Parasolid transmit file import into SW2001+
Jeremy Winston
I continue in my quest to get contours from
a series of CT slices (representing the surface of
a bone) into SolidWorks, with the intention of then
lofting the contours into a solid representing the
bone.
The first step in this process involves extracting
the contours from the CT slice files. This part I have
solved (using the Visualization Toolkit (AKA VTK, free
at http://www.kitware.com/vtk) and Tcl/Tk.
The next step is getting the contours (consisting
of an ASCII flat file (or files) containing a series of
(x,y,z) triples, one series for each contour) into
SW EE 2001+. My recent efforts, which I am about to
abandon for lack of progress, were aimed at using the
Parasolid API (v14.0.110) in some C++ code to turn the
contour data points into an assembly of Parasolid wire
bodies, and then import the Parasolid .X_T file into SW.
At first this step was failing because I saved the
.X_T file as v14 data, but SW2001+ can only "understand"
up to v13.0. Then, with some help from c.c.s readers
(thanks, Richard of EDS), I rewrote my SaveModel() routine
to save in v13.0 format (using the PK_PART_transmit_o_t
option structure). Unfortunately, now the import into SW
fails with this error message:
N part(s) within this assembly cannot be imported.
The component will not be added.
where 'N' is however many wire bodies are in the assembly
stored in the .X_T file. Now, I know that the .X_T file
is not corrupted, since PS/Workshop can read & display it
fine. So my (many) questions are:
1. Any ideas why I'm getting this error message?
2. Is there any other way to get the contour data into SW,
or must I resign myself to manually importing ~100
contours per bone using the "Curve Through Free Points"
tool for each of ~100 separate contour files?
3. When I use "Curve Through Free Points" to import some
contours, and then I loft through those contours, I get
a nasty "twisting" effect which I assumed was because the
number of points in each contour profile were not the same.
So I resampled the data so each contour would have the same
number of points, but I *still* get the "twisting" effect.
Any suggestions as to how to get rid of this?
4. Can someone with a later version of SW than 2001+ try
to load the .X_T file at
http://www.cat.rpi.edu/~winston/SW/LTibia1st7.X_T
(which contains the first 8 contour slices through the
left proximal tibia from a CT series, saved in Parasolid
v13.0 format)?
TIA,
-Jeremy(to reply directly, change dog-->cat & rip-->rpi)
Jeremy Winston
>
> SolidFolk,
> I continue in my quest to get contours from
> a series of CT slices (representing the surface of
> a bone) into SolidWorks, with the intention of then
> lofting the contours into a solid representing the
SW Folks,
It just occurred to me I've been going about this
bass ackwards.
My goal is a Parasolid model of a bone. I was intending
to use SW to edit the model before re-exporting back to
Parasolid format.
It makes a lot more sense to take the ASCII point data
into SW using the SW API (or a macro?) than to create
a PS model using the PS API, import into SW, edit, then
export back to PS.
Could some SW guru either create a macro that uses the
"Curve Through Free Points" tool to create a series of
curves from files named "LTibia.%d.txt" (where %d ranges
from 2333 to 2433), or if that's too much work, a macro
that does so for just a single file? Or point me to
sample code or manual pages?
TIA,
-Jeremy
Mark Mossberg
The curves import into SW2003 SP0 with no errors
Regards
Mark
"Jeremy Winston" <win...@dog.rip.edu> wrote in message
news:3DC1DA0F...@dog.rip.edu...
Mark Mossberg
I tried lofting your contours in two completely different Parasolid based
systems, Solidworks, and Mastercam 9. In both cases I was only able to loft
a couple of sections before getting a failure, (ambiguous dialog in SW2003,
and Parasolid kernal error 999- in MC).
I think Parasolid has problems creating valid topology out of these very
complex curves. You best bet is to simplify some of the curves, or use a
polymesh based system.
Regards
Mark
"Jeremy Winston" <win...@dog.rip.edu> wrote in message
news:3DC1DA0F...@dog.rip.edu...
Jeremy Winston
>
> The curves import into SW2003 SP0 with no errors
Hi Mark,
Thanks for checking. That's useful info.
I guess I'll have to abandon that approach and
take the ASCII coordinates directly in to SW, since
my university isn't going to upgrade to SW2003 until
late Spring of 2003.
Here's a follow-up question:
If I create a "Curve Through Free Points", is there
any way to move individual points of the curve graphically,
or must I edit the Excel-style spreadsheet?
TIA,
-Jeremy
Jerry Steiger
> 3. When I use "Curve Through Free Points" to import some
> contours, and then I loft through those contours, I get
> a nasty "twisting" effect which I assumed was because the
> number of points in each contour profile were not the same.
> So I resampled the data so each contour would have the same
> number of points, but I *still* get the "twisting" effect.
> Any suggestions as to how to get rid of this?
Imagine a very simple loft, with two identical squares as profiles. SW lofts
surfaces between the edges of your profile. When you pick your profiles from
the graphics area, it starts from the closest vertex to where you selected
the first profile and lofts from there to the closest vertex to where you
selected the second profile. If you are looking almost end on to the two
squares and pick the upper right vertices on the two squares, you get a
simple rectangular part. If you pick the upper right vertex on one and the
upper left vertex on the other, you get a twisted part, like a nice piece of
wrought iron work. If you pick two diagonal corners, it won't loft at all.
If you pick the sketches from the Feature Manager, then SW picks which
vertices to start from without your help and may not do a very good job of
it. With complex countours like your bone profiles, I think SW is going to
have a lot of trouble making good lofts.
Jerry Steiger
At Work Computers
Jerry Steiger
news:2rzw9.213$mI2.26...@newssvr21.news.prodigy.com...
> Jeremy,
>
> I tried lofting your contours in two completely different Parasolid based
> systems, Solidworks, and Mastercam 9. In both cases I was only able to
loft
> a couple of sections before getting a failure, (ambiguous dialog in
SW2003,
> and Parasolid kernal error 999- in MC).
>
> I think Parasolid has problems creating valid topology out of these very
> complex curves. You best bet is to simplify some of the curves, or use a
> polymesh based system.
Jeremy,
I think Mark is right on the money here. Trying MC was a good idea, as CAM
software is often better at this type of work than CAD.
Looking at the bone profiles that you generated, SW would have to loft from
a single cross section to three separate cross sections. I don't think it
can do that on its own. (Ed Eaton, Paul Salvador, Mark Biasotti, or some
other surface wizard can tell you for sure.) I think you would have to break
it into a series of lofts, probably five, generating some fictitious split
lines as you transition from one to two to three cross sections. It would be
extremely tedious to get a good looking final shape, with a lot of trial and
error work on tangencies and where you split the sections and such.
Geomagic or some of the other reverse engineering programs may be a much
better fit than SW.
Shandon
I tried it in SolidWorks 2003 pr2 I got it too work (after removing
the txt extension from the file name). SolidWorks created an assembly
file with one part and that part had 12 curves in it. I don't know
anything about the details of the format standard, but maybe you can
change to x_t file so SW2001+ thinks it is a part file and not an
assembly file? Hope this helps.
Shandon
Paul Salvador
I got it to import into SW2003 as the others but I went a bit further by
copying the curves to 3D curves and simplifying them. Also I did it
with sections or at .04" steps and copied the curves again and
simplified them.
It does not look pretty but it is but who knows, it may look close to
your scanned image or part which exception to the peaks.
Here's the file and images...
http://zxys.com/swparts/tibia-Part-1.zip
http://zxys.com/swparts/tibia_scan_2003_fillsurface1.png
http://zxys.com/swparts/tibia_scan_2003_fillsurface2.png
http://zxys.com/swparts/tibia_scan_2003_fillsurface3.png
http://zxys.com/swparts/tibia_scan_2003_fillsurface4.png
..
Paul Salvador
trimmed, looser tolerance otherwise it gets spikey.
http://zxys.com/swparts/tibia_scan_2003_fillsurface5.png
..
Jeremy Winston
>
> Paul Salvador wrote:
> >
> > I got it to import into SW2003 as the others but I went a bit further
> > by copying the curves to 3D curves and simplifying them. Also I did
> > it with sections or at .04" steps and copied the curves again and
> > simplified them.
> > It does not look pretty but it is but who knows, it may look close to
> > your scanned image or part which exception to the peaks.
> > Here's the file and images...
> > http://zxys.com/swparts/tibia-Part-1.zip
Yoiks! Thanks for all that effort.
Could you explain in more detail the steps involved?
I have only been using SW for a couple of weeks, and
I'm not sure how to do the things you describe.
> And for reference, a iges curve export/import into Rhino3D,
> patched and trimmed, looser tolerance otherwise it gets spikey.
> http://zxys.com/swparts/tibia_scan_2003_fillsurface5.png
Wow, Rhino3D's results are much better. Any idea what algorithms
it uses? (I assume it's way too expensive for an impecunious
student researcher to buy.)
Does Rhino3D support exporting to Parasolid format, by chance?
Thanks again,
-Jeremy
Cliff Huprich
pixel the pixel is considered to be "on" (or to exist in the output
data set). Hence the tolerance on the data is almost +/- one 3D pixel
(in ALL 3 dimensions).
This will rarely make for good surface fitting.
Surfaces must also pass *exactly* through the data point - which,
among other things, results in many control points (usually far too many)
and, possibly, a high degree surface (or many patches, choose one).
Ae STL type system is better suited to this need IMHO.
The other option is special surface fitting software (available sometimes
on some systems) or some data smoothing & reduction. An API program
can do this if you know the math (and the design intent) well.
The vendors used to ask > US$ 100,000 for such custom programs.
--
Cliff Huprich
Jeremy Winston
> [...]
> Ae STL type system is better suited to this need IMHO.
Cliff,
My goal is to go from CT slices (in GE Signa format) to
a Parasolid model of a given bone contained therein. I am
open to any suggestions. What are some STL-based systems
I might investigate?
> The other option is special surface fitting software
> (available sometimes on some systems) or some data smoothing
> & reduction. An API program can do this if you know the math
> (and the design intent) well.
I may try this, but since my current need is for just a single
instance of a knee joint, I will probably just labor on this
one in native SW.
> The vendors used to ask > US$ 100,000 for such custom programs.
Cool! So that means if I am successful in writing the API, I can
also be rich!
;-)
-Jeremy
Jeremy Winston
>
> Paul Salvador wrote:
> > [...]
> > And for reference, a iges curve export/import into Rhino3D,
> > patched and trimmed, looser tolerance otherwise it gets spikey.
> > http://zxys.com/swparts/tibia_scan_2003_fillsurface5.png
>
> Wow, Rhino3D's results are much better. Any idea what algorithms
> it uses? (I assume it's way too expensive for an impecunious
> student researcher to buy.)
>
> Does Rhino3D support exporting to Parasolid format, by chance?
Glad tidings of great joy!
1. Rhino3D for students is a mere $195, and
B. Rhino3D supports Parasolid (.X_T)[*]
http://www.rhino3d.com/features.htm
Now, does Rhino3D have a tool similar to SW's "Curve Through Free
Points"?
-Jeremy
[*] By "support" do they mean both import *and* export, or
just import?
Jeremy Winston
> [...]
> And for reference, a iges curve export/import into Rhino3D,
According to the Rhino3D web site, Rhino has native support
for SW. See "File formats supported:" at
http://www.rhino3d.com/features.htm#details
How come you had to export as an IGES curve?
TIA,
-Jeremy
Cliff Huprich
<win...@dog.rip.edu> writes:
>Yoiks!
I missed the start of this thread.
What's the end use of the data?
Some sort of actual physical parts or just 2D drawings?
--
Cliff Huprich
Cliff Huprich
<win...@dog.rip.edu> writes:
>What are some STL-based systems I might investigate?
Might take a look at http://www.stlwork.com/ and the list at
http://www.stlwork.com/3dcad.html - I know cost is important to you.
Some of the limited "free demo" versions might even do what you need
for a one-time use.
Fred Smith (ims...@imsrv.com) could probably advise you.
Look at your final end-use first.
HTH
--
Cliff Huprich
Paul Salvador
Ok, looks like you found some of the other questions you were looking
for.
The answer to this question is: Rhino does not have native support for
SW.
You can not read in a SW file into Rhino.
But you can use a plugin which allows SW to read in native Rhino 3DM
files but not write them out.
So, reading in SW files is usually STEP or IGES.
And for reading in wireframe or curves, IGES is the better option.
Kind of silly hah?
For your use, Rhino is probably perfect?
With your need for x_t export and it's support for data import and
exports or point cloud support and curve support through points, and the
strong support for scripting, Rhino is a very good tool for reverse
engineering.
Also, look at their plugins page for other specialized resources..
http://www2.rhino3d.com/resources/
..
Iain McMillan
> My goal is to go from CT slices (in GE Signa format) to
>a Parasolid model of a given bone contained therein. I am
>open to any suggestions. What are some STL-based systems
>I might investigate?
>
>-Jeremy
Jeremy
I worked on a very similar problem a couple of years back, going from CT
scanners to RP (rapid prototype) and CAD compatible formats.
I spent quite a while assessing several specialist medical reconstruction
software packages. Two of the better programs I investigated were
Mimics http://www.materialise.com/mt.asp?mp=mm_main
Velocity2Pro http://www.image3.com/
Depending on the type of reconstructions you want to do, if is it
bone/cartilage/hard_tissue then Mimics is the better package IMO. However if
you are interested in neurons/blood vessels/soft_tissues Velocity would
probably be the better choice. My work was focussed on orthopaedic patients
hence Mimics would be my choice.
If you are considering evaluating or purchasing these packages give me an
email and I will suggest some things to be aware of and to test for.
Both of these packages will export STL files (surface - facet data), this is
great for RP however not many CAD packages will import this format. Rhino
will open STL files (I believe) but if it is even a medium sized scan it tends
to choke.
Both packages also suggest they are able to export CAD compatible files, by
that they generally mean IGES curve data. Lofting these profiles together
sounds fine in theory but I found that often the curves to be made up of
multiple segments which are not necessary aligned hence you get twists in you
loft. Also then you come to areas like the tibial plateau or femoral head,
where there is saddle geometry you run into problems getting an accurate loft.
Note my interest was in setting up a process that was very accurate and
repeatable (we reconstruct 2-5 patient data sets a week from skulls to ankles
and anywhere in between). Hence mucking around with guide curves, splitting
and joining curves and letting the software guess the geometry where not
options.
The final step in our process to complete the CT or MRI to CAD path will
hopefully one day be
Geomagic http://www.geomagic.com/products/studio/
I have trialled this software (and others like it), and found it to be
accurate and easy to use, the support from the VARs was great; in short it is
fantastic, shame about the price.
In summary it really depends on what you want. If you are after an
good approximation run with what Paul has explained, as always his work is
incredible (shame it is in SW2003 though, I too am stuck with SW2001+ until
the educational release of 2003 is made available). However if you need an
accurate reconstruction then you may need the right tool for the job. As
Cliff has pointed out consider your end application that will determine your
final solution.
Finally if you only have a few data sets perhaps you can FTP or mail me the
raw data and I will do a reconstruction to STL for you using our software. If
you want a Parasolid from that you could try the tool Cliff suggested or try a
service bureau (for a one off it can be economically viable).
Good luck with your work
Regards
Iain
Iain McMillan
PhD Candidate
Mechanical Engineering
University of Canterbury
Christchurch
New Zealand
Jeremy Winston
> Jeremy Winston wrote:
> > My goal is to go from CT slices (in GE Signa format) to
> > a Parasolid model of a given bone contained therein. I am
> > open to any suggestions.
> I spent quite a while assessing several specialist medical reconstruc-
> tion software packages. Two of the better programs I investigated were
>
> Mimics http://www.materialise.com/mt.asp?mp=mm_main
> Velocity2Pro http://www.image3.com/
Iain,
Thanks for the pointers, but unfortunately, my budget is
whatever spare change I can find lying around the house. (Well,
not quite that bad, but I *do* have to pay for all s/w myself.)
> Depending on the type of reconstructions you want to do, if is it
> bone/cartilage/hard_tissue then Mimics is the better package IMO.
Initally, just bone & cartilage.
> My work was focussed on orthopaedic patients hence Mimics would be my choice.
The intended use of the Parasolid model is as input to a meshing
program in preparation for FEA. The meshing program accepts only
Parasolid and Shapes formats. (It is part of an advanced suite of
tools that allow FEA of biphasic materials.)
> If you are considering evaluating or purchasing these packages
> give me an email and I will suggest some things to be aware of
> and to test for.
I'd love to try Mimics, but funds do not allow it. *sigh*
> [...]
> Both packages also suggest they are able to export CAD compatible files, by
> that they generally mean IGES curve data. Lofting these profiles together
> sounds fine in theory but I found that often the curves to be made up of
> multiple segments which are not necessary aligned hence you get twists in you
> loft. Also then you come to areas like the tibial plateau or femoral head,
> where there is saddle geometry you run into problems getting an accurate loft.
This seems to be the tricky part, regardless of which software is used.
SW2001+ lofts of the contours result in horrible "twisting," even with
guide curves. It looks from Paul Salvador's work like Rhino3D does a
much better job of lofting. Rhino3D is free for a trial period, so I
plan to try that.
> Note my interest was in setting up a process that was very accurate and
> repeatable (we reconstruct 2-5 patient data sets a week from skulls to ankles
> and anywhere in between). Hence mucking around with guide curves, splitting
> and joining curves and letting the software guess the geometry where not
> options.
My initial purpose is a one-off, so I can afford to spend a lot of time
tweaking it.
> The final step in our process to complete the CT or MRI to CAD path will
> hopefully one day be
>
> Geomagic http://www.geomagic.com/products/studio/
>
> I have trialled this software (and others like it), and found it to be
> accurate and easy to use, the support from the VARs was great; in short it is
> fantastic, shame about the price.
There's that ugly word again. ;-)
Well, I guess Geomagicians have to eat, too.
> Finally if you only have a few data sets perhaps you can FTP or mail me the
> raw data and I will do a reconstruction to STL for you using our software. If
> you want a Parasolid from that you could try the tool Cliff suggested or try a
> service bureau (for a one off it can be economically viable).
Thanks for the kind offer. I expect it will take me many more weeks
to get something useful; I may take you up on the offer then.
I've been wondering about something, and given your experience,
perhaps you could hazard an answer: Shouldn't it be possible to
use Lorensen and Cline's Marching Cubes algorithm to generate a
polygonal surface mesh of a given bone, and then import that mesh
into a solid modeling package? I mean, shouldn't it be easy to
turn such a mesh into a solid?
I haven't seen any way to get such a mesh into SW, and I haven't
read about any other packages that will allow it, either.
-Jeremy Winston
Cliff Huprich
<win...@dog.rip.edu> writes:
>I expect it will take me many more weeks
>to get something useful;
Had a human mandible once in CAT data for a University dental
college.
You have a *simple* part.
Simplify it. It's for a FEM application ... pick a *few* well-fitted points
per section (like 5 maximum) & fit closed splines. Skip most of the
sections. Assure the starting points on the splines line up (no twist)
& surface.
Remember, your data had a +/- one solid pixel noise in the first place ..
and you have a FEM application, not an exact machining issue.
Feel happy it's not an abscessed mandible <G> It's the most complex
bone in the body.
HTH
--
Cliff Huprich
Iain McMillan
Jeremy Winston wrote:
> Thanks for the pointers, but unfortunately, my budget is
>whatever spare change I can find lying around the house. (Well,
>not quite that bad, but I *do* have to pay for all s/w myself.)
I know the feeling my research grant runs out this month, I have
been fortunate to this point to have really good funding, but things are about
to change... Anyone out there want to sponser some research
>The intended use of the Parasolid model is as input to a meshing
>program in preparation for FEA. The meshing program accepts only
>Parasolid and Shapes formats. (It is part of an advanced suite of
>tools that allow FEA of biphasic materials.)
I would tend to agree with Cliff's latest post you aren't going to want a
really detailed reconstruction just an approximation. Hence the fitting of
splines and tweeking of surfaces and time will probably meet your needs.
>This seems to be the tricky part, regardless of which software is used.
>SW2001+ lofts of the contours result in horrible "twisting," even with
>guide curves. It looks from Paul Salvador's work like Rhino3D does a
>much better job of lofting. Rhino3D is free for a trial period, so I
>plan to try that.
This twisting could well be because your profiles have end points that are not
aligned hence when lofting SW will try to connect these together. Guide
curves help a bit but I found that all they did often was to compress the
twists into one area. By creating new curves (splines) in a new sketch
using your imported data as a reference you can help eliminate this problem as
you set the end point, and can control the number of points on each spline
(ideally each spline should have the same number of points). At least this is
my interpretation of the problem, someone else will probably be able to
elaborate on this better than me (or correct my thoughts on this matter).
If you haven't done a lot of lofting/surfacing before in SW, an amazing
resource is Ed Eaton's SolidWorks World Presentation "Curvy Stuff 101". You
can find it at
www.prodesignsrv.com
under the tutorials section.
>Thanks for the kind offer. I expect it will take me many more weeks
>to get something useful; I may take you up on the offer then.
That is fine, contact me anytime at
i.mcm...@hcem.canterbury.ac.nz
reverse "hcem" to give "mech"
>
>I've been wondering about something, and given your experience,
>perhaps you could hazard an answer: Shouldn't it be possible to
>use Lorensen and Cline's Marching Cubes algorithm to generate a
>polygonal surface mesh of a given bone, and then import that mesh
>into a solid modeling package? I mean, shouldn't it be easy to
>turn such a mesh into a solid?
To be honest I don't know, I didn't really look at that type of solution and
it isn't really my field, but if you have the time and the skills....
Good luck, if I can help anymore let me know
It is nice to be able to help someone rather than my normal cruising the group
for tips and solutions
Regards
Iain
Jeremy Winston
> [...]
> > Paul Salvador wrote:
> > > [...]
> > > And for reference, a iges curve export/import into Rhino3D,
> > > patched and trimmed, looser tolerance otherwise it gets spikey.
> > > http://zxys.com/swparts/tibia_scan_2003_fillsurface5.png
>
> Glad tidings of great joy!
> 1. Rhino3D for students is a mere $195, and
> B. Rhino3D supports Parasolid (.X_T)[*]
In an ironic coincidence, I just received a flyer in the mail
advertizing Rhino, and a *free* 3-D Modeling Workshop a short
drive away from me, scheduled for ...
October 26.
Ah, well. Murphy's Law is inviolate.
I don't suppose the user's manual or tutorial docs are available online?
-Jeremy
Jeremy Winston
>
> Ok, looks like you found some of the other questions you were looking for.
> The answer to this question is: Rhino does not have native support for SW.
> You can not read in a SW file into Rhino.
> But you can use a plugin which allows SW to read in native Rhino 3DM
> files but not write them out.
Paul,
I d/l'd & installed the Rhino eval, and I'm going through the tutorials.
Before I spend a lot more time on this, could you verify something for me?
To wit, if I create a NURBS surface in Rhino, and save it as an IGES file,
will SW be able to import it in a form that can be saved as a Parasolid
.X_T file?
> For your use, Rhino is probably perfect?
I am fervently hoping so.
> With your need for x_t export and it's support for data import and
> exports or point cloud support and curve support through points, and the
> strong support for scripting, Rhino is a very good tool for reverse
> engineering.
>
> Also, look at their plugins page for other specialized resources..
> http://www2.rhino3d.com/resources/
Ta, will do.
-Jeremy
Paul Salvador
You can save the file directly from Rhino as a X_T.
But I think I may understand you concern.. is it because Rhino x_t
export does not not support curve or point export?
If you need to export curve and point data, you will have to export as a
IGES and read into SW then export as a X_T. So, to answer your
question on that, yes, you can import the iges and save as a x_t "but" I
think the think which will export will be curves, not the points.
I mention points because in Rhino, I extracted points from your curves
to create the surface patch. There are some tuts on the web on
converting or extracting point data, like scanned data, or mesh point
data (STL, DXF, LWO, 3DS,..) to nurbs surfaces... like, patch or
drape...
http://market.renderosity.com/~rhino/index.htm
http://www.angelfire.com/myband/coop/biboveralls.html
..
Jerry Steiger
--
Jerry Steiger
At Work Computers
"Paul Salvador" <p.sal...@verizon.net> wrote in message
news:3DC33FF2...@verizon.net...
> Jeremy,
>
> I got it to import into SW2003 as the others but I went a bit further by
> copying the curves to 3D curves and simplifying them. Also I did it
> with sections or at .04" steps and copied the curves again and
> simplified them.
> It does not look pretty but it is but who knows, it may look close to
> your scanned image or part which exception to the peaks.
>
> Here's the file and images...
>
> http://zxys.com/swparts/tibia-Part-1.zip
> http://zxys.com/swparts/tibia_scan_2003_fillsurface1.png
> http://zxys.com/swparts/tibia_scan_2003_fillsurface2.png
> http://zxys.com/swparts/tibia_scan_2003_fillsurface3.png
> http://zxys.com/swparts/tibia_scan_2003_fillsurface4.png
Nice work, Paul! It never occurred to me to try a fill surface. Live and
learn!
Jerry Steiger
Jerry Steiger
> And for reference, a iges curve export/import into Rhino3D, patched and
> trimmed, looser tolerance otherwise it gets spikey.
>
> http://zxys.com/swparts/tibia_scan_2003_fillsurface5.png
How much work was the Rhino part compared to the "native" SW part? It
certainly looks nicer. Maybe I should think more seriously about getting
Rhino for our product design work.
Paul Salvador
Hey, can I get the mandible data? Preferably nurbs (iges, x_t,
step,..) or polymesh (STL, RAW,..) and lastly point data (csv, prn,
txt) will do?
later.. (btw, luv the god and satan story at amc!)
Jeremy Winston
>
> Hello Cliff,
>
> Hey, can I get the mandible data? Preferably nurbs (iges, x_t,
> step,..) or polymesh (STL, RAW,..) and lastly point data (csv, prn,
> txt) will do?
>
> later.. (btw, luv the god and satan story at amc!)
Out of curiosity, I searched the NG hierarchy to figure out what
"amc" was (news:alt.machines.cnc), then subscribed & browsed
through Cliff's posts -- HI-LARIOUS!
Rich with sig fodder and more than a few forwardable humor posts.
Kudos, Cliff.
-Jeremy(or maybe Snickers would be more appropriate?)
Cliff Huprich
<p.sal...@verizon.net> writes:
>Hello Cliff,
>
>Hey, can I get the mandible data? Preferably nurbs (iges, x_t,
>step,..) or polymesh (STL, RAW,..) and lastly point data (csv, prn,
>txt) will do?
Reply Emailed to Paul.
>later.. (btw, luv the god and satan story at amc!)
LOL.
Always happy to brighten a lurker's day.
You should have read the *good* stuff.
--
Cliff Huprich
Cliff Huprich
<win...@dog.rip.edu> writes:
>Kudos, Cliff.
CONTRACTOR'S STANDARD PIPE & FITTING SPECIFICATION
--------------------------------------------------------------------------
-----------------------
(1) All pipe is to be made of a long hole, surrounded by metal concentric
with hole.
(2) All pipe is to be hollow throughout entire length. Do not use holes
of different length than the pipe.
(3) All pipe is to be of very best quality, perfectly tubular or pipular.
(4) All acid proof pipe is to be made of acid proof metal.
(5) OD of all pipe must exceed the ID otherwise the hole will be on
the outside.
(6) All pipe is to be supplied with nothing in the hole so that water, steam,
or other stuff can be put inside at a later date.
(7) All pipe is to be supplied without rust, as this can be more readily put
on at the job site.
(8) All pipe is to be cleaned free of any covering such as mud, tar,
barnacles,
or any form of manure before putting up, otherwise it will make lumps
under the paint.
(9) All pipe over 500 feet long must have the words "long pipe" clearly
painted
on each end so that fitter will know that it is long pipe.
(10) Pipe over two miles long must also have these words painted in the
middle so that fitter will not have to walk the full length of pipe to
determine if it is long pipe or not.
(11) All pipe over six inches in diameter is to have the words "large pipe"
painted on it, so that the fitter will not use it for small pipe.
(12) All pipe closures are to be open on one end.
(13) All pipe fittings are to be made of the same stuff as the pipe.
(14) No fittings are to be put on pipe unless specified. If you do, straight
pipe
becomes crooked pipe.
(15) Fittings come in all sort of sizes and shapes. Be sure to specify the
direction you are going when ordering.
(16) Fittings come bolted, welded or screwed -- always use screwed. They are
the best kind.
(17) Flange must be used on all pipe. Flanges must have holes for bolts
quite separate from the big hole in the middle.
(18) If flanges are to be blank or blind, the big hole in the middle must be
filled
with metal.
(19) All flanges must be cast or forged of the very best iron metal, close
and grained, free from blow holes, lumps, cavities, pock marks, pin
pricks,
and warts, otherwise we can't use them.
(20) Gaskets are to be used to fill spaces between flanges.
(21) Gaskets are to be made of metal, rubber, [plastic, paper or some kind
of goop -- do not use cow or sheep manure, it cracks when it gets dry.
(22) All bolts are to be screwed.
(23) All bolts must have a head on one end and a nut on the other.
(24) Bolts without heads are to be furnished as studs.
(25) Studs without heads are to be screwed all over and have two nuts which
is standard. Studs with three nuts are not to be used since they would
be odd.
(26) All nuts are to be furnished in sacks. Sacks must be whole and sound,
minimum two nuts per sack. Paper sacks will not be tolerated.
(27) All piping must be installed with valves.
(28) All valves must have an opening on each end with a flapper in the middle
that goes up and down or sideways when you turn the wheel or crank so
that it will open or close, otherwise the stuff will run out the ends.
(29) Valves are to be furnished by the kind required as follows:
Ball valves are to have a ball inside
Gate valves are to have a gate inside
Globe valves are to have a globe inside
Check valves are to have a check inside
Angle valves are to have an angle inside
Plug valves are to have a plug inside
Diaphragm valves art to have a diaphragm inside.
(30) Ball valves are not to be used anywhere with a female connection.
(31) Diaphragm valves are to be used with a female connection.
(32) All completed piping lines must go somewhere and connect to something.
The fitter is required to verify this before turnover.
(33) All pipes shorter than 1/8" are uneconomical in use, requiring many
joints.
They are generally known as washers.
(34) Other commodities are often confused with pipes. These include: Conduit,
tube, tunnel and drain. Use only genuine pipe.
(35) Scottish Regiments in the Army use Army pipes in unusual ways. These are
not approved of in engineering circles.
(36) When ordering 90 degree or 30 degree elbows, be sure to specify
right hand or left hand, other wise you will be going the wrong way.
(37) Be sure to specify to your vendor whether you want level, uphill or
downhill pipe. If you use downhill pipe for going uphill the product
will
flow the wrong way.
(38) All coupling should have either right hand or left hand thread, ut
do not mix the threads, otherwise, as the coupling is being screwed on
one
pipe, it's being screwed off the other.
(39) Joints in pipes for water must be water tight. Those in pipes for
compressed air however need only be air tight.
(40) Lengths of pipe may we welded or soldered together. This method is
not recommended for concrete or earthenware pipes.
(41) Pipe over 5 feet in diameter shall have an arrow* painted at 10 foot
intervals
with the word EXIT.
(42) When such pipe is longer than 20 feet all the arrows* must point in the
same direction.
* It has been suggested that the arrow(s) point to the open end of the
pipe.
This option is under study by the committee.
Tom Charron
qualified, manufacturable NURBS surface model (IGES or STEP) called
RapidForm. Using RapidForm Volume, you can convert dicom data into a
CAD model through the following process:
- Open the dicom data, visualize it in 3D and identify your Volume of
Interest
- Automatically convert your VOI into a polygon mesh
- Clean up the polygon mesh with a variety of tools
- Convert the polygon mesh into a NURBS surface automatically
Depending on what is being surfaced, this can take as little as an
hour or less to go from dicom to finished NURBS model (longer for
complex objects).
This has been achieved by adding a 3D Imaging component to RapidForm,
the most popular software for 3D scanning. RapidForm offers the
fastest method to get an accurate polygon mesh or NURBS surface from a
CT or MRI scan.
Go to http://www.rapidform.com/product/pro_07.htm for more
information.
Tom Charron
U.S. Marketing & Business Development Manager
INUS Technology, Inc.
www.rapidform.com