Faceted surface on G2 Lathe command

More
12 Feb 2018 13:17 #105787 by Max_Spencer
I am trying to cut a simple profile composed of a few G2/G3 commands. It seems there are some minor faceting in the middle of the arc commands. Attached is a picture of the cut part where you can see the banding on the arc profiles. It appears this is not at the intersection between a G2 and G3. I have tried various combinations of G61 and G64 commands which appear to do nothing. It appears as if there is some tolerance used in the G2/G3 maneuvers or something needs optimized in my settings. I am running linuxcnc 2.6.13

I am running a sherline lathe with the normal stepper motors. Attached is my ini file, hal, gcode, linuxcncstatus snapshot.

Am I expecting too much from this machine to give a smooth radius without facets that i can feel with my finger?

Thanks, Max
More
13 Feb 2018 14:54 #105823 by Max_Spencer
Tried some more today. Installed a fresh linuxcnc 2.7 on another PC, and created a new INI using stepconf. It cuts the exact same way. Under a microscope it seems like a series of linear lines instead of an arc. Is there a tolerance somewhere for the circular interpolation?
More
13 Feb 2018 16:19 #105830 by skunkworks
what are your full steps/mm ? I wonder is you are seeing your stepper resolution?

sam
More
13 Feb 2018 16:45 #105831 by Max_Spencer
I don't think it is related to steps. running 200 steps/1mm. these facets are much larger than that.
More
13 Feb 2018 20:44 #105836 by skunkworks
I am not seeing anything config wise - other than someone notice that your following error was rather large (26mm) you could try lowering the following error and see if it trips. I would be leaning towards a mechanical explanation. (stiction or play)
More
13 Feb 2018 21:29 #105837 by Max_Spencer
Thanks. I will lower down the following error to make it more realistic. On these open loop steppers i don't know if it will make a difference. But your comment on stiction has me thinking....i've only been testing with arcs such that the diameter is increasing while feeding. depending on tool pressure, it might be stiction holding the cross slide in place until the leadscrew backlash forces it to 'pop' into the next position. I'll reverse the direction tomorrow so the diameter is decreasing which should keep the tool tensioned against the same side of the leadscrew.
More
13 Feb 2018 21:40 #105839 by robh
i would run a lathe in G95 (feed per rev) only using G94 on milling etc

in G95 try a feed rate of around 0.1mm/rev this tends to give food finish on a 0.4mm Rad tool
see if this improves things ..
Time to create page: 0.072 seconds
Powered by Kunena Forum