You might have heard about the 3D-printed prosthetic hands created by Richard Van As and Ivan Owen for about 100 children thus far. That's a great thing they've been doing. There's also a "Snap-Together Robohand" version, http://www.thingiverse.com/thing:92937, which is intended to be easier to build. I'm sure it is... but it still takes half a day to print the parts and assemble them. Unfortunately, I had less than 3 hours total on Saturday, September 28, 2013, to give an overview of making technologies and have YMCA Black Achievers students actually build something. So, I designed a new prosthetic/robotic hand for which all the printed parts print assembled without supports in less than an hour. You still have to thread some fishing line and rubber bands through it, but from print start to working hand is easily done in less than 90 minutes!
The hand is roughly sized to match that of a 18 month old girl. It embodies a variety of compromises, any of which might make it unusable as a prosthetic in practice... but it certainly is functional enough to get the idea across -- which was my goal. I know very little about prosthetics, but I do know that fitting them is a very personal thing, so the primary design doesn't have a mount attached. It is possible to add a simple mount and still have the complete structure print assembled as either a prosthetic hand or a hand for a robot.
Step 1: It Prints Assembled (but you still need to add some stuff)
Although the structure of the hand prints fully assembled, you still need to add the stuff that makes it able to grasp things. It doesn't take much time, nor equipment, but here's what you'll need to build a fully operational hand:
Step 2: Two Hands Are Better Than One?
Grab the design files I've posted on Thingiverse: http://www.thingiverse.com/thing:158843
Initially, I'm posting just the STL file for a pair of hands roughly sized to an 18-month-old girl. Yes, you can actually print both a left and a right hand on the bed of our M2 simultaneously -- and there's still space to spare, although we recommend rotating them at a 45-degree angle. This is what you see in the photo. However, I'll post more design files as I get things cleaned up and/or I hear more about how people would like to use this.
The hand was designed using openscad, and is fully parametric in terms of both overall size and tolerancing. The tolerance issue comes from the fact that this hand has a lot of print-assembled hinges in it. The hinges use a special design that should be printable using most FDM/FFF printers, but how close to each other two threads can be printed depends a lot on what printer you have, the material you're printing with, and how well you've tuned everything.
Although there is a lot of complexity in the fingers and lots of plain solid in the palm, the hand prints in PLA as quite a strong part even with relatively low fill percentages. Something between 20% and 40% should be fine -- most we've printed have been 25% fill. We used 0.25mm extrusion of PLA at a head temperature around 195 degrees C. That's a little hot, but there are no significant spans here, and we need good binding to get the strength desired for a prosthetic. The borosilicate glass bed was kept around 70 degrees C during printing; the hand sticks to the bed very well at this temperature, but literally pops off when the bed has cooled back to room temperature.
Step 3: The Not-At-All-Speckled Band
There are five fingers, each of which has 3 hinges (joints) and four segments. On the back side of the hand, which was printed facing the printer bed and has a UK (University of Kentucky) logo on it, you'll see that each finger segment has a little bridge. The rubber bands pass under the bridges, one per finger. For each finger:
You can trim the excess band ends any time, but I'd suggest delaying a bit because it is easier to slide the knot down to adjust tension when there is a longer end to grasp.
Step 4: No Strings Attached
The rubber bands bring the hand back to the flat position after closing it... but what closes it? Well, it could be "muscle wire," but for a prosthetic, it's generally some type of cord attached to the wrist or a shoulder mount. Here, we'll use fishing line. For each finger:
Once all the lines have been run, grab them behind the wrist and try pulling on them to close the hand. Work it a number of times to remove any minor burrs or imperfections that may have made operation rough. Once operation is smooth, carefully align the lines and twist them together with the fingers all even. Tie a simple knot at least 4 inches behind the wrist to keep all the lines together.
Now you can make final adjustments to the rubber band tensioning. If any of the fingers do not return smoothly, increase the tension on those fingers by nudging the rubber band knot closer to the finger with the rubber band stretched a bit. Don't worry if you break a rubber band -- you can easily replace it.
Step 5: Give Yourself A Hand
Ok, you're sort-of done now.
Recall that we said you could have modified the design so that the printed hand would include a mount... but you didn't do that, did you? So, how do you hold it? I find it easiest to hold it like I did in the intro photo -- from the left and right sides of the palm, using my other hand to pull the fishing line. It works surprisingly well... as a demonstration.
If you really want to try to use this as a prosthetic device, you'll need to figure-out how to scale it and mount it on the user. Scaling isn't hard, but mounting it depends strongly on the anatomical structure of the individual who will be using the hand. For example, if one is only missing some fingers, the design needs to be adjusted to fit around the remaining fingers. A potential user with an intact wrist easily could be fitted so that downward motion of the wrist closes the hand, but routing control up to the shoulder is also possible. Replacing human muscles pulling fishing line with motors or "muscle wire" also could be viable, and this could make a nice, simple, hand for a robot.