Select Page
Design Of An Assistive Upper Limb

Design Of An Assistive Upper Limb

Here is the final report written by 3 graduate students who worked during 4 months full time to build a proof of concept for an assistive system dedicated to ease utilization of e-nable printed apparatus. They together investigated the state of the art in low cost assistive upper limb prosthetics, and proposed and tested a set of technical solutions.
Thanks to Júlia FIGUEREDO DE ALENCAR, Khumbo NYIRENDA and Nicholas TYRIE, who came from the University of Bath (UK) to perform this semester in the school of Industrial Engineering in Grenoble, for their high involvement in this project and for the quality of this report.

Philippe

Thesis advisor, Industrial Engineering School, Grenoble Institute of Technology

You can read or download the report of the 3 students : Responsible_Design_Final_Report. Happy readings.

Wrist Lock System: Relieving wrist bending

Wrist Lock System: Relieving wrist bending

Context:

Classically, bending the wrist enables handling objects with an E-nable hand prosthesis (Raptor or Phoenix). But, we identified that this bending movement, necessary to trigger a grip, led to tire the user of the prosthesis with time.

After Nathalie has experimented her hand prosthesis (see post), it became obvious we had to find a solution to relieve her wrist muscles. Nathalie is an adulte who suffers from an amputation of her right hand (after her wrist) after a healthcare-associated infection. Once she contacted E-nable, she was equipped with a first customised prosthesis. However, as she used it regularly – to handle tableware, purse, or to bike – the fatigue of its muscles becomes too important.

Need and process :

It becomes necessary to enable a grip position of the prosthesis without effort from the user of the prosthesis. Additionnaly, this feature has to be as less bulky as possible and easy to use.

Different solutions were considered to maintain the tension in the finger wires in order to lock the grip position. However, no trade-off was found out. As an alternative, we looked for a system locking the hand in the grip position with no effort. Thus, below proof of concept has been used as a guideline, based on a ratchet working principle.

Row idea, based on ratchet working principle, for locking the grip position of a hand prosthesis with no effort from the user.

To approve such a concept, a first prototype has been designed and printed. Palm and gauntlet proxies have been replaced by simplified parts, printed in blue on below pictures.

 

Proof of concept, every parts were printed in PLA

All the design was realized on « OnShape » for all the advantages listed here.

Design : first version

Once the proof of concept approved, a first functional version has been realised:

First version of the « Wrist lock». On the left, in open position. On the right, in close position.

 

Design

On the kinematic side: the lever cam, as in final version lifts the tooth (cliquet in french) to the open position, when the lever is pulled to the right hand side. The cam is then locked by a small hole in the tooth.  Le flex spring, printed in Ninjaflex, is compressed, to lock the open position. When the user pulled the lever on the left hand side, the flex spring lowers the tooth. This same flex spring maintains the contact between the tooth and the ratchet (roue dentée in french) by compression/extension alternatively.

On the material side, we started by printing tooth and ratchet in PLA and iGlidur (IGUS). But strain wear due to the ratchet/tooth contact were too important. So finally, we machined them in metal (aluminum/steel) to limit wear effect and to support efforts generated by an adult (for instance, the screws used to attach the ratchet to the palm quickly get loose because of these efforts. They have been replaced by two nuts-screws). Ratchet was bought at a specialist and tooth was machined in our hackerspace with a CNC DIY. Notice that the « wrist lock system », in its final version could be nevertherless fully 3D printed if the final user is a child since the generated effort would be lower.

First observations after print and assembly

After assembly, the system works well. However, we observed quickly a permanent deformation of the flex spring, mainly in open position. Likewise, the cam, in PLA wears against the metallic tooth

So we have redesigned these two elements to make the system robust. Purpose was to avoid the wear of the cam while not loosing the elasticity of the spring.

Second version

For the second version, a torsion spring has been introduced to solve both previous drawbacks: 

Second version of the « Wrist lock system ». On the left hand side, the cam and the flex spring have been replaced by a torsion spring. This spring link the lever to the tooth. It lowers the lever and maintains the contact between the ratchet and the tooth in close position. Likewise, when the lever is in open position, the spring raises the tooth. On the right hand side, the spring has been designed with the helix function of Onshape.

Design

We have first considered the wear issue. This wear comes from the friction between the cam and the small hole in the tooth. This contact is necessary to switch between open and close position. So both features have been deleted, replaced by the torsion spring. It is now a part of the lever and slots into the tooth. Thus, both part are solidly linked. So, tooth can switch between open and close position when the lever is pulled toward the right or toward the left respectively, through spring action.

Then, when the lever is pulled toward the left. Tooth switch in close position. Then, when the wrist bends, torsion spring maintains the contact between the ratchet and the tooth by means of its elasticity. Thus, the lower part of the tooth, from version #1, becomes useless and is deleted. 

In this new configuration, the torsion spring does not require to bear high efforts. Indeed, beeing attached under the tooth, it is only used to i-maintain the tooth in open position when the lever is on the right, and ii-maintain the contact between the tooth and the ratchet in close position. Such low efforts limit the risks of wear and failure of the system.

Observations after print and assembly

After assembly, the system works well. However, the torsion spring is not strong enough to switch the lever to the open position when the tooth is in contact with the ratchet. Likewise, in close position, when the wrist bends, the spring is in contact with the ratchet and so wears.
Some final adjustments are then necessary.

Final version and finishing touch

In its final version, lever and tooth have been pulled away from the ratchet (a few milimeters) to avoid generating friction between the ratchet and the torsion spring. The cam lever, from version #1, has been re-introduced to help switching from close to open position. The torsion spring then ensures to remain in open position:

Final verson of the « Wrist lock system ».  The cam lever has been re-introduced but only to help switching from close to open position. The torsion spring then ensures to remain in open position.

After assembly, this version is retained.

Finishing touch

Once the final design retained, two finishing touches have been added:

  • A spacer is mounted on the ratchet to protect Nathalie from the sharp metallic teeth. 
  • A cover is added for aesthetic purpose.

Finishing touch for the final version of the « wrist lock system ». On the left hand side: a protection spacer to avoid any contact with the sharp metallic teeth of the ratchet. On the right hand side, an aesthetic cover.

Why we are using Onshape for designing our projects ?

Why we are using Onshape for designing our projects ?

Most of us, makers of the e-Nable community are not professionals of mechanical design and we are learning this new activity the more we could.

To ease the learning curve, we need to access the most simplest but powerful tools.

From the various posts published on blogs or facebook, most of us are using open software or free version of commercial products whilst very few are using their professional package.

In our case we experiment various software until we reached an obvious agreement to use the same. In the past we were using free and open software like Openscad, freecad, some education license for Solidworks, or free version of commercials products like 123Design or Fusion 360 (Autodesk).

Every software has advantages and odd points, but there were also common issues:

  • Continuous upgrades to be installed
  • Not always the right level of power from our machines
  • Disk capacity always increasing
  • Often crashes, blue screens, white screen of death ….
  • Huge files to exchange with friends
  • Often no sharing capabilities to work on the same project with friends.

There is no heaven in software, still true, but we can find a better life solving most of the issues … unfortunately creating new ones.

For more than a year, team Gre-nable.fr and friends have moved to using Onshape (from Onshape.com). This is a web application, accessible with all browsers, professional grade, with almost all what we need to design our prosthetics.

I will rephrase the 8 points cited by Onshape’s marketing team because I fully agree with them and they were the trigger to write this memo to let our community knows more about it and to consider it for future works.

We are using Onshape because:

  1. Onshape is AGNOSTIC – As Cloud based CAD, it works across Android, iOS, Microsoft Windows, Apple MacOS and Linux – and with phones, tablets, laptops and desktop computers. No more issue with applications that are missing MacOS !!
  2. APPLICATIONS AND UPDATES – No more updating software release with regularly crashes because your system is not 100% sanitized or because the release has not been fully tested by the development team. Everybody knows (and suffered from) this situation!! Onshape is updating his release once a month with new features and bugs fixing. You don’t even see it, they have to push a message in your mailbox to learn about new features.
  3. COLLABORATION – Teams can instantly communicate about design changes and get real-time updates on their colleagues’ progress. The ability to let multiple friends simultaneously work on the same design means that we are more efficient. Counterpart to free license, all our projects are public but who cares in the open world ? We are more than happy if other e-Nable volunteers will copy and improve our designs.
  4. COST MANAGEMENT – The time to buy more powerful computers, more disk space, more memory, is over. We just need a good Internet access (dsl is great), the right graphic card and a clean installed OS. No need to back up tools as well.
  5. MOBILITY – This is not a strong argument for us as we are mostly working from home, but you can create and edit your designs on your phone or tablet, carrying your CAD system with you wherever you go. This is nice to show your design during seminars, meetings, shows where you will meet other members of the community.
  6. SCALABILITY – This criterion is not applicable for us, whilst the free users should not benefit from this ability to require more power for complex design or rendering.
  7. SECURITY – We delegate to the web application the tasks of backup’ing our works, keeping always a safe copy. The developments are journalized so it’s always possible to revert to an older version. The concept of releases and branches is sophisticated and easy to use.
  8. SIMPLICITY – The entire product is based on an internal scripting language ‘featured script’ and the cleverest of us will be able to automate some tasks by editing their own scripts. The basic rules of designing are pretty simple, lots of tutorials exist on Youtube and (as usual) a community is helpful on the support blog.

Of course I have considered the PROS and there are few CONS !

CONS#1 : For the time being, Onshape is still young and is not as features rich as Solidworks or Fusion360. For instance, we needed to use Fusion360 to transform Quad mesh into TSplines, to be then imported into Onshape (see explanations in the post “Adapter une emboiture pour une prothèse“)

CONS#2 : Onshape is said “secure in the cloud”. This is true as long as the company will exist. I bet this company is well founded and will last as long as me. I’m not sure this is a concern, the reality is that since the good ol’ time when I started microcomputing with Apple ][, I lost a lot a important information when changing from one technology to another, and I survived this information chaos. So, the PROS win against the CONS.

At that point, you can think I could have some interest in pushing Onshape on the front scene !! Unfortunately I have none. I’m only enthusiastic by using this powerful tool, and I hope others will do the same, letting us share our projects to build better prosthetics for those who need them.

All our designs are freely accessible for copy, modification (once you will create your own copy) and for sharing whatever the distance is between us.

How to get an account? Nothing could be easier: create an account from the landing page (www.onshape.com), let them few information and that’s it! They are not pushy guys once they understood you will not buy the full paid license. You will receive emails announcing new features, free webinars …

The team behind Onshape is issued from prestigious CAD firms under the leadership of Solidworks’ founders.

To find our project, the usual method is to login at https://cad.onshape.com with your credentials, then to enter the keywords in the Search tools windows.

To find the complete design of our 3D printer, the key words is : “LOGresse”.

Onshape will display few projects including LOGresse within the name, the first one displayed “LOGresse Main” is the correct one, whilst the others are projects of various parts built outside the main stream.

You will be able to visit all ‘Part Studio’ and ‘Assembly’ but to see the scripts, feel free to make your own copy of the project and then, you will have 100% control on each elementary CAD primitives. You will be able to download STL files for 3D printed pieces and DXF files for cutting plates of metal.

For the design of the prosthetics, the keywords to find them will be released in each post of our blog when we will deliver the full details. Coming soon.

Matteo’s Release – Technical Aspects

Matteo’s Release – Technical Aspects

Matteo’s arm: Technical Aspects

Matteo’s arm has been designed mainly based on the Unlimbited Arm , and more precisely the Alfie Edition.

This page will document why and how we proposed some modifications of the UnLimbited Arm. The main set of modifications intends to improve “grip” efficiency.

We then also propose to modify some other details:

  • modification of the forearm shape to make it better aligned with the arm cuff when the elbow is open
  • cuff printed flat and bended in several steps (easier to print and stronger result)
  • insertion of a fun magnetic functionality in one finger.
         Note: The STL files and size information are provided on Thingiverse.

How to get a firm grip with the Phoenix hand mounted on an Unlimbited Arm?

As we said in the “story page“, the main goal was to improve the grip efficiency to allow the user to take a variety of objects. The main ideas are in 3 categories:

  • minimize friction so that a maximum effort is actually transmitted to the fingers;
  • use an efficient whippletree [John Diamond] to ensure almost all fingers come to contact with the manipulated object;
  • put on almost every potential contact surface both a soft material (flexible/deformable) and a high friction coating.

These 3 points are described more in details below.

Reducing friction along the tendons paths

In the Unlimbited arm mounted with Phoenix hand, all tendons go through a complex path from the fingers to the tensioner situated on the cuff. Thus a relatively high friction occurs. For example, if we have a look in the hand, the standard “S-shaped” path (red line on the figure) is needed for the tendon to path around the hand of a standard patient with his(her) palm in the prosthetic hand. But for a person without a wrist, there is enough place in the hand to go straight to the base of fingers (yellow line of the figure). This simple modification can drastically decrease friction within the hand. The first idea is thus to drill the yellow holes and leave the tendons cross the hand straight. The second idea in the same area is to insert PTFE tubing in these holes. This PTFE tube is 2mm outer diameter, and around 1mm inner diameter. It has been sourced from a hobby shop (RC models).

Direct path for tendons in the hand (click to enlarge)

For the thumb, a continuous PTFE tubing drives a kind of Bowden large radius path (yellow dotted line on the picture). This thumb tendon tube has been maintained in place in the hand by PLA deposited with a 3D printing pen (3Doodler for example). Another gluing or welding solution may be used as well.

Thumb tendon path with PTFE tube (click to enlarge)

Finally along the forearm, a PTFE tubing can also be inserted for each tendon (and we will see below that we only need two: see the red zone on the CAD image below).

Holes for PTFE tubes in forearm (click to enlarge)

Integrating a whippletree

On the Unlimbited Arm Alphie version, the fingers are associated by pairs (see stringing process). We expect that a full whippletree system as proposed by John Diamond would be more powerful to ensure the best possible contact of all fingers on the manipulated object.

As we have enough room inside the hand, the second idea is then to place the whippletree in the hand, so that the tendons are as short as possible, with minimum friction. It is simply made of a whippletree bar and a fork. Only one tendon then goes from the hand to the cuff. Note that this tendon should be very strong and rigid as it has to withstand the force of 4 fingers simultaneously. A fishing line made of steel core coated by Nylon (outer diameter 0.5mm) is used. This main tendon is fixed to the fork by a M2 or M2.5 screw (still too long in the picture) and a bowline-knot.

Whippletree integrated in palm cave (click to enlarge)

Note the small hooks at both ends of the whippletree bar: the may receive two elastics that will connect to the central screw, to make a balance in order to ensure that the movements of the two pairs of fingers initiate simultaneously despite a possible difference of friction in pins joints or difference of stiffness of the dental elastiques. They were not required on Matteo’s hand.

Whippletree – 3D design

 

The video below demonstrates how our whippletree is functionning.

 

Combining flexible material and adhesion coating

Case of fingertips

The last phalanx of each finger should be as sticky as possible on the manipulated object. Makers from e-Nable community often use Micro Gel Fingertip Grips, but when we ordered the smallest size available (size 3), it appeared that they were too big for the small hand we made for Matteo. They just fitted its thumb size, not other fingers. Moreover, we didn’t want to add thickness over the fingers, as we expected to keep the fine design of the Phoenix hand and fingers. So we decided to develop a different solution to improve the grip.
Tip-top assembly

After numerous trials, we converged to a combination of rigid PLA with flexible PLA, and a layer of sticky material. We thus separated the distal phalanx in two parts. The first half is made of rigid PLA in order to withstand the movement force (the tendon is attached to this part), and the tip of the finger is made of flexible filament (after trials with a variety of flexible filaments, we used Ninjaflex for the final version). The printing parameters need to be carefully tuned[1] so that the fingertip is rigid enough to support the catching force, but has a deformable surface to maximize the contact area with the manipulated object. The assembly of rigid and soft materials have been initially made by the use of a double extruder printer. But we preferred to publish a version that can be made with a single extruder. So the two parts of the distal phalange are printed separately and assembled by taking advantage of the flexible part. It results in a strong assembly but removable tip if needed. No glue needed here. The new parts have been called “tip” and “top” in the STL files list.

Each distal is split to create a “tip” and a “top”
(click to enlarge)

The next step will be to take advantage of the removable tip part to hide the tendon knot inside this assembly. This is freeing more surface to put adhesion coating on the inner side of the distal phalanges.

Finally in order to improve friction coefficient, these soft fingertips have been immersed in PlastiDip to be covered with a kind of rubber shell.

Case of the palm

The same idea of placing a soft and sticky material is also applied to the palm. We started from the hand generated by the .scad file of the Unlimbited arm, removed the rigid palm, and made a new one that will be fixed by a number of 2.2mm x 6mm screws. We set up the thickness and added a few bosses distributed quite like in a human hand and that will be printed with flexible PLA.

Palm cover made of PLA and Flex

This palm is printed with three 100% infill layers of PLA (that is 0.6mm) to provide a certain rigidity, then the filament is replaced by NinjaFlex to continue printing the flexible layers and flexible bosses. After printing the palm is covered with a thick layer of PlastiDip.

Blue =PLA, green = Flex

On the proximal phalanx

Finally small square pads made of Ninjaflex and coated with plastiDip are glued under each proximal phalanges. The following figure shows the look of the final hand.

 

Put fun and useful add-in: a super power for Matteo

As most of children we were sure that Matteo would like to be a super hero, with super power. We also wanted to provide him with a additional possibility to catch small iron objects with his new hand. This has been done by including a small neodymium magnet into the “tip” part of the forefinger. A void has been reserved in this part, and a pause in printing process allowed to insert the magnet. The magnet size for the forefinger was diameter 3mm and length 6mm. We also made some tests with a diameter 6mm length 4mm included in the thumb tip. Both version work well, the thumb being more powerful (thanks to the bigger magnet), but the forefinger seems to be easier to use.

 

Split view of the forefinger tip.

Pause during print for inserting neodyme magnet

Both STL files are provided in “Thingiverse Matteo release“. Magnets have been provided from this shop (www.sanf.fr).

Matteo picking a badge with the forerunner

Forearm modifications

The main modification we did on the forearm model is linked to a misalignment between forearm and arm cuff.

The problem we faced with

After printing and bending first forearm and cuff, and trying to make the assembly, we discovered that we had to make the forearm quite flat to be able to insert pins.

Original forearm

Pins axes distance

Cuff forearm misalignment

cuff arm aligned

Analysis

This seems to be due because the distance between pins axes is designed to be equal on cuff and forearm before the bending operation, as we can see on the same video (second screen copy). And because of the S-shape of cuff arms, the width of cuff will become bigger than the width of forearm after bending, which makes it difficult to assembly, and also generates a misalignment (see third screen copy aside) of the arm and forearm on the final prosthesis which may hurt the arm in case of intensive use.

Solution proposed

That is the reason why we decided to modify the forearm pins distance, so that the final distance after bending of the forearm is the same as on the cuff. As a result, the forearm becomes well round-bended, and moreover the misalignment is suppressed.

Result

The final arm of Matteo is shown on the fourth picture beside. One can see that the forearm is well round bended, and well aligned with the cuff shape. The STL file provided includes this width modification.

Cuff modification

As the idea of printing parts flat before thermal bending with hot water or hair dryer seems to be a very good process providing both easy printing and really increased strength (no more delamination of PLA layers under load), we wondered why it has not been done for the whole arm cuff, and decided to try it.

Thus starting from the good shape of the “UnLimbited” cuff, we redesigned it with regular thickness to make the two arms bendable, and flattened to make it printable. Then a new jig has been designed to ensure the original dimensions after bending. The green part on the second picture is the new jig. Red arrows show the process of bending the S-shape of the two arms of the cuff.

Flat printable cuff

Bending cuff arm

Note also that we propose to slightly bend the lever on the cuff, which pulls the tendons (see on third figure aside, yellow area). This is just to avoid having this lever prominent from the arm. The drawback of this modification is that it slightly decreases the distance between rotation axis and pulling point, thus it requires a bit more angle of the arm for the same movement of fingers.

Lever and tendons

And finally, a little problem we discovered too late to make the modification on Matteo’s arm, is the thumb tendon is too long when the hand is completely open (see third figure, green area), and we fear that it may catch something (door handle, etc.) when Matteo is playing or running in the house… A solution could be to also put a short PTFE tubing or anything else to close this area, as there is no movement of the cables here except during the tuning phase.


Notes :  Slicing parameters: with a nozzle of 0.4mm and a layer thickness of 0.1mm; 2 perimeters, 3 top/bottom solid layers, 45% infill.