Re: [CAD_CAM_EDM_DRO] Digest Number 1196
Posted by
Andrew Werby
on 2001-03-17 11:55:49 UTC
Message: 19
Date: Fri, 16 Mar 2001 16:46:30 -0700
From: "Smoke" <gordonr@...>
Subject: Re: Re: CAM Software: Opinions Wanted
Hey...now THAT'S good information....thanks a bunch! I'll have to try that.
I'm no where near an expert on Rhino.
[No problem, Smoke. I'm far from an expert myself, but this is something
I've been doing a lot of lately, making cylinder seals on a 4 axis MaxNC. ]
Right now I'm using the MDX 15 for the scanning but eventually I want to
mount a probe on either one of the two machines I'm converting or on the new
ones I'm building so I'll be able to scan some larger artwork. 4" x 6" is
pretty small.
[That 2.4" depth is also a major limitation. How are you going to get the
probe software to work on the larger mill? It seems tied to the Roland
machine, which doesn't use standard g-code.]
As for milling the areas I want to reach...that should be pretty easy as
most of the work is going to be on cylinders or truncated cones with the
gcode set up to mill only in the X-Z direction (along the cylinder/cone
axis) and the part rotated in very small increments using a forth axis.
Smoke
[How are you doing that? I've been using the beta version of DeskProto 3.0,
which automatically does something like this. Have you got a program which
substitutes A values in degrees for Y linear values? Is it smart enough to
make the angle steps smaller when the diameter is greater? Does it stop at
360 and go back the other way, or keep incrementing the rotary values until
it must start dropping off thousandths?]
Andrew
Sculpture, Jewelry, and Other Art Stuff
http://unitedartworks.com
Date: Fri, 16 Mar 2001 16:46:30 -0700
From: "Smoke" <gordonr@...>
Subject: Re: Re: CAM Software: Opinions Wanted
Hey...now THAT'S good information....thanks a bunch! I'll have to try that.
I'm no where near an expert on Rhino.
[No problem, Smoke. I'm far from an expert myself, but this is something
I've been doing a lot of lately, making cylinder seals on a 4 axis MaxNC. ]
Right now I'm using the MDX 15 for the scanning but eventually I want to
mount a probe on either one of the two machines I'm converting or on the new
ones I'm building so I'll be able to scan some larger artwork. 4" x 6" is
pretty small.
[That 2.4" depth is also a major limitation. How are you going to get the
probe software to work on the larger mill? It seems tied to the Roland
machine, which doesn't use standard g-code.]
As for milling the areas I want to reach...that should be pretty easy as
most of the work is going to be on cylinders or truncated cones with the
gcode set up to mill only in the X-Z direction (along the cylinder/cone
axis) and the part rotated in very small increments using a forth axis.
Smoke
[How are you doing that? I've been using the beta version of DeskProto 3.0,
which automatically does something like this. Have you got a program which
substitutes A values in degrees for Y linear values? Is it smart enough to
make the angle steps smaller when the diameter is greater? Does it stop at
360 and go back the other way, or keep incrementing the rotary values until
it must start dropping off thousandths?]
Andrew
>The most important thing I need is not a CAD program (I use Rhino) but athis
>program that will take a complex bas-relief surface that has been "scanned"
>into the computer, saved as a .dxf or .stl or ??? file, wrap or project
>to a curved (possibly a complex curve) and then generate a GOOD gcoed forAndrew Werby - United Artworks
>machining. By GOOD I mean one that doesn't have to be "tweaked"
>considerably to make it work.
>
>Smoke
>
>[These scans are coming from your Roland MDX-15? If so, what I've found
>most effective is to use Rhino's "Drape" command to convert these to NURBS
>surfaces, then do the wrapping, flowing, etc. in Rhino. You use it by
>first maximizing the top view and making the image almost fill the screen,
>Then drag a box around the whole thing to create a "vacuformed" surface
>over your original scan. You can then delete the polygon mesh. Detail will
>be a bit softer, but the file will take up fewer bytes. This NURBS surface
>can be trimmed if necessary. To make it conform to a curve, first draw a
>straight line along the axis of the surface you want to bend. Then draw the
>curve you want to project to (direction matters, so if it comes out
>backwards, flip the curve direction and try again). Highlight the surface,
>then using the "Flow" command pick first the straight line (original
>backbone curve) and next your curve.
>
>Once the geometry is how you want it, export it as a DXF or STL, then use a
>g-code converter like MillWizard or DeskProto to write the toolpath. Of
>course, if you bent it very much, it might be hard for a 3-axis mill to
>reach all the areas you want carved- but you knew that already, right...]
>www.computersculpture.com
Sculpture, Jewelry, and Other Art Stuff
http://unitedartworks.com