Here's a little one I've been trying to get right for a while. It's a ten-bladed impeller, made using G68 angular offsets, nested subroutines and primitive parametrics. As an actual pump impeller it's probably pretty crappy, but it shows the principle well and you could easily get the feel of this code and apply it to any sort of rotational shape.

One thing that hexed me throughout was trying to apply a tool diameter offset to the program ... try it and see what happens, it's extremely wierd. Is there a way round this, or can tool offsets not be used in G68? Let me know, I'm intrigued as to what it is.

(CENTRIFUGAL IMPELLER)
(TEN BLADES)
(NOT DESIGNED AS SUCH)
(THIS IS JUST A DEMO :p )

G40
G64
G21
G0 X0 Y0


#1=0
#2=4.5

M3 M8

G00 Z1.00
M98 P777 L5
G00 Z10
M5 M9
G00 X0 Y0
M30

O777
#2=[#2-4.0]
M98 P333 L10
G00 Z1.0
M99


O333
G68 A0 B0 R#1
G00 Z1.0
G00 X100 Y-60 F700
G00 Z[#2+1]
G01 Z#2 F150
G02 X45  Y0 R50
G03 X100 Y-60 R80
#1 = [#1+36]
G40
G69
M99
%

Ok the code should work but I don't see where you applied the ToolComp.

(;-) TP

this is it with the comp in:

(CENTRIFUGAL IMPELLER)
(TEN BLADES)
(NOT DESIGNED AS SUCH)
(THIS IS JUST A DEMO :p )

G40
G64
G21
G0 X0 Y0


#1=0
#2=4.5

M3 M8

G00 Z1.00
M98 P777 L5
G00 Z10
M5 M9
G00 X0 Y0
M30

O777
#2=[#2-4.0]
M98 P333 L10
G00 Z1.0
M99


O333
G68 A0 B0 R#1
G00 Z1.0
G41<---------------------------------------------------HERE
G01 X100 Y-60 F700
G01 Z[#2+1]
G01 Z#2 F150
G02 X45  Y0 R50
G03 X100 Y-60 R80
#1 = [#1+36]
G40
G69
M99
%

YEP it appears ToolComp is busted in G68 in that version. Brian Has a new TOOLCOMP coming out Lets hope it is fixed. You may want to turn in your problem as a bug report. (;-)

(;-) TP

Ah, ok, thanks. I'll do that.

So far so good with Mach3, I have a lot of confidence in it, and had assumed until now that it was me being dumb. tool comp would be useful in G68 thanks.

ANother thing youmay want to try is to NOT use the R method of arcs. Try the IJ method.

(;-) TP