Maybe fab ladder chain instead of toothed belt?

It seems to me that one of the more complicated (and expensive) mechanical sections of the Darwin design is the Z-axis drive. The toothed drive belt is not cheap and the matching cogs have too fine a pitch to be printed with a RepRap.

I've read of someone experimenting with bead chain. If the cogs are printable on a RepRap and a way is found to create continuous loops then that would be a fair solution.

I think a more 'printable' solution might be a ladder chain. These can be made in a sufficiently course pitch to allow the printing of both the cogs AND the matching chain links. The links would be made to snap together, much like the treads from a Lego tank. If a printable ladder chain design works then we can eliminate one more 'non-printable' part (well, five if you count the four cogs and the drive belt).

That's a nice idea but I would be a little worried about the amount of play in the chain. Even the smallest space would cause the axis to jerk everytime it reversed.

Here's an Idea:

why not try to use simple string as the belt?

a string wrapping 90 degrees of pully would certainly not generate enough torque, but since the friction force increases exponentially with the number of revolutions around the pulley it might be possible to print 4 pulleys with spiral patterns on them.

each pulley would get 3-4 winds of string and if the string would be even lightly tensioned the friction force should be enough. (or at least that's what my intuition tells me... )

-Leav

BrainZap!(tm)
There is a very convenient spiral already located on the RepRap! the actual screw itself! this is too perfect in my mind in order to work. there must be a problem. :/

Edited 1 time(s). Last edit at 07/02/2008 10:01AM by Leav.

The only problem i can see with string is that it has the problem of expanding over time. Which mean you would either need to re tighten it every few days or over tighten it sou it would last longer... that of course might be to much for it and throw off the alignment and cause other problems. But until someone tries it out this is mostly speculation.

I, and I'm sure others, also had the string thought. eD gave it a lot of thought back in 2006
[www.blogger.com]

BitsFromBytes have a wire based Z stage, not much on the site about it. I understand that Ian is sending out a batch of kits based on this imminently. Assuming I can put mine together, and no-one else has done theirs faster/better I'll photo the mechanism and post here ;-}>

Wrapping the string/wire around the Z axis thread might get rid of the normal problem of multiple wrappings, the thread riding up or down. As long as it doesn't hit the table or the bottom it doesn't matter if it changes height.

One thing to watch for is not introducing large side loadings on the Z axis rods as their spacing is critical to the table moving smoothly.

bye
Pete

@ronanwarrior
I'm sure you could use non-stretchy string. maybe very thin steel cable.

@pete
the fact that i'm not the first to think of this takes away from the glory of discovery a bit, but it is out shined by the brilliance of knowing that my idea might actually work!

-Leav

... here - [forums.reprap.org] and in the folowing posts we disputetd some issues of feeding with belts and wires.

For a sample of driving a laser-plotter with steel-wires look at the atached images in this post - [forums.reprap.org]

In this laser-plotter i used a stationary RF-CO2-Laser and moving mirrors/focus-head for cutting thin plastic-sheets and could sense some resonancies with steppermotor-driving in fullstep-mode, but halfstep or even microstep would smooth the moving, and so would a bigger load (or seismic mass) too.


I have some metres of thin elastic steel-wires from 0,3 to 1mm thickness with 7 to twenty-something single steel-fibres - if it's helpful, i could measure the elongation of a meter of this wires with and without a heavy load, so you'll have a glimpse at the real elasticity of steel-strings.


Viktor

This reminds me of an idea I had earlier and almost forgotten about. You can reduce the amount of belt chain that you need in the X and Y axis almost by a factor of two. A large part of the tooths are not used at all (they only run over the idler). This could be a similar string (with low stretching) which is connected to the belt where there ARE tooths needed. If you use an alternative method for the Z-chain, this would mean you can build twice the amount of RepRaps with the same amount of timing belt

If you use a tensioner that would allow you to keep it tight. Ian Adkins and Ed Sells have made a belt tensioner from two plastic parts which can easily be printed.



I don't have the files here, but Ian or Adrian must have them, though the could be easily redesigned. The thing the belt rolls over is a bearing, but I just used a stack of washers instead which works just as well (hardly any extra friction). I've never had problems of my belt slipping. You can adjust the screws to add or remove tension of the belt (or wire in a future version)!

---

Regards,

Erik de Bruijn
[Ultimaker.com] - [blog.erikdebruijn.nl]

1) can't you just use an extra idler wheel on an arm with a spring, keeping the wire tensed all the time?

2) I must be silly I thought "the actual screw itself" was a non-RP part?

Even if the ladder belt had a tenth of an inch of play that would translate to about 1/100 of a rotation on a 3" diameter cog. If the Z axis threads have a pitch of 1.25mm that would cause a vertical offset of 0.0125 mm. Since the specified positioning accuracy of the RepRap is 0.1 mm I think that will be plenty close enough.

Squintz Wrote:
-------------------------------------------------------
> That's a nice idea but I would be a little worried
> about the amount of play in the chain. Even the
> smallest space would cause the axis to jerk
> everytime it reversed.

A problem with using a string for the Z axis is that it provides no means of keeping the four threaded rods synchronized. If one axis slips, just a little, over time the stage will bind because the four corners are no longer in the same plane.