Re: Re[2]: Stratasys wish list, again

From: Atif Mehmet Yardimci (
Date: Tue Sep 16 1997 - 23:39:11 EEST

On Tue, 16 Sep 1997 ralott@CCGATE.HAC.COM wrote:
> .....................
> 4. The FDM hasn't piqued the interest of the academics the way SLA's have.
> I guess without lasers and exotic chemicals, the FDM's are just too
> mundane.

Just a few observations on FDM:

  a. The envelope conditions are far from mundane, Altough envelope
     temperature is more or less homogeneous throughout the volume;
     air velocities are not, due to complicated flow patterns generated
     by two planar jets located at the sides of the hardware, and
     their interactions with machine components (foam platform, elevator
     guides, ...). More on the structure of velocity distributions, if
     there will be interest.

     To test the effects of the air velocity distributions,
     construct a 0.5"diameter x 2" long cylinder *without* containment
     supports at NW/W/SW/N/O/S/NE/E/SE corners of the foam.

     The cylinders on the east/west side of the foam would have more
     shrinkage on the west/east side. Why? differential cooling rates
     imposed with stagnation flow/wake structure would force the
     wake side to remain at elevated temperatures longer.

     Cylinders along N/O/S axis would have reproducible kinks located
     around half an inch from the foam. The air velocities at the foam
     surface near the center are significantly lower for the first
     half an inch of the build, due to the blocking of side jets
     by the foam platform. The sudden change in air velocities-> cooling
     rate may be the cause of the problem.

     (N:north(back), S:south(front), E:east(right), W:west(west),

  b. Imagine you have a 6"x1"x0.26" rectangular prism to be built.
     Let's assume that 6" direction is
     aligned with x axis, and cooling is uniform throughout the
     workvolume. If one would embed thermocouples to the foam or into
     the part, the measurements will indicate that the E side
     of the part will be hotter relative to the W side. The presence
     of the support liquefier and radiative heat transfer from it
     does pre/post-heat the part and surroundings during the build.
     Since the thermal footprint of the traveling extrusion head
     has a non-spherical and rather large spatial extent, thermal
     analysis of the problem becomes significantly involved.

  c. Why do road cross-sections look like a rectangle + two semi-circles?
     Can the answer lie in injection molding? in injection molding
     polymeric melts/liquids injected in-between parallel walls produce an
     elegant flow structure at the flow front, called 'fountain
     flow'. previous experience in this field points that the flow front
     will be near to a semi-circle for most instances. Do the
     thermoplastic melts experience a combination
     of 'fountain-flow'+couette flow between liquefier tip at the top
     and previously deposited material at the bottom?

  d. If one wishes to do adaptive slicing, hence generate a family
     of roads of different thicknesses stacked onto eachother at
     arbitrary angles; how does the microstructure look like? Some
     of the experiments show increased sensitivity of
     thinner roads to substrate topography. Depending on the ratio
     of the slice thicknesses and local stack angle, roads may decide to
     follow the valleys and hills rather than forming a planar surface,
     resembling rivers that meander through the topography at
     well defined wavelengths if viewed from the top.
  e. The build file of FDM is in ASCII, hence it comes with a built-in
     API. i believe it is unique in this aspect among other RP vendors
     and the guys in Stratasys should be encouraged to keep it in an
     open platform. what else can a graduate student ask for?


     Atif Yardimci

Manufacturing Processes Laboratory +1-312-355-0478 / +1-312-413-7408 [voice]
Mechanical Engineering Department +1-312-413-0447 [fax]
University of Illinois at Chicago e-mail:

This archive was generated by hypermail 2.1.2 : Tue Jun 05 2001 - 22:40:25 EEST