#emc-devel | Logs for 2008-02-10

[01:56:06] <SWPadnos> Steve Blackmore posted an interesting use case for jogging while a program is "running" (as he does with Mach)
[01:56:53] <SWPadnos> consider a rool breaks - he basically does feedhold, jogs around a bit so he can replace the insert, re-checks the offset, then moves back and releases feedhold
[01:57:12] <SWPadnos> I think that can be done with stop/set-run-line/resume in EMC2
[01:57:43] <SWPadnos> but it does sound more "natural" to just hit pause when you see the problem occur (to him anyway, I think I'd go for the big red button)
[02:13:57] <cradek> I haven't seen this post yet I guess
[02:14:10] <cradek> but yeah, you could do that with run-from-line if you're careful
[02:35:38] <SWPadnos> it isn't on an EMC-related board, it's on the CCED list
[02:36:34] <SWPadnos> I had mentioned something about using hal_manualtoolchange, and also responded that you can't do machine motion while a program is running (and a toolchange is still a running program)
[02:39:06] <cradek> lots of people would like to jog and set offsets after a tool change
[02:43:34] <SWPadnos> indeed - that's why I mentioned it :)
[02:44:09] <cradek> supporting all the ways people want to change tools is hard
[02:44:20] <SWPadnos> that's certainly true
[02:44:33] <SWPadnos> and I definitely wouldn't want to make feedhold ignore jog commands
[02:44:39] <SWPadnos> (or vise versa)
[02:53:55] <CIA-22> EMC: 03cmorley 07cl_v7124_branch * 10emc2/src/hal/classicladder/files.c: fix compiler error
[02:58:41] <cradek> I wonder if there's still the problem with run-from-line vs spindle-on/off
[02:59:14] <cradek> if so it's kind of useless
[02:59:22] <cradek> I should test that sometime
[02:59:36] <SWPadnos> hmmm. yeah.
[02:59:41] <cradek> if you could turn spindle on in manual mode and have it stay on, that owuld be fine
[02:59:56] <SWPadnos> I had remembered discussions about run-from-line, so I never responded to the CCED list that that would be the "preferred" way of doing it
[03:00:01] <SWPadnos> yep
[03:00:40] <cradek> I think when I have problems like that I just edit the gcode program
[03:00:45] <SWPadnos> though even with spindle working, the "right" thing to do is to move somewhere close to the start point of the screwed up move, then start from the line before that one
[03:00:47] <SWPadnos> heh
[03:01:00] <SWPadnos> with cam-generated G-code, that's less desirable
[03:01:02] <cradek> but some people consider the gcode as sacred
[03:01:27] <SWPadnos> the thread on CCED was about CAM and other stuff, so I'd assume large not-necessarily-readable programs
[03:01:46] <SWPadnos> in a production shop, the g-code is sacred
[03:01:58] <cradek> sure
[03:02:07] <SWPadnos> if a tool breaks and you change the program in the course of recovering, all the other parts are screwed
[03:02:27] <SWPadnos> a/are/will be/
[03:05:39] <cradek> I still think tool length probing is the best way for nonrepeatable lengths
[03:06:20] <cradek> the jog-and-offset way, if you ever machine away the top of your workpiece you're screwed
[03:06:33] <SWPadnos> heh - that's true
[03:07:01] <SWPadnos> is there a way to get the measurement to happen as part of the M6 function?
[03:07:05] <SWPadnos> * SWPadnos doesn't think so
[03:07:16] <cradek> nope
[03:08:13] <SWPadnos> bummer
[03:16:52] <jmkasunich> whats the difference between tool length probing, and touching off on the table?
[03:17:10] <jmkasunich> (vs. touching off on the top of the workpiece, which might get machined away)
[03:17:18] <SWPadnos> not much I think
[03:17:34] <jmkasunich> except that probing needs a probe
[03:17:40] <SWPadnos> they all require motion during/after the toolchange
[03:17:54] <CIA-22> EMC: 03cmorley 07cl_v7124_branch * 10emc2/src/hal/classicladder/Submakefile: add define for modbus i/o
[03:18:37] <jmkasunich> I think the idea behind the probing is that it can be done by the program, and does NOT require manual actions in mid-program
[03:18:58] <SWPadnos> right, it could be automated, whereas any touch-off can't be
[03:20:31] <jmkasunich> for one program I did touch off by having the program prompt "insert drill high in the chuck", then after you hit continue, it moves to a spot above the work, and prompts "loosen chuck, drop drill to touch work, tighten chuck"
[03:21:00] <SWPadnos> ok - that's cheating for Z motion ;)
[05:29:01] <cradek> on the BP for drilling I set the tool change position near, but outside, the edge of the work (EMC needs flexible tool change position like this!) then the first move after tool change is to Z0.5 and then M0 pause. I set a .5 dowel pin on the work and move the knee so the drill touches it, and then continue.
[05:31:28] <cradek> that means I have to always do drilling last since my real tool lengths are lost
[05:31:39] <cradek> but that's been OK so far I guess
[05:32:04] <cradek> huh, it's 11:30 and I'm talking to myself
[05:32:14] <cradek> this is what I get when I nap in the middle of the day
[05:42:09] <jmkasunich> if its any consolation, I'm still up
[05:52:24] <cradek> how's your e-week project coming?
[05:56:34] <jmkasunich> slowly
[05:56:44] <cradek> how much time is left?
[05:56:45] <jmkasunich> I fart around too much - didn't really get started until after 9pm tonight
[05:56:53] <jmkasunich> the 22nd is the contest day
[05:57:05] <jmkasunich> but this is the last weekend, since I'll be in wicheta next weekend
[09:10:42] <CIA-22> EMC: 03cmorley 07cl_v7124_branch * 10emc2/src/hal/classicladder/config_gtk.c: add define to stop display of direct i/o data-we don't use it in EMC
[18:16:48] <CIA-22> EMC: 03jmkasunich 07TRUNK * 10emc2/src/hal/components/comp.comp: improve HAL comparator man page
[18:19:27] <CIA-22> EMC: 03jmkasunich 07v2_2_branch * 10emc2/src/hal/components/comp.comp: improve HAL comparator man page
[19:13:40] <cradek> +pin out bit equal "Match output. True when difference between in1 and in0 is
[19:13:43] <cradek> less than 'hyst'.";
[19:13:53] <cradek> this is incorrect
[19:14:25] <cradek> this defines a hysteresis region of 2*hyst
[19:14:59] <jmkasunich> oops
[19:17:22] <jmkasunich> should say "True when difference is less than 'hyst'/2"
[19:17:31] <jmkasunich> fixing now
[19:17:43] <jmkasunich> that looks weird with the single quotes
[19:18:06] <jmkasunich> but I wanted to quote it to indicate that it is a param name, not just a word
[19:20:28] <cradek> aren't the special names highlighted somehow in the manpage?
[19:21:09] <jmkasunich> not when they appear in the text
[19:21:14] <jmkasunich> http://www.linuxcnc.org/docview/html//man/man9/comp.9.html
[19:21:28] <jmkasunich> there is markup that could be embedded in the .comp file I guess
[19:21:37] <jmkasunich> that's probably the right way to do it
[19:21:51] <jmkasunich> although I dunno if comp uses that text for anything other than man pages
[19:26:02] <jmkasunich> grrr - man file markup doesn't survive processing by comp
[19:26:11] <jmkasunich> probably needs some extra escapes
[19:31:05] <jmkasunich> well, it seems that \\fB and \\fR work (normal manpage markup is \fB and \fR)
[19:31:20] <jmkasunich> I wonder if anything else uses those strings, and might get broken
[19:31:40] <jmkasunich> comp,g is not very easy for me to read, I can't tell if those strings are used anywhere else
[19:38:40] <CIA-22> EMC: 03jmkasunich 07v2_2_branch * 10emc2/src/hal/components/comp.comp: Hystersis band is +/- half of hyst value. Added man page markup codes.
[19:39:51] <CIA-22> EMC: 03jmkasunich 07TRUNK * 10emc2/src/hal/components/comp.comp: Hystersis band is +/- half of hyst value. Added man page markup codes.
[19:43:49] <jmkasunich> how/when does the online copies of the manpages get updated
[19:43:51] <jmkasunich> ?
[19:44:11] <jmkasunich> I was pleasantly surprised to see my changes online already
[19:44:31] <jmkasunich> (the first, incorrect changes - I assume the corrections will appear soon)
[20:40:25] <cradek> jmkasunich: cool, sounds like all you have to do to get a graphical HAL editor is learn scheme
[20:44:17] <jmkasunich> heh
[20:44:40] <jmkasunich> that wasn't true 3 years ago, and it isn't true now
[20:47:27] <jmkasunich> just put a tire on my 3 spoked wheel
[20:47:45] <jmkasunich> turned a 0.041 groove in the edge of the 0.075 thick wheel, and stuck a rubber band in it
[20:48:03] <jmkasunich> gotta have that traction!
[20:48:04] <cradek> cool
[20:48:08] <cradek> nice flat floor?
[20:49:02] <jmkasunich> yeah - the atrium at work, polished terrazzo
[20:50:34] <cradek> did the wheel turn out to be fairly well balanced?
[20:52:12] <jmkasunich> haven't spun it fast
[20:52:32] <jmkasunich> the floor wheel with the tire isn't even on a shaft yet
[20:53:01] <jmkasunich> the 0.125 thick one that winds up the string came out OK, not perfect - when I pressed it onto a hub it wasn't square
[20:53:12] <jmkasunich> maybe 0.020 out a the 3.3" diameter OD
[20:53:29] <jmkasunich> thats OK for there, but I need to do better for this one
[20:53:57] <jmkasunich> the first hub didn't have a shoulder, but this one will - that should help
[20:54:37] <jmkasunich> once I get this wheel on a hub, and make a shaft, I can make the fusee
[20:55:09] <jmkasunich> that should be fun
[20:55:19] <cradek> yep I'm anxious to hear whether that works
[20:55:58] <jmkasunich> my stumbling block at the moment is how to support the shaft
[20:56:21] <jmkasunich> bearings at each end is the natural approach, but would require some odd brackets to go around the wheel
[20:56:48] <jmkasunich> cantilevered seems better - then the fusee can screw on the end of the shaft easily, and both bearings can be on the other end of the wheel
[20:57:40] <jmkasunich> but NEFS is kicking in, as always
[21:00:27] <jmkasunich> stuart might take the bait ;-)
[21:00:41] <cradek> heh
[21:01:08] <cradek> yeah what he needs is another language to learn
[21:01:46] <cradek> oh the other stuart
[21:01:54] <cradek> people should have individual identifiers
[21:02:00] <skunkworks> heh - the new stuart
[21:02:07] <jmkasunich> yeah, new stu