#linuxcnc-devel | Logs for 2014-05-17

[11:45:08] <cradek> ugh, got a bunch of emc-users unsubscribes all at once again
[11:45:10] <cradek> wonder what happened
[11:47:03] <cradek> bounce processing is still off (and sf hasn't done anything at all about the dmarc problem)
[12:15:20] <kwallace> Hello, I'm planning on adding a D flag to G76 for a thread taper feature. Fanuc seems to use R = (start_dia - end_dia) / 2. I'm wondering if I should use the same or use a slope or angle instead?
[12:18:33] <archivist> a lot of taper threads are specified as a diameter ratio like 1 : 16
[12:23:15] <archivist> this was something I was looking at a few years ago but life got in the way :(
[12:24:35] <cradek> the interaction of L and (your proposed) D makes my head hurt
[12:25:31] <cradek> I think having L was a mistake
[12:28:39] <cradek> NPT is 3/4 inch per foot diameter change
[12:28:49] <cradek> I guess that's 1:16 as archivist says
[12:29:27] <archivist> BSP and NPT are the same taper as it happens
[12:29:32] <kwallace> I was thinking of applying D over the length of the drive line, so L would not matter?
[12:29:40] <cradek> so you'd use D0.0625 (if in G7) or D0.03125 (if in G8)?
[12:30:01] <cradek> so disallow L if using D? that sounds utterly reasonable to me
[12:32:34] <kwallace> I think the path between the end tapers would follow a model path that had no tapers.
[12:33:10] <kwallace> What are the end tapers for?
[12:34:31] <archivist> tool in/out
[12:35:07] <archivist> no need for safety groove
[12:35:09] <cradek> well for one thing, you can cut a thread that's not at the end of something
[12:36:32] <cradek> imagine you want to make an adapter in one chucking - you can cut both threads (use an entry taper for the left one), then part off in the middle of the left one
[12:36:47] <cradek> I bet it's not used much
[12:37:06] <archivist> I use the exit taper all the time :)
[12:37:12] <kwallace> An entry in X only could still work, or I would tend to cut a channel before threading.
[12:37:15] <cradek> I bet it's used all the time
[12:37:58] <cradek> yes you could cut a channel, but then you have really limited space to get accelerated and sync up
[12:38:53] <kwallace> Ahh, yes.
[12:41:55] <cradek> if you're using D, you'll have to check for and forbid G41 Dx G76 ... but I think that's fine
[12:42:24] <cradek> I bet people don't use cutter comp with G76 anyway
[12:44:45] <kwallace> Yeah, I found this in interp_check: " if (block->d_flag) {
[12:44:45] <kwallace> CHKS(((block->g_modes[7]... " and thought it might need attention.
[12:45:01] <cradek> yep bet so
[12:46:49] <kwallace> Oops, that just checks if D was in a block and was tied to something appropriate.
[12:52:04] <kwallace> I suppose after getting D in and working, it's just as easy to make it work as an angle, slope or delta X.
[12:53:42] <cradek> that's true
[12:53:57] <cradek> delta X seems the most fiddly for the gcode programmer, to me
[12:54:14] <cradek> if you decide you need one more thread (more Z) you have to recalculate it and adjust it
[12:54:20] <cradek> same if you decide you need some more lead-in
[12:54:42] <cradek> for one particular taper, the delta X would be different for every thread you do
[12:54:48] <cradek> that seems pretty silly
[12:55:09] <cradek> to me, one of the schemes that lets you specify the taper directly would be better
[12:56:43] <cradek> on the other hand, I've only ever cut NPTs with dies, so what do I know
[12:57:05] <cradek> time for coffee break
[13:00:18] <kwallace> Well Mach seems to use a degree value, so that must be right. A Fanuc example looks like a pain: http://www.helmancnc.com/tapered-threading-with-fanuc-g76-threading-cycle/
[13:49:30] <archivist> fanuc R does look a bit silly
[13:51:05] <archivist> I usually use my manual lathe and set the angle or x:12 has taper turning attachment