Experienced user based advice for investors in the domain of prosthetic arms [technical guide for understanding the field]

Investors may need to understand what prosthetic arms really are, how they come about, and what is there to be considered, before investing their investments, or money, in a prosthetic arm component manufacturer or research spin-off.

A current list of this type of products or hopeful startup device ideas is:

  • Mobius Bionics https://www.mobiusbionics.com/ https://www.google.com/search?q=mobius+bionics
  • Open Bionics https://openbionics.com/ https://www.google.com/search?q=open+bionics
  • Atom Limbs https://atomlimbs.com/ https://www.google.com/search?q=atom+limbs
  • Motorica https://global.motorica.org/eu https://www.google.com/search?q=motorica
  • ...

With a realistic estimate of around 85% rejection rate, the prosthetic arm industry so far is one of the most unsuccessful industries that there are both in high-tech and medical technology.

Now, what usually keeps developments, market introductions, and user feedback, as well as improvement circles going, is a successful social setting. Such a really successful circle, where respect combines with a striving for technical proficiency, in the context of prosthetic arms, appears to be largely absent. To a great part, this explains why the current status with regard to marketing, improving, or successfully selling prosthetic arms is not a lot better than maybe sixty years ago.

Marketing problems

Once you face a particular solution, very carefully check the people behind these solutions.

History and what can be learned maybe

The market is largely absent, and so, developers such as Touch Bionics that came out with the iLimb seemed to have had problems staying in business and were then sold to other companies. So, Touchbionics was sold to Ossur. With that, Ossur first of all closed down the social media presence of Touch Bionics. As the main selling point of the iLimb is a multi-grip control leveraged via a mobile phone app, Ossur now canceled out the Windows/PC software, the Android app, and made it impossible for us users using Windows and Android to access our prosthetic hands the way they were advertised. None of the manufacturers seem to sell a lot of "bionic" hands.

The product that could be successful is a product that works, and the problems to be covered are the following:

  • I wear a prosthetic hand if it makes me look good, and that possibly where other folks are that may look, inspect, and check. So for parties, for going out, for giving presentations or workshops. But the problem is, these prostheses cannot deliver that impression. With the slightest stress and sweat, the electrodes stop working. With the grip patterns, holding an olive snack plate or wine glass reliably, so it never ever falls down, is just not possible. So any malfunction is unacceptable at that level. All myoelectric devices excel mostly in totally stress-free environments, where the user never sweats, and they are best used in forgiving environments, where all grasps are accepted to malfunction. For all error-critical tasks, body-powered or cosmetic prostheses, or wearing no prosthesis, is the way to go.
  • I wear a prosthetic hand if it alleviates my asymmetry. The asymmetry is strained particularly when I work heavy bimanual or repetitive tasks. So the prosthetic hand must sustain very heavy use - and no myoelectric socket or myoelectric hand really does that. These devices are far too delicate. For all real heavy-duty tasks, body-powered prostheses are the way to go.
  • I may wear a prosthetic device possibly if it looks totally stunning and if it really offers great visual effects. The best thing to have is a hand no one recognizes as prosthetic. So far, that is best offered by my cosmetic hand. The iLimb is funny to look at, but not stunning. The device that I got by far the most unsolicited comments for so far is the TRS Jaws. Far more than for any other device. People seem to absolutely love that. Go figure.


Medical products should comply with CE-marking legal requirements in Europe.

You may want to make sure, beyond some doubt, that what you try to understand as a product actually follows the requirements.

Hardly any currently sold prosthetic arm component is, from my personal view, free of doubt in this regard, which is why I can not at all recommend investing in so many commercial attempts. The risk you run is as follows:

  • Product is overadvertised - i.e., promises are made that are not kept: first of all, CE-marking requires promotional material and product descriptions to be accurate to the point. If products are over-advertised, consumers buy them with wrong expectations. You will then experience furious customers if things go smooth, or even class-action lawsuits if things escalate.
  • Product limitations are not technically proofed and comprehensively listed, with error margins, in the documentations. If this is not done, consumers will invariably do their own testing and reporting and after that has taken place, you can probably forget trying to promote or advertise these products. Then, "product management by banana" (product matures at customer site) has gone full circle and you won't sell many anymore. Could have as well not invested, to begin with.
  • Product failure points and weaknesses are not openly communicated and declared. While CE-marking legally requires problematic aspects to be made available to any consumer, we do not buy prosthetic devices because they are perfect in each and every regard. Hell, no. We know they all suck in their own way. But if you do not make it explicitly clear that a certain prosthetic glove cover falls apart by itself after 3 months of storage, and the customer finds out, you do make yourself the laughing stock at least of the user community. Whereas if you put that into the "caveat" section of product documentation, you can tell any user to watch out. In a world of fragile overpriced items, visual awareness and great respect for the customer is the only way you can go, ever. Has not been tried so far, but I suggest that if you do go such a path, you can surely avoid a run-in with outspokenly and aggressively critical users such as me - and I can say for sure that I would have never made that many super negative experiences with prosthetic components had it not been for CE-non-compliant documentation and advertising.

If you tell me that a device is robust, then I will use my own definition of robust to test it. Simple as that. And if you do not know my definition of "robust" (and mind you: a product being CE-marked does not mean the consumer has to clarify their definitions... no, the manufacturer has to do that) then you may as well call it quits right there.

So if you are interested in whether a manufacturer actually complies with CE requirements, check the details of the requirements and then investigate your situation. It is worthwhile to check that before investing any money. You do your own research, all I can do is raise relevant questions.

Referring to prosthetic hooks by using Peter Pan and Captain Hook metaphors

Both the subject of prosthetic arms and their technical intricacies and the subject of fairy tales with their symbolic aspects and gory realities, are difficult and multi-layered subjects.

And anyone can tell you that actual understanding of dynamic, non-linear, or otherwise twisted aspects requires mental capacities above average.

So, unreflected use of a complex pointer may be a dead give-away that you are dealing with someone that has no clue, hidden agendas, or both. In particular, if ever you run into people that tell you that prosthetic split-hooks are bad because of Captain Hook being an evil guy, please make sure you ever so thoroughly check these people and their arguments.

Because the Captain Hook metaphor is one of the most wicked and unfortunate pointers anyone could give you, ever [read this: link].

So if you start pretending that body-powered technology is outdated or has no potential, you clearly exhibit a very serious lack of technological understanding:

  • The technology that does have significant potential for further improvements along many directions is body-powered technology, whereas myoelectric technology has arrived at a functional stand-stillinks. This is detailed in this publication [link]:
  • Most notably, myoelectric technology and "bionic" hands suffer a range of inherent problems that are largely unsurmountable.

Problems of the technology that are inherent

The problem of selling a prosthetic arm is that currently, out of the box or built with commercial conventional items, it does not really work. Sure, they are also expensive, but making them cheap does not make them really functional.

The prime issue is that they initially lack the functionality to a point where they become dispensable and unnecessary while causing a lot of problems and cost - particularly follow up cost and efforts with used up time - as well as anger, frustration, or disappointment.

  • Myoelectric control and socket: A myoelectric arm will stop working once you sweat on your stump. That is it. It is an inherent severe limitation. The electrodes also won't work for a few other reasons, so they are not reliable. One sweats when running on a bus, when being nervous, or when feeling just a bit warm otherwise - so one sweats relatively fast even during a so-called sedentary everyday life. The electrodes risk causing skin burns if used with less sweating. These may take 4-6 weeks to heal or so, during which you cannot wear the prosthesis. With no sweating and just a day of typing at the office, the prosthesis causes my stump to run up blisters and abrasions that are so significant they take a week to heal, after which I am not inclined to repeat the experience. This applies to a sufficient number of people. With that, a realistic >85% of arm amputees are non-users. A further large portion of amputees that start using a myoelectric prosthesis drop out, as the systems are never reliable in any comfortable way, really. In over 40 years, research and development never attained any actually useful control error rate for myoelectric control. Any promises that "this is improved" anytime soon cannot be taken seriously, as that actually is technically impossible. If it were, don't you think we would know that by now? You can always try to look forward, but literature analysis of almost fifty years if research and development tell a totally different story. A more detailed technical analysis that describes in understandable terms what the intractable, unsolvable, inherent problems of myoelectric technology are, is found here [link].
  • Myoelectric hands: You can wear a prosthetic hook gripper on your myoelectric arm, but really, everyone wants a "bionic" hand. They are fragile, the gloves may tear up as fast as within 10 minutes or so, they are relatively heavy and if not, they are really weak. Their grip may really be not configured well to a degree, where gripping itself becomes awkward, cumbersome, difficult, hard, due to the ill devised geometry.
  • Failure aspects of myoelectric arms: A realistic downtime after wearing the "bionic" hand prosthesis for just 1 day, just to recover from the strain and skin damage, is 1-2 days to 6 weeks. A realistic time for the myoelectric prosthesis to break is some 10 minutes (glove: torn) until a few months (other parts). A realistic count of uncomfortable problematic control situations over a day is maybe around 20.
  • Body powered arms: They are usually built wrongly, using a figure 9 harness with a thin strap, and, using a wrong type of cable housing. Their hooks, particularly those of Otto Bock, have some funny features that may make them fail surprisingly fast. Their wrists fall apart surprisingly fast, too. While a body-powered prosthetic arm may not be that ridiculously expensive, you can only really improve them with substantial good engineering. Large German and other companies do not sell parts that really hold up. We developed our own parts and have two patents. They are not really for sale, we patented these to protect us from larger companies so they do not sell our ideas back to us. All details here [link].
  • 3D-printed arms: their sockets and grippers are not useful for realistic everyday work tasks at all. The sockets that are of the usual 3D printed materials are likely to tear up an arm stump, over a day of hard garden work, quite painfully. The grippers are not expected to reliably operate dangerous or heavy machinery. The use of 3D-printing has been declared to be extremely experimental, so while a 3D-printed hand is relatively cheap, there is often an almost entire absence of true usefulness, so it is still overpriced no matter what. It is the experience of me that only a really comfortable and really robust arm will actually be of great use in everyday life, as a ground rule, whereas a prosthesis built for light everyday activities is likely to fail even these on a regular basis [link].
  • A passive or cosmetic prosthetic arm is one of the more functional concepts from the view of durability and reliability. It is possible to mount that on the arm with a relatively comfortable suspension even though even there, I did experience a relatively severe friction rash after only 90 minutes of slow ballroom dancing, that took 2 days or so to heal.

Requirements for a successful prosthetic arm


  • The socket must not cause abrasions or blisters even if worn for heavy work with sweating for at least 1-2 weeks without break. Hardly any socket can be survived with an intact stump when wearing it even for just 1 day. This causes the amputee to experience severe downtimes - not being able to wear the prosthesis. After a year of two mostly not wearing the prosthesis, you as a user usually find out that it is better and more functional that way, anyway. That is the death sentence of the idea of "wearing a prosthetic arm". That death sentence may start with having to take long breaks to let the overused stump skin recover.
  • The socket is the main part of the prosthesis that has to both fit the stump perfectly well and transmit the control commands, either physically in body-powered arms, or electrically in myoelectric arms.
  • A myoelectric socket contains electrodes and with that, ridges. That alone is a problem.
  • A suspension that does not damage the stump under work constraints necessarily is layered with sophistication. It may contain a first tubular gauze layer, then a soft thick gel liner layer, and then only a carbon or epoxy socket. Other sockets are available but they are not unknown to cause real trouble for the skin - and do not survive long, as a prosthetic arm under continuous use.


  • Control of a prosthetic arm must attain industrial levels of quality in grip performance [link].
  • An acceptable quality level will offer failure or error rates below three or four sigma, not just the usual academic "success rates" of 90-98% representing error rates of 2-10%. That spells out as 1 out of 10 broken coffee cups when unloading a dishwasher. Details of these considerations are found here [link].
  • Myoelectric prostheses have inherent serious control issues and thus one may have a hard time boosting their control error rates down to below some 0,1% which still is not great. Sweat, body or limb posture, electric interference, and skin issues are the most frequent problems there.

Grip geometry and forward reliability

  • Grip geometry must allow for realistically plannable grips.
  • An ill-designed prosthesis that has unplannable grip geometries such as the iLimb may even be a liability to wear.


  • Load testing and failure testing has easily to be as rough as you can think. An elderly lady even asked a presenter of a Michelangelo Hand by Otto Bock whether one could do push-ups on that prosthetic hand, a while ago. The man tried to dodge the question at first but the answer is clearly no. So, there.
  • Repairs: the prosthesis ideally is easy to fix, or better, can be repaired by the user. Prostheses that require costly remote repairs are likely to strand sooner or later.

Past actual experiences

Devices for developing countries

Devices for "bionic" fans

  • Current "bionic hands are either too heavy (Michelangelo) or far too fragile (BeBionic, iLimb). Status: unquestioned [i.e., link] in the community of users.
  • A few follow the fates of "brand ambassadors" for prosthetic hands, that are advertising representatives and that thus are getting their extremely expensive devices cheaper, but do not take their word for bare value - they just provide optimistic loyal sales talk. We all respect that. It may be really hard to smuggle too crappy parts past people that know what they see when they see it. Never fall for their assertions.

Devices and parts for users of body-powered technology

  • The TRS Prehensor, a body powered device, is very, very popular among users. It has also won the Cybathlon. Status: unquestioned.
  • The Becker Hand (body powered as well) can make manual workers very, very happy. Status: so far, single cases, that product is not too well known. No reports of that hand not working. Becker hands have been reported to be too weak by people that failed to go ahead to actually increase their spring settings (the hand can be tuned to higher spring settings but you have to do it in order to achieve that... it will not tune itself all by itself).
  • Alpha liners and Molnlycke tubular gauze works like a charm in sweaty heavy environments. Status: confirmed over about 5 years of continuous heavy use, by myself, confirmed by others.

Word of mouth gets around fast usually and most of us users are quite utilitarian in our needs.


Cite this article:
Wolf Schweitzer: Technical Below Elbow Amputee Issues - Experienced user based advice for investors in the domain of prosthetic arms [technical guide for understanding the field]; published 26/07/2018, 15:17; URL: https://www.swisswuff.ch/tech/?p=8702.

BibTeX: @MISC{schweitzer_wolf_1604181069, author = {Wolf Schweitzer}, title = {{Technical Below Elbow Amputee Issues - Experienced user based advice for investors in the domain of prosthetic arms [technical guide for understanding the field]}}, month = {July},year = {2018}, url = {https://www.swisswuff.ch/tech/?p=8702}}