On 6/1/2012 7:28 AM, gene heskett wrote: > On Friday, June 01, 2012 07:20:37 AM John Thornton did opine: > >> No wonder that file takes so long it makes one pass at the OD > :) To check fixture clearances. > >> then >> rapids to a slot makes one pass then rapids to the next slot... then >> rapids to the holes then rapids to the ID. And wow is it complicated and >> full of neat code. It would make much more sense from a machining view >> point to cut each slot fully then move to the next slot then change >> tools > This spindle is #2 morse collets, loosen drawbar about 2 turns, beat it > loose with a piece of steel. Hard on bearings. The only fixed length is > the rings on the drills, so my tool changing z accuracy if I change the > collet to fit the bigger mill, really needs my autoz code. Unforch, without > making a melamine jig to insulate the whole thing, that's out. It uses > G38.2 for that. This way its all done with one z home& one bit if I don't > break it... Yea, for multiple tools have multiple G code files, set Z run one to completion, change tools set Z ... >> and drill the mounting holes and with a larger end mill cut out >> the center and OD of the part. The beauty of Lawrence's file is that it >> is completely configurable. > Almost, I've added about 6 or 8 more vars so it is considerable more > adjustable. Lawrence's original code is also available on the net. If you > need that for comparison I'll dig up the bookmark. That's ok your rendition was interesting enough to look at... > >> How in the world do you hold the material to machine ID, OD, slots, and >> holes at the same time? > Clamps on each of the 4 corners. Because this stuff is so thin, I've > considered making a jig fastened to the table, then using one of the > leftovers in 1/16" brass as an overlaying holddown with its own bolts to > hold the stuff flat while machining. That could also serve as a dam to > hold cutting oil in place a bit better, as it is, it creeps away and has to > be replenished about 2x a loop thru the main loop. If I see debris stuff > on the bit when it goes to safeZ, its time for more oil. I have made a > mister that could keep the swarf blown away, but even with the shop door > open, it turns into a glasses fogging fog of oil in there in about 30 > minutes. And my lungs don't need that safflower oil in them either. > > This between the clamps warpage is why I have to start about 15 thou high, > and run to about 10 thou into the 1/4" oak sacrificial its clamped down to. > It wasn't near the problem with 1/16" brass, but with a #60 drill it would > be 2-3 days to carve because the brass tends to work harden. I've already > broke about $50 in these bits, but if I don't push them, they'll cut with a > 0.0025" touch forever in alu if I keep an oil film on the work to seal out > the oxygen. A higher speed spindle would help, this one is tapped out at > 2500. Right now the 'touch' is about 0.0032, so I've dropped the feed > override to about 75% where I don't see the bit flexing as much. Time I > have, accuracy is whats needed. I only have to do it once IF I get it > right. Keyword=right... > >> If I had to make that part and I might one day, I'd drill the holes then >> mill the ID. Then I'd prepare a mounting plate with a boss the size of >> the ID then bolt the disk down and mill the slots and the OD. > I've considered that. If I was to make them for sale I'd at least make a > better holddown jig. In my case, a mini lathe, the mounting holes could be > dispensed with as the finished disk is trapped between the spindles preload > adjusting nut, and an identical locking nut. So the concentricity demands > means that center hole is -0.000 to +0.002 max. It's about right when I > have to fuss a bit to get it over the threads and seated against the > preload but. At the diameter of that bolt circle the bolts would interface > with the locknuts. So if to actually be used, the bolt circle would need > enlarged but the limit there is when the bolts or rivets hit the center, > index pulse generating interrupter. There isn't a lot of room to play in > that area. :) > >> By "duty cycle" does that mean the width of the slot verses the width >> between the slots? > Yes, exactly, in order to get a 50% duty square wave. Lawrence's original > cut at 50% with a /2.0 there, no comp for opto aperture. Besides, mulls > are quite a bit faster then divides for the computer. My last attempt was > at .45, and it was about 60/40 high/low since the opto's don't have an > infinitely small aperture. Lawrence's original code also tapered the index > slot, but the placement of it at the same fixed width made the inner end of > the AB slot wider at slot 0, big bump in encoder velocity output. Spindle > goes whump whump as that slot went by the opto's out of time. Not good. > Listen to Z going crazy when running a G76, sounds like its being run by > pink noise! Pink Floyd noise! The subroutine I cobbled up could be done in the same way with all the variables. The radius of the arcs at the end of the slot is so big that a G1 move would be the same really and a lot less complex to write the code for.
John ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
