Re: Proposal: 5-Axis machining in STLWork, Millwizard and Deskproto

--- In CAD_CAM_EDM_DRO@y..., vrsculptor@h... wrote:

> I think there is a relatively easy way to implement 4 or 5 (or X)
> axis machining in point and shoot CAM programs such as STLWork,
> Millwizard and Deskproto without complex re-writes or post
> processors.
>

No this is NOT 4 or 5 axis machining. It is 5 axis positioning and 3
axis machining. Actually only 2 as all the raster programs only
generate planar cuts. Even the waterline Strategy in StlWork2 is
planar.

> Consider the case of machining a rolling pin lying on a cutting
> board. The rolling pin is aligned with the x-axis and the cutting
> board is square and several times the width of the rolling pin. A
> reasonable manual machining strategy would be to set up the blank
> with the rotary head aligned on the centerline of the rolling pin.
> You would use the X and Y axis to surface the cutting board and the
> rotary axis to surface the rolling pin and to do the undercuts on

the

> cutting board that are not visible from above.
>

A better approach in this case is to redesign the model and make a
cutting board and a rolling pin as two separate parts. You will be
hard pressed to find a tool that can reach under a rolling pin and
also finish the top of the cutting board with the side of the tool.
At a minimum you will want to position to about 45 degrees instead of
90. this way you will be tangent to the radius of the cutter, high
up on the side and the cutting speed will result in a much better
finish on the flat sweeping surfaces. This is a good strategy for
any situation that you have to machine with a raster pattern. Tilt
the flat surfaces at about 5-10 degrees. You will never be finishing
the most visible part of your part with the slowest and nastiest part
of your cutting tool, the DEAD center of the ball. Things change if
you use a bull cutter, which is a flat end tool with radiused corners.

> STLWork (and BobCad's) apporach of simply replacing the Y-axis with

a

> rotary or wrapping my geometry around an axis won't let me do this.
>

There is some confusion about this. StlWork has no provision for
rotary axis output. It will allow you to position the model with
respect to the output axes, but it will only generate X-Y-Z code.

Neither Bobcad nor Vector will develop tool paths from complex
tessellated surfaces like that used for STL, vrml, 3ds, dxf mesh,etc.

Vector however will generate X-Z-A(pick your rotary axis) code from
any wireframe. It ALSO has a function that allows one to wrap a flat
surface around a radial axis. It does this visually and actually
creates the wireframe, which CAN then be used to generate X-Z-A code.
There are two distinct functions described in Vector.

> An alternative to this using STLWork's current features is to
> generate three separate sets of G-code. One each, top, left and
> right. Mount the blank as described above and do three separate

runs

> manually rotating the table 90 degrees left and right. The only
> problem with this approach is that the operator needs to reposition
> the work's origins to compensate for the axis rotation offest and I
> don't think STLWork can be convinced not to re-machine surfaces

that

> were cut in the prior orientation.
>

StlWork2 will not automatically prevent or avoid remachining, however
it will let the user define the boundaries of any number of cutting
strategies by lineal axis boundaries and also by the surface
condition in the case of the waterline strategy. The cutting
boundaries are independent of the model orientation.

> Suppose you were able to tell STLWork where the rotary axis in the
> blank was located. You could then generate G-Code to machine from

the

> top, ask the program to rotate the object 90 degrees around your
> defined axis, machine just the material that was not removed in the
> prior pass, rotate back 180 degrees and remove the rest of the
> material. The operator would then not have to manually reset the
> origin.
>

That is exactly how StlWork presently works.

> If the rotation feature was implemented with three rotation options
> the functionality would be extended to most everyones machines: A)
> Automatically generate G-Code to do the rotation; B) Allow the

manual

> entry of G-code to do the rotation assuming that the program left

the

> cutter above the axis in a retracted position; C) Generate a G-Code
> operator prompt to tell the operator to do the rotation manually.
> This should address about everyone's needs.
>

Use a text editor or the NC code editor in Vector to manually add
this kind of information. We do not plan to have an editor in
SltWork. It is designed to complement not supplant Vector.

> The 5th (and Xth) axis is a variation on the theme. If you could
> define a second axis in the blank you could ask the program to

rotate

> around this axis

You can rotate the model in SltWork using the X, Y, or Z axes. The
generation of the G-code and the cutting strategies to cut each side
are calculated independently of each other, and would be seperate
programs. Stlwork can append these into a single file, but more than
likely there would be a couple of minor edits required, at best.

A90 B90 (after the first cut)

A-180 B-180 (after the second cut)


Problems that you still need to consider are interference and gouge
checking if there is a hole in your cutting board, and at an angle
NOT parallel to any axes, or with a kink inside. Also how do you
handle overhanging cutouts, like the inside of a toilet bowl. More
food for thought.(true 5 axis problems)

Best Regards,

Fred Smith
IMService