#emc | Logs for 2009-12-23

[00:00:18] <ries> We did make the V-groove ourselve
[00:00:29] <ries> then we presses the two ball bearings in there
[00:00:46] <frallzor> but when you mount them to the car how did you do
[00:01:05] <frallzor> check pdf 4
[00:01:08] <ries> Nut and a bolt, and a very thin shim
[00:01:09] <frallzor> side 19 and 22
[00:01:14] <ries> let me open that
[00:01:25] <frallzor> those guys =)
[00:02:00] <ries> Ahh.. I need to do these two still
[00:02:28] <frallzor> then how could you assemble the stuff you have so far?
[00:02:43] <frallzor> arent the page 19 needed for the bearings
[00:03:40] <ries> gosh... where was M1 20 220 needed again...
[00:04:26] <frallzor> between the parts and the v-bearings?
[00:04:36] <frallzor> on the screw holding it in place
[00:07:56] <frallzor> now when you say it i cant find a place for it either
[00:10:02] <frallzor> page 22 I can place, but suddenly 19 isnt findable in the drawings =)
[00:12:56] <ries_> ries_ is now known as ries
[00:13:13] <frallzor> as I said, now I cant find where 19 should be either =)
[00:16:26] <ries> frallzor: I think it was on teh z plate
[00:16:41] <frallzor> ah the spiderplate?
[00:16:49] <ries> so, took me 10 minutes to start the generator and get going again :D
[00:16:54] <ries> yeaa spider plate
[00:17:10] <ries> I need to do the last part of my spider plate still ;)
[00:17:11] <frallzor> not really needed then
[00:17:27] <frallzor> since many just solder nuts to it
[00:18:34] <ries> PS: I also know how the clamp strip works and the whole bigger, we talked about that... remember?
[00:18:51] <frallzor> yup
[00:18:56] <ries> indeed, they need to get soldered
[00:19:51] <ries> PDF 3, page 8... there you go :)
[00:20:09] <frallzor> so we need 6x page 19 and then like.... 2x 22?
[00:20:48] <frallzor> need to get em fabricated asap =)
[00:21:27] <ries> 1x 22 I think, since you will have only one idler
[00:21:51] <frallzor> isnt it 1 for each side?
[00:22:24] <ries> No, teh idler will be mounted on the z-car
[00:22:35] <ries> or y-car (whateher they call it...)
[00:22:48] <ries> one side will have the idler, the other site will have the stepper
[00:22:49] <frallzor> yeah but isnt it 1 on each side of it?
[00:22:55] <frallzor> ah
[00:23:04] <frallzor> you mean the little car with the motor in it
[00:23:14] <frallzor> not the long gantry
[00:23:32] <frallzor> pretty obvious though, gantry has 2 steppers =P
[00:23:50] <frallzor> the idles just act as another stepper not working
[00:23:52] <frallzor> *r
[00:24:27] <jt-plasma> Bridgeport_II: are you around?
[00:25:47] <ries> sorry, I had to replug my router :)
[00:25:53] <ries> one side is a notmal stepper comfiguration,
[00:26:12] <ries> then other side is just the idler so the y car will not top over and stays sturdy on the rails
[00:26:38] <frallzor> yeah I realized that
[00:26:47] <frallzor> only 1 stepper on the y axis =)
[00:31:48] <ries> yup...
[00:32:03] <ries> I was planning to go tomorrow, but that is changed to friday...
[00:32:11] <ries> to work on the MM again
[00:32:33] <ries> I have been playing around with EMC2 a lot (recompiling and stuff) but I think I have to buy a new computer for it :s
[00:36:24] <ries> I got setsmi working though, but then I can only start emc2 once :) then it complains about rtai not being available and all sort of stuff
[00:36:41] <frallzor> I found a kickass touchscreen pc on ebay for $100
[00:37:00] <frallzor> dust proof and all
[00:39:31] <ries> cooool
[00:39:36] <ries> do they have any more?
[00:40:08] <frallzor> seems he had a few more dont know
[00:40:16] <frallzor> but the specs varied a bit it seemed
[00:40:29] <frallzor> my only had 40gig hdd when it said 100 but i was ok with it
[00:40:42] <frallzor> so i hope it doesnt take too long to get here =)
[00:55:42] <oPless> linky ?
[00:56:57] <frallzor> do a "industrial computer" search on fleagay
[00:59:24] <oPless> ta
[00:59:57] <lilalinux> lilalinux is now known as lilalinux_away
[01:23:01] <MarkusBec> MarkusBec is now known as MarkusBec_away
[01:31:49] <jfigie> Is there a list of all of the system variables, for example 5181 = stored x axis absolute position by G30? A complete list would be helpful.
[01:32:44] <jfigie> Is there a way to tell from a Gcode program if G20 or G21 is active?
[01:33:11] <jfigie> From within a Gcode program
[02:51:47] <krushia> http://www.popsci.com/science/article/2009-12/3-d-bio-printer-fabricates-organs-order
[02:53:15] <krushia> now where to we buy the head?
[02:53:24] <krushia> *do
[04:58:16] <cradek> jthornton: thanks for fixing the counterbore program - it's useful
[05:49:52] <Jymmm> So, I go to the shop I usually go to to grab some smokes. On the counter is a display of these lil tubes with a purple rose inside it. The gf like purple and thought it be a nice stocking stuffer. I ask the gal behind the counter what are these for (as they look a tad bit strange) she says "you dont want to know", then it dawns on me, I almost got my gf a crack pipe as a stocking stuffer.
[06:09:32] <Omni5cience> lol
[06:35:27] <Omni5cience> Omni5cience is now known as Omni5
[06:37:05] <Omni5> Omni5 is now known as Omni5cience
[06:37:18] <Omni5cience> Omni5cience is now known as Omni5
[06:37:44] <Omni5> Omni5 is now known as Omni5ci
[06:38:23] <Omni5ci> Omni5ci is now known as Omni5cience
[08:12:45] <MarkusBec_away> MarkusBec_away is now known as MarkusBec
[09:32:34] <lilalinux_away> lilalinux_away is now known as lilalinux
[10:17:26] <lilalinux> lilalinux is now known as lilalinux_away
[10:19:13] <lilalinux_away> lilalinux_away is now known as lilalinux
[11:20:27] <archivist_emc> * archivist_emc tests counterbore, seems fixed :)
[11:23:46] <tarzan> cbore with chamfer?
[11:26:45] <Valen> made my first really 3d part today
[11:28:58] <Valen> came out pretty good
[11:29:06] <Valen> well in MDF anyway ;->
[11:48:34] <archivist_emc> wee 6 compile proce running as well as emc and firefox
[11:50:41] <archivist_emc> although firefox doesnt get much time to do anything, best part of unusable
[11:51:05] <Valen> you know you can own more than one computer ;->
[11:53:07] <archivist_emc> http://www.collection.archivist.info/cbore_works.png
[11:53:59] <archivist_emc> I actually running this as a buildslave to test mysql in a realtime environment
[11:54:17] <archivist_emc> so abuse and it works :)
[11:54:18] <Valen> so the compile is on another machine?
[11:54:25] <archivist_emc> no this
[11:54:38] <archivist_emc> controlled from another
[11:54:46] <Valen> i meant another machine is compiling and this is helping
[11:55:03] <archivist_emc> oh ant this is xchat and web serving also
[11:55:28] <Valen> your bad ;-P
[11:55:57] <Valen> but yeah, Like i was saying, my toolchain has been validated w00t
[11:56:05] <archivist_emc> buildbot is a system to build and test on multitudes of computers
[11:56:31] <Valen> rhino cad + rhino cam > custom post > emc and it makes stuff like its sposed to look even
[11:56:39] <archivist_emc> emc is tested that way too
[11:58:09] <Valen> heh I wonder about passing out compiling jobs to google
[11:58:18] <Valen> via wave or whatever it is
[11:58:29] <Valen> a C compiler written in python
[11:58:31] <Valen> scary
[11:59:25] <archivist_emc> I would never contemplate compiling C with python, pythos is too slow
[11:59:59] <Valen> yeah, but you could throw several hundred/thousand CPU's at it for the few seconds it'd take to compile
[12:00:13] <Valen> probably best suited to something like a kernel I spose
[12:00:49] <Valen> (lots of files to process)
[12:01:24] <Valen> upload them individually all zipped up, and the reply to your http upload is a download with the compiled object in it
[12:01:48] <Valen> probably wouldn't have enough bandwidth to be much faster than just compiling it yourself
[12:01:54] <archivist_emc> plus the shitload of includes :)
[12:02:08] <Valen> nah they would all share amongst themselves for those ;->
[12:02:11] <Valen> upload once
[12:02:21] <Valen> I think a kernel compile takes me something like 20 minutes or so
[12:02:56] <archivist_emc> but all the slaves have to collect, this runs a repositry at each box
[12:03:38] <Valen> I think it works such that each request you make spawns a new process, and generally that will get its own CPU
[12:03:54] <Valen> make each one block until its includes are uploaded
[12:04:06] <archivist_emc> this takes 4-8 hours including all the tests depending on load as its only a 32 bit box
[12:04:27] <Valen> 23/64 shouldn't make a drastic compile time difference?
[12:05:00] <archivist_emc> about 2 hours on the 64bit box
[12:05:13] <Valen> ahh different machine, not different mode ;->
[12:05:26] <Valen> I've got 4x quad cores to throw at things
[12:05:48] <archivist_emc> this is a little P4
[12:11:10] <archivist_emc> hehe m,anaged to crash firefox :)
[12:11:38] <Valen> not that hard in some versions
[12:13:10] <archivist_emc> some...ALL
[12:15:59] <archivist_emc> its really struggling to update ebays realtime action update
[12:22:33] <archivist_emc> some lucky bugger got a hardinge at 500 quid
[13:09:05] <Valen> where would you look for a decent spindle with VFD?
[13:09:25] <frallzor> ebay
[13:09:31] <frallzor> damencnc has some too
[13:09:58] <jthornton> what kind of spindle?
[13:10:06] <Valen> that is a good question too lol
[13:10:08] <Valen> a cheap one
[13:10:17] <frallzor> cheap = what? :P
[13:10:25] <Valen> $5? would be nice lol
[13:10:37] <frallzor> then the one I got isnt a good tip :P
[13:10:49] <jthornton> a rummage sale :P
[13:10:49] <frallzor> but damen has a few others too =)
[13:11:00] <Valen> a friend of ours has got a set of 4x 400W brushless servo motors with drives for about $30 each
[13:11:13] <Valen> each one has a 100:1 harmonic drive gearbox on the front
[13:12:01] <Valen> I'm after something we can put a 20mm end mill in if we are feeling keen but most of its work would be in the 6-12mm range and If it could do engraving sized jobs as well that'd be nice
[13:12:19] <Valen> want to be able to remove stainless or better at a decent rate
[13:12:43] <frallzor> 20mm, cheap, steel
[13:12:53] <frallzor> that doesnt seem to mix to well :P
[13:12:56] <Valen> I dont ask much ;->
[13:13:06] <Valen> ey, if it was easy I wouldn't be asking now would I ;->
[13:13:27] <Valen> currently we are limited to 1600RPM on the head of the current thing
[13:13:48] <Valen> which is really limiting things in terms of feeds and speeds
[13:14:02] <Valen> we can push the 20mm end mill through at 350mm/minute
[13:14:13] <frallzor> http://img4.imageshack.us/i/p1090501g.jpg/ do you want something like that?
[13:14:17] <frallzor> or other type of spindle
[13:14:30] <Valen> yeah something along those lines
[13:14:37] <Valen> very open design at the moment
[13:14:54] <Valen> if it can run an automatic tool changer then that would be *really* nice
[13:15:36] <Valen> the ebay store we used to get stuff off had http://cgi.ebay.com.au/WATER-COOLE-MOTOR-SPINDLE-2-2KW-AND-MATCHING-INVERTER_W0QQitemZ230378759914QQcmdZViewItemQQptZAU_B_I_Electrical_Test_Equipment?hash=item35a3a4e6ea for around $400 and that was nice
[13:15:36] <frallzor> well some can
[13:15:49] <frallzor> yeah those are common on ebay
[13:16:01] <Valen> any good for metal you think?
[13:16:09] <Valen> 8000RPM is pretty fast on a larger cutter
[13:16:20] <frallzor> intend to drill?
[13:16:26] <frallzor> pure drilling that is
[13:16:26] <Valen> and no hope of ridgid tapping I'd think
[13:16:39] <Valen> do we or you mean they are?
[13:16:44] <frallzor> you
[13:16:54] <Valen> presumably there would be some drilling
[13:18:05] <frallzor> Id say that one isnt meant for axial forces
[13:18:06] <Valen> why?
[13:18:14] <frallzor> bearings
[13:18:32] <frallzor> pretty sure it doesnt pack double angular bearings
[13:18:32] <Valen> wouldn't be a lot in comparison to the rest of it
[13:19:12] <Valen> I just see people using them for wood more than metal
[13:19:33] <Valen> I'd rather 1600-15000RPM range
[13:20:26] <Valen> those things all seem to be min speed 8000RPM or so
[13:21:25] <Valen> hmm
[13:21:30] <DaViruz> presumably they could be run down to 0rpm if you wish, but they would be very weak
[13:21:39] <Valen> I wonder about using one of those brushless servo motors as the spindle
[13:21:55] <Valen> they should be good for 5000+ RPM
[13:22:00] <DaViruz> if it produces 2kW at 16000 RPM it will produce like 100W at 1000 RPM
[13:22:22] <Valen> yeah
[13:22:29] <frallzor> milling cheese doesnt sound too bad
[13:22:42] <Valen> why would one do that?
[13:22:49] <DaViruz> i routinely make swiss cheese out of ordinary cheap cheese
[13:22:55] <frallzor> 100W :P what else would it be good for
[13:23:07] <Valen> aluminium from bunnings?
[13:23:15] <Valen> (local hardware chain store)
[13:34:10] <Valen> http://www.cncathome.com/spindles.html this guy is such an aussie ;->
[13:34:22] <Valen> I think I used to live a few minutes up the road from him
[13:37:26] <Valen> check the video at the bottom of the page
[13:47:18] <DaViruz> "now that may just be t he coolest thing i have ever made!"
[13:47:41] <DaViruz> i have to agree with him, it is very neat
[14:03:24] <jepler> awww now we have cats using emc http://wiki.linuxcnc.org/uploads/cam2d_mini.jpg
[14:04:50] <archivist> helicopter frame /me assumes
[14:04:59] <jepler> no idea
[14:14:35] <jthornton> tiny electric rc helicopter
[14:31:55] <archivist> jthornton, in your counterbore prog you dont seem to overlap cut passes, you seem to use the full width of the cutter is that right?
[14:33:06] <jthornton> it should use a percentage of the cutter during the spiral out
[14:36:06] <archivist> one test was 1/5" with a 1/4" cutter and it spiraled in and missed the spiral out
[14:39:20] <jthornton> I think there is a bug that if the cutter is too close to the hole diameter it does that, but I can't find it :/
[14:40:04] <archivist> the spiral in was not on center in that case as well
[14:41:02] <jthornton> it does start off center so it can complete the spiral
[14:41:19] <jthornton> I have to run bbl
[14:41:23] <archivist> ok
[14:44:30] <jepler> A .200 hole with a .250 cutter? I must misunderstand what you're talking about.
[14:46:21] <skunkworks_> .2 must be radius
[14:55:35] <archivist> I mean 1/2"
[14:56:43] <archivist> if the spiral in is on center it would have to spiral out and do overlapping cuts
[15:02:29] <jepler> hm. unicode defines a code point for FRACTION NUMERATOR ONE but not for any other digit. There are superscripted versions of all digits, though.
[15:02:54] <SWPadnos> 1/x is much more common than y/x, I bet
[15:03:01] <jepler> yeah
[15:03:21] <SWPadnos> but the real question is - why are you looking at unicode characters? :)
[15:03:26] <jepler> but if you want to write a function to print any fraction as y/x using unicode, what should you do?
[15:03:50] <Jymmm> jepler: maybe some of the math chars?
[15:04:01] <SWPadnos> print Y, then '/', then X
[15:05:09] <Jymmm> Fractions
[15:05:09] <Jymmm> The fraction slash character (U+2044) allows authors using Unicode to compose any arbitrary fraction along with the decimal digits. Unicode also includes a handful of vulgar fractions as compatibility characters, but discourages their use.
[15:05:42] <SWPadnos> do vulgar fractions argue about who should be on top?
[15:06:13] <archivist> I prefer polite fractions
[15:06:31] <Jymmm> jepler: does it HAVE to be unicode?
[15:06:54] <Jymmm> archivist: liar
[15:06:57] <jepler> actually Unicode and I are having a spat; I think Unicode's decision to represent diaeresis and umlaut with the same code points is unforgiveable
[15:07:09] <Jymmm> ASCII RULES!
[15:08:57] <SWPadnos> EOT
[15:09:16] <jepler> hm, unicode does define U+2044 to do something useful http://en.wikipedia.org/wiki/Mapping_of_Unicode_characters#Fraction_slash but wikipedia notes that hardly any software does that
[15:09:51] <skunkworks_> the video chipset of the atom board works quite well with opengl. It runs the glxgears full screen and pretty smooth.
[15:19:42] <jepler> ugh, gtk2 and tk8.5 both do something different with U+2044 than the standard suggests
[15:20:50] <jepler> but there is a combination that produces not-bad-looking fractions. superscripted digits + u+2044 + subscripted digits
[15:24:08] <skunkworks_> Jymmm: atom - 14,855 so far.
[15:24:34] <jepler> left=tk8.5 (ubuntu 9.10) right=gtk2.8.20 (ubuntu 6.06) http://emergent.unpy.net/files/sandbox/fraction-layout.png
[15:25:15] <Jymmm> skunkworks_: cool beans!!! loaded down and overnight?
[15:25:58] <skunkworks_> nope - about 30 minutes so far.
[15:26:06] <skunkworks_> runing glxgears and youtube.
[15:26:39] <cradek> jepler: that looks quite useful
[15:27:07] <cradek> jepler: might look nice in jthornton's counterbore program
[15:27:18] <jepler> in fact that's exactly why it was on my mind
[15:29:03] <jepler> unfortunately it looks bad in tk8.4 http://emergent.unpy.net/files/sandbox/fraction-layout-tk84.png
[15:29:27] <jepler> that's with font "fixed"; with tk's default helvetica font on Ubuntu 6.06 many of the characters aren't even visible
[15:29:27] <cradek> bleh
[15:34:38] <jepler> sigh. tk didn't look so bad back when X servers were configured with a range of bitmap fonts and nobody expected more than latin1 characters
[15:34:56] <jepler> now that most fonts are scalable fonts and everybody (including me!) wants full unicode, it looks sooo bad
[15:35:39] <cradek> have the bitmap fonts gone away?
[15:35:56] <cradek> I still always seem to have lucidasanstypewriter without doing anything special
[15:37:04] <cradek> (wow, I've used the same font for most of my day-to-day work and play for 15 years)
[15:37:42] <jepler> yes, but it doesn't have superscripted 0, 4..9 or subscripted 0..9 or fraction slash
[15:38:32] <jepler> er, lucidasans doesn't
[15:40:16] <jepler> er, no, it does too
[15:40:33] <jepler> Helvetica doesn't, typo'd fonts don't, but lucidaxxx seems to
[16:02:40] <skunkworks_> Jymmm: 17,734 (about what I remember - works its way close to 20us but doesn't touch it.)
[16:27:04] <Valen> skunkworks, is that single or dual core?
[16:31:57] <skunkworks_> dual core
[16:32:08] <skunkworks_> I am as we speak installing erics smp kernel
[16:32:12] <skunkworks_> just for grins
[16:37:13] <skunkworks_> Is the repository down?
[16:38:59] <skunkworks1> samco@samco-desktop:~$ git clone git://git.linuxcnc.org/git/emc2.git emc2-dev
[16:38:59] <skunkworks1> Initialized empty Git repository in /home/samco/emc2-dev/.git/
[16:38:59] <skunkworks1> fatal: The remote end hung up unexpectedly
[16:38:59] <skunkworks1> fetch-pack from 'git://git.linuxcnc.org/git/emc2.git' failed.
[16:40:06] <skunkworks_> I must be doing something wronge
[16:40:08] <skunkworks_> wrong
[16:45:50] <SWPadnos> do you need to put some anonymous name in there somewhere?
[16:46:29] <SWPadnos> or are you doing a developer checkout (with all the appropriate ssh keys in your ~/.ssh/ directory
[16:46:33] <SWPadnos> )
[16:49:28] <skunkworks_> hmm - I was just following the wiki ;)
[16:52:34] <skunkworks_> ah - that might be the problem... ;)
[16:52:58] <skunkworks_> welcome back cradek
[16:53:06] <cradek> well thanks
[16:53:14] <skunkworks_> yep - that was it.
[16:53:40] <skunkworks_> * skunkworks_ just happened to try a git checkout while you where away..
[16:55:33] <skunkworks_> what are the odds. ;)
[16:56:42] <MattyMatt> Sod's law says he was away checking in a vital patch so you are now out of date
[16:59:42] <MattyMatt> Sod's Evil Uncle's Law say that the battery in his RTC has just died, and his checkin has the wrong timestamp so nobody can touch the repo for 8 hours without causing a desync (that happened to me once)
[17:00:15] <cradek> ugh. must have been in the bad old days.
[17:00:29] <MattyMatt> cvs i remember that much
[17:12:00] <skunkworks_> ok - stupid quesion - how can I tell if realtime is on one core?
[17:12:26] <cradek> if using emc, it always puts all the realtime tasks on one core
[17:13:01] <skunkworks_> I am running latency_test
[17:13:06] <skunkworks_> I am running latency-test
[17:13:36] <SWPadnos> top can tell you where all the processes are
[17:13:50] <SWPadnos> I don't remember the keystroke to show which CPU is being used, but it's in the help
[17:15:03] <celeron55_> pressing '1' will show load on individual cpus
[17:15:28] <SWPadnos> that's overall, there's a different way of adding the CPU NR column to the display
[17:15:39] <SWPadnos> (or whatever they call it :) )
[17:16:33] <danimal_garage> morning
[17:17:35] <celeron55_> f j <enter> will do that
[17:17:51] <celeron55_> i guess
[18:07:40] <lilalinux> lilalinux is now known as lilalinux_away
[18:11:15] <lilalinux_away> lilalinux_away is now known as lilalinux
[19:32:17] <Jymmm> skunkworks: good to hear =)
[19:34:23] <skunkworks> well - I am not getting any better latency with the smp kernel
[19:39:41] <SWPadnos> with or without isolcpus?
[19:49:28] <Jymmm> isowhat?
[19:57:40] <cradek> you can tell linux to not run processes on a certain cpu. then if emc uses it for the realtime threads, there is no competition.
[19:57:52] <Jymmm> ah
[19:58:14] <cradek> isol = isolate
[20:24:40] <MarkusBec> MarkusBec is now known as MarkusBec_away
[20:41:42] <MarkusBec_away> MarkusBec_away is now known as MarkusBec
[21:25:51] <skinnYPuP> I didn't see anyone mention htop so i'll throw that in
[21:37:56] <skunkworks> I will try (don't know if isolcpus=1 is set in the menu.lst)
[21:43:51] <skunkworks> if that is right...
[21:53:14] <SWPadnos> it probably isn't by default
[21:53:53] <SWPadnos> and yes, it's a comma-separated list (used to be a bitmap, but in a 2-processor system it makes no difference :) )
[22:24:52] <MarkusBec> MarkusBec is now known as MarkusBec_away
[22:25:29] <MarkusBec_away> MarkusBec_away is now known as MarkusBec
[23:52:45] <Jymmm> http://www.cnn.com/video/#/video/offbeat/2009/12/22/moos.gift.wrapped.apartment.cnn