HSM on STL-files

From: Jan Willem Gunnink (j.gunnink@ind.tno.nl)
Date: Tue Jul 09 2002 - 10:30:30 EEST

Dear all,

Regarding the questions Chang-Shik.Min posted to the list on STL-based
HSM please read our points of view. If you are not interest at all in
milling deleted the message but you then ignore one of the best RT
technologies at the present moment for real plastic injection moulding
and die casting. All the answers can be found more extensive in a
downloadable paper from our website:


1.Tool path accuracy problem resulting from triangulation.
When a STL-file is outputted with a tolerance of e.g. 0.001 mm then
theoratically you can mill a geometry with this accuracy. But as the
most of you know a CAM system is also using a calculation tolerance.
Normally we use 0.005 mm for that. When you have your milling (HS-)
proces totally under control you can reach this kind of accuracies but a
small change in diameter of cuttertools will effect this. Normally we
guarantee 0.01 mm on this. So the problem is not at all in the STL file
or the CAM system. That is all just maths by calculating with good
algorithmes. Output the geometry in a fine enough tolerance (0.001 mm)
and calculate with the right tolerance (0.005 mm) and have your milling
proces in control and you will be fine. As you can read in our study on
this subject it is even better to use STL-files directly to calculate
on! Allmost all CAM systems do that internally anyhow, but they don't
tell you as you then become afraid of inaccuracies and SLA like
geometries with staircases etc.

2.Speed for tool path generation and machining.
This is normally exactly the same as using "conventional IGES rubbish".
As you can read in our study or as said in point 1, most of the CAM
systems use triangulated data internally anyhow. But now there is
something very good on the market to make life easier with respect to
(HS-)milling for moulds and dies. That is PowerMILL-AutoCAM. Have a look
for more information at www.delcam.com. This is really one-push-button
milling for high quality moulds without knowledge needed as really all
technology (cuttertool databases with all technology for speeds and
feeds, depth/width of cut etc) is already in the system. Just like our
Actua-MJM machine: plug and mill! After importing your mould geometry
the system is chosing cuttertools, feeds, speed, depth and with of cut,
strategies it self on a reliable basis. It will also tell you what the
smallest detail is it can mill so that you know that you need to do some
small detailling yourself with EDM or really exotic milling strategies.
So no more looking in catalogues or calculating n=1000*V/(3.14*D),
f=fz*z*n etc.

The only thing you have to do is import your geometry, chose the
blockmaterial, chose the machine and the reference plain and push the

This kind of developments will speed up developments on RT more and
more. We are waiting for and working at the next step!

3.Problems caused by the fact that STL data can not be changed once
   in a CAD system.
The old problem; let people at once make a good mould geometry and then
there is no need for changing any file format what so every. I always
become "angry" when I hear this. The mould engineer just should make the
right design of the mould and then all is fine! However developments are
there of course to help when a STL-file is there and a change is needed.
Look at te Materialise software for Rapid Tooling and you are helped.
But nr 1 rule is: good mould designs and no excuses from the engineer
when he did it wrong!

So our conclusion: try to use STL-files as much as possible, avoid Iges
and use HSM for (Rapid) Tooling. However as a milling addict I fear a
bit of technologies like DMLS with 20um and emerging technologies like
3D-printing and LENS for more advanced tooling but in all cases for the
right accuracies and surface quality milling is needed so I have enough
to do.

When some one wants to know more or wants to comments: be free

All the best from Holland

Jan Willem

(PS: yes we were involved in the miracle as mentioned in point 2)

Ir. Jan Willem Gunnink
Senior Project Manager
TNO Industrial Technology
Dept. of Industrial Prototyping
De Rondom 1
P.O.Box 6235
5600 HE Eindhoven
The Netherlands
Work: +31-40-2650609
Fax: +31-40-2650678
Cellular: +31-6-20353750

Visit us at: http://www.ind.tno/homepage.html

For more information about the rp-ml, see http://rapid.lpt.fi/rp-ml/

This archive was generated by hypermail 2.1.4 : Tue Jan 21 2003 - 20:14:00 EET