Work with probe
Сheck just in case that all files from the archive from folder "macros" are placed in folder specified in your ini file as SUBROUTINE_PATH.
fairngs wrote: I was having trouble with accessing the subroutines in the single probe file. I have not had time to try it further but I think that problem was due to not putting .ngc on the end of the symlink names.
very nice tool!! Looks really great.
I have been using Z-probing for tool length compensation for a while. Using the probe for edge finding is very interesting to me, too, I will try it out.
I've seen in the screenshots that you are able to detect the angle of the workpiece, too? (I mean the angle along Z). Are you able to correct the angle, too? I guess no, right?
We all know, that it can be hard to properly align a workpiece on the machine (e.g. when working of the backside, or for many other purposes). Is there any possible way for semi-automated alignment of x and y AND PHI by doing a probe.
As far as I know linuxcnc uses the concept of joints and axis. So a trivial machine (with joints equal to axis) could use a rotation matrix with angle phi. Phi is 0 by default, but can be set by a probe.
Does anyone tried something like this already?
Could be even more interesting for engraving operations to probe the z-plane in a couple of points, fitting a surface and auto-correcting the plane in all three axis. And could be really really usefull when working with 4 or 5 axis.
Has someone experience with such things?
schorsch wrote: I've seen in the screenshots that you are able to detect the angle of the workpiece, too? (I mean the angle along Z). Are you able to correct the angle, too? I guess no, right?
It should be trivial. G10 L2 allows you to rotate the coordinate system around the Z axis.
I didn't know, It seems the documentation of G10 L2 is a bit "poor", only for my understanding:
G10 L2 R90
rotates the xy plane 90deg? I'm asking, because there is no unit mentioned in the doc. that means x becomes y and y becomes -x, right? And it works for any angle?
Lets assume a more complex example:
A 5axis machine, xy table, ontop is an a-axis ontop of the a-axis is a c-axis, spindle has fixed angle, but moves in z
I'm placing a workpiece in the stack, its surface is 10 deg rotated around Y-Axis (in mcs). Now I'm starting to probe the surface to align it properly (three points). From the gradient I'm detecting my highest slope (should be in the direction of the x-axis, because its tilted around y).
That means the corresponding alignment proceduce needs to turn the C-axis by 90 deg and the A-axis by 10 deg. that means that the x-axis in tcs is now the y-axis in wcs and vice versa. Please understand, this is a simple example, it needs to run for all kinds of surfaces
How can I do this with G10 L2??
schorsch wrote: How can I do this with G10 L2??
I am pretty sure that G10 L2 Rnn only knows about X and Y.
On a non-trivial machine I think that the behaviour would need to be experimentally determined.
I think the units are degrees, as is usual in G-code. But again, testing will make it clear pretty quickly.
I tried to remove what was there since it was entering in conflict with the input hook up command
#net probe-in => motion.probe-input
and place a command to hook up the input
net probe hm2_5i25.0.7i77.0.0.input-31 => motion.probe-input
When I try it, I get "must be in MDI mode to issue MDI commands"
When I touch the probe, I hear the relay triggering but don't see anything on screen confirming contact ... I am missing something ?