#emc | Logs for 2010-11-27

Back
[00:04:07] <theorb> theorb is now known as theorbtwo
[01:34:43] <skunkworks> http://www.machsupport.com/forum/index.php/topic,16620.0.html
[01:36:44] <alex_joni> sad
[01:37:10] <skunkworks> * skunkworks hugs emc again
[01:46:08] <skunkworks> http://www.youtube.com/user/PomplamooseMusic#p/c/F125407272F3C1A4/0/Uolz7V12evc
[01:50:20] <skunkworks> Like this one also http://www.youtube.com/user/PomplamooseMusic#p/c/F125407272F3C1A4/2/z9KMgg7T_sg
[02:33:17] <WesBaker> Is there a configuration option to change the default filename extension for g code files from .ngc to something else in Axis under the file open dialog?
[02:41:42] <skunkworks> I think you would add a line like this to the filter section of the ini file. PROGRAM_EXTENSION = .NGC XYZ Post Processor
[02:41:58] <andypugh> He was too impatient
[02:42:02] <skunkworks> well - crap - he is gone
[02:44:27] <andypugh> He stayed for 5 mins
[02:47:38] <skunkworks> kids these days
[10:47:58] <billykid> hello
[10:48:54] <archivist> \o/
[10:50:02] <archivist> we have a bot factoid to catch users like that in #mysql
[10:52:44] <psha> * psha was too shy to say something
[10:56:10] <archivist> !wench learn hello is hello <nick>, you have a question?
[10:56:10] <the_wench> I have learnt hello is hello <nick>, you have a question?
[10:57:03] <psha> archivist: he's not greeting me :(
[10:57:15] <archivist> hello
[10:57:15] <the_wench> hello archivist, you have a question?
[10:57:23] <psha> hello
[10:57:23] <the_wench> hello psha, you have a question?
[10:57:27] <psha> hi
[10:57:29] <psha> :(
[10:57:50] <archivist> !wench learn hi is hello <nick>, you have a question?
[10:57:50] <the_wench> I have learnt hi is hello <nick>, you have a question?
[10:59:22] <psha> hi
[10:59:22] <the_wench> hello psha, you have a question?
[10:59:26] <psha> :)
[11:55:15] <jthornton> morning
[12:15:25] <ries> good morning... is it possible in Axis to turn of the preview, or at least the loaded g-code file?
[12:16:07] <ries> Currently when I load a 4MB g-code file, axis starts to consume a couple of 100MB and loading of the file takes at least 5-10 minutes... I think it's because of the preview...
[12:16:27] <jthornton> yes
[12:17:50] <jthornton> ries: http://www.linuxcnc.org/docview/html//gui_axis.html#r1_11_7
[12:18:06] <ries> jthornton: let me take a look...
[12:19:22] <ries> jthornton: I can also use that in MDI ?
[12:19:42] <ries> Hmm guess not, but I can edit the gcode file...
[14:38:44] <wobblybootie> Hi folks, a quick question from a complete novice (Machine still under construction) anyone know if this http://www.poscope.com/product.php?pid=3 will work with EMC?
[14:45:48] <archivist> for the joystick probably, for the stepper drive no
[14:47:10] <wobblybootie> yep joystick/pendant control only ... Stepper drives already taken care of ...
[14:48:01] <archivist> Im not aware of anybody using that with emc so far
[14:48:27] <wobblybootie> Thank you, I shall invest a little time and do some reading ... I'll let you know the outcome
[14:50:53] <archivist> I cant see mention of linux on their page
[14:55:28] <jthornton> wobblybootie: yes that works, I have one
[14:56:02] <jthornton> you have to set it up on a winblows machine though
[14:56:23] <wobblybootie> I have just been informed that the developer?? has no interest in LInux/EMC preferring to stick with Mr Gates and Mach3
[14:57:03] <jthornton> for jogging this is what I use on my plasma http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Simple_Remote_Pendant
[14:57:16] <archivist> tell said dev that the number of converts mach to emc has been rising lately :)
[14:57:34] <jthornton> * jthornton wanders off to be a lumberjack for a while
[14:58:16] <wobblybootie> Don't want Winsh#t if that's the case thats a lost sale ...
[14:59:48] <wobblybootie> Cheers I will have a look at that ....
[15:07:13] <dimas> hi
[15:07:13] <the_wench> hello dimas, you have a question?
[15:07:17] <dimas> :)
[15:07:30] <dimas> just got axis.ngc milled
[15:07:48] <dimas> my first mill
[15:08:20] <dimas> thanks to all helping here instantly
[15:23:48] <elmo401> dimas: congrats :)
[15:23:56] <elmo401> any pics?
[15:24:01] <dimas> elmo401, thanks
[15:27:15] <dimas> http://imagebin.org/125054 - the bot
[15:27:48] <dimas> http://imagebin.org/125055 - test mill
[15:27:52] <elmo401> beautiful :)
[15:27:58] <dimas> the text was stripped
[15:28:22] <elmo401> russian?
[15:29:19] <dimas> yep
[15:29:41] <psha> dimas: spb?
[15:29:48] <elmo401> do I see limit switches?
[15:30:12] <dimas> psha, vladimir
[15:30:35] <dimas> elmo401, i have three switched installed for homing
[15:30:41] <dimas> switches
[15:30:44] <psha> смутила улица Луначарского :)
[15:31:08] <dimas> psha, i need relogin to read russian
[15:31:30] <psha> dimas: i was confused with Luncacharsky street ;)
[15:32:14] <psha> have strong association Lunacharsky st == SPB
[15:32:20] <psha> don't know why
[15:32:24] <dimas> psha, they our common heroes for every town :)
[15:33:35] <psha> yes, like Sovetskaya street, Lenins square and so on :)
[15:33:40] <dimas> psha, phone code makes it clear
[15:34:09] <dimas> yeah soviet union forever... :)
[15:34:18] <psha> missed that...
[15:34:22] <psha> have to look on it first
[15:34:44] <psha> my grandparents are from Murom so i've spend ~50% of my time in childhood there
[15:35:36] <dimas> small world as always.. :)
[15:36:17] <elmo401> I am Estonian. ;)
[15:36:41] <dimas> elmo401, just was about to ask you :)
[15:37:12] <psha> now is time of eourpean part of community :)
[15:37:27] <elmo401> I am now in Canada. 10:37am here.
[15:37:44] <psha> logger_emc: bookmark
[15:37:44] <psha> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2010-11-27.txt
[15:38:36] <psha> there was heated discussion yesterday about using flash as GUI for emc
[15:40:36] <dimas> what's the need?
[15:41:22] <psha> there is guy who's considering from mach to emc
[15:41:28] <psha> and he has lot of flash code done for mach
[15:41:50] <psha> it seem that he don't want to trash it and start from the scratch
[15:41:50] <elmo401> flash = bloat
[15:42:12] <psha> elmo401: tell it to poppabear :)
[15:42:16] <elmo401> windows people... they don't mind buying octo-core computers just to power their OS :P
[15:42:40] <psha> i'm working on gladevcp for a month and i really impressed how Gtk is flexible
[15:43:54] <psha> for example i've drawn interface with widgets from stock gtk, emc and camunits only with glade
[15:44:03] <psha> and it works!
[15:44:05] <psha> :)
[15:48:09] <elmo401> any screen shots?
[16:07:53] <skunkworks> logger_emc: bookmark
[16:07:53] <skunkworks> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2010-11-27.txt
[16:45:37] <mhaberler> psha: for your kind review: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?GladeVcpSetup - comments esp on 'Grey areas' welcome
[17:28:38] <psha> mhaberler: here
[17:28:40] <psha> great work!
[17:28:54] <psha> mhaberler: touchy also support EMBED_TAB_* commands
[17:29:30] <psha> running gladevcp via halcmd is needed if you want to run it as separate window
[17:30:25] <psha> elmo401: it's not full functional yet but is equivalent to two camview-emc's in axis :)
[17:30:41] <psha> elmo401: but with small subset of needed controls
[17:48:59] <psha> mhaberler: i've added note about touchy + gladevcp
[17:49:28] <psha> but i'm not able to make a screenshot since i don't know how to resize windows in my WM :)
[17:50:18] <mhaberler> psha: thanks - never tried touchy; no preview :(
[17:50:19] <mhaberler> mail shot, will resize
[17:51:39] <mhaberler> psha: any use in doing loadusr gladvcp any more? can one pass the windowid that way at all?
[17:58:00] <psha> for standalone gladevcp - sure it's useful
[17:58:06] <psha> for embeded - not
[17:58:51] <psha> i've created screenshot from Xephyr with Xfce4 :)
[17:59:03] <psha> but it's too large i think
[18:00:26] <mhaberler> psha: mail it to me
[18:03:16] <psha> it's on wiki
[18:03:50] <psha> resized to 1024
[18:04:51] <psha> thanks for docmentation! )
[18:07:44] <psha> mhaberler: may you check you repo for commit 2b78027d29827fa5a2cc1a21101580dac535aaa3
[18:07:50] <psha> it's there or not?
[18:08:02] <mhaberler> psha: yes, resizing..
[18:08:32] <psha> if not - i'll kill it from gladevcp-modules barnch
[18:08:34] <psha> branch
[18:08:37] <psha> it's pretty useless
[18:08:44] <psha> or maybe not? :)
[18:09:08] <psha> http://psha.org.ru/cgit/psha/emc2.git/commit/?h=gladevcp-modules&id=2b78027d29827fa5a2cc1a21101580dac535aaa3
[18:11:52] <psha> i think i'll leave it there
[18:12:08] <mhaberler> I scaled it down a bit, its in now
[18:13:39] <mhaberler> well, not just yet,,,
[18:16:04] <mhaberler> but now
[18:16:25] <psha> nice
[18:17:56] <psha> btw any missing widgets in gladevcp?
[18:18:12] <psha> something you need in pyvcp and is missing in gladevcp?
[18:21:40] <mhaberler> you mean in terms of widgets? need to look, hold on..
[18:22:07] <psha> yes
[18:22:56] <mhaberler> I think I'm good with respect to my setup, no voltmeters or jogwheel eye candy. "just need to" migrate
[18:25:06] <psha> on your screenshot there are vertical bars?
[18:28:07] <psha> mhaberler: also it's not correct that gladevcp is in 2.4.6
[18:28:30] <mhaberler> uh. will fix.
[18:28:32] <mhaberler> which screenshot youre referring to?
[18:29:10] <psha> your screenshot with example gladevcp panel
[18:29:13] <psha> has pyvcp panel on the right
[18:29:16] <psha> with bars :)
[18:29:32] <psha> i'm editing wiki page now, i'll tell you when i'm finished
[18:31:24] <mhaberler> you mean the two under 'Spindle RPM'? couldnt that be a HAL progressbar?
[18:33:04] <psha> may be but progressbar is a bit buggy :)
[18:33:16] <psha> it's not inteded to be used as bar widget...
[18:33:32] <psha> i've added something to make it at least usable
[18:33:40] <psha> but i don't think it's good idea
[18:38:10] <psha_> psha_ is now known as psha
[18:38:15] <psha> split
[18:39:06] <psha> mhaberler: i'm finished with edits on wiki
[18:40:09] <kb8wmc_> kb8wmc_ is now known as kb8wmc
[18:40:44] <mhaberler_> anoher things which bugs me.. I'd rather have them on the side like the former pyvcp.. screen is big enough and I can still see the preview
[18:40:46] <mhaberler_> summary: optionally having it in the former pyvcp place would be great
[18:40:56] <mhaberler_> mhaberler_ is now known as mhaberler
[18:41:46] <andypugh> There seem to be fairly regular requests to add specific buttons to the left-hand pane too.
[18:42:03] <SadMan> * SadMan is gone. Gone since Fri Oct 17 17:10:00 2008
[18:42:04] <psha_> mhaberler_: i think when gladevcp will be a bit more mature it's possible to do
[18:42:21] <andypugh> (For example, not having coolant, but plasma or waterjet-related buttons instead)
[18:43:08] <mhaberler> yeah. and probing for workpiece, ref tool and current tool, for us unwashed with no toolchanges and whathaveyou.
[18:43:53] <andypugh> There are people without toolchangers? Is such a life even worth living?
[18:44:15] <acemi> acemi is now known as Guest32937
[18:44:47] <psha_> psha_ is now known as psha
[18:44:48] <mhaberler> I'm running a hobby shop here, not a metal dump ;-)
[18:45:26] <psha> andypugh: dreaming if axis was gtk based... :)
[18:45:26] <atmega> looking for ninny pics?
[18:45:31] <atmega> <urk>
[18:46:53] <mhaberler> talk to me... The tcl stuff is impenetrable for me. I dont read Postscript as well.
[18:48:40] <cradek> gtk AXIS would sure be cool. I'd be sorry to give up years worth of hard-won stability though. AXIS works pretty well now.
[18:50:49] <morfic-> morfic- is now known as morfic
[18:51:29] <psha> some major split on freenode servers..
[18:54:54] <atmega> psha: noticed lots of new options in camview/camview-emc, but it still doesn't like my cheap camera
[18:55:36] <psha> atmega: please remind me details :)
[18:55:50] <psha> lucid?
[18:56:02] <psha> and what are sympthoms
[18:56:04] <atmega> yuv422 reversed byte order, no display, only works with mplayer
[18:56:49] <psha> unplug camera and do 'sudo modprobe vivi'
[18:56:53] <psha> is it working?
[18:57:01] <mhaberler> atmega: the proper way to shop for linux-compatible webcams is the Shramov procedure: take laptop to shop and plug in cameras until you find one that works ;-)
[18:57:23] <atmega> gotta go out to the garage to do it.. doing some stuff right now.
[18:57:33] <atmega> I hate laptops
[18:57:33] <psha> ping me when you try it
[18:57:37] <atmega> ok
[18:57:43] <psha> there are some issues with rtai + opengl
[18:57:52] <psha> it may be not related to camera
[18:58:03] <mhaberler> you can always forklift your real machine ;-)
[18:58:28] <atmega> only place to buy one locally is walmart or best-buy...I'd rather get one online
[19:01:17] <psha> if it's working with mplayer it will work with camunits :)
[19:16:25] <mhaberler> psha: thinking about gladevcp integration.. if one were to add it like pyvcp it would be great to have the gladevcp stuff run in the axis process context; this would be really useful for getting an extension into axis without jumping the create-hal-pin-and-poll-it-frequently hoops; other than with pyvcp one could have real callbacks
[19:18:12] <psha> mhaberler: i don't see way to do it...
[19:18:21] <psha> gladevcp needs glib main loop
[19:20:48] <psha> that's why i'm so sad that axis is not gtk based:)
[19:23:23] <mhaberler> duh.. and a thread?
[19:24:30] <psha> python thread? :) it's a joke and not real thread... also gtk and tk will definitely have conflicts
[19:26:58] <mhaberler> well, the problem I'm referring to is really an Axis extension mechanism beyond communication through a hal pin. If there's another way to do it that's fine; the hal method makes one bend over backwards if the extension is beyond bit/float/integer i/o to/from the panel
[19:28:00] <mhaberler> once you get for instance into executing an MDI command from the extension then you get into the sync problem of two UI's
[19:28:59] <psha> you may run it directly
[19:29:54] <psha> import emc; c = emc.command(); c.mdi('G0 X10')
[19:30:01] <psha> axis is not needed
[19:30:20] <psha> i think it will be useful to create some 'EMC' widgets, not HAL ones
[19:31:14] <psha> mhaberler: btw i'm pretty close with HAL_Bar :)
[19:34:14] <SadMan> * SadMan is gone. Gone since Fri Oct 17 17:10:00 2008
[19:35:27] <mhaberler> some RPC mechanism could do that too
[19:38:18] <mhaberler> psha: 'EMC widget' might carry what I'm trying to to
[19:38:19] <mhaberler> simple example: touchoff Z at workpiece; doing that in py with MDI is simple. BUT then you need to ket axis know z reference has changed, and redraw
[19:38:34] <atmega> psha: vivi loaded, now what?
[19:39:02] <cradek> mhaberler: how would your desired function be different from the touch off that AXIS already has?
[19:39:27] <mhaberler> by running a probe command (g38.2) from a button
[19:40:02] <mhaberler> I'd rather have that canned than typing an an MDI every time. or load a new .ngc
[19:40:15] <cradek> ah - like I do in touchy (but I load a file - one click)
[19:40:20] <cradek> poke
[19:40:36] <psha> atmega: run camview-emc
[19:41:20] <mhaberler> same for ref tool and current tool touchoff on a switch
[19:41:22] <mhaberler> that's pretty much 'all' of it
[19:41:23] <mhaberler> together with the 'abort program/restart program at line X' toolchanger that would be really useful
[19:41:57] <mhaberler> cradek: and after loading the probe file,and executing it, you reload the original file?
[19:42:37] <cradek> 'loading' in touchy is a single screen poke - switching programs is no real burden
[19:42:53] <cradek> so yeah if I want to probe, I poke probe, then push the green button
[19:43:41] <psha> cradek: you have 'buttons' for all progs laying around as i remember
[19:44:20] <mhaberler> I havent tried touchy, but I assume you can load a ..ngc job to execute; when you load a probe.ngc and are done with that you reload the job you originally wanted to do?
[19:44:26] <cradek> psha: unfortunately when I converted my probing programs to O-call mdi macros, they don't work right anymore
[19:44:41] <mhaberler> ah, welcome in my pit
[19:44:58] <psha> cradek: that's bad... i thought i've fixed it?
[19:45:07] <mhaberler> O-calls are broken with probing, period
[19:45:09] <cradek> psha: they work most of the time, but sometimes fail
[19:45:20] <cradek> psha: I thought so too :-(
[19:45:20] <psha> mhaberler: it was fixed, comma :-P
[19:45:38] <mhaberler> psha: ?
[19:45:40] <cradek> psha: I wish I had a good way for you to see the problem. I'd say it works right 80% of the time.
[19:46:07] <psha> mhaberler: that was my first patch for emc :)
[19:46:31] <psha> cradek: any hints how to hit non-working case on sim?
[19:47:03] <mhaberler> psha: are you referring to o-word subs and probing?
[19:47:12] <psha> yes
[19:47:40] <mhaberler> psha: halleluja and a HUGE vodka !!
[19:47:50] <cradek> the problems I saw were: (frequently) last move after probing aborted early (final position was wrong) and (very infrequently) probe hit was missed, so the probing motion kept going, endangering probe
[19:48:11] <cradek> I am really unsure how you could see either of these in sim, sadly
[19:48:24] <cradek> psha: please visit nebraska usa... :-)
[19:49:02] <psha> i have no foreign passport (don't know how to tell it in english) and will be able to get one only next summer :)
[19:49:22] <psha> missing probe hit is extremly bad...
[19:49:28] <cradek> yep
[19:50:13] <cradek> these routines work 100% reliably when run directly, loading ngc in auto mode. I have used them for a year +
[19:50:23] <psha> how probe hit is reported to interp? maybe there is some race?
[19:50:41] <cradek> motion polls the probe input, and does tpAbort() or similar
[19:50:43] <cradek> let me find it for ou
[19:50:45] <cradek> y
[19:51:06] <mhaberler> are you guys telling me its ok to call a mdi command from py referencing a o-word probe subroutine and have that work?
[19:51:56] <atmega> psha: no luck with vivi eithe
[19:52:02] <atmega> psha: no luck with vivi either
[19:52:03] <cradek> motion/control.c:process_probe_inputs()
[19:52:29] <psha> atmega: install libgl1-mesa-swx11
[19:52:34] <psha> cradek: thanks
[19:52:42] <psha> mhaberler: it seem not
[19:54:32] <mhaberler> I'm referring to this bug: http://sourceforge.net/tracker/index.php?func=detail&aid=2847594&group_id=6744&atid=106744
[19:54:36] <cradek> psha: yes it is tpAbort() that it calls
[19:55:11] <atmega> psha: doesn't seem to help.
[19:55:20] <cradek> mhaberler: psha recently made all this work - but seems it is not quite 100%
[19:55:49] <mhaberler> the problem is some queue sync issue - interpreter proceeds whn it should wait for probe result
[19:56:20] <psha> atmega: what is your camview-emc version?
[19:56:53] <cradek> mhaberler: we know - if you can help test master, please do
[19:57:15] <atmega> latest
[19:57:15] <mhaberler> ok, will try on next shop visit
[19:57:53] <mhaberler> cradek: what was your observation - worked 'most of the time' or so?
[19:58:05] <cradek> mhaberler: please read above
[19:59:12] <psha> cradek: what is tpAbort? and why setting probe flags are done _after_ calling abort?
[19:59:13] <atmega> 0.0.33-lucid
[20:00:11] <psha> atmega: that's it. so you have libgl1-mesa-swx11 installed and camview-emc shows... what? black screen? is there anything in console?
[20:00:20] <cradek> psha: it means stop motion - set requested velocity to 0, wait for machine to stop
[20:00:38] <atmega> camview-emc shows a black screen
[20:00:57] <atmega> err... make that... it segfaults
[20:01:25] <psha> something useful in console? some error messages? do you start it without params?
[20:02:23] <atmega> Running glut init
[20:02:23] <atmega> RTAPI: ERROR: could not open shared memory (errno=2)
[20:02:23] <atmega> HAL: ERROR: rtapi init failed
[20:03:10] <mhaberler> atmega: do first a /etc/init.d/realtime start
[20:05:34] <atmega> Warning: /dev/video0 does not support VIDIOC_CROPCAP
[20:05:45] <atmega> ** (camview-emc:20634): WARNING **: Duplicate enum value 1 in control [standard]
[20:06:16] <atmega> then: Active:0... Active: 1..., Active:1... then seg fault
[20:08:49] <atmega> whole camview output in: http://pastebin.ca/2003979
[20:31:04] <psha> atmega: uninstall camunits-plugins-emc for a while
[20:31:11] <psha> it will simplify testing
[20:32:08] <psha> mhaberler: may you take a look on HAL_Bar widget?
[20:33:00] <atmega> ok.. I'm in the middle of cutting something right now.
[20:36:22] <psha> let me know when you've finished
[20:48:20] <psha> mhaberler: take a look on gladevcp-bar branch
[21:07:18] <atmega> psha: camunits-plugins-emc removed... camview-emc still segfaults
[21:07:34] <psha> with vivi and swx11?
[21:07:45] <psha> try running stock camview
[21:09:12] <atmega> stock camview, vivi doesn't turn the cam on (no LED)
[21:09:28] <psha> vivi won't turn cam on
[21:09:31] <psha> it's virtual video deviec
[21:09:32] <psha> device
[21:09:35] <psha> just for testing
[21:09:39] <atmega> oh... heh
[21:09:40] <mhaberler> psha: ah, here you are! ok, will do
[21:09:42] <psha> that everything except cam works
[21:09:42] <atmega> still blank
[21:10:02] <psha> vivi input with rgb format and opengl?
[21:11:20] <atmega> vivi input, 2000x2000 RGB 24bpp, that time it displayed something wrapped in the top 1/5th of the output
[21:12:17] <psha> what's your version of libcamunits6?
[21:13:19] <atmega> 0.2.0-7~lucid
[21:14:03] <psha> hm, i've fixed vivi in that version already...
[21:14:29] <psha> wait, 2000x2000 means that try_set_fmt is disabled
[21:14:34] <atmega> err... 0.2.0-5~lucid for plain libcamunits6
[21:14:40] <psha> upgrade it
[21:16:27] <atmega> cool, that makes camview-emc look better
[21:17:59] <atmega> hey, it makes the cam work too!
[21:18:53] <psha> :)
[21:27:05] <atmega> * atmega salutes psha
[21:28:12] <mhaberler> psha:works great! minor nit: hard limit on max 30.0 is a bit low for rpm ;-)
[21:29:50] <psha> mhaberler: it's adjustable from glade
[21:30:27] <psha> mhaberler: i'm cooking second version with both VBar/HBar's
[21:30:58] <mhaberler> psha:in normal HAL Bar properties its a hard limit on 30?
[21:31:14] <psha> oops
[21:31:20] <psha> it have to be MAX_INT
[21:32:15] <mhaberler> psha: make negative hard limit MIN_INT (ccw spindle = regative speed)
[21:33:32] <psha> sure
[21:45:44] <psha> mhaberler: check this one
[21:45:51] <psha> i've added HAL_VBar
[21:46:22] <psha> and updated gladevcp-test.ui
[21:47:33] <psha> i feel that yellow/red zones will be useful?
[22:14:19] <AeroTec> hello
[22:14:20] <the_wench> hello AeroTec, you have a question?
[22:14:25] <AeroTec> yes
[22:14:34] <psha> :)
[22:14:36] <AeroTec> lathe encoders and index
[22:14:40] <psha> archivist: working :)
[22:14:45] <AeroTec> I am running mach
[22:14:55] <AeroTec> it uses index only
[22:15:17] <AeroTec> I want to run EMC but have no encoder on my lathe yet
[22:15:37] <AeroTec> can I do threading with index only with EMC?
[22:16:01] <cradek> yes, but it will not work as well as with an encoder
[22:16:22] <cradek> you have to read your pulse with the software encoder module, then use its interpolated position output
[22:16:47] <AeroTec> LOL, I am using Mach, I would think it would work better then Mach I would hope
[22:16:56] <cradek> if you want to cut threads and have them turn out right, you should fit an encoder
[22:17:04] <cradek> why play games
[22:17:42] <AeroTec> connecting a encoder would not be easy
[22:17:51] <atmega> what count encoder would be sufficient?
[22:18:04] <AeroTec> is there examples of connecting encoders on lathes?
[22:18:31] <cradek> a few hundred lines is the sweet spot for reading it with the software module
[22:18:46] <cradek> if you have a hardware counter like mesa/ppmc/etc you can use as many lines as you want
[22:19:21] <cradek> AeroTec: you can get hollow shaft encoders but they are expensive. cheaper (not easier) is to use a 1:1 ratio toothed belt
[22:19:50] <cradek> some people make their own encoder wheels (large, with a large hole for the spindle) because neither option is all that great otherwise
[22:20:07] <AeroTec> also I am running a P port
[22:20:18] <AeroTec> can one connect a encoder to it?
[22:20:22] <andypugh> EMC2 will work at least as well as Mach with index-only, but it has the option of working a lot better (for example, stopping if the spindle stops, or coping with inconstant spindle-speed (interrupted-threading?)
[22:20:45] <cradek> http://www.wallacecompany.com/cnc_lathe/HNC/00011-1a.jpg
[22:20:51] <andypugh> You need 2 more pins, but yes, encoders and p-ports works
[22:20:57] <cradek> here is an example of a 1:1 geared kit encoder with a toothed belt
[22:21:12] <atmega> just 2 pins?
[22:21:20] <cradek> 3 pins because you need index
[22:21:21] <andypugh> He already has an index
[22:21:30] <cradek> right
[22:22:14] <andypugh> http://picasaweb.google.com/lh/photo/_hlZumHKbsbV7YoKPcEmOQ?feat=directlink
[22:22:31] <andypugh> home-made encoder ring direct on the spindle.
[22:23:00] <cradek> another? http://members.shaw.ca/swstuff/spindle-encoder.html
[22:23:12] <cradek> google finds a lot of information about emc2 + spindle encoders
[22:23:19] <atmega> would a 2 line encoder be any better than index only (same, but on 2 sides)?
[22:23:34] <cradek> andypugh: that looks nice
[22:24:45] <andypugh> I guess you could use an encoder in counter-mode with just one extra pin
[22:24:47] <andypugh> ?
[22:24:59] <cradek> yes, if you didn't want to tap, that'd be fine
[22:25:18] <cradek> but why not do it right the first time and give yourself the ability to tap
[22:25:21] <andypugh> And, in fact, could use a handy gear in the headstock and a prox
[22:25:30] <cradek> (easy for me to say since I don't have to do the work)
[22:25:58] <cradek> yes counter mode is a big improvement over 1ppr
[22:26:17] <cradek> AeroTec: have any pics of your lathe/headstock?
[22:26:55] <andypugh> AeroTec: Do you have any gears permanently on the spindle? You could use those as an encoder ring, you just need to offset the detectors by N + 1/2 pitches to get quadrature.
[22:27:17] <atmega> I have some 750 encoders
[22:27:19] <cradek> assuming you can get close to a 50% duty cycle with the proxes
[22:27:28] <AeroTec> it is a grizzly
[22:27:57] <andypugh> I am guessinf pretty much identical to mine then?
[22:28:04] <AeroTec> belt drive witht he extra gear down greas in the head stock
[22:28:44] <AeroTec> gears
[22:29:17] <andypugh> My very first attempt at an encoder (which worked fine) was a laser-printed strip of paper wrapped round the spindle between the bearings. I had some reflective photo-detectors looking at it.
[22:33:24] <AeroTec> cool
[22:34:09] <AeroTec> inside the head stock it gets very dirty
[22:34:36] <AeroTec> belt wear and oil
[22:35:02] <AeroTec> would the paper now get realy messed up?
[22:35:14] <AeroTec> not
[22:35:40] <AeroTec> need to proof read before hitting enter
[22:37:09] <AeroTec> so EMC will work as good as mach with index only
[22:37:34] <AeroTec> and that I can add the encoder later
[22:37:41] <AeroTec> cool
[22:38:01] <AeroTec> also, emc can do 9 axis
[22:39:54] <AeroTec> the manual talks about some running in perailel
[22:40:38] <AeroTec> can EMC run 9 seprate axis that move with out other axis moving?
[22:41:00] <andypugh> Yes, at the moment you can have 9 cartesian/rotary axes and 9 motors.
[22:41:43] <AeroTec> and control one at a time, none slaved to the other
[22:41:45] <andypugh> But that is a historical limit which is being worked on, the plan is 9 cartesian/rotary axes and as many motors as it takes to move them
[22:42:14] <AeroTec> cool
[22:43:07] <AeroTec> I want to make some robot arms, is there some cool way to controle them and gen code to get them to move and do what you want?
[22:43:11] <andypugh> EMC2 allows you to use kinematics modules that separate the cartesian movements of the controlled point from the movements of the motors required to achieve that movement.
[22:43:18] <andypugh> Absolutely
[22:44:04] <X704> X704 is now known as WalterN
[22:44:31] <andypugh> This is a Puma-style robot running the standard EMC2 G-code splash-screen, with kinematics to convert elbow angles to XYZ: http://www.youtube.com/watch?v=80YhX73DuSg
[22:44:36] <AeroTec> so if I made a arm and wanted to control it and or program it, what is the best way to do that?
[22:45:09] <cradek> wow that's a bit of an open-ended question
[22:47:01] <andypugh> If you are willing to learn about (or know about) Denavit Hartenberg parameters then EMC2 has a kinematics module that will work an arbitrary robot. But you might not always want to program a robot in cartesian space.
[22:51:16] <andypugh> Following links blindly, This one has one heck of a rack of elctronics! http://www.youtube.com/watch?v=pX5uYjo6AsA&feature=related
[22:51:25] <AeroTec> I want to make arms to load and unload from lathe and mill
[22:52:04] <AeroTec> any advice you can give me would be great
[22:52:42] <AeroTec> I would build them my self so I can get the inferface drivers that would best do the job
[22:53:07] <andypugh> It probably makes sense to pick up a dead industrial robot with a duff control and retrofit.
[22:53:10] <AeroTec> all I need is to be pointed in the right direction
[22:54:35] <psha> mhaberler: i've added color zones with gradients to HAL_Bar
[22:55:18] <mhaberler> psha: back from dinner and will check.. can we talk later about this widgets idea?
[22:56:04] <psha> hm, i'll be asleep after your dinner :)
[22:56:32] <mhaberler> psha:fish+rice so brain stays intact
[22:56:41] <andypugh> AeroTec: EMC2 can control most interfaces, so you have a pretty wide choice of hardware. USB is a "never", SPI is a "not yet" and Modbus is an "experimental" bit anything driven by PWM, pdm, frequency or bits can be used.
[22:56:43] <AeroTec> so casting parts and making one would not be a good thing to do?
[22:56:53] <psha> mhaberler: it's ~02:00 here :) nothing will help :)
[22:56:57] <Connor> anyone have any info on modding a Bosh Colt so that EMC can control the RPM?
[22:57:15] <psha> mhaberler: in brief: bar maybe split into color zones (up to 3)
[22:57:27] <psha> default colors are green - yellow - red
[22:57:30] <andypugh> Casting parts and making one sounds like a great project. If that is what you want to do, then have fun.
[22:57:59] <mhaberler> I see. Have a good one. Not urgent.
[22:58:01] <mhaberler> That probing bug fix is a bit of a gamechanger because mdi+oword allows a lot to happen
[22:58:20] <andypugh> Connor: It's a universal motor with an integral PWM controller?
[22:58:30] <psha> mhaberler: btw cradek pushed that in master
[22:58:33] <AeroTec> thought maybe I might cast extra and sell some for other to play with as well
[22:59:07] <mhaberler> see it, thnx
[22:59:21] <mhaberler> thats where that one belongs *now*
[22:59:26] <Connor> It has a slow-startup speed controller on it.. I think the motor is a universal though.. It has no perm magnets in it..
[23:00:05] <andypugh> Might be easiest to turn it to max at the unit and control the input voltage
[23:00:08] <AeroTec> I do not need large industrial arms, just some smaller ones for moving parts
[23:00:19] <joejaxx_> joejaxx_ is now known as JuniperJaxx
[23:00:24] <psha> mhaberler: so take a look on bar after dinner :) if it's ok i'll merge it to gladevcp-modules
[23:00:41] <psha> and it'll be in master on next gladevcp merge
[23:00:43] <Connor> It's AC, I think you loose major torque when you drop voltage with it..
[23:00:53] <mhaberler> psha: will do and let you know. have a good one.
[23:02:03] <andypugh> Connor: Do you have a VFD you could try it on? I suspect that would work poorly. They run perfectly well on DC though, so perhaps one of the generic Mini-mill DC drives?
[23:02:12] <AeroTec> what is the best way to go to connect computer to CNC, not high end but best bang for the buck?
[23:02:27] <Connor> No, I don't have a VFD.
[23:02:38] <andypugh> AeroTec: Probably the Mesa 7i43
[23:02:39] <skunkworks> AeroTec: that kind of cnc?
[23:02:58] <AeroTec> lathe, mill and Arms
[23:03:23] <AeroTec> want to add active tooling and milling to lathe as well
[23:03:33] <AeroTec> so lots of axis of control
[23:04:43] <andypugh> Connor: I think one of these would _probably_ drive your router. Bit of an expensive experiment though.
[23:04:44] <andypugh> http://www.littlemachineshop.com/products/product_view.php?ProductID=1211&category=6
[23:05:12] <andypugh> Connor: Cheapest is probably a stepper and cam onto the trigger :-)
[23:05:29] <Connor> Holy crap, that's more than my router costs.. :)
[23:06:31] <andypugh> http://www.cnczone.com/forums/diy-cnc_router_table_machines/101823-controlling_speed_bosch_1617_via_cnc.html
[23:07:35] <andypugh> But, if you do it according to that thread, you are probably going CNC-PWM-Voltage-PWM-Voltage-Router
[23:07:44] <AeroTec> On my lathe I have 2 gears I am no longer using in the head stock. what would be the best way to make that a encoder?
[23:08:04] <AeroTec> On my lathe I have 2 gears I am no longer using in the head stock. what would be the best way to make that a encoder?
[23:08:17] <AeroTec> oops
[23:08:26] <andypugh> Possibly like Jon did on his Bridgeport.
[23:08:32] <andypugh> Let me google
[23:09:07] <skunkworks> I bought 6 of those gear tooth sensors - I have not tried them yet
[23:09:32] <andypugh> http://pico-systems.com/bridge_spindle.html
[23:09:55] <Connor> "I can positively tell you that I own a HF speed controller, and a Bosch Colt with the slow start feature, and the HF speed controller does not work with the Colt."
[23:10:35] <andypugh> My concern is that they don't count at very low speeds. But I think Jon is ridgid-tapping with that setup so perhaps it is not that important
[23:10:45] <Connor> I've had it apart.. If I could figure out how to bypass the internal control circuit.. I would be good.
[23:11:33] <andypugh> Can you put a scope on the internal circuit? There will be a PWM somewhere, with any luck it will be a conveniently low voltage too.
[23:13:00] <andypugh> Simplest, is a big power mosfet linked to an EMC2 PWM output, then a big cap and rectifier to give you 120VDC and connect direct to the motor winding.
[23:24:31] <Connor> Looks like they're saying you by-pass the speed controller by shorting the 3 pins on the TRIAC
[23:29:00] <andypugh> Maybe drive the TRIAC directly?
[23:30:04] <Connor> I dunno..
[23:33:51] <andypugh> I was not entirely joking when I talked about a little stepper onto the manual controls.
[23:34:20] <Connor> yea. I know.. that's just a bit much.. I think by-passing the speed control is a better idea.
[23:37:44] <Connor> http://www.cnczone.com/forums/diy-cnc_router_table_machines/112658-super-pid_new_low-cost_router_speed_controller-12.html
[23:39:25] <andypugh> These guys spend far too much money
[23:39:52] <andypugh> Why use a "Super PiD" when you can do it all in HAL?
[23:39:55] <Connor> yea. I'm all about doing cost effective.. Looking to see what can drive the TRIAC direct
[23:40:05] <Connor> I think they're using Mach
[23:40:31] <andypugh> Surely Mach does arbitrary PWM too?
[23:41:26] <andypugh> Hmm, a TRAIC switches off when the input is greater than the control pin? (time for Wikipedia)
[23:42:53] <skunkworks> I thought the pwm out of mach was in the HZ range..
[23:42:59] <andypugh> So the voltage on the control pin is variable 0-240V DC?
[23:43:51] <andypugh> Or 110V perhaps in the US?
[23:44:46] <Connor> dunno yet. still looking into it.
[23:47:46] <andypugh> You can get opto-couplers that will control 5000V with a 5V logic supply, so that is eMC2 PWM to mains voltage PWM in a $0.32 chip.
[23:48:36] <Connor> One thing I've read is.. the TRIAC doesn't cut off till the AC on the inductive side hits 0 (from the AC wave form)
[23:48:52] <Connor> So, the PWM might need to be timed with the AC Current?
[23:49:28] <Connor> I dunno, I may have miss read or not understood what it was saying..
[23:55:41] <andypugh> No, that is how it is. You would need to filter the PWM (on the HV side) to create a control voltage.
[23:56:09] <andypugh> ie, 5V PWM -> 110V PWM -> 0-110v DC
[23:56:20] <Connor> http://www.onsemi.com/pub/Collateral/AND8011-D.PDF
[23:57:34] <andypugh> Yes, not sure how they get <50% duty cycle
[23:57:50] <andypugh> I am not very good at electronics
[23:58:41] <Connor> I've already made a RPM encoder for my router..
[23:59:08] <Connor> Mine just goes on the end of the spindle, no internal mods needed