Monday, August 11, 2008


Larger scale positioning system

The (additive) RepRap doesn't have the rigidity requirements that subtractive CNC machines do. This makes various positioning systems possible. Several have been mentioned (Matt's hexapod 1, 2, Viktor's tripod and Dylan's 4 axis design to name a few).

There is a very simple plotter design based on hanging a pen by two threads. The threads are rolled up or down through pulleys. I won't explain more, just look at these video's:

Another nice spritesmod project project page.

More video's of hektor, here.

The reason this system works (no upward rigidity) is because the write head is pretty light.

A big advantage is that you can work on a large canvas, or, tatada daaa, a LARGE BUILD AREA! Because this approach is not limited to two axes, you could also put a write head on three strings. I know it wouldn't have the precision you'd like, but often the tolerances for larger scale objects are also bigger. I can imagine that someone could pretty easily make a 3 string version of hektor with instead of spray paint a PUR spraycan (Polyurethane)

It would not compete with RepRap's resolution, but it will be manufacturing on larger scale. I can imagine a nice playful version to investigate this method:
You have a three axis system and put the pulleys on three poles, on a beach. There, you connect, besides the threads, a hose to the write head that transfers sand that is colloid. This makes it possible to make 3D sand-sculptures. Nice?

I was also thinking of automated construction sites (being reminded of a patent to print buildings). If you make it VERY big the poles would need to be able to handle a lot of force. This positioning system wouldn't work in that case.

More pro's:
More con's:
I know this isn't a drop-in replacement for the RepRap Darwin (it is not intended as such!), but it certainly has its interesting characteristics. Let me know what you think!

Erik de Bruijn
My RepRap Blog



my gut feeling is that the heaver your print head, the more stable this system would be.

i want to try hacking one together sometime!
My thought would be that if the print head were heavy you'd run into problems with inertia and backlashing. I suppose that really depends on how fast you're moving it.

I like the idea of a RepRap inspired sand castle building machine. Imagine the crowd that would draw (no pun intended).
Regarding inertia, looking at the videos and page in more detail, it seems they have a driver system that always moves the spray can in smooth paths - so seems like moving in arbitrarily tight angular corners would cause oscillation of the print head. This complicates the pathing algorithm, and would add a lot of time if inefficiently implemented, when printing a lot of infills.

A heavier print head might indeed reduce any jitter - maybe some kind of dampening / self correcting logic could be built in also, if there is feedback on the actual position of the print head - but this would get complicated.

It also has problems near the left, right (and presumably upper) edges, where the weight is very unevenly divided on the motors. This could be corrected by restricting the working area to the center region.
wow, man. That's a great idea. Get to it and show us a reprapped prototype!

A few years back, I remember seeing something where a system similar to this used some kind of portland cement to build structures. Can't find the reference right now, but it was basically a similar setup, with 3 chains carrying the cement 'extruder', and depositing blobs of cement as controlled by the computer.
Hi Eric,

last year we've already had some talk about string-robots - look here through the linked threads:,7522,7522#msg-7522

Especially Alex Joni's toy is a simple 3-string-thingy ...

Somewhere there too i mentioned a 'top-down' system, where a string-robot is fixed upside-down to the ceiling and three motorized and synchronized blinds (for the stability) around elevates the building-bed ...

On the sand castle scale, you could use some simple speakers and microphones to get position feedback. Transmit a coded sound pulse from the head, and measure the time of flight to each motor/pulley unit. Either do a double ping, or synch the clocks via wires or radio, and you've got your 3 dimensions.

A free hanging head wouldn't be well constrained by only 3 cables though; it would rock and twist a bit. That whiteboard head is constrained by the whiteboard to only move in 2 dimensions, so two strings work great in that case.

6 cables might work, but it gets messy quickly.
i think my solution is best, of course:

you need six wires to fully constrain the print head. adding mass will just slow the system down overall, shifting resonances to lower frequencies.

also see
fenn: I just saw your idea here. I really like it.
I think with the bowden extruder (light extrusion head) this type of mechanical bot would become feasible!

The lighter the head, the faster the moves can be.
I found microfactory by
Dae Kyung Ahn:

He implements this string-based system to create a ultra portable cutting-plotter.
Post a Comment

Links to this post:

Create a Link

<< Home

This page is powered by Blogger. Isn't yours?

Subscribe to
Posts [Atom]