#emc-devel | Logs for 2008-06-23

[01:48:03] <jepler> hum, if the servo thread dies but the base thread lives on, and you use doublestep, you could get the behavior Andrew Ayre describes.
[01:52:07] <jepler> hi SWPLinux
[01:52:26] <SWPLinux> hi jepler
[01:53:16] <SWPLinux> kinda funny - I saw a commercial for HGR at the restaurant during dinner
[01:53:30] <jepler> did you decide whether to go?
[01:53:40] <SWPLinux> it didn't seem like the kind of place that would advertise during prime time
[01:53:52] <SWPLinux> I don't think I will - I'd have to backtrack 35 miles or so
[01:54:02] <jepler> yeah I agree
[01:54:15] <SWPLinux> then there's the full Jeep ...
[01:59:47] <SWPLinux> anyone heard from jmk? he never passed me on the highway
[02:00:10] <jepler> not me
[02:00:45] <SWPLinux> did you guys notice that I pulled onto I-74 right in front of you?
[02:01:00] <SWPLinux> then you passed me :)
[02:01:54] <jepler> yes
[02:02:18] <jepler> chris said "I think there's a guy following us"
[02:02:23] <SWPLinux> heh
[02:02:39] <SWPLinux> did you end up going east on Main street until you hit 74?
[02:02:44] <jepler> yes
[02:02:50] <SWPLinux> ah - ok
[02:03:04] <jepler> it's a couple miles
[02:03:08] <SWPLinux> I think the Seminary Street approach could be faster, unless there's a train like today
[02:04:12] <SWPLinux> ah - there he is
[02:04:18] <jepler> huh -- I just realized that while 2 threads of rendering with POV have been going on, my laptop fan was either off or on low -- but when I went to a web page with flash earlier, the fan came on high.
[02:04:22] <jepler> hi jmkasunich
[02:04:30] <jepler> so I guess java really is a hot language
[02:04:30] <jmkasunich> hi
[02:04:31] <SWPLinux> funny
[02:04:39] <jepler> er, not flash -- java
[02:04:49] <SWPLinux> either
[02:04:53] <jepler> maybe it also turned on all the extra transistors in my video card, or something
[02:05:03] <jepler> (though running a video fullscreen doesn't start the fan)
[02:05:12] <jepler> Render Time: 1 hours 3 minutes 55 seconds (3835 seconds)
[02:05:18] <SWPLinux> per frame?
[02:05:19] <jmkasunich> for one frame?
[02:05:20] <jepler> ugh that's a long time for one frame
[02:05:24] <SWPLinux> heh
[02:05:30] <jepler> (by 2008 standards at any rate)
[02:05:34] <cradek> hi everyone
[02:05:38] <SWPLinux> hi
[02:05:38] <jmkasunich> what happened? I thought it was doing a frame every few minutes
[02:05:43] <jmkasunich> hi cradek
[02:05:49] <jepler> jmkasunich: it gets slower as more cylinders are added
[02:05:58] <SWPLinux> maybe the workpiece is now translucent
[02:06:00] <jmkasunich> yeah, but 20x slower?
[02:06:16] <SWPLinux> I think it's O(n^2) or worse
[02:06:18] <cradek> I haven't found anything missing or extra yet!
[02:06:23] <jmkasunich> yay
[02:06:39] <jepler> also compared to the original renderings, i added textures (including making the "ground" reflective), enabled antialiasing, and increased the number of pixels twofold
[02:06:40] <jmkasunich> I just unpacked the computer - the rest can wait
[02:07:00] <jmkasunich> cpu cycles are a terrible thing to waste, eh?
[02:07:03] <SWPLinux> I only brought in the computer and coffee pot. the rest will have to wait
[02:07:15] <SWPLinux> best to unload at home :)
[02:07:17] <jmkasunich> well, you're only halfway home
[02:07:23] <SWPLinux> yep
[02:07:47] <SWPLinux> I expected you to pass me somewhere before Chicago, but didn't see you
[02:08:08] <jmkasunich> I was driving 62-65 mph
[02:08:16] <SWPLinux> ah, even slower than me
[02:08:32] <SWPLinux> I was at 65-ish, and managed 22MPG average
[02:08:36] <jmkasunich> the trip totalled something like 1150 miles, and I have 3/4 of the last tank I bought
[02:08:45] <jmkasunich> haven't calc'ed the mileage yet
[02:09:13] <SWPLinux> I think if I'd go 55-60, I'd get 26 or so
[02:09:27] <jmkasunich> at 55, you'd get rear-ended by a truck
[02:09:33] <SWPLinux> but I;d get rear-ended 17 times, so it might not be worth it
[02:09:35] <SWPLinux> rifght
[02:09:37] <SWPLinux> uh
[02:09:39] <SWPLinux> right
[02:10:42] <jmkasunich> wrt the "exceeds limits" bug report - I think I've seen that behavior here too
[02:10:45] <cradek> arg, this shell mill is 20mm, not 3/4
[02:11:00] <jmkasunich> some of my tool offsets are negative, since my ref pt is 4" past the spindle nose
[02:11:02] <SWPLinux> oh - that could be a bummer
[02:11:55] <jepler> cradek: what was the name of the page layout device that Samuel Clemens was said to have invested in?
[02:11:57] <SWPLinux> sigh. my cell phone refuses to charge
[02:12:25] <cradek> heck I wish I could remember.
[02:12:37] <SWPLinux> damn. they just mentioned it
[02:12:51] <jepler> ah, "page compositer"
[02:12:56] <SWPLinux> right!
[02:13:02] <SWPLinux> I knew it began with a P
[02:13:02] <jepler> or maybe "paige", based on the google hits
[02:13:28] <SWPLinux> was that on PHC or some other show today?
[02:13:35] <jepler> http://www.marktwainhouse.org/themuseum/archivist.shtml
[02:13:43] <SWPLinux> right - news
[02:13:44] <jepler> SWPLinux: all things considered
[02:13:49] <jepler> the show that passes for news
[02:13:56] <SWPLinux> considering
[02:13:59] <jepler> did you hear the terrible light story on ethanol in brazil?
[02:14:22] <SWPLinux> yeah
[02:14:34] <jepler> "ethanol: why can't we have it like brazil? well, for one thing, we use more gas in america than in brazil (but we'll quantify how much very poorly)"
[02:14:42] <SWPLinux> "Gas is expensive, Ethanol is cheaper, so people use it"
[02:14:50] <SWPLinux> 7 times as much!
[02:14:57] <SWPLinux> is that per capita?
[02:15:42] <jepler> beats the f*** out of me
[02:15:44] <SWPLinux> they also neglected to mention that the price of fuel in Brazil is about what it was in England a couple of years ago
[02:16:14] <SWPLinux> gasoline at least
[02:17:11] <jmkasunich> cradek: did you commit (or at least capture) the version of mmount.ngc that cuts to 0.550 instead of 0.600 with the 1/4" tool?
[02:17:52] <cradek> jepler has a copy of it
[02:18:18] <jmkasunich> my checkout has those changes, and I can commit them, but I don't know if any other changes were made after we copied it to the mazak
[02:18:49] <CIA-34> EMC: 03jepler 07TRUNK * 10emc2/nc_files/mmount.ngc: as milled at cnc workshop 2008
[02:19:03] <jmkasunich> that solves that, thanks jeff
[02:19:15] <cradek> heh, I should duplicate this: 5 digits marked DISPLAY with a knob to select N/F/X/Y/Z
[02:19:34] <jmkasunich> those panels?
[02:19:52] <cradek> yes
[02:20:05] <cradek> as far as I can tell, that was the UI
[02:20:10] <jepler> N/F?
[02:20:14] <jepler> oh, N/F
[02:20:15] <SWPLinux> NFC
[02:20:18] <jmkasunich> N = rpm I bet
[02:20:20] <jmkasunich> F = feed
[02:20:27] <jepler> N might be N-number
[02:20:33] <jepler> (program line number)
[02:20:34] <jmkasunich> true
[02:20:35] <SWPLinux> yeah, line number
[02:20:50] <jepler> cradek: uh oh
[02:20:51] <jepler> cvs [commit aborted]: could not open lock file `/cvs/emc2/configs/demo_mazak/,demo_mazak.hal,': Permission denied
[02:20:56] <cradek> I'm sure it was line no
[02:21:00] <cradek> jepler: anon?
[02:21:12] <jepler> oh -- oops, yes it probably is now
[02:21:21] <jepler> (it's the untarred config and it had a CVS directory in it!)
[02:21:48] <CIA-34> EMC: 03jepler 07TRUNK * 10emc2/configs/demo_mazak/ (5 files): configuration files updated at fest 2008, with adjustment to tool change height missing
[02:21:57] <jepler> we modified the tool change height after I grabbed those files
[02:22:05] <jepler> my recollection is that we added some number to Z
[02:22:11] <cradek> .010
[02:22:12] <jepler> but I wasn't going to guess
[02:22:17] <CIA-34> EMC: 03jmkasunich 07TRUNK * 10emc2/lib/python/vismach.py: fix box surface normals, add a triangle primitive, add a scale operation
[02:22:38] <jepler> cradek: if you're confident, will you please do it?
[02:23:15] <jepler> goodnight all
[02:23:21] <jmkasunich> goodnight jeff
[02:23:40] <jmkasunich> I think I'm gonna call it a night too - walk Mr. Dog, and then sleep
[02:23:45] <jmkasunich> work tomorrow :-(
[02:23:46] <cradek> can't right now, but sure
[02:23:52] <cradek> jmkasunich: ouch
[02:23:54] <jmkasunich> I can do it
[02:24:37] <jmkasunich> was -0.032, should be -0.022, right?
[02:24:46] <jmkasunich> (less negative)
[02:25:32] <CIA-34> EMC: 03jmkasunich 07TRUNK * 10emc2/configs/demo_mazak/demo_mazak.ini: tweak tool change height
[02:26:10] <jmkasunich> cradek: not really that ouchy - I've been home about 3 hours, and this is a relatively early bedtime for me
[02:26:46] <SWPLinux> I'm wondering how early I'd end up getting up if I go to bed now
[02:26:52] <SWPLinux> since it's a bit early for me too
[02:28:03] <cradek> "RESET - Dumps Everything CLEAR - Resets a fault"
[02:28:17] <cradek> (handwritten sticker next to the reset/clear switch)
[02:28:30] <SWPLinux> heh
[02:28:45] <SWPLinux> sounds like those were mislabeled originally
[02:30:05] <jmkasunich> oh, one last thing
[02:30:13] <jmkasunich> could somebody tell me if jmkasunich.com is working?
[02:30:19] <cradek> it is handwired. maybe this one was wired backward!
[02:31:15] <SWPLinux> "nothing to see here. well, maybe you could look at my blog"
[02:31:24] <jmkasunich> ok, thanks
[02:31:29] <SWPLinux> np
[02:31:34] <jmkasunich> (that means the dyndns got updated and the server started)
[02:31:43] <SWPLinux> ah, ok
[02:36:20] <SWPLinux> well, it's bedtime. night guys
[04:06:03] <tomp> jepler: if of interest, i now have eeepc virtual screen at 1024x768 with hardy.
[04:06:11] <tomp> ´╗┐followed notes written for gutsy at http://forum.eeeuser.com/viewtopic.php?id=8242&p=2.
[04:06:12] <tomp> ´╗┐ basicly is installing intel 915 driver, edit xorg.conf and and "xrandr --size 1024x768"
[04:10:41] <cradek> tomp: sometimes lines you type have \ufeff at the beginning of them. maybe you have a funny setting somewhere?
[04:13:39] <tomp> ? like unicode? sorry, but no clue how to fix
[04:14:56] <tomp> maybe from cut & paste? i was trying not to type my usual paragraphs.
[04:16:26] <tomp> tomp considers having a funny setting, imagines its a knob turned way up :)
[04:21:09] <cradek> does seem like something to do with unicode but I have no idea either.
[04:21:17] <cradek> no big deal.
[04:33:27] <tomp> well, gnite
[12:28:59] <skunkworks> did everyone make it home ok?
[12:30:46] <jepler> skunkworks: since leaving I've heard from cradek, jmkasunich, jon elson, swpadnos (though he's not all the way home yet)
[12:42:55] <skunkworks> Cool. Fit performed flawlessly?
[12:43:13] <jepler> yep
[12:43:26] <jepler> slightly lower fuel economy
[12:43:29] <jepler> on the way home
[12:44:20] <skunkworks> I bet :)
[12:45:10] <jepler> more time on the interstate (higher speed limits) probably contributed too
[12:46:12] <jepler> breakfast time bbl
[13:45:25] <skunkworks> Did you know the help file for the puma robot config says for a 'puma 560'
[13:45:27] <skunkworks> :)
[14:49:26] <tomp> jepler: i cant seem to reply to bug report #1966633, regarding sound loss on ubu8.04/emc2.
[14:49:26] <tomp> Bug seems closed already,
[14:49:26] <tomp> Alex's suggested lib fixed it.
[14:49:26] <tomp> I reported to Alex and not the bug report system, sorry.
[19:23:19] <CIA-26> EMC: 03seb 07TRUNK * 10emc2/debian/emc2.files.in: Finish removing the old removed "blocks" component.
[19:26:31] <CIA-26> EMC: 03seb 07TRUNK * 10emc2/debian/ (emc2-dev.files emc2.files.in): more packaging fixes
[19:46:09] <skunkworks> so - when you home the puma config.. Then move to world coordinates.. If I do x,y or z moves - they are not linear.
[19:46:23] <skunkworks> the are some sort of arc
[19:46:51] <skunkworks> (trying to wrap my head around it all) OUCH
[20:00:53] <cradek> seems like the geometry is slightly wrong then
[20:01:16] <cradek> somewhere there are some lengths that you need to fix so they match the model
[20:02:51] <skunkworks> ok
[20:02:55] <skunkworks> I will dig into it more.
[20:03:04] <skunkworks> Thanks
[20:03:25] <skunkworks> just making sure the moves should be strait ;)
[20:04:00] <cradek> sure seems like it to me
[20:04:25] <skunkworks> The manual of the robot has most of the joint lengths also. The only one that doesn't seem to be there is the distance between joint 3 and 4.
[20:05:09] <cradek> looks like pumakins.A2, A3, D3, D4 are the hal params you should set
[20:05:13] <skunkworks> probably because the manual is for both the 550 and 560 - the 550 doesn't have joint 4
[20:05:33] <cradek> #define DEFAULT_PUMA560_A2 300.0
[20:05:33] <cradek> #define DEFAULT_PUMA560_A3 50.0
[20:05:33] <cradek> #define DEFAULT_PUMA560_D3 70.0
[20:05:33] <cradek> #define DEFAULT_PUMA560_D4 400.0
[20:06:04] <skunkworks> maybe the model doesn't match the joint lenghts?
[20:06:15] <cradek> yes, sounds like
[20:06:22] <skunkworks> ok - I will play with it.
[20:08:10] <cradek> it would be neat if you could fix the model to match your real robot.
[20:08:20] <cradek> that would help if others want to do any work on it.
[20:10:42] <skunkworks> I will see what I can do :P
[21:38:43] <CIA-26> EMC: 03seb 07TRUNK * 10emc2/src/Makefile:
[21:38:43] <CIA-26> EMC: Install the hostmot2 firmware files in /usr/share/emc/firmware as part
[21:38:43] <CIA-26> EMC: of the emc2 debian package.
[21:38:44] <CIA-26> EMC: 03seb 07TRUNK * 10emc2/debian/emc2.files.in:
[21:38:44] <CIA-26> EMC: Install the hostmot2 firmware files in /usr/share/emc/firmware as part
[21:38:46] <CIA-26> EMC: of the emc2 debian package.
[21:45:40] <micges> gcode properties doesn't work on trunk
[22:06:09] <CIA-26> EMC: 03jepler 07TRUNK * 10emc2/src/emc/usr_intf/axis/scripts/axis.py: fix exception showing properties dialog: ValueError: too many values to unpack
[22:33:24] <CIA-26> EMC: 03jepler 07TRUNK * 10emc2/share/axis/tcl/axis.tcl: fix joint/axis switch criteria -- again