Re: Geometrical errors of SLS parts

Date: Fri Jun 11 1999 - 19:10:05 EEST

> Dear RPML;
> I am a research student working on analysis of geometrical errors in SLS
> part. I have to develop a mathematical model to describe such errors. As
> a result, I would like to know and have information concerning about any
> errors such as part disorientation, layer dislocation, layer displacement,
> and so on... , which might be occurred during either fabrication process
> or data preparation process.

2 Words -- Good Luck.

There are so many variables to consider in trying to perform this
task. Here are a few geometrical issues, the cause, and such. . .

* Sink - This one is pretty easy. If you have a "T" shaped part, one
wall will cause a shrinkage sink in the bisecting wall. Easy to
understand, easy to predict, not so easy to solve.

* Curl - Again pretty easy. Downward facing surfaces are affected,
especially if they are "flat." This can be corrected in the process for
the most part by running part at the "right temperatures."

* Duraform Skin Disease - Real easy. Recycling of duraform
without addition of any "virgin" powder can cause a horrible acne-
like pitting on vertical surfaces. This is corrected by proper powder
mixing. No mathematical model can predict the shape of the acne,
'cause it's just too ugly.

*Warpage - thin wall parts - Duraform - If you build a thin wall part,
very flat with no ribs, the part is usually flat. But, if you add ribs
around the outside or on the edges, the part will tend to "potato
chip." (Hey, notice I spelled potato without the "e") The added
volume on the edges causes a variable shink in the part (the walls
shrink more). Thus, if a part is 1 inch, it becomes .99 on the edges
and 1 in the middle. (not real numbers, I just made 'em up for the
illustration). Anyway, it ain't pretty.

*Warpage - large "square" parts - A rectagular box can tend to look
more like a smiley face if built horizontal. Dimensions may measure
out, but there is an arc.

* Neighborhood shrink - If a part is built next to another, they will
affect each others shink value. For instance, you build a "donut"
around a cylindrical shaft, you shaft will "offset" inside the donut.
Kinda like if the moon got too close to the earth, there would be a
big crash. Hard to predict, and even harder to model.

These are just a few observations from a mind that is slightly less
than sound. One thing that is cool, pull out a thin walled part when
its too hot, and you can watch it warp and straighten back out as it
cools. The outsides reach room temperature first and shrink at a
faster rate than the inside. You think the part's gonna crap out, but
as the inside cools, it straightens back out. This is recommended
for experienced personnel only and should not be tried at home.

David K. Leigh
Harvest Technologies (254) 933-1000
Rapid Prototyping Services
1000 Industrial Park Road Belton, TX 76513

For more information about the rp-ml, see

This archive was generated by hypermail 2.1.2 : Tue Jun 05 2001 - 22:51:52 EEST