On Wednesday 19 December 2018 15:02:57 Gene Heskett wrote: > On Wednesday 19 December 2018 01:27:46 David Berndt wrote: > > I could write a big email about this, and I will if you want to, > > just let me know. But I'll jump in here and agree with Chris. You're > > missing out not having a cad/cam package. You're not able to use > > your tools, or your time to optimum effect, and you're not getting > > the best end result that you could possibly get in most cases. > > Modern CAM software outputs some awesome tool paths and is worth > > having/using. Especially for a hobbyist imo. > > > > > > Regards, > > Dave > > I guess I'm a stubborn old cuss. Right now a simulated run looks > great, and its about 280 active lines of code if the comments are > skipped. You saw the screen snap I posted yesterday. Can your cad/cam > program do that in 300 LOC? 100k LOC maybe. > > One question remains: I'd like to drill holes for connectors with a > 4mm dia tool. All are under .625" in dia. The cutter comp section > doesn't mention the g42 enable blending moves in anything but the xy > plane, so my question is: > > Can I go to the center of a hole with a 4mm tool a half inch above the > work, enable the g42. and descend to the work on the z axis while the > comp is being phased in? If I can do that, my backlplot visualization > would be considerable more accurate. Right now I'm doing cutter comp > in the code, but that means the trace is tool_dia undersized visually. > Those moves then could be incorporated into the SUB at the expense of > 5 seconds of cutting air above the actual cutout, and only have to be > written once. I don't see any reason why it couldn't work, and the > radius of these holes isn't long enough to do the "blending" by a > loooong ways in the xy plane, and being able to do it with a z move > would be a huge help. Seeing as how its freezing out, and nearly so in > the shop where this machine is, I may as well find out. And I guess > the answer is no, no amount of Z is enough. So the calls will have to > pass the x & y's too and the SUB will need to do more math. > > Next question, will it work at all in G91? No, don't even try. > So all my calls to drill holes are now absolute to the map in use. > So the drill a hole call now looks like this: > > o130 SUB > G40 > M3 > F15 (feed rate) > (arrives here at center of hole, change this to take a passed argument > for radius) > G0 Z#<_ZUP> ( clear work) > G0 X[#2 - 0.3] Y[#3 - 0.3] (move away for G42 blend in) > G42 D15 X[#1 + #2] y#3 > G0 Z0 > G3 F#4 I-#1 J-0.000 Z#<_ZDN> P9 > f15 > G0 Z#<_ZUP> (back to clear) > G40 > M5 > > o130 ENDSUB > > The extra, cut air moves, to switch the G42 in and out add around 15 > or 16 minutes to the run time. This routine gets called 19 times. > > Its expecting 4 arguments, the last one being a feed rate so I can > turn it down nor nearly straight drilling, the little holes that are > almost a straight drill can push pretty hard, and there no support > under the workpiece except where its clamped at the ends. Plus theres > now 3 M0's in it, to move clamps at one end or the other, and 2 more > M0's to allow clamps to be loosened and the work re-aligned in case I > have to take it off the table. So yes, it will need baby sitting but I > don't see anybody here except me to do it. > > The idea is, I put a workpiece a half inch bigger each way than I need > on the table, and when its done, I take a complete back panel ready > for parts off the table, needing only a bit of deburring if the tool > gets dull. But it will be running up to its kneecaps in cutting oil > soaked swarf. I moved the ac cord from the too flexible, middle of the > panel to the extreme left end as that puts an S curve into that heavy > cord, absorbing any panel flexibility by about 99.9 % before it gets > to the terminals of a 12 volt 1.25 amp supply that runs it all. > > Theres 116+ inches of rapids, and 395+" of cutting, on a panel that's > about 11.875 by 3.125. 320 LOC according to file->property's. I doubt > if your cad/cam's gcode output can touch that. If it can't write gcode > a heck of a lot more efficiently than pcb-2-gcode for eagle, it can't > play in the same ballpark. I don't have an eagle sourced file thats > not something in the 100k territory for a 1.6" x 2.2" board. > > [...] > > And I see its up to 53F, so I might hit the shop and try it if my > right leg will get me up the hill. Except I haven't been out to get my > lady her daily dose of crosswords from the local fishwrap yet. Some > things are more important than others. See rule 2 about the boss is > always right. :) So now I have a leg with a pulled muscle, I woke up > around 7ish with a leg cramp, so I've not been getting around the > house too well all day. Enough to take care of my lady, but not much > else. > > I don't recommend this business of getting old, its nothing like AARP > claims it is. I could cheerfully throttle the guy who claims these are > the Golden Years. Or he could die a long lingering death from COPD > like my wife is doing. > > Thanks David, and have a Merry Christmas. I'm going to go get my lady > her cross-words, before she uses them on me. :)
Well I got that, then at the end I have a trip around the panel to cut it out. But I got to the end, using f42 and discovered my climb cut was cutting the panel dimension so the intended cut size was the inside edge of the cutout, not the intended outside. So I reversed it to a g41, but now its cutting the horizontal lines downward by a few thou on the right in almost a foot of run, so obviously I don't have something coded correctly. I am limited on table to right because of the limit switch, so I can't start the G41 from the outside in the x direction. Humm I thought I had arrived at the starting x.5 y.5 air offset in the g40 mode, but a g40 before that .5 inch inside run, then the g41 fixed it. Damn it would be nice if the mdi status line box actually followed the state the interpretor was in as the code was executed, but it doesn't follow either a map change or the g40-41-42 status. On checking the codes flow , the last thing it does is drill the mounting holes in the ends of the panel. It appears the sub that drills a hole does Not issue a move to leave the g42 state after a canceling G40, so I'll have to add that move to the subroutine, but it is not calling that an error, and the next move is more than enough for an exit blending move. So I reversed the outline cut sequence, and used a g41 so its still using a climb cut. I think now its actually done. One panel per completed gcode run and all I need to do is switch the connector hole size from 16mm to 12mm for the second panel. I have a software switch for that. The only thing I might adjust yet is the locations of the db25's once I've made the first one and can test fit it. Go back to bed and get the rest of what sometimes passes for enough sleep. Merry Christmas everyone. -- Cheers, Gene Heskett -- "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author) Genes Web page <http://geneslinuxbox.net:6309/gene> _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
