Circular path (or any path that require two axis movement at one time)

Watch the springs while the machine is cutting, they are likely moving while cutting. Getting the right ammount of pressure on those things is very hard as that depends on a lot of mechanical stuff, like how far is the axis that the reduction is mounted on, what angle it is set at etc.

Speed variation in a cut on only one axis will not show up in the work, it is only when you have two axes moving that you see the problem.

So, this is because one or more axes are not moving at a steady speed.

This could be mechanical, it could also be in the motor controllers if these are closed-loop steppers or step-servos.

I did at first think that it could be the PID tuning in LinuxCNC (there is a PID even with steppers) but the Mach3 swap eliminated that possibility.

So, what motors and drivers are you using?

I see now that you said it was steppers and simple step/dir drivers.

I think that analysing the pitch of the waves (orthogonal to each axis) is likely to be a very instructive step. Does it match one turn of a shaft, or one turn of a gear.

After some analysing it looks like you have something sloppy there as the distance of the anomaly is not exactly the same always.
Check well everything, there is no easier way to say it.
BTW, did you edit anything in the INI and HAL files ?

tommylight wrote:

After some analysing it looks like you have something sloppy there as the distance of the anomaly is not exactly the same always.

The wiggle pitch will vary depending on the direction of motion. (but, if mechanical) should be constant if projected to an axis.

This picture shows the result of a constant wiggle on X combined with a constant Y at various angles.



You will see nothing in a straight Y move (as X doesn't move) and nothing in a straight X move (As the wiggle is along the cut line)