How would you draw this?

And
What would you use to draw it?

I was thinking about what to use for drive belts and have tried string, bowden cable, toothed belts and screws but I thought that this:

http://www.ebay.co.uk/itm/2309588040...%3AMEBIDX%3AIT

would do the job really well and extremely cheaply.

I bought some and it is easily strong enough to shovel a head and carriage about so I now need to print some pulleys for it which is where my problems start.

I have to have a circumference that is an integer number of pitches, the balls are around 2.3 mm, the rod is 1mm so the pitch is 3.3mm or so. On the face of it it is a trivial problem.

I thought of setting up a great big spread sheet for all possible diameters but the arithmetic is not too easy.

Attachment 3543

I think there should be an easy way in Openscad but I am not familiar enough to use it. That picture is done in Freecad but it is less than ideal and involves loads of working out for each repeat.

No idea about how to do that in Openscad, and a bit dubious over the idea of using that for a drive, in my experience the balls have a little play in those things. On the other hand it would work on a steampunk themed printer. :D

It's a chain, all chains have play until they are under tension. It takes a lot to snap it, I think more than the narrow toothed belts.

look at http://www.123dapp.com/design its easy to use and free

Just a cautionary word before you spend too much time designing a pulley.

Have you examined the accuracy of movement of the extruder you would expect by using this ball link chain and compared that the the accuracy against the expected accuracy of a belt or threaded screw system?

I'm not poo-pooing the experience gained through the design and drawing process, but you might be designing something that does not result in sufficiently accurate movement for good prints.

Old Man Emu

Well I can't test it until I've drawn and printed some pulleys. :)

I can't see any reason why it should be any less accurate than any other system unless I am missing something. If you have anything specific that may make it suffer then please elucidate. :)

I don't really have a use for it anyway, just messing.

Quote:

---

Originally Posted by Mjolinor

I don't really have a use for it anyway, just messing.

---

Ah well! In that case, go for it. It will be a good exercise.

I agree that on first glance, it should be accurate enough, but I'd do the calculations.

I have a similar device which rolls a blind up and down at my front door. I was going to pull it down and photograph it, but the blind is inside the house and that's not my territory. I'll have to get approval from the wife to play with it.

OME

Quote:

---

Originally Posted by old man emu

I'll have to get approval from the wife to play with it.

OME

---

:eek: You have a wife that gives approval, where did you find that particular make/model? Mine is strictly the disapproving type, I thought wives all were.

Funny about that. She was born in Bedford, raised near Burnham-on-Crouch, so she started off like yours. Must've of been the mellowing under the Southern Sun that changed things.

OME

I have built a polargraph, which uses this very mechanism to move a pen holder around on a vertical drawing surface (check it out - it's extremely cool) http://www.polargraph.co.uk/.

Handily, there is an openscad design for a parametric ball pulley, specificallly designed for use with stepper motors, on thingiverse here: http://www.thingiverse.com/thing:12403

My experience using these is that accuracy suffers as the engineering tolerances on the ball chain are not that great. YMMV, but it's worth giving it a go.

Quote:

---

Originally Posted by 3dkarma

Handily, there is an openscad design for a parametric ball pulley, specificallly designed for use with stepper motors, on thingiverse here: http://www.thingiverse.com/thing:12403

My experience using these is that accuracy suffers as the engineering tolerances on the ball chain are not that great. YMMV, but it's worth giving it a go.

---

I see your point about the accuracy of the ball chain. The chain that Mjolinor showed seems to be at the back end of the quality queue. The ball chain on my blind roller looks a lot better that that. Once again it's a case of paying peanuts.

Since you are both in the Old Dart, you should be able to cooperate easily on this project.

What got me with the Polargraph was the use of Polar geometry rather than Cartesian. If polar geometry allows you to work in 2D, could it be adapted to 3D? Would the adoption of polar geometry required the development of new CAD and CAM software to work with polar geometry?

Old Man Emu

Note: I didn't see what 3dkarma posted until after I started making this. The files he posted are really nice. I decided to leave this post because this OpenScad code is much simpler and easier to understand. But if you are actually going to do something like this, the files 3dkarma pointed to are much better!!!!

If you plug this into OpenScad:

Code:

---

pully();

module pully() {
    difference() {
        cylinder(h=15, r=30, $fn=25);                    // main pully block

        translate([0,0,-.1]) cylinder(h=25, r=4, $fn=10);        // drill center hold

// add a path for the wire connecting the balls on the chain
       
        translate([0,0,(15/2)-(2/2)]) difference() {
            cylinder(h=2, r=32, $fn=20);   
            cylinder(h=2, r=29, $fn=20);   
        }

// add holes for ball chain  ---  assume 25 balls fit around the pully
// also assume the balls are 2mm in radious

        for(i=[0:360/25:360-.1]) {               
            rotate([0,0,i]) translate([30,0, 15/2]) sphere( r=2, $fn=15);
        }

    }
}

---

You will get something that looks like this:


Attachment 3654

Quote:

---

Originally Posted by Roxy

. I decided to leave this post because this OpenScad code is much simpler and easier to understand.

---

That's what I hate about OpenScad. If you read the code after reading the comments and looking at the image, OpenScad looks like a very clear and concise way to go from the mind's eye to the printer plate. However it is finding out about the syntax and terms used that slows me down.

Am I correct in saying that if I watched a few tutorial videos , it would not be long before I uninstalled my Rhino?

OME

Excellent.

That is just perfect.

Roxy, you Jezebel! You are enticing me to the Dark Side

I watched some OpenScad tutorials tonight and I think I'm getting the idea of it. It is very logical, and once you understand creating modules I think it would be a quicker way to go from a sketch on the back of a beer coaster to an stl file for slicing.

Maybe Old Man Emu's Rhino might become extinct.

OME

It is a dog to learn is Openscad. I have had a few goes but I think it needs a few hours concentration which I never seem to get without interruption.

I have a bit more incentive now and gese that pulley making looks so simple once you have the solution, like most things in life.

openscad is seriously easy to learn.

Admittedly it's hard to look at a really elegant script and work out what the hell is going on. But I was never in the school of elegant programming anyway.

I could easily produce a fully parametric script to make those pulleys. But it wouldn't look anything like roxy's script. That's way to elegant for me :-)

People make the mistake of thinking you have to work out all the maths first and then write the script.
Nope.
You can bodge it all together just like a drawing cad package.
Just hit the f5 button each time you do something and it'll draw what you typed.
I often move things into place by changing a number, redrawing, looking at it all ways up and then repeating.
Just like dragging things around in a gui package.

I find if I do try and work things out, I'm invariably wrong and the suck-it-and-see method is often easier and quicker :-)

One of the great things about it, is how easy it is to move bits of the script around, or copy them. With just the basic rotate, translate, hull, difference and union commands you can make pretty much anything.

Adding loops is great for repetition, but copy and paste works as well for longer less elegant scripts.

I honestly do not understand why people think it's difficult to use. The syntax is simply remembering to add the right brackets at the right place. Why some of them are there - who knows, and it doesn't matter.

And you're always learning stuff. For example, I had no idea you could vary the facet numbers within a script. Thought that was a global function only. That will come in useful :-)

@Roxy why did you use 25 facets for the pulley ? and only 20 for the spheres (well at 20 facets they're not actually spheres). I always go for the max 100 so round things are actually round.

One thing you will need to be mindful of is chordal action - just like chains they are susceptible to this with increasing magnitude as the pulley diameter decreases. If you have a 30 tooth "sprocket" it really a 30-sided polygon where each segment is the pitch, not a wheel with a circumference of 30 * pitch.

That is also true for toothed belts to a certain extent.

Quote:

---

Originally Posted by curious aardvark

@Roxy why did you use 25 facets for the pulley ? and only 20 for the spheres (well at 20 facets they're not actually spheres). I always go for the max 100 so round things are actually round.

---

Mostly this was just thrown together... At the size of the pully I specified you are going to see edges unless you bump the facet number way up. Also, I was kind of thinking that each ball is connected to the next one via a straight line. So having the pully have flat edges all the way around sort of made sense to me. Like I said, it was kind of just thrown together.

Both the pully and spheres could have a big number like $fn=100 but you really don't want to do that while developing the logic. I have a really fast machine with lots of memory and that short little script would bog down with those numbers in it.

Remember, OpenScad is working with lots and lots of surfaces. If you specify $fn=100 that is way more work for it than to do things at $fn=20. Maybe its worth it when you have the logic right and are ready to render and print.

Quote:

---

Originally Posted by curious aardvark

I often move things into place by changing a number, redrawing, looking at it all ways up and then repeating.

---

I work out the math, realise I got it wrong, then move stuff around using your method until I get it right.

I've recently dropped an amd am1 mini-itx board into my little wooden box and a real cheap quadcore chip.
Openscad renders reasonably quick.

My point was actually that the round balls and round linking rods would fit better in round sockets. Okay so the final f6 render might take a few minutes. But you end up with a more functional model. And the f5 redraws are pretty instant even with fn=100.

So no biggy.

But have to thank you for using individual facet functions. I had no idea that was possible and even wrote a polygon script to get round the fact that I could have a six sided cylinder.
Now I know different :-)

I'd have to say the biggest isue with openscad is the manual. It's written by someone who knows the program inside out and doesn't understand the simple fact that most people don't.
So any time I've looked at the manual I've come away thoroughly confused.
Whereas tutorials and forum answers have always made sense - thank god lol

One thing that does make it easier is to use diameters rather than radiuses. That way I don't have to remember that double or halve the radius. Just simplifies calculations.

That $fn - is great I can set it as a global variable so everything has 100 facets, and also assign a lower number for polygons to individual shapes.
You have no idea how much that helps me :-)

It works well with the chain. The pulley openscad file had to be changed a bit to make it remove less of the sphere so it held the chain better but other than that it seems fine.

For the first couple of hours running I had to keep adjusting the tension but after that it has stayed tight. I left it running like the video overnight with no problems or slipping.

http://homepage.ntlworld.com/dknpower/chain.mp4