Pulley / Fraying - "Mechanical Flower" Project

[Fuichris]

Hi all,
"MECHANICAL FLOWER PROJECT"
I'm working on a personal hobby art project (non-academic, non-commercial, non-governmental) where I'm mechanizing petals of a plastic flower. Basically, making petals curl open and close.
*Please see picture and video.

GOAL:
To pull continously without fraying or breaking the string (duration: 6-8 months or longer).

MATERIALS:
1) braided fishing line (20lbs, .24mm, Spiderwire brand, ultracast)
2) Parallax servo horn (4 pronged turning pulley)
3) Tube #1 (surgical - syringe, .5mm thick, very soft)
4) Tube #2 (thick hobby craft tube, 1mm thick, rougher)

As you see from the picture and video, I have used tubes for the two (2) 'friction' points: (POINT A) Coming out from the flower within the copper tube above; and (POINT B) as the string loops into the Servo which pulls the string.

(Note: I have not tied any knots on the servo because I was afraid of the friction at POINT B, so instead, I just looped the string around the servo a million times.)

Desperately seeking some engineering magical expertise.

Many thanks!
Chris

https://player.vimeo.com/video/109680749

 

[Danger]

Hi, Fuichris.
Meaning no insult, I can't quite figure out the point of your question or the elaborateness of your setup. To start with, there's virtually no chance that 20# braided fishline is going to fray sliding through a latex tube, or a copper one (unless you leave a ragged edge). Secondly, there should be no friction at point B other than the back-and-forth flexion of the string. That would contribute to metal fatigue (marginally) in a wire, but won't affect string. Your tubing and wrapping, therefore, are extraneous. All that you need to do is bring the string back up from the back side of the servo arm hole and tie it to itself, or hot-glue it in place, or run a bolt and nut with a washer through the hole as a clamp.
From the beginning of the clip, I didn't actually see any movement of the flower (although I don't have my reading glasses on), and the servo point of connection barely moves.
Can you please be more specific about your concerns?

 

[Fuichris]

Hi Danger,
Appreciate the help and comments. :D

STRING POINT A: Previously, the string at point A was slanted and it rubbed on the copper tube and frayed a bit, which is why I put on that surgical tube at the exit point of the copper tube, and hot-glued it together. Sometimes the angle of the string gets moved unexpected, so I had to be proactive and protect it. I need the project to run indefinitely for 6+ months if need be without me fidgeting around it so I was putting all these fail-safes. It frayed twice previously because it was angled too much.

STRING POINT B: The yellow hobby tubing at point B was to avoid friction between the string and the servo horn. Previously, I had inserted the string through the 'easy-connector' (see pic) and screwed the string down as you mentioned, but the constant pulling and tension eventually broke the string within 24hrs (three times already). So this is also a problem. The reason for wrapping the string around like a madman is because: a) I can't knot it since I need to adjust the tension often and untying was difficult ; b) and I thought wrapping would alleviate the pressure/tension at anyone point and avoid causing it to fray/break. The pressure of the pull is constant and very strong.

I'm making a lot of uneducated assumptions in this setup so I'm hoping to get some advice :D

MAIN QUESTION: Is my setup sustainable for the long term or are there areas where the string will break somewhere in my setup?

Tks!

 

[Danger]

Okay... I'm still a bit lost, but two things that you mentioned this time (not in original) are the magnitude of the load and the need for adjustability. Those make a huge difference. I'm going to retire with my TV, beer and pizza, and ponder this some more. Right now, the only thing that comes to mind is using a turnbuckle to attach the horn and string.

 

[Danger]

I got interrupted by real life, but I'm back now. So... you're not going to like some of the things that I say, but I'm honestly trying to give you my best advice.
Point #1 is that no matter how often or how closely I look at that video, I can't see any movement whatsoever in the flower. It isn't on camera long enough. Your written explanation is lacking detail. There's a big difference between having some curling at the petal tips and having them close up like a rosebud.
I can't see where the copper tube is. In order to protect the string from scraping across its ends, you have to have your latex tubing inserted inside those ends, not just crammed on from the outside. Is that what you did? Otherwise, you're just hiding the wear-points from view.
There's no point in tossing a perfectly good servo since you already have it, but you'd have been much better off going with a crankshaft on a normal motor.
The wrapping is totally unnecessary other than to store up a lot of spare string. The stress relief that you mentioned is non-existent, since there is essentially no stress upon the string once it's through the hole and secured. For the tension adjustment, I still think that a turnbuckle would be best, unless a progressive spring would suit.
While it would be very dirty, I would recommend firing graphite powder into the tube. Since the guts are out of sight, the mess shouldn't be a problem. Just be sure not to let it get inside the servo housing or into contact with other electrical bits; it's highly conductive.
My own approach to the whole thing, if the budget allows, would be to forget the whole string thing and insert shape-memory metal wires into the petals. Whenever you apply current, they'll curl up; turn it off and they'll return to the rest state.
I'll wait for your response to this before commenting further. I might be way off-base, and don't want to continue without consultation.

 

[256bits]

It looks like solid nylon fishing line rather than braided, which doesn't make that much of a difference except flexibility on tight corners, but that is not the problem per se.

You have already seen that any sideways motion of the line against another surface will cause failure. So eliminate the contact will surfaces moving lateral with one anothe by having a clearance between the line and the part.

Cable works best in tension. Eliminate any places where there is a bending motions in operation. One such place is where the line is glued to the motor arm. Attach a wire to the end of the line, and then the wire to the arm. Danger mentioned a spring - add one somwhere to smooth out tension forces, possibly instead of the wire.Make sure the wire, or spring, at that located works smoothly with no rubbing or bending. If you have to offset by attaching to a bolt or pin through the arm, then do so.

Flexible tubing. If the line wears through then you will get binding. The flex could be for the dust bin. Sure the flexing looks like it is doing something but what really other than flexing. Pass the line through a smooth end grommet of harder material that won't wear out. Make sure your line runs always straight into the copper tube through a grommet after passing over a large radius curve would be better, if the angle of the line is excessive.

Points of summary
- no, or minimum bending of the line, except over large radius
- no sideways rubbing of the line against other surfaces

 

[Fuichris]

Hi Danger!

- "I can't see any movement whatsoever in the flower."
Sorry about that. Try this link, you will need to scroll through to the end of the short video:
https://player.vimeo.com/video/108456552

- "I can't see where the copper tube is."
Yes, the latex tubing is inserted inside the copper tubing.

- "wrapping string ...essentially no stress upon the string once it's through the hole and secured.."
My worry was how how it is secured through the hole of the servo horn. A knot is difficult to adjust, and putting it through the easy-connector will eventually break the string.

- "...insert shape-memory metal wires into the petals."
Wow! What's this? As you will see from the video, the petals are super tiny, about 1-2cm. Are the wires quite small? Any advice on where to buy and test out?

Super advice!

 

[Fuichris]

Thanks 256bits!
Can I clarify with you on your comments:

- "Eliminate any places where there is a bending motions in operation. "
Did you mean even the string going through the hole of the servo arm is bad as well? (like the way i did it?)

- "One such place is where the line is glued to the motor arm."
Actually, my bad, my pic is not in the best angle. It's the yellow tube that is glued to the motor arm as my way of cushioning the string thru the hole.

- "Attach a wire to the end of the line, and then the wire to the arm."
I just tried your idea of the wire + string... see pic... is this what you mean? Problems: tying a knot to the wire makes it hard for me to adjust, and I think the wire may be too soft because I noticed that if I pull hard enough, the wire will loosen? Am I doing this right? Danger's idea of a shape-memory wire is sound great at this point ;p

 

[Fuichris]

NEW: LOOSENING GRIP :L
I also just noticed that the line is also loosening it's grip as well. So my 'wrapping' the string around the servo arm and securing it through the easy-connector is not tight enough. I notice it just by looking at the tension of the strings, but also because the petals don't close or move as much.
The energy is going elsewhere.

 

[Danger]

Nice video; I really enjoyed it and it certainly clears up several of my questions. The scale took me by surprise. Initially, I thought that it was much larger. Given that, NiTiNOL wire probably won't work. If you want to check it out, though, there are dozens of sources on line. The type that I particularly like is called BioMetal, which was (and I hope still is) supplied by Toki Industries. Rather than having to supply heat from an external device to heat it into shape-shifting mode, it has a calibrated built-in resistance so you just run a pulse-modulated current through it and it heats itself. I think that it's too large for your project. The "maple connectors" to anchor one in place and supply power are probably about the same size as your petals.I did see somewhere (probably Daily Planet) that there's now a shape-memory polymer that might be applicable, but I don't know where to find it.
I can't help thinking that some sort of cloth would make a better hinge than the plastic for your "spines". You can't break it.
It's bed time for me now (8:45 am; I've been up for about 24 hours), so I'll hold further research and posting until I wake up.

edit: Okay, I'm awake again. Although I sleep 10—14 hours out of 24, it's seldom in stretches of more than 2—6. This is probably going to be a short period of wakefulness, though.
Anyhow, Toki still makes BioMetal, and it's available through Images Scientific Products. One of their pages (home is www.imagesco.com) gives a small sub-index of their NiTiNOL products. The home page is the main menu with all sorts of other cool robotic-type gizmos indexed. I didn't bother opening any specific descriptions, except the one for BioMetal springs. Those didn't exist back when I was designing things around the stuff; it was only in 1 or 2 mm diameter wires. The other problem back then was that after being deformed by whatever force or mechanism, the wires snapped back to their original shape in a couple of milliseconds. I forgot about that until now. I'm still too tired to try reading any of the stuff, but I do know that the problem has been dealt with because I once saw on Daily Planet a high-tech evening gown that gracefully altered its form by the model tweaking the control module. As for the plastic stuff... I Googled "shape-memory polymer" and got close to 2,000,000 hits. I didn't click any of them, but you should have more information at your disposal than you could possibly need.
I'm off to harass others on PF for a while. I might check in here again before going back to sleep. If not, it could be a few more hours.
By the bye, the name NiTiNOL derives from Nickel-Titanium alloy developed by the Naval Ordnance Laboratory.

2nd edit: I just decided to take a peek at your bio and clicked on your website link. Holy crap! I had figured that you were a youngster still in school, not a world-renowned installation artist! I don't think that I'm anywhere near qualified to help you with such a project (but I'm going to keep trying).

 

[Fuichris]

Hi Danger!
Thanks for the great advice! My brain and skill level is still as at a very junior level, sadly ;p that's why I need expert advice :D Need help!
Look what I found!


I love the idea of a shape-memory alloy to curl the petal. I really want this to work because the string and pulley is quite inconsistent but I still don't see anything that fits yet due to the size and stimuli, but i could be totally be wrong.

FLEXINOL WIRE ~ 70 DEGREES: This will burn through my plastic and fabric no?
BIOMETAL FIBER : the wire is thin (0.15mm) enough but the crimps and are too big I think.
Shape-Memory Polymer: I checked it out online but heat, water and ethanol seem to be the only stimulus. Not electrical.
ON/OFF: How would i keep it plugged in, yet have it control the amount of heat that goes to it? a controller of sorts? Pulse Width Modulation circuit is needed so that it can turn the electricity on and off, yes? So basically it's an AUTO on/off switch, but I still can turn the whole thing off.

http://www.imagesco.com/catalog/nitinol/index.html

 

[Fuichris]

FLEXINOL + copper tape seems to be the answer! :D
I'm not sure how I will activate it through the copper tube yet.
I'll do what they did in the video and weave the flexinol around copper tape and thru the petal fabric, and then ... and then ... haven't gotten to the activation part yet... ;) Jie Qi videos from MIT media lab looks perfect, but I'm still uncertain abt activation though.

 

[Danger]

Well, now...
At this point, you've probably learned more about the stuff than I knew to start with just from researching on-line. (There was no internet when I was into it; my information came from Toki's distributed data sheets.)
I can't see that any reasonable plastic or cloth would have a problem handling 70°. That's less than 3/4 of the way between the freezing and boiling points of water. (Unless you mean degrees Fahrenheit, in which case it's slightly above the boiling point of my ex-wife.)
As for the heating mechanism for normal NiTiNOL, I would just expect to use a resistance heater such as one scavenged from a hot glue gun. As mentioned previously, BioMetal heats itself just by being part of an electrical circuit. The pulse-width controller is supplied by the same company that makes the metal. Once you set it to the right modulation, you could put a timer between it and the wall plug to take care of actuation times. (Unless newer models have timers built in; that would be a nice touch.)
I have a tickle in the back of my brain as to something that might overcome the problem of connector size for BioMetal. If you clamp one end of the wire under the table, run it up through the spine to the far end of the petal, loop it around and run it back down... you can then clamp the other end under the table as well, thus eliminating the need for compactness. I'm not sure that it would work properly, and it would require a lot of planning and experimentation, but it might be worth a try.
I can't see any way that pumping ethanol through your flower could be a good thing, so maybe avoid the polymer...

edit: I see that you posted again while I was composing this one. I'm wondering why you need the copper tube for guidance. If it's such a pain in the ass, have you thought about trying to set up some nylon pulleys or bushings instead, only at turning points and leaving the rest open? It would certainly minimize any frictional losses and wearing.

 

[Fuichris]

Hey Danger!

NEW LAYOUT: hmmmm... I thought maybe the copper tube could not be used to funnel the + and - wires of power to reach the Flexinol? And the tube would stretch under the table where it would plug into my microcontroller and the power adapter?

Thoughts? I wanted to keep it as simple as possible for longevity purposes.

:D

 

[Danger]

Right, I see what you're referring to, but it's actually more complex than what I have in mind. With BioMetal, there's no need to have any mechanical design beyond the opening in the table, since there are no bulky motors to keep out of each other's way.
Keep in mind that this is just a rough suggestion, not a solid plan. Drill your access hole through the table and line it with a nylon bushing (or put pulleys on both sides of it). Insulate the wire (dipping it in latex or liquid vinyl should suffice) so that it remains flexible but won't short out against itself. Screw you're first maple connector, or whatever type of clamp they use these days, to the underside of the table right beside the hole. Screw the second connector nearby. Fold the wire double and thread it up to the end of the petal. Connect one end of the wire to each maple, and then run the power leads from those to where ever you want to stash the controller.
As for longevity, that used to be listed on the data sheets. I can't remember specifically now, since I haven't seen the things in over 30 years, but I believe that the 1 mm wire was rated at well over 100,000 flex/extend cycles continuously at high speed (such as moving the wings of a model dragon fly at real-time rates).

 

[Fuichris]

Hey Danger,
I'm trying to visualize what you are describing. I'm going to try and draw it out. I'm going to put an order of the Flexinol, but then I'm not sure the difference between the BioMetal and the Flexinol still. The videos I sent you are all Flexinol.

I'm drawing and visualizing ... before I start buying all the wires and copper tape. :D

 

[Danger]

This is very frustrating in that I used to just draw stuff up in Illustrator, move it to Photoshop for translation to jpeg format, upload it to ImageShack, and post a link. None of my Adobe stuff will work on this computer, so I have immediate access to Inkscape and TurboCad only but I do still have one of my old G3's and even a 200c with the software. The G3 is still on-line, but hard to get at physically. ImageShack has changed so much that I haven't been able to view the stuff that I've had stashed there for years, let alone add new ones, so I'd have to find a new image hosting site or figure out some way to afford a paid upgrade to ImageShack.
Oh... hey now! Maybe I can draw the thing up and then just e-mail you an attachment. I think that I've done that once before with a picture to a friend, so I'll give it a try. If I'm not back within 3 days, send in a search party.

 

[Fuichris]

Hehehe... technology moves way to fast for me as well!
Check this out... i like the sewing method. But not sure about the powering it up. i have a feeling I don't need all the servo and homework board anymore. hmmm...

 

[Danger]

Well, we still seem to be thinking in different directions, but it's your project and I don't want to mess it up.
The one thing that I'm sure you've noticed is that artists and scientists don't generally think the same way. The former tend to be right-brained (and therefore often left-handed) while the latter are usually left-brained (and right-handed). I'm in an odd-ball limbo. I'm ambidextrous with no brain hemisphere dominance. My artistic and mechanical aspects intermingle without much control on my part, but I'm not really good at either one. It gives me a lot of creative freedom, such as my unhesitating willingness to, say, incorporate a ferret into an exoskeleton design, but there's always an awful lot of editing that needs to take place along the way (such as how to feed it. :D) Anyhow, my point is that I have a bit of trouble communicating fluently with either artists or scientists because I'm not really either one. I'm going to seriously dive into Inkscape and draw up what I'm trying to convey, and then find some way to get the image to you. I'll continue checking in here during rest breaks. (Drawing shouldn't take long once I get started, but trying get it to you might be a bugger.)

 

[Fuichris]

Thinking in different directions is a good thing ;p Hehe!
I really want to know what you are imagining most importantly. The drawing i sent above is not feasible... it's a hodgepodge of what I saw with the Flexinol videos and my own setup... but I think i have to throw my own setup out the window for a different setup altogether, which I'm very open to!

Looking forward to seeing what you are thinking of! :D
I'll send you my email in a PM now...

 

[Danger]

Okay, I've got it and stashed it in my address book. Now I'll go and start drawing. While I'm doing that, could you please mail me a new ferret? For some reason, nobody in North America will sell me one any more. :(

By the bye, there's a weird aspect to that ambidextrousness that I'm wondering if you've ever encountered in your circle of acquaintances. When using a pencil and paper rather than a computer, I do all of my draughting and mechanical sketching right-handed, but if I want to draw something artistic like a horse I have to use my left. (Of course, it doesn't actually look anything like a horse, but it's a lot better than I can do with my right.) Likewise, I write fiction with my left hand and non-fiction with my right. (I can do both with either, but that's what's most comfortable and productive.)

 

[Fuichris]

Goodness! Ferret?! Why don't sell it to you anymore? What do you do with them ;p

I've never encountered this kind of ambidestrouness... interesting though. You're utilizing every part of your intuition together with you physicality! That's good and important!

 

[Danger]

Goodness! Ferret?! Why don't sell it to you anymore? What do you do with them ;p

I've never encountered this kind of ambidestrouness... interesting though. You're utilizing every part of your intuition together with you physicality! That's good and important!

Interesting, perhaps, but annoying. It sure cuts down on the excitement of my sex life; I can't switch hands to make it seem like a stranger since I'm used to both.
As to the ferret... there was an actuator feedback loop malfunction in that last exoskeleton, and, well... it turns out that ferrets conduct electricity a lot better than you might expect.

 

[Danger]

I'm bumping this thread just for the sake of updating it. I'm still working on the drawing. A combination of health issues, terrible time-management skills, and a couple of mains power losses have caused some delays. It shouldn't be long now...

 

[Fuichris]

I'm bumping this thread just for the sake of updating it. I'm still working on the drawing. A combination of health issues, terrible time-management skills, and a couple of mains power losses have caused some delays. It shouldn't be long now...

Hey Danger,
Now worries, but hold on. I'm just getting my flexinol wires today and will experiment with it first. My various 'string/fishing line' tests are breaking still and the tension varies. So the consistency is not happening. Plan B is the muscle wires. Need something that runs 1 million times without fail. Hmmm...anyway, learning pbasic and breadboarding. I'll keep you up to date. Could it be as simple as muscle wire in the petals activated by wires thru the stem? Hope so.
Tks!

 

[Danger]

Could it be as simple as muscle wire in the petals activated by wires thru the stem?

That's what I was aiming for. The difference is that (at least in the case of BioMetal), the clamps that connect the active wire to the terminal leads are too big to fit in the stem so my drawing shows them under the table. Of course, that means that the stem would have to be made of metal or otherwise reinforced to keep it from curling as well. Anyhow, I'll hold off any more input until you see what you come up with.

 

[Fuichris]

That's what I was aiming for. The difference is that (at least in the case of BioMetal), the clamps that connect the active wire to the terminal leads are too big to fit in the stem so my drawing shows them under the table. Of course, that means that the stem would have to be made of metal or otherwise reinforced to keep it from curling as well. Anyhow, I'll hold off any more input until you see what you come up with.

OK :D
I'm literally working it as we speak, and reading up on the What's a Microcontroller pdf. All making sense so far... more soon.

 

[Danger]

O
I'm literally working it as we speak

No, you're figuratively working as we speak. You are literally working as we communicate. We've never spoken a word to each other. :p

 

[Fuichris]

Well! My first test of the flexinol wires didn't go so well. If you know of a link to how to hook up flexinol wires to a breadboard / microcontroller, let me know ;p

Crash and burn!

 

[Danger]

Bummer! :(
Did you use the wire manufacturer's controller design, or something else? The pictures make it look as if the thing just snapped into shape instantly and ripped through the fabric rather than curling gently.
If you can't control the wire's inherent transition speed, all that I can think of is to replace your return spring with a much stronger progressive one to provide velocity moderation.

 

[Fuichris]

Howdy!
The Flexinol wire manufacturers don't have their own design...I've been looking online on how others have managed using the muscle wire and the hooking it up to the PC board, but still no luck ;( The snapping into shape is because there is too much current. Stilll tinkering ...

 

[Danger]

The snapping into shape is because there is too much current.

Are you sure about that? I haven't looked into Flexinol; I've been leaving that up to you. BioMetal, as I mentioned, has a hysteresis loop in its thermal cycle. At low temperatures, you can pull it to about 105% (or whatever disappointingly small percentage) of its normal length and it will stay there. When you heat it, it doesn't react at all until you hit its transition temperature (let's say 150° C as an example, but I really have no idea) at which point it spontaneously snaps back to its "resting" length within a millisecond or two. You can't deform it again until it drops to maybe 120° (again, just an example). I thought that since you had settled on Flexinol, it must be more controllable, but maybe it isn't. Does the literature state that it is, or did you just buy some to experiment with? If the latter, maybe what I was thinking of as the way to go with BioMetal is worth trying. Instead of what you're currently using as a return spring, (I would be using a leaf-type (no, not because you're making a flower, smartass...)) try one that perhaps has a more aggressive progression factor. I don't know whether or not that can be accomplished with a single leaf, or if you would have to go with multiple leaves like in a car suspension system. The bad aspect of that is that unless you have some very tiny shock absorbers in your parts bin, it will uncurl just as rapidly at it would curl without the spring. That would be okay if the flower that you want to model is a Venus flytrap, but your pictures sure don't indicate that. :p The whole issue is why I was hoping that the metallurgy had improved significantly since I dealt with it, so that a gradual transformation could be achieved. It's obvious from that one video that it's possible with some memory substance, but I don't know what it is.
Aw, nuts! Now I'm starting to think that my design won't work either. :(
I'll keep thinking on it.

edit: Actually, I'm currently too tired to keep thinking on it. Instead, I'll resume tomorrow (well, later today).

 

[Danger]

Hey now! I've been over-thinking this thing all along. Would a simple bimetallic strip such as from a thermostat work? Just heat the whole thing and let the differential reactions do the work...