Make arc in W axis

Hi..new here....switching from mach 3...I set up 8 axis so far and want to run a test using my W axis to do an arc ....Should just changing axis name in file fro Z to W work or is there more to it than that?

Thanks for any help,

Rick

As far as I know arc moves are only possible in XYZ. Maybe you could describe your setup in more detail as there may be another way to get what you want.

[edit]
 
linuxcnc.org/docs/html/gcode/g-code.html#gcode:g17-g19.1

Hi..thanks for reply..on a wood lathe I would have rough cut on one side opposed on other with finish cutter...both would need to do arcs but at slightly different times ..also finish cutter acting like skew would have to reverse while rough waits and the both move forward again though offset slightly still. So that is at least two axis on opposing sides sometimes doing arcs simultaneously...finish cutter actually has another axis to rotate cutter to keep bevel angle to wood....would using angular instead of linear work?

Thanks again,
Rick

So that is at least two axis on opposing sides sometimes doing arcs simultaneously...

It is not possible to execute two arcs simultaneously. What you could do is use the switchable kinematics feature to switch the X and Z axes to different joints. So you could run regular gcode with arcs in the XZ plane and switch in gcode on which side of the work piece the movement takes place. Remember though that this is possible only one after the other not both simultaneously.

Arc moves are really only defined for linear axes XY, XZ or YZ. If you need different linear axes to perform arcs then switching the joint/axes mapping using switchable kinematics seems to be the only way to go.

To add to that, if you are wanting simultaneous synchronized motion of both your roughing and finish heads, you are not going to be able to use arcs at all. You will probably have to use a cam program to output sudo arcs made up of small line segments. This is really the only way to have two heads doing different synchronous movements. You may also need to use G93 inverse time feed commands (especially if you will also be using a rotary axis to control the finish head's angle.)

Imagine each line of g-code as a frame of a movie film, with each new line of g-code as an image of each axis' position at the frame rate's time interval. The difference being, unlike in a film the size of your time interval can be variable. For all linear moves you would only need a frame at the beginning and end. But for anything else the interval can be as small as you need to achieve your desired resolution.

Hi...Thanks for help...not sure why I thought that would even be possible.....will look into switching though ..may be valuable down the road..

Thanks!!!

Thanks for help!!...I will give that method a try.....also thought that I may be able to do both cutters off same gcode..only reverse one...the finish cutter may be able to live in the larger cutting point of the rough...working right behind trail edge of rough cut..

Thanks again!!

The problem isn't so much with Linuxcnc, it is more an inherent limitation of G-code itself.

To do what I think your asking for, you would almost have to have two separate G-code interpreters executing two separate, yet some how synchronized g-code files. I think there are some high end control systems that can do such things, but it is currently beyond the scope of Linuxcnc.

That said, I think I can imagine two separate Linuxcnc control computers on one machine linked with some of digital io bits to synchronize a set of parallel processed files. It could work, but would require some experimentation, and there may be unforeseen pitfalls.

Hi...so one more question...I was thinking to drive 2 tools off same axis from computer...the only thing is that they have different scales..meaning one is screw...other is gear....almost 10 to 1 difference...I was thinking of using 2 different drivers with different microstepping...does not seem to work...any other workarounds?

Thanks,

Rick

I was thinking of using 2 different drivers with different microstepping...does not seem to work...any other workarounds?

I don't see why that would not work. Basically a tandem axis, one axis mapped to two joints.