Falcon Hand

For discussion of hand designs.
AdamArabian
Posts: 32
Joined: Mon Feb 16, 2015 2:04 am

Falcon Hand

Postby AdamArabian » Sat Apr 11, 2015 2:43 pm

Jason suggested we start up a dedicated thread for the falcon hand work we're doing.

hero-falcon-handa.jpg
hero-falcon-handa.jpg (69.16 KiB) Viewed 4600 times


Quick summary: The falcon hand is a test bed we built at Seattle Pacific University. Problems we wanted to (and did) solve:
1. Maintainability - we want the design to be repairable by a one-handed individual
2. Hardware free - we wanted to be 100% printable
3. Fully printable in PLA or ABS - I love many of the new materials but many makers don't have them or their printers are not friendly to them.
4. Positionable thumb - we want the thumb to be able to create both a cylindrical (like you're holding a water bottle) grip and key (current raptor hand) grip.
5. Easy to assemble - no (or very few) knots, specifically no knots that are position sensitive
6. Modular Bracer - Along the lines of the picatinny rail system, we want to be able to easily cover up the tensioner system with something more artistic should the user desire
7. Modular palm cover - we want to enable artist and/or designers to easily create custom palm covers without having to redesign the entire palm; we are also using this to test our locking schemes for grip.
8. More anthropomorphic geometry - make it more human like, less robotic like

Our current version is located here:
http://www.thingiverse.com/thing:603039

All portions of the design work well - it is demonstrably easier to build, easier to maintain, and has the functionality we want. The palm cover system isn't good yet though.

It has not yet been fit to a person; in order to do that I have to submit an IRB to the university which I will be doing in the next couple of days.

The design has been idle for the last few months (we're on the quarter system at the university and all of the students who wanted to work on it were unavailable winter quarter) but we're back to it.

Our plan for spring is to try to drive the design home. The design schedule is:

April 10 Palm cover test: Reconfigured design suitable for replacement palm
April 17 Final reconfigured palm / cover meeting target requirements below (including improved aesthetics and snap-in design)
April 24 Aesthetically improved fingers and horizontal snap in configuration
May 1 Replacement thumb snap in and testing of complete system
May 8 Complete system testing
May 15 Documentation
May 22 Documentatin
May 29 Watch band test
June 5 Unified / tethering of tendons
Final Delivrable Completed falcon hand, including
• Modular palm cover
• Aesthetically improved fingers
• Horizontal snap in fingers
• Urethane glove friendly configuration (without thumb rotation)
• Aesthetically and functionally improved palm cover
• Replaced thumb clip for improved maintainance
• Documentation/instructions and video on assembly
• Documentation on shape for replacement palms
Experiments conducted on
• Watch band bracer
• Unified/tethered tendons (single tendon over wrist)

Jason M Bryant
Posts: 198
Joined: Mon Feb 16, 2015 1:59 am

Re: Falcon Hand

Postby Jason M Bryant » Sat Apr 11, 2015 3:52 pm

I printed out a few of the parts so I could play with the individual elements. I had a bit of trouble with that. I had to re-orient them and add in some supports. What kind of printer are you guys using? It must be quite different from mine. The fingertip really has no flat surfaces on it, so it was hard to find a good orientation for printing it.

First off, I printed out a finger and a phalange. I tried putting a rubber band around the post on the finger, and it broke instantly. That's a tiny little post, and the way I had to print it made the layers prone to sheering.

Right now I'm working on building a version of the Raptor Reloaded with these kinds of fingers. Based on what happened with the Falcon finger, I'm making the posts much, much bigger.

DSCF3260.JPG
DSCF3260.JPG (57.38 KiB) Viewed 4597 times


Note that I don't have any walls around the post on the phalange. It pretty much stays covered by the walls of the fingertip, so instead I used that space to make the post bigger.

This will take some more work, especially with figuring out ways to modify the palm. How to design the connection at the thumb is vexing me a bit. I'll post more about that as I make progress.

Second, I printed out the part of the gauntlet that has the tensioner box on it. Crimping is interesting.

I had a lot of trouble getting the crimpers into their slots. I'd strongly suggest a little more rounding on all the edges, both on the crimper and the slots. That would help it get started, and once it is started, getting it in is fine. Plus, is the little cylinder on the bottom really necessary? If the walls surrounding the crimper were higher, as high as the crimper itself, would that make it more solid or mess something up?

The fishing line that I put in there seemed very securely fastened. However, I'm still a little dubious. It seems like this would weaken over time, and the wire would start slipping out. If a little groove gets formed in the slot, then a whole new gauntlet would be needed. Do you think wear is a concern?

I like the idea of crimping. If it works well, then I have a couple of ideas how it can be incorporated into the Gripper Box. Since that only requires two wires, I could fit two crimpers onto the box. Removing the tensioner pins would make for a smaller, sleaker design. On the other hand, one of those crimpers would be getting all the pull of four fingers. That seems like a lot to ask it to hold.

AdamArabian
Posts: 32
Joined: Mon Feb 16, 2015 2:04 am

Re: Falcon Hand

Postby AdamArabian » Sat Apr 11, 2015 7:04 pm

Jason,

We've printed it on an expensive machine (uprint SE) and a cheap one (printrbot).

Some tips: The fingers should be printed on their side, if necessary with internal supports (I get about 50% success without). Sorry if thats not in the instructions. I have to use a pick to clean them out which I don't like but I've not come up with an alternative. The reason thats required is that otherwise you get these two interlocking horizontal layers that are basically grooves that want to grab on each other when the finger flexes.

The posts can be easily made bigger - its a good idea and one we can do as well.

Can you clarify the issues you're having with modifying the palm? What are you going for?

I like your recommended mods to the crimper posts - all easy to do and I agree they are a good idea.

As for wear, so far all I can say it has not been an issue; that does not mean it WONT be an issue; I tend to think you will have wear at the initial insertion point but since the material crimps all the way around as it twists I think it will continue to hold, but until we get long term testing its hard to know.

One of the things my students and I are working on this quarter is exactly what you're suggesting - tying all the lines together at the palm and then sending a single one back to the bracer. This is actually for another reason; cosmetically the folks in Jordan report that the tendon look isn't liked, and we think a single, heavier line running back will be more accepted. If we do that we will probably go back to a prior design that was bigger and used gear type engagement to more positively lock in place but we don't know. So far we have simply not been able to unspool any of these crimps so they may work just fine (perhaps more rotation of crimping will be better). Overall crimping is a solution I like as it allows for the zero knot tying and easy finger replacement.

Jason M Bryant
Posts: 198
Joined: Mon Feb 16, 2015 1:59 am

Re: Falcon Hand

Postby Jason M Bryant » Sun Apr 12, 2015 1:02 am

Printing the finger on the side is a pretty good solution. It will also add strength, since the layers won't sheer off as easily. I didn't do it that way because of exactly the support problems you mentioned. I do my testing at 100%, and it is very difficult to cut out supports in a 1mm gap.

Here are my suggestions for printing it on its side.

Untitled-1.jpg
Untitled-1.jpg (26.49 KiB) Viewed 4591 times


1) As I mentioned before, I suggest cutting away the walls. The walls on the phalange will still be on either side of it, so it will look about the same. You can actually cut away quite a lot, making room for a bigger post.

2) Make the post bigger, but also make the base of the post bigger. Your current design is a wedge shaped post, getting bigger at the top. That's good for getting the band on, but not so good for structural support. If it flares out again at the bottom, that makes the piece stronger.

3) Make your own supports. Some slicer software is good at generating support. Most are bad, and they're bad in different ways. If you figure out exactly where you need supports, the prints will have the least amount necessary to cut away.

Palm modifications for the Raptor:

The issue I'm looking at with the Raptor Reloaded is how to attach the elastic bands to the palm. It looks fairly straightforward for the four fingers, but more problematic for the thumb.

Untitled-2.jpg
Untitled-2.jpg (38.17 KiB) Viewed 4591 times


The shape of the palm at the thumb joint is tricky. There's a squared off gap where it contacts the thumb that is necessary because printers can't print at shallow angles. Just putting a post on top of that joint is going to stick out quite a bit. So that's going to require some thought.

AdamArabian
Posts: 32
Joined: Mon Feb 16, 2015 2:04 am

Re: Falcon Hand

Postby AdamArabian » Sun Apr 12, 2015 1:55 am

I like your ideas for the band posts - we want to come up with a way to hide them away a bit if we can but we're working on some pop-off covers too which should work regardless (perhaps even a bit better as they would have more engagement surfaces)

As for the palm, I'd suggest just running a palm with the finger elastics replaced just for testing. You may find the spacing isn't quite right or some other interference issue - what we usually do is just cut away the rest of the palm and use just a finger block to make a quicker print then see how it works out. We print very few full palms but lots of test blocks.

Once you get the spacing right you can try a full palm and solve the thumb issue.

Jason M Bryant
Posts: 198
Joined: Mon Feb 16, 2015 1:59 am

Re: Falcon Hand

Postby Jason M Bryant » Sun Apr 12, 2015 2:14 am

It would actually be fairly easy to go into my 3D software and cut off the part of the palm that attaches to the thumb. A little 20cm square block of the palm would give me everything I need for testing. That's a good suggestion.

Jason M Bryant
Posts: 198
Joined: Mon Feb 16, 2015 1:59 am

Re: Falcon Hand

Postby Jason M Bryant » Sun Apr 12, 2015 2:27 am

By the way, I had a crazy though for finger hinges with your design. The little holes where you put filament to act as hinges between the joints? Without drilling them out, they're the perfect size for a straightened out paper clip. Put the paperclip in, bend down the sides, and clip off the excess. If you've got a couple little grooves on the sides of the fingers, the part of the paperclip you bent down can fit right in them and be flush with the sides.

You'd have a very smooth, metal pin for the fingers to swivel around. No drilling, so heating up the ends, you just need a pair of pliers to snip off the ends.

AdamArabian
Posts: 32
Joined: Mon Feb 16, 2015 2:04 am

Re: Falcon Hand

Postby AdamArabian » Sun Apr 12, 2015 3:52 am

We thought about that but I didn't think I could cut them close enough to the finger. I wouldnt want to leave sharp edges hanging out to grab on fabric or scratch someone.

The filament seems to work pretty well though.

Jason M Bryant
Posts: 198
Joined: Mon Feb 16, 2015 1:59 am

Re: Falcon Hand

Postby Jason M Bryant » Sun Apr 12, 2015 2:53 pm

Adam, have you seen what Skip and Jeff have been doing with a whippletree?

https://plus.google.com/+SkipMeetze/posts/7EnhmxiMvmc

Something like this might be an excellent approach for your hand. It would give you the single line that you want, plus adaptive grip.

AdamArabian
Posts: 32
Joined: Mon Feb 16, 2015 2:04 am

Re: Falcon Hand

Postby AdamArabian » Sun Apr 12, 2015 3:48 pm

I have, and I really like it for a transradial prosthesis. I don't think its a good solution for partial hands though since it requires bulking up the geometry. All the feedback I have (which again, is all international) is that overall our hands are too big, so adding a whippletree box just makes it bigger. I do think though that when/if we use the design for a xradial prosthetic terminal device its a great idea.

I have some concerns about terminating the control lines - I don't have a lot of faith long term in knots for that. May need to come up with a crimping system there.

Jason M Bryant
Posts: 198
Joined: Mon Feb 16, 2015 1:59 am

Re: Falcon Hand

Postby Jason M Bryant » Tue Apr 14, 2015 8:37 am

Here's another question about crimping. From looking at the files, it looks like the hole for the crimper is a slightly bigger radius than the crimper. Is that right? If so, wouldn't that difference cause a problem as you scale up? It seems like what fits perfectly at 100% would have problems at 135%.

AdamArabian
Posts: 32
Joined: Mon Feb 16, 2015 2:04 am

Re: Falcon Hand

Postby AdamArabian » Tue Apr 14, 2015 5:17 pm

Oh absolutely - the design is fully parametric in solidworks but prior to transitioning it to public use it would need to have some intelligent scaling; I'm not sure how openscad works as far as converting the files over but we have some brute force ideas of how to scale things if its not easily converted to openscad.

Long and short, you're right, but thats an issue for a lot of the hand (same problem for scaling the holes used for filament hinges for instance)

If you've got a good solution we'd be glad to hear it.

Jason M Bryant
Posts: 198
Joined: Mon Feb 16, 2015 1:59 am

Re: Falcon Hand

Postby Jason M Bryant » Tue Apr 14, 2015 11:33 pm

No, I don't have a solution, I was just wondering. It sounds like you're working on it. I don't have a 3D program that works with parametric stuff, so I get a little jealous about that.

AdamArabian
Posts: 32
Joined: Mon Feb 16, 2015 2:04 am

Re: Falcon Hand

Postby AdamArabian » Wed Apr 15, 2015 2:18 am

You should try Autodesk Fusion. Its remarkably powerful and free for makers/inventors/hacker types. Andreas Bastian posted some excellent tutorials on how to use it for the eNABLE community. I can link them if you can't find them. I think they're on youtube.

Jason M Bryant
Posts: 198
Joined: Mon Feb 16, 2015 1:59 am

Re: Falcon Hand

Postby Jason M Bryant » Wed Apr 15, 2015 4:53 am

I can't download it. I live in China and autodesk's system for streaming doesn't seem to play well with that.

Jason M Bryant
Posts: 198
Joined: Mon Feb 16, 2015 1:59 am

Re: Falcon Hand

Postby Jason M Bryant » Mon Apr 27, 2015 8:36 am

Above I gave three suggestions for the posts on the fingers. Number 2 was to make the base of the posts flare out so they'd have more strength at the base.

I did that on my Raptor remix, but I ended up taking it out. The posts pushed the rubber bands too far to the side, which made them get caught in the joints. That may not be an issue with your design, depending on how you do it. So it might still be worth experimenting with, just keep an eye out for that problem.

droomurray
Posts: 69
Joined: Mon Feb 16, 2015 4:04 pm

Re: Falcon Hand

Postby droomurray » Mon Apr 27, 2015 8:36 pm

AdamArabian wrote:I have some concerns about terminating the control lines - I don't have a lot of faith long term in knots for that. May need to come up with a crimping system there.


Adam,

I am evaluating a crimping system this week and hope to post results by the weekend :)

Drew.

AdamArabian
Posts: 32
Joined: Mon Feb 16, 2015 2:04 am

Re: Falcon Hand

Postby AdamArabian » Mon Apr 27, 2015 8:43 pm

Crimps for the whippletree? That would be awesome! I don't know that ours would work well (actually, I don't know that they wouldn't either) but the form factor has to be so small I'm not sure how to do it well!

droomurray
Posts: 69
Joined: Mon Feb 16, 2015 4:04 pm

Re: Falcon Hand

Postby droomurray » Mon Apr 27, 2015 10:14 pm

Ah, crimps for normal tendons atm is what I am testing, they are 5mm long.

One test might be to make them smaller as they are weight rated ALLOT higher than needed.

Drew.

AdamArabian
Posts: 32
Joined: Mon Feb 16, 2015 2:04 am

Re: Falcon Hand

Postby AdamArabian » Mon Apr 27, 2015 10:55 pm

Oh ok. There are two crimp designs floating around I assume you've seen - one is ours on the falcon hand, the second is... er... I don't recall who had it but it had sort of a gear tooth design.

droomurray
Posts: 69
Joined: Mon Feb 16, 2015 4:04 pm

Re: Falcon Hand

Postby droomurray » Tue Apr 28, 2015 7:06 am

Jason M Bryant wrote:I can't download it. I live in China and autodesk's system for streaming doesn't seem to play well with that.


Jason have you raised this on autodesk support forum ?
Alternatively you could try a vpn service, i had to use one when in China

Drew

Jason M Bryant
Posts: 198
Joined: Mon Feb 16, 2015 1:59 am

Re: Falcon Hand

Postby Jason M Bryant » Tue Apr 28, 2015 12:33 pm

I seriously doubt that Autodesk could do anything. China has been cracking down on VPNs lately. A ton of stuff that used to work doesn't anymore.

Right now I'm happy using blender 3D. When I head home in early July I'll look into getting Fusion 3D again.

Jason M Bryant
Posts: 198
Joined: Mon Feb 16, 2015 1:59 am

Re: Falcon Hand

Postby Jason M Bryant » Tue May 05, 2015 8:07 am

Adam, you mentioned on the google group that you moved the hinge point down very slightly for the second joint so it has a slightly smaller moment arm. Which joint do you mean, the joint connecting the fingertip to the proximal phalanx?

AdamArabian
Posts: 32
Joined: Mon Feb 16, 2015 2:04 am

Re: Falcon Hand

Postby AdamArabian » Tue May 05, 2015 1:31 pm

Correct. The lever arm (distance between the tendon and the hinge pin) is what causes the movement so be reducing that distance you reduce the torque and thus delay operation of that joint with respect to the other.

We were actually working with two joints in the finger so made that change progressively between the first and second and then second and third.


Return to “Hands”

Who is online

Users browsing this forum: No registered users and 1 guest