Re: [CAD_CAM_EDM_DRO] Re: Help needed with G code

Alan Marconett KM6VV wrote:

>Hi Jon,
>
>Arc moves in two planes at once sounds like a neat math problem! Any
>idea how it's done? I understand (and have implemented) moves in each
>of the single planes. How often is something like this used? OK,
>probably not necessary in a hobby controller!
>
>

Well, it depends on your hobby. Yes, of course, even many commercial CNC
controls don't do this, or require an expensive option to be enabled.

Some of the higher-end controls can do NURBS surfaces or any surface
that can be defined mathematically with standard mathematical functions
such as sine, tangent, etc. They can accept the formula right in the
G-code.
If you do exotic molds and stuff, it can allow 10 lines of code to replace
500,000 blocks of small linear moves.

>R radius parameters are not that hard to implement, but I'm told they're
>not always useable (or accurate for all arcs). The do appear to make it
>easier to write arc blocks in an "interactive mode". I guess it just
>depends on what you know about what you're drawing.
>
>

Really, all CNC controls create an internal radius value when
interpreting the
G-code. They generally test that the radius is the same between the defined
center and the starting and ending points for the arc. If the mismatch is
greater than some limit value, the program aborts the toolpath. The one
place
the R is ambiguous is a full, 360 degree circle. If you are at some
point, and
are told to make a 1" circle, should you start left or right? The I,J
parameters
completely specify the center of the arc, while the R does not.

>Adding a linear move along with an arc move is also not that hard
>(helical), but I haven't made much use of it... yet.
>
>

Great for thread milling!

Jon