#emc | Logs for 2007-03-26

Back
[00:00:36] <roltek> what kind of material you usaully work with
[00:00:40] <anonimasu> :)
[00:02:01] <skullworks-PGAB> 303 stainless, A36 mild steel, lots of A2 tool steel, and tonnes of aluminum.
[00:02:23] <roltek> molds
[00:02:27] <skullworks-PGAB> work 2 jobs :(
[00:03:25] <skullworks-PGAB> so during the week I battle robots and other machines - and often spen the day running a CNC metal spinner
[00:03:49] <skullworks-PGAB> roll forming journals in tube stock
[00:04:36] <anonimasu> ouch
[00:05:05] <anonimasu> lots of stuff to do
[00:05:13] <roltek> never seen the cnc spinners but have seen hydraulic
[00:05:28] <skullworks-PGAB> nights and weekends I work part time for a company I used to for full time building fixtures and doing all the CNC mill programming.
[00:06:26] <skullworks-PGAB> well the CNC is hydraulic
[00:06:41] <roltek> mostly work with 9310 9310var 4340 300m
[00:06:43] <anonimasu> o.
[00:06:45] <anonimasu> oh..
[00:06:47] <anonimasu> I got a idea.
[00:06:56] <skullworks-PGAB> material comes out about 185 deg.
[00:07:00] <anonimasu> I know how the geometry of the changer arm should look
[00:09:10] <anonimasu> ^_^
[00:32:17] <skullworks-PGAB> * skullworks-PGAB must go to the shop - some minor crisis - overwritten offsets...
[01:09:57] <ejholmgren> woo
[01:46:13] <CIA-18> 03jepler 07TRUNK * 10emc2/src/emc/usr_intf/axis/scripts/axis.py:
[01:46:13] <CIA-18> bugfix: if the current axis is changed while the "-"/"+" buttons are pressed,
[01:46:13] <CIA-18> remember which axis to start jogging. Same for the "-"/"=" keystroke commands.
[02:08:41] <CIA-18> 03jepler 07TRUNK * 10emc2/src/emc/usr_intf/axis/scripts/axis.py: fix occasional warning message: invalid command name "-1211088452callit"
[02:13:50] <CIA-18> 03jepler 07TRUNK * 10emc2/src/emc/ini/emcIniFile.cc:
[02:13:50] <CIA-18> second fix for this error: INIFILE: ERR_CONVERSION, section=TRAJ, tag=LINEAR_UNITS, num=0, lineNo=129
[02:13:50] <CIA-18> must catch the IniFile::Exception since we don't know whether it's in
[02:13:50] <CIA-18> exception-throwing mode or not!
[02:25:29] <renesis> hey can i set axis gui to anything besides g54 workspace?
[02:25:50] <renesis> seems to just ignore g53-g59 stuff
[02:26:36] <jepler> axis 'touch off' always sets the g54 coordinate system offset. the coordinate system offset in use is always shown if you are showing relative coordinates.
[02:26:49] <jepler> at least, that's how it's supposed to work
[02:26:59] <jepler> I never use anything but the g54 coordinate system, personally
[02:27:04] <renesis> i know how do i mdi in g53 or g55-g59?
[02:27:15] <renesis> okay yeah usually but if i want to
[02:27:47] <CIA-18> 03jepler 07v2_1_branch * 10emc2/src/emc/usr_intf/axis/scripts/axis.py:
[02:27:47] <CIA-18> merge rev 1.67, 1.68:
[02:27:47] <CIA-18> - fix occasional warning message: invalid command name "-1211088452callit"
[02:27:47] <CIA-18> - fix changing axes with keyboard while jogging with the jog button
[02:28:04] <renesis> like, i wanna setup 6 fixtures, use the same gcode for all 6 fixtures, just switching workspaces between cycles
[02:28:38] <cradek> just program the fixture locations using G10 L2 P ...
[02:28:51] <jepler> and switch to the coordinate system you want with 'g54', 'g55', etc
[02:28:58] <cradek> right
[02:29:10] <cradek> AXIS doesn't really get involved at all
[02:29:32] <jepler> axis may always show the preview as though the g54 coordinate system was in effect
[02:29:42] <renesis> active gcodes: g53 g54
[02:29:47] <renesis> thats a bug, no?
[02:29:48] <jepler> I'm not sure about that
[02:29:58] <cradek> the G54 coordinate system IS the one in effect after a program ends (M2)
[02:30:14] <renesis> okay
[02:30:18] <cradek> during the program run you'll see the axes move around to the other coordinate systems
[02:30:21] <renesis> say i never wanna use g54, ever
[02:30:30] <cradek> then ignore the 'touch off' button
[02:30:53] <jepler> what did you do to get g53 and g54 both in 'active gcodes'?
[02:30:54] <renesis> so emc will do it, but axis md1 or manual control wont
[02:31:07] <renesis> g0 g53
[02:31:16] <renesis> cuz it wont let you just do a g53
[02:31:17] <cradek> sure you can switch systems in mdi
[02:31:28] <renesis> it says you need to specify a g0 or g1
[02:31:42] <jepler> g53 is not modal
[02:31:41] <renesis> which just seems bunk, cuz usuallu a g5x is set in first block
[02:31:43] <cradek> yes g53 is not sticky and must be put on every line
[02:31:58] <renesis> in emc?
[02:32:06] <renesis> thats not very normal, heh
[02:32:08] <cradek> yes, in rs274ngc
[02:32:25] <renesis> okay ill check cuz thats weird
[02:32:44] <renesis> g54-g59 are still modal?
[02:32:46] <Guest720> Guest720 is now known as skunkworks
[02:32:55] <cradek> yes
[02:33:04] <renesis> yeah it switched
[02:33:04] <cradek> http://www.linuxcnc.org/handbook/RS274NGC_3/RS274NGC_33a.html#1002573
[02:33:26] <renesis> ty
[02:33:35] <renesis> isnt there a nist link?
[02:33:37] <jepler> tkemc will also display 'g53 .. g54' in the 'active g-codes' area, so it's not a GUI-specific behavior
[02:33:53] <cradek> to find that I used Applications/CNC/Gcode quick reference, then clicked on G53
[02:34:21] <renesis> good job
[02:35:13] <CIA-18> 03cradek 07TRUNK * 10emc2/src/emc/usr_intf/axis/scripts/axis.py: remove debug output
[02:35:35] <renesis> okay so g53 needs to be specified every line, with a linear movement, and g54-59 are still modal
[02:35:55] <renesis> oh i guess any movement
[02:36:01] <skunkworks> that is how it works when I played with it.
[02:36:22] <CIA-18> 03cradek 07v2_1_branch * 10emc2/src/emc/usr_intf/axis/scripts/axis.py: remove debug output
[02:36:45] <cradek> the idea is that nobody would want to machine in g53 mode - those coordinates only make sense when locating home position, tool change position, etc.
[02:36:46] <jepler> cradek: thanks for fixing that up
[02:37:18] <cradek> welcome
[02:37:43] <renesis> its still weird
[02:38:19] <cradek> if I had to pick the weirdest things about gcode, g53 would be down on the list a ways...
[02:38:23] <renesis> like, setting a workspace, then doing another movement next line, and not being in that workspace, ph33r
[02:40:49] <CIA-18> 03jepler 07v2_1_branch * 10emc2/debian/changelog: items I did
[02:41:38] <renesis> g53 motion in machine coordinate system
[02:42:01] <renesis> heheh, okay it says so
[02:42:19] <cradek> ??
[02:43:27] <renesis> from linuxcnc.org, is the words they use for g53
[02:50:10] <toastydeath> why is g53 weird?
[02:50:23] <toastydeath> lathes make extensive use of g53
[02:50:29] <toastydeath> and some milling machines do, as well
[02:52:07] <toastydeath> also g53 is modal on most machines
[02:52:27] <toastydeath> production machines, anyway
[02:53:33] <cradek> google says on Haas, it's not
[02:53:50] <toastydeath> the Haas manual says it is
[02:53:54] <toastydeath> and i program a Haas
[02:54:05] <cradek> also Fadal?
[02:54:12] <toastydeath> there may be an option to turn it off on a Haas
[02:54:19] <toastydeath> Fadal, no idea
[02:54:22] <toastydeath> Fanuc does
[02:55:19] <toastydeath> older controllers make heavy used of g53 as a modal command
[02:55:51] <toastydeath> and many experienced CNC programmers still do it that wayt
[02:55:53] <toastydeath> *way
[02:56:03] <cradek> this is not interesting to argue about - in NGC it's not modal, and that's also apparently the case for some other controllers
[02:56:15] <toastydeath> a minority of controllers
[02:56:17] <cradek> incompatibilities like that are a big suckage about gcode
[02:56:32] <toastydeath> the original standard and the Fanuc standard have it as modal
[02:56:53] <toastydeath> the move some controllers make to a single block command is unorthodox
[02:58:27] <jepler> the point is, it's not up for debate what emc will do: it will follow the ngc "spec" when changing would change the meaning of currently valid, useful gcode
[02:58:44] <toastydeath> sure, as EMC is not a production controller
[02:58:48] <jepler> this has a meaning now, whether you prefer that meaning or not: g53 g0 x0 / g1 x1 f1
[02:59:01] <jepler> I don't even know what that word means -- is it like "enterprise"?
[02:59:11] <toastydeath> it means usable in a production environment
[02:59:28] <jepler> * jepler rolls his eyes.
[02:59:42] <cradek> (your troll is showing)
[02:59:42] <toastydeath> * toastydeath shrug
[03:00:00] <toastydeath> you want to deviate from the original gcode standard, and the fanuc standard
[03:00:02] <toastydeath> be my guest
[03:01:20] <toastydeath> it only adds to the confusion.
[03:03:54] <cradek> http://www.haascnc.com/training/MillProgram_PDF/Mill%20G%20and%20M%20Card.pdf
[03:04:57] <toastydeath> interesting, I haven't seen a haas mill from the factory where g53 was non-modal.
[03:05:15] <toastydeath> but if the card says so, i'll take it.
[03:05:39] <jepler> goodnight all
[03:05:50] <cradek> g'night jepler
[03:06:34] <cradek> http://www.google.com/search?q=haas+g53+non-modal
[03:06:56] <toastydeath> like i just said, if the Haas card says so, I'll take it.
[03:07:29] <maddash> 2.6.20.4!
[03:07:44] <ds3> oh no
[03:08:11] <ds3> another EMC vs Haas vs Fanuc debate
[03:08:23] <toastydeath> head for the hills!
[03:08:33] <ds3> long live Haas!
[03:08:34] <toastydeath> G53 is legacy to support older programs.
[03:08:42] <toastydeath> where work shift is used instead of a work offset.
[03:08:53] <ds3> G53 == use machine coordinates?
[03:08:56] <toastydeath> yes
[03:09:02] <ejholmgren> "1/2 od x .109w 1026 dom steel tube" ... would .109" be the wall thickness?
[03:09:11] <ds3> then what does it mean to call out G53 if it is not modal?
[03:09:37] <toastydeath> it only works on the one line.
[03:09:43] <ds3> Oh
[03:09:54] <toastydeath> you can't set the machine to move in the machine coordinate system
[03:10:22] <toastydeath> which reduces it's functionality for parametic programming and legacy support
[03:10:28] <ds3> so... then, on startup, if I do a G00 X0. Y0. Z0. as the first (and only) line, what does it mean?
[03:10:45] <toastydeath> it will run to whatever work shift it starts up on.
[03:10:51] <toastydeath> er, work coordinate.
[03:11:03] <ds3> Hmmmm that doesn't sound good
[03:12:02] <toastydeath> i mean, if you have it set to the right coordinate system it should be fine.
[03:12:17] <toastydeath> or just leave g54 or whatever your default is to the machine default.
[03:12:54] <ds3> I suppose you can do the same thing by clearing out another coordinate, say for G55, and use that in lieu
[03:13:18] <toastydeath> indeed, but that's not a substitute when you're doing macros/parametric programming or have a legacy program
[03:13:28] <ds3> *nod*
[03:13:48] <ds3> it is for reasons like this that i have a perl script to EMCify things
[03:14:15] <toastydeath> lol
[03:14:20] <toastydeath> epic perl maneuver
[03:14:35] <ds3> there are other quirks but so be it
[03:14:42] <ds3> (i.e. R planes)
[03:14:48] <toastydeath> ?
[03:14:55] <toastydeath> how does emc handle that
[03:15:04] <toastydeath> (differently, anyway)
[03:15:13] <ds3> R planes are not modal in EMC; they just carry it within the can cycle
[03:15:17] <ds3> whereas Haas treats it as modal
[03:15:23] <toastydeath> so does fanuc
[03:15:48] <toastydeath> less crash-prone that way.
[03:16:12] <ds3> so my answer to it is to make sure the R is called on in all relevant canned cycle (done automatically by a perl script)
[03:17:04] <ds3> would prefer to have a global setting (a parameter or such) that would enable 'Haas emulation' or "Fanuc emulation"
[03:17:14] <ds3> but I don't have time to write that patch
[03:17:14] <toastydeath> yeah
[03:17:45] <ds3> it would also give me better confidence that a CAM post won't blow up if it doesn't have a EMC specific postprocessor
[03:18:18] <toastydeath> has anyone written a postrpocessor for any program in EMC?
[03:18:26] <toastydeath> *postprocessor
[03:18:47] <ds3> I heard the Mach3 one is very close to EMC
[03:18:54] <SWPadnos> I twiddled with a TurboCADCAM EMC post
[03:19:05] <SWPadnos> Mach3 is based on the PD EMC interpreter
[03:19:16] <SWPadnos> as is DeskCNC, I believe
[03:19:34] <ds3> the $1M question there is... how much has Mach3 diverged from EMC
[03:19:45] <SWPadnos> loads
[03:19:51] <ds3> sigh
[03:19:56] <SWPadnos> it should be based on the very old NIST code
[03:20:33] <SWPadnos> they have extra canned cycles for various things (like routing a circle)
[03:20:50] <ds3> Hmmm
[03:21:19] <ds3> what about the ISO Gcode? is EMC suppose to be 100% compatible with that?
[03:21:46] <SWPadnos> if they stuck to the RS274NGC rev3 spec for ISO, yeah ;)
[03:21:55] <ds3> hahahaha
[03:22:08] <SWPadnos> the EMC interpreter was the reference interp for G-code
[03:22:12] <ds3> is this 'RS' the same 'RS" as in RS232?
[03:22:21] <SWPadnos> dunno - could be
[03:22:32] <toastydeath> fairly certain
[03:22:36] <SWPadnos> isn't that "Radio Standard" or something?
[03:22:41] <ds3> so it will eventually become a EIA274NGC?
[03:22:47] <toastydeath> i bet it has to do with hip hop
[03:22:50] <ds3> RS == Recommended Standard
[03:22:53] <SWPadnos> ah
[03:24:14] <ds3> don't care much for the ISO folks.... have more respect for EIA
[03:24:25] <SWPadnos> I guess I was thinking of RG "Radio Guide"
[03:25:28] <ds3> So, SWPadnos, where in San Jose are you going to be staying?
[03:25:41] <SWPadnos> the Hilton that's attached to the convention center
[03:25:49] <ds3> Oh lucky you
[03:25:56] <SWPadnos> indeedy
[03:26:24] <ds3> getting a rental to get around town?
[03:26:35] <SWPadnos> nope. that's why I can (almost) afford the Hilton ;)
[03:26:42] <maddash> what's in san jose?
[03:26:43] <SWPadnos> no $30/day parking fees either
[03:26:56] <SWPadnos> ESC http://www.esconline.com
[03:27:07] <ds3> $30/day!?! I'll rent you a space for $25/day!
[03:27:11] <SWPadnos> heh
[03:27:16] <SWPadnos> thanks ;)
[03:27:25] <SWPadnos> when are you going to be at the show?
[03:27:47] <ds3> all 3 days, going to a bunch of the free classes
[03:27:53] <SWPadnos> cool
[03:28:23] <a-l-p-h-a> SWPadnos, how was Cali?
[03:28:55] <ds3> if you want to meet up or something, I am up for that but I'll be coming via the train
[03:29:09] <SWPadnos> a-l-p-h-a, dunno, haven't been there yet this year
[03:29:24] <a-l-p-h-a> thought you went to visit jymmm.
[03:29:24] <SWPadnos> ds3, sounds good to me. I'll be there the whole week
[03:29:32] <SWPadnos> going next week
[03:29:38] <a-l-p-h-a> oooh
[03:29:43] <SWPadnos> but not to visit him - that's just an added benefit (?)
[03:29:46] <a-l-p-h-a> thought you were there last week.
[03:29:52] <SWPadnos> nope, that was Florida
[03:29:56] <a-l-p-h-a> SWPadnos, free foodage. ;)
[03:30:02] <SWPadnos> oops, no, that was Manchester NH
[03:30:16] <SWPadnos> yeah - it's a great deal
[03:30:23] <a-l-p-h-a> house of prime rib?
[03:30:27] <SWPadnos> $1200 for the conference fees, and I get 6 free lunches!!
[03:30:38] <a-l-p-h-a> $200/lunch.
[03:30:40] <SWPadnos> of course - I have to go there whenever I'm anywhere near SF
[03:30:50] <a-l-p-h-a> better be all you can eat lobster, and oysters, and angus beef.
[03:31:01] <SWPadnos> or a sandwich, soda, and chips /:
[03:31:03] <SWPadnos> :/
[03:31:09] <ds3> eatting at the convention center is steep... it is like $15 for a coke + vending machine sandwiches
[03:31:25] <SWPadnos> they provide lunch for conference attendees
[03:31:30] <a-l-p-h-a> $15/coke??
[03:31:32] <a-l-p-h-a> yikes.
[03:31:35] <a-l-p-h-a> thought Raptor games were bad.
[03:31:39] <ds3> coke from the machine is like $5
[03:31:41] <toastydeath> holy crap
[03:31:43] <a-l-p-h-a> $4/bottle of pop.
[03:31:49] <toastydeath> you might as well stay home and get addicted to crack
[03:31:50] <SWPadnos> I've paid $12 for a small caesar salad, plus the $3.50 coke
[03:31:57] <toastydeath> that is insanity
[03:31:58] <ds3> think the round up so you can use a $5 bill ;)
[03:32:03] <SWPadnos> trade shows are insanity
[03:32:06] <ds3> I'm bring lunch or walking over to quiznos
[03:32:14] <SWPadnos> heh
[03:32:15] <ds3> at least when I pay $10, I get something edible
[03:32:27] <SWPadnos> I'd just go to the executive lounge at the Hilton
[03:32:44] <SWPadnos> I'm not sure they have food at lunchtime though
[03:34:26] <a-l-p-h-a> night folks
[03:34:31] <SWPadnos> see you
[03:35:22] <ds3> there are lots of places w/i walking distance
[03:35:48] <toastydeath> chipotle!
[03:36:32] <SWPadnos> there's a mostly psdestrian mall about 5 blocks from the convention center with some nice enough places for dinner
[03:36:37] <SWPadnos> pedestrian
[03:37:05] <ds3> there is? Hmmm wonder what they are talking about
[03:37:26] <ds3> you can always get on the train and up a few stops to get to stuff like Japan Town
[03:38:22] <ds3> SWPadnos: is that SMSable?
[03:38:28] <SWPadnos> I don't think so
[03:38:41] <SWPadnos> lemme check
[03:39:34] <CIA-18> 03cradek 07v2_1_branch * 10emc2/VERSION: 2.1.4 release
[03:39:34] <CIA-18> 03cradek 07v2_1_branch * 10emc2/debian/changelog: 2.1.4 release
[03:42:09] <CIA-18> 03cradek 07v2_1_branch * 10emc2/VERSION: bump after release
[03:43:06] <SWPadnos> ds3, SMS OK, just not in the basic plan
[03:43:17] <SWPadnos> what carrier are you on?
[03:43:46] <ds3> TMO
[03:43:56] <ds3> I'd msg you but this server is being annoying
[03:43:59] <SWPadnos> ok. I'm on VZW, so no help there
[03:44:06] <SWPadnos> it's off anyway
[03:44:22] <ds3> can I try it? I can send to VZW people just fine
[03:44:27] <SWPadnos> it
[03:44:30] <SWPadnos> it'll work
[03:44:37] <ds3> oh okay
[03:44:55] <SWPadnos> I just wasn't sure about the charges, but $0.15 each isn't bad
[03:47:09] <SWPadnos> hmm. I guess I was thinking of 1st or 2nd street, maybe it's not a pedestrian mall
[03:47:55] <ds3> Oh that side
[03:48:00] <SWPadnos> heh
[03:48:03] <cradek> cradek has changed the topic to: Welcome! EMC (Enhanced Machine Controller) is a linux-based CNC control. | Latest release: EMC 2.1.4 | http://www.linuxcnc.org | http://wiki.linuxcnc.org
[03:48:07] <cradek> EMC 2.1.4 is released
[03:48:12] <ds3> I guess you can call it that.... there is the bunch of stuff on the other side near SJSU
[03:48:20] <SWPadnos> cradek, yay!
[03:48:31] <SWPadnos> ok - saw some of that
[03:48:47] <SWPadnos> well, I've got to finish up some work before I leave town tomorrow
[03:48:49] <SWPadnos> see you later
[03:48:57] <ds3> speaking of which, are there any other machine people using EMC as their controller besides the BDI folks w/Sherline?
[03:49:00] <ds3> later, SWPadnos
[03:49:06] <SWPadnos> smithy, I think
[03:49:33] <ds3> BDI based?
[03:49:41] <SWPadnos> I'm not sure
[03:50:03] <SWPadnos> actually, they may be using something else right now
[03:50:27] <SWPadnos> they've hosted EMC Fest on 1 or 2 occasions, and I think they also used to host an ISO
[03:53:36] <ds3> I see.
[03:53:53] <ds3> I am guessing no one is putting together turnkey controller systems based off EMC either?
[03:54:23] <cradek> "the cool tool" is
[03:55:02] <cradek> (also, due to the nature of free software, it's not likely that we would know about all the folks using EMC)
[03:55:45] <ds3> why? isn't that needed for GPL compliance?
[03:56:22] <jmkasunich> GPL doesn't say "you must inform the authors if you are using their code"
[03:56:24] <cradek> that users tell the authors they're using the software? certainly not
[03:56:45] <cradek> using the software secretly is one of your rights
[03:56:54] <ds3> but they must publish the code they used when selling their controller
[03:57:17] <cradek> yes if they redistribute EMC they must provide corresponding source with it
[03:57:26] <ds3> while we may not know about all the controllers, we should be able to deduce who is not using it
[03:57:56] <toastydeath> and then hold those people hostage!
[03:58:00] <toastydeath> USE EMC OR DIE
[03:58:34] <toastydeath> i have the best ideas.
[03:58:49] <jmkasunich> goodnight all
[03:59:12] <ds3> would be interesting to sell a complete turn key EMC based controller
[04:15:55] <toastydeath> ds3: that would be v. cool
[04:16:06] <toastydeath> as like, a retrofit kit
[04:31:42] <ds3> yep
[04:32:03] <ds3> no traditional KB/PC/Monitor; just a single unit
[06:48:12] <renesis> ds3: would be sex
[06:48:41] <renesis> could do it with an itx mobo and lcd
[06:48:49] <renesis> custom case to hold it all
[06:49:23] <renesis> with like kb and trackball or touchpad
[07:01:14] <skullworks-PGAB> hey toasty, Why does the Haas "Machinist's CNC Reference Guide" state on pg11 (mills) and again on pg 16(lathes that G53 is a non-Modal command.
[07:03:51] <skullworks-PGAB> (its the 2005 edition, but they were handing them out at last years WESTEC show)
[07:04:29] <skullworks-PGAB> WESTEC opens tomorrow... and I can't go :(
[08:21:16] <anonimasu> morning
[08:22:44] <anonimasu> ds3: yes, I have a mastercam postprocessor
[08:22:53] <anonimasu> ds3: I used to have a visualmill one
[08:22:58] <anonimasu> and I've used something else also..
[08:23:12] <anonimasu> the fanuc posts are very similiar to what emc wants...
[08:29:49] <Jymmm> morning anonimasu
[08:40:27] <anonimasu> morning
[08:40:30] <anonimasu> *hunts sensors*
[08:55:15] <ds3> Hmmmm
[09:28:05] <alex_joni> anonimasu: hi
[09:28:12] <alex_joni> you have some CAM package?
[09:31:22] <jlmjvm> alex: good morning
[09:32:02] <alex_joni> hi jlmjvm
[09:35:34] <jlmjvm> i was wondering if glass scales could be used on a second parallel
[09:36:11] <jlmjvm> to use for homing to an index and ferror
[09:36:54] <jlmjvm> that is if the scale has index marks
[09:37:54] <alex_joni> sure.. why not
[09:38:02] <alex_joni> but it depends on the speed you need
[09:38:05] <alex_joni> the parport might be a bit slow
[09:38:20] <alex_joni> the glass scales I know have a pretty high resolution (um)
[09:39:18] <jlmjvm> these are from a dro i think,figured they were graduated in tenths
[09:41:29] <jlmjvm> when you say paraport might be slow,how slow are you meaning
[09:41:44] <alex_joni> just as fast as the pulses come oout of it
[09:41:49] <alex_joni> usually 40kHz or so
[09:42:12] <jlmjvm> thats plenty fast for this mill
[09:42:18] <alex_joni> so if you have 20000 edges on the scale / inch, you can max go to 2IPS (I would recommend way less)
[09:42:27] <alex_joni> it all depends what the scales are
[09:43:31] <jlmjvm> the same would be true for encoders also,correct?
[09:44:11] <jlmjvm> the more counts,slower rapids
[10:13:33] <alex_joni> jlmjvm: correct
[10:13:50] <anonimasu> alex_joni: yes
[10:20:03] <anonimasu> ^_^
[10:21:37] <anonimasu> alex_joni: still there?
[10:31:12] <alex_joni> yeah
[11:06:40] <anonimasu> hm
[11:07:36] <alex_joni> hm?
[11:08:23] <anonimasu> I need to go for a bit..
[11:08:25] <anonimasu> laters..
[11:35:36] <SWPadnos_> SWPadnos_ is now known as SWPadnos
[12:09:24] <skullworks-PGAB> ug - off to work I go.
[12:13:40] <anonimasu> alex_joni: wake up
[12:29:23] <RichiH> Hi all. As you surely noticed, we just had a largish split. A sponsor had a route flap which took a part of freenode from the net. Sorry for the inconvenience and thank you for flying freenode air!
[12:29:25] <skunkworks> well - I am not going to move again - ever.
[12:43:47] <alex_joni> skunkworks: haha
[13:17:25] <skunkworks> I ache all over ;)
[13:36:26] <alex_joni> skunkworks: haha
[13:41:53] <skunkworks> great job everyone on the latest release :)
[13:42:58] <alex_joni> oops.. that reminds me.. still need to do SF & linuxcnc.org
[14:06:09] <wallyGL> hi everyone.. i would just like to ask... if i compile emc2 with '--enable-simulator' option... is there a way to still run emc2 together with the realtime system..
[14:06:29] <cradek> simulator means realtime is not used
[14:08:23] <wallyGL> ok thanks... ill probably just have another copy compiled without that option...
[14:09:11] <cradek> yes with --enable-run-in-place you can have any number of builds
[14:40:14] <tomp> Dallur: I'm studying your http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl/emcinfo.pl?Torch_Height_Control.
[14:40:20] <tomp> It appears that mux2.3 feeding the Zpos-cmd gets either the Zpos-instructed-cmd or the Zpos-fb according to the selector 'moveZEnabled'.
[14:40:26] <tomp> If so, it appears that emc gets the current position as the command sometimes, allowing non-emc control over the axis that is now 'monitored' by emc.
[14:40:35] <tomp> Is this the correct interpretation?
[14:40:37] <tomp> btw: thanks for the wiki pages.
[14:42:02] <jepler> tomp: yes I think that's right
[14:51:02] <jepler> I'm not sure if the net graph clears anything up for you: http://emergent.unpy.net/files/sandbox/dallur.svg
[15:48:07] <tomp> jepler: thanks! i'm looking at editors that let you drag those bits around. to let you untangle one idea, or re-arrange stuff yet keep the connections
[15:48:45] <jepler> tomp: let me know if you find something
[16:43:57] <Dallur> tomp: Hey, I'm at work atm and working my ass off because of a deadline I had Yesterday :P
[16:44:38] <Dallur> tomp: Yes that is correct, btw i'm rewriting the whole thing, cleaning it up and making better documenation
[16:46:01] <Dallur> tomp: just need to quit my job to get the time I need to finish it :P
[16:51:52] <wayne609> anyone home today ;)
[16:52:16] <jepler> wayne609: looks that way
[16:52:49] <wayne609> just starting to look at emc again after a long time with mach myself.
[16:52:56] <wayne609> Got some serious toys to retrofit
[16:53:13] <wayne609> emc seems to have come a long wa since my last look at it
[16:53:44] <jepler> thanks, we believe it has too
[16:53:56] <wayne609> quite intested to know what options I have to fully implement my Hardinge HXL cnc lathe
[16:54:42] <wayne609> I do find to docs hard to follow at this point though
[16:55:12] <wayne609> wish I had more time to dijest them lol
[16:55:31] <jepler> first thing is to know what lathe-specific g-codes we have -- g33 "spindle synchronized move" and g76 "canned threading cycle". We're missing constant surface speed, to name one important lathe feature
[16:56:03] <wayne609> well thats not so important to me, my Ahha system never had it either
[16:56:07] <wayne609> but it's nice to have
[16:56:13] <jepler> OK
[16:56:37] <wayne609> but i'd like to be able to use the existing 1024 cpr encoder on the spindle
[16:56:51] <jepler> does it have index pulse?
[16:57:13] <wayne609> yes, it's a real heidenhain encoder lol
[16:57:18] <jepler> OK
[16:57:30] <jepler> servo motors or steppers for X and Z?
[16:57:42] <wayne609> there seem to b e afewmotion cards I could chose, but which one |I don't know
[16:58:12] <wayne609> i have dc brushed servo's wih step/dir inputs, the intention was to use mach
[16:58:17] <wayne609> but not so suire now
[16:58:49] <cradek> do you still have the original amps so you could run emc closed loop instead of step-servo?
[16:58:49] <wayne609> I'm using italian motors and cnc teknix drives
[16:59:13] <wayne609> nope they were ancient SCR amps, no way I was going to reuse that liability
[16:59:42] <wayne609> old amps can be bad news in my experiance
[16:59:46] <cradek> gotta run, lunchtime
[17:00:21] <jepler> wayne609: if you weren't considering step+direction output, I'd direct you towards the mesa 5i20 PCI card
[17:00:47] <wayne609> well I can get analogue amps but modern ones LOL
[17:00:50] <jepler> the mesa 5i20 doesn't yet do step+direction output with emc, only PWM (or analog with a separate adapter board 7i3x)
[17:01:32] <wayne609> The old seimens amps were just too risky, I need to use this machine 7 days a week lol
[17:02:07] <jepler> it's lunchtime here too .. take a look at the 5i20 on http://mesanet.com/
[17:02:11] <wayne609> and not have to pay $100/hr for when they went down
[17:02:15] <jepler> category "motion control"
[17:02:32] <jepler> also the 7I33 "Quad Analog servo interface"
[17:03:04] <wayne609> no tacho's on the new motors though
[17:46:36] <skunkworks> wayne609: not tachs on the new motors - but encoders?
[17:48:05] <tomp> jepler: Dallur: the editor for dot files is dotty! http://www.graphviz.org/Documentation/dottyguide.pdf
[17:56:28] <Dallur> hmm anyone here know if Krum milling machines are any good ?
[18:04:55] <tomp> jepler: try 'which dotty' to see if it's installed. Then 'dotty somedotfile.dot' to test. Right click on the canvas for options, and left drag to rearrange. i cant get it to save yet, and 'lefty' is having issues with some auto-generated files.
[18:09:55] <maddash> "while (fabs(etime() - start_time) < 10.0 &&" in emcsvr.cc -- why the call to fabs? time progresses in only one direction, last time i checked..
[18:15:53] <jepler> does seem like a silly way of writing it, I agree
[18:19:29] <jepler> I am having trouble imagining a motivation for fabs() that isn't silly, though
[18:20:36] <jepler> tomp: this 'lefty' looks like an interesting message
[18:20:39] <jepler> er, interesting language
[18:22:04] <maddash> er, if you have a variable that could be on either side of a constant?
[18:22:39] <maddash> also, the subsequent checks for NULL are a bit inefficient -- can I rewrite this?
[18:23:31] <alex_joni> maddash: yeah, sure .. as long as you keep the code readable
[18:24:01] <alex_joni> we prefer easy to follow code vs. high performance hacks which are really obscure
[18:26:23] <maddash> hm, then maybe I shouldn't rewrite
[18:27:23] <jepler> maddash: these NULL checks? if (NULL == emcCommandChannel) {
[18:28:02] <maddash> yes. it seems that you're making one extra check for each of those.
[18:29:07] <jepler> the condition of the 'while' loop tests whether any of them are NULL; the if tests whether a particular one is NULL
[18:31:31] <lerman> Well, in the old days, we would just AND all of them together and test the result. If any is NULL (zero), the AND will be zero and we can do just a single test.
[18:32:05] <lerman> But I suppose the compiler would complain if we tried that and the people who read the code would (should) not be very happy with that.
[18:32:05] <cradek> anything with a sleep 0.2 seconds in the loop probably doesn't need much optimization
[18:32:15] <jepler> cradek: hah
[18:32:21] <jepler> cradek: FTW
[18:32:33] <lerman> Particularly if it the main program that is executed at startup.
[18:32:48] <lerman> ...if it is...
[18:32:53] <alex_joni> jepler: ftw?
[18:32:54] <cradek> yes it's at the beginning of main
[18:33:12] <jepler> "for the win" -- that is, chris "won" the argument, in my opinion
[18:33:23] <alex_joni> oh sure he did
[18:33:34] <cradek> haha I didn't even know it was an argument
[18:33:39] <cradek> guess that's a double win
[18:33:46] <alex_joni> win-win?
[18:33:47] <jepler> "discussion" then
[18:33:54] <lerman> I don't know where it is from, but: "There are things worth doing that aren't worth doing right."
[18:33:58] <alex_joni> cradek: thanks for the release :P
[18:34:05] <cradek> welcome
[18:34:15] <cradek> thanks to you guys who update all the various docs every time
[18:34:18] <jepler> any problem reports from the new release yet?
[18:34:21] <cradek> no
[18:34:23] <alex_joni> nope
[18:36:13] <maddash> ok, I'm done
[18:36:45] <alex_joni> someone turn him over :D
[18:37:34] <maddash> cradek: have you seen the code?
[18:38:11] <alex_joni> * alex_joni feels proud
[18:38:19] <alex_joni> did some machining with my 'mill' today
[18:38:36] <maddash> the main goal here isn't to optimize, it's to make the code look less silly
[18:38:43] <anonimasu> yay
[18:38:46] <anonimasu> my box came today
[18:39:20] <jepler> I sure wouldn't mind seeing the fabs() thing removed
[18:40:06] <jepler> the rest looks like a matter of style. Personally, I throw up inside my mouth when I read 'if(NULL == x)' but I accept that others write in this style.
[18:41:52] <tomp> jepler: i cant get lefty to open any dot files do far, and i've set the LEFTYPATH and moved dot file onto that path. i've turned on -V verbose and see no errs in Dotty. So, Dotty wont save and Lefty wont open :-/
[18:41:55] <lerman> I, too, have a hard time with it. But I have written if(x=NULL) when I didn't mean to.
[18:42:01] <maddash> start_time = etime()+10.0;
[18:42:05] <maddash> while (etime() < start_time &&
[18:42:24] <maddash> did that show up for you guys?
[18:42:26] <jepler> I'd rather understand why the author thought it was useful to try for ten seconds to create these buffers -- why might that fail during the first 9.8 seconds?
[18:42:38] <alex_joni> jepler: it fails sometimes
[18:42:41] <alex_joni> or takes a long while
[18:42:42] <jepler> alex_joni: why?
[18:42:44] <alex_joni> if they are remote
[18:42:51] <alex_joni> bad network?
[18:42:53] <lerman> Nope. If you are going to do that, you can't call it 'start_time'. It is actually end time.
[18:43:17] <lerman> (or stop trying time).
[18:43:32] <maddash> er, it is "start_time"
[18:43:42] <maddash> it is correctly named...
[18:43:54] <lerman> But it isn't after you add ten to it.
[18:44:08] <maddash> ah, good, so you did get it
[18:44:13] <alex_joni> lerman: current time + 10 = time needed to start
[18:44:16] <maddash> 3 lines, rah?
[18:44:22] <alex_joni> maddash: yeah
[18:44:37] <lerman> NOpe.. time allowed to start.
[18:44:45] <maddash> lerman: i didn't bother renaming the variable "start_time"
[18:44:49] <alex_joni> jepler: I agree it feels 'wrong' to just use an arbitrary numer (like 10)
[18:45:06] <lerman> I was going to point that out, but enough is enough.
[18:45:25] <pier_> pier_ is now known as pierp
[18:45:28] <maddash> who wrote this module, anyway?
[18:46:06] <lerman> Do you really want to put that in an include file? Or perhaps at the beginning of the file. So that when you read the code, you have to go search somewhere to find that ten seconds are allowed for startup (or was it tne minutes).
[18:47:59] <jepler> maddash: looks like it is a very old source file from the NIST days
[18:48:03] <lerman> jepler is listed as the author. Derived from a work by...
[18:48:24] <lerman> Let's just give jepler the credit.
[18:48:29] <jepler> maddash: it was added to the emc1 CVS back in 2000 by "wshackle"
[18:48:42] <alex_joni> probably written in 98 or earlier
[18:48:45] <jepler> while(fabs(etime() - start_time) < 10.0 &&
[18:48:53] <jepler> that version already has the fabs
[18:48:59] <alex_joni> ;-)
[18:49:04] <jepler> if the code's worked for 7 years it'll probably work for another 7
[18:49:31] <alex_joni> I'll say..
[18:49:34] <lerman> That was a long time ago. Maybe time sometimes ran backwards in those days.
[18:50:06] <alex_joni> who knows.. maybe we'll get there in the future
[18:50:11] <alex_joni> 'there' = back in the past
[18:50:36] <lerman> If you get rid of the fabs(), maybe you can get rid of #include <math.h>.
[18:51:30] <maddash> i made many more changes, but if i paste them over, i will disconnect
[18:51:39] <alex_joni> maddash: use pastebin.ca
[18:51:46] <alex_joni> or any other pastebin site
[18:55:20] <jepler> use "cvs diff -u" or "cvs diff -u filename.c" to show the differences between the original version and your version.
[19:16:17] <jackc> jackc is now known as crepincdotcom`
[19:16:46] <crepincdotcom`> damn services on large networks
[19:16:50] <crepincdotcom`> all the short nicks are taken
[19:19:26] <alex_joni> use &32*1
[19:19:31] <crepincdotcom`> heh
[19:20:02] <alex_joni> alex_joni is now known as \_
[19:20:11] <\_> :P
[19:20:21] <\_> \_ is now known as alex_joni
[19:20:27] <skunkworks> I was really supprised that 'skunkworks' wasn't taken.
[19:20:47] <alex_joni> who would want such a crappy name anyways?
[19:20:54] <crepincdotcom`> lol
[19:20:55] <skunkworks> ;)
[19:20:56] <alex_joni> :-P
[19:21:10] <alex_joni> * alex_joni likes samco better
[19:21:22] <skunkworks> it was samcoinc for a while also ;)
[19:22:02] <alex_joni> samcoinc "Your heavyweight moving expert" ?
[19:22:17] <skunkworks> riight.
[19:22:39] <skunkworks> * skunkworks take some more ibuprophen
[19:22:51] <skunkworks> 'takes'
[19:23:47] <skunkworks> I really should take some picture.
[19:24:32] <alex_joni> you should :P
[19:43:03] <maddash> s/ibuprophen/ibuprofen/g
[19:43:07] <maddash> http://pastebin.ca/410961
[19:43:15] <maddash> jepler: sorry, no cvs installed here
[19:43:39] <maddash> heh cool, pastebin automatically formatted my c++ file
[19:43:50] <skunkworks> http://www.usgyms.net/ibuprophen.htm
[19:43:59] <maddash> no, wait, nvm.
[19:46:13] <maddash> in rewriting the source, i assumed that a) each of those emc*Channel and tool*channel variables needed to return true on the valid() calls and b) the errors couldn't be resolved in 10sec, then reinstantiate a new class in a last ditch effort . I extrapolated all this from the v2.1 emcsvr.cc, so I could be wrong.
[19:46:36] <maddash> skunkworks: http://www.google.com/search?hl=en&safe=off&client=opera&rls=en&hs=cXF&q=ibuprophen&btnG=Search
[19:46:54] <maddash> the canonical way is to spell it with an "f"
[19:48:01] <maddash> also, someone should take a look at valid() in nml.cc, seems like those if's should be else-if's
[19:49:56] <jepler> maddash: all proposed changes should be submitted in "diff -u" form. the easiest way to do that is to use cvs.
[19:50:20] <jepler> maddash: "this looks fishy"-type changes will not be added to 2.1, which is receiving bugfixes only
[19:51:06] <skunkworks> maddash: not saying it is wrong.. Just glad I am not the only one ;)
[19:51:48] <maddash> so, does that mean that you won't look at the pastebin? or does it mean that i just wasted the past 5 minutes?
[19:53:39] <alex_joni> "emcErrorChannel bungalo-ed, attempt to reinstantiate" ?
[19:54:42] <jepler> maddash: I don't owe you anything, but I'm more likely to consider a proposed change when it's in a format that lets me easily see what the changes are, and if I like them to automatically incorporate them in my version of the file.
[19:54:49] <jepler> maddash: that is what "diff -u" is for
[19:56:28] <maddash> jepler: how hard is it to "diff" yourself?
[19:57:40] <lerman> I just looked at the patebin. The indentation seems foobarred.
[19:58:32] <skunkworks> jepler: go diff yourself ;)
[19:59:01] <maddash> lerman: yeah, i didn't want to break my space bar like the other guy, so i started tabbing, but it came unglued in leafpad
[19:59:12] <maddash> jepler: what should I diff against?
[19:59:55] <jepler> you have to use an editor that doesn't modify tabs, spaces, and newlines at random
[20:00:04] <lerman> Also, I noticed that you did a #define MAXTIME 10.0 but used something called MAX_TIME. You also continued to use .200 instead of #defining RETRY_E
[20:00:26] <jepler> maddash: against the file you started with, i.e., the one from the CVS tree or from the source tarball
[20:00:26] <maddash> lerman: whoops, thanks for that.
[20:01:01] <maddash> jepler: I downloaded emc v2.1_branch (think that's the name). should I diff against HEAD instead?
[20:01:06] <lerman> ^RETRY_INTERVAL^
[20:01:28] <jepler> maddash: you want to produce a file that shows the differences *YOU* added
[20:01:36] <jepler> maddash: that means using 'diff' on the original file and on your modified file
[20:02:00] <jepler> maddash: using 'diff' between a file you modified, and a file some other developer modified, will confuse the issue
[20:02:02] <maddash> jepler: you realize that I'm asking you what the original file should be, rah?
[20:02:27] <maddash> emcsvr.cc in head has the same problem
[20:02:52] <jepler> maddash: clearly I'm being unclear.
[20:03:03] <jepler> maddash: I want to have a "diff -u" format file that shows the changes *YOU* made
[20:03:18] <jepler> maddash: to do that, you run "diff -u" on the file you started with, and the file you ended up with
[20:04:52] <maddash> yes, but it's meaningless for me to diff against 2.1's emcsvr.cc, given what you've told me: 15:50:18 <jepler> maddash: "this looks fishy"-type changes will not be added to 2.1, which is receiving bugfixes only
[20:05:02] <skunkworks> wouldn't you want the latest file first before modifing it (head or trunk or whatever it is called). other wise any changes that where made between when you checked out and now would not be in there. (talking from a non programmer)
[20:05:21] <jepler> maddash: please try to understand what diff and patch do
[20:05:34] <jepler> maddash: have you ever looked at a "diff"?
[20:06:22] <lerman> So, someone could take the diffs that maddash generates and apply them against ANY version. Generally, that will do the appropriate thing -- unless the changes are too convoluted.
[20:07:04] <jepler> Here's an example of a diff: http://pastebin.ca/411001
[20:07:06] <lerman> The best thing would be for maddash to apply his changes to the current 'trunk'.
[20:07:39] <jepler> if another developer changed the file in the meantime, *but didn't change these particular parts of the file*, then 'patch' will still be able to incorporate my changes
[20:07:46] <jepler> even though they were made against a different version of the file
[20:08:38] <lerman> maddash: You probably need to become a developer and learn to use cvs, etc. That's particularly true if you are going to be making cosmetic changes.
[20:08:59] <jepler> maddash: the other thing you should do before wasting my time is make sure your change even compiles at all.
[20:09:16] <jepler> I copied your file into my emc2.1 source tree, but I get this error when I try to compile:
[20:09:19] <jepler> I copied your file into my emc2.1 source tree, but I get this error when I try to compile:emc/task/emcsvr.cc: In function 'int main(int, char**)':
[20:09:21] <lerman> People seem to be more willing to help us make changes that THEY view as important.
[20:09:22] <jepler> emc/task/emcsvr.cc:79: error: 'MAX_TIME' was not declared in this scope
[20:09:25] <jepler> emc/task/emcsvr.cc:99: error: expected primary-expression before ')' token
[20:09:28] <jepler> emc/task/emcsvr.cc:99: error: expected `;' before ')' token
[20:10:02] <jepler> "cvs diff" shows that your editor destroyed all the whitespace in the file, too
[20:10:15] <jepler> it believes that you've changed every line of the file, because the editor changed tabs to spaces or vice versa
[20:10:47] <jepler> or maybe it's pastebin's fault that there are DOS line endings, I don't know or care exactly
[20:11:14] <lerman> That is to say, if it improves the functionality, people are more likely to want to help. That's not to say that a general code cleanup wouldn't be useful and valuable. Each of us has to determine our own priorities, though.
[20:12:52] <robin_sz> didn't we agree on running stuff through some sort of code layout manager at some point?
[20:13:02] <robin_sz> to arrive at a common layout style?
[20:13:39] <jepler> no
[20:14:04] <robin_sz> I thought at one pointit was done pre-cvs checkin
[20:14:08] <robin_sz> as a script ...
[20:14:12] <jepler> never in emc2
[20:14:22] <alex_joni> jepler: actually there was a time for emc2
[20:14:28] <alex_joni> it was done by the build system
[20:14:31] <jepler> oh really
[20:14:33] <jepler> that's fucking insane
[20:14:36] <alex_joni> automatically :)
[20:14:44] <robin_sz> yeah, should be done before check in
[20:14:45] <alex_joni> indent for c files
[20:14:52] <robin_sz> right
[20:14:58] <alex_joni> I think nothing for c++
[20:15:11] <alex_joni> or the same thing.. but it had some issues with c++
[20:15:27] <robin_sz> that way you just run indoent when you check out to have it in your preferred style for display
[20:15:42] <robin_sz> and the system shoudl put ti right again before diffign it
[20:15:50] <robin_sz> theoretically ;)
[20:15:54] <alex_joni> robin_sz: yeah, but that surely breaks diff & co
[20:16:01] <maddash> http://pastebin.ca/411014
[20:16:28] <robin_sz> nope ... in theory it means it shoudl indnet it back and there shoud be no diffs ..
[20:16:31] <jepler> maddash: is this a change you made?
[20:16:31] <jepler> - IniFile inifile;
[20:16:31] <jepler> - const char *inistring;
[20:16:31] <jepler> + Inifile inifile;
[20:16:31] <jepler> + const char *inistring;
[20:16:31] <lerman> I've never seen an automatic tool that works well with long lines, comments, etc.
[20:16:40] <robin_sz> quite
[20:16:49] <jepler> maddash: (I know the answer is "no")
[20:17:01] <maddash> jepler: so wtf do you want?
[20:17:15] <robin_sz> just running indent on checkin and then encouraging people not to fuck up the formatting unless they need to is better
[20:17:18] <jepler> will someone else explain to maddash what I want?
[20:17:18] <maddash> jepler: i've ran the fucking source through `indent`, what else do you need?
[20:17:39] <alex_joni> maddash: the point is to *NOT* run it through indent
[20:17:41] <robin_sz> * robin_sz notes theres a lot of the f word flying about tonight
[20:17:44] <lerman> maddash: you are getting every line because you mucked with the spaces. See the -b option of diff.
[20:18:11] <alex_joni> maddash: when we make changes we try to keep code changes separate from beautifying changes
[20:18:25] <maddash> jepler, lerman, and whoever wants to nit-pick about indentation: read trunk's original, and note the inconsistent tabs and spaces
[20:18:39] <alex_joni> maddash: it's not a matter of right vs. wrong
[20:18:53] <lerman> Its a matter of useful and useless.
[20:19:01] <alex_joni> it's just that we don't want to change it if we don't absolutely have to
[20:19:12] <alex_joni> maddash: if there's a bug in emcsvr.cc in a couple months
[20:19:21] <robin_sz> surely, it was all run through indent at some point, no?
[20:19:34] <alex_joni> and someone starts looking for the bug.. they happen to know version 1.12 works OK, and 1.23 doesn't
[20:19:39] <maddash> robin_sz: only after the bitching about indentation started.
[20:20:02] <alex_joni> if the diff is 20 pages it's lots harder to spot the problem then if only the changes are in the diff
[20:20:22] <jepler> maddash: if you want a change made in emc, you should submit exactly that change, nothing more or less
[20:20:27] <maddash> alex_joni: i'd be glad to conform to your indentation specs, as soon as you've decided what they are.
[20:20:32] <lerman> Some editors let you use spaces and tabs and still look good. The version I'm looking at (1.11) looks just fine, and it mixes spaces and tabs.
[20:20:44] <robin_sz> maddash, however it happens, the trunk should be well indented, .. probably that shousl have happened shortly after the original commits, before real work began
[20:20:48] <alex_joni> maddash: I don't care how you indent.. I only care if you change an existing file
[20:20:50] <jepler> maddash: because you ran "indent" your diff shows a bunch of whitespace changes, which is not useful to a developer who is trying to decide if your change is correct and desirable
[20:21:06] <jepler> maddash: (also from the first file you posted, it looks like your text editor may also make unrequested changes to whitespace)
[20:21:34] <jepler> maddash: then you muddied the waters earlier by using as the other operand to "diff" some other version of emcsvr.cc than the one you started with. That file contained changes made by other developers in the meantime.
[20:22:12] <jepler> maddash: so your diff shows "remove all the changes made by the other developers between 2.1.4 and TRUNK" and "change spaces and tabs in a bunch of places" in addition to the change you actually want the developers to consider
[20:24:19] <alex_joni> re indent: let's just switch over to python :P
[20:24:50] <anonimasu> Ā§haha
[20:24:52] <anonimasu> * anonimasu punches alex
[20:24:59] <lerman> :P yourself.
[20:25:23] <alex_joni> anonimasu: ??
[20:25:30] <anonimasu> 22:28 < alex_joni> re indent: let's just switch over to python :P
[20:25:37] <alex_joni> anonimasu: and?
[20:25:43] <lerman> I actually find python fun... but hate one aspect of it.
[20:25:48] <jepler> alex_joni: anonimasu must have thought you were joking
[20:26:09] <alex_joni> anonimasu: I was half serious (maybe more than half)
[20:26:23] <alex_joni> I am astonished how efficient the code is..
[20:26:34] <alex_joni> and you're forced to indent it properly ..
[20:26:59] <jepler> lerman: only one aspect of it? you must not have used it enough
[20:27:18] <lerman> jepler: true
[20:27:34] <jepler> bbl
[20:28:08] <lerman> I have code that crashes on one machine and not another. A seg fault inside the interpreter is not very useful.
[20:29:14] <alex_joni> * alex_joni has an urge to answer an email to Michel on the mailing list
[20:29:15] <lerman> alex_joni: how do you measure efficiency?
[20:29:32] <alex_joni> "we'll take your patch, but for a 1-year subscription.."
[20:29:35] <alex_joni> lerman: code size
[20:29:50] <alex_joni> and time spent on writing that
[20:30:07] <alex_joni> (not me.. cause I just started to learn some python..)
[20:30:14] <maddash> http://pastebin.ca/411035
[20:30:34] <anonimasu> alex_joni: I were kidding
[20:30:40] <anonimasu> alex_joni: that's a good idea for the interpreter..
[20:30:49] <anonimasu> alex_joni: I remember the discussion a while back
[20:31:28] <alex_joni> maddash: the diff is way better now :)
[20:31:59] <alex_joni> one question though.. if emcCommandChannel is NULL, then emcCommandChannel->valid will barf
[20:32:23] <alex_joni> segfault I think.. null pointer dereferencing or something like that.. (I think)
[20:32:45] <maddash> alex_joni: jepler couldn't pull his head out of his arse and realize that HEAD's indentation is decidedly unpredictable (so how the fuck am i suppose to conform to that?), so i ran both the original and mine through `indent`
[20:33:08] <anonimasu> maddash: calm down.
[20:33:10] <maddash> alex_joni: emcCommandchannel is never null when valid is caled
[20:34:06] <alex_joni> maddash: you're supposed to not touch parts you don't change..
[20:34:32] <alex_joni> re ->valid().. I beg to differ
[20:34:36] <maddash> alex_joni: i was changing whole
[20:34:41] <maddash> alex_joni: lines
[20:34:42] <lerman> The way to conform to HEAD's indentation is to not change the indentation when you are making small changes.
[20:34:48] <alex_joni> the first time the loop runs it can be NULL
[20:35:08] <maddash> alex_joni: how?
[20:35:12] <alex_joni> the emcCommandChannel = new... can fail
[20:35:23] <alex_joni> if there's not enough memory to alloc that
[20:35:34] <maddash> alex_joni: then the same would hold on the original.
[20:36:49] <alex_joni> you mean the second part of the while?
[20:36:53] <alex_joni> probably..
[20:37:17] <maddash> no, i mean inside the while loop of the original.
[20:37:32] <alex_joni> anyways.. not sure why you want to waste time on something that works OK
[20:38:06] <maddash> because the original code looks ridiculous
[20:39:15] <alex_joni> guess I've seen way worse..
[20:40:00] <alex_joni> * alex_joni wanders off
[20:40:52] <anonimasu> hm..
[20:42:38] <lerman> To repeat myself: I don't know where it is from, but: "There are things worth doing that aren't worth doing right."
[20:43:07] <maddash> where's the "hal_axis" program that accepts "axis.0.pos-lim-sw-in"
[20:43:07] <anonimasu> hm, is there a wiki page on axis compensation
[20:43:45] <lerman> The existing code: (1) did what it was supposed to do (2) was very readable (3) was efficient (enough for its purpose)
[20:44:36] <lerman> Whenever I think something looks ridiculous, I pause and take a look at myself in a mirror. In this case, I would look at some of the code I had written.
[20:44:58] <maddash> whatever. i gave something better, so if you choose to muddle in something worse...
[20:46:42] <alex_joni> lerman: lol, I'll remember that one
[20:46:54] <maddash> ...that would be "ignorance" in the fullest sense.
[20:48:32] <skunkworks> anonimasu: i remember reading about it - but don't remember where, (I assume your talking about lead screw comp)
[20:49:10] <skunkworks> (for correcting inconsistancy over the length of the lead screw)
[20:49:32] <alex_joni> anonimasu: I know a bit about it.. what's the question?
[20:49:56] <anonimasu> alex_joni: pondering if you can do pitch compensation..
[20:50:18] <anonimasu> alex_joni: like if my machine travels 0.01 over due to leadscrew accuracy errors..
[20:50:58] <anonimasu> or well 0.05..
[20:51:00] <maddash> grep -r hal_link * |less
[20:51:36] <anonimasu> alex_joni: do you understand what I mean?
[20:51:41] <jepler> lerman: do you know about the "standalone interpreter"? It's relatively new in the TRUNK, and makes it easy to run the interpreter in the debugger, without the rest of emc going on at the same time
[20:51:41] <alex_joni> anonimasu: sure..
[20:51:44] <anonimasu> :)
[20:51:56] <alex_joni> anonimasu: you just need to get a good measurement of the machine
[20:52:02] <anonimasu> yeah
[20:52:03] <alex_joni> and usually it's hard to do that
[20:52:13] <anonimasu> dial gauge..
[20:52:15] <alex_joni> but if you have done that emc2 can work with the data
[20:52:30] <alex_joni> you basicly supply triplest (nominal forward reverse)
[20:52:30] <jepler> lerman: it's called bin/rs274 and if you run it without arguments it prompts you for stuff like the tool table, var file, and ngc file to use
[20:52:43] <anonimasu> alex_joni: I dont know what that means
[20:52:47] <lerman> Never heard of the standalone interpreter. I usually use the sim version.
[20:53:13] <jepler> lerman: are you working with TRUNK? if so you should have it available.
[20:53:17] <alex_joni> anonimasu: nominal position is where the machine _should_ be, forward is where the machine goes if moving forward to that position, reverse...
[20:53:21] <lerman> Then I use axis to see the error messages. I enable debugging messages to a log and use that to trace what I'm doing.
[20:53:34] <anonimasu> ah ok
[20:53:36] <alex_joni> anonimasu: this approach also gathers backlash..
[20:53:48] <alex_joni> difference between forward and reverse
[20:53:49] <skunkworks> alex_joni: so in effect it fixes backlash ?
[20:53:50] <skunkworks> right
[20:54:00] <lerman> I'm busy these days, so I haven't done anything with the interp in the last few weeks.
[20:54:21] <lerman> But I want to get back to finishing my "named oword" stuff.
[20:54:24] <anonimasu> alex_joni: really nice
[20:54:39] <anonimasu> alex_joni: the mapping is the real problem
[20:54:42] <alex_joni> anonimasu: there's an ini param to enable differential mode in the data
[20:54:54] <alex_joni> nominal d(forward) d(reverse)
[20:54:57] <anonimasu> ok
[20:55:05] <alex_joni> where the d(forward) is forward-nominal
[20:55:45] <anonimasu> * anonimasu nods
[20:55:52] <anonimasu> im not doing it, but im just curious
[20:56:18] <anonimasu> need to find a linear scale first...
[20:56:23] <anonimasu> though the market's flooeded with cheap dro's..
[20:56:35] <alex_joni> cheap != good ?
[20:57:06] <anonimasu> 0.005 accuracy..
[20:57:14] <jepler> lerman: I understand .. but feel free to ask if you ever want help running the standalone interpreter under gdb
[20:57:27] <maddash> find ./ -name motmod*
[20:57:31] <anonimasu> it's not like the heidenhain scales on the big mill..
[20:57:42] <lerman> Really. I bought a microwave oven last year for $60. It is much better than the one I bought for $400 around 1970 :-)
[20:57:47] <anonimasu> but thoose are probably worth a leg.
[20:57:48] <lerman> Thanks, jepler.
[20:58:13] <anonimasu> alex_joni: they have them down to 0.5um
[20:58:20] <alex_joni> anonimasu: I know...
[20:58:20] <anonimasu> the cheap ones
[20:58:31] <alex_joni> I'm troubled about linearity though :P
[20:58:40] <lerman> alex_joni: you are corrent cheap != good. But also cheap != bad.
[20:58:44] <alex_joni> but I guess they are way better than what we need
[20:59:08] <alex_joni> lerman: so good == bad?
[20:59:11] <anonimasu> alex_joni: I still think down to 0.001 is better then most hobbyists can do..
[20:59:26] <anonimasu> oh..
[20:59:27] <alex_joni> anonimasu: still talking mm.. right?
[20:59:36] <anonimasu> theese scales are delivered with measurement protocols..
[20:59:48] <anonimasu> individual ones..
[20:59:59] <alex_joni> anonimasu: most I've seen lately are either Modbus or rs485 (older ones)
[21:00:07] <anonimasu> theese ones are "sino"
[21:00:14] <alex_joni> yuck :D
[21:00:24] <anonimasu> yuck?
[21:00:34] <alex_joni> wth is a sino?
[21:00:46] <anonimasu> a scale manufacturer
[21:00:57] <alex_joni> ahh.. thought you're talking about the protocol
[21:01:14] <lerman> sino == chinese ?
[21:01:30] <jepler> maddash: I looked at 411035. It doesn't compile for me. I don't see the declaration of 's'.
[21:01:34] <anonimasu> yeah..
[21:01:36] <anonimasu> or japanese
[21:01:49] <anonimasu> sinus output from the scales..
[21:02:03] <anonimasu> 40um period..
[21:03:30] <maddash> jepler: {int|byte|bool} s; . it's in my edit emcsvr.cc, but diff isn't including it.
[21:03:53] <alex_joni> now that would be a first
[21:04:16] <jepler> $ diff --help | grep bugs
[21:04:17] <jepler> Report bugs to <bug-gnu-utils@gnu.org>.
[21:04:25] <anonimasu> ^_^
[21:04:42] <jepler> make sure you attach the file with the declaration of {int|byte|bool} s;
[21:05:31] <lerman> jepler: cut that out. play nice.
[21:05:46] <lerman> at least include a :-)
[21:06:43] <lerman> Remember that last year's virgin is this next year's ho^H^H major contributor. :-)
[21:07:28] <anonimasu> hm.
[21:07:35] <alex_joni> lerman: that fits jepler :-P
[21:07:53] <jepler> I'm a ho^H^H?
[21:08:06] <alex_joni> no, but a major contributor :P
[21:08:31] <lerman> Oh. Did I accidently leave that in? :-) [equivalent to did I say that out loud?]
[21:08:38] <alex_joni> heh
[21:09:14] <lerman> maddash: You probably diffed the wrong files.
[21:12:11] <alex_joni> hi michael
[21:12:39] <maddash> jepler: can you pastebin the patched emcsvr? I lost mine...
[21:13:20] <jepler> maddash: you want the diff or the working file? My working file is from TRUNK, so it won't work with your 2.1.
[21:14:02] <maddash> jepler: diff?
[21:17:05] <alex_joni> 23:29 < maddash> http://pastebin.ca/411035
[21:17:49] <jepler> http://emergent.unpy.net/files/sandbox/emcsvr.cc-maddash-patch http://emergent.unpy.net/files/sandbox/emcsvr.cc-maddash
[21:19:44] <jepler> the strange 'fabs(etime() - ...)' appears at least two other places in emc
[21:19:45] <jepler> ./libnml/cms/cms_srv.cc: if (fabs(etime() - time_of_last_key_request) > 30.0) {
[21:19:45] <jepler> ./libnml/nml/nml_mod.cc: return (fabs (etime () - Dclock_start_time) < Dclock_expiration);
[21:19:50] <jepler> there must be some motivation for it that we can't guess
[21:19:50] <maddash> jepler: while I'm at it, what source should I start from?
[21:20:35] <jepler> maddash: non-bugfix changes are only made on the CVS TRUNK.
[21:20:54] <lerman> I'm not familiar with the fn etime() . Where is it defined? (man etime doesn't work).
[21:21:05] <jepler> lerman: it's from somewhere in libnml I think
[21:21:17] <jepler> lerman: I think it returns a representation of the time or elapsed time as a floating-point number. 1.0 = 1 second
[21:21:18] <alex_joni> * alex_joni goes to bed
[21:21:24] <jepler> see you alex
[21:21:27] <alex_joni> good night all
[21:21:54] <anonimasu> night
[21:21:59] <jepler> lerman: libnml/os_intf/_timer.c
[21:22:46] <maddash> jepler: trunk_premerge7?
[21:22:53] <jepler> maddash: no.
[21:23:05] <jepler> maddash: TRUNK is what you get when you give no "-r" argument to 'cvs co'
[21:23:53] <lerman> Thanks, jepler. I was thinking it might be worth taking a look to make sure that the code was actually wrong.
[21:24:04] <lerman> (or unnecessary).
[21:24:37] <lerman> Note that it will return zero if unable to get the time.
[21:25:02] <jepler> yes but the number of reasons that gettimeofday can fail is fairly small
[21:25:20] <jepler> one reason I can think of is this: suppose that the system time is reset to be days or hours different (earlier) while one of these loops is running
[21:25:35] <jepler> e.g., the date was set to 2010. I start emc. Set the date to 2007.
[21:25:48] <maddash> jepler: i'm not using `cvs`...what RCS of emcsvr.cc is this?
[21:25:50] <jepler> without fabs() the loop will run for 3 years
[21:25:53] <alex_joni> how about running during DST changes?
[21:26:10] <jepler> alex_joni: no, DST changes don't make etime() go backwards -- gettimeofday() returns UTC
[21:26:21] <alex_joni> ntp fixes?
[21:26:33] <lerman> In that case, we can get a negative difference. If we ignore that possibility, the code can loop forever.
[21:27:00] <jepler> so the fabs() means "if 10 seconds has passed, or we went more than 10 seconds into the past because of one of ntp or a manual clock change..."
[21:27:11] <jepler> maddash: use cvs.
[21:28:01] <lerman> OK. It makes perverse sense to me. But I'll bet that the level of perversion varies all over the system, and is rarely this bad.
[21:28:23] <skunkworks> alex_joni: go to bed :)
[21:28:32] <alex_joni> * alex_joni is gone already
[21:28:41] <alex_joni> skunkworks: this is an autoresponder
[21:28:48] <skunkworks> impressive ;)
[21:28:50] <jepler> you could view it as extreme defensive programming
[21:29:00] <jepler> It would not occur to me without seeing the fabs() that you might even plan for this situation
[21:29:32] <jepler> but "gives up before 10 seconds wall time have elapsed" seems better than "tries much longer than ten seconds -- potentially for years"
[21:30:25] <maddash> why not just bypass etime altogether?
[21:30:43] <maddash> ie, {#of tries} instead of {delay time}
[21:30:51] <jepler> beats me
[21:31:59] <maddash> http://pastebin.ca/411147
[21:32:03] <jepler> but I am biased towards not touching code that works when the changed code doesn't seem to offer an advantage
[21:33:50] <jepler> time for the drive home ..
[21:34:32] <CIA-18> 03petev 07TRUNK * 10emc2/src/hal/drivers/m5i20/hostmot5_src/ (Hostm54e.pin Hostm58e.pin): -Initial revision.
[21:43:24] <maddash> then why'd you ask me to submit the changes?
[21:51:51] <jepler> maddash: because at the beginning it looked like you'd found a genuine WTF, and I thought it would be a teachable moment. By the time it devolved into "and i wouldn't personally have written the rest of this perfectly-working file in the same way" I regretted it.
[21:53:32] <robin_sz> the interent is a wonderful thing .. it saves lives
[21:54:11] <robin_sz> i get the distinct impression that being connected only by tcp/ip has saved physical harm tonight
[21:56:10] <Dallur> robin_sz: Im just glad were not programming killer robots, or wait ....
[21:56:10] <robin_sz> heh
[21:56:10] <robin_sz> * robin_sz reads the news ...
[21:56:10] <robin_sz> lets see ...
[21:56:10] <robin_sz> war
[21:56:10] <robin_sz> war,
[21:56:10] <robin_sz> oil
[21:56:10] <robin_sz> and ..
[21:56:10] <robin_sz> err ... war.
[21:56:10] <robin_sz> nothing new then
[22:02:25] <renesis> i think i broke axis gui
[22:02:45] <renesis> i hit reload instead of load, with its default splash showing, and now its stuck
[22:03:02] <renesis> which is sad cuz i did much setup just now
[22:04:54] <renesis> it has stolen control of all of gnome
[22:05:36] <cradek> is there a menu popped down?
[22:05:38] <renesis> mouse works, doesnt respond to kb
[22:05:40] <renesis> nope
[22:05:56] <renesis> dont even get capslock lights
[22:06:08] <renesis> cant ctrl+alt+bksp
[22:06:23] <cradek> ok, that's an OS crash
[22:06:28] <renesis> i have stuck gnome wait pointer
[22:06:36] <renesis> well, x crash
[22:06:48] <cradek> did you just change anything like maybe BASE_PERIOD?
[22:06:58] <renesis> wonder if the livecd install has sshd on by default...
[22:07:03] <renesis> no
[22:07:07] <cradek> no it doesn't
[22:07:07] <anonimasu> no
[22:07:09] <anonimasu> it does not
[22:07:18] <anonimasu> it's strange..
[22:07:20] <renesis> well, thats good/bad
[22:07:30] <anonimasu> the crash..
[22:07:30] <cradek> did you change anything in your config right before this run?
[22:07:36] <cradek> your emc config I mean
[22:07:35] <renesis> nope
[22:07:45] <anonimasu> I use reaload all the time
[22:07:53] <renesis> i spent about 20min jogging it around
[22:07:55] <cradek> I suppose Ctrl-Alt-F1 doesn't work either
[22:07:59] <renesis> nope
[22:08:03] <renesis> no kb input
[22:08:20] <renesis> ima write down the xy coordinate on the screen
[22:08:24] <renesis> touch off with those offsets
[22:08:28] <renesis> disable limits
[22:08:35] <renesis> i guess thatd work
[22:08:37] <cradek> is this trusted hardware or is it a new setup?
[22:08:50] <renesis> pretty trusted, new for emc
[22:09:09] <renesis> but like ive compiled stuff with it and used it for general stuff besides emc
[22:09:32] <renesis> axis usually works neat (like it alot, still getting used to it tho)
[22:10:00] <cradek> I'll be interested to hear whether you can reproduce it
[22:10:00] <renesis> mouse still works, is the weird part
[22:10:20] <cradek> I'm a little suspicious that what you did in AXIS isn't really the cause
[22:10:24] <renesis> i dont think so, all i remember doing was hitting reload button when i meant to hit load
[22:10:26] <cradek> well more than a little supsicious
[22:10:56] <jepler> it should work fine to hit 'reload' on the splash g-code
[22:11:07] <jepler> it's just a .ngc file in a special location, after all
[22:11:09] <renesis> i know ive done it
[22:11:44] <renesis> button is highlighted, all of x is stuck except mouse
[22:12:02] <cradek> hard disk light stuck on?
[22:12:16] <renesis> no is blipping sometimes
[22:12:36] <cradek> you might be lucky enough to see something in the system log after you reboot then
[22:12:57] <renesis> if i could get a terminal, but sshd is off so prob have to kill system
[22:13:01] <renesis> yeah, maybe
[22:16:16] <robin_sz> try ctrll-alt-f1
[22:16:25] <renesis> no kb input
[22:16:37] <renesis> ctrl alt bksp no good either
[22:16:46] <renesis> had to reset
[22:16:45] <robin_sz> dead then
[22:16:54] <cradek> try your serial terminal :-)
[22:17:11] <renesis> hax0r, i only had a compy since 2000
[22:17:22] <renesis> i dunno with your silly hyperterm stuff
[22:17:29] <robin_sz> ??
[22:17:55] <anonimasu> lol
[22:17:56] <robin_sz> hyperterm is a bit sucky
[22:18:05] <anonimasu> yeah, terminal.exe is better.
[22:18:08] <renesis> be like doing terms thru com ports at baudratish speeds, crazy old people!
[22:18:10] <anonimasu> "bray terminal" or whatever..
[22:18:11] <cradek> I haven't had a machine with a serial console for a while, but I had a few
[22:18:24] <robin_sz> nearly as bad as minicom
[22:18:37] <renesis> okay i just found emc2 updates
[22:18:44] <robin_sz> that is the suck
[22:18:46] <renesis> so i prob gotta reboot again?
[22:19:09] <renesis> i update brand new? i thought i update the ubuntu thing like 2 days ago
[22:19:16] <cradek> reboot for updates? hahaha
[22:19:21] <renesis> s/s/is
[22:19:28] <cradek> err, I mean, no
[22:19:33] <cradek> :-)
[22:19:36] <renesis> dont haha at me i was a gentoo whore for years id rebuild my whole system and not reboot
[22:19:38] <robin_sz> bastards .. ! someone just sent me a link worse than goatse.cx and tubgirl combined
[22:19:39] <renesis> !
[22:19:40] <jepler> renesis: emc 2.1.4 was released within the last 24 hours
[22:19:52] <renesis> okies ty
[22:20:19] <cradek> renesis: sorry, no offense meant
[22:20:23] <robin_sz> I think im going to be unable to sleep now :(
[22:20:33] <renesis> gentoo was like an imaginary cat chasing an invisible tail
[22:20:39] <renesis> cradek: is okies wasnt being serious
[22:21:24] <renesis> yeah dunno reload works fine now like before
[22:22:43] <renesis> okay so i guess ill home axis, do setup again, and hit reload for no reason, and see if it does it again
[22:22:49] <renesis> * renesis shrug'
[22:23:45] <anonimasu> yep
[22:33:15] <renesis> yeah nothing in kernel log
[22:33:28] <twice2> i was wondering if changing par port read time might eliminate the infrequent but bothersome false inputs
[22:33:40] <renesis> it has parport and stepper init stuff from 45min previous
[22:33:54] <renesis> then it has when i hot[plugged my kb after x locked
[22:34:01] <renesis> then it has the reboot stuff
[22:35:46] <anonimasu> hm
[22:36:07] <anonimasu> did you plug it into the wrong port?
[22:36:11] <anonimasu> if it's ps2?
[22:36:50] <renesis> no right port, it works now and before x locked up
[22:36:59] <anonimasu> this toolchanger design is comming along
[22:37:10] <renesis> that kernel entry is couple min before reboot, is when i unplugged the kb trying to get x to respond to it again
[22:37:17] <renesis> neat
[22:37:27] <renesis> what kind of toolholders does it use?
[22:37:28] <anonimasu> I cracked how to do the rotary motion earlier today..
[22:37:32] <anonimasu> iso30
[22:37:39] <renesis> is that big?
[22:37:40] <anonimasu> im working on the arm/swivel part..
[22:38:00] <anonimasu> 50mm
[22:38:04] <anonimasu> dia..
[22:38:10] <anonimasu> at the end of the holders..
[22:38:26] <renesis> thats too big =\
[22:38:29] <anonimasu> nah
[22:38:31] <anonimasu> it's too small
[22:38:32] <anonimasu> :D
[22:38:42] <renesis> hehe, for my mill cuz is little er16 spring collets
[22:38:53] <anonimasu> you can probably scale it..
[22:38:54] <renesis> i have to make collars for my tools or something
[22:39:39] <renesis> okay i go rezero the z, hit reload for no reason, see if it fucks up
[22:39:43] <anonimasu> ok
[22:40:32] <twice2> he's fast
[22:40:48] <anonimasu> ?
[22:41:04] <twice2> maddash
[22:41:58] <anonimasu> hehe
[22:41:59] <anonimasu> yeah
[22:42:06] <twice2> hey i just tried sheetcam, it's sweet
[22:42:06] <renesis> okay clear cellulose rolling paper kinda sucks for setting Z
[22:42:10] <renesis> i keep losing them
[22:42:53] <twice2> i touch off with .130 piece of neoprene
[22:43:19] <renesis> the papers are .001"
[22:43:30] <twice2> why
[22:43:46] <renesis> why what?
[22:44:00] <renesis> cuz thats how fat they are i didnt design them i just buy them and then sell them
[22:44:01] <twice2> why touch so close
[22:44:10] <renesis> why not?
[22:44:25] <twice2> ouch, sorry man
[22:45:40] <renesis> cuz is know theyre .001", i dont have to measure, theyre always .001"
[22:45:50] <twice2> gotcha
[22:46:02] <renesis> also smoking brand red rice paper
[22:46:10] <renesis> i dont lose those, tho...
[22:46:15] <twice2> ic
[22:49:01] <renesis> okay i did setup and hit reload and it still works
[22:49:05] <renesis> * renesis shrug
[22:51:10] <twice2> cya
[23:00:54] <RichiH> Good news everyone! We are holding a design contest for both pdpc.us and freenode.net including all art, designs and logos. Submit your mockups, logos, css or crayon drawings to design@freenode.net and win the secret prizes. You can submit as little as a logo or a full-fledged website. Full details can be found at http://blog.freenode.net/?p=28 and its comments, plus you can join #freenode-design - As always, thank you for using freenode and have a great d
[23:03:05] <anonimasu> hm
[23:04:10] <anonimasu> hm
[23:04:16] <Jymmmm> hmm
[23:04:18] <anonimasu> I'll post the image of what i've drawn..
[23:04:35] <Jymmmm> anonimasu I thought we were humming here
[23:04:47] <Jymmmm> "The EMC Tenors" lol
[23:06:10] <anonimasu> toolchanger..
[23:06:24] <anonimasu> http://imagebin.org/7853
[23:06:37] <anonimasu> the actual arm isnt complete..
[23:06:43] <anonimasu> that's the part with the 4 boltholes..
[23:08:20] <anonimasu> *getting a better picture
[23:10:19] <renesis> wow axis is neat
[23:10:36] <renesis> i can see where my tool is when its in motion (turbocnc sucked in this respect)
[23:10:53] <anonimasu> turbocnc sucks totally..
[23:10:57] <anonimasu> always..
[23:10:59] <anonimasu> :D
[23:11:03] <anonimasu> it never made nice circles :/
[23:11:10] <Jymmmm> it does for me
[23:11:37] <anonimasu> it also kept outputting too few steps..
[23:12:37] <Jymmmm> that I can't comment on
[23:12:50] <Jymmmm> I'd have nfc if it did or dind't =)
[23:14:04] <anonimasu> oh my parts got small/big..
[23:14:33] <eric_u> anyone have a copy of "using mach 3 mill" want to offer an opinion about their estop circuit?
[23:14:36] <eric_u> http://www.artsoftcontrols.com/documentation/M3M1_84-A2.pdf
[23:14:55] <anonimasu> http://imagebin.org/7854
[23:15:06] <anonimasu> that's a better view
[23:15:22] <eric_u> it's on page 12-1 or 145
[23:16:50] <eric_u> I have a theory that a 3d model is more impressive than a copy of the actual item
[23:17:23] <anonimasu> eric_u: why is that?
[23:18:12] <anonimasu> the estop circuit looks sane..
[23:19:34] <anonimasu> eric_u: it's kind of hard to just _make_ something without a good drawing :)
[23:22:39] <anonimasu> *yawns*
[23:31:03] <anonimasu> well, im going to bed..
[23:31:11] <anonimasu> as everyone else seem to have died.
[23:38:46] <eric_u> anonimasu, I'm not sure why a 3d rendering of a part is more impressive, but that's my observation
[23:39:25] <eric_u> from now on, if I have a 3d drawing, that's what I'm publishing
[23:40:32] <anonimasu> well, what else should you publish when your ideas are on the planning stage?
[23:41:11] <eric_u> I mean, if I have the choice between a 3d rendering and an actual object, I'm publishing the rendering
[23:41:22] <eric_u> before you make it, there is no choice
[23:43:51] <eric_u> it's just an observation I made recently, nothing to do with you
[23:44:33] <anonimasu> ok
[23:45:24] <anonimasu> this is a part of pain.
[23:46:13] <eric_u> what kind of tooling is that for?
[23:46:39] <anonimasu> iso30
[23:46:42] <anonimasu> it's just the changer arm so far..
[23:46:49] <anonimasu> there's a linear axis to move it up and down too
[23:47:12] <eric_u> so the d shaped part is the tool end?
[23:47:43] <anonimasu> no
[23:47:49] <anonimasu> that's just the pivot of the arm..
[23:47:55] <anonimasu> svivel..
[23:48:01] <anonimasu> that moves the arm in and out to the spindle..
[23:48:17] <anonimasu> the D shaped part is what mounts to the machine
[23:48:26] <eric_u> tool holder left as an excercise
[23:48:34] <anonimasu> it's 01:48 at night ;)
[23:48:41] <anonimasu> going to draw that another day..
[23:48:46] <anonimasu> going to check what kind of bearing I need..
[23:48:55] <anonimasu> that is like 20x too big I think..
[23:49:13] <anonimasu> probably
[23:49:18] <eric_u> actually, that is the part of a robot arm that stopped me from designing it, so I feel your pain
[23:49:31] <anonimasu> the thing with holes is where the arm mounts to the motor..
[23:49:41] <anonimasu> it's not a big pain..
[23:49:58] <eric_u> I see now
[23:50:04] <anonimasu> http://imagebin.org/7826
[23:50:10] <anonimasu> that's something like a changer arm..
[23:51:33] <eric_u> tool goes in the 50mm cutout?
[23:52:25] <anonimasu> yeah
[23:52:36] <anonimasu> it's big too, but im only doing concept stuff so far..
[23:53:02] <anonimasu> need to check out how thick material I really need and what kind of forces ther'll be
[23:53:40] <anonimasu> a nema23 electric motor will be enough to move it in/out
[23:53:41] <eric_u> bearing deflection is important, the rest will be overkill probably
[23:53:47] <anonimasu> yep
[23:54:18] <anonimasu> the motor is a nema23 motor..
[23:55:05] <anonimasu> I were thinking about pneumatic but, the geometry is a pain
[23:55:42] <anonimasu> as I need to swing it 270 degrees
[23:56:01] <eric_u> I think a motor is better
[23:56:15] <anonimasu> this is on a pretty small machine..
[23:56:30] <anonimasu> so, space is a big issue..
[23:57:15] <eric_u> pneumatic motor
[23:57:21] <anonimasu> they are very high rpm..
[23:57:41] <anonimasu> there are step cylinders, but they are hard to get by
[23:57:58] <eric_u> true, but there is one on my bridgeport for the spindle speed
[23:58:44] <anonimasu> im in sweden :/
[23:59:18] <anonimasu> though maybe I should check that out..
[23:59:34] <eric_u> are you using a stepper, or dc?
[23:59:37] <anonimasu> dc