#emc | Logs for 2009-03-19

[01:41:39] <toastydeath> pandora radio on the blackberry is killer
[01:41:43] <toastydeath> just in case anyone is wondering?
[01:41:44] <toastydeath> it is killer.
[02:10:19] <dareposte> hi all
[02:11:21] <hypa7ia> hi dareposte
[02:11:33] <dareposte> wouldn't you know it, all set to go pick up a milling machine, and forecast is for rain
[02:12:04] <hypa7ia> oh no :(
[02:12:07] <hypa7ia> bring a tarp!
[02:12:15] <dareposte> heh
[02:12:26] <dareposte> i'm guessing knee mills and rain showers don't mix very well
[02:13:01] <dareposte> i wonder if a tarp would actually be effective...
[02:13:10] <dareposte> its about a 60 mile drive on the highway
[02:14:34] <dareposte> sigh
[02:15:12] <dareposte> looks like it should clear up by noon though, so maybe a later trip would be okay
[02:20:48] <hypa7ia> where are you, dareposte ?
[02:20:59] <hypa7ia> oh later not necessarily today
[02:21:01] <dareposte> kentucky
[02:21:06] <hypa7ia> * hypa7ia was trying to figure out time zones :p
[02:21:11] <dareposte> oh its evening now
[02:21:19] <dareposte> i was planning on getting it when they opened at 8am
[02:21:32] <dareposte> but the showers are scheduled to go until about 11
[02:21:46] <hypa7ia> gotcha
[02:21:57] <hypa7ia> good time to be buying this kinda thing in the US eh
[02:22:02] <dareposte> yeah great time
[02:22:20] <dareposte> picked up a 9x42 bridgeport copy for $1050 usd
[02:22:38] <dareposte> allegedly in good condition, but that remains to be seen
[02:22:54] <dareposte> he offered a 30 day return policy if there were any problems with it, so it may be pretty good
[02:24:32] <hypa7ia> i don't know what that is :s
[02:24:40] <hypa7ia> * hypa7ia is just here due to the hacklab.to lazzor
[02:24:59] <dareposte> hacklab?
[02:25:44] <dareposte> pretty standard milling machine, they usually go for $3000 or so used
[02:26:02] <dareposte> mine's a copy, maybe $2500 would be a normal price around here
[02:26:50] <dareposte> like a knockoff
[02:26:53] <dareposte> not the original
[02:28:26] <hypa7ia> gotcha
[02:28:30] <hypa7ia> http://hacklab.to
[02:28:36] <hypa7ia> we rebuilt a ULS laser engraver
[02:29:26] <dareposte> oh it looked like a plotter at first
[02:29:42] <dareposte> until the laserbeam shot out
[02:29:49] <dareposte> and started scorching stuff
[02:29:51] <dareposte> very cool
[02:29:52] <dareposte> co2 laser?
[02:31:27] <hypa7ia> yar
[02:31:28] <hypa7ia> 30W
[02:33:20] <dareposte> yar?
[02:33:25] <dareposte> not sure wat that is
[02:33:44] <hypa7ia> sorry
[02:33:46] <hypa7ia> yar = yes
[02:33:47] <JymmmEMC> I thnk he meant YAG
[02:33:53] <hypa7ia> she, actually
[02:33:54] <hypa7ia> :)
[02:34:12] <JymmmEMC> =)
[02:35:05] <hypa7ia> plan for this weekend is to work on a gcode plugin for inkscape
[02:35:16] <JymmmEMC> Don't bother
[02:35:17] <dareposte> i am so turned on right now
[02:35:45] <dareposte> a girl coder who is making a gcode plugin for inkscape??
[02:35:47] <dareposte> that is unreal
[02:36:10] <JymmmEMC> Inkscape has a very old bug in it that the math is incirrect
[02:36:16] <JymmmEMC> incorrect
[02:36:23] <dareposte> with a co2 laser gantry
[02:36:30] <hypa7ia> dareposte: http://xkcd.com/322/
[02:36:33] <hypa7ia> don't be that guy :(
[02:36:52] <hypa7ia> JymmmEMC: interesting, tell me more?
[02:36:52] <cradek> hypa7ia: dangit, I was just furiously trying to find that comic
[02:37:08] <hypa7ia> cradek: hehe, thanks, backup appreciated :)
[02:37:08] <dareposte> lol
[02:37:49] <JymmmEMC> Ok, but if one is NOT joking and asks for naked pics, what does that make them then?
[02:38:09] <hypa7ia> JymmmEMC: plain old creepy :/
[02:39:17] <hypa7ia> JymmmEMC: i'd much rather hear more about the inkscape math bug
[02:39:53] <dareposte> me too actually, i thought vector drawing was the cats meow these days
[02:40:13] <JymmmEMC> Just something I found years ago, Export to SVG and look at the source.
[02:40:58] <dareposte> cann't inkscape export to dxf though
[02:41:17] <dareposte> i was pretty sure i'd seen that in there
[02:41:38] <hypa7ia> yeah, but we want gcode
[02:41:50] <dareposte> for your laser?
[02:42:10] <dareposte> there are several good dxf cam packages for free
[02:42:32] <hypa7ia> we have a good workflow with gcode now
[02:42:38] <hypa7ia> using emc2
[02:43:49] <dareposte> burning graphics?
[02:46:39] <hypa7ia> yup
[02:47:34] <dareposte> line art only, or shading too
[02:49:19] <dareposte> for something like that laser it would be really cool if you could get a postscript interpreter for it and do full out dot matrix printing with it
[02:51:03] <hypa7ia> we're working on shading now
[02:51:39] <dareposte> http://www.chiark.greenend.org.uk/~alanb/ps2gc.html
[02:51:49] <dareposte> there's a ps to gcode translator someone wrote
[02:52:14] <dareposte> looks sort of similar to what you're working on actually
[02:53:20] <dareposte> gtg
[02:53:22] <hypa7ia> we actually already have the code working
[02:53:45] <hypa7ia> but one of the members wrote it in VB :/
[02:53:52] <hypa7ia> so it needs translating to python
[03:13:29] <fenn> hypa7ia: you should try this out http://dp.jstenback.com/gcode/
[03:13:52] <fenn> it converts curves to biarcs, for better smoothness in the output
[03:15:27] <hypa7ia> cool, thanks fenn :)
[03:17:31] <fenn> i suppose i should try it too
[03:22:57] <fenn> hot diggety it works
[03:26:02] <jmkasunich> hypa7ia: how many different people are involved in the laser project?
[03:26:12] <jmkasunich> seems like I've seen half a dozen names lately
[03:27:06] <fenn> oops.. almost works at least. seems to have fumbled a coordinate for some reason
[03:27:32] <jmkasunich> hi seb
[03:27:57] <garage_seb> hi there
[03:28:00] <fenn> axis missed the bad coordinate on the preview because it's an arc?
[03:28:04] <garage_seb> just the man i wanted to meet :-)
[03:28:11] <jmkasunich> uh-oh
[03:29:44] <jmkasunich> that explanation of 1.5*dt probably didn't shed a whole lot of light, did it?
[03:30:09] <garage_seb> http://highlab.com/~seb/emc2/Screenshot-HAL%20Oscilloscope.png
[03:30:23] <jmkasunich> sweet
[03:30:52] <garage_seb> i just copied your code without understanding it... :-/
[03:31:10] <jmkasunich> at this point, I'd be doing the same
[03:31:49] <jmkasunich> did you try it with the 1.5 set to zero for grins?
[03:31:59] <jmkasunich> I think you get ringing or overshoot or some such
[03:32:04] <jmkasunich> I just remembered another test case
[03:32:19] <jmkasunich> use a square wave
[03:32:36] <jmkasunich> I remember seeing it overshoot by 1 step, and then take a step back
[03:32:54] <jmkasunich> you really don't want to see dir change when you move to a constant target
[03:33:45] <jmkasunich> maybe it didn't actually overshoot by a whole step, but it went far enough that dir changed (even tho no step was issued while dir was in the opposite state)
[03:35:04] <jmkasunich> * jmkasunich shuts yo
[03:35:08] <jmkasunich> up even
[03:36:52] <garage_seb> http://highlab.com/~seb/emc2/Screenshot-HAL%20Oscilloscope-square.png
[03:37:09] <garage_seb> gotta zoom in to see if it overshoots...
[03:39:12] <jmkasunich> I forgot you can't look at dir with halscope on the 5i20, it isn't a HAL pin
[03:39:46] <jmkasunich> vel fb should be usefull tho, blow it way up
[03:40:02] <jmkasunich> damn slider doesn't work on that image ;-0)
[03:40:16] <jmkasunich> I can't even type ;-) right tonight
[03:40:31] <garage_seb> lol
[03:40:55] <jmkasunich> one of these days I want halscope to be able to store/retrieve data so you can zoom and pan while viewing
[03:41:24] <garage_seb> http://highlab.com/~seb/emc2/Screenshot-HAL%20Oscilloscope-square-zoomed.png
[03:41:26] <garage_seb> looks good to me
[03:41:32] <jmkasunich> gets complicated tho, since the person viewing almost certainly won't have the same pins, sigs, etc, and really doesn't even want to have hal running
[03:41:41] <jmkasunich> purty
[03:42:12] <garage_seb> heh, all i had to do was copy jmkasunich and it start working ;-)
[03:42:15] <garage_seb> thanks :-)
[03:43:31] <jmkasunich> you're welcome
[03:43:58] <jmkasunich> I just learned something - my 0.2497 reamer reams >0.2501 when it isn't lubed
[03:44:19] <jmkasunich> dowel pins are a snug slip fit instead of the tap fit I expected
[03:44:26] <fenn> is that good or bad?
[03:44:33] <jmkasunich> slightly annoying
[03:44:46] <jmkasunich> these pins are just to give me something to clamp down, they're coming out later
[03:44:52] <fenn> the hole is probably not round
[03:45:27] <jmkasunich> reamers usually make fairly round holes, but they can be finicky about holding perfect diameters
[03:45:43] <jmkasunich> the WD-40 that I usually use for lube when reaming is outside by the mill
[03:45:52] <jmkasunich> (which is where I gotta go soon)
[03:46:37] <jmkasunich> bbl
[03:52:38] <hypa7ia> jmkasunich: about 5
[03:52:58] <hypa7ia> the lab has 25 members, but only a few are working on it so far
[03:56:55] <cradek> I've had that problem too. what was the speed?
[03:57:18] <cradek> oh, he left
[04:00:04] <cradek> I was draining the water from the compressor today. it didn't sound right - it wasn't hissing enough for as far as I had turned the drain. I looked and the drain was stuck and the NPT was unscrewing from the tank
[04:00:50] <eric_unterhausen> that could have gotten exciting
[04:00:54] <cradek> I don't know what had happened if I had unscrewed it, but I'm glad I noticed
[04:01:12] <cradek> valve - hand - one inch - concrete floor
[04:01:52] <garage_seb> http://highlab.com/~seb/emc2/Screenshot-HAL%20Oscilloscope-ferror.png
[04:02:06] <cradek> 60 gals of 175 psi air...
[04:02:38] <eric_unterhausen> you would have expereinced some water injection, no doubt
[04:02:49] <eric_unterhausen> what kind of compressor is it?
[04:02:57] <garage_seb> that's starting at X=0 and running G1 X1 F9999
[04:03:04] <SWPadnos> 175PSI shouldn't be enough for injection
[04:03:04] <cradek> upright tank bolted to the floor
[04:03:18] <cradek> two stage oil sump
[04:03:20] <SWPadnos> I think they use 1200 or so for air needles
[04:03:36] <garage_seb> oops g1 x2
[04:03:41] <eric_unterhausen> I think the asme wants you to have below 40 psi to avoid water injection
[04:03:45] <SWPadnos> (then again, they could have been lying - it was the army after all)
[04:04:07] <eric_unterhausen> the difference being the army wanted to get it subcutaneous
[04:04:14] <SWPadnos> yeah, big time ;)
[04:04:46] <eric_unterhausen> you just reminded me of my least favorite air force vaccination story which I will keep to myself
[04:04:57] <cradek> and thanks for that
[04:05:04] <SWPadnos> heh
[04:05:07] <cradek> (eek)
[04:05:34] <cradek> I wonder how many times the valve could have hit me before it stopped bouncing off things...
[04:05:50] <eric_unterhausen> probably twice
[04:06:02] <eric_unterhausen> second time it would be stuck
[04:07:01] <cradek> goodnight folks
[04:07:05] <SWPadnos> heh
[04:07:12] <SWPadnos> see you later
[04:07:39] <JymmmEMC> cradek: Gives a totall new menaing to body piercing
[04:09:01] <garage_seb> goodnight
[04:41:06] <jmkasunich> hypa7ia: hacklab sounds like a fun place
[04:41:23] <hypa7ia> it's teh awesome
[04:41:51] <jmkasunich> most of us are scattered all over the place, and only get together once a year or so
[04:43:19] <jmkasunich> http://www.youtube.com/watch?v=l5q1rVKYLTQ
[04:43:26] <jmkasunich> the annual get-together ^^^
[04:44:08] <hypa7ia> cool, where does that happen?
[04:44:18] <jmkasunich> galesburg IL
[04:44:34] <hypa7ia> neat :)
[04:44:35] <SWPadnos> well, it did anyway ...
[04:44:41] <hypa7ia> we're up in toronto
[04:44:41] <jmkasunich> unfortunately the guy who hosted it had some health problems, so it won't happen this year
[04:44:47] <hypa7ia> oh, that sucks
[04:44:50] <hypa7ia> you should run it in toronto
[04:44:55] <SWPadnos> heh
[04:44:57] <hypa7ia> at the hacklab
[04:45:09] <SWPadnos> hey, that's within driving distance at least :)
[04:45:11] <jmkasunich> we're planning a smaller get-together in wichita kansas
[04:45:17] <SWPadnos> unlike Wichita
[04:45:21] <hypa7ia> hehe
[04:45:21] <jmkasunich> yeah
[04:45:26] <jmkasunich> for me and you at least
[04:45:33] <SWPadnos> and there's always Rodney's in Toronto
[04:45:58] <SWPadnos> I wonder if the "Nine of Cups" is still there
[04:46:01] <jmkasunich> the wichita gathering will be at this place: http://jmkasunich.com/cgi-bin/blosxom/shoptask/wichita-trip-02-20-08.html
[04:46:33] <jmkasunich> is hacklab basically just a collective workshop?
[04:47:16] <hypa7ia> jmkasunich: pretty much
[04:47:32] <hypa7ia> we're a non-profit member space
[04:47:38] <hypa7ia> with lots of open events
[04:47:43] <hypa7ia> a geek community centre :)
[04:47:56] <jmkasunich> * jmkasunich reads hacklab.to
[04:49:01] <jmkasunich> mostly electronics?
[04:49:24] <hypa7ia> and software
[04:49:29] <hypa7ia> and security stuff
[04:49:37] <jmkasunich> CNC tends to be larger, heavier, and dirtier
[04:49:49] <toastydeath> for those with true grit.
[04:49:56] <toastydeath> and, my friend, we are chock full of that.
[04:50:07] <hypa7ia> the hackerspace in seattle has a ton of CNC gear
[04:50:21] <fenn> take anything toasty says with a grain of grit
[04:50:24] <hypa7ia> they converted a lathe and are working on cnc-ifying an old hitachi milling machine
[04:50:31] <jmkasunich> using EMC?
[04:50:56] <hypa7ia> no... they used it years back and weren't impressed, but they may give it another try given the success we've had with it
[04:51:34] <jmkasunich> bigger and dirtier: http://www.youtube.com/watch?v=mxxdq6y8z8M&feature=related
[04:51:49] <jmkasunich> that machine is in wichita
[04:52:09] <SWPadnos> did we take any videos of the G&L moving?
[04:52:24] <jmkasunich> no, I don't think anyone had a video-ish camera
[04:52:30] <SWPadnos> bummer
[04:52:46] <SWPadnos> maybe this year, I think Stuart got all 4 axes tuned
[04:52:49] <hypa7ia> * hypa7ia must head bedwards
[04:53:07] <jmkasunich> I thought it was on hold while he worked on the cinci?
[04:53:08] <SWPadnos> see ya
[04:53:12] <jmkasunich> goodnight hypa7ia
[04:53:25] <SWPadnos> could be, but he was really making a lot of progress on it soon after we left
[04:56:40] <SWPadnos> oh. it is bedtime, isn't it
[04:56:42] <SWPadnos> night folks
[04:56:45] <jmkasunich> yes
[04:56:52] <jmkasunich> gotta walk the dog first...
[04:56:54] <jmkasunich> goodnight
[07:12:27] <Tottish> OK, anyone out here?
[08:06:46] <JymmmEMC> alex_joni: You HAVE to see this, you'll LOVE IT!!! ... http://www.opb.org/programs/ofg/videos/view/56-Electric-Drag-Racing
[08:11:45] <Tottish> Sorry, jimmy. I seem to be the only one awake in here and I have no soundboard. =)
[08:12:52] <micges_emc> good morning :)
[08:13:10] <Tottish> Howdy
[08:13:27] <maddash> edt
[08:13:40] <maddash> where's the cvs for emc3?
[08:14:08] <micges_emc> you mean 2.3 ?
[08:14:31] <maddash> no, i mean emc3
[08:14:44] <maddash> there was an email going around in the mailing list about emc3
[08:25:00] <micges_emc> maddash: that are only plans
[08:25:44] <micges_emc> to put rt-ethernet, usb and so on into EMC2
[08:25:53] <micges_emc> not official
[08:29:38] <maddash> i see.
[09:19:35] <maddash> what's the cheapest hardware stepgen method?
[12:47:58] <alex_joni> http://www.telegraph.co.uk/news/newstopics/howaboutthat/5005022/Teens-capture-images-of-space-with-56-camera-and-balloon.html
[13:28:27] <skunkworks> jeepers - why do I keep hitting the back button on this web page? duh
[13:28:44] <skunkworks> alex_joni: that is awesome
[13:29:36] <Tottish> Yeah, makes you wonder how much luck/skill was involved though. =)
[13:30:36] <archivist> alex_joni, there are some brits doing some similar stuff
[13:31:16] <Tottish> Did the sensor equipment have any parachute or what?
[13:32:07] <eric_unterhause2> people are doing that all the time nowadays
[13:33:06] <archivist> http://www.ukhas.org.uk
[13:33:25] <eric_unterhause2> it's certainly a cool project
[13:33:49] <eric_unterhause2> there was some guy up in Canada that was dropping autonomous gliders
[13:34:42] <archivist> and if you get hooked #highaltitude
[13:39:02] <Dallur> I would just send the signal via wifi with a parabolic antenna, no need to recover the camera, just let the whole thing float off into oblivion
[13:39:48] <Dallur> real time image and altitude information
[13:40:11] <Optic> moo
[13:40:31] <Tottish> You don't think it would be to heavy? I doubt you could fit a 20mile WiFisender with batteries in a 1,5kg package. =)
[13:40:42] <Tottish> Hi optic
[13:41:42] <archivist> 20 miles is trivial with the right setup, Ive done 136 km with 8mw on 10ghz
[13:42:16] <Tottish> No way!? Now thats impressive. and what about the weight and atmospheric conditions?
[13:42:55] <archivist> I admit thats using dishes both ends carefully aligned
[13:43:20] <archivist> but was wideband TV
[13:43:37] <Tottish> yeah,but still its really cool. What baudrate did you reach?
[13:43:50] <archivist> TV video
[13:44:07] <Tottish> ok, so thats an analogue signal or what?
[13:44:20] <archivist> we FM the carrier
[13:44:46] <Tottish> okidoke
[13:45:13] <archivist> balanced waveguide mixer
[13:45:32] <Tottish> Thats greek to me, baby... =)
[13:51:51] <Tottish> OK, so who's up for a little Q&A about how I can get my laser-programming-thingy started? It's about HAL-programming and something more, I guess
[13:52:52] <Tottish> Is there a parameter that can be read from a HAL that contain info on which row of G-code theprogram is currently on?
[13:53:44] <cradek> Tottish: http://www.linuxcnc.org/docs/devel/html/config_emc2hal.html
[13:55:23] <Tottish> OK, it doesn't seem to be in there, but this is all the accesible components from HAL?
[13:55:44] <cradek> I see it...
[13:56:09] <cradek> this is all the realtime state of emc that is available. halui can give you some more but it is not up to date in realtime.
[13:56:25] <cradek> (motion.program-line)
[13:57:01] <Tottish> ah, thnx! II searched for gcode... =)
[14:01:55] <Tottish> What I want to do is when I run a program, the HAL (or whatever has to be programmed) should go into a textfile containing a table. And at a certain point at some Gcode-rows it should turn the laser on and then off again at a later point. This is to burn a multi-shaded image with the laser. I found the component containing the DRO-value and now you helped me find the "G-row"-component but how do I access a textfile? is it possible from HAL?
[14:02:29] <cradek> in realtime, not really
[14:03:16] <archivist> code it into the gcode
[14:03:25] <Tottish> OK, what about in the currently loaded G-code? I'm thinking it might be faster?
[14:03:49] <Tottish> maybe I could place the info for the laser in comments
[14:04:05] <cradek> HAL doesn't have access to that, either
[14:05:29] <Tottish> Ouch. So, can you think of any other way to do it?
[14:05:51] <archivist> Tottish, I would create a program that writes the gcode
[14:07:21] <Tottish> Yeah but its not possible to switch the laser on/off while X/Y-axis is still moving and that is neceserry to get an even color. Acceleration/deceleration=longer exposure=darker shades around the edges..
[14:07:57] <Tottish> This is why I've abandoned the natural G-code-solution
[14:10:18] <archivist> care in code writing and there will be no speed changes
[14:10:18] <Tottish> http://www.cnczone.com/forums/showthread.php?t=75822
[14:10:20] <SWPadnos> alex_joni thought of one solution yesterday: create a HAL component that has a bitmap loaded (by a userspace helper), has X and Y inputs, and outputs a number based on those coordinates from the bitmap
[14:10:23] <Tottish> If you have any additional ideas I'm all ears. I discussed this in here last night and 'the guys' seemed to agree with me that it was not possible in G-code...
[14:10:35] <SWPadnos> you'd also need limit and resolution information
[14:11:19] <SWPadnos> the G-code then becomes a loop which just does scan lines, but has a preamble that sets up the image mapper component
[14:11:26] <SWPadnos> (you can use custom M codes for this)
[14:11:31] <eric_unterhause2> is there any profit in increasing the allowable acceleration?
[14:12:13] <SWPadnos> his idea is an extension to an idea I had, which was to use a linear lookup for each raster line
[14:13:10] <cradek> SWPadnos: how would that component get the image?
[14:14:14] <SWPadnos> userspace helper, using custom M codes, and written similarly to the way halstreamer gets its data from userspace
[14:14:51] <cradek> interesting idea...
[14:15:33] <SWPadnos> hmm. is there a number in the TP that represetns the fraction of the current move that's been completed?
[14:15:54] <SWPadnos> like from 0 to 1, not from (start point) to (end point)
[14:16:27] <Tottish> I'm not exactly down with the lingo here but maybe you could point me in the right direction. Anything to reccomend reading to start working on this?
[14:16:46] <SWPadnos> uh - I'm not sure really :)
[14:17:11] <cradek> SWPadnos: dtg is sort of like that
[14:17:29] <SWPadnos> I think there's some developer documentation for HAL, then you should look at some simple HAL components, and then take a look at the component "halstreamer"
[14:17:45] <SWPadnos> cradek, yeah, sort of
[14:18:08] <cradek> the tc progress could be exported to hal, but it's 0..len, not 0..1
[14:18:24] <Tottish> OK, i spent all morning reading the integrators manual so I know the basics about HAL. I will check halstreamer out.. in the wiki maybe?...
[14:18:24] <SWPadnos> right, which could be useful if len is also available
[14:18:30] <cradek> yes
[14:18:31] <SWPadnos> no, in the source
[14:18:56] <BJT-Work> http://www.linuxcnc.org/docview/devel/html//man/man1/halstreamer.1.html
[14:19:11] <cradek> of course tp could generate 0..1 if it wanted
[14:19:29] <SWPadnos> yeah. I'm not sure why I asked actually, so no worries :)
[14:21:58] <Tottish> OK, so a quick question: "streamer(9) and halstreamer are used together to stream data from a file into the HAL in real time" Why can't this component be used to stream from a textfile?
[14:22:36] <SWPadnos> you must make sure that the realtime component always has data. RT tasks must not block for any reason
[14:22:58] <SWPadnos> so you could do it, but you have to use a userspace helper to get the data
[14:23:49] <SWPadnos> (maybe we answered your question too literally - the RT component can't get the data from a text file, but with the userspace helper reading the data and sticking it in a buffer, an RT component can effectively stream data from a text file)
[14:24:21] <Tottish> OK, thats great!
[14:24:38] <SWPadnos> that's why I suggested you look at halstreamer ;)
[14:24:55] <Tottish> So then I guess I should concentrate on understanding the userspece next then...
[14:25:41] <Tottish> Yeah, problem is that I barely knew what HAL was up until this morning.
[14:25:58] <SWPadnos> it's OK, you have until Monday :)
[14:26:09] <Tottish> lol
[14:26:42] <archivist> the examination is Saturday
[14:26:58] <SWPadnos> shhh. it was supposed to be a *surprise* quiz
[14:27:23] <Tottish> thnx for tipping me off achivist, =)OK, so where do I get the relevant info on 'userspace'?
[14:27:56] <SWPadnos> years of programming experience
[14:27:58] <SWPadnos> err
[14:28:08] <SWPadnos> have you downloaded the source yet?
[14:28:26] <Tottish> no, should I? =)
[14:28:31] <SWPadnos> yeah
[14:28:41] <SWPadnos> on the wiki there's a guide to contributing
[14:29:01] <SWPadnos> also the installation page, where some of the options are "from source"
[14:29:17] <SWPadnos> you said you do microcontroller work, right?
[14:29:39] <Tottish> Yeah sure, assembler on PICs
[14:29:46] <SWPadnos> ah, ok. use C much?
[14:29:55] <Tottish> no, sorry =(
[14:30:20] <SWPadnos> then I'd definitely study up on that, since you're about to download several megabytes of C source code :)
[14:30:50] <Tottish> Yeah, I was kinda hoping I didn't have to go there...
[14:30:55] <SWPadnos> you do
[14:31:01] <Tottish> just bite the bullet, iguiess
[14:31:08] <Tottish> I guess..
[14:31:24] <SWPadnos> you can look at "comp", which lets you more or less write the code and not deal with a lot of the details
[14:31:47] <Tottish> OK, I will
[14:32:15] <SWPadnos> note that I said "more or less". this may be one of the applications that comp isn't suited for - I don't know
[14:32:24] <Tottish> Thnx alot you guys, I don't know if I can grasp this but I'll give it a shot and see where it takes me...
[14:32:47] <SWPadnos> sounds like a good plan
[14:33:41] <Tottish> yeah, =) Now i will have to go do some cooking! SeeYa!
[14:33:51] <SWPadnos> see you
[14:34:04] <jepler> with shared memory, you have to do extra setup -- comp may get in your way more than it helps. I'd be more tempted to start with streamer_rt/halstreamer and hack it until it takes the data in your desired format and can be "clocked" appropriately..
[14:34:34] <jepler> but you probably covered all that in the scrollback I didn't read :-P
[14:34:45] <SWPadnos> no, that was good and useful info :)
[14:34:47] <cradek> do you think one scanline worth of buffer is enough? you have the 'retrace' time to load the next line
[14:34:59] <SWPadnos> yes, I think a single line is fine
[14:35:10] <jepler> I
[14:35:14] <SWPadnos> and may have other applications as well
[14:35:15] <jepler> 'd think in samples and not scanlines
[14:35:33] <cradek> I don't follow
[14:35:35] <jepler> no reason to wait until reaching the end of a scanline in order to transmit more dat
[14:35:38] <jepler> a
[14:35:40] <jepler> argh my typing
[14:35:43] <SWPadnos> heh
[14:35:44] <SWPadnos> laptop?
[14:35:45] <cradek> oh, well that's true
[14:35:52] <jepler> streamer will continuously transmit data when the buffer's not full
[14:35:54] <cradek> it's an epidemic today
[14:36:00] <SWPadnos> you don't have a linear progession of samples though
[14:36:05] <jepler> SWPadnos: no, desktop machine with nice ibm clicky keyboard
[14:36:13] <SWPadnos> oh. more coffee then :)
[14:36:18] <jepler> SWPadnos: no, but your userspace program and your rt will agree on the zigzag pattern to use
[14:36:41] <SWPadnos> I think the positional idea Alex had is better
[14:37:04] <SWPadnos> you could actually trace over the same line several times if you want (maybe with a tiny offset in Y)
[14:37:28] <SWPadnos> a "scanline doubler"
[14:38:49] <jepler> or just put the whole thing in a shared memory area. An 8x10 at 300DPI is less than 8 megs
[14:39:07] <SWPadnos> yeah, that was Alex's idea
[14:39:22] <SWPadnos> X and Y inputs for a lookup into the 2-d array
[14:49:01] <alex_joni> with some smart closeness factor you even get antialiasing ;)
[14:51:25] <SWPadnos> lending
[14:51:29] <SWPadnos> +b
[14:51:53] <SWPadnos> like screw comp
[14:52:02] <SWPadnos> only you have to change it all the time ;)
[14:53:13] <alex_joni> what I mean.. if you have a black dot in the image (assuming 0/1 only as teh info)
[14:53:24] <alex_joni> then you can turn the laser power on based on the distance to the point
[14:53:42] <SWPadnos> ah, interesting
[14:53:51] <SWPadnos> but you'd want it to be a blend from the next point
[14:53:58] <alex_joni> obviously ;)
[14:54:14] <tomp> hypa7ia: for some kool vector driven robotic s, google up Hektor, the grafitti machine ( a bipod like emc's alex has ) it uses 2 cables to move a can pf spray paint, and fontagrapher to generate paths
[14:54:15] <SWPadnos> well duh!
[14:54:53] <alex_joni> SWPadnos: get the 4? 8? nearest pixels, and set the power based on all 4/8
[14:55:13] <SWPadnos> 9
[14:55:30] <SWPadnos> at that point you might as well use a kernel
[14:55:32] <alex_joni> err.. yeah :D
[14:55:36] <SWPadnos> (like image processing)
[14:55:41] <alex_joni> yup
[14:55:49] <alex_joni> the possibilities are endless ;)
[14:56:00] <alex_joni> in b/w the "picture" won't be that big anyways
[14:56:02] <SWPadnos> endless^2 :)
[14:56:08] <alex_joni> ^9
[14:56:17] <SWPadnos> no, 2-dimensional data ;)
[14:56:36] <alex_joni> couple hundred k
[14:56:45] <alex_joni> * alex_joni runs home
[14:56:47] <alex_joni> see you later
[14:56:50] <SWPadnos> see you
[15:10:56] <tomp> still grokking the streamer map idea to turn laser on/off. this is realtime? and not at begin/endpoints of gcode? what syncs the data stream to the path?
[15:11:13] <SWPadnos> it doesn't work yet
[15:11:15] <SWPadnos> :)
[15:11:46] <SWPadnos> streamer is time based at the moment, though the idea to add external clocking has been floated before
[15:11:48] <cradek> aside from not yet being written, it sounds promising
[15:12:05] <tomp> i get that, its a new idea, but what might sync the stream to the path?
[15:12:25] <cradek> it's synced to position, not path
[15:12:38] <tomp> i thought of going fwd bwd in a prerecorded stream for edm
[15:12:41] <SWPadnos> I'd make a different component that takes in a position and does a lookup based on that and some offset/size parameters
[15:13:03] <SWPadnos> no, the positions aren't recorded, the output just depends on them
[15:13:10] <tomp> oh, posn, so use xy as indices
[15:13:17] <SWPadnos> yes, or do a line at a time
[15:13:30] <SWPadnos> either way would work
[15:13:31] <tomp> line = use x :)
[15:13:34] <SWPadnos> yep
[15:13:41] <SWPadnos> or Y if you want :)
[15:13:41] <tomp> cool idea
[15:13:50] <tomp> chines Y english X
[15:13:52] <hypa7ia> tomp: cool, thanks!
[15:14:08] <tomp> arabian -x
[15:15:24] <tomp> hypa7ia: yeah, i imagined a bosons seat Hektor project for the emc fest, or a 2d window washer seat for hi-rise
[15:15:25] <SWPadnos> hmmm. I wonder if I can whip up a 12V -> 20V, 5A power supply in the next week
[15:15:37] <SWPadnos> that doesn't look like a bomb, so they'll let me take it on an airplane :)
[15:16:35] <tomp> i've brought some scary prototypes onto planes, you get a 'special' guy to check it out, they call him up and you wait a while.
[15:17:10] <SWPadnos> this would just be so I could run my laptop off plane power
[15:17:29] <SWPadnos> the Radio Shack 140W inverter I have doesn't quite do it
[15:17:49] <SWPadnos> the power brick is woefully inefficient, rated for 2A at 120V
[15:17:53] <tomp> ah, a big pouch of LiFePo4 cells?
[15:17:58] <SWPadnos> heh
[15:19:40] <tomp> today i got 0.5mm x 400mm tubes running nice on the new edm, they steam :) and the servo doesnt bend the tube ( tho i made sevral glow like filaments during testing )
[15:21:54] <tomp> hmm, the prob was in the pic edm oscillator. i gotta look for open src pic programmers ( read not off the shelf frkn chinese, i cant rtfm ;)
[15:23:55] <tomp> a colleage decide to do me a favor and added an enable on the oscillator output, well it inverted the downstream data and got me some hefty current thru that tiney tube
[15:24:31] <tomp> bye bye, off to google pic prgmrs
[15:28:31] <cradek> can't pics be programmed with a few pins of a parport, like avrs?
[15:28:55] <cradek> I guess I used my old eprom programmer back when I used them
[15:30:27] <archivist> there are some designs out there for parport pic programming
[15:30:37] <archivist> but he left!
[15:31:59] <JymmmEMC> SWPadnos: It doesn't matter if it looks like a bomb. Well, if you get a clueful agent that is
[15:33:04] <JymmmEMC> SWPadnos: Oh and if you do build one, make it two =)
[15:40:40] <JymmmEMC> SWPadnos: Ah, here ya go... http://www.generatorpower.com.au/images/Cummins%20P1700i%20Petrol%20Portable%20Inverter%20Generator.jpg
[15:51:34] <SWPadnos> JymmmEMC, thanks, but I'm not sure that would fit in my laptop bag
[15:51:39] <SWPadnos> what with the laptop in there and all
[15:51:51] <JymmmEMC> SWPadnos: Carry on luggage =)
[15:52:05] <SWPadnos> that's what my suitcase is for
[15:52:07] <SWPadnos> :)
[15:54:38] <JymmmEMC> SWPadnos: Don't airlines limit to 75W outlets?
[15:54:48] <SWPadnos> hmmm. could be
[15:55:47] <SWPadnos> ok, it was 74W, but it's now 200W
[15:55:49] <SWPadnos> 75
[15:56:35] <JymmmEMC> and you want to build a 250W?
[15:56:42] <SWPadnos> no
[15:56:50] <SWPadnos> 20V * 6A = 120W
[15:56:51] <JymmmEMC> 20V@5A=250
[15:57:00] <SWPadnos> 20 * 5 = 100
[15:57:28] <JymmmEMC> stupid calculator
[15:57:31] <SWPadnos> heh
[15:57:44] <SWPadnos> I use my head for things that simple - it's more accurate ;)
[15:58:04] <JymmmEMC> Still on first cup of coffee
[15:58:30] <JymmmEMC> then why isnt your 140W inv enough?
[15:58:47] <SWPadnos> I think it's because the AC power supply for the laptop is inefficient
[15:59:00] <SWPadnos> it lists 120V @ 2A as the input power
[15:59:08] <SWPadnos> those could be peaks though
[15:59:23] <SWPadnos> and sticking it on my power meter shows that it never seems to get above 90 or so
[15:59:51] <skunkworks> you need a wind - up generator.. ;)
[16:00:10] <SWPadnos> no, a portable windmill I can stick out the window ;)
[16:00:45] <JymmmEMC> SWPadnos: Well if you find/build something let me know. I'd like to have a 12VDC to 20@2A for my netbook
[16:00:57] <SWPadnos> ok
[16:01:19] <SWPadnos> that's actually quite a bit easier, the 6A spec for mine eliminates a lot of switch controllers
[16:04:12] <skunkworks> you could use pwmgen out of emc to control the mosfets with a a/d converter to sense the voltage... (you just need enough battery power fo the laptop to start hal ;)
[16:04:41] <SWPadnos> don't forget the voltage doubler
[16:05:09] <cradek> don't forget that the TSA will say it's a bomb
[16:05:19] <JymmmEMC> SWPadnos: Just use TWO airline jacks and wire them in series =)
[16:05:26] <JymmmEMC> <rimshot>
[16:05:36] <SWPadnos> "um, excuse me. can I just plug this into your seat?" ...
[16:05:39] <SWPadnos> thanks
[16:05:40] <cradek> (make sure it's off-white and very finished-looking)
[16:06:33] <SWPadnos> or shiny black, with blue LEDs
[16:06:53] <cradek> not too many LEDs - bombs have lots of those
[16:06:57] <SWPadnos> well, at least I can charge the thing, so if worst comes to worst, I'll just eat or something :)
[16:07:00] <cradek> in the movies, you know
[16:07:10] <SWPadnos> only ones that count down or do the Cylon/Kitt thing :)
[16:07:20] <SWPadnos> so no bar graphs, that's for sure
[16:07:21] <cradek> oh right, no counting-down 7 segment LEDs
[16:07:28] <cradek> that's a bomb for sure
[16:07:53] <cradek> and exposed wires
[16:08:01] <SWPadnos> and I'd also leave off the user friencly 1-second beep
[16:08:09] <SWPadnos> friendly
[16:08:14] <cradek> be sure you can look at it and say "this would never in a million years be used as a prop for a bomb on TV"
[16:08:22] <SWPadnos> right
[16:08:25] <SWPadnos> or in the movies
[16:08:34] <cradek> which is a tall order, seems like
[16:08:42] <SWPadnos> hmmm. yeah
[16:09:00] <SWPadnos> I guess I'll just eat and charge, and not use the laptop the whole trip
[16:09:11] <SWPadnos> err, flight
[16:09:47] <JymmmEMC> SWPadnos: Or...get yourself a netbook
[16:10:03] <cradek> http://www.gizmodo.com.au/2008/07/tsa_confiscates_homemade_battery_and_water_bottle_declares_victory_over_terror-2.html
[16:10:07] <SWPadnos> that's the other option
[16:14:57] <JymmmEMC> SWPadnos: 8GB SSD + 160S hdd http://www.pcconnection.com/IPA/Shop/Product/Detail.htm?sku=9481197
[16:15:09] <JymmmEMC> s/S/GB/
[16:15:28] <JymmmEMC> on order though
[16:16:28] <SWPadnos> still $300 though, interesting
[16:16:34] <JymmmEMC> SWPadnos: 512MB is soldered on, max is 1.5GB
[16:16:59] <JymmmEMC> and you can buy a internal BT adapter for $16 shipped from ebay
[16:18:50] <JymmmEMC> SWPadnos: http://cgi.ebay.com/2-0-Bluetooth-cable-Module-screw-for-Lenovo-S10-S10e-S9_W0QQitemZ320349143688QQcmdZViewItemQQptZLH_DefaultDomain_0?hash=item320349143688&_trksid=p3286.c0.m14&_trkparms=72%3A1234|66%3A2|65%3A12|39%3A1|240%3A1308|301%3A1|293%3A1|294%3A50
[16:27:01] <JymmmEMC> SWPadnos: Oh, and make sure to bring a binary clock powered by four 7.2V RC battery packs
[16:27:19] <JymmmEMC> painted dark red of course
[16:28:06] <SWPadnos> I thought we had been through that already :)
[16:28:40] <JymmmEMC> SWPadnos: If you haven't had a full body cavity search, then no, you haven't =)
[16:28:58] <SWPadnos> what if I don't want a full body cavity search?
[16:29:32] <JymmmEMC> SWPadnos: I'll be sure to call MrsSWPadnos then
[16:29:48] <SWPadnos> what if she doesn't want one either?
[16:29:57] <JymmmEMC> SWPadnos: sucks to be you
[16:50:15] <Guest748> hello !
[16:54:37] <Guest748> Is there someone who can point to more info about a system with 2 x-axis (using 2 x-motors) and let emc do the work concerning homing etc. ?
[16:56:40] <archivist> homing is awkward, google gantrykins eg http://axis.unpy.net/01162326817
[16:58:25] <SWPadnos> Guest748, there are several ways to do it, though I don't think any are perfect
[16:58:37] <SWPadnos> you'd just have to choose one that has drawbacks you can live with
[16:59:39] <SWPadnos> it's also dependent on what kind of motors you use and how they're driven, among other things
[17:01:15] <Guest748> true. Now I'm using 2 seperate stepper drives which uses the same input signals
[17:02:25] <SWPadnos> ok. if you have separate home switches, you can use an or2 or and2 (depending on the polarity) to decide whether steps get sent to each drive
[17:02:54] <SWPadnos> you'd have to use separate step/dir pins for the two drives (or at least separate step pins)
[17:03:26] <SWPadnos> the step output should get combined with the limit switch input, and the step only passed out to the drive if the limit switch is not hit
[17:03:32] <SWPadnos> oops, home switch
[17:04:32] <SWPadnos> the home input would be the combination of the two home switches, so EMC keeps trying to move the motors until both are at home, but either one will stop as soon as it hits its switch
[17:05:02] <Guest748> nice idea...very simple
[17:05:12] <SWPadnos> let us know if it works :)
[17:07:01] <skunkworks> that is neat
[17:08:29] <skunkworks> how do you back off then?
[17:08:30] <Guest748> i've 2 x axis wich drive together the y-axis connection between x and y is flexible. So a little offset doesn't cause mechanical problems. But the home position of both x axis is always al little different (ok it's to solve manual but not that nice sollution)
[17:08:46] <SWPadnos> skunkworks, that's left as an exercise for the reader :)
[17:08:54] <skunkworks> heh
[17:09:00] <SWPadnos> (but you could also use the DIR line in the logic)
[17:09:33] <SWPadnos> port pin = ((switch | dir) & step)
[17:09:45] <SWPadnos> suitably adjusted for positive/negative logic
[17:11:28] <SWPadnos> hmmm. is that right? (switch&!dir || dir) may be it
[17:11:54] <SWPadnos> ah, they're equivalent enough for this purpose
[17:12:14] <SWPadnos> maybe. time for more coffee
[17:12:17] <micges> logger_emc: bookmark
[17:12:17] <micges> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2009-03-19.txt
[17:13:36] <Guest748> thanks :)
[17:13:52] <SWPadnos> sure. like I said, let us know if it works
[17:16:10] <Guest748> I checked google but didn't find much info about this topic ...lot's of questions not much useful info. I will let you know what works the best
[17:19:49] <Guest748> thanks all for the response!
[17:20:22] <SWPadnos> sure. good luck with it
[18:59:00] <motioncontrol> good evening.i have one problem with emc2 cvs. in the axis interface in g code active list the active code not change.example if i have active g90 and start in part program g91 the active code not change is g90. because?
[19:00:45] <cradek> those show the state of the interpreter, which reads ahead past the current motion
[19:01:02] <cradek> it is really only useful in MDI mode
[19:02:09] <motioncontrol> ok.in automatic mode not fuction equal another control ? only on axis interface or all interface (tkemc) ?
[19:02:41] <cradek> the active code readout works the same in all the GUIs
[19:04:20] <motioncontrol> ok thanks cradek.you thing in the futhure version the automatic change the g active code in automatic mode will possible?
[19:05:47] <cradek> I don't know, sorry
[19:06:30] <motioncontrol> ok thanks for info
[19:22:17] <jensor> How does one close axis from a terminal window?
[19:22:47] <cradek> axis-remote --quit
[19:22:58] <jensor> thanks
[19:25:43] <jensor> doesn't work - from the same terminal window that emc was started in
[19:26:08] <jensor> tried open ing another terminal window and didn't work there either
[19:26:27] <jensor> cant do it normally because of error
[19:27:26] <jensor> no axis display appearing. If I try to restart it claims it is still running and asks if I want to restart
[19:27:53] <SWPadnos> the terminal window in which you started emc won't work, because the terminal doesn't respond to anything until emc exits
[19:28:21] <SWPadnos> if you are using run-in-place, then you'll need to source emc-environment in the window in which you want to use axis-remote
[19:28:29] <jensor> Appearently the error was generated because it couldn't find the comp file
[19:28:41] <jensor> I inadvertently erased
[19:28:53] <SWPadnos> oh
[19:30:31] <jensor> So I guess the onoly way is to shut down and reboot?
[19:30:43] <BJT-Work> * BJT-Work heads out to work on his tractor
[19:30:54] <SWPadnos> I'm pretty sure you aren't telling us everything, so it's impossible to answer
[19:31:08] <SWPadnos> closing AXIS has nothing to do with a missing comp file
[19:36:30] <jensor> I inadvertently deleted my comp files. Wheen I start emc I get the flash screen and then nothing. The error is:emc/ini/iniaxis.cc 244 bad return from emcAxisLoadComp
[19:37:10] <jensor> emc/task/emctaskmain.cc 2605 can't initialize motion
[19:37:37] <SWPadnos> you could try halrun -U
[19:38:13] <alex_joni> jensor: cvs up ;)
[19:38:23] <alex_joni> you surely have your config in CVS
[19:40:21] <jensor> alex_joni what is cvs?
[19:42:52] <alex_joni> a versioning system
[19:43:00] <alex_joni> that tracks changes (especially to text files)
[19:44:20] <jensor> I think I can probably access the files in the trash bin I haven't checked yet. Really don't need backlash comp for what I am doing now
[19:44:42] <alex_joni> ah, ok.. but then you need to remove the reference from the ini too
[19:44:46] <alex_joni> at least comment it out
[19:45:32] <jensor> I can do that but I still need to somehow stop emc from thinking it is still running. I am about to shut down and reboot
[19:49:33] <alex_joni> http://www.youtube.com/watch?v=9EPfyID0nZ4
[19:51:30] <jensor> what does donovan-colours have to do with the subject?
[19:52:45] <jensor> Bye _ I'm going to reboot
[21:30:13] <JymmmEMC> alex_joni: did you see it?
[21:34:38] <alex_joni> JymmmEMC: shortly
[21:34:48] <alex_joni> but I was at work, didn't have time to pay attention
[21:35:02] <alex_joni> some electric car with good accel, didn't really cause a tingle :D
[22:17:06] <RayAcayan1> RayAcayan1 is now known as RayAcayan
[22:17:29] <alex_joni> now this is a great lullaby
[22:17:34] <alex_joni> http://www.youtube.com/watch?v=BgOxR6ZDB4Q
[22:17:52] <alex_joni> nothing better than singing pi to 53 decimals to your kid :D
[22:24:55] <alex_joni> fenn: you probably know that song ^
[22:26:47] <fenn> yep
[22:27:34] <fenn> it's aoi tada
[22:28:13] <alex_joni> yup
[22:28:22] <alex_joni> not much on her on youtube though
[22:29:32] <fenn> there are like 5 CD's on amazon.jp
[22:31:58] <fenn> hm maybe that was someone else
[22:32:14] <fenn> i guess this is all there is http://www.last.fm/music/Aoi+Tada
[22:32:55] <alex_joni> http://www.amazon.co.jp/Shirley-.../dp/B001714C3M/ref=sr_1_7?ie=UTF8&s=music&qid=1209545815&sr=1-7
[22:33:34] <fenn> yeah, that.
[22:46:58] <fenn> reprap is on-par with commercial RP machines now: http://1.bp.blogspot.com/_Cii6PZ78wNc/ScD8n-LvOhI/AAAAAAAAAFE/6-o4ezLUsuI/s1600-h/WineG_ABS.jpg
[22:47:01] <fenn> assuming someone can duplicate this person's methods
[22:47:05] <fenn> http://4.bp.blogspot.com/_Cii6PZ78wNc/ScD9RMyB_xI/AAAAAAAAAFU/TMP-8zS8yRQ/s1600-h/WineG_ABS10.jpg
[23:09:22] <hypa7ia> fenn: those are neat!
[23:09:45] <hypa7ia> what's the source?
[23:22:00] <fenn> hypa7ia: http://builders.reprap.org
[23:25:22] <hypa7ia> fenn: thanks!
[23:29:19] <alex_joni> seems fenn has a little brother called fennec ;)
[23:32:21] <fenn> does fennec work yet?
[23:32:53] <alex_joni> fenn: beta1 released
[23:34:02] <fenn> i'll try it when i get my beagleboard
[23:47:21] <alex_joni> no way.. this is crazy ;) http://www.theregister.co.uk/2009/03/19/keyboard_sniffing_demo/page2.html
[23:50:18] <fenn> also they can cross the beams and read your brain directly
[23:50:32] <fenn> but aluminum foil stops that
[23:51:08] <alex_joni> ha
[23:54:36] <alex_joni> somehow I really doubt you can scope the grid to see variations induced by keypresses
[23:55:00] <fenn> why not just use a telescope
[23:55:33] <fenn> if you can sneak a laser into whatever facility a camera should be doable
[23:56:25] <fenn> i dont really see a need for the laser though
[23:56:31] <fenn> could just use a microphone