On 01/26/2012 05:52 AM, gene heskett wrote: > Which is why it seems more better an idea to grab what is in #5063 > after the G38.2, and apply enough math to arrive at where you want it > to think it is, and do a G92 Zmath-result. To keep track of where the > machine needs to be, back at its own referenced home position, do a > G92.1 to clear the offset before going back to the tool change > position above the gage, for the next tool change. The way pcb-gcode > issues those commands is consistent, so a search for the M06 T, than > back up 2 moves and do the G92.1 Then after the M6 T# insert "call > <z_cal.ngc>", which will do the G38.2, process the result and apply > diff as G92 Z<result> The .drill files have 5 such change places each.
I'm a visual person, I need to "SEE" something to understand what's needed. Looking at the above, gives me a headache. BUT, I still think that creating a script to adjust the g-code would be easier. >> I have a hard enough time with EMC2/LinuxCNC, as it is, For me at >> least, the only other option would be to edit PCB-GCode, but I'm not >> well versed in ULP files. >> >> Due to moving my shop, and having my CNC mill sitting on the floor in >> pieces due to having worn out leadscrews, I'm not set up to test >> PCB-Gcode. > Ouch. Bummer. That day will come for my little toy too. The XY nuts are > such a kludge, I should have spares on the shelf for when I break one > trying to take up the backlash. :( Mine's a Speedway series Mill/Drill. Cast iron nuts, on steel leadscrews works well in manual mode, but, running them back, and forth, on my hand coded g-code, for the last 2 years, wore the leadscrews out. Been seriously considering ballscrews, but Disability doesn't pay a third of what I used to make. I also have the spindle apart, to put new bearings in, and to redesign the convoluted way of moving it. Waaaaayyy too much slop, the pinion would occasionally jump a tooth on the rack. > >>> An old friend and engineer back in about 1960 was fond of the phrase >>> 'simplicate' and I'd think this qualifies. :) >>> >>> Cheers, Gene >> For me, this would be simple... >> >> >> DISCLAIMER: I've been using Linux since '94 (Early SLS version), and >> I can write (and have written) BASH scripts in my sleep. They might not >> be the most efficient scripts, but they work for me. I've been doing >> this for so long, I do most things without really thinking about HOW to >> do them. > There was a time when I dreamed in 6809 assembly. So I know well how that > works. :) Like you, now I write system daemons in bash. Except for its > lack of floating point operations, its a pretty good OS (but don't tell > that to an emacs fan) :) BASH not directly handling floating point, is IMO it's only major drawback. I still have a few old bash scripts on my system, that are over 100,000 lines long, and I use that same perl script throughout them for math calculations. They are Linux installers, that will completely create a custom OS, from source, calculate the time that it took to install each program, as well as the percentage of time it took compared to the first program that was created, and give me feedback, with fancy colored output. All this from a bare command prompt, no X required. I had been using Linux for 2 years, before I even looked at installing X. > > Cheers, Gene -- -Mark Ne M'oubliez ---Family Motto Hope for the best, plan for the worst ---Personal Motto ------------------------------------------------------------------------------ Keep Your Developer Skills Current with LearnDevNow! The most comprehensive online learning library for Microsoft developers is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3, Metro Style Apps, more. Free future releases when you subscribe now! http://p.sf.net/sfu/learndevnow-d2d _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
