I've been working on some probing routines in Mach3, and have been having issues with the probe plowing into and through the objects being probed. I noted that despite having an error check to validate probe is not 'pre-triggered' before issuing a G31, if the starting clearance was small (Say between .005" to .0015") the UC300 would not stop when it hit the object, it would just keep plowing ahead.
As a method of trouble shooting, I created a pair of scrips that did the same thing in Mach3 (3.043.062, uc300 plugin v1.015)) and in UCCNC (2106). Essentialy the routine probes the part, retracting a starting clearance value, probes again, then retracting to a smaller and smaller value iteratively until the probing either issued a 'probe active' error or crashed though the part.
In Mach3, the probe would consistently crash 'thorugh' the part if the clearance between part and tip was between .010" or .015" as the G31 is issued. As the feedrate increased, the clearance at which a failure occurred also increased a bit. Going to a very slow feed of 0.5 in/min also still results in a 'crash' This leads me to believe that when the UC300 initiates the G31, it is ignoring the probe signal during the acceleration portion of the move, or during some amount of time. I can also manually position the probe within a small distance, issue a G31 by MDI, and have it plow through the part.
UCCNC behaved perfect and as expected, it would iteratively probe until the tip was in contact with the part, then stop and issue the warning that the probe was triggered. Exactly what I wish Mach3/UC300 was doing.
Based on the behavior, I'm guessing it is possibly in either the firmware for the UC300 (when running for Mach3), or in the plugin driver. This machine used to run via parallel port, and is recently converted to UC300, so I may attempt converting back to parallel and repeat the test to verify it isn't purely a Mach3 hiccup.
I'm attaching for reference the scripts I used for this quick testing in UCCNC and Mach3, if anyone cares to help verify where the problem lies in Mach3/UC300 coordination. (Note: The scripts are written for 'inch' setup, speeds and clearances are all set as variables that should be changed for metric. They are sloppily written, so test carefully ).
Anyone else had issues like this that they've confirmed? Ideally, I'd like convert all my probing to be done in UCCNC and escape Mach3, but I'm still wishing for being able to G31 probe in combined X/Y/Z moves with UCCNC (pretty please )
regards,
Eric Brust
Mach 3 scripts to test in Z and X:
UCCNC scripts to test in Z and X: