So. I’m trying to machine a large aluminum part on my Nomad (many hours) and worried about it crashing and having to start from scratch.
After much searching, there doesn’t appear to be a good way to resume from where you left off.
After some experimenting however I’ve come to an acceptable workaround.
What I was doing initially was coming back every 15 minutes or so to check the current gcode line in case I needed to restart, my plan was to edit the gcode manually in a text editor to the last line I recorded to avoid having to “air machine” and waste time. After about 45 minutes of this I was like “eff this!” and after a bit of searching and found an app that lets you do timed screen captures. (https://sourceforge.net/projects/autoscreen/ I have zero affiliation with this app, just googled it).
So no, I have my machine running and the app taking screenshots every 10 minutes in case it crashes. If/when it does, I’ll edit the gcode to the right spot and restart the workflow.
Thought it might help someone in my position! Might be interesting if Carbide Motion could had a function to “record last known line” or something…
What sort of ‘crash’ are you experiencing - software or the endmill ramming somewhere it shouldn’t - or is this just potential risk mitigation (healthy paranoia is understandable:) )?
I’ve a similar setup and run many hour jobs but as yet haven’t had any sort of software crash.
It is mostly a “healthy paranoia” thing as you put it but my main concerns are.
-) Chips/heat/wear over a longer run causing the endmill to ram/bind up/stall out.
-) My computer timing out somehow and losing the connection to the machine.
-) Some unforeseen issue.
Understood. The second issue is probably one that one could collect anecdotal evidence for using this forum, since there’s plenty of people here running very long jobs. My data sample would be “not a problem”.
Biggest problem I’ve had is an endmill breaking during the job. The photo approach of the latest g-code instruction wouldn’t work in this case since it might have been successfully routing air for 30 minutes.
I can see that it would work in those cases where the computer failed rather than the machine or routing.
My much slacker approach for restarting jobs on failure is to look at the part itself and then, in CC, disable the toolpaths that appear to have been done already and generate a new nc file… possibly adjusting the start depth for the last pass that was running.