Less non-replicated parts using rope

I had an idea on how to have a larger percentage of the
parts of a RepRap being printed.
Why not get rid of the steel-rods, threaded rods and pulleys?

Instead of printed gears that never fit perfectly and thus
have backlash and a but of free motion when vibrating
I thought about using rope instead as an alternative to threaded rods.

We can easily build cable-pulls instead
of gears. They can be fastened very
easy and can create a lot of force.

Take a slide that shall be moved.
On both ends you have a cable-pull
attached to it.
The ends go around a cable-roll with to meet below in one point.
There a motor has 2 spools.
One giving rope to the left cable-pull
and one taking it from the right one.
Adjusting the tightness is trivial
and can be done in serveral ways.

This would provide the accuracy
of threaded rods without their
length-limit and without their
issues of the rods are not perfectly straight.
It only requires parts that are trivial to print as opposed to
inperfect gears.

---

-------------------------------------------
* homeprototype free 3d design repository
* Blog
* Google+

Hi Marcus,

... in past i built a laserplotter with steel-wires for moving the axes (look here: [forums.reprap.org] ) - you need high tension for the wires, so normal plastic isn't suitable for motor-spools and feeding rolls.

I used lathed aluminium-spools and nylon-coated bearings where i lathed a V-groove in the nylon for the wire.

Solid Nylon, POM or similar plastics can be used, but the reprapped body isn't so rigid ...

Viktor

From what I see on your design you used the wire directly, without a cable-pull
that will easily provide any tension required.
With flexible Nylon-rope instead of rigid wires you will not require as much tension.

The rolls can contain a hole for a skateboard-bearing to give little resistance.

As for the rigid frame. That is something that is easy to change using
the maker-beams. After all, we are free to change any part of the design.

Now I won't say it uses less material then e.g. the Mendel but apart from
motors, elektronics, rope, skateboard-bearings, screws/nuts/washers, cables, extruder-tip there seem to be no parts left that are not printed.

With the reduction provided by cable-pulls it would even be worth
a try to use cheap DC-motors with an optical or magnetic rotary-encoder (thus a servo-motor)
instead of stepper-motors, making the electronics very trivial and the
motors and electronics much cheaper.
It would only have one speed but cannot miss any steps and you get
instant feedback for an auto-shutdown if anything jams. You can even counter
any force applied to the toolhead (e.g. resistance in a milling-head).

PS:
Cable-pulls can have any length. So there is no limit to the size of
the work-area you want. Think parametric reprap-model here. Just scale
it for metric or imperial screws and what size of the printer you prefer
for your workshop.

---

-------------------------------------------
* homeprototype free 3d design repository
* Blog
* Google+

Interesting; I was just thinking about the same thing last night.

Perhaps I've misunderstood your idea of a 'cable-pull'... Ropes do require tension because they stretch. If you have a long rope, and you're trying to get quick accelerations on a heavy mass at the end of it, cable stretch becomes significant. And the longer the cable, the more significant it is. The stiffness of the cable is inversely proportional to its length, and for stretchy materials like nylon, it probably will matter quite a bit. That's why timing belts have steel cable molded into them.

Do you have a picture of a "cable-pull" that you have in mind? The term is too generic for me to look up on google.

> Do you have a picture of a "cable-pull" that you have in mind? The term is too generic for me to look up on google.

No problem.
[www.systemdesign.ch]

The mass should be as small as possible.
I have some design-ideas that only have the tool-head as a moving
mass.

Streching is a property of the type of string you use and some testing
should be done.
A sensible rope that is not hollow streches very little
and if you have N cable pulls on the N sides of an objekt
you are moving (N=2 for a slider or N=4 for free 3D-movement in
a pyramid) then the streching should already be at maximum at
all times. The rope is never loose.

Of cause you can use a kind of bicycle-chain instead of rope
if streching is a real issue but I don't think it will be.

---

-------------------------------------------
* homeprototype free 3d design repository
* Blog
* Google+

I've been giving this some thought as well, because it'd be possible to do away with the belts and cog gear. Elsewhere in the forums, it's been proposed to use the cable that hand-control combat model planes use, but I think that cable is too heavy, unnecessarily so. Nylon-covered, stranded-stainless-steel cable is sold in fishing stores, it's used to make leaders that attach the hook to the nylon, so fish don't bite through the nylon line and swim away. This cable is available in a range of breaking strains, is widely available and dirt cheap. I chose 15kg from 15, 25 and 50 in the store.


I've drawn a couple of rough layouts to get the idea across. These are proposed to eliminate the x and z belts, the y-belt can be eliminated with an adaption of the x, and one of the z drive screws could be deleted as well, I think. The x-drive cable is wind-on, wind-off, cable length is involute, so it doesn't jump across coils and there's no stretch through non-linear winding. The movement is taken care of by the windlass screwing onto the screw at the end, the pitch of the screw being the thickness of the cable.

I think it's better leaving one of the z-screws in place, because there's no chance the z-gantry would settle under gravity if power is lost, in the same way that a worm-and-wheel drive can only transmit power one way.

Murray

Edited 5 time(s). Last edit at 11/11/2009 08:06PM by murd.

Sorry but your cable_pull_x_drive will not work with the precision required.
I'm currently designing a carthesian-bot that should work but AoI is crashing
on me all the time and boolean operations take minutes after you subtracted
3 objects already.

---

-------------------------------------------
* homeprototype free 3d design repository
* Blog
* Google+

Hi Marcus,

... in the images of my linked previous post, especially the image "X-axis.jpg", i used nearly the same geometry (only rotated 90 deg) for the X-axis of a laser-plotter.

With the first prototype and 0.3mm thick steelwires i had some elasticity resulting in oscillation along diagonal paths (visible in the cutting trace) with an amplitude of nearly 0.2mm at a step-resolution of 0.04mm.

The plotting range was dimensioned bigger than DIN-A3 (or something around 460x340mm), so you can estimate the specific lengths of the wires.

Then i changed to 0.6mm/0.75mm high elastic steel-wire (i think there were 49 single fibres in the wire arranged in 7 bundles á 7 fibres) with the same mechanical setup and reduced the elasticity/oszillation below my measuring (or visible) accuracy of 10 microns.

The mass of the moving parts should be something around 2kg with optics, so this could be comparable to the reprap too ...

Viktor

Ah, mea culpa.
I thought about rope, not steel wire. I thought steel wire would cut into
printed parts too easily.

How it that setup tensioned?

---

-------------------------------------------
* homeprototype free 3d design repository
* Blog
* Google+

Hi Marcus,

in the X-axis, where the wire forms an endless loop, i inserted a wire-tensioner ... and in the y-axis both ends are simply fixed with screws to the left-side columns ...

What's not visibele in the images: - i wound the wires several times around the motorized spools, so with tension the friction was enough to avoid slippage ...

Viktor

Edited 1 time(s). Last edit at 11/12/2009 04:49AM by VDX.

In the setup I drew, I envisaged the cable being tensioned by the return pulley at the other end of the drive from the motor, because the cable is locked onto the drive windlass.

I built this scroll saw about thirty years ago. It's driven by a hand-held cake mixer (it had a variable speed control) with reciprocating linkage and cable drive and tension It looks a bit poorly and cobwebby because it's been neglected in the depths of the storage shed for ten or fifteen years. It worked quite well but shook considerably, until I put some bobweights to compensate for the reciprocation linkage. Point five mm cuts in metal up to 10mm if I was patient. If I was impatient the blades got hot and snapped within 20 seconds....it taught me some patience, because the blades weren't easy to change!

MarcusWolschon Wrote:
-------------------------------------------------------
> I'm currently designing a carthesian-bot that
> should work but AoI is crashing
> on me all the time and boolean operations take
> minutes after you subtracted
> 3 objects already.

Try the HeeksCAD that Viktor left a pointer to elsewhere. For a freeware program, it's very, very good.

my mendel will use ball chain because the timing belts are too expensive here for reasons unknown to me

I have had good luck making complicated things with AOI if you use version 2.7 (not later versions) and you simplify the mesh with the solid editor plugin after every two or three boolean operations. Before I learned to do that, it was very frustrating.

Fdavies

Thanks for the hint.
Someone also suggested using the new Beta2 as
boolean seem to have improved in that one.

---

-------------------------------------------
* homeprototype free 3d design repository
* Blog
* Google+