Re: Another Shapeways report
--- In STMFC@..., "Rob Kirkham" <rdkirkham@...> wrote:
That's what I'm wondering, Rob. Hey, I don't have any inside track on this, I just submitted my part file through the automated front end of the web site, like anybody else. I would have thought they would have built the part with the flat disk surface that makes the end of the valve facing up, to give it the best quality, but apparently not. I actually submitted a dozen of these, joined together on a "runner", and I think they built it standing on end. It really doesn't matter, since the purpose of my test is to see what kind of quality can be had on ALL of the surfaces. This is the perfect part for a test, as far as I'm concerned, there is no "show face", they are ALL equally important.
Let me address Dave's comment about "noise". It appears that the best quality surface is the one that is "normal" to the print process, that is, flat and parallel, like a sheet of paper relative to the print head of a ink jet printer. These surfaces will be flat and smooth, as the face of Tom's boxes are. The next best quality will be on surfaces that are perpendicular to the normal, because each layer should end at the same point. The problem is they don't; there are minute differences in both the size and position of each dot of resin deposited, and the result is a certain amount of surface noise. The next worst surfaces are going to be those that are round, or at an angle to normal, because there is the possibility that the edges of the layers will actually show.
I am now realizing that the process used by Shapewys has another pitfall; the basic process can't build parts with overhangs, since there would be nothing to support the first layer of resin that hangs past the previous layer, so the process used by Shapeways builds a wax structure to support this first layer. Unfortunately, it appears that the wax construction is done at a considerably coarser resolution than the actual part, and any surface built against the wax picks up this coarser resolution. Tom pointed this out when discussing a unsuccessful earlier attempt of his, and If I can borrow a link to his photo:
You can clearly see the striations where the part was built against the wax supports that were needed to build the overhanging features. I believe I'm seeing this same effect on the bottom angular surface of the bowl on the bottom of my valve:
I have no idea why they would degrade the resolution of the process by using a coarser resolution for the support material, but it appears they have.
I'm wondering if these fused disposition process (FDM) is ever going to meet our needs. However, I see stereiolithography (SLA), where the part is grown from liquid polymer in a tank, has considerably finer layer resolution. Here's a video of the process. Scroll to the end to see the finished parts:
Asiga is a new low cost (relatively) entrant into the field, here is some comments from their discussion forum:
Our system has two main factors that affect build quality of the parts, these are listed below:
1. Image Resolution
This refers to the size of the pixel in the exposed image. The Pico and Pico Plus39 have a 39 micron pixel size, the Plus33 has a 33 microns pixel and the Plus27 a 27 microns pixel. The smaller the pixel the sharper the exposed image and the higher quality surface finish and detail resolution.
2. Layer Thickness
The thinner the print layer, the less visible the stepping. Our range of printers have presets from 25 microns through to 150 microns and you can select which layer thickness you require depending on the part you are printing. The 25 microns layer thickness will achieve a much better part resolution compared to the 150 microns style. You can also customize this yourself and print in 1 micron increments if you wish.
A combination of these will determine the resolution of a printed part.
Please let us know if you require clarity. If you are unsure which machine suits your application then let us know what industry/product you are producing.
The build envelope on these machines is somewhat limited; 35 x 22 x 75mm for the finest resolution printers. That's 1 3/8 x 7/8 x 3 inches, approximately. Big enough for parts, not for complete carbodies.
Unfortunately, SLA is a single component system, so overhangs require support structures, as mentioned below:
Our Freeform Pico range of 3D printers are a single material system which means you do need to 'break-away' supports by hand after printing. Our Composer software automatically places these support points in the exact positions in which they are required so processing is quick and easy.
It's also unfortunate that I don't as yet know of any service bureau that offers to build parts on one of these machines. Perhaps Shapeways will see the need to offer a material/process that will reproduce finer detail?
I can see that this is definitely coming, but we're not quite there yet.