After my problems with rather small bolts and a cylindrical spring-based squeeze-hold mechanism for wrist/connector unit (Otto Bock part 10V9/10V10, cost 429.95 CHF) and after I did not detect any subsequent technical innovation being mentioned in the answers I obtained from the manufacturer (Otto Bock), we decided to move on to build our own setup.
MSM wrist 0.1 [prototype, open prosthetics]
Our parts are designed so they do not preclude a subsequent return to the original Otto Bock parts for a number of reasons, they allow for minimal cost for moving to this type of wrist and they allow for a very tight and stable fit while using pre-installed Otto Bock parts with variable specifications.
A first step was to get to a stable 'tight fit' (Formschluss) - further steps were intentionally omitted at this stage. It is intended to be a specifically stable solution for a long forearm stump, and force distribution was optimized by getting the terminal device adapter deep inside and as peripherally as possible so as to maximize the area over which forces are transmitted.
The wrist fits into the existing prosthetic socket (following image of aluminum demo: part on the right) and the terminal device-sided part containing a sprocket connector will fit onto the existing bolt of the terminal device by squeeze-on using three pin-pointed M3 inbus screws (following image of aluminum demo: left).
The first set was milled using Ramax (chromium alloyed stainless holder steel in prehardened condition). Below there are photos of the wrist fit sitting inside my prosthetic socket (center line: pin protruding from pin lock). The wrist features six bolts that catch the sprocket.
The terminal device is fixed to a sprocket connector with three pin-pointed M3 inbus screws.
Lacking a lock mechanism (currently under consideration/design, see further prototypes along development roadmap), the wrist locks with supination and unlocks with pronation. More importantly, the cable that pulls the hook does not pull in a straight line so we used this pull to automatically pull the sprocket into the wrist sided bolts.
As I am not using forearm rotation anyway we figured it would be good to see where that causes problems if at all. The sprocket sliders narrow a tad bit towards their ends so with brute force, the bolts can be jammed and that provides a lock of the wrist for most practical purposes.
Video showing initial handling that is so far button or lever free:
Analysis of application considerations
I will provide a systematic analysis of everyday performance and problems before we'll move to the next version. This prototype has no rotational lock yet and so it is interesting to see under what conditions the terminal device falls out.
Typing: This wrist is virtually wiggle free as opposed to the standard issue OB wrist. I like the precise and secure feel I get when I use a fast 5 finger + hook typing style. Typing is a major task at work and so this is important.
Carrying heavy objects: I work in a job where occasionally, heavy weights (up to 100 kg) need to be moved. As our disability insurance inspectors agreed that a cable controlled prosthesis was the correct choice for me, and as our orthopedic technician used default Otto Bock products, and as Otto Bock now appears to not support heavy usage of their hooks and wrists after all as they now specifically warned me to avoid heavy usage (without a more concise definition though), this is really a problem. It may be news to some but I am not wearing this for decoration so I have to be very frank about performance limits. So our new MSM wrist connector solves this problem: I can give full weight to lifting operations and the connector does not increase wiggle, it does not bend or otherwise give way. Today, we moved a 95 kg weight and the prosthetic wrist did not exhibit the slightest problem. No rotational unlock issue turned up. This is not a frequent but important area of usage for my prosthetic arm.
Driving (hook): No problem at all; the wrist allows for a very precise control of the steering wheel using the hook.
Ironing clothes, laundry: Not a problem at all as most small and frequent pincer grabs occur under almost permanent cable pull on the 'heavy' hook spring setting and under small supination load.
Garden works: Cutting hedges and broom cleaning the path did not cause any trouble caused by the wrist whatsoever.
Carrying a plate: Carrying a plate will trigger a pronation load but for practical purposes, I had no accidents as of yet. Theoretically though this could cause problems.
Opening refrigerator: Opening the fridge to put an object inside is a manipulation I sometimes perform with the prosthesis. When extending the arm, using the door handle and pulling close, pronation can occur and I had the wrist unlock once. So here, a missing rotation lock is a problem. However, there is no objective need to open refrigerators with the hook.
Pulling up trousers when dressing: Fast outward rotation while pulling up trousers also unlocks the wrist. However, it is not a really important task as I can just as well pull them up otherwise.
Driving (hand): Firstly, the new Otto Bock silicon glove does not provide a good grip on the steering wheel. Secondly, the Otto Bock system hand has a different cable pull direction so the wrist is differently loaded by the hand. Due to the orientation, the sprocket of our new wrist easily falls out with the hand. Even though, there are some dirty workarounds. Also, manufacturing a mirrored sprocket would be a no-brainer. So while this is a real prototype problem / issue, it's not playing out as biggie.
Analysis of manufacturing considerations
Our wrist design is shallow and could be built even more shallow - to allow for longer stumps and shorter wrists. Wearing a prosthetic hand or hook on a quick release wrist connector does not mean that I am willing to wear parts that make my prosthesis too long in comparison with my other arm.
Three M3 inbus screws are not a good way to fix a bolt inside a sprocket. M3 is so small the inbus key starts rotating freely rather fast when standard steel screws are used. We thus move our 0.1-revision setup to six M4 screws.
March 6th 2009. Concluding so far, it seems as if work related and relevant non-work related tasks are mastered very well already with this prototype. Full loads are stabilized easily with our sturdy design. Even the unlocked sprocket type lock is no problem at all if one knows that there is one rotational direction (supination) that locks it and the other (pronation) that unlocks it. Tasks that I perform that require a locked wrist seem to employ light forces and those are negotiable tasks, stuff I don't need to necessarily do. Rotational locking needs to be easily accessible and provide a tight fit.
April 8th 2009. MSM 0.1-revision fixes the most important issues right away and will continue to be the main wrist unit that I now wear on both of my prosthetic sockets (current and new one).