Home » Community » Work in progress »

# Pyramid points

Pyramid points [message #53171] Sun, 26 August 2012 19:58 UTC
I'm attempting to produce a 3d tree fractile (l-system, if you know the terminology), and am running into an unexpected limitation. The ends of my fractile come to points, and apparently those points are causing automated auditing software to fail in a way that doesn't make sense.

The audit is telling me that it can't be printed because the ends of the tree structure come to points that are smaller than the minimum wall thickness of .7 mm. My reaction to that is, well, ya, they're points. Is it impossible for the 3d printer to produce anything with pointy bits?

Their concern is that, if you make walls to thin, parts might break off and damage other models in the tray. I'm good with that. I wouldn't want to disrupt their industrial process just because I'm looking for a cool shape. However, I'm having a hard time imagining the tips of a pyramid (even a pointy pyramid) being structurally unsound, and can't imagine that, even if the tip did break off, that the grain-sized pieces that would result from it could damage anything else.

Actually, to be more specific, if this shape is possible, I'm uncertain why my shape is restricted.

What am I missing here?

(Attached, for reference, a fifth order fractal of the form I'm trying to produce. The software has the ability to vary all height, width, order, and truncation parameters)

Re: Pyramid points [message #53177 is a reply to message #53171 ] Sun, 26 August 2012 21:41 UTC
What is the desired overall height of this model?
What Material are you trying to print it in?

The STL you shared above has a height of 2149 units. That's too big for any of the printers (or would be extremely expensive). I divided it by 25.4 to convert mm to inches - That gave me a reasonable thickness on the final branches.

Something as "spikey" as this IS likely to have difficulties during printing. The analysis that you are using is partly correcct.. you can indeed print things that come to an "edge", but.. it works a lot better if it's an EDGE, not a SPIKE.

What is very difficult about this particular model is that you've got points sticking both up and down. It's the downward points that tend to be very troublesome for the printers. If you think about it, to print a spike that is pointed upward, you can attach the thinner/smaller layers to the ones below, and they're supported by the material underneath. But... when trying to print a downward-pointing spike, you are trying first to print a single "dot" of material, followed by a layer above it that is only slightly larger, etc. Even a half or third of the way up along the spike, the thing is stil just a sliver of material, and has no mass side-to-side to support it. It's too easy for that small peice to be caught by the printhead and shifted out of position.

If your model could be oriented with the spikes in the horizontal plane, then they might not have as much trouble, but with spikes going "all" directions, it becomes more of a problem.

Patience, Persistance, Politeness - the 3Ps will help us get us to Perfect Printed Products
Re: Pyramid points [message #53179 is a reply to message #53177 ] Mon, 27 August 2012 00:33 UTC
Thanks for the reply, Stony. That isn't the specific model I was uploading, it's just an example. The program I wrote takes parameters to build the model. I'm familiar with the raw formatting limitations of the printer (scale, resolution).

The downward spike problem makes sense. I could probably generate the shape with better truncation on the downward pointing spikes.

This still leaves the question of how sharp I can make the spikes, which apparently is an independent number for upward and downward spikes, and maybe another for side-to-side spikes. My original shape was designed to determine those limitations (I fully expected it to come out looking somewhat like cauliflower), but they refused to print it for fear of damaging other models.

When i asked customer service, the only answer they could give me was that the minimum wall thickness is .7 mm. This makes for some seriously blunt spikes, though. Do you have a better estimate?
Re: Pyramid points [message #53180 is a reply to message #53179 ] Mon, 27 August 2012 03:39 UTC
Note: The explanation below is not using "official" measurements.. Someone from the Shapeways Production Team will have to supply the correct definition. I'm just trying to help explain the topic.

A good bit of this discussion depends upon what you call a "point". Both shapes below come to a sharp point. It's very likely that the green shape would pass the QC checks, but the pink shape might not. The difference between the two is the distance from the end of the point back to where the supporting body is at least as thick as 0.7mm.

You should be able to get away with a point that is a bit deeper than 0.35mm, but I don't know what the maximum depth can be. My suspicion is that you can't have it deeper than 0.7mm (or, an angle between sides of about 54 degrees)

• Attachment: Points.jpg