#emc-devel | Logs for 2008-06-16

[02:37:25] <dmwaters> {global notice} Good day all, I need to do some quick rerouting, this should not take long.
[15:03:27] <skunkworks_> come on now.. the how are we supposed to keep tabs on you guys?
[15:05:02] <BigJohnT> kinda dark
[15:05:44] <BigJohnT> I think they put a picture up in front of the camera
[15:06:13] <skunkworks_> that could be... maybe some terrorist have taken over..
[15:20:14] <alSMT> the twilight zone
[15:56:09] <skunkworks_> woo whoo - back up :)
[16:00:25] <jmkasunich> the webcam driver seems to be buggy, it causes intermittent crashes of my computer
[16:00:34] <skunkworks_> yeck
[16:00:43] <jmkasunich> the last crash was last night, and I didn't restart the cam (or IRC) then
[16:00:49] <jmkasunich> cam is running now tho
[16:04:43] <skunkworks_> cool
[16:13:53] <BigJohnT> I see you now cool
[17:18:29] <skunkworks_> jepler: how is the turnout? (compared to last year?)
[17:24:21] <jepler> skunkworks_: sparser so far I think
[17:24:49] <skunkworks_> ok. Thank
[17:25:26] <skunkworks_> s
[18:04:32] <tom1> bbl
[19:07:46] <skunkworks_> jepler: what is up on your screen?
[19:08:17] <skunkworks_> looks like your machining something out of a solid.
[19:10:52] <skunkworks_> or was it a screen saver? ;)
[19:12:21] <jepler> skunkworks_: jmk made a raytraced preview of the thing they're machining on max5. the image you saw might have been something like this: http://emergent.unpy.net/files/sandbox/output0922.png
[19:12:27] <jepler> halfway milled
[19:12:45] <jepler> in that image
[19:12:59] <skunkworks_> aw - that is cool.
[19:14:31] <skunkworks_> Yes - that is what I think I was seeing
[19:19:15] <jmkasunich> he's making a movie of that rendering
[19:19:23] <jmkasunich> takes much longer than actually cutting the part
[19:19:40] <jmkasunich> (many hours)
[19:19:44] <skunkworks_> yikes.
[19:20:08] <jmkasunich> I hope to make a movie of cutting the actual part (out of wood) and then combine them
[19:20:17] <jmkasunich> coming soon to a youtube near you
[19:20:36] <skunkworks_> that would be awesome - can't wait to see it in person too. Just don't use up all the wood ;)
[19:20:50] <skunkworks_> we might have to bring our own ;)
[19:21:53] <skunkworks_> jmkasunich: are you cutting the sphire with the edge of the mill? (the flutes - not the end)
[19:45:43] <jmkasunich> the finishing cut is made with the side of the mill
[19:46:15] <jmkasunich> speaking of wood - we're using pieces of 2x2, about 3" long
[19:57:03] <jepler> temporary location of incomplete video: http://linuxcnc.org/c_f/5-axis-milling-simulation.avi
[19:57:25] <jepler> it does the first third, then jumps ahead to near the end, then quits abruptly
[19:59:28] <awallin> nice, what tools did you use to produce it?
[20:00:29] <jepler> awallin: jmkasunich has a special-purpose C program that outputs a .ngc file and a .pov file.
[20:00:48] <jepler> the pov file is basically a CSG difference between the original material and each successive tool position in the program
[20:01:20] <jepler> I modified it a bit so that each frame of the animation is the addition of one new cylinder to the csg
[20:01:42] <awallin> ok, I've been running the vpython demo programs and that can also do animations and simulations quite nicely
[20:01:47] <jepler> the later frames are progressively slower to render -- I think that there are ultimately 1200 frames, so the main solid is a CSG of 1200 elements.
[20:02:05] <awallin> is this the same C-code that is used for a simulated machine view with emc?
[20:02:13] <jepler> no, unconnected
[20:03:16] <awallin> how are joint positions relayed to the simulated machine from emc?
[20:03:39] <awallin> I could try it in vpython if the communication to emc is straightforward to set up
[20:03:49] <jepler> awallin: this current animation doesn't use emc at all -- the C program directly writes .pof
[20:03:52] <jepler> er, .pov
[20:04:18] <jepler> for the machine visualization stuff, the visualization program is a hal component
[20:04:39] <jepler> so in effect it is polling the position of each joint before each redraw
[20:04:55] <jepler> if you want it "live", that's a good way to go
[20:05:35] <awallin> and the visualization component is hooked up to axis.N.xxx in normal HAL?
[20:05:41] <jepler> if you want to record it and do something offline, halsampler is the way to go -- you can sample each joint position and whatever else you want that is in hal (like tool length or commanded spindle speed) and each servo cycle produces a sample in an output text file
[20:05:45] <jepler> yes
[20:06:10] <awallin> does the visualization component need to be real-time? or could it be done in python?
[20:06:47] <jepler> this older visualization is done in python: http://axis.unpy.net/01196094602
[20:06:56] <jepler> python can be fine for this kind of thing
[20:07:48] <jepler> look at the program pumagui / scaragui installed by emc 2.2 and later, that's just python source
[20:08:21] <jepler> errr
[20:08:52] <jepler> they should be, anyway
[20:10:17] <awallin> ok, I'll take a look, I'm mostly interested in toolpath visualization and cutting simulation for CAM, but connecting it to emc would be nice too
[20:55:10] <CIA-33> EMC: 03jmkasunich 07TRUNK * 10emc2/src/emc/usr_intf/axis/scripts/maho600gui.py: added a vismach model of a Maho 600C mill with 2-axis rotary
[21:06:51] <tomp> so the csg object is a boatload (at completion) of invisible cylinders subtracted from the initial workpiece?
[21:07:24] <jmkasunich> yes
[21:07:33] <tomp> way cool
[21:08:08] <tomp> i'd like to hack it for a formed tool to simulate sink edm ( please :)
[21:09:41] <jmkasunich> I will post the code on my blog when I get home
[21:10:05] <jmkasunich> (my blog is hosted on my computer, which is here - so my site is offline)
[21:10:18] <tomp> np: i'm sure you're buried. hope Ray's classes are goin ok
[21:40:57] <SWPLinux> hello?
[21:42:38] <jepler> hi SWPLinux
[21:56:47] <SWPLinux> oh good, the intarweb is more or less working
[22:07:04] <jepler> yeah once everyone cleared out it got fast again
[22:07:06] <jepler> ponder that
[22:11:50] <skunkworks> is everyone downloading emc iso's? ;)
[22:12:26] <jepler> I hope not
[22:13:40] <skunkworks> maybe they are too that they bought mach but want to try emc..
[22:13:46] <skunkworks> oops
[22:13:54] <skunkworks> maybe they are too embarrassed that they bought mach but want to try emc..
[23:01:44] <jmkasunich> standin' around watching coolant flow
[23:01:49] <jmkasunich> almost as exciting as golf
[23:04:45] <skunkworks> Heh - I hear you
[23:06:00] <skunkworks> have you guys started machining the servo mounts?
[23:06:22] <skunkworks> assuming that is what I understood was one of the to-do's
[23:10:40] <jmkasunich> not yet
[23:11:18] <jmkasunich> getting rid of the old coolant and flushing the sump is one step in that direction
[23:11:32] <jmkasunich> a workholding fixture and tooling are two other steps, in process
[23:11:44] <jmkasunich> we probably won't make parts till weds or so
[23:15:24] <fenn> lean manufacturing
[23:17:12] <skunkworks> heh
[23:17:16] <skunkworks> Hi fenn