Isel Post Processor

[hmatyas]

I am trying to help a friend with an old CNC machine called Isel. It use to use MasterCam.
In Windows 10, I have a program (forgot the name) where I can home and jog the machine.
When it comes to loading the gcode it wants a NCD extension. I wrote a simple program to make a
small circle Saved it as a Techno Isel post processor).
It looked good until the router crashed into the table. (It actually stopped on the table).
I then looked at one of the old gcode files. In the file is referenced a MCX file and a NC file. I believe they
are both MasterCam files one I think references the size of the bit being used, I don't know what the other file is for.


I have a gcode file that works (attached)

Is this thing just to old to work with new software (Vcarve or Aspire)? He does not want to spend a lot of money to convert
the controller to something new. Any ideas?

Thank you

(NO good dead goes unpunished)

[LittleGreyMan]

Hi,
Could you post the gcode that crashed?
The age of the machine vs age of the software is not an issue. A CAM software internally computes the path in its own "language" then translates it into the machine language according to the postprocessor you choose.

[hmatyas]

Sorry that it took so long. I deleted the files off my flash drive. However, I created basically the same file.
What is weird is that I thought maybe the z height or something like that was the problem. I don't understand why
the router went straight down to the table. I could see it going up because of safe z height. If you need the program that
I simulated the gcode I'll have to call to get the name of it. The simulation works fine, the actual action is the problem.

[LittleGreyMan]

The result of the test file in a gcode viewer looks odd. Could you post the Vcarve file please?

What is the exact full name of the Isel PP you choose?

The working gcode shows that the machine controller accepts circular interpolation, which is not the case with the PP you selected.

What is the model of the machine? I have an Isel PP that works with a CPM 2018 with an IMC-4 controller.

[adze_cnc]

Both test files look reasonable in this online NC Viewer.
 

 

 
What interests me is the "o0000" line in the test.nc file. I wonder what it does?

[LittleGreyMan]

Steven,

The plunge in the center of the cylindric cut seems strange, that's why I asked for the Vectric file.

Same question for the "o0000" line.

[Adrian]

o0000 is a memory storage command. 0000 means it's not stored whereas a number would mean that's the location it's stored in memory of the controller.

[hmatyas]

Attached is a picture of the nameplate Isel Model II EP 1090
The crv file may not be exactly what the gcode that I sent.
I did not save the file but recreated it as best that I could.
All it is is a square with a hole in it. That should not be the problem.
techno gcode interface ver 3.02 gcode interface for Devavici c series

[hmatyas]

I used the Techno Isel ATC (metric or imperial) used both at one time or another.

[LittleGreyMan]

OK, I'll have a look to-morrow.

[hmatyas]

I'm getting closer for it to work. I believe right now it has to do with the limit switches and maybe the safe z height.
We worked on it for 3 hours. We started by using a file created about 3 years ago that we had the code from
an old computer. That worked on the new computer and new version of control software. We got as close as getting an error, saying okay and then the thing "air cut". However. the x and y coordinates were no longer correct. So it started and where it should have cut, we got the error. When we hit okay, it moved to a new location and ran all of the code. When it went to the home position, it was off by the distance of where it started.

[adze_cnc]

Where is the CircleC.ncd file from? It hasn't been generated from the TestIsel.crv3d file. It's not related to the test.ncd file posted earlier.

Is this a file that did or did not work?

I say this because the test.ncd file and TestIsel.crv3d files are ones where the z-zero is set to the machine bed. The CircleC.ncd was made with the z-zero set to the material surface.

If the machine is set to z-zero being at the machine bed and you ran the CircleC.ncd file then it's going to try and cut below the bed level.

The PDF file shows an ncd file called CIRCLE5102023.NCD. Why not post that here instead of CircleC.ncd?

[hmatyas]

I'm sorry, I must have posted the wrong file. Yesterday, we worked on the machine for 3 hours and I created different versions with different toolpaths. It was my mistake. We were working with 3 different laptops, his old one, his new one and mine where I created the file and toolpaths.

[LittleGreyMan]

Sorry, I couldn't take time to have a closer look.

But reading the last posts, I'll wait having reliable and coherent information. Please understand it's not possible to help if the data you provide is not correct.

[hmatyas]

I understand, but there were so many unknows we were going through. However, we now know that the cable works with the new computer. For the most part, it is reading the gcode. I know that the controller is the Davinci controller. I found a drop down in the configuration menu. The Davinci works, the Wizard reverse the x and the y axis. The other 2 versions did not work at all. The machine would say an error, hit okay, the machine would move to a new location and start to run. It would then home, but it would end up in a spot that was not the home position.