Fadal Retrofit - ATC cycle

If you start linuxcnc from the command line you should see the Python errors in the terminal (command "linuxcnc" )

I tend to use "linuxcnc -l &" to open the last config and return to an interactive shell.

Then you can enter the Python command line, and see what you have.

My suspicion is that HAL_BIT pins are not boolean, so "is True" and "is False" will fail, these are likely to be integer values 0 and 1. Try "==" instead of "is"
(I am not at a LinuxCNC machine at the moment so can't test this theory)

For some strange reason, some basic python commands were not executing, for example; "sys.exit()"

Do you see anything in the terminal output?

It gives me a "sys" is not defined. But in regards to the hall sensor line, it does not give any errors.

Just to provide another avenue, the auto tool change in question was ultimately achieved by having the sequence run via G code but having a subset of user defined M codes which compartmentalized the whole auto tool change sequence. Within these was also the usual "check for carousel in extended position" etc fault checking.

Then these were all brought together (again via G code) and made into the full M6 sequence (the M6 itself being re-defined to run the complete tool change G code).

It was felt that there would be advantages in having compartmentalized M codes for the individual segments of the procedure in case there were any crashes during the tool change sequence so that one could call a certain M code and carry out any rectifying actions in a step by step way.

For example, I left the probe in the spindle when I executed a tool change to get the tool I needed. Not the brightest thing to do, and because I didn't want to damage to probe I hit the E Stop. So, the carousel is sitting half way out and the whole machine is somewhere in the middle of a tool change. Having the ability to call smaller subsets of the tool change sequence via user defined M codes made it easy to retrieve the probe then send the carousel back to home. In other words, sort out the error and place the machine back into a known state.

Not knowing well the way classic ladder or python integrates with LCNC, can segments of code be run outside the entirety of the program? In other words with full and complete auto tool change routine written in classic ladder, would it still be able to send the carousel back to home position using the above example? Or can it only be run linearly from start to finish?

There are a million ways to skin a cat, one way no better or worse than the other. But having seen things like this first hand, I'm glad I have the facility to call smaller M codes to fix the rare occasion of an incomplete tool change.

Nitram, could you send your linuxcnc files?

COlger81 wrote: It gives me a "sys" is not defined. But in regards to the hall sensor line, it does not give any errors.

I am pretty sure that the Python script will give up entirely at that point.

Do you see the HAL pins appear?

Pretty sure you need to add at the beginning of your file. I certainly had to on Windows as the output below shows
I'm still very much a noob with python

COlger81 wrote: For some strange reason, some basic python commands were not executing, for example; "sys.exit()"

Python does not randomly not execute code. If some bit of code is not executed then it is not being called.

JT

COlger81 wrote: I'm going through a retrofit of an AC Fadal 3016L. I have the mill running perfect with manual tool changes. I just started writing code for the non servo ATC in python. Is python the right method for the ATC or should i go with a different method?

You would get better exposure if you started your own thread in a more appropriate section like Advance Configuration. Most folks that come in this section are looking for Classicladder help or to answer Classicladder questions not Python.

JT