Pediatric Trainer
The Idea
This project is creating an add-on to a pediatric body-powered arm that would help children learn to use their prostheses faster. The idea comes from Robert Haag, whose two year old son Michael uses a body-powered gripper. Here’s a video of a typical training session with a physical therapist:
In the video, you can see the adults giving him positive feedback when he does the right thing. This is essential to learning; the quantity, quality, and promptness of feedback directly affect the development of a skill. Rob’s idea is to build a small device that would measure the tension in the cable and make friendly sounds to tell Michael that the gripper is opening and closing. Now, instead of just being in short sessions, the feedback would be instantaneous and constant, hopefully helping him learn faster and better.
Jack Walker, an experienced product design engineer, has volunteered to help with this project. He and Rob are corresponding and we’re publishing the results here. If you want to volunteer or offer suggestions too, please contact us and we’ll put you in the loop.
Technical Information
The device will essentially consist of a sound chip that can play audio samples, a sensor that measures tension on the cables, and possibly some component that sits between them and handles the logic. With the right kind of switch, it might be possible to make the logic mostly mechanical.
Rob has found some sound chips that might work from AGC Sound. Below are pictures of them and Rob’s comments. On the left is the 1301 and on the right is the TAS:
The chips pictured here are samples sent to me by Les Zubli at AGC Sound. They are currently activated by closing the connection. The 1301 is the most attractive because of it’s small size. It’s also the least versatile (oh well, it’s a first prototype, right?) The TAS can be re-recorded by the user. The metal disk on the bottom of the 1301 distributes the sound. If you hold it with your fingers, you can barely hear it. Lay it on a resonator like a table top and it’s sounds nice and loud.
Here are some pictures of Michael’s prosthesis and Rob’s comments:
Ideally the new device would be located close to the hook. Limiting any improvements we make to the terminal end of the device means that more people could take advantage of it without consulting their prosthetist. How much force to close the hook? I’m just a dad, man :) I wouldn’t know how to measure that. Michael’s only two so it can’t be that much, although he’s pinched me a couple of times by accident and it hurts like %$@@^#. It has to be enough pressure to overcome the elastic that holds it open and can still hold something. Hanging three VHS tapes from the shoulder strap seems to be a good threshold. In the future we may want to make that threshold adjustable or make it more than binary (the more strain the louder the sound, for instance).
The Trautman Hook
The Trautman Hook is an upper-extremity terminal device that was invented in the 1920’s or 1930’s and produced until the company went out of business a few years ago. Kenneth Heide, CPO brought this device to our attention in the interests of getting it produced again, and generously loaned us two used devices and two unused devices for reverse engineering. The Trautman Hook backlocks when closed, uses fewer rubber bands than other models, and packs a high mechanical advantage into a small package. Its simplicity, at three metal parts and two screws, makes it a promising platform for customization.
We reverse engineered the old hooks and made a CAD model in Alibre Design. Below is the original hook and a picture of our CAD model:
We made some small changes to the design based on the areas where the used hooks had been broken and welded back together, and there are probably more opportunities for stengthening and weight reduction. As soon as we had finished the model, we emailed it to Bill Watson at Anvil Prototype & Design, who printed it on his Z Corp rapid prototyping machine and filled it with cyanoacrylate (super glue) for strength. We we able to assemble the parts into a moving model to test the design:
The next step was to try to get the device made without investing a lot in tooling. The best option seems to be rapid manufacturing, so we got the device quoted for different processes and quantities. Here are the resulting quotes:
Prometal (3D printing of stainless steel powder)
- $435 for 1 device (subject to a minimum $500 order)
- $370 each for 10 devices
- $275 each for 100 devices
Rapid Tool Inc. (SLS Laserform ST-100)
- $954 for 1 device
- $338 each for 10 devices
- $289 each for 100 devices
Quickparts.com (plaster casting process)
- $5,651 for 1 device in Aluminum
- $793 each for 10 devices in Aluminum
- $217 each for 100 devices in Aluminum
American Precision Prototyping (investment casting, not including final machining of holes and threads):
- $701 for 1 device in Stainless Steel, $759 for Aluminum
- $449 each for 10 devices in Stainless Steel, $467 for Aluminum
- $166 each for 100 devices in Stainless Steel, $170 for Aluminum
Precise Cast (casting with CNC machining):
- $7,213 for 1 device in Aluminum
- $1,257 each for 10 devices in Aluminum
- $448 each for 100 devices in Aluminum
Synergeering Group (direct fabrication in Titanium)
- $2,325 each for any number of devices in Titanium
Note: The rapid prototyping and manufacturing industry is very dynamic, and these prices may not be accurate within a few months of February 2006.
We’re working on getting the device back into production using one or more of these processes. If you want to be notified when the device is available, please contact us.
You can download the CAD model (4.8M zipped) if you want to experiment with the design or STL files (2.8M zipped) if you want to get your own quotes. Warning: This version is largely untested and may not work as intended.
We sent the part to ProMetal for a first functional prototype. Here are some pictures of the parts when we got them:
As you can see in the closeup, the printed surface is a bit grainy, but not unattractive, and vibratory tumbling would help it look smoother. After a little drilling and threading, it went together without a hitch. Here’s the assembled device:
The action is smooth and the backlock is tight, the only issue is that the grippers don’t line up perfectly, probably due to the fingers warping a little during heat treatment. The total error is about 0.050 inches.
The next step is to test it with a body-powered harness, then we’ll try to break it to see where the it needs to be stronger. Also, we’re already planning on making a few revisions that would make the device more durable. The cost of this prototype was $500 (ProMetal’s minimum order) plus about $50 for finishing tools. Thanks to Kenneth Heide, CPO for funding it.
All content and designs on this site are in the public domain, and we place no restrictions on their use. We encourage any derivative works, but all designs are registered periodically so that our work cannot be kept from the public by patents.
Adaptive Grasp Idea
When a human hand grasps an object the hand comforms or adapts to the object and distributes the grip force around it. A conventional prehensor grasps an object in 2 or 3 relatively small places. Because of the limited contact with the object, the mechanical hand requires more force and the precision required in the placement of the contact is critical. Several researchers are pursuing hands that adapt to objects by bending the fingers at multiple places. This paper out of the University of Toronto describes one such aproach involving a set of hinged linkages and this paper out of Stanford describes a simpler approach using a Spectra cord that acts like a tendon and runs through the joints of fingers. Both efforts have shown that as the hand conforms to the variety of object shapes encountered, less overall grip strength is needed to grasp the object.
We’ve been working on a mechanism for adaptive grasp that uses a flexible bag filled with small particles. Normally, the particles can slide past each other and the bag can easily change shape, but when the air is removed from the bag, it squeezes the particles together and they behave like a harder substance. Our first prototype of this concept was the finger of a dishwashing glove filled with sushi rice. Here’s our second prototype of the concept, made from a surgical glove filled with tiny glass beads:
This video (2.2M) shows the glove gripping an object. There is no structure to it yet and it has to be manually positioned onto the object, but when vacuum is applied, it hardens and holds the object securely. This is a good demonstration of adaptive grasp because there is virtually no grip force exerted on the object itself. It is able to conform so readily to the object that it grasps easily even if it is placed without care.
There would be a lot of details to work out on this concept such as the source of vacuum (a piston?), how it actually closes and how to make it durable but it is an interesting direction. I envision a hand made from a simple material that uses vacuum to close around an object and then hardens.
All content and designs on this site are in the public domain, and we place no restrictions on their use. We encourage any derivative works, but all designs are registered periodically so that our work cannot be kept from the public by patents.