Turning a 6mm sphere produces a bullet shape.

I am experiencing a problem trying to machine a 6mm dia ball. I am using a parting tool to machine front and back curves. The front curve is fine, but after moving tool to compensate for the tool width, and then continuing with the corrected arc, the result is a failure, ending up as a bullet shape. I have used ECAM (free on weekends) to produce the gcode, tried both the radius and IK methods, but same result. The back curve just does not come out correctly.

Anybody got any idea's. Is it my cheap eastern breakout board, or does LinuxCNC have an as yet undiscovered flaw?

The Fusion turning CAM is pretty good now. There are a few features still under development (such as adaptive roughing) but it's all free and the preview is pretty powerful.
knowledge.autodesk.com/search-result/caa...8F0A5FA28B5-htm.html

Hi,
The front is fine, the rear deformed, a parallel band then a sharp curve to the stem the ball is mounted on. However, I have detected some backlash in the X ballscrew, so I will add a backlash move, X 7, X6 at the quadrant point after tool compensation z move. The preview in LinuxCNC shows an elongated ball, circular front and back with a parallel portion, as expected.

(I have found that Mach3 and LinuxCNC don't always interpret the code the same way.)

The problem was found to be the result of a failing ballscrew nut on the X axis.
Thank's for your interest. I shall replace the nut.
I'm pleased to know LinuxCNC and the BoB are not at fault. I enjoy working with LinuxCNC

Interesting job.
Backlash would seem to be my problem. I have implemented it now on LinuxCNC and Mach3. At the moment I have made enough satisfactory items. I am awaiting a carbide parting tool and holder as advertised in eBay and others.