More Z-Axis Issues

I have reported this before but now I have actual steps and more info on what is going on.

I am having problems with my machine losing the arc at the end of a cut. I have tried to cut the same thing 3 different times and it always fails in the same spot. It is at the very end of the cut, when the piece falls into the water pan I get the error. I call also see which line the gcode fails on and its always the line right before M5 to turn off the torch.

This causes all sorts of z-axis issues because when I hit resume, LinuxCNC almost never resumes. Maybe one out of 20 tries it will actually resume. When this happens, all of the axis freeze and I can't move the gantry at all. The only way I have found to resume is to hit the Z Move To Top button, then right clicking on the second line of gcode after the line it failed on. The reason I have to click on the second line after the failure is because the next line after M5 is Z0.039. If I right click on it and hit "Run From Here" the z-axis crashes into the work piece, luckily my torch pops out of its holder instead of doing damage.

How can I keep this from happening? I think the main issue is losing the arc signal at the very end of the cut. Is there a way to avoid this? I did not have this problem when running THC-Toma so I don't think it is a THC issue. Is there a way to say wait so many milliseconds after the arc lost signal before throwing the error?

I am using sheet-cam. I think I can make it remove the z-axis commands. Here is the g-code I had to save it as a txt file so I could attach it. I will go back out in the shop and see if the z-axis is commented out on the LinuxCNC machine and let you know.

The arc is always lost on the line before M05. For instance, below it would fail on line 300 and the z-axis will raise up 3 inches. Then I would have to click Move Z To Top to unlock gantry(I think the float switch is causing the gantry to freeze), then I would right click on line 330 and select run from here and cross my fingers that the z-axis doesn't crash or give an error that says float switch tripped while moving to probe height. If I right click line 320 or 310 and select run from here, the z-axis crashes 100% of the time.

N0300 X6.9827 Y7.3268
N0310 M05
N0320 G00 Z0.3937
N0330 X6.2158 Y7.3853
N0340 Z0.0000
N0350 M3 S1

Not sure what you mean by g-code parser. How do I disable it?

I went out to the shop and looked and the z-axis lines in the g-code do not appear to be commented out. All of the lines in the g-code window are blue.

I will remove the z-axis commands tomorrow and see if it works any better.

Is there anything that can be done about losing the arc signal?

Everything is put away for the night. I will try your post processor tomorrow. I was also ale to mdify my post processor to remove the z-axis commands. I will try that too.

I don't understand why it would lose the arc before it had finished a cut. It doesn't make sense to me. Does it happen on every cut?

No not every cut but each time it fails, the part it was cutting falls into the water pan. So I know that it actually finished the cut.

I was just wondering why you did'nt get errors on loading a GCode file.

I think I know what the issue is. My post processor saves the files with a .tap extension. It appears that the g-code parser is looking for .ngc files. Maybe it is sneaking past validation because it has a different extension?