Strange lathe motion.

I have to admit not completely understanding what you mean. Do you mean the program will run complete without any issues 2 times, but on the 3rd time it fails? Always on the third time? Or is it just random failures?

Is z not moving at all when it fails or does it move straight?
If it moves straight, does it go to the arc endpoint as if you had programed a G1? (I see its G2, just trying to be clear)

Thanks for the complement. (Gentlemen)

John types faster I suppose

When I try and run your code on the lathe sim I get this error.

image missing http://www.linuxcnc.org/media/kunena/attachments/legacy/images/aike.png

John

BigJohnT wrote:

While I can't test here I do know that even if you have something in your RS274NGC_STARTUP_CODE it can be overwritten by any g code file or a MDI command. So if you ran another file that changed them even just running an example file it can change your current modes. Things like that should always be in the preamble of your g code file.

Do you mean the problem occurs the 3rd time you run the code?

John

Thank you for explain, I will insert this sequence of command in the header of g-file.
The error occures randomly.
Looks like the error in interpolation (float point operation) caused in not controlled moving.
I think so
The program becomes to live its own life

aike wrote:

The error occures randomly.
Looks like the error in interpolation (float point operation) caused in not controlled moving.
I think so
The program becomes to live its own life

I ran your file in my back plotter. Each of the arcs has an issue with the end point. Is my back plotter crazy or are the arc centers/radius really out of whack? The code likely would run but the results would be unexpected. I attached the plot.

image missing http://www.linuxcnc.org/media/kunena/attachments/legacy/images/arcs.png

If I understand what your trying to do... I re-wrote the arcs. I left the radius unchanged but ended each arc properly by moving the X end point. I then added a move to the original end point. Does this look more like what you'd expect?

Original code:
N20 G2 X46.5 Z36.5 I0. K3.
N28 G2 X49.1 Z36.5 I0. K4.3
N30 G3 X54.7 Z40.8 I2.8 K0.
N37 G2 X50.5 Z36.5 I0. K5.
N39 G3 X54.7 Z40.2 I2.2 K0.


Changes:
N20 G2 X43.5 Z36.5 I0. K3.
G1 X46.5

N28 G2 X44.8 Z36.5 I0. K4.3
G1 X49.1

N30 G3 X51.85 Z40.8 I2.8 K0.
G1 X54.7

N37 G2 X45.5 Z36.5 I0. K5.
G1 X50.5

N39 G3 X52.7 Z40.2 I2.2 K0.
G1 X54.7

I've attached the plot
image missing http://www.linuxcnc.org/media/kunena/attachments/legacy/images/arc2.png

Yes, looks like you are right.
The start and end point not concerned, the software doesn't know what to do, try to search the point, hence the moving is unpredicted.
I will test your changes in lathe, Great THX!