#emc | Logs for 2008-05-23

[00:07:52] <als> still no joy tring to install Hardy Heron tried burning @ 2x no joy errno 5 error bad cd or hard drive
[00:09:01] <als> I think its maybe the driver for the cd drive makes the drive read to fast but I would be guessing
[00:10:01] <als> the dapper install works great on the same sys. but cd drive doesn't wind up as fast?
[00:11:26] <als> I see others with the same error on the ubuntu forum
[00:12:59] <als> just thought I would report on it
[01:47:21] <cradek> good evening folks
[01:48:28] <rayh_> Hi
[01:48:58] <LawrenceG> ciao
[01:49:00] <SWPadnos> hwody
[01:49:05] <SWPadnos> umm. hi
[02:11:52] <cradek> how is it possible that I have so many things and no single metal outlet box?
[02:12:19] <DanielFalck> are you doing some wiring in the shop?
[02:13:53] <cradek> I would be...
[02:14:07] <rayh> handybox
[02:14:46] <rayh> Hi DanielFalck
[02:14:55] <DanielFalck> hi there rayh
[02:14:56] <nineteenfingers> hi folks... i see on linuxcnc.org's FAQ that EMC needs a minimum of 128MB of RAM to run (with an older version) - i was hoping to resurrect an old laptop to use for CNC work - but it only has 96MB, is there another way to do it rather than using EMC?
[02:15:22] <DanielFalck> turbocnc ?
[02:15:35] <cradek> even high power laptops are problematic for machine control
[02:15:36] <DanielFalck> dinner time for me here. bbl
[02:15:54] <rayh> One sec let me see how much I've got on the laptop.
[02:15:54] <nineteenfingers> cradek - how so?
[02:15:57] <cradek> it's true you might be able to use a dos control, but even that may work badly if the bios of the machine is uncooperative
[02:16:18] <cradek> power saving features built into laptops often harm their timing-critical performance
[02:16:28] <cradek> a human using the machine never notices but stepper motors sure do
[02:17:05] <nineteenfingers> DanielFalck - looking at their site now - thanks
[02:17:06] <rayh> 196meg and 6.06 installed there okay
[02:17:22] <nineteenfingers> tyah - only 100meg more than mine - haha
[02:17:28] <nineteenfingers> ryah^
[02:17:41] <rayh> I knew it was a bit less than the advertised minimum.
[02:18:17] <nineteenfingers> cradek - hadn't considered the power saving stuff - think it might be time to reconsider using this thing...
[02:19:12] <nineteenfingers> ok - if i was to set up a box just for machine control (cnc router) - what sort of spec is best to aim for? (just a rough idea)
[02:19:28] <nineteenfingers> rayh - i'm already at the maximum for this thing! haha
[02:20:07] <jepler> nineteenfingers: if you are a linux pro, you can start with the ubuntu 6.06 server install cd (text mode), then add X and emc packages after the fact. You won't be able to use the live cd installer with 128 megs RAM, but you may be able to get X + tkemc running in that kind of memory. few people are doing this.
[02:20:18] <rayh> Mine is a gateway laptop and that's the absolute max for it as well.
[02:20:36] <cradek> for an EMC machine I use old P3 with plenty of RAM (384 MB or more)
[02:21:04] <jepler> emc likes a CPU with at least high hundreds of MHz and not the low-cache varieties (sempron, celeron, etc)
[02:21:20] <nineteenfingers> jepler - i only have 96meg! :S - i'm no linux pro but i play with it - happy hacking on openSUSE when I have the time
[02:22:03] <cradek> I think that machine is not at all suitable for any recent OS
[02:22:15] <jepler> I don't know specificially how much or little RAM it takes to get to X in ubuntu these days .. but they aren't catering to machines with what are now such meagre specs
[02:22:37] <cradek> it could happily run DOS, windows 3.1 or 95, or extremely old Linux distributions
[02:22:40] <jepler> you could also start with bare debian and add what you need, if you follow the kernel compilation instructions on our wiki.
[02:22:43] <nineteenfingers> i'd been thinking about using a 2.4 series linux kernel (think: DamnSmallLinux) or something else of that scale
[02:22:45] <cradek> but, none of these are well suited for machine control
[02:22:50] <jepler> debian may give more of a nod to underpowered machines than ubuntu
[02:23:29] <jepler> 'night all
[02:24:46] <nineteenfingers> thanks for the input jepler
[02:48:30] <nineteenfingers> thanks for your input folks... looks like i've got a bit more to think about on this - gonna note down what you all suggested and mull it over - "i'll be back" - cheers
[02:51:12] <cradek> good luck and come back anytime.
[02:56:57] <nineteenfingers> count on it :p
[02:57:04] <nineteenfingers> adios
[03:22:11] <SWPadnos_> SWPadnos_ is now known as SWPadnos
[03:35:47] <fenn> piggy piggy ubuntu.. is hardy heron buggier than usual? or is it just that i'm actually using it instead of saying 'oh thats pretty' and using debian for real work
[03:38:03] <cradek> I haven't used it day-to-day yet. I am disappointed they chose to trade properly-working opengl applications for eyecandy though.
[03:38:37] <cradek> I'm also disappointed the gnome focus bug isn't fixed, but that's not exactly their fault (though, they could have fixed it in ubuntu)
[03:39:07] <fenn> well i've turned off all the desktop effects and still there is a lot of little weird things like mime-types being set wrong and kde stuff crashing a lot
[03:39:13] <cradek> otherwise I bet it's probably fine...?
[03:39:33] <cradek> oh I have not tried kde
[03:40:01] <SWPadnos> are you using KDE 3.5 or 4.0?
[03:40:05] <fenn> 3.5
[03:40:14] <SWPadnos> oh. that's supposed to be the less buggy one
[03:40:14] <cradek> I was affected by a serious [kernel] bug and was happy with the response when I put it in launchpad. you could try reporting things that bother you and they might get fixed.
[07:02:46] <K`zan> Night all.
[09:12:11] <pjm> good morning
[09:13:17] <dimas> good morning
[09:45:56] <pjm> btw perhaps some one could give me a quick pointer re the coordinates system in EMC. Say for example I've homed the table and my XY0 is in the bottom left corner, but I put a lump of metal in the vice and wobble it to get the centre, how can I then tell EMC where the new centre is?
[09:46:13] <pjm> I know its in the manual somewhere but I just wanted a quick idea of how to do it..
[11:58:48] <rayh> pjm, you still around?
[12:24:52] <micges> hello
[12:28:19] <micges> new pics of my new machine http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Micges_Laser_2
[12:28:20] <micges> another prototype :D
[12:31:30] <JanVanGilsen> Hi guys, I'm tring to get the TLO_IS_ALONG_W to work, but the tlo is still along Z
[12:42:13] <BigJohnT> cool machine micges:
[12:43:29] <micges> thanks
[12:43:40] <micges> its still in alpha stage :)
[12:45:03] <JanVanGilsen> could it be possible to use halui.tool.length-offset? I could link it with a halpin to my kinematics and recalulate tool.compensations :)
[12:45:16] <skunkworks> skunkworks is now known as skunkworks_
[12:47:05] <rayh> Fascinating info in the sourceforge privacy policy. "For example, when a user visits a web page, the code for the page may include instructions to go to another server to gather a single pixel graphic image. Web beacons are not placed on a user’s computer, and users remain anonymous."
[12:52:29] <Poincare> micges: did you solve the parallel port issue?
[12:55:58] <micges> yes
[12:56:13] <Poincare> and it was what I told you or something else?
[12:56:32] <rayh> Nice work micges
[12:56:51] <micges> I dont know how is it in english..but
[12:57:08] <micges> port in PC can't trigger logic 0
[12:57:32] <micges> becouse we connect signals to ground
[12:57:49] <micges> (sth like that)
[12:58:42] <micges> pull to ground or sth like that
[12:59:00] <micges> rayh: thanks
[12:59:42] <micges> laser is controlled by velocity not position
[12:59:53] <micges> move is VERY smooth
[13:01:21] <micges> Pointcare: we disabled pulling to ground (added to eliminate distortion) and everything is fine now
[13:05:21] <alex_joni> JanVanGilsen: it would be cool, but not possible atm
[13:05:31] <alex_joni> the tool compensation is done by interp/canon calls
[13:05:49] <alex_joni> the halui stuff is during the execution part, so it's different
[13:05:59] <alex_joni> one would have to move the tool compensation from the interp to motion
[13:10:04] <alex_joni> haha "Theory is when you know something, but it doesn't work.
[13:10:12] <alex_joni> Practice is when something works, but you don't know why.
[13:10:17] <JanVanGilsen> indeed
[13:10:21] <alex_joni> Programmers combine theory and practice: Nothing works and they don't know why."
[13:10:35] <JanVanGilsen> can you help me with the tlo along W stuff?
[13:12:51] <hasta2003> alex_joni: hi, I'm the guy that wrote to you for help on create an Ubuntu RTAI live cd.. may I ask you some questions?
[13:13:05] <alex_joni> hasta2003: sure, but maybe we should do it privately ;)
[13:13:19] <alex_joni> it will be boring for others
[13:13:32] <alex_joni> hasta2003: come to #emc-de
[13:13:38] <hasta2003> ok
[13:13:40] <hasta2003> thanks
[13:13:47] <alex_joni> no problem ;)
[13:14:06] <JanVanGilsen> i added that line to the TRAJ section but the "show EMC status" reports that is along Z
[13:14:23] <alex_joni> JanVanGilsen: bug cradek
[13:14:59] <JanVanGilsen> mmm, haven't seen any response from him :)
[13:15:47] <JanVanGilsen> i can't go into private conversation since i'm not registered
[13:22:54] <alex_joni> JanVanGilsen: no, just wait for him to respond here
[13:23:04] <alex_joni> cradek knows most about TLO_ALONG_W
[13:23:27] <alex_joni> (but if my memory serves me right, he did the work on a special branch.. is that what you are using?)
[13:24:24] <JanVanGilsen> no, im using the live cd version
[13:25:04] <alex_joni> then it won't work ;)
[13:25:15] <alex_joni> tlo along w is still experimental in a branch in CVS
[13:25:26] <alex_joni> so if you want it.. you need to grab it from CVS and compile it
[13:27:31] <JanVanGilsen> now i only have to figure out where he added that ..
[13:27:32] <Poincare> JanVanGilsen: ben jij van de HA afdeling Paardemarkt?
[13:27:37] <JanVanGilsen> Ja
[13:28:13] <Poincare> tof, wist niet dat men daar met emc2 bezig was
[13:29:06] <Poincare> Hi again :-D
[13:29:33] <JanVanGilsen> Pointcarre: Ja, sinds dit jaar, ik maar er mijn eindwerk over
[13:30:07] <Poincare> 't is Poincare zonder T
[13:30:14] <Poincare> in welke richting zit je?
[13:30:36] <Poincare> For those that don't understand us: we're talking Dutch :-D
[13:31:25] <JanVanGilsen> okidoki, ik zit in elektromechanica
[13:31:57] <Poincare> 'k Ken volk bij de elektronica mensen
[13:32:12] <Poincare> en heb daar zelf nog gezeten, 15 jaar geleden
[13:33:06] <JanVanGilsen> ah, das grappig we hebben hier een 5-assertje gebouwd, en een paar machines van het kerkhof gerecycleert met emc
[13:33:57] <Poincare> een paar dan nog, ik moet nog aan men eerste beginnen
[13:34:27] <JanVanGilsen> weetj jij perongeluk hoe je de cvs branch vindt met tlo_along_w ?
[13:34:42] <Poincare> nee, ik ben geen developer maar eerder een gebruiker
[13:36:33] <JanVanGilsen> Oh, mss ken je mijn vader, die geeft les aan de HA.
[13:37:29] <Poincare> Van Gilsen zegt me niets... Wat geeft die?
[13:39:02] <JanVanGilsen> Robotica, heeft ook nog EEM, informatica en operationeel onderzoek gegeven
[13:39:38] <Poincare> Zegt me nog steeds niets, maar ik ben daar al sinds 1995 weg he
[13:45:36] <JanVanGilsen> waar werk je nu?
[13:46:25] <skunkworks_> logger_emc: bookmark
[13:46:25] <skunkworks_> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2008-05-23.txt
[13:48:53] <JanVanGilsen> I found a w_tool_length branch, I'm hoping that will do the trick
[13:51:02] <cradek> TLO_IS_ALONG_W is in trunk
[13:51:31] <cradek> the branch is less complete. you should use trunk.
[13:51:44] <JanVanGilsen> okey, thanks alot :)
[13:55:22] <cradek> Andre says use timing gears/belts for a 40-50 : 1 motor reduction. wouldn't this take 3 or even 4 stages?
[13:56:02] <SWPadnos> or a 3-foot diameter pulley
[13:56:51] <cradek> ok so it's not just me :-)
[13:56:56] <SWPadnos> heh
[13:57:07] <cradek> I don't think you can reasonably get over about 3:1 with a single stage
[13:57:16] <SWPadnos> that's what I hear
[13:57:20] <cradek> because the engagement on the small gear is too little
[13:57:23] <Poincare> JanVanGilsen: www.ampersant.be
[13:57:34] <SWPadnos> you can use idlers to increase the engagement
[13:57:39] <lerman> Unless you use some idler wheels.
[13:57:52] <cradek> true
[13:59:29] <fenn> differential gearing?
[14:00:00] <lerman> Differential screw.
[14:01:25] <lerman> A 4 pitch screw advances .250, while a 5 pitch retards .200. That gives .050 per rotation of the screw.
[14:02:08] <lerman> Unfortunately, since it corresponds to a 20 pitch screw, it takes 20 turns per inch or 240 turns per foot.
[14:02:27] <SWPadnos> hmmm. and at some point the screws fall out of the carriage
[14:02:50] <lerman> 240 turns of a 4 pitch screw means that the screw is moving 60 inches to get 12 inches of travel.
[14:02:53] <JanVanGilsen> Poincare: webosting, ben zelf al veel bezig geweesd met php, mysql en apche
[14:03:58] <SWPadnos> sure looks complicated: http://www.wipo.int/pctdb/images/PCT-IMAGES/04112004/JP2004005598_04112004_gz_en.x4-b.jpg
[14:05:37] <lerman> Patent drawings always look complicate. :-
[14:05:45] <fenn> roller nut
[14:05:47] <SWPadnos> oh, here's a simpler model: http://brunelleschi.imss.fi.it/genscheda.asp?appl=LIR&xsl=modello&lingua=ENG&chiave=100805
[14:06:20] <alex_joni> how about using a planetary gear?
[14:06:30] <fenn> i think the idea was to use timing belts
[14:06:40] <alex_joni> http://en.wikipedia.org/wiki/Epicyclic_gearing
[14:08:26] <fenn> http://www.darali.com/page17.html
[14:09:15] <fenn> crappy page, but i cant find the good page i'm thinking of (story of the 21st century)
[14:10:47] <fenn> its like a harmonic drive without all the pesky deformation
[14:11:55] <lerman> And the backlash is...???
[14:13:10] <SWPadnos> deformation within the elastic limit of the metal is apparently just peachy, more or less forever
[14:14:06] <fenn> yeah, well, those things are friggin expensive
[14:15:06] <fenn> lerman: i think you could use two inner gears with a spring preload between them
[14:16:25] <lerman> Lots of precision contact areas. I suspect that these are power devices; not precision devices.
[14:17:07] <renesis_> fuck backlash!
[14:17:16] <renesis_> just machine from the same direction all the time!
[14:17:30] <renesis_> together we can defeat this!
[14:17:32] <fenn> just use really sticky ways!
[14:17:34] <renesis_> oh, hai
[14:17:35] <cradek> that works great (until you want a circle)
[14:17:40] <renesis_> hahahaha @ stickyways
[14:17:52] <renesis_> cradek: =(
[14:18:37] <renesis_> ok ya we prob need circles my thinger doesnt hold endmills that big
[14:18:52] <cradek> all you need is a boring head
[14:19:05] <fenn> lerman: the spring preload reduces the need for precision
[14:19:08] <renesis_> alot of them?
[14:19:37] <renesis_> * renesis_ was wanting one of those single insert holders w/ 3/8" straight shank
[14:19:56] <lerman> fenn: thats as far as backlash is concerned; not as far as precise rotary motion. Oh, we can correct for that with our linear encoders.
[14:19:56] <renesis_> boxes of TiN Carbide ninja stars are so cheap!
[14:20:05] <JanVanGilsen> http://www.lifetime-reliability.com/tip_003.html
[14:20:36] <fenn> lerman: well, for a robot there's no linear encoder, but you could map the errors i guess
[14:21:14] <cradek> JanVanGilsen: wow that's neat
[14:21:51] <fenn> yep that's pretty much what i was thinking
[14:21:57] <JanVanGilsen> its sometimes used in robots
[14:22:34] <renesis_> janvangilsen: wtf
[14:22:39] <renesis_> omg they use that
[14:22:40] <renesis_> ?
[14:22:45] <renesis_> neat
[14:23:34] <JanVanGilsen> it has a huge reduction :)
[14:23:54] <JanVanGilsen> you should see an animation of it to see how it works
[14:24:05] <JanVanGilsen> http://www.youtube.com/watch?v=vPPhwoe2QMU => thats neet too :)
[14:24:06] <cradek> I think it would be the same reduction as the number of lobes, so 40-50 is probably possible
[14:24:44] <renesis_> http://smcyclo.com/media/cyclostory/cyclo-large-wmv.html
[14:25:04] <JanVanGilsen> cradec: you'r right
[14:25:18] <renesis_> oh neat they have two cams on theirs
[14:25:20] <fenn> sumitomo uses the double gears
[14:26:13] <renesis_> ha i bet they bounce nice when they get old
[14:26:34] <fenn> not if it's spring loaded
[14:27:00] <fenn> unless you mean "really really old"
[14:27:08] <pjm> rayh, yeah i'm here, out in the garden planting beans!
[14:27:20] <renesis_> hmm
[14:27:55] <renesis_> wait how come we havent thought of that rack and pinion linear thing yet
[14:28:25] <fenn> like a cycloidal roller gear but linear?
[14:28:47] <renesis_> no click the youtube its just rack and pinion
[14:29:07] <fenn> what youtube?
[14:29:25] <SWPadnos> http://www.youtube.com/watch?v=vPPhwoe2QMU
[14:29:26] <renesis_> 7:27pst
[14:30:24] <renesis_> but yeah i can think of multiple ways to backlash compensate that
[14:49:44] <JanVanGilsen> i've buit the trunk but now it complais about the kinematics component that doens'nt exist, I recompiled it with comp but it gives the same error
[14:50:02] <fenn> what's the error?
[14:51:47] <JanVanGilsen> Can't find module 'tablekins' in /home/student/emc2-trunk/rtlib
[14:52:41] <jepler> make sure that you invoke the right "comp" when you compile and install your 'tablekins'.
[14:53:09] <jepler> Either execute ". /home/student/emc2-trunk/scripts/emc-environment" in the terminal before you run comp, or give the path to comp: /home/student/emc2-trunk/bin/comp --install tablekins.c
[14:53:51] <JanVanGilsen> i executed em-environment,
[14:54:03] <JanVanGilsen> ill try the complete path :)
[14:54:15] <jepler> is tablekins.ko (or .so if --enable-simulator) in /home/student/emc2-trunk/rtlib ?
[14:58:43] <JanVanGilsen> the complete path fixed it, thx
[15:01:07] <JanVanGilsen> arg, the tooloffset is still in Z
[15:03:02] <JanVanGilsen> would it be useful to put my ini-file on pastebin?
[15:03:07] <cradek> yes
[15:04:03] <als> could there be some reson why I cannot install CodeBlocks IDE in the dapper emc live cd install?
[15:04:31] <cradek> if you can't, yes, I'm sure there is some reason
[15:04:54] <als> /lib/tls/libc.so.6: version `GLIBC_2.4' not found (required by codeblocks)
[15:05:17] <cradek> where did the package come from?
[15:05:32] <JanVanGilsen> http://pastebin.ca/1027124
[15:05:54] <als> code blocks latest stable release -v
[15:06:18] <cradek> that means the binaries are incompatible with your system
[15:06:26] <cradek> can you compile it yourself? that would fix it.
[15:06:51] <als> i'll give it a shot
[15:07:32] <als> should I uninstall first?
[15:07:48] <cradek> yes I'd remove the incompatible binary install first
[15:07:56] <als> thanks
[15:08:19] <JanVanGilsen> tablekins.c: http://pastebin.ca/1027128
[15:08:21] <cradek> JanVanGilsen: looks right to me. can you tell me what you see
[15:08:28] <pjm> btw i have put a utube vid of my XY table build to date if anyone wants to look, its at http://www.youtube.com/watch?v=tJkPLrSkfCc
[15:09:14] <SWPadnos> shouldn't there be some sort of kinematics= line?
[15:10:14] <SWPadnos> ah, in the hal file. nevermind
[15:10:38] <JanVanGilsen> loadrt tablekins
[15:10:38] <cradek> JanVanGilsen: I don't see where you apply pos->w to joints[2]
[15:10:46] <JanVanGilsen> i don't
[15:10:49] <cradek> but still, you should see W change on the screen when you apply a tool offset
[15:11:04] <JanVanGilsen> not a Z change ...
[15:11:32] <cradek> are you sure you are running the trunk version you built?
[15:12:16] <cradek> wait you don't even have a W axis defined
[15:12:32] <cradek> put W in your ini so you can see it onscreen
[15:13:09] <cradek> and I'm pretty sure you have to apply pos->w to joints[2] which is not 'Z'
[15:13:20] <cradek> when you have a longer tool, you have to lift the head higher (joint 2)
[15:13:27] <JanVanGilsen> it works
[15:13:40] <cradek> which part?
[15:13:45] <JanVanGilsen> yes the "frees" variable is the tlo
[15:14:13] <JanVanGilsen> Z doen't change anymore, i don't know how i did it tough
[15:14:21] <cradek> frees is fixed
[15:14:29] <JanVanGilsen> atm yes
[15:14:33] <cradek> I guess it is the height when TLO=0
[15:14:41] <cradek> but when TLO (W) changes you need to change this height
[15:14:57] <cradek> that is the whole point here
[15:15:29] <JanVanGilsen> its the height for my current tool, but its not verry usefull when you need to recompile the kinematics when changing tool :)
[15:15:39] <cradek> yes
[15:15:44] <cradek> that's why you use TLO
[15:16:28] <JanVanGilsen> yes, i know i'm goin to ad the pos->W and joint[8], i removedit because it didn't work :)
[15:25:57] <tomp2> the new issue of Control Engineering has article on NIST.
[15:25:57] <tomp2> The last page editorial has photo of NIST Hexapod.
[15:25:57] <tomp2> Very proffesional execution, very new looking, tho undated photo.
[15:25:57] <tomp2> The control screen is visible on a pendant screen.
[15:25:57] <tomp2> whole article LESS photo is at...
[15:25:58] <tomp2> http://www.controleng.com/article/CA6559113.html?q=NIST
[15:26:00] <tomp2> will try to scan & imagebin.
[15:36:21] <JanVanGilsen> the tlo is now in w, but i replaced frees with pos->w and joint[8] but that doens't work
[15:36:53] <JanVanGilsen> arg, sorry for my bad english
[15:39:13] <tomp2> re: NIST hexapod http://imagebin.ca/view/Tnf3N1x.html
[15:40:15] <JanVanGilsen> in "Show EMC status" both joint_position and joint_actual_position are 0 for joint 9
[15:41:54] <JanVanGilsen> *joint 8
[15:45:49] <tomp2> re: NIST Hexapod, that machine may have been built by Ingersoll http://www.atp.nist.gov/eao/sp950-3/immc.pdf The connection between the photo and an Ingersoll Hexapod is not clear.
[15:46:39] <SWPadnos> only $3.499 Million - what a deal! :)
[15:47:26] <tomp2> could've gon eto the emc2 crew ;)
[15:47:33] <SWPadnos> yeah, darnit
[15:49:30] <tomp2> hey they're in Rockford, on the way from ORD to Fest, maybe they got some old iron to donate
[15:49:45] <SWPadnos> or nice motion components ...
[15:54:34] <jepler> and you also have to command a W move (even one that does nothing) for a length offset on W to become effective.
[15:54:38] <jepler> (this is the same rule foras for Z)
[15:55:33] <renesis_> renesis_ is now known as renesis
[15:59:32] <K`zan> Morning all.
[17:18:22] <BigJohnT> http://www.youtube.com/watch?v=KvhvQu5Xxtk
[17:22:23] <LawrenceG> BigJohnT, nice work..... that would be such a fun toy to have in the shop!
[17:22:58] <BigJohnT> thanks LawrenceG:
[17:24:12] <BigJohnT> come over and play with anytime you want
[17:25:02] <LawrenceG> not sure... will plasma cut stainless? I know oxy-xxx doesnt work well as the O2 doesnt burn the steel due to the nickel
[17:25:32] <BigJohnT> yes but it leaves an ugly edge on air
[17:26:02] <LawrenceG> more of an air-arc type of cutting than a clean burn
[17:29:50] <BigJohnT> there is nothing clean about plasma cutting LOL
[17:41:51] <klickrr> is there a good explanation of FF0, FF1, FF2 anywhere? the PID tuning doc says a brief one word description of each...
[17:42:37] <klickrr> oh i found something... still not to much though
[17:42:57] <klickrr> FF1 = 2.000 a velocity feedforward, helps reduce following error proportional to velocity
[17:43:20] <klickrr> that appears to be the exact issue i'm tackling, i'll play with FF1 then, cool
[18:03:49] <hasta2003> alex_joni: hi
[18:04:33] <dmess> hi all
[18:04:52] <dmess> yabadabba-dooobie.... its friday....
[20:34:22] <BigJohnT> I have an ellipse. One program generated a series of line/arcs and one gave me all short lines. No matter what I set G64Pn to the file that has line/arc runs real jerky.
[20:34:49] <BigJohnT> I'm assuming it is stopping or slowing down during the transition from arc to line or the other way around
[20:35:07] <BigJohnT> The second file that is only short lines runs very smooth
[20:35:33] <BigJohnT> Is this the expected behavior in EMC?
[20:35:55] <BigJohnT> feed is 125 IPM acc is 200
[20:36:26] <cradek> not expected by me. I bet you have zero length moves in the jerky one.
[20:36:54] <BigJohnT> I didnt' see any but I'll look closer
[20:36:56] <cradek> see if you can figure out which lines cause a jerk.
[20:37:27] <BigJohnT> it's constant more or less throught the whole file
[20:37:56] <SWPadnos> is the output line/arc/line/arc, or line/line/line/arc/arc/line/line ... ?
[20:38:10] <BigJohnT> line/arc/line/arc
[20:38:19] <SWPadnos> more or less alternating all the time?
[20:38:23] <BigJohnT> yep
[20:38:30] <cradek> oh wait a sec, I'm wrong
[20:38:36] <cradek> g64Pn only works with all lines
[20:38:37] <SWPadnos> ok. I don't know what that means, but thought it could be relevant :)
[20:38:46] <BigJohnT> now that makes sense
[20:38:46] <SWPadnos> ther etou go :)
[20:38:48] <SWPadnos> uh
[20:38:50] <SWPadnos> there you go
[20:38:58] <cradek> sorry, I forgot until I thought about it more
[20:39:09] <BigJohnT> I don't like but I understand now
[20:39:17] <SWPadnos> can't you just memorize the code already? geez
[20:39:26] <BigJohnT> me?
[20:39:30] <alex_joni> no, cradek
[20:39:33] <SWPadnos> heh
[20:39:36] <SWPadnos> yeah, you too!
[20:39:53] <BigJohnT> I have a mind like a steel trap
[20:39:57] <SWPadnos> me too
[20:40:04] <BigJohnT> a lot gets in but nothing gets out
[20:40:09] <SWPadnos> too bad I have to gnaw off a leg to get anything out
[20:40:24] <alex_joni> SWPadnos: and you don't have any left?
[20:40:31] <SWPadnos> nah, they keep regrowing
[20:40:41] <cradek> alex_joni: what's the no?
[20:40:55] <SWPadnos> no to bigjohnt's question
[20:41:02] <SWPadnos> with the correct answer, "cradek"
[20:41:22] <cradek> I'm confused, but that's ok
[20:41:39] <alex_joni> cradek: nm.. sorry I confused you even more than usually
[20:41:44] <SWPadnos> (responding to BJT question about my comment about you memorizinv the code)
[20:41:49] <SWPadnos> s/v/g/
[20:41:58] <rayh> I wrote several ellipse programs for a mazak years ago. Used the Machinery Handbook's formulas so they were all arcs tangent where they connected. .
[20:41:59] <BigJohnT> so is G2/3 a slow down on processing or just the look ahead?
[20:42:25] <SWPadnos> it stops that kind of blending
[20:42:28] <BigJohnT> rayh: does that improve the movement
[20:42:47] <rayh> Oh yea. No change in speed at all.
[20:43:13] <BigJohnT> to bad my software can't do that
[20:43:17] <cradek> BigJohnT: G64Pn lets short lines combine into one, if the resulting motion is within the tolerance. with arcs, they don't combine, so the planner has to slow down to do each arc
[20:43:25] <rayh> I did it with a Vic-20.
[20:43:39] <BigJohnT> did you guys see the video ?
[20:43:42] <rayh> So your software should be able to handle it ;)
[20:43:52] <cradek> BigJohnT: which one?
[20:44:02] <BigJohnT> running the plasma
[20:44:31] <BigJohnT> http://www.youtube.com/watch?v=KvhvQu5Xxtk
[20:45:17] <BigJohnT> cradek: does the speed of the computer make any difference?
[20:45:57] <cradek> BigJohnT: nope
[20:46:16] <cradek> BigJohnT: machine acceleration would make the biggest difference
[20:46:31] <BigJohnT> mine is at 200 now
[20:46:39] <BigJohnT> IPSPS
[20:46:40] <kohlsva> Hello, I've tryed to find out how much disk space a standard installation of EMC2 will take as I'm going to use SSD but wondering I'll need 1- or 2Gb, I'm guessing its more than the .iso size it self ? I've had no luck finding an answer on linuxcnc.org.
[20:47:00] <alex_joni> kohlsva: well.. 2G will be crowded I think
[20:47:06] <alex_joni> but you can tweak it to work
[20:47:16] <cradek> kohlsva: for the ubuntu installation I think it's 3G+
[20:47:21] <alex_joni> 1G is certainly not enough (unless you want to run it like a livecd_
[20:47:33] <BigJohnT> my first hard drive was 10 mb
[20:47:42] <alex_joni> mine was 40MB
[20:49:12] <rayh> I stripped out a lot of the stuff and got it on a 2G flash drive.
[20:49:22] <kohlsva> ok. hm..
[20:49:56] <rayh> It could run on a lot less by using lighter window managers and such.
[20:50:09] <cradek> openoffice is a great one to strip out first
[20:50:15] <alex_joni> kohlsva: might be easier to start with xubuntu
[20:50:21] <alex_joni> then use the install script to get emc2
[20:50:26] <cradek> very true
[20:50:27] <kohlsva> ah, ok.
[20:50:27] <rayh> During the puppy days I put it on a 125 Meg stick
[20:50:44] <alex_joni> yeah, too bad puppy is so much pita to get right
[20:50:55] <rayh> Exactly
[20:51:05] <alex_joni> and the one that's working is waaay too old
[20:51:56] <alex_joni> it's pre 2.0.0 iirc
[20:52:18] <rayh> I've been testing here with the blackbox window manager.
[20:52:29] <rayh> It's less than a meg.
[20:52:43] <rayh> There is also a small version of X.
[20:53:20] <alex_joni> * alex_joni remembers building a linux distro for ARM a while ago
[20:53:29] <alex_joni> it was about 4MB total :D
[20:53:38] <alex_joni> including kernel and busybox :)
[20:54:01] <SWPadnos> 6.06 with all dev tools fits in 3G
[20:54:15] <SWPadnos> that's the desktop install + dev tools
[20:54:25] <SWPadnos> plus other stuff
[20:54:29] <alex_joni> ok, so removing openoffice & co will save a lot of things
[20:55:11] <kohlsva> if i eliminate x, will emc2 work fine just with the keystick interface ?
[20:55:49] <jepler> I sure wouldn't want to use emc that way
[20:56:02] <alex_joni> kohlsva: you might have to fix the runscript
[20:56:02] <rayh> Sure.
[20:56:09] <alex_joni> but other than that, it should work
[20:56:22] <alex_joni> (I think the runscript runs keystick inside a xterm)
[20:56:35] <rayh> You can also use X without a window manager if you fix mini just a bit.
[20:56:36] <kohlsva> jepler: well the thing im building is going to very limited things.
[20:56:43] <alex_joni> kohlsva: the other thing you can do is running the GUI remotely
[20:56:53] <SWPadnos> this install has about 400M of extra junk, in addition to all the junk that I didn't remove :)
[20:57:11] <alex_joni> SWPadnos: mine has lots more :)
[20:57:12] <SWPadnos> so I'd bet 2G is enough if you think about what you're doing
[20:57:19] <alex_joni> mostly hardy sandbox .. 4-5G
[20:57:26] <SWPadnos> well, that's the one that runs on the 8G SSD
[20:59:22] <kohlsva> btw if i install xubuntu and then the install script wont there be issues since normaly(?) ubuntu use a diffenrent kernel ?
[20:59:58] <alex_joni> kohlsva: the emc2 install script will install the proper kernel
[21:00:08] <kohlsva> ah, ok
[21:00:15] <alex_joni> (you will have to remove the old one if the rtai one is working ok)
[21:03:26] <jepler> the prebuilt packages have lots of X stuff in their dependencies. you'll have to either not use a packaged version, deliberately break package dependency management, or modify the packaging so that you can create an emc2-nox package or the like.
[21:04:10] <kohlsva> I've been reading some of the information regariding HAL, but I'm uncertain of it accessability, can i readaly do hal operations in a runing program, ie, go to a subprobram that will prefrom hal functions to home a axis?
[21:04:45] <BigJohnT> do you mean from a push button ?
[21:05:12] <alex_joni> jepler: emc2-nos sounds nicer
[21:05:20] <kohlsva> well, no. ill try to explain what im up to.
[21:05:43] <SWPadnos> homing isn't a HAL function, just so you know
[21:05:58] <alex_joni> you can trigger homing from HAL, but not while running a program
[21:06:03] <alex_joni> (or at least harder..)
[21:06:09] <SWPadnos> (a G-code program)
[21:06:23] <alex_joni> are there other types?
[21:06:27] <kohlsva> i might missued home.
[21:07:05] <kohlsva> i mean to use hal to confirm that the rotary table trips a switch
[21:07:26] <BigJohnT> as in probing?
[21:07:35] <BigJohnT> I do that with my plasma torch
[21:07:55] <BigJohnT> all in G code
[21:07:56] <kohlsva> i whant the program to stop moveing the rotary table when it trips and continue with the regular program
[21:08:17] <BigJohnT> that sounds like a probe move
[21:08:21] <cradek> I agree
[21:08:43] <cradek> it's not the same as homing though. you cannot home from gcode.
[21:08:49] <BigJohnT> AFAIK you will have to back off the switch after you trip it
[21:09:18] <BigJohnT> G28 is return to home...
[21:09:22] <kohlsva> im trying to get a threaded part into the rotarty table by spining it and then i whant it to move backwards til it trips a switch and then continues the program
[21:10:01] <BigJohnT> you can use a G38.2 probe move to do that
[21:10:18] <kohlsva> the threaded part is suposed to have a hole in relation to the thread. :)
[21:10:18] <BigJohnT> if it works on a rotary???
[21:10:21] <alex_joni> BigJohnT: that doesn't reset the A position though
[21:10:50] <alex_joni> but I guess you can use some G92 or G10 to make it 0
[21:11:00] <kohlsva> ok
[21:11:50] <kohlsva> the reson id whant it to reset the value of the rotatry table is so that i could do this with automation and just loop the program.
[21:11:54] <BigJohnT> could he do this with M66?
[21:12:06] <alex_joni> BigJohnT: I don't think so
[21:12:40] <alex_joni> kohlsva: I'd try G38.2 and then Gxx for resetting the position
[21:13:00] <alex_joni> although I would expect to know what the A position is after threading
[21:13:56] <BigJohnT> so if I'm jumping back and forth between G1 and G2/3 would faster acceleration be better or slower?
[21:14:04] <alex_joni> better
[21:14:09] <BigJohnT> smoother
[21:14:16] <BigJohnT> more constant speed
[21:14:25] <alex_joni> smaller times it stops :P
[21:14:50] <kohlsva> hmm. well yes, but i had the thought of the switch to confirm that the lotation would be repeatable even after long runs and prehaps los of steps.
[21:15:20] <alex_joni> kohlsva: you can confirm it with M66
[21:15:33] <alex_joni> go to switch position, check for digital input, if it's not there.. error out
[21:15:49] <alex_joni> (blindly fixing lost steps is not a solution imho)
[21:17:11] <kohlsva> well im consered about torking the threaded part to stay in the threaded part in the rotaty table, removeal and thighning is likely to give me issues.
[21:20:08] <cradek> BigJohnT: I finally watched your video. You can make a very passable ellipse out of four tangent arcs: http://www.tpub.com/content/engineering/14069/css/14069_152.htm
[21:20:58] <cradek> it's not a perfect ellipse BUT it will be totally smooth and show no "points" which you may be getting with your line approximation
[21:22:06] <skunkworks_> how about this? http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Oword#Sample_2_Subroutine_Sample_ellipse
[21:22:10] <skunkworks_> * skunkworks_ just got here
[21:22:15] <kohlsva> I'm sorry if I'm going around this the wrong way, I do appreciate all the help and good information.
[21:23:16] <cradek> skunkworks_: cool, but still a line approximation so it will probably end up with (minor) flats and points
[21:23:20] <rayh> kohlsva May I ask where you are located?
[21:23:51] <kohlsva> Sweden, sorry about the constant spelling errors.
[21:25:06] <rayh> Glad you are with us.
[21:25:21] <alex_joni> kohlsva: no problems, we're not all native english speakers here
[21:25:29] <alex_joni> kohlsva: and there are a couple guys from sweden
[21:25:35] <rayh> No problem about your typing or english.
[21:26:00] <alex_joni> kohlsva: not right now, but you can look for anonimasu, awallin, lernaenhydra, and a couple more
[21:26:26] <alex_joni> hmm.. awallin is finnish, not swedish (my bad)
[21:26:28] <BigJohnT> I'm not english either
[21:26:39] <BigJohnT> I'm redneck
[21:27:07] <kohlsva> Heh, well Ive lived stateside for about two years.
[21:27:33] <kohlsva> (Im back in Sweden now)
[21:27:43] <alex_joni> BigJohnT: now that explains a lot
[21:27:52] <BigJohnT> what does?
[21:28:04] <alex_joni> < BigJohnT> I'm redneck
[21:28:10] <BigJohnT> LOL
[21:28:12] <alex_joni> BigJohnT: just kidding ;)
[21:28:50] <BigJohnT> I just tried different accelerations and no difference only slowing down to below 50IPM does it move smooth :(
[21:29:52] <BigJohnT> I took the ferry from denmark to sweden once but it was closed as it was Sunday
[21:31:03] <kohlsva> err, that dont add up ?
[21:31:38] <BigJohnT> cradek: the only problem I have is with jerking going from G1 to G2/3 moves at high speeds
[21:31:59] <BigJohnT> kohlsva: me?
[21:32:54] <kohlsva> yes, sorry, but how can one take a ferry with it dosent run
[21:33:25] <BigJohnT> it was some time ago
[21:34:17] <kohlsva> do you live in germany or wore you on a trip thou europe ?
[21:34:29] <BigJohnT> I was in .... one moment I'll tell you the cities
[21:34:53] <BigJohnT> I was going to Middlefart Denmark to visit a factory that made machines for making nails
[21:35:28] <BigJohnT> That day I was in Copenhagen
[21:36:03] <BigJohnT> we went to Helisngborg
[21:36:26] <BigJohnT> I think
[21:36:31] <kohlsva> ah ok
[21:36:43] <BigJohnT> make sense now?
[21:37:08] <BigJohnT> 1990 or so
[21:37:17] <BigJohnT> crap I'm old
[21:37:38] <kohlsva> ah yes, btw they finaly made a bridge
[21:37:42] <BigJohnT> Oh wait I saw Ray's picture and don't feel so old now...
[21:37:52] <kohlsva> oh ?
[21:38:06] <BigJohnT> ok
[21:38:20] <kohlsva> what did you think of Copenhagen ?
[21:38:28] <rayh> Just wait. You'll get there!
[21:38:35] <alex_joni> kohlsva: where are you in sweden?
[21:38:51] <BigJohnT> it was nice we stayed at an old wharehouse converted to a hotel
[21:39:13] <BigJohnT> LOL but your going get there first Ray
[21:39:30] <rayh> Nah you'll catch up pretty quick.
[21:39:40] <BigJohnT> the warehouse was built during the sailing ship days from real timbers
[21:39:53] <kohlsva> Alex_joni: Lund
[21:40:34] <alex_joni> kohlsva: cool.. I was there a couple times
[21:40:37] <alex_joni> at lth
[21:40:58] <kohlsva> very good uni
[21:41:08] <alex_joni> Lund is a great city
[21:42:33] <kohlsva> well, its a nice city but there are some issues, houseprices for exsample
[21:43:24] <kohlsva> btw ppl, this might be of intrest, http://org82.zorpia.com/0/4372/27986152.71b7c9.jpg its one of the parts im trying to work on.
[21:44:10] <BigJohnT> are you assembling the part or making it
[21:44:17] <kohlsva> doing the hole
[21:44:26] <kohlsva> or makeing a machine to do the hole for me
[21:44:47] <BigJohnT> the hole in the stickmeout part?
[21:45:08] <kohlsva> no, but the one near next to the thread
[21:45:15] <kohlsva> its in relation to the thread
[21:45:28] <BigJohnT> ok the spotface next to the thread?
[21:45:49] <kohlsva> yes
[21:46:11] <kohlsva> it screws into another part and when indext properly it will tension a oring just right
[21:46:12] <BigJohnT> you screw this in a number of turns then make the spotface?
[21:47:10] <kohlsva> atm i have a fixture on a cube in a horizontal paletchanger.
[21:47:50] <BigJohnT> aircraft part?
[21:48:16] <kohlsva> marine and bio
[21:48:24] <BigJohnT> cool
[21:48:57] <kohlsva> its one of several simular looking sparts that are used in different machines for cleaning tanks.
[21:49:47] <BigJohnT> rotating spray nozzle?
[21:50:39] <kohlsva> im not sure of the exsact workings of the part since ive been with the company less than 6 months.
[21:51:50] <BigJohnT> well, time for me to get my strechey pants and slippers on and pet the dog awhile while I chill out. have a good weekend everyone
[21:52:10] <kohlsva> oh, well take care.
[21:52:28] <rayh> I met a fellow from Sweden at the NAMES show several years ago. He use EMC to make embossing rolls for paper.
[21:52:36] <alex_joni> * alex_joni is also off to bed
[21:52:41] <alex_joni> good night all
[21:52:52] <kohlsva> good night alex joni
[21:54:07] <alex_joni> God natt
[21:54:32] <alex_joni> kohlsva: Vq^ is also from sweden
[21:55:03] <kohlsva> thanks
[21:55:44] <kohlsva> Rayh, New England NAMES (http://www.neme-s.org/) or The North American Model Engineering Society?
[21:56:24] <rayh> Yes. They meet in the midwest Detroit then Toledo now.
[21:58:31] <kohlsva> is there prehaps a site that lists the memebers different projects ?
[21:58:43] <kohlsva> i mean, the linux cnc memebers
[21:59:09] <kohlsva> i meant to say ppl active here.
[22:02:45] <rayh> There are some links and projects on www.linuxcnc.org and wiki.linuxcnc.org
[22:03:20] <rayh> Sorry distracted NAMES I was thinking of is The North American Model Engineering Society?
[22:03:41] <rayh> We used to have quite a large area for folk to show machines and projects.
[22:05:20] <rayh> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Case_Studies
[22:05:57] <kohlsva> ah yes, thanks, i should have done more research
[22:06:34] <rayh> http://www.linuxcnc.org/component/option,com_weblinks/catid,13/Itemid,7/lang,en/
[22:06:46] <rayh> a couple of those links are broken.
[22:07:31] <rayh> It's not complete at all but http://www.frappr.com/emc2 shows where some folk are located.
[22:09:50] <rayh> Wow guys. Just got the raffle box from Pete at Mesa. Nine ROSH compliant boards.
[22:13:25] <rayh> Dinner calls, bbl.
[22:41:37] <kohlsva> Jeeves And Wooster
[23:14:04] <micges> logger_emc: bookmark
[23:14:04] <micges> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2008-05-23.txt
[23:55:19] <tomp2> kohlsva: Jeeves & Wooster, one of my favorites, http://www.gutenberg.org/browse/authors/w scroll down to Wodehouse for free books
[23:55:52] <kohlsva> oh :)
[23:58:39] <kohlsva> uhm.