prints made from them are just so remarkably tough.

I hope that helps. I realize that using OpenSCAD might be a bit new if you haven't seen it before. It does, however, allow for parametric adjustment of parts - which was critical to this project. The learning curve to be able to open the scad files, change some parameters, and regenerate STLs is not horrible. I can't see it while commenting here, but I think I might even have provided some brief instructions on how to do it with the main thing text.


I received some questions that I thought might help others - here are my responses:

What is the knuckly joint material and what is the purpose of the water?
The knuckle joints are made with a piece of standard 3mm PLA filament. You will find numerous posts on the Internet from people explaining how to make rivets out of small pieces of filament. The technique generally involves heating a metal part, tapping a bit to shape the head of the rivet, reheat the metal, tap some more, and keep working until you have a nicely shaped head. Then repeat the process on the other side if it is a two headed pin like used in these finger joints. I had issues with that. First, I found it to be very labor intensive to make a nice pin. Second, I found that PLA tended to stick to the metal tapper and ruin the rivet. I had less than a 1 out of 10 success rate for making them that way. It would probably work better if I was using ABS filament, but it would still be a pain and the metal would need to be kept at a higher temperature to make it work. Because of that, I came up with a different approach. First I cut the filament to just slightly more than needed. Then I used files or sandpaper to clean up the ends - so that they would be more likely to press into a nice, even rivet head. I then started the sink running cold water. I heated the metal to slightly on the hotter than needed side (so that it could fairly quickly melt PLA.). I then brought the finger with the rivet filament already inserted into position close to the sink, with the hot metal in my other hand. With a single motion, I pressed the PLA rivet head completely flat and then immediately quenched it under the cold water. That instantly froze the PLA in a nice, clean head. I tried hard not to move the metal or filament relative to one another while putting it under the water. That is, once the metal was pressed in, I kept it right where it was until the plastic was quenched. Once quenched, I could easily snap the metal off the solid rivet head, and repeat on the other side of the finger. Since the fingers are PLA as well, I included a slight indent for the rivet head to fill. I basically just push it flat and call it a day. That technique worked for me just about every time - I think I had one fail. Although you might think that a piece of PLA filament is not good for this, it is actually very strong - much more so than a printed pin like used in many other designs.

What are the spring tensioners made from?
The spring tensioners are just springs and washers on long screws inserted into the plastic parts (in the box on the upper rear of the gauntlet, and through to the printed tuner rods that you have to tap to accept the screws. The springs are a bit of a problem - one of the weaknesses with the design. I could not find any other satisfactory way to accomplish what they provide, so that is what I went with. I spent a lot of time at the local hardware store poring over their spring selection and then bought a bunch of any size that seemed like it would be about right. It is a balancing act. You want the springs stiff enough that they do not compress when just closing the fingers in open air, but light enough that they will start to compress when there is resistance against one or more fingers. That is really the key - the force needed to compress them. You want them to be somewhat small in diameter so that they will be trapped on a screw and not interfere with one another. For length you want them to be long enough to reach all the way through the spring box, but you can cut them with a tin snips if they are too long. I don't remember the exact springs I used, I am afraid - and I am not sure the same ones would work for you unless the size you are building is the same. If I find something saying the exact spring, I will see about adding it for you..

What type of cord is used?
For the cords, I used 0.5mm nylon Chinese knotting cord. I got mine from an online retailer called Tangles and Knots. I tested how much weight that cord could support, and it was more than the strength of the average adult finger, so that is probably fine. I have heard of others using a slightly larger nylon cord used for window blinds. You might need to tweak the hole sizes if you use something much bigger though. The elastic is, unfortunately, another trial and error thing. (The nylon cord pulls the fingers closed when the wrist is bent, the elastic extends the fingers when the wrist is released.) I used "Thick Elastic Cord" by Jewelry Fundamentals. I bought 6yds of it on a small cardboard card at the local JoAnn Fabrics and that was more than needed. I tried several other elastic cords. This is again a balancing act. The cord needs to have enough spring force to return the fingers to the extended position. If there is too much spring to it, it will be hard to bend them at all and will be exhausting for the user. You want something just strong enough to get the fingers back straight - and no stronger if that is possible. The elastic also has to have enough give to allow the full range of motion. You really cannot change either of those - it is just the physical properties of the elastic you get.

I did look for all sorts of other solutions for the elastic cords, but didn't find any of them satisfactory. I tried some experiments with strips of spring steel that I thought might have more uniform spring and could be tuned per knuckle. They needed to slide, however, and that didn't work so well (and they really aren't that much more regulated than elastic cord). I Thought I had a clever solution that worked sort of like a catapult tensioner where you would add twists to increase tension, and take out twists to reduce it. That didn't work because the twists made bumps that would get caught and it was hard to find a good material for the torsion part. I tried orthodontic bands thinking they would be manufactured to more exacting standards, but they broke far too quickly. Some people are trying to print a hinge out of flexible filament, but that requires special printing capabilities and is still not easy to tune the tension.

I considered other types of string, rather than the nylon cord. Most other materials do not seem to work very well. One obvious option seemed the high-tech, braided fishing lines. The cord has to slide through the joints, however, and those lines fray quickly. (They are excellent when they don't have to slide - like as the drive system for printers - but not for this application.)

I have tried printing this model without supports, but I am not having any success. Do I need to use supports and rafts? I have a Replicator 2x using ABS material. I have cleaned and leveled the build plate. I am running the table at 110 degrees and the extruder at 230 degrees as I do for all my builds using this material, but still things fingers de-laminate where they are at 45-degrees and some of the edges of my rafts are lifting. Any help would be greatly appreciated.

I have not printed it in ABS, only in PLA. I have never needed to use a raft, or a brim, or support for it to print - except for the gauntlet. The overhang part of the fingers sometimes came out a little rough, but never so much that it could not be cleaned up with just a little bit of light sanding. I could probably have even just left them as they were, but I wanted them to be as nice as possible. Rafts lifting generally means that the bed temperature is not where it should be. With ABS it is often hard to tell if it is too hot or too cold - both will look very similar. Based on the numbers you list, it sounds like the bed is a bit hot. I print ABS at a colder bed temperature than that but around that extruder temp. (My printer, however, is a heavily customized Mendelmax that I built myself, so it might need very different settings from what you need.) Have you printed anything the size of these parts in ABS? I generally find that delamination either means that there are drafts (ABS hates drafts), or the extruder temp is too low for that filament (but 230 should be in the right ballpark). Of course, it is hard to tell if the raft is lifting since that can throw everything off. I am generally printing at 0.2mm layer height. For drafts, I make sure that all of the fans are off for the print and sometimes print a skirt for maybe as many as 50 layers. The skirt then takes the drafts and delaminates, but protects the part. Hopefully something there helps. I am not exactly sure what is happening since I did not have an issue printing this. You mentioned fingers, so I assume you are not trying the gauntlet. For the gauntlet, however, I do use support - no brim or raft. I am afraid I don't remember the angle limit I set - something fairly normal like anything more than 45 degrees should be about right.

Thank you for your reply. I changed the kapton tape on my build plate,and set my "Settings" to 4 shells and 50% infill. I left my extruder at 230 degrees and my table at 110 degrees because this is normally what I use for ABS and the spool of ABS that I am printing with has worked fine in the past. I am printing the fingers at the moment using a "raft" but no supports and so far... so good so maybe it was the kapton tape just getting old. It did not have any tears and I usually wipe it down with Windex prior to printing. I am afraid to use acetone because it will most likely eat through the kapton tape. If you know of anything else to use on the kapton tape, I would certainly appreciate knowing what else works well or better..... Thanks again

I've printed and assembled one, and most of it was pretty clear so far from the photo's. I plan on delivering it to the user at the Orlando Maker Faire in September.

A few notes from the process so far:

I had to trim the finger joints' "round" ends to let the fingers bend, as there were corners on the bottom of the curve (on the print bed) that prevented the fingers from bending, but cutting the corners off lets the fingers curl in freely, so I guessed that was your intention. Is that right?

Two of the parts feel fragile relative to the stress - the gimbal and the back of the hand. You said that you printed yours using PLA. But what slicing profile did you use? I'm using 0.2mm layers, 2 shells, 30% infill. The fingers and gauntlet seem fine, but there's a lot of stress around the single pivot point, and I can already see the PLA delaminating at the hand. Should I print 100% infill? Or should I switch to a stronger material? I have some Taulman t-glase that make stuff that's almost indestructible.

BTW, I did make one tweak to the design. The two outer bolts that hold the cable guide to the back of the hand were in the cable channels, which would require cutting the bolts and even then seemed quite tricky to assemble. So I moved the bolts to be between the cable runs and everything went together nicely.

The Chinese knot tying thread that I found feels like it would break if forced. I'm wondering if perhaps it would be better to use wire? Or is it stronger than it looks?

Finally, when I rendered the OpenSCAD it by default gave me holes sizes for M3 bolts (machine screws) and tiny bearings. There's also an option for printed pins, but I went for bolts as being stronger. But in reading your post, I see that you're using 3mm filament for the pivots. That's great - fewer parts to order. Is there a setting in the OpenSCAD for that? And does it replace the bearings, etc.? And are you using them as rivets? I've been doing that in other designs by taking 3mm filament and pressing one ends against a hot surface to form the ends of the rivet. Is that what you're doing?

I found that M3 countersunk machine screws 15mm long fit almost everywhere in the design - fingers, hand, main pivot, wrist. I am using 3mm filament for the knuckle joint, as I can't find any M3 bolt long enough for an adult hand. But the filament looks like it's fine there. Then the rest are M3 nuts - one in the trap, and "acorn" caps on the screws in the wrist and the hand. And washers on the pivot between the hand and the gimbal.

For regular use, how does the elastic hold up?

Lastly, I can't find any springs the right size locally. But even without the springs, the tensioners seem to work pretty well. I'll probably break down and order some springs, but with shipping, etc., it seems expensive for four springs.

Hi Laird,

Progress is looking good.

The blocks on the round parts is intentional. I find that very few printers can handle a cylinder making tangential contact with the bed. They come out very jagged and misshapen. Because of that, I add the blocks to make them print clean and then file to a nice round. You can turn that off in Include-HandParameters.scad if you want. Just change Use_Support to false.

I printed entirely out of PLA. I used 0.2mm layer height, 2 perimeters, 5 solid layers top and bottom, and then varied the infill based on the part. For the wrist and main palm that make the gimbal I just made them 100%. I actually tried to break the palm part at the wrist and could not with my bare hands - even bracing the part against the floor and trying to force it. The palm part is a little weaker on that wing that connects the knuckle block, but still should be plenty strong for the actual forces in that area. The wrist piece is a little more fragile, but seems strong enough at 100% for real world forces. Most of the other parts I do at 30-40% honeycomb infill. I have at time done a higher infill on the gauntlet to make sure it was good and strong as well. I have had some that were fine at lower infill, but it takes so long to print that I want to make sure I don't have to repeat because of it being too weak. The nylon might be nice, but is probably unnecessary.

I would be interested in getting the scad files you are using so that I can see why the screws are in the string path. My best guess is that there is some dependency on some scale factor that isn't working optimally. Mine were in the right place, so there is some difference there. Although I am glad that you were able to fix it, I am a little worried that others won't know how.

The Chinese knot string should be pretty strong. I had, at one point, hung a 20 pound weight from a single strand of the stuff I have as a test. That said, you certainly could use other things. If you are to use something else, my best recommendation is larger braided nylon string - like the blind pull string. The material I most strongly caution against is Spectra fishing line. That is a miracle material, but it is not good when sliding over surfaces (which it is in these hands) and it will fray and break.

I am, indeed, using 3mm PLA filament and riveting them in place. I still use the small bearings. I then leave the settings same as for the screws and set parameters to inset the rivet heads. I use, for example, Inset_PIJ_Rivets=true, PIJ_Rivet_Diameter=5.0, PIJ_Rivet_Depth=1.0. That just makes a cylindrical inset on both sides of the joint that fills with the rivet head. I make the rivets a little differently, and you will see it in the photos. Most people head something and then tap at the filament. I first sand the end of the filament so it is a nice, smooth end. I then heat an implement and press the head completely flush in a single press. I do that press right by running water and then immediately quench the PLA. Then I can snap the implement off the part and leave a clean rivet every time. Only downside is the rivet head is welded to the finger or knuckle block. I actually like that - but it means you can't easily replace a rivet. You can see all that in the pictures - but it is a little unclear what is going on.

I haven't had any issues with the elastic I am using. I did make it so that you can change an elastic cord in a matter of minutes without any disassembly of the hand. I think it will break over time, but it lasts for at least a while.

For the springs, I can never find ones exactly the right size. I mainly look for the right gauge and the right diameter, and then end up cutting to the right length with a snip. It is hard to find just the right one. I am sure more than one hardware store employee has been concerned about me standing there staring at their spring selection for 20 minutes trying to decide what I might be able to make work.


Hello orgemd,
I'm willing to try and print this hand for someone who needs it in my town, but I'm wondering if I need to have openscad skills (which I don't, at all).

Hi lucascorato,

You need a little openscad skill - but not a lot. These hands need a lot of custom parameters (things like measurements specific to the person for whom it is made, or options like if it is the right or the left hand). You make those customizations by opening the openscad scripts (which are basically text files) and changing the relevant lines. I have tried to use at least somewhat understandable names for the variables, and have included a lot of comments in the files to help make sense of what they control. I have separated many of the key measurements and options into the hand parameters file - so changing them in one place should work for most things. That said, there are some parameters in the individual files (especially the gauntlet). I have placed those near the top of the file to make it easier to know what to change. Once changes are made, you need to know how to compile it into an stl file. That is all pretty basic stuff. I expect you can find a lot of information on how to do those things with an Internet search. You should not need to actually do any openscad programming. That is, you don't have to know how to make openscad produce a sphere of a certain size in a certain location with a cutout of a certain shape... None of that sort of knowledge should be necessary. The caveat would be if you try to create a hand much bigger or smaller than my original - then more tweaking might be needed. If you run into that, I might be able to help you straighten it out.


Fantastic, I'll give it a try, thanks a lot!
p.s.: amazing design, congratulations!

I am trying the V3 design. and there's one piece that I can't get to export as STL:


Compiles, but exporting to STL gives me:

Compiling design (CSG Tree generation)...
ECHO: "Scale Factor is", 1.3
Rendering Polygon Mesh using CGAL...
Geometries in cache: 1457
Geometry cache size in bytes: 4304624
CGAL Polyhedrons in cache: 170
CGAL cache size in bytes: 154446888
Total rendering time: 0 hours, 0 minutes, 13 seconds
Top level object is a 3D object:
Simple: no
Vertices: 618
Halfedges: 1934
Edges: 967
Halffacets: 702
Facets: 351
Volumes: 2
Rendering finished.
Object isn't a valid 2-manifold! Modify your design. See http://en.wikibooks.org/wiki/OpenSCAD_User_Manual/STL_Import_and_Export

I changed only the measurements, to values a bit larger than your defaults.

Any suggestions?

Interesting. It exports fine in version 2014.03 but not in the new version 2014.06.21 (a nightly build). Guess it's an OpenSCAD bug. Do you want to report it or should I?

If you could report it, that would be great.

Hi David,

Thank you for the quick reply. This is exactly what I need. We will get the supplies and print the hand in the coming weeks. Looks pretty straight forward, but I may have additional questions as we get into it. I'm not the handiest person (no pun intended). Will keep you posted.

I was born with Poland Syndrome on my right hand. I became a Shriner about 3 years ago because I wanted to help children with similar disabilities. A coworker/friend of mine found this website and we started looking at designs to possibly create a printed hand for me that I could learn to use and show families/patients of the Shriners Hospitals for Children. I think this design would work really well for me and I've downloaded the files and plan to create one in the next few weeks. I was wondering if you might have a parts list for all of the other materials needed to create the hand?

Any help would be much appreciated.


Steve Sturm
Bloomington, IL

Hi Steve,

I have never compiled a list. I just now grabbed my box of stuff and went through it to try to make a list for you. Here is my best accounting of what I used for the latest version:

2 to 3 feet of 1.5 inch wide Velcro strap (loops on one side, hooks on the other) - used for the gauntlet closure and also to close the hand in - I bought mine on-line from Amazon

1 package of Velcro Brand Industrial Strength 4 inch x 2 inch strips - these have a strong adhesive on the back - I add them to the gauntlet closures to avoid having to sew the Velcro - I got mine at Michael's Crafts

1 to 2 feet of elastic cord - this is used to return the fingers to straight position when no longer bending the wrist - I used Jewelry Fundamentals Thick Elastic Cord #SUL51978 Ecru from Advantus Corp - I purchased it off the shelf at JoAnn's Fabric (6 yds wrapped around a small card - much more than needed for a hand)

around 5 ft of 0.5mm braided nylon cord - this is the pull string used to curl the fingers when the wrist is bent - I used Chinese knotting cord that I purchased on-line from Tangles-n-Knots - it came in a 50 yd spool

8 small, sheilded bearings (3mm x 6mm x 2.5mm) - these are inserted into the finger joints to make the motion a bit more smooth - I bought mine on-line from vxb in a set of 10

4 small springs - these are used to allow the adaptive grip to work - I have used multiple sizes and tensions - the most recent ones were part# 88011 5/32 x 1-5/16 0.023 wire dia compression springs from Menard's - I cut them down to the desired length with tin snips (use safty glasses, the bits can really fly)

3M Gripping Material TB400 - used on some of the finger surfaces to improve grip - mainly on the proximal phalanges - I bought mine on Amazon and got 6 sheets of 6 inch x 6 inch size - that is more than you will need in a very long time - I think I use maybe 6 square inches per hand

Some form of rubber dip - maybe it was Plasti Dip - dip the finger tips in this and let it dry so that the fingers will have better grip - I got a can of it off the shelf at Home Depot - not sure the exact brand since I used the last of it on this hand - it starts to dry after you open it, so you only get so many treatments from a single can

medical-grade adhesive-backed foam padding - put this anywhere skin might contact the parts for protection - I ordered mine on-line from Patterson Medical and use about 1 square foot per hand

5 x stainless steel M3 acorn nuts - hold together the hand parts and cap off the screw ends to minimize exposed sharp parts

3 x stainless steel M3 lock nuts - these hold the screws that have to pivot in the wrist

3 x stainless steel M3 button head socket screws - sorry, I don't know the length - these made the wrist hinges - I cut them to length with a diamond dremel bit and polished to remove sharp ends

1 x stainless steel M3 button head socket screw to hold the palm to the knuckle block - maybe 20mm long

1 x stainless steel M3 jam nut to lock that screw into place in a trap in the knuckle block

2 x stainless steel M3 fender washers used in the gimbal joint of the wrist

5 x stainless steel M3 flat head socket screws - used to hold the other parts together - size varies though, so you probably need a variety of sizes to find the one that works - these are the screws that go into the acorn nuts - which are all trapped by the printed parts - because they go into acorn nuts you either need the length to be about right or you need to cut them

4 x stainless steel M3 socket head screws - longer ones - maybe 25mm or so - you want them to have a top portion without threads for smoother motion - these will hold the tuners with the springs

I have occasionally used some other M3 washers in various places, but not on the latest version.

Hope that helps. I think it is a pretty complete list of what I have in this latest model.


I'm going to be printing one of these for someone, and I wanted to check that this was the latest version of the OpenSCAD file.

Hi Laird,
Wait for about a week. I have a new version with a number of improvements (though nothing terribly radical). It is done and I was going to post it. Right now, however, I am traveling. I will put up the new files when I get back.

Will do! What material are you printing in - Bridge?

Hi Laird,

I have posted the new version 3 files. Read the update in the main description, take a look at some of the new photos, and check out the files. The fingers do need to be run through something like Netfabb after creating your stl files. There is some small error in the SCAD that I have not tracked down yet. I actually just use PLA. PLA has excellent wear characteristics. Although you will hear a lot about it being too brittle, that just is not a valid concern for most applications (it is for some, but only for some). I have not seen any PLA failures in these hands to date. That might in part be because I agonize over trying to get a print orientation and form that will allow for the greatest strength. I could make the hand look a lot cooler cosmetically, but only by sacrificing part strength.

I have been very busy lately. I know I have not provided as great of details on what I did here as is my norm. If you have any questions, please feel free to ask and I will try to get you the answers.

Beautiful work, David! Your work offers great potential to these mechanical hands.
Yes Cameron, I'm pretty sure the gimbal on the wrist that offers the lateral 2nd degree of freedom was inspired by the Bob Roth Hand. Bob is a wonderful collaborator and an outstanding machinist/designer.

wow. This is exactly what i need. i really like the axis and the separate digit control...now, i just have to find a printer . It shouldn't be too hard i work in the printing industry.

If the tensioner was closer to the wrist swivel and attached so like is seen here: http://www.thingiverse.com/thing:220942http://www.thingiverse.com/thi... , you could probably add the ability to pinch with the thumb and fore finger. Nice hand!

Roth Hand (Progressive and Independent Finger Movement)
by fflood