I'll need to study your file closer,including your post processor. Are you sure this is a Path G1 move to start-depth rather than the beginning of a canned cycle?
No, looking at the unexpanded Gcode before the post processor ( inspect g-code ) the canned cycle starts from clearance on the first hole, there is no move down to safe ht, first. So this is not related to any of the post proc code.
Code: Select all
(bar-fix-Drilling)
(Begin Drilling)
G0 F50.000000 Z21.000000
G90
G99
G83 F1.000000 Q3.000000 R-0.500000 X170.000000 Y-15.000000 Z-20.000000
G80
G0 Z3.000000
G83 F1.000000 Q3.000000 R-0.500000 X320.000000 Y-15.000000 Z-20.000000
So if there was a G0 from clearance to safe ht, before G83 a lot of this would disappear. Moving up to clearance between each hole would not be a problem. This is nothing to do with canned cycles since if I deselect Peck, the G81 does the same : much drilling of air. This is a defect I've noticed before but never got around to bringing it up.
If you need to put clamps in an area that Path would otherwise expect to be clear, what else would you suggest? How is Path expected to know that you have a clamp in the working envelope?
There may be a case for this ( though I remain to be convinced ) but my point here was that the doc is incorrect. If this is the thinking behind it, the concept of the "envelope" should be a caveat to the current statement this space is "safe".
Good question. Maybe someone with experience in other cam software can weigh in. Maybe we need to create another kind of object or dressup to indicate a no-go or clamping area but implementing it is going to be a nightmare.
Assuming that the user can tell what route FC will take and anticipate the envelope is not guaranteed, it not always that obvious. I already have a few lumps missing from my clamps where I got caught out by this unsafe "safe space".

One thing I did intend to raise at some point is ordering of "base geometry" being taken into account.
Yes, if someone with trade experience could provide input on existing conventions it would be helpful.
I would have thought where there are separate volumes to be removed , as in this case, retraction to clearance would be preferable. I have not thought it through very much for other tool paths but in principal I would think the same would apply. Pocket and helix are pretty directly analogous, maybe you can come up with a geometry where this does not make sense.
I would think a dressup similar to holding tags could be feasible, if needed. I wasn't sure that there was not already a facility for clamps which I had not discovered yet. I'd seen "fixations" appearing the Gcode comments.