Non-circular boring. (Linuxcnc fun)

[skunkworks]

I have two questions;
1. Is the cutting mainly done by the XY table movement and the tool rotation is principally to present the tool at the required rake angle to the line of movement at each location on the path?

Kind of - it would be impossible to keep the tool at the correct rake angle unless the tool had no thickness. Even then. The code just keeps the tool perpendicular to the center of the polygon. It works surprisingly well it seems (even if the tool isn't exactly at the right angle compared to the material.) So - half the time it is negative-ish and half the time it is positve-ish

2. I am not familiar with Linux nor LinuxCNC, but I notice that you use a HAL file for the control. Is that because LinuxCNC does not use Gcode (I thought it did) or does a HAL file have advantages over Gcode?

The Hal component is what does the realtime stuff to do the motion. This can be controlled by gcode. The code is run 1000 times a second and reads the spindle position and calculates where x/y need to be. There are a bunch of parameters that you set (number of sides, cutter diameter, polygon radius and so on) I will get my gcode example to show you. The next step would be to do what is called a gcode re-map. So you could make it more user friendly. (one line of gcode maybe sets and enables the component)

I have been thinking of making a rotary broach but this is a much better approach for general work because you do not need a new broach for each size and shape of hole. A small and large boring bar is all you need for any size and shape. Now I'll have to write some Gcode, although a search of the net might turn up a wheel already invented by another.
Nice work BTW.

PS. I just realised that I'll have to add a rotary encoder to my spindle before I can follow your example. I have been meaning to do that for a while now, this is just the motivation needed.

thanks!