your opinions (right or wrong)

From: Elaine Hunt (Clemson University)
Date: Wednesday, July 5, 1995

From: Elaine Hunt (Clemson  University)
To: RP-ML
Date: Wednesday, July 5, 1995
Subject: your opinions (right or wrong)
RP'ers

Here's your responses.  Agree or disagree......IS this how we picture the
world of rapid prototyping?    Are we talking the same talk when we do ever
TALK.  Do we even silently agree or disagree even when we just
"""listen?"""
Many thanks to those who do speak and express their opinions.

Accuracy: measure of how close the final RP model is to the original CAD
model: The precision of dermining locations relative to one reference
location: An estimate of the one-time precision to which a process is
capable of reproducing an original model: The dimensional tolerance
acquirable by the RP device.  How close the part comes to what is shown on
the drawing or in the CAD model.    Dimensional fidelity.
     1) PART ACCURACY is the maximum difference between measured dimensions
     and nominal dimensions specified per part drawing or CAD model.
     2) SYSTEM ACCURACY is maximum error between commanded position and
     actual position, sampled over the entire work envelope of the system.
     A reference coordinate system must be specified from which all
     measurements are taken.

Repeatibility: measure of how often the same accuracy can be obtained: The
precision of going back to the same location in space. The location may not
be accurate.: An estimate of the precision to which a process is capable of
reproducing an original model on sequential runs.: The ability to build
identical parts at the same tolerances. Whether the machine produces the
same output given the same input consistantly over time. Ability of the
system to return to a given location in space.  This usually expressed
statistically to describe the sampled population of deviations. Eg. Avg.
Error +/- 6 sigma (or whatever sigma level you  feel comfortable with for
your application).  Consistancy.

Warpage: distortion of any kind: Distortion of a two dimensional surface.:
An estimate of the out-of-plane deviation from an original model which is
caused by end-use: "twist" that occurs due to shrinkage. The distortion of
large areas in plane.  Usually seen in the middle of large areas.  Large is
a relative term.  The measured distortion of a planar surface.  This is
similar to the  geometric tolerance "flatness" feature.  The warpage may be
expressed   as the distance between two theoretical planes, parallel to the
specified plane, which envelope the actual surface at its' upper and
lower extremities.  Unexpected bending.

Curl: warpage that follows a circular pattern : A distortion that is
intrinsically one-dimensional and characterized by progressively increasing
deflection as a function of distance along the axis.: An estimate of the
out-of-plane deviation from an original model which is strictly process
induced by shrinkage. : bending in the x,y, or z direction also due to
shrinkage. The lifting of an edge from the horizontal plane usually caused
by shrinkage of successive layers.    Edges of upward or downward facing
surfaces which warp due to internal stresses or insufficient support
structures.  Non-flat flat areas.

Distortion: warpage of curl: Out of dimension-ness.: An estimate of the
out-of-plane deviation from an original model from all process induced
causes.: inaccurate tolerances in a part (due to process parameters?). Any
other mangling of the part not covered above.  Often due to uneven
postcuring.  Includes twisting, bubbling, etc.   Systematic multi-axis
deviations in part accuracy.  Geometric deviations.

Rapid Prototpying: physical models without prints : This is a toughy! Rapid
- by whose determination? Prototypes - models made from end-product
material using processes that yield end-product mechanical properties. RP -
a pretty fast way to make a model that is cann't be used for functional
tests, in most cases. Personnaly, I like Rapid Product Realization for the
field and RP machines for the material additive fabricators. RP doesn't
have to stand for anything (e.g., NCR, IBM, 3M).: The additive manufacture
of product or tools: layer-by-layer production of parts from CAD. The art
of producing a model of a design in an expedient manner. AKA holoforming.
Holoforming, making solid objects directly from CAD data with little
mechanical setup or manufacturing planning.  Essentially a "fire and
forget" 3D plotter or imager.  The RP parts are generally not produced  in
materials which can be directly used for product applications.   Pronto
model making.

RMS errors: Something went wrong and I didn't do it: Repeat the same
process N times, measure the errors, and use the standard statistics
definition to determine rms error: A statistical term to describe accuracy
and repeatability. It is the average of the absolute error measurements.:
bumps or crevices in surface finish (due to stair-stepping, etc. ). Root
Mean Square errors:  This really mean you square the errors, average them,
and take the square root.  It is a technique for describing average errors
that is not dependant on sign (i.e. the error can be either + or -).  It is
best used for relative comparison  amongst parts.  A statistical method
used by physicists to make errors appear smaller than they are.  (See "How
to Lie With Statistics" by Darrell Huff) RMSerrors are typically less than
maximum errors, and frequently are less  than actual errors measured on
real parts.   "Average" errors.

Elaine Persall: A conscientious, hard working, humorous human being known
to make a point by driving a large nail with a pneumatic hammer. Your
standard net.troublemaker who is always trying to expand the state of the
art often by unconvential means.  Also, a sweet southern belle.  And I'm
not just saying that because she can get her hands  on me.

Other comments:
>>>>>>Concerning the definitions of CURL and WARPAGE I see a problem to
seperate those two things, both observations seems to be coupled.  One way
to seperate CURL from WARPAGE would be to define  WARPAGE as deformation of
planes which are supported during build and  CURL as deformation of
unsupported parts,  however those definitions are not very practical, since
most parts are  build with a support-structure when it is
nescessary...Therefore another suggestion to define CURL and WARPAGE: CURL
refers to the deformation of the first few bottom (downfacing) layers,
whereas WARPAGE refers to the deformation of the top layer of a plane or
feature. This definiton would also define the measuring of those build
defects, since CURL can be easy measured as the flatness of the bottom
layer and  WARPAGE accordingly as the flatness of the top layer. Currently
some  people measure CURL on the bottom layer, others measure the top
layer.  CURL usually refers to the build defect caused by shrinkage of the
solidifying  material in succesive layers. However, the term WARPAGE is not
really coupled with a certain build defect and cause, (please correct me if
I am wrong in that), so this term is felt unnecessary and can be replaced
by FLATNESS of a plane, which is already  defined for conventional
manufacturing processes. who uses other definitions of curl and warpage or
has suggestions how those terms could be defined precisley and applicable
to measure those build defects ?????>>>>>>

Elaine
What's your opinions......they are important even if they are wrong.


Previous message | Next message
Back to 1995 index