#emc | Logs for 2007-01-16

Back
[00:38:37] <fenn> vtk would do very nicely as a cam preview plot.. you could show cutter physics, toolpath, and stock remaining, all at the same time
[00:45:18] <ejholmgren> what's vtk?
[00:45:57] <ejholmgren> never mind .... ooooooo pretty
[00:48:18] <fenn> i'm playing with python-vtk right now
[00:53:51] <ejholmgren> you may have a class action lawsuit on your hands if you can implement all the things in that earlier post
[00:54:05] <fenn> shh
[00:54:19] <ejholmgren> ... for a large dry cleaning bill ;)
[01:56:21] <ds3> is the lack of 3D primitives like spheres, cyclinders, cones, etc an intristic feature of a parametric modeler? (Trying to figure out why I need to extrude stuff to create say a cube in Alibre but Rhino has primatives like a cube)
[01:57:57] <SWPadnos> I've only seen parametric modelers use the "sketch" then "3d-ify" approach
[01:58:44] <fenn> there's no reason a parametric modeler couldn't use spheres/cylinders etc
[01:59:07] <fenn> brlcad has a vast assortment of primitives (but its parametric support is pretty skimpy)
[01:59:34] <SWPadnos> there could be macros or something, but basically you don't need that kind of primitive
[02:01:37] <ds3> I see.
[02:01:52] <ds3> this alibre is still confounding me
[02:01:57] <fenn> so, a bevel around a circular extrusion is good enough for a cone?
[02:02:37] <SWPadnos> in a parametric modeler, you'd draw your lines (with constraints that the endpoints are coincident), then constrain the Y dimension to be the same as the X, then after the extrude, you can constrain the extrude height to be the same as the X dimension
[02:02:46] <SWPadnos> then, if you change X, the whole thing changes size ...
[02:03:09] <SWPadnos> no - I'd make a cone. you should be able to loft a circle to a point
[02:03:37] <SWPadnos> and then you can actually have the point off the circle centerline
[02:03:52] <ds3> still isn't sinking in... how would a sphere work then?
[02:04:11] <SWPadnos> make an arc in some plane, and sweep it in a circular arc
[02:06:47] <ds3> Hmmm
[02:10:33] <fenn> this is something off along a different vector, has nothing to do with whether it's parametric or not
[02:13:05] <ds3> hmm
[02:29:05] <SWPLinux_> SWPLinux_ is now known as SWPLinux
[02:51:26] <CIA-8> 03jmkasunich 07TRUNK * 10emc2/src/hal/utils/halcmd.c: bugfix: loadusr searches in multiple places for the program to run - but didn't use the search result, only the program name
[02:53:18] <jmkasunich> why is CIA-8 saying "TRUNK" instead of "HEAD"?
[02:54:48] <SWPLinux> methinks somebody changed it to say the right thing
[02:56:40] <jmkasunich> probably that cradek fellow
[02:56:53] <SWPLinux> he's not here - he's the asshole :)
[02:58:50] <ejholmgren> hrmmnn... how long will a regular 12v dc computer fan last at 24v ?
[02:58:58] <jmkasunich> not very
[02:59:13] <ejholmgren> I finally finished the heater for my orchid enclosure downstairs
[02:59:22] <ejholmgren> it looks like a sputnik
[03:00:01] <ejholmgren> I think the heating elements will work at 12v ... but they're 80W each
[03:00:32] <ejholmgren> might be time to hack up a psu
[03:01:12] <SWPLinux> you could - gasp - use a computer PSU ...
[03:03:12] <ejholmgren> that assumes that I have a pile of them in the basement
[03:03:20] <ejholmgren> which I do :)
[03:03:54] <SWPLinux> by george, I think he's got it!
[03:04:01] <SWPLinux> or is that Jove?
[03:04:13] <jmkasunich> george is jove
[03:04:16] <jmkasunich> oops
[03:04:27] <jmkasunich> george is jove's younger, smarter brother
[03:06:13] <SWPLinux> jumpin george-hosephat!
[03:34:58] <owad> Has anybody gotten a chance to look at Digital Machinist magazine?
[03:36:54] <SWPLinux> sadly, no. it's only sent to subscribers to "Home Shop Machinist", I think
[03:38:26] <owad> That was the case with the first issue. It's now a separate magazine (or is about to be - their website isn't clear if it's out yet).
[03:38:42] <owad> http://www.homeshopmachinist.net/?page=main.features.digitalmachinist
[03:40:01] <SWPLinux> anyway - do you have it?
[03:40:20] <cradek> hey
[03:40:26] <SWPLinux> their site is mostly not functional for me
[03:40:35] <SWPLinux> heh - oops :)
[03:40:40] <owad> no, I haven't seen it.
[03:40:44] <jepler> hi cradek
[03:40:53] <cradek> hi
[03:41:02] <skunkworks> http://www.cnczone.com/forums/showthread.php?t=30423
[03:41:17] <cradek> hope I didn't put us on the spot by saying a date on the list
[03:42:10] <owad> I asked HSM about it back in may and they had this to say: "Our special edition of HSM is the
[03:42:11] <owad> forerunner. As the new title suggests, the subjects will cover the digital
[03:42:11] <owad> controlled aspects of machining (CNC, CAD, CAE, CAM) on one end to machining
[03:42:11] <owad> for digital applications (robotics, mechanotronics) on the other." Sounds like it has the potential to be a great magazine.
[03:43:43] <cradek> skunkworks: I have a similar pwm signal for the spindle control on the nist-lathe.
[03:44:15] <skunkworks> cradek: I know. Your using a simple filter to do it - right?
[03:44:34] <SWPLinux> not for 0-10V ...
[03:44:37] <skunkworks> pwmgen out of the printer port?
[03:45:01] <SWPLinux> it's funny that people think that Mach is great because it does threading with a 1PPR signal
[03:45:06] <skunkworks> swplinux: right
[03:45:07] <cradek> skunkworks: yes (it had a jumper that allowed 0-5v input
[03:45:16] <skunkworks> nice.
[03:45:20] <cradek> )
[03:45:59] <cradek> does anyone know if mach has lathe tool shape compensation and offsets?
[03:46:13] <skunkworks> swplinux: mach pays for advertising on that site. it is very mach biased.
[03:46:29] <SWPLinux> so's almost everyone on the CCED and geckodrive gourps
[03:46:34] <SWPLinux> it sometimes gets annoying
[03:46:39] <skunkworks> very
[03:46:47] <skunkworks> I try to ignore it ;)
[03:47:34] <skunkworks> http://www.cnczone.com/forums/showthread.php?t=11665
[03:48:06] <SWPLinux> I sometimes post responses :)
[03:48:19] <SWPLinux> (always professional and responsive, of course ;) )
[03:48:53] <owad> How do Mach3 and EMC2 compare? I've yet to use either.
[03:49:08] <SWPLinux> funny - I just responded to a similar question on CCED or Geckodrive ...
[03:49:13] <cradek> you're in the wrong place for an unbiased answer :-)
[03:49:17] <skunkworks> very
[03:49:17] <owad> heh
[03:49:18] <SWPLinux> owad: well, in here, we think EMC is much better ;)
[03:49:30] <cradek> (I haven't used Mach)
[03:49:50] <SWPLinux> I've tried it, and it basically wreaked havoc on the machine (PC) while it was running
[03:49:51] <skunkworks> but but mach can do threading with 1 spindle encoder pulse.
[03:50:06] <owad> I read an article, I think in HSM a while back that implied emc didn't support nearly as many G-Codes as Mach3 did. Is that the case?
[03:50:13] <SWPLinux> * SWPLinux gets the wet noodles out
[03:50:21] <SWPLinux> that's partly true
[03:50:51] <cradek> do you know what functionality corresponds to those gcodes?
[03:51:07] <owad> no. I'm not sure I'd even be able to find the article. This was at least a year ago.
[03:51:10] <SWPLinux> there are non-standard codes that are in Mach, to do things like drill down, move out to a radius, do a full circle, then go back to center and raise
[03:51:41] <SWPLinux> however, EMC2 has subroutines, which can do the same thing (and are also nonstandard)
[03:51:45] <cradek> interesting (and easy to do)
[03:52:09] <SWPLinux> right. G12 and thereabouts, I think (I looked it up when there was a question on some list)
[03:52:27] <cradek> I have not heard anyone ask for those, but if users wanted them, we could probably do them pretty easily
[03:52:28] <owad> What CAM software do you guys use?
[03:52:31] <SWPLinux> the Mach manual is freely downloadable. it might be good to make sure there's a macro for any of the useful additions
[03:53:03] <SWPLinux> that's the main problem with emc comapred to Mach (or DeskCNC)
[03:53:31] <SWPLinux> getting better, but it's still not quite as easy to just load a bitmap or dxf or stl file and machine it
[03:53:54] <cradek> have you tried bitmaps lately?
[03:54:03] <SWPLinux> I think both Mach and DeskCNC also have some rudimentary conversational programming mode
[03:54:06] <cradek> jepler's work on that is pretty amazing I think
[03:54:31] <SWPLinux> no, but I can't load the bitmap - the emc method is to filter it first (so I always look at G-code in Axis ...)
[03:54:40] <SWPLinux> I agree - just not as seamless
[03:55:05] <SWPLinux> I don't think that's a bad thing, it's a philosophical difference between the Windows crowd and the Unix crowd
[03:55:06] <skunkworks> emc is better - free and I can talk to the developers at almost any time. Even about my personal problems.
[03:55:11] <SWPLinux> shut up
[03:55:14] <SWPLinux> :)
[03:55:18] <cradek> haha
[03:55:44] <cradek> I seem to either write gcode by hand or generate it with very simple programs
[03:56:12] <SWPLinux> right. either method is great for some people, and completely frightening to others
[03:56:15] <owad> I'd like to be able to go SketchUp Pro -> ??? -> EMC2
[03:56:44] <SWPLinux> what does SketchUp Pro output? (dxf, stl, ...)
[03:57:49] <owad> in 3d: dwg, dxf, obj, 3ds, fbx, vrml, xsi
[03:58:17] <owad> in 2d: dwg, dxf, and various image formats
[03:58:34] <cradek> triangle dumps maybe
[03:59:07] <cradek> (I'm nervous about any software other than autocad that generates dwg files)
[03:59:13] <skunkworks> mostly acad and some sort of post proccessor that takes dxf and converts it to gcode. Like ace converter or the dedicated post we have a work.
[04:00:55] <owad> Is Ace Converter Windows only?
[04:01:27] <A-L-P-H-A> SWPadnos?
[04:01:32] <SWPLinux> what?
[04:01:38] <A-L-P-H-A> SWPadnos?
[04:01:40] <SWPLinux> why?
[04:01:42] <A-L-P-H-A> SWPadnos?
[04:01:52] <SWPLinux> duh!
[04:01:58] <A-L-P-H-A> SWPLinux?
[04:02:03] <SWPadnos> duh!
[04:02:11] <skunkworks> http://www.dakeng.com/ace.html
[04:02:13] <A-L-P-H-A> owad, www.winehq.com or was it .org
[04:02:38] <SWPLinux> I've successfully run SolidWorks in VMWare
[04:02:42] <SWPLinux> and that's a hog
[04:03:12] <A-L-P-H-A> SWPadnos, but don't you have 64bit?
[04:03:17] <SWPLinux> I haven't tried STLWorks yet though
[04:03:21] <SWPLinux> yes
[04:03:21] <owad> I do most of my work on the Mac, but I don't mind running X11 apps
[04:03:34] <SWPLinux> quad cores, 4G, Opterons, 7800GT ...
[04:04:03] <cradek> ace convertor seems to run just fine in wine
[04:04:57] <A-L-P-H-A> SWPadnos? and you don't play UT99?? :(
[04:05:04] <SWPLinux> nope
[04:05:13] <SWPLinux> pretty much only wolf:ET for me
[04:05:14] <A-L-P-H-A> SWPadnos, jesus... I'll tried you my 5200 for your 7800....
[04:05:23] <A-L-P-H-A> you'll still be my moving target in that.
[04:05:29] <SWPLinux> no. I think I'll keep it (and the dual Quadro FX3500s)
[04:06:01] <SWPLinux> I only get around 90 FPS at 1920x1200 with wolf:ET
[04:06:17] <A-L-P-H-A> SWPadnos. mofo.
[04:06:31] <SWPLinux> sadly, it won't run on both monitors :)
[04:06:52] <SWPLinux> unlike the Parhelia (which I've never gotten to work in surround gaming mode)
[04:06:56] <owad> TurboCNC calls their product "the best open-source CNC control around…" and then they write: "Source code and free support is yours when you register this program for $60 US."
[04:07:26] <SWPLinux> right - you can pay to get the source, so it's open / viewable, but not free/libre ...
[04:07:32] <cradek> RMS would say "open source" means little enough that it's proper usage
[04:07:40] <cradek> EMC2 is real Free software
[04:07:42] <A-L-P-H-A> owad, it's still open sourced... just not free open sourced.
[04:07:48] <cradek> http://timeguy.com/cradek-files/emc/ace-convertor-sample-nc.png
[04:08:02] <cradek> this is what ace convertor gives me from its sample.dxf
[04:08:08] <cradek> no idea if it's right - seems unlikely.
[04:08:35] <owad> oh wow, I thought it was a command line tool.
[04:08:36] <A-L-P-H-A> cradek, it was to show general crap... it doesn't do a bad job at all in 2d... it does not do 3d.
[04:09:09] <cradek> A-L-P-H-A: I don't follow - are you saying this looks right?
[04:09:25] <SWPLinux> owad: it is - that's the preview when you load it in EMC using the AXIS front end
[04:09:30] <cradek> owad: what is?
[04:09:38] <A-L-P-H-A> cradek, yeah... that file was dumb... it was just to show you rapids and arcs, and stuff... it's useless, but ace does work.
[04:09:48] <cradek> A-L-P-H-A: ok I understand
[04:10:04] <cradek> it generates bad gcode (no F specification and no program end code) but that's easy to fix
[04:10:37] <owad> cradek: I'm not sure what you're asking.
[04:10:57] <cradek> sorry
[04:11:02] <cradek> you thought what is a command line tool?
[04:11:19] <owad> Ace Converter
[04:11:31] <owad> then SWPLinux clarified that it was
[04:11:31] <cradek> no, it is a gui app, but only just barely
[04:11:40] <owad> ah ok
[04:11:53] <cradek> it pretty much has "open" and "go" buttons
[04:12:10] <cradek> no meaningful gui interaction
[04:12:15] <owad> I was looking at this free competitive upgrade to madCAM today, but I think it requires a Rhino license: http://www.madcamcnc.com/competitive_upgrade.html
[04:12:37] <owad> good deal if any of you have Rhino, though
[04:12:38] <cradek> my screenshot is EMC2 showing a preview of its output
[04:13:07] <cradek> heh, wonder if REALIZE counts
[04:13:38] <A-L-P-H-A> you know what's fucked up... when I started univeristy, in 97... I'd been on the net for 3 years already... so since 94... and bbs probably since 1990. Can you remember back when you didn't have internet?
[04:13:45] <A-L-P-H-A> wth did you do with your time?
[04:13:56] <cradek> A-L-P-H-A: BBSes
[04:14:03] <jmkasunich> real life?
[04:14:30] <A-L-P-H-A> cradek, die spelling nazi llama :)
[04:14:35] <A-L-P-H-A> jmkasunich, yeah...
[04:14:40] <cradek> eh?
[04:14:48] <A-L-P-H-A> I remember watching cartoons... cause I was 10 in 1990.
[04:14:53] <cradek> no I mean that's what I did before I had internet access
[04:14:53] <SWPLinux> die spelling dalai lama?
[04:15:12] <SWPLinux> heh
[04:15:15] <A-L-P-H-A> cradek, oh thought you were correcting me.
[04:15:16] <cradek> ugh, ACE cuts stuff in a very strange order
[04:15:34] <A-L-P-H-A> cradek, no it doesn't actually... it converts based on draw order of the dxf.
[04:15:53] <A-L-P-H-A> cradek, it just parses the DXF, without priority to the next nearest line, etc.
[04:15:58] <SWPLinux> does it do pockets correctly?
[04:16:04] <cradek> in the sample it has a tool path that's cut in three depths
[04:16:09] <A-L-P-H-A> SWPadnos, it only does 2d.
[04:16:14] <cradek> it does three depths of the first segment, then three depths of the second segment, etc
[04:16:21] <SWPLinux> so it doesn't do pockets then?
[04:16:22] <A-L-P-H-A> SWPadnos, but if drawn in a certain order, yeah, it does pocket
[04:16:24] <skunkworks> I just played with legos befor the internet.
[04:16:34] <SWPLinux> fischertechnik, baby!
[04:16:41] <A-L-P-H-A> SWPadnos, but only at a 2d level... so you'd have to repeat given sections manually.
[04:16:43] <cradek> it seems like REALIZE but a little dumber, and requires windows
[04:16:47] <SWPLinux> but my uncle never let me use his
[04:17:37] <skunkworks> I guess I did copy games with 'fast hackum' on my commador
[04:17:52] <SWPLinux> ok - that's the main thing I see with a lot of the dxf and STL converters - they either don't do the right thing with areas to be removed (with only the outline in the DXF file), and/or they don't do multiple passes at reasonable cut depths
[04:18:05] <owad> What's realize?
[04:18:23] <cradek> owad: gcode output that runs inside autocad (also real Free)
[04:18:39] <cradek> SWPLinux: how do you define the area to be removed with dxf?
[04:19:20] <SWPLinux> you'd probably need a GUI so soemone could select an entity (or chain) nad say "this is a pocket, cut it to a depth of 1.5" in 0.1" passes"
[04:19:22] <cradek> to me, if you use a vector format to represent areas, you're not going to be very happy
[04:19:33] <skunkworks> Time for bed. Night guys.
[04:19:37] <cradek> yeah I suppose
[04:19:39] <cradek> bye skunkworks
[04:19:59] <SWPLinux> the idea would be to be able to make an engineering drawing of soemthing (like you'd hand a machinist), and generate the code you need to make the part
[04:20:17] <A-L-P-H-A> SWPadnos... should I be checking my mail box every day now? :D
[04:20:28] <SWPLinux> it's not that hard - lerman's sample circular pocket does the right thing (from the image he had)
[04:20:47] <cradek> with offset/array it's pretty easy to generate toolpath for a pocket in autocad
[04:20:49] <SWPLinux> A-L-P-H-A: no - the post office was closed today for the holiday
[04:21:00] <cradek> then REALIZE can export it directly
[04:21:03] <A-L-P-H-A> oh... so I may get it by weeks end.
[04:21:11] <A-L-P-H-A> or early next week.
[04:21:15] <SWPLinux> maybe - or saturday since it'll be the mail
[04:21:28] <A-L-P-H-A> Canada != saturday mail delivery.
[04:21:33] <SWPLinux> bummer for you
[04:21:37] <A-L-P-H-A> yuppers.
[04:21:59] <SWPLinux> then maybe next week - I dunno how paranoid they with things going north across the border
[04:21:59] <A-L-P-H-A> surprised USPS has sat delivery still... but I guess they have to compete against UPS, FEDEX, DHL, etc.
[04:22:10] <A-L-P-H-A> SWPadnos, to Canuckia... not that bad...
[04:22:18] <SWPLinux> fedex actually didn't charge any extra for the saturday deliveries ...
[04:22:22] <A-L-P-H-A> We're the ones shipping drugs to you remember.
[04:22:27] <SWPLinux> (ground, residential delivery)
[04:22:31] <SWPLinux> yes, I know
[04:23:17] <SWPLinux> cradek: the idea would be for a person with a machine / EMC2, but not Autocad, to be able to machine something from a dxf file
[04:23:35] <SWPLinux> the same file that would be used to generate a print for a machine shop
[04:23:37] <cradek> ah
[04:23:43] <cradek> that's a bigger problem
[04:23:45] <SWPLinux> yep
[04:23:48] <cradek> you do need some kind of gui thing
[04:23:57] <SWPLinux> that's the problem people like me have ;)
[04:24:07] <SWPLinux> (not that i'm planning to hire out as a machine shop any time soon)
[04:24:33] <cradek> you could plot the dxf and look at the drawing, then write your gcode
[04:24:35] <SWPLinux> isn't there a CHAIN element in dxf?
[04:24:38] <SWPLinux> heh
[04:24:44] <cradek> what's chain?
[04:24:48] <SWPLinux> load it in dia or whatever, then do it all by hand ..
[04:25:11] <SWPLinux> a series of primitives that form a larger element
[04:25:13] <SWPLinux> sort of
[04:25:23] <cradek> like group?
[04:25:27] <SWPLinux> could be
[04:25:31] <cradek> autocad has block/explode
[04:25:38] <SWPLinux> what would you get if you drew a rectangle then filleted the corners?
[04:25:42] <cradek> but for toolpaths polyline is the obvious entity to use
[04:25:47] <cradek> a polyline
[04:25:51] <SWPLinux> ok - then a polyline
[04:26:19] <SWPLinux> this would be 2.5D only, of course
[04:26:24] <cradek> right
[04:26:46] <SWPLinux> within autocad, the file would ideally have things to be milled to a certain depth on the same layer
[04:26:54] <SWPLinux> multiple layers for multiple depths
[04:27:09] <jmkasunich> my 2 cents: drawing to g-code without review by a human machinist (and likely much tweaking) = busted tools
[04:27:09] <cradek> I use layers for tools. depths are represented as ... depths
[04:27:36] <cradek> jmkasunich: that's very true.
[04:27:46] <SWPLinux> ok - that works also, but that's assuming that the person doing the drawing knows the machine it'll be made on
[04:28:12] <SWPLinux> if I use a BP, I can probably use different tools than someone using a sherline
[04:28:25] <SWPLinux> and / or different depth passes
[04:28:42] <SWPLinux> (hence the need for human review ...)
[04:28:44] <cradek> my goal was to use autocad as an interactive tool to draw paths/shapes to make gcode for my machine. sounds like your goal is different.
[04:28:52] <SWPLinux> yep - I think so
[04:29:15] <cradek> you just can't take any old dxf and make gcode from it... the person drawing has to play nice with you.
[04:29:24] <SWPLinux> I'm talking about a tool to let one load a dxf file, whereas you made a different thing - to allow one to use autocad to make G-code
[04:29:32] <cradek> right
[04:29:41] <SWPLinux> it certainly can't be automatic
[04:29:55] <cradek> to me, dxf files are for plotting. until they're read by eyeballs, they're a terrible part description
[04:30:07] <cradek> (and then sometimes still)
[04:30:08] <SWPLinux> hmmm - more projects. I guess I should see why halcmd segfaults for me now ;)
[04:30:10] <SWPLinux> heh
[04:30:24] <jmkasunich> are you running the latest?
[04:30:31] <SWPLinux> even later than that :)
[04:30:42] <cradek> you may as well use postscript or hpgl or any other 2d vector format to describe your parts. same problems.
[04:31:02] <SWPLinux> yep
[04:31:07] <SWPLinux> or bmp, for that matter
[04:31:16] <cradek> well that's worse still
[04:31:18] <SWPLinux> heh
[04:31:35] <cradek> but some people do expect to be able to load a BMP into their machine controller and get ... something? what they want?
[04:31:42] <SWPLinux> as is STL (which does have the saving grace of actually describing a solid)
[04:31:54] <SWPLinux> that's what they can do with deskCNC and Mach
[04:31:55] <cradek> I have no interest in writing mind-reading software. that's for the commercial folks.
[04:32:08] <SWPLinux> load BMP, set resolution / milling area / height scaling, and go
[04:32:37] <cradek> afaik, STL describes triangles, not solids
[04:32:52] <SWPLinux> yes - it's faces (arbitrary number of points, actually)
[04:32:56] <cradek> you might get closed volumes, but you have no guarantee of that
[04:33:10] <cradek> oh ok, not just triangles
[04:33:29] <SWPLinux> yes - it's random as to what happens if the solid isn't closed, or if any face contains more than 3 points :)
[04:33:38] <cradek> haha
[04:33:39] <SWPLinux> but it's in the spec
[04:33:49] <cradek> people generally expect riangles?
[04:33:50] <cradek> t
[04:34:08] <A-L-P-H-A> are "beans" in the veggetable food group, or grains?
[04:34:09] <SWPLinux> there are also two (indistinguishable but different) methods of specifying the color of a face
[04:34:14] <SWPLinux> no, they're legumes
[04:34:15] <cradek> dxf 3dfaces can be quads too, but I bet few softwares handle it
[04:34:26] <A-L-P-H-A> SWPadnos, but there's like 4 food groups, what do beans go under?
[04:34:45] <cradek> A-L-P-H-A: they go with the lentils
[04:34:48] <SWPLinux> starch / veggies / I don't know
[04:35:08] <cradek> and peanuts?
[04:35:11] <SWPLinux> as Roman Moronie would say: "That's a fargin' silly question"
[04:35:16] <SWPLinux> ~
[04:35:18] <SWPLinux> !
[04:35:21] <A-L-P-H-A> http://en.wikipedia.org/wiki/Image:MyPyramid1.png
[04:35:27] <A-L-P-H-A> Means & Beans
[04:36:00] <SWPLinux> beans comtain incomplete protein (often completed by either rice or milk), so they are similar to meats in a way
[04:36:23] <A-L-P-H-A> wonder where mushrooms come in...
[04:36:26] <cradek> * cradek cries for the tax dollars that paid for that picture
[04:36:31] <SWPLinux> vegetables (fungi)
[04:36:33] <SWPLinux> heh
[04:36:47] <SWPLinux> I wonder if the dairy association had to lobby to get milk its own category
[04:36:49] <jmkasunich> depends on what kind of mushrooms
[04:37:08] <cradek> where do sausage with chocolate chip pancakes go?
[04:37:11] <A-L-P-H-A> SWPadnos, why lobby, they're subsidized by the government.
[04:37:21] <SWPLinux> graimeatables
[04:37:48] <SWPLinux> oh - chocolate chips, not blueberries
[04:37:51] <cradek> on a stick
[04:37:57] <SWPLinux> that's the wood group
[04:38:11] <jmkasunich> A-L-P-H-A: the government gives the dairy org subsidies, so they can give them back as lobbying
[04:38:30] <cradek> http://www.jimmydean.com/products.asp?p=9
[04:38:45] <SWPLinux> that is a sick product
[04:39:05] <cradek> you don't like chocolate chip sausage pancake on a stick?
[04:39:05] <jmkasunich> yep, I thought you were making it up
[04:39:10] <cradek> nope.
[04:39:17] <SWPLinux> dunno - never tried it
[04:39:18] <A-L-P-H-A> if you're really think about milk, it's a gross thing... it's mammory secretions from another species of mammals... you're not supposed to drink it... except for the fact that Europeans have a mutant gene that allows for them to do so.
[04:39:25] <jmkasunich> I thought I liked chocolate everything, but not that
[04:39:29] <A-L-P-H-A> but we've gone through this conversation before
[04:39:36] <owad> Anybody heard of CADPACK, built using Open Cascade? Lower left-hand corner: http://www.opencascade.org/showroom/screenshots/
[04:39:40] <SWPLinux> at least a corollary conversation
[04:39:44] <cradek> A-L-P-H-A: that's why I prefer it congealed into cheese
[04:39:51] <SWPLinux> or yogurt
[04:39:51] <jmkasunich> A-L-P-H-A: say what you will, nothing goes better with a warm chocolate chip cookie
[04:40:03] <A-L-P-H-A> cradek... I like cheese... I can handle cheese, but not milk, as I'm lactose intolerant.
[04:40:06] <cradek> jmkasunich: I disagree - another warm chocolate chip cookie
[04:40:12] <SWPLinux> with more milk :)
[04:40:15] <SWPLinux> or coffee
[04:40:21] <cradek> mmm
[04:40:21] <A-L-P-H-A> jmkasunich, yes there is... a hot cup of coffee with baileys in it.
[04:40:31] <SWPLinux> that'll do!
[04:40:33] <jmkasunich> coffee = yuck
[04:40:34] <cradek> I don't like milk - I drink soy or grain milks
[04:40:48] <SWPLinux> I don't like soy or grain milks - I drink milk
[04:40:49] <A-L-P-H-A> homo soy milk is good... if sweetened...
[04:40:59] <A-L-P-H-A> or if you mix it with a protein shake, instead of milk.
[04:41:02] <SWPLinux> ewww
[04:41:07] <cradek> some are good and bad, depends on the brand
[04:41:11] <SWPLinux> you're talking like my wife now
[04:41:17] <SWPLinux> and my sisters
[04:41:19] <cradek> I think some straight soy milks are very gritty
[04:41:22] <SWPLinux> and my mother sometimes ...
[04:41:28] <A-L-P-H-A> cradek, I like the chinese ones... the ""Soy Good" brand tastes nasty to me
[04:41:40] <cradek> haven't tried that
[04:41:48] <A-L-P-H-A> SWPadnos, any of your sisters, single, and 20 something? :D
[04:41:48] <SWPLinux> edensoy is the one we always seem to have around
[04:41:56] <SWPLinux> hah - they're all older tahn me
[04:41:59] <cradek> I think edensoy is one of the bad ones
[04:42:00] <A-L-P-H-A> gah....
[04:42:01] <SWPLinux> except for my half sister
[04:42:16] <cradek> don't know why people like it... but I know some who do
[04:43:11] <cradek> http://www.pacificfoods.com/products-nut-grain.php
[04:43:15] <A-L-P-H-A> http://en.wikipedia.org/wiki/Image:USDA_Food_Pyramid.gif
[04:43:25] <cradek> I like the pacific multi-grain, or Silk
[04:43:51] <A-L-P-H-A> cradek, I'd give it at least a try...
[04:44:02] <A-L-P-H-A> silk... eew... I don't like that one either.
[04:44:15] <cradek> it's ok, but not as good as the pacific
[04:44:16] <A-L-P-H-A> fresh sweetened soy bean drinks are yummy... but only fresh.
[04:44:21] <cradek> better than edensoy :-)
[04:44:38] <A-L-P-H-A> stuff in a box doesn't seem right to me.
[04:44:41] <cradek> there's such a variety I bet everybody can find one they like
[04:44:44] <A-L-P-H-A> milk in Canada comes in a bag.
[04:45:03] <A-L-P-H-A> 4L of milk = 3 bags of milk, that fit inside a milk holder.
[04:45:24] <A-L-P-H-A> no cardboard... oddly enough.
[04:45:29] <cradek> we have gallons in plastic jug, half gallons in waxed cardboard boxes
[04:45:49] <jtr> anyone know why xrdb is interpreting the numbers in axis_light_background as octal?
[04:45:51] <SWPLinux> wow - I actually found a "poem" of my -alf sister's that I like
[04:45:52] <A-L-P-H-A> we have gallon plastic jugs too... but that's a lot of freak'n milk to go through...
[04:46:15] <A-L-P-H-A> SWPadnos, as opposed to the half sister(s) you don't like?
[04:46:28] <cradek> jtr: how are you loading it?
[04:46:29] <SWPLinux> err - no :)
[04:46:48] <A-L-P-H-A> SWPadnos, does you having long hair, influenced by being surrounded by so much estrogen?
[04:47:13] <SWPLinux> it's a comparison of our dying (now dead) aunt to John Glenn in space
[04:47:23] <SWPLinux> I don't think the hair is influenced by women at this point
[04:47:31] <jtr> I'm not - just looking at Manfredi's question. The numbers in the file are actually in hex.
[04:47:38] <cradek> I'm glad there aren't photos of me, online, that are old enough to show when I used to have hair
[04:47:42] <SWPLinux> heh
[04:48:05] <jmkasunich> you had hippie hair?
[04:48:15] <A-L-P-H-A> cradek, aren't there photos of you looking like bigfoot, with a beard and glasses?
[04:48:20] <jmkasunich> thats me
[04:48:23] <cradek> yeah, to my belt
[04:48:29] <cradek> haha, that's all of us
[04:48:42] <A-L-P-H-A> uuuuuuuuuuummm... not me...
[04:48:45] <jtr> but the command line he used looks the same as jepler recommended.
[04:48:47] <cradek> you have to be more specific about which bigfoot
[04:48:51] <A-L-P-H-A> I can't grow a beard if my life depended on it.
[04:49:20] <A-L-P-H-A> I got ID'd saturday night by a bouncer. :(
[04:49:29] <A-L-P-H-A> I think I'm going to be ID'd for a few years to come.
[04:49:57] <cradek> jtr: I get it too
[04:50:01] <SWPLinux> hey - I didn't know she was a college professor
[04:50:14] <A-L-P-H-A> who?
[04:50:17] <jmkasunich> not in touch with family much?
[04:50:19] <SWPLinux> weird what the web brings up on your siblings
[04:50:21] <A-L-P-H-A> she who?
[04:50:22] <SWPLinux> heh
[04:50:28] <SWPLinux> my little sisterlet
[04:50:32] <A-L-P-H-A> I didn't know this girl I was interested in, was my profs daughter...
[04:50:39] <A-L-P-H-A> went out a few times, and the prof didn't know.
[04:50:44] <jmkasunich> uh-oh
[04:50:44] <SWPLinux> heh - oops (until that class is over)
[04:50:55] <A-L-P-H-A> but somehow I think he found out... but he liked me... so I got a awesome mark in that course.
[04:51:49] <cradek> jtr: I have no idea. I find it hard to believe that xrdb is broken but it looks that way...
[04:52:35] <A-L-P-H-A> it was a 4th year, artificial intelligence course... I wrote this mastermind playing game in delphi, and it was good, and couldn't lose... he was impressed, so I got perfect, and he had the descrition of awarding an extra 10% for some messed up reason... so I ended up getting 110% on that assignment which boasted up my mark a little. :D I was pretty damn happy.
[04:53:37] <jmkasunich> anybody familiar with waitpid()?
[04:53:46] <jmkasunich> I think loadusr has a couple problems
[04:54:14] <jtr> cradek: I was about to tell him that 8 wasn't an octal digit until I saw that the file had hex in there anyway.
[04:54:47] <jmkasunich> if you've called waitpid once, and it has returned that the process has terminated, what happens if you call it again?
[04:55:04] <jmkasunich> man says it waits for a change of state, and there ain't gonna be another change
[04:55:08] <cradek> jtr: pretty sure X colors are always specified in hex. I don't know what it's thinking
[04:57:05] <jtr> cradek: I'm not very familiar with X - was just now skimming the man page for xrdb.
[04:58:39] <jtr> or x configuration, rather. Oops - S.O. says it's bedtime. Night, all.
[04:58:53] <jmkasunich> goodnight
[04:59:22] <SWPLinux> see you jmkasunich
[04:59:29] <SWPLinux> oh - wat :)
[04:59:35] <SWPLinux> see you jtr :)
[04:59:34] <jmkasunich> I'm not leaving (yet)
[04:59:47] <SWPLinux> missed that last bit of readback
[04:59:57] <jmkasunich> still wrestling with loadusr
[05:01:23] <cradek> the xrdb source doesn't give up its secrets easily
[06:40:19] <CIA-8> 03jmkasunich 07TRUNK * 10emc2/src/hal/utils/halcmd.c: call hal_ready() after loadusr, make loadusr deal with -w, -W, and even -wW correctly, handle case where -W program exits without ever becoming ready
[06:50:29] <CIA-8> 03swpadnos 07TRUNK * 10emc2/src/hal/utils/halcmd.c:
[06:50:29] <CIA-8> Fixed several possible buffer overflows in replace_vars (pointed out by jmkasunich)
[06:50:29] <CIA-8> Fixed detection of missing close parens/brackets
[06:50:29] <CIA-8> Improved error messages from failed variable replacement
[06:55:42] <K`zan> Night folks
[07:03:18] <K`zan> Night all.
[07:11:21] <tomp> pyvcp-dro-Zonly diagram http://imagebin.org/6930
[09:10:35] <alex_joni> SWPadnos: ohh.. a commit
[09:37:00] <A-L-P-H-A> oh hey! scribus looks neat.
[09:48:30] <anonimasu> hm
[09:48:37] <anonimasu> I found linear motors at my reseller..
[09:48:50] <anonimasu> 3m/sec max speed ;)
[09:49:10] <anonimasu> +/- 2um
[09:49:21] <anonimasu> repetability
[09:49:24] <anonimasu> :d
[09:51:23] <alex_joni> ooh
[09:51:32] <alex_joni> http://www.youtube.com/watch?v=6CsVkYmznVY
[09:52:43] <anonimasu> I'm going to check the price :D
[09:55:40] <slundell> http://www.yankodesign.com/product_info.php?products_id=641
[09:59:48] <awallin> gday
[10:00:05] <alex_joni> hi
[10:00:31] <awallin> some interesting exchange of comments with the Julian Todd of freesteel: http://www.freesteel.co.uk/wpblog/2007/01/embarrassing-toolpath-ordering/
[10:00:47] <awallin> and http://www.freesteel.co.uk/wpblog/2006/12/the-offset-ellipse/
[10:00:54] <awallin> let's see what he thinks about my math...
[10:01:34] <anonimasu> brb
[10:01:37] <anonimasu> lunch
[10:11:06] <A-L-P-H-A> anonimasu, 2 micro meter?
[10:15:19] <A-L-P-H-A> Is that you Anders?
[10:31:28] <alex_joni> A-L-P-H-A: awallin
[10:31:36] <alex_joni> is Anders wallin :)
[10:32:16] <A-L-P-H-A> Oh.
[10:32:23] <A-L-P-H-A> What's anonimasu's name then?
[10:32:34] <A-L-P-H-A> Smokey the bear?
[10:38:33] <alex_joni> anders somethingelse
[10:38:50] <alex_joni> it's an anders coalition
[10:38:58] <A-L-P-H-A> kinda freaky
[10:42:44] <lerneaen_hydra> 'lo all
[10:43:03] <A-L-P-H-A> hi
[10:44:42] <alex_joni> hi lh
[11:14:49] <anonimasu> hm
[11:14:57] <anonimasu> my slides are being delivered today
[11:15:11] <anonimasu> so much for the 3 weeks of delviery ;)
[11:16:24] <lerneaen_hydra> anonimasu: don't complain :p
[11:16:44] <anonimasu> hyaha
[11:16:57] <anonimasu> I called them to ask about when they could deliver..
[11:17:06] <anonimasu> "oh, they are on the truck already should arrive today"
[11:17:05] <anonimasu> :D
[11:17:21] <anonimasu> hm
[11:21:28] <alex_joni> * alex_joni cusses xp
[11:21:49] <alex_joni> stupid people who wrote that
[11:22:14] <alex_joni> it seems XP decides to mark an drive as bad (and only able to use PIO mode) if it fails 6 times to read in time
[11:22:35] <alex_joni> so if you stick a bad/scratched CD in your drive/burner, after 6 bad attempt it won't ever run as DMA
[11:23:06] <anonimasu> define ever..
[11:23:07] <alex_joni> ever
[11:23:18] <alex_joni> until you get pissed at it, and edit the registry
[11:23:20] <alex_joni> fixing it
[11:23:23] <anonimasu> heh
[11:24:00] <alex_joni> I was wondering why my burner had problems writing dvd's .. now I know
[11:24:09] <alex_joni> http://users.bigpond.net.au/ninjaduck/itserviceduck/udma_fix/
[11:26:20] <anonimasu> I made toolpaths for a limit/home switch yesterday
[11:26:32] <anonimasu> :)
[11:26:46] <anonimasu> I just need to hook up my index of the encoder and the sensor :)
[11:32:31] <alex_joni> this is for UxC ?
[11:33:11] <anonimasu> yeah..
[11:33:36] <alex_joni> nice
[11:33:41] <alex_joni> got them pictures yet?
[11:33:41] <alex_joni> :D
[11:33:44] <anonimasu> haha
[11:33:46] <anonimasu> no
[11:33:54] <alex_joni> get busy then :D
[11:34:02] <anonimasu> my vibration and stuff stopped when I aligned the spindle..
[11:34:07] <anonimasu> trammed it.. :9
[11:34:08] <anonimasu> :)
[11:34:18] <alex_joni> trammed?
[11:34:21] <anonimasu> \ /
[11:34:38] <anonimasu> I cant tighten down the machine as much as I'd like due to small servos :)
[11:34:47] <anonimasu> so I end up with vibration on deep cuts..
[11:34:50] <A-L-P-H-A> anonimasu buy big or go home.
[11:35:05] <anonimasu> thugh it may be the servos oscillating..
[11:35:07] <anonimasu> I'm not sure
[11:35:32] <anonimasu> it's not like it matters.. the cut
[11:35:42] <anonimasu> cut's ends up with a mirror finish :)
[11:36:54] <alex_joni> this is cutting lexane.. right?
[11:39:53] <anonimasu> haha
[11:39:54] <anonimasu> aluminium
[11:40:06] <alex_joni> styrofoam?
[11:40:10] <Dallur> balsa wood
[11:40:24] <alex_joni> butter
[11:40:38] <Dallur> jello
[11:40:47] <anonimasu> *sigh*
[11:40:48] <alex_joni> cutting jello is hard
[11:40:53] <alex_joni> it rips into pieces
[11:41:08] <Dallur> hmm might be better off to freeze it first
[11:41:19] <anonimasu> yeah
[11:41:22] <alex_joni> then it's too hard ;D
[11:41:41] <alex_joni> and I'm not sure it stays jello after melting it up
[11:42:14] <Dallur> alex_joni: hmm I know some of it does (my mom used to freeze some)
[11:42:38] <alex_joni> Dallur: I'll take your word for it :D
[11:42:43] <alex_joni> never tried it myself
[11:43:15] <Dallur> next emc meet I suggest everyone brings one item of food they have cut/made/sculptured with a cnc machine
[11:43:30] <Dallur> just keep the lubricant off
[11:43:41] <alex_joni> * alex_joni brings 1-year old parma ham
[11:44:01] <Dallur> milled into the shape of ?
[11:44:07] <alex_joni> parma cheese of course
[11:44:17] <Dallur> lol
[11:44:40] <Dallur> anyone ever tried milling ice ?
[11:45:02] <anonimasu> hm, should work nicely
[11:45:18] <anonimasu> the hard thing would be to keep it cool enough the whole process..
[11:45:32] <Dallur> liquid nitrogen nozzles ?
[11:46:01] <anonimasu> yeah :D
[11:46:25] <Dallur> that's what I call cooling :D
[11:46:39] <anonimasu> Dallur: do you understand sheetcam?
[11:46:47] <anonimasu> know..
[11:46:47] <anonimasu> :D
[11:46:50] <Dallur> anonimasu: I tried using it a bit
[11:47:02] <Dallur> anonimasu: I was not imipressed :(
[11:47:03] <anonimasu> what do you use for plasma?
[11:47:15] <Dallur> anonimasu: i'm making my own
[11:47:15] <anonimasu> I want something that can nest stuff esaily
[11:47:27] <anonimasu> shouldnt be hard :)
[11:49:01] <anonimasu> ^_^
[11:49:18] <anonimasu> sheetcam dosent impress me either.
[11:49:39] <Dallur> anonimasu: feel free to try it, it's a work in progress -> http://dallur.com/fileadmin/user_upload/dxf2nc_beta4.zip
[11:50:08] <Dallur> anonimasu: requires vpython and python 2.4/2.5
[11:50:17] <anonimasu> ah I have it somewhere..
[11:50:29] <Dallur> anonimasu: been fixing it the last couple of days
[11:50:29] <anonimasu> Dallur: why not write it as a plugin for axis?
[11:50:45] <anonimasu> so you can load a dxf directly?
[11:50:51] <Dallur> anonimasu: it might end up going into axis
[11:51:04] <anonimasu> Dallur: with some code for nesting and stuff it'd make a great addition to it :)
[11:51:24] <anonimasu> *rant*
[11:51:36] <anonimasu> awallin: hey, I saw your post on the freesteel blog :)
[11:51:55] <Dallur> anonimasu: im going to add route optimization, nesting and automatic sorting and plate placement
[11:52:11] <alex_joni> automatic nesting sounds great :D
[11:52:27] <anonimasu> * anonimasu nods
[11:52:31] <anonimasu> that's all I need/want
[11:52:35] <anonimasu> import 20 dxf parts..
[11:52:43] <anonimasu> and say NEST!
[11:52:51] <anonimasu> and it stuffs them togther as well as it can
[11:52:52] <anonimasu> ;)
[11:53:30] <awallin> anonimasu: yep, figured those guys know what they are talking about, so why not ask them about CAM stuff
[11:53:57] <awallin> the nesting problem is np isn't it?
[11:54:02] <Dallur> yup
[11:54:24] <Dallur> but there are ways to simplify by making assumptions
[11:54:58] <anonimasu> awallin: I'm getting more interested in this cam stuff, but my math's really shitty :/
[11:55:27] <anonimasu> awallin: yeah, i dont think it's a problem..
[11:56:04] <anonimasu> awallin: go through the list and compare what the maximum area you can get out of a part is..
[11:56:20] <anonimasu> awallin: you could make bounding boxes around the parts and stuff them togther, that'd work..
[11:56:24] <awallin> anonimasu: as long as you are good at something else required by CAM !
[11:56:25] <anonimasu> crude, but it's a quick one ;)
[11:57:50] <alex_joni> * alex_joni yawns
[12:07:11] <anonimasu> awallin: I have no idea what that would be;9
[12:07:17] <anonimasu> ;)
[12:10:10] <alex_joni> CAD?
[12:10:24] <anonimasu> that might be one thing :)
[12:15:02] <awallin> if someone would setup a barebones vtk view where an algorithm can just say point(X,Y,Z) or line(...) etc. and have it display in the view and be able to rotate zoom etc, then that would be a start. from there it's more about the math and CAM algorithms.
[12:16:02] <awallin> then there is the question of programming language :P python is quick to prototype, but C or C++ might be required for reasonable speed
[12:16:09] <anonimasu> awallin: also, I do know some opengl ;)
[12:16:44] <anonimasu> well I can manage with python and I know c/c++
[12:17:20] <awallin> I'm just afraid that with C or C++ a lot of work will go into creating basic data structures
[12:17:26] <anonimasu> yep
[12:17:29] <awallin> lists, trees etc
[12:17:40] <anonimasu> though if you use python, for the processing
[12:17:48] <anonimasu> and use c++ for the algorithms..
[12:17:51] <anonimasu> the computing heavy stuff
[12:17:53] <Dallur> There are plenty of CAD programs out there which can be used as as base though, pycad for example
[12:18:19] <anonimasu> that would require autocad right?
[12:18:46] <Dallur> pycad does not
[12:19:02] <anonimasu> ah "pyacad"
[12:19:15] <Dallur> http://www.pythoncad.org/
[12:20:59] <anonimasu> yeah
[12:21:30] <anonimasu> looks sane
[12:21:42] <anonimasu> that
[12:21:56] <anonimasu> I know im going to run that on the plasma
[12:22:27] <anonimasu> seems a neat way to draw small stuff
[12:23:25] <Dallur> merge axis+pythoncad+CAM functionality and you have something very scary
[12:23:40] <anonimasu> hm
[12:23:57] <anonimasu> pythoncad is 2d only right?
[12:25:18] <Dallur> as far as I know
[12:25:29] <anonimasu> :/
[12:25:57] <Dallur> which is a bit silly since adding a third dimension is not that much extra work
[12:28:32] <anonimasu> awallin: may be I should have a look at vtk
[12:30:48] <anonimasu> i'll be back in a few hours
[14:36:09] <skunkworks> http://www.cnczone.com/forums/showthread.php?t=30568
[14:37:30] <awallin> dear acondit, please rtfm
[14:37:47] <awallin> f is a word of your own choice...
[14:38:29] <anonimasu> :d
[14:44:15] <awallin> skunkworks: now I pointed him to the manual
[14:45:30] <alex_joni> skunkworks: good :)
[14:46:40] <skunkworks> Thanks awallin: I got busy
[14:53:55] <alex_joni> awallin: bet he won't figure out how to connect the limit signal to multiple pins
[14:54:11] <anonimasu> lol
[14:54:53] <maddash> does jack bauer freak out when I use it in concert with a userspace program that accesses pins 18-25?
[14:55:05] <alex_joni> jack bauer?
[14:55:28] <maddash> oh god. s\jack bauer\emc's hal
[14:55:38] <alex_joni> that's a common typo
[14:56:12] <alex_joni> maddash: if you have an userspace app that exports some pins, which you connect to pins of the parport, then it's ok
[14:56:14] <Dallur> WHUT
[14:56:17] <maddash> especially after 2 hrs of "24" from the previous night
[14:56:21] <alex_joni> doesn't matter if it's userspace or realtime
[14:56:38] <alex_joni> but 18-25 are just GND pins.. so those will only work if you want to write 0 to them :D
[14:57:49] <alex_joni> heh.. parport pins 18-25 are half pins.. you can only write 0 to them
[14:58:04] <alex_joni> that's a nice quote :D
[15:00:01] <maddash> erm, what about pins within 2-17 that aren't used by emc (ie, pins that are not used for [X,Y,Z]step, [X,Y,Z]dir, etc)?
[15:00:18] <alex_joni> maddash: you can connect to any unused pins
[15:00:38] <alex_joni> but it has to be a HAL application, you can't code a simple app which writes directly to 0x378
[15:00:41] <alex_joni> that won't work
[15:00:51] <maddash> define " won't work "
[15:01:12] <alex_joni> you write to the parport from userspace (maybe once / second or a bit faster)
[15:01:25] <alex_joni> but the RT code will run a lot faster and simply overwrite your values
[15:01:26] <maddash> I'm using the parportpin library...writes/reads from parport fine, but I've only tested my program when emc is inactive
[15:01:36] <alex_joni> maddash: that won't work
[15:01:54] <alex_joni> when emc is active you'll get only the values the emc/HAL writes to parport
[15:02:20] <alex_joni> what you need to do is add some hal pins to your application (not that complicated), and link those to the needed parport pisn
[15:02:24] <maddash> but emc's hal only writes to certain pin - not all of them, right?
[15:02:24] <alex_joni> pins even
[15:02:43] <alex_joni> emc's hal writes to the pins you tell it to
[15:02:51] <alex_joni> but the parport driver updates all pins at once
[15:02:56] <maddash> ah.
[15:02:57] <alex_joni> based on it's internal reference
[15:03:09] <maddash> "parport driver" = hal_parport.c?
[15:06:28] <maddash> did I disconnect spontaneously?
[15:08:27] <awallin> yes
[15:16:11] <mtedad> when homing where do I inject the number I want the axis to read. It homes at X -73.00 and that's what i like it to read.
[15:16:42] <anonimasu> that adds up as 0.. dosent it?
[15:16:48] <anonimasu> then you offset your fixture from ti..
[15:16:49] <anonimasu> it..
[15:17:55] <cradek> mtedad: g10L2P1x0y0z0, G92.1
[15:18:23] <cradek> with offsets cleared, you'll see the actual value
[15:18:34] <anonimasu> hm..
[15:18:41] <anonimasu> does anyone have a clue what the inputs on a USC will handle=?
[15:18:47] <jepler> mtedad: do you mean that the position of the home switch is -73.00? Then you want HOME_OFFSET=-73.00 in your .ini
[15:19:04] <jepler> mtedad: you may also want HOME=-73.00 so that it doesn't move to 0 after homing, but stays at -73
[15:19:15] <cradek> hmm, maybe I don't understand the question
[15:19:15] <mtedad> position of home sw is -73.
[15:19:33] <jepler> http://linuxcnc.org/docs/2.1/html/config/ini_homing/index.html (all but the last two inifile items apply to 2.0 as well)
[15:20:37] <anonimasu> so, nobody has a clue about the usc?
[15:21:01] <jepler> anonimasu: unfortunately, no
[15:21:16] <anonimasu> the inputs are optoisolated but I cant remember the ratings.
[15:22:18] <jepler> this page doesn't seem to give an exact rating .. http://216.239.51.104/search?q=cache:JwEgtuUX944J:jelinux.pico-systems.com/univstep.html
[15:22:23] <mtedad> tried setting [TRAG ] home to -73.0 0 in ini
[15:22:47] <anonimasu> mtedad: home_offset?
[15:23:17] <jepler> I believe you should set [AXIS_0] HOME and HOME_OFFSET
[15:23:36] <mtedad> don't want to leave it up to the operator to set offset.
[15:23:48] <anonimasu> mtedad: well, set your offset up in the config file :9
[15:24:11] <jepler> HOME_OFFSET is not a G92 or G5x offset. It gives the axis position corresponding to the home switch location.
[15:27:33] <jepler> anonimasu: my guess is that you need to choose a proper resistor to give the right current to the optoisolator, depending on the input voltage
[15:29:02] <anonimasu> hm..
[15:29:15] <mtedad> whats the ranges of P I D in units.
[15:29:46] <alex_joni> P is usually up to a few thousands
[15:29:57] <alex_joni> even 10000 isn't that uncommon
[15:30:04] <alex_joni> I is usually up to 1
[15:30:07] <jepler> anonimasu: any markings on the optos?
[15:30:20] <anonimasu> jepler: im not close to the board right now
[15:30:26] <anonimasu> pico-systems.com is down otherwise I'd check
[15:31:22] <alex_joni> seems it's been down for a few days
[15:31:36] <anonimasu> disturbing :/
[15:33:32] <Guest999> hello?
[15:34:10] <jepler> Guest999: welcome
[15:34:27] <anonimasu> I need 5v output inductive sensors.
[15:34:47] <jepler> anonimasu: 10mA current with a 1.5V drop over the opto might be a good first guess. e.g., http://www.fairchildsemi.com/ds/4N/4N35.pdf
[15:35:16] <anonimasu> they do handle 5v atleast..
[15:35:35] <anonimasu> that much im sure of..
[15:35:50] <alex_joni> bbl
[15:35:56] <jepler> if in doubt choose a larger resistor and work down
[15:36:04] <jepler> 380 ohm might be good for a 5V input signal
[15:36:28] <anonimasu> I wonder if I can use the plc if I home slowly.
[15:36:42] <anonimasu> though that wont be realtime.
[15:36:50] <anonimasu> but it'd work until jon gets he's site working
[15:36:51] <anonimasu> :)
[15:37:55] <maddash> where can I get a .tar.gz of 2.1a?
[15:38:15] <anonimasu> jepler: or maybe I should try mailing him
[15:38:16] <cradek> maddash: you can get a tar of the v2_1_branch on cvs.linuxcnc.org
[15:39:00] <maddash> cradek: yes, I'm browsing the cvs, but I can't find any links to a tar/tgz...
[15:39:15] <maddash> ach nvm
[15:39:21] <anonimasu> jepler: might be a easy way if in doubt
[15:39:21] <anonimasu> ;)
[15:39:23] <jepler> maddash: from http://cvs.linuxcnc.org/cvs/emc2/ choose Show only files with tag: v2_1_branch, click GO (if it doesn't GO automatically), then use the link Download this directory in tarball
[15:39:23] <cradek> :-)
[15:40:52] <Guest999> Can help me, how I can send serial port command to EMC ? (sorry by my poor english)
[15:42:33] <awallin> Guest999: I don't think there is an RS232 driver for HAL (yet). but please describe what you want to do
[15:43:28] <jepler> Guest999: awallin is right -- there's not an RS232 serial driver. however, you can write your own HAL drivers to do exactly what you want to do.
[15:43:35] <Guest999> i try to connect a PC (windows) and send commmand trought RS232 to the EMC
[15:44:01] <jepler> or software that sends NML messages, for that matter.
[15:44:30] <Guest999> ?
[15:44:49] <maddash> thanks jepler
[15:45:06] <jepler> The User Interface programs send "NML messages" to make emc do things -- for instance, to begin running the program
[15:45:26] <jepler> is that what you want to do?
[15:46:05] <Guest999> yes , I think ...
[15:48:11] <jepler> you can do this kind of thing by writing a program in C++ or Python that reads from the serial port and sends "NML messages".
[15:48:42] <jepler> I am not familiar with how to read from serial ports on Linux, so I can't give you specific advice.
[15:50:21] <Guest999> THANKS A LOTS,
[15:50:34] <jepler> here are two simple Python programs which send NML messages. The first one loads a g-code file, runs it, and then exits. The second accepts MDI commands typed by the user and executes them. http://axis.unpy.net/files/01167419757/jdi.py http://cvs.linuxcnc.org/cvs/emc2/src/emc/usr_intf/axis/scripts/mdi.py?rev=1.2;content-type=text%2Fplain
[15:51:39] <cradek> I have used this: http://pyserial.sourceforge.net/
[15:52:51] <Guest999> Thanks, i will to try..
[16:00:36] <paragon36> Hello All how was the weekend?
[16:00:48] <jepler> not long enough!
[16:01:17] <paragon36> :-)
[16:06:32] <paragon36> I just tried to install Gcam for testing but on the configure it gives the following error 'GTK+ 2.10 or higher is required' I guess I have the wrong version I'm using the install from the live EMC2 CD.
[16:08:45] <paragon36> apt-get install libgtk1.2-dev ??
[16:10:26] <maddash_> yay.
[16:10:31] <maddash_> maddash_ is now known as maddash
[16:10:52] <maddash> would any of you, by any chance, be expert spot welders?
[16:11:03] <paragon36> No did not work 'checking for GTK+ - version >= 2.10.0... no'
[16:11:25] <awallin> maddash: nope, just seen it done for NiMH cell batteries
[16:11:42] <cradek> paragon36: looks like libgtk2 is version 2.8.20 on dapper
[16:13:06] <maddash> aesh. I'm having a helluva time getting my steel to stick.
[16:13:22] <lerneaen_hydra> join #kaffeine
[16:13:27] <lerneaen_hydra> oops
[16:16:28] <paragon36> Hello Cradek o it should find it ok as it is above 2.1 right?
[16:17:05] <paragon36> Here the full error :-
[16:17:09] <cradek> I think 2.8 < 2.10
[16:18:09] <paragon36> Oooop 10 (TEN) .... lol I been working to hard I was seeing a one
[16:20:28] <paragon36> Mmmm can't find a package that high with apt-get is there a way too search with apt?
[16:21:03] <jepler> I doubt there is a gtk+ 2.10 package for ubuntu 6.10
[16:21:04] <jepler> er, 6.06
[16:22:47] <paragon36> Oh ok
[16:36:12] <jepler> paragon36: I can't test whether it works right now, but with this patch gcam compiles on my dapper box: http://emergent.unpy.net/index.cgi-files/sandbox/gcam-2007.01.09+dapper.patch
[16:37:05] <maddash> paragon36: I successfully compiled a lesser version of gcam (2006-something) yesterday...not sure how it differs from the latest version, though
[16:37:17] <paragon36> Thanks Jepler ... I was just manualy installing the latest version of GTK2 with all it deps
[16:37:32] <paragon36> Thanks for the info Maddash
[16:39:40] <CIA-8> 03jepler 07TRUNK * 10emc2/docs/src/config/ini_config.lyx: clarify [AXIS_n]HOME_OFFSET and [TRAJ]HOME
[16:39:42] <paragon36> Just getting the latest glib and pango dep's
[16:40:14] <CIA-8> 03jepler 07v2_1_branch * 10emc2/docs/src/config/ini_config.lyx: merge rev 1.14: clarify [AXIS_n]HOME_FFSET and [TRAJ]HOME
[16:44:09] <maddash> alex_joni, or anyone who knows how to do this: how do "add hal pins" to my application?
[16:45:12] <jepler> maddash: You use the hal functions such as hal_init() and hal_pin_bit_new()
[16:46:13] <jepler> and link with libemchal.so
[16:46:15] <maddash> jepler: so I'd have to link to hal.o?
[16:47:36] <maddash> jepler: can I do this while sendProgramOpen, sendProgramRun, etc?
[16:47:50] <jepler> I'm not sure what sendProgramOpen is
[16:47:59] <awallin> anyone installed enthought python on windows? I get an error when starting IDLE: tk84.dll was not found
[16:48:02] <jepler> the same program can use NML and HAL
[16:50:13] <maddash> jepler: so the fact that emcsh.cc (or any module from usr_intf) can send signals through the parport implies that hal_pin_bit_new is called with each launch of emc, right? what module makes this call?
[16:51:11] <jepler> maddash: the component which creates the pin calls hal_pin_bit_new during its startup
[16:51:33] <alex_joni> emcsh.cc is not hal related
[16:53:08] <maddash> alex_joni: yes, but it could be made so, according to jepler
[16:54:46] <alex_joni> sorry.. maybe I missed something
[16:55:03] <jepler> http://pastebin.ca/318591
[16:55:16] <maddash> is there any doc on the hal_*() calls? couldn't find "pin_bit_new" in either the wiki or hal_documentation.pdf
[16:55:46] <jepler> maddash: yes, in 2.1 there are Unix-style manpages for all the HAL and RTAPI functions. HTML versions linked from here: http://linuxcnc.org/docs/2.1/html/
[16:56:34] <alex_joni> maddash: I happen to know a bit about spot welding :)
[16:57:57] <alex_joni> jepler: any reason to make data a pointer?
[17:00:06] <jepler> alex_joni: yes, that memory has to be in HAL's shared memory. If it's not, you get an error like this at hal_pin_bit_new(): HAL: ERROR: data_ptr_addr not in shared memory
[17:00:42] <alex_joni> oh, right.. silly me ..
[17:01:59] <maddash> alex_joni: oooh, yay. does that bit include tempered steel sheets ~0.30 mm thick?
[17:02:42] <alex_joni> what spot welder do you have?
[17:02:46] <alex_joni> kVA rating
[17:05:19] <alex_joni> maddash: 0.30 mm is fairly thin, so 2-3 cycles at 50kA should be more than plenty
[17:08:10] <maddash> alex_joni: AC, 2.76 kVA
[17:09:37] <maddash> alex_joni: the welds are quite poor in quality. either I get a hole in the weld, or the resulting nugget can be broken apart easily with bare hands
[17:09:57] <alex_joni> hole means too much current
[17:10:07] <lerman> awallin: I'm running python on windows.
[17:10:12] <alex_joni> nugget broken is too little.. or too little pressing force
[17:11:23] <awallin> lerman: ok, I used enthought python once before, seems that package comes with a lot of useful modules
[17:11:43] <awallin> but IDLE is broken... I've spammed their mailing list with a quesiton, let's see what happens
[17:11:45] <maddash> alex_joni: could coating be a problem? the steel is blue tempered, but when I sand it down a bit, it becomes chrome-like
[17:11:51] <lerman> Not using enthought, though. Whatever that is.
[17:12:07] <lerman> I am using IDLE, though.
[17:12:25] <alex_joni> maddash: it can be
[17:12:25] <SWPadnos> activestate has a Windows Python
[17:12:33] <Dallur> At home I am using IDLE with Python 2.5 without any problems
[17:12:40] <alex_joni> maddash: if it's galvanized, chromed then it's a pita to spotweld
[17:12:46] <awallin> lerman: enthought is just a package with python and a lot of useful modules
[17:12:48] <alex_joni> you can try lower power, but more time
[17:13:03] <alex_joni> is this a manual welder? one where you do the pressing by hand?
[17:13:07] <awallin> code.enthought.com
[17:14:45] <lerneaen_hydra> awallin: just out of curiosity, how many hits per day do you get at anderswallin.net?
[17:14:57] <alex_joni> 2 nasty ones :)
[17:15:01] <alex_joni> and the rest normal ones
[17:17:47] <awallin> lerneaen_hydra: no idea, I guess I could check from the ISPs logs. I had some counters installed in wordpress, I think they counted unique IP addresses and those went up to about 7-800 if I posted something interesting
[17:18:13] <lerneaen_hydra> oh, ok
[17:18:23] <lerneaen_hydra> not bad, how much bandwidth is that eating?
[17:18:32] <lerneaen_hydra> lots or just a little?
[17:19:01] <awallin> no idea of that either, it sits on a service providers server and has nothing to do with my broadband at home
[17:19:27] <paragon36> Is there an apt-get package for opengl that you know of?
[17:19:51] <maddash> alex_joni: manual welder.
[17:19:59] <lerneaen_hydra> awallin: oh, ok
[17:20:12] <alex_joni> maddash: usually you can adjust the force on those (there's a screw on the top rocker arm)
[17:20:33] <awallin> lerneaen_hydra: if I could just remember how to access the site control panel
[17:20:52] <alex_joni> anderswallin.net/cpanel/
[17:21:04] <maddash> alex_joni: is pressing too hard a possibility?
[17:21:12] <alex_joni> maddash: unlikely
[17:21:38] <alex_joni> although it can be a problem if you have too much current
[17:21:40] <awallin> alex_joni: thanks, now can you remember my username and password for me also ;)
[17:21:50] <alex_joni> awallin: that was just a lucky guess
[17:21:52] <SWPadnos> awallin *****
[17:22:03] <lerneaen_hydra> anyone know if you can set up something like <a href=veryveryverylongurl>shorter text</a href>
[17:22:08] <alex_joni> yeah, it only shows as stars to us
[17:22:11] <lerneaen_hydra> </broken html code>
[17:22:25] <alex_joni> lerneaen_hydra: what's the verylongurl?
[17:22:31] <lerneaen_hydra> http://lerneaenhydra.shacknet.nu/index.php?option=com_content&task=view&id=15&Itemid=28
[17:22:34] <alex_joni> a http:// link?
[17:22:42] <lerneaen_hydra> yeah
[17:22:44] <SWPadnos> lerneaen_hydra, you would get rid of the "href" in the </a> tag
[17:22:50] <lerneaen_hydra> awallin: http://www.bash.org/?244321
[17:23:02] <lerneaen_hydra> SWPadnos: can you use html in the wiki?
[17:23:07] <SWPadnos> so it looks like <a href="http://lerneaenhydra.shacknet.nu/index.php?option=com_content&task=view&id=15&Itemid=28">My Link</a>
[17:23:17] <SWPadnos> dunno
[17:23:18] <alex_joni> that's no problem
[17:23:20] <SWPadnos> probably
[17:23:23] <alex_joni> lerneaen_hydra: yes, you can
[17:23:32] <lerneaen_hydra> oh, nice
[17:23:35] <alex_joni> depends on the wiki though
[17:23:41] <alex_joni> but most have <html> tags enabled
[17:23:44] <SWPadnos> I just copied it - I have no idea how it works ;)
[17:24:35] <lerneaen_hydra> hmm, no go
[17:24:41] <lerneaen_hydra> alex_joni: this is the linuxcnc wiki :p
[17:25:47] <jepler> paragon36: everything to run opengl apps should be installed by default on ubuntu. to compile gcam I had to add libgtkglext1-dev or something like that
[17:25:58] <alex_joni> lerneaen_hydra: then yes, some html is enabled
[17:25:59] <SWPadnos> lerneaen_hydra, http://www.usemod.com/cgi-bin/wiki.pl?TextFormattingRules
[17:26:14] <maddash> jepler: in http://pastebin.ca/318591 , why's it necessary to -DULAPI?
[17:26:55] <alex_joni> maddash: because hal relies on rtapi, which comes in 2 flavours
[17:27:04] <jepler> maddash: ULAPI means "UserLand API", as opposed to RTAPI (RealTime API), which is used only in kernel modules
[17:27:11] <alex_joni> -DRTAPI for realtime stuff, -DULAPI for userspace
[17:27:21] <SWPadnos> and rtapi.h requires that exactly one of those be defined
[17:27:25] <lerneaen_hydra> SWPadnos: oh, it's apparently wiki-specific code
[17:27:44] <maddash> so if compiled a kernel mod, I'd -DRTAPI?
[17:27:57] <SWPadnos> yes, so they do it with brackets [http://lerneaenhydra.shacknet.nu/index.php?option=com_content&task=view&id=15&Itemid=28 short link]
[17:28:10] <jepler> yes, for a "realtime" HAL module, you use -DRTAPI
[17:28:16] <lerneaen_hydra> * lerneaen_hydra smacks himself over the head
[17:28:25] <lerneaen_hydra> someone else had done exactly the same thing on the same page
[17:28:32] <SWPadnos> heh
[17:28:39] <lerneaen_hydra> somebody kill me now
[17:31:10] <alex_joni> * alex_joni kills lerneaen_hydra
[17:31:16] <lerneaen_hydra> * lerneaen_hydra dies
[17:31:23] <Dallur> * Dallur cries
[17:31:25] <lerneaen_hydra> * lerneaen_hydra grows back the lost organs
[17:31:34] <lerneaen_hydra> * lerneaen_hydra regenerates some more
[17:31:40] <lerneaen_hydra> iab
[17:31:57] <awallin> lerneaen_hydra: now I dug out the password for the control panel. For December it tells me 4044 unique visitors and 5.23Gb Bandwidth
[17:32:07] <lerneaen_hydra> 5gb?
[17:32:18] <lerneaen_hydra> you hosting videos?
[17:32:31] <awallin> nope, just pictures and pdfs
[17:32:39] <jepler> 5gb over how much time?
[17:33:05] <awallin> jepler: one month
[17:33:22] <jepler> emergent.unpy.net is about 1GB/month, 9000 visits
[17:33:27] <lerneaen_hydra> how large pics?
[17:34:02] <jepler> I don't post too often, and I rarely have large pics
[17:34:14] <awallin> previously I had 1024 wide pics, but now they are 500 pixel wide. Some of the pdfs are scans, so they are 1-3 Mb
[17:34:53] <awallin> it calculates for me a value of 719kb per visit, which doesn't seem that much. if you just load the front page you could reach 700kb
[17:35:06] <jepler> axis.unpy.net is 1GB/month but only 5000 visits -- every month some smart-arse manages to find and download all the old daily snapshots of the axis source code
[17:35:12] <lerneaen_hydra> oh, that much on the front page?
[17:35:40] <alex_joni> jepler: lol
[17:35:50] <lerneaen_hydra> then again, lots of images
[17:36:08] <lerneaen_hydra> haha
[17:36:12] <maddash> jepler: oh no, you've discovered my secret http transactions!
[17:36:48] <alex_joni> ROFL
[17:37:18] <awallin> when my ISP advertizes 2000 Gb of transfer, is that for one month ? or per year ?
[17:37:22] <lerneaen_hydra> damn, you gotta love jon's URL
[17:37:31] <maddash> jepler: I'm have OCD *and* I'm a history buff
[17:37:51] <maddash> s\I'm\I
[17:39:24] <maddash> geez, did I get disconnected again?
[17:39:34] <jepler> no
[17:40:00] <maddash> oh, you're just simmering in your anger silently.
[17:40:36] <maddash> ugh. bad joke.
[17:40:56] <lerneaen_hydra> psst! maddash, offer jepler one of your cakes
[17:41:06] <lerneaen_hydra> bribery is always effective
[17:42:28] <jepler> as long as they're not crabcakes -- yech
[17:43:06] <maddash> lerneaen_hydra: stop mocking me!! I don't, have not, and never will produce cakes!
[17:43:46] <maddash> crabcakes reek of fish, anyway
[17:44:59] <paragon36> Just seen your reply ... Thanks Jepler
[17:46:27] <maddash> bah. we'll always love you, Mr. Hydra
[17:46:31] <alex_joni> awallin: I think per month
[17:47:12] <lerneaen_hydra> jepler: that stippler thing you've done looks really nice, is it acrylic that you've painted black and then drilled through slightly?
[17:47:41] <jepler> lerneaen_hydra: oh I hadn't thought of painting the acrylic -- it's got black cardstock behind it
[17:48:03] <lerneaen_hydra> oh, so the cardstock is black to begin with?
[17:48:17] <lerneaen_hydra> just paper or something more exotic?
[17:48:43] <tomp> were some ppl having probs with video codecs? just solved mine.. the ubuntu binary doesnt get the win32codecs needed by newer wmv's
[17:48:44] <jepler> the dots are in plexiglass
[17:49:06] <tomp> i got codecs from ftp://ftp.fu-berlin.de/unix/X11/multimedia/MPlayer/releases/codecs/ and put them in /usr/lib/win32/
[17:49:11] <maddash> tomp: w32codecs, you mean?
[17:49:15] <tomp> yes
[17:49:21] <maddash> tomp: aren't they available in the ubuntu repos?
[17:49:23] <tomp> no
[17:49:25] <alex_joni> tomp: google for dapper w32codecs
[17:49:29] <tomp> k
[17:49:30] <lerneaen_hydra> hmm, dots in plexiglass and black cardboard behind it? where does the light come from?
[17:49:35] <alex_joni> they are available in some certain repo
[17:49:41] <lerneaen_hydra> through the side of the plexiglass?
[17:49:47] <jepler> lerneaen_hydra: yes, from the edges of the plexi
[17:49:55] <lerneaen_hydra> ah, ok
[17:50:00] <lerneaen_hydra> how well does it work?
[17:50:32] <jepler> lerneaen_hydra: def
[17:50:33] <jepler> er
[17:50:53] <jepler> lerneaen_hydra: on this page, you can see the "LED sticks" I made: http://emergent.unpy.net/01109127822
[17:50:59] <lerneaen_hydra> well, is it easily visible? is there leakage light where there's not supposed to be light etc
[17:50:59] <jepler> third image
[17:51:30] <jepler> there's not much leakage, and it's very clearly visible except in the brightest direct sunlight
[17:51:38] <lerneaen_hydra> hey, nice
[17:52:05] <tomp> "how to install the win32 codecs the easy way " http://www.ubuntuforums.org/showthread.php?t=79449 (but mine already works :)
[17:52:08] <lerneaen_hydra> do you think that acrylic, painted black, with just enough of a hole to get through it, and LEDs behind would work well?
[17:52:28] <jepler> I think you have to light from the edges
[17:52:31] <alex_joni> lerneaen_hydra: if you have enough leds
[17:53:25] <jepler> when edge lighting, the light is internally reflected to a certain degree, until it hits one of the drilled cones. When it hits a cone, it's directed out and you see it
[17:53:42] <lerneaen_hydra> that's true
[17:54:03] <lerneaen_hydra> I still don't think you'd get much leakage with my alteration
[17:54:16] <SWPadnos> edge lighting is much more efficient in that you only need a line of LEDs instead of an area
[17:54:17] <jepler> but most of the light will be absorbed by the black paint
[17:54:49] <lerneaen_hydra> yeah, that's true
[17:55:00] <SWPadnos> hmmm - it would depend on the paint. internal reflection works because of the dielectric of air - the paint may not cause those reflections
[17:55:14] <lerneaen_hydra> hmm, what if you paint it black first and then something reflective
[17:55:14] <jepler> I think the power dissipation of all the LEDs is around .9W
[17:55:17] <jepler> ooh lunchtime
[17:55:32] <maddash> nearly forgot
[17:55:41] <lerneaen_hydra> jepler: have you got the stippler app somewhere?
[17:55:46] <eholmgren> you could use EL
[17:55:50] <lerneaen_hydra> or is it available somewhere else?
[17:55:58] <eholmgren> but that would be expensive probably
[17:56:15] <eholmgren> cat looks awesome though, good job!
[17:56:22] <SWPadnos> http://emergent.unpy.net/software/01133127497-stippler-source-release
[17:56:42] <lerneaen_hydra> ooh, shiny
[17:56:44] <eholmgren> my cat Phil got "fixed" the other day ...
[17:57:10] <eholmgren> now he's on teh drugs and missing his poor cat balls
[18:01:11] <maddash> jepler: why do you declare data as a struct? and not a simple pointer?
[18:01:42] <maddash> jepler: also, what is hal_bit_t? there's no doc on that type.
[18:01:50] <SWPadnos> it
[18:01:59] <SWPadnos> it's in hal.h, I think
[18:02:09] <SWPadnos> (or possibly hal_priv.h)
[18:02:35] <tomp> jepler: the stippler is very nice. the stippler2dac sounds interesting too, ever use galvos?
[18:02:44] <alex_joni> maddash: as jepler pointed out the structure needs to be in hal (that's why the hal_malloc)
[18:02:57] <alex_joni> to be in hal shared memory
[18:04:55] <alex_joni> tomp: it seems the plf project is down at the moment, that means no apt packages for w32codecs for ubuntu
[18:05:02] <alex_joni> or at least I can't find them
[18:05:04] <eholmgren> jepler: did you ever make any progress with the board for futurlec?
[18:06:17] <tomp> ok, temporary ( just when I look :)
[18:06:55] <tomp> jepler at lunch maybe... but me too i got 2 futurelec 8255's
[18:09:28] <alex_joni> tomp: which codecs did you get from that ftp?
[18:09:52] <tomp> alex_joni: windows_all (when you dont know which, grab em all :)
[18:10:43] <tomp> alex_joni: also, there was no /usr/lib/win32 on my sys, so all wouldnt clash with anything
[18:10:58] <alex_joni> there is also an all-20061022
[18:11:16] <tomp> yes, didnt try all the combos, just the one
[18:11:39] <alex_joni> ok
[18:11:56] <tomp> oh, thats the one windows-all-20061022.zip
[18:12:16] <tomp> now i can watch the fab@home movies
[18:12:45] <maddash> alex_joni: yeah, but hal_malloc returns void* -- pointer, not struct*...
[18:13:09] <alex_joni> void * matches struct*
[18:14:42] <alex_joni> you cannot expect one function for every struct* you pass to it
[18:17:10] <maddash> well, from reading hal.h as SWPadnos suggested, hal_bit_t is just a char, so the struct data is essentially a pointer to a pointer. so why not use hal_bit_t **?
[18:17:42] <SWPadnos> what structs are you looking at?
[18:18:19] <alex_joni> maddash: I see no point in arguing
[18:18:37] <alex_joni> if you're not happy with the solution jepler gave you simply use another
[18:19:34] <maddash> alex_joni: no, no, not arguing, I'm just confused...I've compiled it without the struct, but I'm hesitant to run it for fear of perma-kernel-panic damage
[18:19:44] <alex_joni> there is no such thing
[18:19:56] <alex_joni> nothing you can do as an user can damage linux
[18:20:00] <maddash> alex_joni: I'm running in sudo.
[18:20:07] <alex_joni> unless you are reckless enough to run wiht sudo
[18:20:12] <alex_joni> and why would you do that?
[18:20:33] <maddash> alex_joni: otherwise, I can't access the parport
[18:20:46] <SWPLinux> if you want to make a component with more than one pin, you'd put more than one hal_<x>_t * in the struct
[18:20:48] <alex_joni> you won't have to access the parport
[18:20:49] <maddash> SWPadnos: http://pastebin.ca/318591
[18:21:00] <SWPLinux> yep - found it
[18:21:01] <alex_joni> maddash: you'll have to change a hal pin
[18:21:09] <alex_joni> which you can do as an user
[18:21:14] <maddash> ah ok
[18:21:23] <alex_joni> then using HAL you link that pin to the parport.0.pin-xx-out
[18:21:37] <alex_joni> and the driver (which runs inside the kernel) will access the parport
[18:23:29] <maddash> alex_joni: so I'll have to modify one of the *.hal files in order to link whatever pins I want to the parport driver??
[18:23:37] <maddash> alex_joni: s\??\?
[18:23:38] <SWPLinux> maddash: yes
[18:23:51] <alex_joni> yes
[18:24:07] <alex_joni> you need to modify the file to load your component
[18:24:13] <alex_joni> then link your pins to parport
[18:37:30] <awallin> is motion.enable in EMC2 the same thing as iocontrol.0.emc-enable.in was in some old version?
[18:37:54] <awallin> http://www.cnczone.com/forums/showthread.php?p=243302#post243302
[18:38:16] <alex_joni> no
[18:38:28] <alex_joni> awallin: motion.enable has always been there
[18:39:33] <awallin> the forum user asks a valid question: if e-stop is time critical, why is it in iocontrol (userspace?)
[18:39:34] <alex_joni> emc-enable.in is estop related
[18:39:45] <alex_joni> estop is not time critical
[18:39:53] <alex_joni> that's why you have one on your gui
[18:40:02] <alex_joni> at least emc's estop is not time critical
[18:40:14] <awallin> well, when I press e-stop I'd like my machine to stop pretty quickly
[18:40:19] <alex_joni> any sane design includes an external estop which is done using relays etc
[18:40:26] <alex_joni> which usually cuts the powers to the motors
[18:40:41] <alex_joni> when I press e-stop I DON'T want to rely on software
[18:41:04] <alex_joni> software is a separate issue
[18:41:12] <alex_joni> but motion should be long stopped
[18:42:35] <SWPLinux> Software: taking the "E" out of E-Stop for 30+ years :)
[18:43:14] <alex_joni> SWPLinux: not over here :)
[18:43:33] <alex_joni> 24V relays with brakes all over the place :)
[18:43:56] <SWPLinux> right. Hardware: Leaving the "E" in E-Stop, regardless of software ;)
[18:44:05] <alex_joni> if done properly
[18:45:11] <alex_joni> awallin: to answer your question: iocontrol.0.emc-enable.in is merely a pin to let emc know there's an external ESTOP which happened, and emc shouldn't be too mad if motors don't move :D
[18:45:23] <SWPLinux> now this is weird. I just got a pop-up window asking me for my POP server password
[18:45:39] <SWPLinux> which might be normal if I had an email app running
[18:45:43] <awallin> alex_joni: ok, I tried explaining that on cnczone
[18:45:54] <alex_joni> saw that
[18:46:16] <alex_joni> SWPLinux: oopsy
[18:46:35] <SWPLinux> I guess there are a few Evolution related things running in the background
[18:47:31] <SWPLinux> interesting. it's fallout from a couple of days ago when I clicked on the "contact us" link on xylotex.com
[18:47:38] <alex_joni> lol
[18:47:46] <SWPLinux> (and then closed the email window in some way)
[18:49:25] <SWPLinux> all hail evolution --force-shutdown :)
[18:49:27] <alex_joni> awallin: as a sidenote.. I thought about integrating all iocontrol into motion/RT
[18:50:33] <awallin> alex_joni: yeah, there are not that many pins left in iocontrol... and how many actually use stuff like lube? is there a G-code for lube? seems like old baggage from the nist code
[18:50:46] <alex_joni> M<mumble? is lube
[18:50:58] <alex_joni> same for flood
[18:51:21] <robin_sz> awallin, by any chance, do you build life-size models of cows out of match sticks?
[18:52:18] <awallin> robin_sz: nope...
[18:52:33] <robin_sz> or collect stamps?
[18:52:34] <alex_joni> but boats.. that's an idea :)
[18:54:20] <robin_sz> its just that I wondered if you had many pointless hobbies as well as trying to discuss technical stuff with CNCZone users ...
[18:54:26] <robin_sz> ;))
[18:55:29] <awallin> you think everyone on cnczone is stupid?
[18:55:36] <robin_sz> not everyone
[18:55:37] <awallin> there are a few OK threads over there...
[18:56:06] <robin_sz> they key pharse is "a few" ...
[18:56:55] <awallin> ah, the thread continues...
[18:57:58] <awallin> perhaps our docs are not complete?
[18:58:01] <robin_sz> one problem they have is some of the so called "experts" sell really crap gear and advertise on CNCzone, so they get to peddle their crazy ideas and crap gear with impunity
[19:01:25] <awallin> alex_joni: you probably know where the limit pins are??
[19:02:10] <awallin> seems they are not in the html docs
[19:03:04] <alex_joni> sure they are :)
[19:03:29] <awallin> show me
[19:04:21] <alex_joni> axis.N.neg-hard-limit
[19:04:22] <alex_joni> The negative hard limit for the joint
[19:04:22] <alex_joni> axis.N.neg-soft-limit
[19:04:22] <alex_joni> The negative soft limit for the joint
[19:04:22] <alex_joni> axis.N.pos-hard-limit
[19:04:23] <alex_joni> The positive hard limit for the joint
[19:04:26] <alex_joni> axis.N.pos-soft-limit
[19:04:29] <alex_joni> The positive soft limit for the joint
[19:04:36] <alex_joni> but I think they ended up as params for whatever reason :/
[19:04:40] <awallin> those are listed as Parameters...
[19:04:56] <alex_joni> oh, not those
[19:05:25] <alex_joni> axis.*.neg-lim-sw-in and
[19:05:30] <alex_joni> axis.*.pos-lim-sw-in
[19:05:58] <awallin> those don't seem to be in the docs??
[19:06:51] <alex_joni> awallin: seems like it :0
[19:06:59] <alex_joni> you found a bug .. yay :P
[19:07:02] <alex_joni> * alex_joni adds them
[19:07:15] <jepler> lerneaen_hydra: http://emergent.unpy.net/software/01133127497-stippler-source-release
[19:07:49] <lerneaen_hydra> hmm, seems like linpmg isn't something very common
[19:08:02] <lerneaen_hydra> wasn't in ubuntu's universe or multiverse rep
[19:08:03] <jepler> eholmgren, tomp: I have a driver for the futurlec 8255 on one of my home machines. It does digital outs only. An early, and I think completely non-working, version is in the emc2 CVS
[19:08:37] <lerneaen_hydra> btw, what type of fileformat is pmg?
[19:09:15] <maddash> lerneaen_hydra: dmg you mean?
[19:09:20] <jepler> lerneaen_hydra: probably a typo for 'libpng'
[19:09:27] <lerneaen_hydra> ...
[19:09:28] <jepler> not sure though
[19:09:30] <lerneaen_hydra> hahaha
[19:09:33] <lerneaen_hydra> right
[19:09:45] <lerneaen_hydra> so file-input is png too?
[19:10:37] <jepler> oh you mean pgm
[19:10:37] <tomp> jepler: thanks, someday later, I'd like to try it
[19:10:53] <lerneaen_hydra> hmm, maybe I do
[19:11:23] <jepler> lerneaen_hydra: the "convert" utility can turn greyscale images into .pgm, and gimp can save in that format
[19:11:31] <jepler> .pgm is a simple uncompressed format
[19:11:31] <lerneaen_hydra> oh
[19:11:33] <lerneaen_hydra> nice
[19:11:42] <lerneaen_hydra> bmp-similar?
[19:12:53] <CIA-8> 03alex_joni 07v2_1_branch * 10emc2/docs/src/config/emc2hal.lyx: add forgotten pins
[19:13:02] <jepler> looks like on ubuntu you want libnetpbm10-dev
[19:13:28] <jepler> on older machines you may have to change it to -lpnm not -lnetpbm
[19:15:48] <maddash> jepler: in http://pastebin.ca/318591 , how do you know which realworld pin data is linked to?
[19:16:18] <awallin> run it and look with halmeter?
[19:16:41] <awallin> test.test.bit ?
[19:17:12] <SWPLinux> no -just test.bit
[19:17:32] <jepler> maddash: you know what it's linked to because you have to write a 'link' line in your .hal files
[19:17:36] <alex_joni> maddash: test.bit, and it's not linked to anything
[19:17:45] <awallin> hm, then hal_pin_new works differently in C and Python
[19:17:49] <alex_joni> not unless you link it specifically first
[19:18:13] <jepler> awallin: yes, in "C" it doesn't automatically prepend the component name
[19:18:18] <SWPLinux> awallin: the python hal library may prepend the component name to all pins, but it's not required by HAL
[19:18:37] <awallin> SWP: it might as well be required, so as not to confuse the user...
[19:18:38] <SWPLinux> it is a source of confusion in some cases, like axis.* coming from motion
[19:18:55] <SWPLinux> it can be more confusing in complex cases
[19:19:53] <SWPLinux> I'd love to see the HAL naming become a hierarchical data struct, but that is probably not in the cards any time soon
[19:21:56] <jepler> someday when I have a lot of time, I'll write 'comp2' with arrays and variants
[19:23:00] <jepler> personality 1 { pin out bit dout[16]; } personality 2 { pin out bit dout[8]; pin in bit din[8]; } personality 3 { pin in bit din[16]; }
[19:23:49] <jepler> for a hardware driver that can be configured as inputs, outputs, or a mix
[19:25:22] <jepler> (this would work for a small number of variants, but be very cumbersome to support e.g., futurlec 8255 with 4096 distinct combiniation of inputs and outputs)
[19:30:53] <CIA-8> 03alex_joni 07TRUNK * 10emc2/docs/src/config/emc2hal.lyx: add forgotten pins
[19:31:06] <alex_joni> huh.. TRUNK again :)
[19:32:36] <alex_joni> so.. who wants to test some 2.1.0~alpha0 packages ?
[19:33:43] <SWPLinux> dunno about that, but I do have a question
[19:34:06] <SWPLinux> I think the halcmd change I did last night should probably be in 2.1 - what do you think?
[19:34:18] <alex_joni> what change was that?
[19:34:34] <alex_joni> I was so astonished there was a change from you I failed to read what it was :P
[19:34:42] <SWPLinux> fixed some potential buffer overflows and improved error checking/reporting in replace_vars
[19:34:46] <SWPLinux> heh :P
[19:34:55] <alex_joni> *kidding*
[19:35:04] <alex_joni> bugfixes can be backported :)
[19:35:38] <alex_joni> is it safe?
[19:35:48] <SWPLinux> hmmm. I didn't see a backport from jmk of his fixes either (some issues with loadusr -w / -W)
[19:36:01] <alex_joni> jmk plans to do a net backport
[19:36:14] <SWPLinux> I think it is. I tested with some strange cases, and correctly loaded pyvcp_demo (after fixing that)
[19:36:50] <alex_joni> can you convince someone else to review it ?
[19:37:19] <alex_joni> if you grab me a link I'll do it :)
[19:37:38] <SWPLinux> link to the diff?
[19:37:51] <alex_joni> yeah..
[19:37:56] <alex_joni> found one now
[19:38:00] <SWPLinux> ok
[19:38:58] <alex_joni> * alex_joni puts i18n on swp's todo list
[19:39:01] <alex_joni> :P
[19:39:18] <SWPLinux> add that to hal_lib, and I'm happy ;)
[19:39:51] <SWPLinux> cradek, jepler: can cvsweb be configured to show the commit message at the top of a diff?
[19:39:51] <alex_joni> http://cvs.linuxcnc.org/cvs/emc2/src/hal/utils/halcmd.c.diff?r1=1.107;r2=1.108;f=h
[19:39:56] <alex_joni> I'm looking at that
[19:39:58] <SWPLinux> (like here: http://cvs.linuxcnc.org/cvs/emc2/src/hal/utils/halcmd.c.diff?r1=1.106;r2=1.107;f=h )
[19:40:04] <alex_joni> shouldn't that test be the other way around?
[19:40:13] <SWPLinux> which one?
[19:40:15] <alex_joni> if ((retval < 0) && (retval >= -7)) {
[19:40:37] <SWPLinux> no, I want numbers in the range of -7 to -1
[19:40:39] <alex_joni> oh wait.. those are negative
[19:40:56] <alex_joni> didn't read it fully :)
[19:40:58] <SWPLinux> he
[19:41:00] <SWPLinux> h
[19:41:40] <SWPLinux> I guess I should fix the comment for strlimcpy - it actually returns an error if the dest_space is too little
[19:43:52] <alex_joni> the rest looks good to me
[19:44:13] <alex_joni> skunkworks: around?
[19:44:20] <SWPLinux> ok. I don't like all the constants sprinkled through the code, but it works
[19:49:40] <alex_joni> I would have used #define FOO_ERROR -1
[19:50:19] <SWPLinux> yep. that and the limits (128, 2048, 2*2048 ...)
[19:57:06] <jepler> alex_joni: huh, I wonder how I missed the limit switches...
[19:57:15] <maddash> where in the cvs is scripts/emc located? is it scripts/emc.in?
[19:57:32] <jepler> maddash: yes, 'configure' generates scripts/emc from scripts/emc.in
[20:00:00] <alex_joni> maddash: there are a number of files parsed by configure
[20:01:34] <maddash> but does the core of scripts/emc (ie, how it links *.hal to $(DISPLAY), which is what I want to know) already reside in emc.in?
[20:01:51] <alex_joni> it's all there
[20:02:14] <alex_joni> only certain varialbles like @THIS@ get replaced by configure
[20:13:52] <maddash> I've got the rtai kernel for debian; can it be added to the repos?
[20:14:07] <maddash> .deb packages, I mean
[20:14:32] <alex_joni> what kind of kernel?
[20:14:42] <alex_joni> and what kind of rtai?
[20:14:57] <maddash> the end result of this page: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Debian_Etch_Compile_RTAI
[20:15:23] <maddash> I figured that it'd faster and easier to install a .deb than to go through that page from top to bottom
[20:15:37] <alex_joni> I'm afraid I can't add anything to an etch repo
[20:15:48] <CIA-8> 03jepler 07TRUNK * 10emc2/docs/man/man3/rtapi_get_time.3rtapi: add note about how to compute time deltas with integer math
[20:15:48] <CIA-8> 03jepler 07TRUNK * 10emc2/docs/man/man3/ (5 files): document hal types better
[20:16:06] <alex_joni> cradek: you know what's up with HEAD vs. TRUNK ?
[20:17:22] <SWPadnos> heh
[20:17:41] <SWPadnos> HEAD is the most recent part of a particular branch, TRUNK is the main branch
[20:17:53] <jepler> but why did it start saying TRUNK instead of HEAD in the CIA messages?
[20:17:57] <SWPadnos> (from esoteric CVS discussions a few days ago)
[20:18:00] <jepler> I don't care if TRUNK is more correct :-P
[20:18:03] <SWPadnos> heh :)
[20:18:08] <maddash> who is this CIA-8?
[20:18:15] <cradek> I changed it
[20:18:15] <jepler> maddash: it's a bot that reports changes made to the CVS
[20:18:31] <maddash> alex_joni: by "repos," I meant http://linuxcnc.org/emc2/
[20:18:52] <lerneaen_hydra> well, the devs will say it's a bot, but in reality he's a chinese slave alex has in his basement and feeds rice occasionally
[20:19:01] <maddash> rofl
[20:19:14] <cradek> the wrong use of that word caused (big) confusion, so that's what I can do to train us to use the right word
[20:19:19] <lerneaen_hydra> poor little cia-8
[20:19:38] <lerneaen_hydra> the number after denotes which slave he's up to
[20:19:44] <lerneaen_hydra> so this would be his 8th one
[20:19:50] <CIA-8> 03jepler 07v2_1_branch * 10emc2/docs/man/man3/ (6 files): merge documentation improvements from HEAD
[20:19:52] <lerneaen_hydra> they don't survive very long :(
[20:19:53] <maddash> what's "CIA", then?
[20:20:08] <SWPadnos> http://www.navi.cx I think
[20:20:11] <alex_joni> * alex_joni will only commit to branches form now on
[20:20:13] <jepler> http://cia.navi.cx/ "CIA - The open source informant"
[20:20:17] <alex_joni> cia.navi.cx
[20:20:25] <lerneaen_hydra> chinese informant (of) alex
[20:20:50] <alex_joni> maddash: we only have repositories for ubuntu
[20:21:00] <alex_joni> any other platform is not supported by EMC
[20:21:17] <alex_joni> sure emc works on it, but ther won't be any support for etch for example
[20:21:17] <SWPadnos> at least, not with packages
[20:22:17] <CIA-8> 03jmkasunich 07TRUNK * 10emc2/docs/src/config/emc2hal.lyx: corrected type of axis.N.index-enable from OUT to IO. (note: motion.spindle-sync should also probably be IO, but thats not just a docs change)
[20:22:31] <cradek> whoah, is jmk still on vacation?
[20:23:52] <alex_joni> "I'd like to download the universe..." LOOOOL
[20:24:02] <alex_joni> http://www.ubuntuforums.org/showthread.php?t=20217
[20:24:25] <jepler> alex_joni: I don't think we should rule out putting kernel images for other distributions on linuxcnc -- there's room and bandwidth.
[20:24:43] <maddash> jepler: caveat emptor?
[20:24:46] <cradek> I think there's already some FC kernels there
[20:24:48] <jepler> alex_joni: but we would have to make it clear what system we "support" (ubuntu 5.10 and 6.06) vs "host"
[20:24:56] <jepler> cradek FTW
[20:25:10] <maddash> !define FTW
[20:25:11] <jepler> yeah there's an FC2 or FC3 kernel in there
[20:25:13] <jepler> "for the win"
[20:25:19] <lerneaen_hydra> for the win
[20:25:20] <alex_joni> jepler: I think there are some RPM's in there
[20:25:24] <lerneaen_hydra> oh so slow
[20:25:32] <alex_joni> right
[20:25:44] <alex_joni> I guess.. we can do that
[20:25:45] <cradek> are there some kernels there? like maybe for FC?
[20:25:50] <cradek> maybe they are in RPM format
[20:25:53] <alex_joni> :P
[20:26:02] <maddash> there are rpm rtai kernels?
[20:26:12] <jepler> yes, two or three years old by now
[20:26:14] <cradek> SWPadnos: I tried to add the log to the diff page, but failed
[20:26:26] <maddash> does that mean I could `alien` instead of going through that etch compile page?
[20:26:26] <SWPadnos> ok. thanks for trying
[20:26:40] <alex_joni> maddash: most likely those won't work on etch
[20:26:40] <cradek> SWPadnos: since the diff can be between any two versions, it's not clear what the log should be
[20:27:00] <SWPadnos> actually, I was thinking of the commit message from the latest version in the diff
[20:27:18] <cradek> actually, if they are on different branches, no log/logs is/are appropriate
[20:27:23] <SWPadnos> true
[20:27:33] <SWPadnos> then the commit message of the first one in the diff :)
[20:27:36] <jepler> maddash: http://www.linuxcnc.org/index.php?option=com_content&task=view&id=2&Itemid=4&lang=en "There are experimental kernel & RTAI packages available here for FC3. These have been built by Jonathan Stark, please let him know (on the developers mailing list ) if you encounter any problems" http://www.linuxcnc.org/www.linuxcnc.org/rpm/ [hm that URL looks broken]
[20:27:45] <maddash> SWPadnos == SWPLinux?
[20:27:49] <SWPadnos> yes
[20:28:02] <cradek> SWPLadnox
[20:28:03] <SWPadnos> two computers on ATM
[20:28:15] <SWPadnos> here in SWPLand, we do things a little differently
[20:28:18] <lerneaen_hydra> cradek: haha
[20:28:42] <SWPadnos> I could have used SWPBigOpteronMachine, but I didn't want to make people jealous :)
[20:29:28] <cradek> I lost the only pen I had that writes well
[20:29:42] <SWPadnos> I went to staples and bought more pens
[20:29:56] <cradek> now I have only the half dozen that the marketing department gave me, with my own company's name
[20:29:59] <SWPadnos> I don't like them much though - they're a little different from the last set I got
[20:30:00] <cradek> and they all suck
[20:30:04] <cradek> one blinks
[20:30:13] <maddash> where's SWPLand? where people end up when they look for SMP-land?
[20:30:15] <SWPadnos> ooooh -a dashed pen. you should patent that
[20:30:19] <cradek> and looks like a "marital aid"
[20:33:12] <skunkworks> :)
[20:33:12] <cradek> really - it does
[20:33:12] <jepler> he's not kidding
[20:33:12] <cradek> no.
[20:36:07] <lerneaen_hydra> err...
[20:36:23] <lerneaen_hydra> this is the last channel I expected to see stuff like that on
[20:36:37] <lerneaen_hydra> great :D
[20:37:47] <maddash> I'm so dumb. "marital aid"??
[20:37:58] <lerneaen_hydra> it took me a while too
[20:38:01] <jepler> maddash: sex toy
[20:38:58] <cradek> sorry, I thought everyone knew that code
[20:39:39] <maddash> ohhhh
[20:39:50] <cradek> * cradek snaps to his senses just in time to not search for "marital aid" on google from work
[20:40:24] <lerneaen_hydra> hahaha
[20:40:42] <lerneaen_hydra> that's what ssh is for ;)
[20:40:47] <alex_joni> jepler: thanks for pointing out the broken link... I fixed it
[20:42:01] <maddash> for someone who retains a stoic and deadpan facade, cradek sure knows his innuendos.
[20:43:24] <maddash> alex_joni: on the topic of broken links, does the link to the code that "needs massive tweaking" work, (http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Cam) ?
[20:44:22] <alex_joni> maddash: a wiki is a comunity based information Knowledge base
[20:44:30] <alex_joni> so I can't be bothered to look for broken links
[20:46:55] <maddash> alex_joni: so...should I go ahead and remove it myself?
[20:47:31] <jepler> before editing you have to follow the BasicSteps: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?BasicSteps
[20:47:45] <jepler> we are always pleased when someone improves the wiki
[21:00:32] <CIA-8> 03alex_joni 07v2_1_branch * 10emc2/debian/control.in: remove a duplicated line, replace/conflict/provide emc2-axis too
[21:07:48] <CIA-8> 03alex_joni 07TRUNK * 10emc2/debian/control.in: remove a duplicated line, replace/conflict/provide emc2-axis too
[21:07:57] <maddash> is there a reason why register arguments are disabled in the rtai kernel?
[21:08:30] <alex_joni> register arguments?
[21:09:09] <maddash> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Debian_Etch_Compile_RTAI during the make menuconfig step
[21:09:49] <maddash> if enabled, then system calls would utilize the registers instead of the stack, thus making those calls faster
[21:10:05] <jepler> maddash: or line in global_assignments.split("\n"):
[21:10:05] <jepler> a, b = line.strip().split(None, 1)
[21:10:05] <jepler>
[21:10:08] <jepler> argh
[21:10:14] <jepler> maddash: https://mail.rtai.org/pipermail/rtai/2005-November/013367.html
[21:10:33] <jepler> maddash: it appears that for at least some versions of rtai, this option must be turned off to get a working system
[21:12:01] <alex_joni> "Just a flag in kernel config, not compatible with RTAI own syscalls."
[21:12:09] <alex_joni> by Paolo
[21:15:44] <maddash> the http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Debian_Etch_Compile_RTAI page needs some editing...it currently assumes that you're doing everything under `su`, including the make-kpkg. this is an unsafe practice. should I rewrite it to use fakeroot?
[21:16:12] <alex_joni> if you're sure it works.. please do
[21:22:10] <maddash> ugh. compilation error.
[21:33:10] <daniel_br> hi guys
[21:36:09] <alex_joni> hi daniel_br
[21:36:58] <daniel_br> hello alex
[21:37:34] <daniel_br> i have a question about the pluto config
[21:37:48] <daniel_br> i thinking buy one
[21:38:32] <daniel_br> can some input be use for "read" a analog value?
[21:38:36] <alex_joni> jepler is the one who did all emc2-pluto stuff ;)
[21:39:00] <alex_joni> daniel_br: but I seriously doubt that
[21:39:07] <alex_joni> it's hard to emulate an ADC
[21:39:22] <jepler> the FPGA chip on the pluto has only digital I/O pins
[21:39:48] <jepler> the firmware has no privision for reading from an external ADC, either
[21:40:14] <daniel_br> for convert some analog input what is better?
[21:40:22] <daniel_br> some pic?
[21:40:48] <alex_joni> there are a few serial ADC's
[21:40:53] <jepler> many microcontrollers have ADCs
[21:41:10] <jepler> you could hook an ADC to the Pluto, but you'd have to write your own FPGA firmware & HAL driver to actually use it
[21:41:25] <alex_joni> is there still space in the FPGA?
[21:41:39] <jepler> alex_joni: not much, you might have to sacrifice something else
[21:41:52] <jepler> or optimize it better than I have
[21:41:58] <alex_joni> I doubt that :)
[21:44:04] <maddash> hmmm this is odd - using fakeroot with make-kkpkg returns an ipipe error...
[21:44:24] <daniel_br> the drivers for the mesa board have some internal adc?
[21:44:27] <jepler> alex_joni: I am far from an expert on FPGA
[21:45:13] <alex_joni> daniel_br: you mean the 5i20 ?
[21:45:16] <daniel_br> yes
[21:45:30] <alex_joni> I don't think so
[21:45:36] <jepler> I think it also would use an external ADC
[21:45:42] <alex_joni> but I'm still waiting for mine :)
[21:45:57] <jepler> bbl
[21:46:49] <daniel_br> but is possible add some adc inputs or not?
[21:47:04] <awallin> daniel_br: the 5i20 has only digital IO
[21:47:24] <awallin> you could have a seprate ADC, and hook that to the 5i20
[21:47:40] <awallin> if you don't need fast sample rates, you could write a driver in HAL that reads the ADC
[21:47:57] <awallin> if you need fast sample rates you need to read the ADC with the FPGA
[21:48:23] <awallin> if you can sacrifice many pins on the m5i20 then the ADC could give out it's value in parallell
[21:48:46] <awallin> if you need to conserve pins, you need a serial ADC (SPI, I2C, etc.)
[21:48:51] <daniel_br> i think this is one interesting feature for future drivers
[21:49:22] <daniel_br> for 5i20
[21:52:47] <daniel_br> http://zone.ni.com/reference/en-XX/help/371599B-01/lvfpgaconcepts/performing_basic_i_o/
[21:53:06] <daniel_br> If the FPGA target you use includes analog input resources, you can configure an FPGA I/O Node to read an analog input value.
[21:55:59] <awallin> daniel_br: yes, we use NI fpga cards (R-series) at work, they are very easy to program using labview, but they cost 2-4k, which is over hobby budget for most
[21:57:19] <daniel_br> in these fpga used in the pluto and 5i20 is this not possible?
[22:02:16] <alex_joni> daniel_br: I bet it's not
[22:05:08] <awallin> daniel_br: the NI cards have the ADC onboard, the 5i20 and pluto don't have any ADCs
[22:05:48] <alex_joni> awallin: you could build a sucky ADC out of a PWM-pseudo DAC and a comparator
[22:06:29] <awallin> alex_joni: yes, but to be of any use I am guessing it would have to be done on the fpga (for speed)
[22:07:10] <daniel_br> analog -> pwm ->hal component -> float?
[22:08:16] <alex_joni> daniel_br: no..
[22:08:33] <awallin> daniel_br: I think it's called successive approximation AD
[22:08:33] <alex_joni> pwm -> filter -> analog comparator -> bit feedback
[22:09:12] <awallin> here http://en.wikipedia.org/wiki/Successive_Approximation
[22:10:22] <awallin> daniel_br: what do you need the analog input for?
[22:11:43] <daniel_br> i thinking about some edm ..;
[22:12:14] <daniel_br> home build one
[22:17:16] <daniel_br> bbl
[22:23:10] <anonimasu> hello :)
[22:23:34] <alex_joni> hey anders
[22:23:46] <awallin> who, me? ;)
[22:23:55] <anonimasu> hey
[22:23:57] <anonimasu> what's up?
[22:24:05] <anonimasu> the usc had 5v to switches and stuff :/
[22:24:05] <anonimasu> damn.
[22:24:30] <anonimasu> means I need inductive sensors with universal output's
[22:24:34] <anonimasu> for my limits
[22:25:41] <awallin> why not microswitches?
[22:26:06] <anonimasu> because mounting them is a pain
[22:26:17] <alex_joni> those get confused by HF plasma around them
[22:26:46] <anonimasu> alex_joni: well, it's not for the plasma..
[22:27:00] <anonimasu> the plasma has limit switches in all axes already :)
[22:27:02] <anonimasu> microswitches..
[22:29:11] <CIA-8> 03alex_joni 07v2_1_branch * 10emc2/debian/changelog: bump version to 1:2.1.0~alpha0
[22:29:31] <anonimasu> alex_joni: stuff should arrive tomorrow I think
[22:29:39] <alex_joni> nice
[22:29:41] <alex_joni> any pictures yet?
[22:29:58] <anonimasu> on
[22:29:59] <anonimasu> no
[22:29:58] <anonimasu> :)
[22:30:26] <alex_joni> * alex_joni wonders why anonimasu is wasting time in here instead of doing the "right" thing
[22:30:38] <anonimasu> alex_joni: becaues it's 23:30
[22:30:50] <anonimasu> I were to work 10 minutes ago and let a machine in there..
[22:31:02] <anonimasu> didnt feel like staying :)
[22:31:32] <CIA-8> 03alex_joni 07v2_1_branch * 10emc2/debian/.cvsignore: new package name files to ignore
[22:31:39] <alex_joni> 00:29 < anonimasu> alex_joni: becaues it's 23:30
[22:31:44] <alex_joni> that's no reason :P
[22:31:48] <anonimasu> because..
[22:31:49] <alex_joni> some are still working now :P
[22:31:49] <anonimasu> yeah
[22:31:53] <anonimasu> lol
[22:31:55] <anonimasu> well, I'm not
[22:34:14] <CIA-8> 03alex_joni 07v2_1_branch * 10emc2/debian/ (configure control.in):
[22:34:14] <CIA-8> change naming of the packages:
[22:34:14] <CIA-8> the main package will be named 'emc2' so it can easily upgrade 2.0.x packages
[22:34:14] <CIA-8> the sim package is called 'emc2-sim'
[22:35:03] <cradek> yay
[22:35:10] <cradek> that makes me happy
[22:35:16] <anonimasu> alex_joni: my geckos are still a bit away
[22:35:43] <lerneaen_hydra> BØRK BØRK BØRK
[22:36:40] <alex_joni> 5263 messages since the first one, 2.35 years ago, for an average of 3.9 hours between messages
[22:36:43] <alex_joni> wow !
[22:37:32] <alex_joni> quite a bit of work on emc2 :)
[22:37:40] <anonimasu> yeah
[22:37:41] <alex_joni> mostly jepler .. judging by the stats
[22:37:44] <alex_joni> :-P
[22:37:50] <anonimasu> alex_joni: well, it still sucks according to paul_c ;)
[22:38:02] <lerneaen_hydra> 3.9 hours between average commit?
[22:38:08] <alex_joni> lerneaen_hydra: go figure..
[22:38:13] <lerneaen_hydra> well well
[22:38:18] <lerneaen_hydra> not bad at all
[22:38:43] <alex_joni> http://cia.navi.cx/stats/project/emc
[22:39:47] <lerneaen_hydra> lots of big projects seem to be using ciabot
[22:40:54] <alex_joni> and irc channels
[22:41:01] <lerneaen_hydra> yeah
[22:41:22] <lerneaen_hydra> IRC seems to be quite well suited actually for devving over distances
[22:45:20] <cradek> yes it sure is
[22:45:25] <anonimasu> * anonimasu nods
[22:45:33] <alex_joni> * alex_joni seems to nod
[22:45:45] <alex_joni> but actually I fell asleep and my head fell down :)
[22:45:47] <cradek> if it would somehow warp time zones to match it would be better
[22:45:52] <alex_joni> heh.. yeah :D
[22:46:05] <alex_joni> or if I wouldn't need to wake up in the morning
[22:46:11] <cradek> goodnight alex :-)
[22:46:39] <alex_joni> I put the new packages online: http://dsplabs.cs.utt.ro/~juve/tempdebs/
[22:46:54] <cradek> yay, that's very exciting
[22:46:54] <alex_joni> still alpha though.. they install/upgrade just fine for me
[22:47:02] <cradek> I may try them tonight
[22:47:09] <alex_joni> maybe someone around here cares to try ..
[22:47:11] <K`zan> My belt drive kit arrived!!!!!!!!!!
[22:47:25] <alex_joni> cradek: those are not final, I didn't tweak any folder paths or whatever
[22:47:42] <alex_joni> built from CVS
[22:47:57] <cradek> ok
[22:48:25] <alex_joni> maybe skunkworks is later online
[22:48:34] <alex_joni> usually he tries these things out :D
[22:53:13] <anonimasu> K`zan: nice!
[22:53:13] <anonimasu> :9
[22:53:42] <K`zan> Sure should be, I can now use my flycutter without cringing :-)
[22:54:48] <K`zan> http://www.tj-cnc.com/index.html from these folks
[22:55:19] <alex_joni> omg is that a bad webpage design
[22:55:19] <K`zan> Eager to do it, but my wrists and shoulders just are not up to it ATM :-(.
[22:55:45] <anonimasu> now just some motors ;) and emc2
[22:56:08] <K`zan> motors, drivers and the emc2 box is already in there :).
[22:56:24] <K`zan> and mounts...
[22:56:57] <jepler> 8000RPM? darn, pcb trace isolation milling is best at even higher RPMs
[22:57:17] <K`zan> Still thrashing on the drives and feeling braver about making mounts. So many things needed, sigh.
[22:57:39] <K`zan> Yep, either that or make a mount for the dremel for that on the side of the head.
[22:58:52] <K`zan> Duct tape it to the side of the head :) LOL.
[23:04:53] <anonimasu> :)
[23:07:30] <awallin> good night.
[23:09:20] <alex_joni> good night all
[23:09:37] <anonimasu> night alex
[23:11:48] <daniel_br> anonimasu: I also have one usc
[23:11:59] <anonimasu> how do you like it?
[23:12:22] <daniel_br> i have a stepper closed loop working fine
[23:12:37] <jepler> other emc users would love to hear about that
[23:12:46] <jepler> we had a user in here yesterday (I think) asking about it
[23:12:51] <anonimasu> im running step servos..
[23:13:15] <daniel_br> after some pid tunning all fine
[23:14:25] <daniel_br> for home sensor i m using this http://www.farnell.com/datasheets/234.pdf
[23:14:32] <anonimasu> :)
[23:14:37] <anonimasu> I like the usc
[23:15:01] <anonimasu> though if I were buying my setup newly I'd buy that pwm stuff jon's made
[23:15:36] <daniel_br> that is good also
[23:18:38] <daniel_br> these hall switches are cheap and very good
[23:29:13] <lerneaen_hydra> 'night all
[23:29:35] <anonimasu> night lh
[23:47:24] <daniel_br> 'night
[23:47:44] <dastumster> bbs
[23:47:46] <anonimasu> night