Re: Request for some informations

From: David Retallick (
Date: Thu Apr 30 1998 - 17:02:43 EEST

Hi Volker, will try my hand at a couple of your questions.....

Volker Schillen wrote:

> 1. A rp-system build a physical object layer by layer. There are at least
> two possibilities to estimate these layers:
> a) Slicing with a layer thickness, which is constant
> b) Slicing with a layer thickness, which is variable
> Q1.1: Have the different rp-systems the ability to manufacture layer with a
> variable layer thickness?
Basically yes, but it is not that easy: You must realize the difference
between layer and slice. 2 examples:
a) Stereolithography and sinter systems operate with a constant LAYER
thickness, which also dictates a constant SLICE thickness.
b) Lamination systems, such as LOM (Helisys) or PLT (Kira) cannot
control the layer thickness, because the process is additive in Z,
minute errors accumulate through the part. In this case, the height is
measured each layer, and the slice made accordingly at this position.
This needs an ONLINE-slice algorithm.
If you need 3D/2D software with online slicing, contact me directly. You
will probably see that, due to a modular concept, we can do you a taylor
made solution!
(sorry list, that was not advertising, that was a "new product
> Q1.2: Is manufacturing with a variable layer thickness a general practice
> and if not, why!
It is not general practice to my knowledge, but many people have thought
about implementing it. lets hear what the list says ;-)

> 2. A rp-system tries to manufacture a physical object as accurately as
> possible.
> Q2.1: Which accuracy have the current rp-systems (layer thickness and point
> solution; especially the stereolithography systems)
IŽll get into trouble here, but nevertheless....
accuracy approx 0,1 mm, layer thickness 0,05 to 0,3 mm.
The first problem is the the scanner resolution, as already explained,
and the second is the recoating, producing thin even layers. If you run
with thin layers and get curl, you remove the part with the wiper. i.e.
purely mechanical/hydrostatic problems.
> Q2.2: What are the main problems which are responsible for inaccuracy
> (especially if you manufacture very small parts or parts with filigree
> elements).
Most systems use scanners ie X/Y units with galvos. These are analog
units and
have a range of about 10-14 degrees. If you think about it, it is clear
that we are pushing the technology way over the reasonable expectations.
In other words, big machine, long focal length, poor focus. You can
counteract this with a larger aperture, but lose speed due to the
heavier mirrors.
Congrats, your system doesnŽt have these problems.
Small machine, short focal length, super focus, super accuracy.
This is the reason for NOT buying a big machine and making lots of small
parts on it.
> Q2.3: Are software systems able to estimate the exact volume error between
> the original and the sliced object?
Sure, if the shrinkage factors are known.
> Q2.4: Is the volume error important for development/manufacturing of products?
RP tends to use linear measurements, like on the drawings.
> Q2.5: Are support structures necessary if you want to manufacture small
> parts (especially if you use stereolithography systems)?
Always, yes, yes, yes.
> Q3: What are you missing most at your software system?
Nothing, I developed and configured it just as I wanted it.
> Q4: Is there anybody who manufacture microparts with stereolithography?
Q4 back: what do you mean by micro? define your terms, or even better,
mail me directly.

Sorry folks, could have sent it directly, but am interested in the
opinions of the list.


Dr. David Retallick
Retallick Systems, Baumgaertle 36, 
D-86505 Muensterhausen (Germany)
phone +49/8281-1763 - fax +49/8281-5556

For more information about the rp-ml, see

This archive was generated by hypermail 2.1.2 : Tue Jun 05 2001 - 22:45:27 EEST