Work with probe

mariusl wrote: What is the other piece of software next to you probing stuff on your picture a couple of posts back?

This is Features for linuxcnc from Nick (cnc-club.ru) - construction of g-code directly from linuxcnc

mariusl wrote: Hi
Gmoccapy loads with the probing stuff loaded in the usertab but when I run it the system goes to MDI mode and nothing further happens. Is there some other stuff that must be installed or not?

I'm a little touched up installation instructions.

Check at this part

1. .ini
2. The following files from the archive placed in:
....
your-folder-configuration/your_subroutine_path
all from folder "macros"
...

I have those settings in the .ini file.
I see that your version of Gmoccapy is a bit old and not standard any more. Did you make a lot of changes to gmoccapy?

mariusl wrote: I have those settings in the .ini file.
I see that your version of Gmoccapy is a bit old and not standard any more. Did you make a lot of changes to gmoccapy?

Just a bit old screenshot. No any changes to gmoccapy, only custom additional panels.

I cannot get the code to work but something else might be the problem. I have Ngcgui loaded as well and for some reason when ngcgui is loaded the Gmoccapy cannot find any files that are in the macro directory. I have this on my lathe and on my mill.
I cannot get Andy's Lathemacro to work when I have ngcgui either.


Would you mind sharing how you got Nick's Features to load and work with Gmoccapy?

mariusl wrote: I
Would you mind sharing how you got Nick's Features to load and work with Gmoccapy?

My translation from this (for sim gmoccapy)

Installation:
features.zip unzip to ~/linuxcnc-dev/lib/python/gladevcp/
features_gmoccapy.zip unzip to ~/linuxcnc-dev/configs/sim/gmoccapy/
In file ~/linuxcnc-dev/configs/sim/gmoccapy/gmoccapy_4_axis.ini (or your-config.ini) in section [DISPLAY] add: Do: Insert in ~/linuxcnc-dev/lib/python/gladevcp/hal_pythonplugin.py, after "calculator" : Insert in ~/linuxcnc-dev/lib/python/gladevcp/hal_python.xml
after first "calculator" after second "calculator":
Then, update only folder ~/linuxcnc-dev/lib/python/gladevcp/subroutines from github

Inside circle routine isn't working for me. The first -X move is executing twice, once in the python code(line 975) and once in the macro xminus_in.ngc (line 10).

LinuxCNC version is 2.6.4.

elf wrote: Inside circle routine isn't working for me. The first -X move is executing twice, once in the python code(line 975) and once in the macro xminus_in.ngc (line 10).

This is bug. It's corrected. Thank you for accurate description.

The updated code fixes the double move bug. There is another bug in the python code where the program assumes both probed positions have the same sign [line 1028: cxres=0.5*(xres+xres1)] Note: the docs don't say if the probed_position is relative or absolute: linuxcnc.org/docs/html/common/python-interface.html

The home position on my machine is in the center for all three axes so when the circle is centered on the home position, I'm getting wildly different results. When I move the circle to +X and +Y, the code works as expected.

A minor issue is xplus.ngc and yplus.ngc contain Z axis moves. I think these should be moved to the python code.

elf wrote: The updated code fixes the double move bug. There is another bug in the python code where the program assumes both probed positions have the same sign [line 1028: cxres=0.5*(xres+xres1)] Note: the docs don't say if the probed_position is relative or absolute: linuxcnc.org/docs/html/common/python-interface.html

The home position on my machine is in the center for all three axes so when the circle is centered on the home position, I'm getting wildly different results. When I move the circle to +X and +Y, the code works as expected.

Screenshot in the header of this topic by the way shows the result of measurement of hole. We see that true equality "Xc"=0.5*("X+" + "X-"). In Probe Screen application probed_position matches with absolute coordinates . I do not quite understand why it is wrong.

elf wrote: A minor issue is xplus.ngc and yplus.ngc contain Z axis moves. I think these should be moved to the python code.

Now I think so too