Cost reduction (fewer driven axes)

I think that it would be a good idea to cost reduce these printers further so as many people can buy them as possible.

One thing I wondering was if it was possible to have less axes; so I wondered what the bare minimum is.

It looks like you can theoretically make do with a single axis and a plastic mechanism.

If that sounds impossible, consider that with a combination lock on a safe a single knob can dial in millions of combinations, so it's clearly possible to have a single knob set up many different mechanical states.

For example, you could have a stepper motor drive that, when turned clockwise, moves in a particular preset direction (left, right, forwards, backwards, diagonally, up, down) but when you do anticlockwise followed by a sequence of clockwise/anticlockwise it changes the preset state.

Doing this would slow it down a bit, and you would need software changes, but the electronics and motors would be much cheaper. Also you'd need more plastic printed parts for the mechanism, but I think in many ways that that's a good thing, you're swapping cheap plastic for expensive electronics. Although it would print more slowly I think it would increase the overall replication rate, since more people would buy one if it can all be made cheaper.

Till someone came with such mechanism we can only speculate. But I expect it to be complex, error prone and innacurate. Motors arent that expensive, prusa is using two stepper motors on Z axis instead of one used on original mendel because its cheaper than belt drive.

... we had this 'lower-count-of-axes'-discussions several times in the last four years - it's theoretically possible, but won't be effective, as much to complex to realize and way to slow

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

I don't think it would be significantly slower. Most of the time the head is moving in one particular direction and extruding. Any extra time only occurs when something changes, but that's going to be only a small fraction of the time.

Most of the time though, you /are/ doing rapid changes in speed and direction - infill. To test this, run too many programs at once on your pc while printing - if you can slow the comms down, the head will pause slightly from time to time, which shows up as big blobs on the print. On more organic curves as well, you'd need to change X and Y speeds continuously the entire time. Plastic isn't dirt cheap either - a full set of prusa parts costs me more than two stepper motors to print, and the added volume of such a complex mechanism would probably not actually make it cheaper at all.

That's another area of potential improvement, I doubt that a lot of that infill is doing so very much, most of the strength and rigidity of a part is in the skin- making the parts bigger, with more air inside, generally makes them cheaper and stronger.

wolfkeeper Wrote:
-------------------------------------------------------
> I doubt that a lot of that infill is doing so very
> much, most of the strength and rigidity of a part
> is in the skin- making the parts bigger, with more
> air inside, generally makes them cheaper and
> stronger.

I'm not so sure about that assertion - the solid fills/perimeters support/hold the part together and give it form while the infill provides strength. I would say that the method of infill matters much more in terms if structural strength rather than the infill level itself at relative infill densities. Until someone does stress testing on infill patterns and densities for parts printed at varying speeds, layer heights, materials and nozzle sizes I'd say that this is a moot point.

---

- akhlut

Just remember - Iterate, Iterate, Iterate!

[myhomelessmind.blogspot.com]

If you look laminates, tubes, and bones, etc, you see material at the edges is much more important than material in the middle for adding strength.

---

[www.hydraraptor.blogspot.com]

Nophead, I agree with you. But aren't the process for constructing those items are very different from RepRap? Laminates is probably the most apt example, but the process of making a laminate floor is very different, even if the concept of laying down the material in layers is the same. Tubes are naturally strong - bamboo or extruded ABS are good examples there. And bones are more complex than anything that can be produced by RepRap.

The processes that create those items are what makes them strong, not the materials that make them.

In our case the issue is one of maximizing strength while minimizing cost. And since no one has undertaken this aspect of strength vs. cost (as far as I've seen anyway) we can't really get a sense of what the truth is. Maybe that's something someone with access to the proper equipment should start working on. Basic stress testing of Reprapped parts. Then we can do a cost-benefit analysis to determine a printing sweet-spot or zone that achieves this goal.

I just know that I've made several nearly hollow items with minimal infill (<25%) and they are nowhere near as strong as the higher infilled parts when put under load. Also, I've printed low infill parts with the same number of shells and differing infill patterns - honeycomb and rectilinear - and the honeycomb part was superior in stiffness compared to the rectilinear part.

So even if shells matter more than infill how do we get away from infill? In our process it doesn't seem possible for larger parts. Small bridges fine, but what about a 20cm or longer bridge? And generating support for the bridges defeats the purpose - might as well have infill - at least it's structural.

Just a thought, but what about varying the amount of infill density? I guess you could slice a model multiple times and disassemble-reassemble the part based on what density you wanted where and with transition layers to bridge the gaps.

Just my 2 cents.

---

- akhlut

Just remember - Iterate, Iterate, Iterate!

[myhomelessmind.blogspot.com]

You could fairly easily algorithmically generate a fractal/hierarchical foam with spherical voids if you wanted. That would give you arbitrarily big holes and may help handle the bridging (spheres have some degree of self support, and the end result should be pretty damn rigid. Otherwise it's very difficult to produce structures like that using non 3D printing techniques.

Edited 1 time(s). Last edit at 05/07/2012 10:20PM by wolfkeeper.

I've wondered about building mostly hollow objects, but with struts supporting large faces or overhangs. For example, a hollow cube might have struts from the centre base to the middle of the four outer walls, these would be joined to each other by smaller overhangs (and would be internal so wouldn't matter if they were messy) and then from there to the centre top. A bit like a smaller wireframe cube inside. I might mock one up in openscad, but I don't know how easy an algorithm to extrapolate this to weirder shapes would work.

Quote

the process of making a laminate floor is very different,

I didn't mean decorative lamination like floor laminates. They are a paper thin veneer on an MDF like material, so are probably no stronger than MDF.

The things I mean are foamboard that model makers use which is thin card with a sparse foam in between and Dibond that is PE with thin aluminium outer layers. That is almost as stiff as solid aluminium of the same thickness but much lighter. A reprapped part with thick walls and sparse infill is very similar.

---

[www.hydraraptor.blogspot.com]

Human bone is a very good example of nature minimising weight and maximising strength, I know that a lot of research is being done in this area, developing algorithms and multi material structures.

---

Random Precision

You can design this part cost-reduction manually, just take out spherical (preferably) or cylindrical shapes (failing that) in the interior in a hexagonal format. If you use spheres and the holes overlap slightly, then it's open cell. Avoid removing material very near to stress risers (i.e. holes in the part where you apply force). If the part fails due to buckling, doubling the diameter of the part increases rigidity by a factor of 4, but if it fails due to simple lack of material, then you've overdone the material removal!

I did some more thinking about this, you could have a single motor axis that sequentially connected to each of the printer axes.The idea is that a 3D printer is not an analogue device it's digital, so you're moving in discrete steps. You can ensure that with various devices to give a clean step each time.

So it would output:

L
R
L
L
R
L
L
L

and this would be translated by the mechanical hardware as something like:

x:L
y:R
z:L
p:L
x:R
y:L
z:L
p:L

for the z axis presumably rotating right would do nothing, and rotating left would raise the print head. If you want to stay where you are on one axis, just do a left-right.

There's lots of variations on this, which can make the idea work better. The point is not to reduce the number of motors, it's to reduce the electronics.