#emc | Logs for 2005-12-27

[00:43:04] <Jymmm> Jymmm is now known as Red70sShow
[00:43:04] <Red70sShow> Red70sShow is now known as Jymmm
[00:44:10] <Jymmm> Jymmm is now known as Red70sShow
[00:48:00] <Red70sShow> Red70sShow is now known as Jymmm
[02:13:02] <Jymmm> Jymmm is now known as Red70sShow
[02:13:02] <Red70sShow> Red70sShow is now known as Jymmm
[02:35:55] <skunkworks> cradek:
[02:37:08] <skunkworks> the apt-get install xlibmesa-dev libxmu-dev on your install section of you site needs to be run sudo.
[02:37:22] <skunkworks> unless I am doing something wrong ;)
[05:02:31] <CIA-1> 03cradek 07simple_tp * 10emc2/src/emc/kinematics/ (tc.c tc.h tp.c tp.h): beginning simplification: remove unused stuff
[05:39:03] <CIA-1> 03paul_c * 10emc2-auto/wiki/ (5 files in 4 dirs): "Auto update wiki from a cron job. Tue Dec 27 05:30:01 GMT 2005 "
[06:14:33] <K`zan> Night folks
[06:50:47] <Jymmm> Jymmm is now known as Red70sShow
[06:50:47] <Red70sShow> Red70sShow is now known as Jymmm
[15:02:41] <skunkworks> cradek ?
[15:03:03] <skunkworks> never mind - looks like you changed your site. thanks
[15:05:05] <jepler> changed his site?
[15:09:09] <skunkworks> used to be "apt-get install xlibmesa-dev libxmu-dev" now "sudo apt-get install xlibmesa-dev libxmu-dev"
[15:15:49] <skunkworks> installed bdi - emc2 and axis on a new computer last night
[15:16:54] <jepler> oh, on the axis website?
[15:17:33] <skunkworks> yes - sorry wasn't clear
[15:17:54] <skunkworks> the only other thing I had to do was "sudo apt-get install python2.3-dev python2.3-opengl"
[15:18:18] <jepler> I'll add those packages to the list
[15:19:03] <skunkworks> was going to right something up but I am really not a linux person. between the wiki page and cradeks site it works
[15:19:26] <jepler> good
[15:19:32] <skunkworks> so I downloaded the latest axis and extracted it.
[15:19:42] <skunkworks> sudo apt-get install python2.3-dev python2.3-opengl
[15:19:55] <skunkworks> sudo apt-get install xlibmesa-dev libxmu-dev
[15:20:12] <skunkworks> then changed to the axis directory and ran
[15:20:25] <skunkworks> sudo env EMCROOT=/path/to/emc2 python setup.py install
[15:20:37] <skunkworks> with the correct path installed
[15:21:49] <skunkworks> think that is all I did - pretty smooth
[15:21:53] <djb_rh> is a 500Mhz celeron gonna be too painfully slow for EMC?
[15:22:22] <skunkworks> we are running it on a 450 with axis and it runs ok for steppers
[15:22:39] <skunkworks> pentium II 450
[15:22:44] <djb_rh> oh
[15:22:45] <djb_rh> cool
[15:23:12] <djb_rh> found a good deal on a used 1U rackmount box
[15:23:20] <djb_rh> hoped it would work
[15:23:50] <jepler> cradek is now running his stepper machine on a 667MHz Pentium II
[15:24:07] <jepler> his newer, faster machine stopped working
[15:24:19] <djb_rh> do servo machines require more hp or something?
[15:24:25] <jepler> it's OK but AXIS is not as responsive as it was on an Athlon 2700+
[15:24:52] <jepler> I haven't ever worked with a servo machine
[15:25:26] <djb_rh> well, the way both of you specified steppers explicitly made me wonder
[15:25:52] <jepler> steppers are where all my experience is
[15:26:06] <skunkworks> never used servos either yet.
[15:26:26] <djb_rh> ah
[15:26:28] <djb_rh> okay
[15:26:36] <djb_rh> well, the first machine I'll be building is stepper based
[15:26:38] <jepler> servo may actually need less computing power; the "PID loop" runs from 1kHz to 10kHz, while steppers might run up to 40kHz step rate.
[15:26:44] <djb_rh> but ultimately I'll be doing a servo machine, too
[15:27:07] <jepler> (frequency of the realtime task as controlled by PERIOD)
[15:28:16] <djb_rh> need to stumble across a small rack for cheap now
[15:29:09] <jepler> why go rackmount?
[15:29:25] <djb_rh> because I found rackmount stuff cheap and it's cool
[15:29:27] <djb_rh> no good reason
[15:29:41] <djb_rh> I picked up a 15" rackmount LCD for $45
[15:29:46] <djb_rh> which is about the going rate for a "normal" one
[15:30:12] <djb_rh> most of my machine stuff is staying semi-portable right now, so mountable stuff like that is good
[15:30:12] <jepler> used?
[15:30:20] <djb_rh> actually, it was new at a surplus place
[15:30:29] <jepler> oh nice
[15:30:33] <djb_rh> but it isn't one of the nice flip up ones that takes 1U
[15:30:35] <djb_rh> it's a full 8U
[15:30:52] <djb_rh> but I want it up all the time anyway for this application
[15:30:59] <djb_rh> unlike a server room where you only need it occasionally
[15:31:04] <jepler> yeah
[15:31:22] <jepler> not one of those situations where you need the screen only when something's wrong
[15:31:23] <cradek> jepler: my machine is a PIII 667 (not PII 667)
[15:31:54] <djb_rh> my mini-mill and mini-lathe are on a butcher block workbench that's got two wheels on it on one end...if you can lift the other, it's portable. :)
[15:32:47] <djb_rh> so then I was gonna stick a set of those bread-rack shelves (chrome wire things) that are on industrial casters behind it for "stuff", including the PC
[15:33:12] <jepler> cradek: oops, sorry
[15:33:24] <jepler> I wonder if II vs III makes any real difference
[15:33:31] <cradek> I'm not offended, but it might make a difference to them
[15:33:44] <djb_rh> well, a 500Mhz celeron is still fairly comparable to a PIII 667, I'd guess
[15:33:59] <djb_rh> I'm gonna see when I get the machine if I can upgrade the CPU at all
[15:34:08] <djb_rh> if so, another CPU for it is gonna be pretty cheap
[15:34:22] <djb_rh> since at most it'll be a couple hundred megahertz upgrade
[15:34:34] <djb_rh> the box already has 512M of RAM, which is good
[15:36:38] <skunkworks> I am able to run a period of .00004 pretty good. that is a maximum of 12500 pulses (maximum of 300 ipm on my x-y axis and 75 ipm on my z)
[15:38:08] <skunkworks> cradek - did you see my post above?
[15:38:42] <cradek> which?
[15:38:58] <skunkworks> how I installed axis on the bdi 4.3
[15:39:05] <cradek> great
[15:39:10] <skunkworks> <skunkworks> so I downloaded the latest axis and extracted it.
[15:39:11] <skunkworks> <skunkworks> sudo apt-get install python2.3-dev python2.3-opengl
[15:39:11] <skunkworks> <skunkworks> sudo apt-get install xlibmesa-dev libxmu-dev
[15:39:11] <skunkworks> <skunkworks> then changed to the axis directory and ran
[15:39:11] <skunkworks> <skunkworks> sudo env EMCROOT=/path/to/emc2 python setup.py install
[15:39:11] <skunkworks> <skunkworks> with the correct path installed
[15:39:12] <cradek> are the instructions pretty much right?
[15:39:13] <skunkworks> <skunkworks> think that is all I did - pretty smooth
[15:39:17] <jepler> cradek: I made a few changes to the instructions
[15:39:23] <cradek> thanks
[15:39:23] <jepler> cradek: and removed the inflamatory "dead end" statement
[15:40:34] <cradek> the emc makefiles sure make me sad
[15:44:44] <skunkworks> cool - didn't know you could string the packages togather. <- linux newbie
[15:50:47] <jepler> yeah, "apt-get install" can take as many package names as you like
[15:51:03] <skunkworks> learn something new every day.
[15:51:24] <skunkworks> It is nice having it all on your page now.
[18:00:02] <rcsu> hi all
[18:00:30] <rcsu> is someone awake who can give me a helping hand for emc2 ?
[18:01:19] <rcsu> it wont fire up cause some shm and sem werent created
[18:27:43] <skunkworks> what version of linux are you installing it on
[18:28:29] <rcsu> gcc 3.3.4 under 2.6.10
[18:29:26] <skunkworks> ok - I only have experience installing it on the 4.3 bdi install
[18:29:47] <rcsu> skunkworks: thanks
[18:29:51] <rcsu> hi alex_joni
[18:29:54] <alex_joni> hello
[18:30:16] <rcsu> the current cvs emc2 wont compile on my box
[18:30:34] <rcsu> the stuff which handles the shm and semaphores are broken
[18:31:26] <alex_joni> what box?
[18:31:50] <alex_joni> shm and semaphores are handled by wrappers to RTAI or RTLinux
[18:31:55] <alex_joni> so maybe there's the problem
[18:32:03] <alex_joni> that code hasn't changed in the last year
[18:32:41] <rcsu> i dont think so, the somewhat old cvs (some months ago) ran til that update
[18:33:05] <alex_joni> ok, describe the actual problem..
[18:33:27] <rcsu> what ive seen so far, the shmget isnt invoked with IPC_CREAT
[18:33:44] <rcsu> can i paste some lines here (6 or so)
[18:33:48] <alex_joni> sure
[18:37:35] <rcsu> mmt, to make sure to get the right responses, i have to take out my changes
[18:40:31] <alex_joni> ok.. can we do this a bit later?
[18:40:35] <alex_joni> I have to run for a bit
[18:40:40] <rcsu> sure
[18:40:40] <alex_joni> back in 1-2 hours
[18:40:44] <rcsu> k
[18:57:26] <rayh> rayh is now known as rayh_away
[20:01:41] <rayh_away> rayh_away is now known as rayh
[20:34:34] <alex_joni> rcsu: still around?
[20:38:12] <rcsu> alex_joni: yep
[20:38:41] <alex_joni> ok.. you may paste now :P
[20:39:02] <rcsu> cvs -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/emc login
[20:39:02] <rcsu> cvs -z3 -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/emc co emc2
[20:39:09] <rcsu> ups
[20:39:23] <alex_joni> authentication failed
[20:39:24] <rcsu> not what i meant
[20:39:43] <rcsu> sizeof(emcmot_comp_t): 6164
[20:39:43] <rcsu> emctask.cc 270: interp_error: Unable to open file
[20:39:44] <rcsu> Unable to open file
[20:39:44] <rcsu> Interpreter stack: - Interp::restore_parameters - Interp::init
[20:39:44] <rcsu> emctaskmain.cc 2548: can't initialize interpreter
[20:39:44] <rcsu> _shm.c 238: shmget(1001(0x3E9),8192,0) failed: (errno = 2): No such file or directory
[20:39:46] <rcsu> _shm.c 257: No shared memory buffer exists for this key and the IPC_CREAT was not given.
[20:39:48] <rcsu> cms_cfg.cc 908: cms_config: -4(CMS_NO_MASTER_ERROR: An error occured becouse the master was not started.) Error occured during SHMEM create.
[20:39:51] <rcsu> nml.cc 369: NML: cms_config returned -1.
[20:40:20] <rcsu> thats better :)
[20:40:24] <alex_joni> rcsu: that's easy ;)
[20:40:30] <alex_joni> it can't find the emc.nml file
[20:40:38] <alex_joni> configuration stuff has been moved around
[20:40:47] <alex_joni> and probably you have your own version of the .ini
[20:41:05] <alex_joni> basicly you need to have '../common/emc.nml' as the NML_FILE in your ini
[20:41:19] <alex_joni> check a recent inifile for changes
[20:41:20] <rcsu> mmt, let mecheck the nml setting
[20:41:22] <alex_joni> and ask
[20:41:35] <Imperator_> Hi alex
[20:41:41] <alex_joni> hi Martin
[20:41:49] <rcsu> NML_FILE = common/emc.nml
[20:42:05] <rcsu> and the file is in configs/common
[20:43:00] <rcsu> PCSSol:/home/su/CAD-Fraese/emc2 # ls -l configs/common/emc.nml
[20:43:00] <rcsu> -rw-r--r-- 1 root root 5210 Dec 11 23:40 configs/common/emc.nml
[20:43:37] <alex_joni> where is your ini?
[20:43:46] <rcsu> in configs
[20:44:03] <rcsu> called with scripts/emc configs/rcsu1.ini
[20:44:37] <rcsu> cwd is emc2
[20:45:36] <alex_joni> you are running as root?
[20:45:42] <rcsu> sure
[20:47:11] <alex_joni> hrmm.. strange
[20:47:52] <rcsu> IMHO its all due to the lack of IPC_CREAT options to the shm and sem calls
[20:48:09] <rcsu> the dont have any harm if the keys exists
[20:48:24] <rcsu> but create em if they werent there
[20:48:58] <rcsu> i put IPC_CREAT into the shm alloc
[20:49:07] <rcsu> and that did that error
[20:49:31] <rcsu> but the sem allocation is somewhat mor complicated
[20:49:53] <rcsu> and i dunno why this changed at all
[20:50:11] <alex_joni> strange
[20:50:17] <alex_joni> can you give me a date that works?
[20:50:29] <rcsu> sorry, no
[20:50:46] <rcsu> how do you test this stuff
[20:51:11] <rcsu> its not quite old (the change)
[20:51:27] <alex_joni> well.. you take a new dir
[20:51:32] <alex_joni> and checkout from CVS again
[20:51:41] <alex_joni> maybe using -d 'date' for an older version
[20:51:45] <rcsu> mt
[20:51:47] <rcsu> mmt
[20:54:46] <rcsu> grrr, how can i get access to the cvs repo ?
[20:55:20] <rcsu> and please update the linuxcnc.org site :)
[20:56:51] <alex_joni> rcsu: are you registered at sourceforge?
[20:58:20] <rcsu> alex_joni: nope, getting an account now
[21:10:29] <rcsu> oh man, what conplicated procedure on sf.net, but now i have one
[21:11:03] <rcsu> alex_joni: so now i can co with rcsu as account ?
[21:11:35] <alex_joni> rcsu: you need to get added as a developer to check out that way
[21:11:46] <alex_joni> for that you need to drop an email to the developers list stating who you are, what you want to do
[21:12:53] <rcsu> alex_joni: so why did you ask for an account in sf ?
[21:13:15] <alex_joni> because you need that to get added to emc
[21:13:26] <rcsu> alex_joni: i know co's are mostly anonymous possible
[21:13:28] <alex_joni> there is also some description you can enter (what skills you have, etc)
[21:13:38] <alex_joni> anonymous for non developers
[21:13:44] <alex_joni> and non-anonymouse for developers
[21:13:56] <rcsu> thats clear
[21:14:41] <alex_joni> but non-anonymous means you can also commit
[21:14:43] <alex_joni> change, delete, etc
[21:15:47] <rcsu> ok, so you want to convince me to debug that stuff ?
[21:16:25] <alex_joni> dunno.. if you like.. but that's not necessary
[21:16:43] <alex_joni> anyways.. you should checkout a clean version
[21:16:47] <alex_joni> in a different directory
[21:16:50] <alex_joni> and test that.
[21:16:55] <rcsu> grr, how ?
[21:16:59] <alex_joni> mkdir
[21:17:02] <alex_joni> md
[21:17:03] <alex_joni> :)
[21:17:22] <rcsu> thats done with cvs directly AFAIK
[21:17:32] <alex_joni> you can do it with cvs too
[21:17:50] <alex_joni> but it's easier to just make a new dir, then cd to it and follow the instructions at SF
[21:20:07] <rcsu> cvs -z3 -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/emc co emc2
[21:20:17] <alex_joni> that's it
[21:20:25] <rcsu> results in EOF from server
[21:20:39] <alex_joni> well.. seems SF has some problems once in a while
[21:20:42] <alex_joni> you need to try again
[21:20:44] <alex_joni> a few times
[21:20:48] <alex_joni> till it works
[21:20:59] <rcsu> ey, not the whole evening them
[21:21:13] <rcsu> ok
[21:21:22] <rcsu> ill search another stuff
[21:23:37] <Imperator_> rcsu: you are also from germany ! Where are you located ?
[21:27:43] <rcsu> Imperator_: yes, Goettingen
[21:27:54] <rcsu> * rcsu had to calm down
[21:28:43] <Imperator_> ah im from Aalen
[21:28:46] <rcsu> so, is there a viewcvs available ?
[21:29:06] <alex_joni> yes
[21:29:06] <alex_joni> at sourceforge
[21:29:09] <alex_joni> cvs.sourceforge.net/viewcvs.py/emc/
[21:29:14] <alex_joni> iirc correctly
[21:30:07] <rcsu> ahh nice
[21:39:53] <rcsu> whats the master process ?
[21:40:08] <alex_joni> rcsu: master?
[21:40:28] <alex_joni> ahh.. in your error message?
[21:40:34] <rcsu> cms_cfg.cc 908: cms_config: -4(CMS_NO_MASTER_ERROR: An error occured becouse the master was not started.) Error occured during SHMEM create.
[21:40:39] <alex_joni> well.. emctaskmain.cc starts some NML channels
[21:41:04] <alex_joni> look at emc.nml
[21:41:53] <alex_joni> # Top-level buffers to EMC
[21:41:53] <alex_joni> B emcCommand SHMEM localhost 8192 0 0 1 16 1001 TCP=5005 xdr
[21:41:53] <alex_joni> B emcStatus SHMEM localhost 10240 0 0 2 16 1002 TCP=5005 xdr
[21:41:53] <alex_joni> B emcError SHMEM localhost 8192 0 0 3 16 1003 TCP=5005 xdr queue
[21:41:53] <alex_joni> the process owning the channel, defined further below, creates it
[21:42:09] <alex_joni> and all other processes (which need to be run after that) access it without IPC_CREATE
[21:42:51] <alex_joni> can you run 'scripts/emc -v -d configs/rcsu.ini' and send me the output?
[21:43:52] <rcsu> emc.nml looks excactly the same as yours
[21:44:26] <alex_joni> yes I know.. was just pointing you towards master
[21:52:15] <rcsu> its not good to try cvs.sf.net so much times, but thats not your fault
[21:52:29] <rcsu> i now have a anon login and a co
[21:57:43] <alex_joni> great
[22:01:50] <rcsu> how must i set my configs ?
[22:02:13] <alex_joni> well.. there is a config picker right now
[22:02:32] <alex_joni> that lets you pick one of the existing configs, and use that as a base for your config
[22:02:43] <alex_joni> but you could test with configs/stepper/stepper_in.ini
[22:02:58] <alex_joni> if it hasn't already been renamed to stepper_inch.ini
[22:03:39] <rcsu> ok, my issue was a config file prob
[22:03:57] <alex_joni> it surely sounded like that, but it's strange..
[22:04:10] <alex_joni> care to look at it in more detail? or shall we scrap it
[22:04:29] <rcsu> if you want, why not
[22:04:41] <alex_joni> I'm kinda tired :)
[22:04:52] <rcsu> grin, me too
[22:05:05] <rcsu> but there is another day tomorrow
[22:05:14] <alex_joni> I
[22:05:21] <alex_joni> I'm off to the mountains tomorrow
[22:05:23] <rcsu> ok, g8
[22:05:33] <alex_joni> so I guess only in the evening I'll be around
[22:05:36] <rcsu> when are you back here ?
[22:05:43] <rcsu> ok, np
[22:05:53] <rcsu> its not a productive env
[22:06:03] <rcsu> so i have no hurry
[22:06:18] <alex_joni> rcsu: every one gets the same treatment :D
[22:06:42] <rcsu> can y gimme a starting point for steppers with all axes deflns ?
[22:07:33] <alex_joni> all axes?
[22:07:51] <cradek> to mny ltrs mssng frm sntnce
[22:08:04] <cradek> what is deflns?
[22:08:09] <rcsu> lol
[22:08:09] <alex_joni> cradek: 2 l33t 4 me
[22:08:16] <alex_joni> defs?
[22:08:32] <alex_joni> it's like.. the keys are right next to each other :D
[22:08:35] <alex_joni> a common typo
[22:08:49] <rcsu> can you give me a starting point for steppers with all the axes definitions
[22:09:03] <alex_joni> all meaning 6 ?
[22:09:17] <rcsu> 3 of them are ok for me :)
[22:09:19] <cradek> rcsu: the starting point for steppers is the "stepper" config
[22:09:30] <cradek> rcsu: it's 3 axes on parport step/dir
[22:09:40] <alex_joni> stepper/stepper_inch.ini (or stepper_in right now) is a good config
[22:09:48] <alex_joni> there should (will) be one for mm too
[22:09:54] <alex_joni> or did I add that?
[22:10:03] <alex_joni> * alex_joni keeps forgetting..
[22:10:23] <cradek> yes there is _mm and _inch
[22:10:28] <alex_joni> ok
[22:10:51] <alex_joni> so rcsu: just pick one (mm or inch) and you should be set
[22:11:05] <alex_joni> best would be to use the config-picker to make a new dir
[22:11:19] <rcsu> * rcsu is looking @ stepper_mm.ini
[22:15:17] <rcsu> and says g8 for today
[22:30:55] <alex_joni> later everyone
[22:30:58] <alex_joni> I'm off
[23:33:06] <robin_sz> meep?
[23:36:52] <jepler> all gone
[23:37:46] <jepler> sorry
[23:37:46] <robin_sz> nope ...I can see them hiding
[23:40:33] <skunkworks> how do you like the burny controller? I had to find a post proccessor for one a while ago.. (on the cheap)
[23:40:51] <skunkworks> maybe a 3 - I don't remember
[23:41:02] <robin_sz> I like it a lot, for a basic, reliable, simple controller
[23:41:09] <robin_sz> I hate it a lot for its price.
[23:41:14] <skunkworks> ;)
[23:41:20] <robin_sz> WAY too expensive
[23:41:36] <skunkworks> this was a machine that was purchased used - I think he got a good deal
[23:42:03] <robin_sz> the base model is over 6K, and doesnt even have homing switches!
[23:42:48] <robin_sz> but it is monkey proof
[23:42:55] <skunkworks> ;) it is funny as our laser we use for cutting die boards is never homed. the first line in the program is g92x0y0 ;)
[23:43:05] <skunkworks> in every program
[23:43:05] <robin_sz> heh
[23:43:20] <robin_sz> just jog to the top left corner?
[23:43:25] <skunkworks> right
[23:43:36] <robin_sz> ok top right corner ...
[23:43:36] <skunkworks> bottom left - what ever ;)
[23:43:55] <robin_sz> a LED pointer is SOOO useful there
[23:44:58] <skunkworks> I have to run - be back on later.
[23:45:01] <robin_sz> k