Re: [CAD_CAM_EDM_DRO] Watch finishes, by CNC

Chuck Knight wrote:

>http://mywebpages.comcast.net/timconnor/ham922/ham922_07.jpg
>
>I want to do this type of finishing to a watch movement, just as an
>experiment...isn't it pretty? This style of decoration is done by using
>a rotating emery disc, and a watchmaker's lathe...lots of tedious
>handwork. I'm thinking of attaching an emery disc onto the end of a
>dowel, and chucking it into a router. Needless to say, there is no
>pressure involved, but only light contact and a precise and slightly
>overlapped toolpath. Given a small, reasonably accurate CNC router,
>this should be simple to achieve. FYI: This watch is from the turn of
>the last century -- one with this level of decoration would cost a
>minimum of $2500, and most are closer to $4000, in today's market. This
>one cost >$100, 100 years ago...this is one of the super-accurate
>railroad watches that kept the trains on time.
>
>Back to CNC. As I see it, once I have a tool made, it should simply be
>a matter of programming the toolpath. Does anyone have a program that
>would automatically generate a spiral, concentric circles, straight
>parallel lines (cotes de geneve) and other simple "filler" patterns
>which I could use for this purpose? I don't want to "reinvent the
>wheel" if I don't have to. I already know that the dust generated will
>require me to clean the movement afterwards, but for the purpose of this
>experiment, I'll happily do it. I could probably do it myself, but I've
>never programmed in G-Code.
>
>I know there are simple "pocketing" programs floating around...what
>types of paths do they generate?
>
>

I have a program available for download that will do something quite close.
See http://jelinux.pico-systems.com/gcode.html and it is the first
program in
the list, "Boring holes". You would set it so the bore depth and depth
per pass
would bring the tool to the right depth for the emery disc. So, if you had
the machine set so that the tool would touch the work at Z=-0.1, then put
0.1 in for bore deoth and depth per pass, and it will generate one
sort-of-spiral
with the tool at Z=-0.1 and stop. The width of cut is the radial
advance per
full turn. It starts at the center and works outward, as you'd want for
boring a
hole. It looks like your sample might have started at the outside and gone
inward. If you set the diameter larger than the piece, then the finish
settings
are not important. Otherwise, I think you'd set the finish width of cut to
zero, but the way the code is written, I'm not sure you can get rid of the
finish pass. That is done at constant radius.

Anyway, you can try out the program to see if it gives the moves you want.
I have both a dos executable as well as portable C code available for
download. You can choose your favorite dialect, using arcs with I and J or
radius.

When I say "sort-of-spiral" above, it is done by computing a set of arc
quadrants of increasing radius. The center of the arcs are offset by the
increase in radius from one to the next, so the endpoints meet. I don't
know
if that would show up as an undesirable inflection at each 90 degree point.
The larger the expansion of the spiral per turn, the more this would make
a difference. Most CNC controls can't do a true spiral except by a huge
number of short vectors or arc segments. The technique should work for
arcs smaller than 90 degrees, I think.

Jon