I just can not get it to work correctly

[skyguynca]

I have tried it on two of my machines. One is vista 64 and the other vista 32.
I load the gcode and zero .03 in above the board, and it creates the log file thru the probing, I reload autoleveller and create the modified gcode, load in mach3 and hit start, it zeros its reference, I remove the clips and hit start, it runs the code and it way too deep in some parts and way to high in others. Here are my files.

[skyguynca]

here are pictures of how it is cutting. I stopped it before it ruins the entire pcb blank

[daedelus]

I think this could be for a number of reasons. A possibility could be an inaccuracy in the probing values retrieved, again the reasons for this could be numerous, could be something to do with anti-backlash nuts on your axis or some setting in Mach. I remember a problem a while ago that affected a few individuals where AL didnt like the probe values when fed back to AL because some values on the X and Y axis were slightly off. I adjusted AL to compensate but it would have been better if the machine had been 'tightened up' rather than the software working around the problem.

I did notice a few of these discrepancies in your log file but I'm not sure if its enough to affect anything. From your log file:

Code: Select all

4.92941,0.44475,0.00461 
4.54935,0.44477,0.00373 
4.17032,0.44475,0.00302 
3.79118,0.44477,0.00356 
3.41198,0.44477,-0.00143 
3.03290,0.44475,-0.00126 
2.65381,0.44477,0.00143 
2.27473,0.44475,-0.00121 
1.89559,0.44477,-0.00027 
1.51645,0.44475,-0.00077 
1.13736,0.44477,-0.00143 
0.75817,0.44475,0.00005 
0.37914,0.44477,0.00164 
-0.00000,0.44475,0.00488 

A couple of Y values are 0.44475 wheres they should be exactly the same as the others (0.44477). X and z could be worse, not sure. Again, dont know if these minor differences are enough to cause a problem.

Another thing which you could try is to increase the cut depth. I noticed your cut depth was set to .002", maybe try .004" and see if that helps.

If anyone has any other ideas, please chip in.

[skyguynca]

my backlash on all axis is less than .0001 checked.
I noticed while probing sometimes the x axis would move, don't know where that command is coming from. it would move maybe .001 to .002 occasionally. really weird and I thought it came from your software rewriting the gcode to do the probing.

I am going to add a capacitor to the probing leads to prevent false signals and prevent bounce, but other than that I don't have a clue either. This is the eighth machine I have build for cnc work. My other two big mills and lathes I still use are acurate withing .0005 when cutting metal, this one tested milling out objects is within .0002 so and I have not encounter the random x move when cutting other projects, just when probing with the autoleveller.

[daedelus]

Very weird as you say, X should not budge during probing at all. Has anyone else experienced this?

As a suggestion, try generating a 'probe only' file in mm not inches and try that. It should not make a difference to the X axis but I would try anyway just to rule it out.

Maybe Mach rounds up some values when inches are used which might account for the X axis movement. Or maybe I am clutching at straws.

[skyguynca]

Ok, I will give it a try and get back to you.