#linuxcnc-devel | Logs for 2015-12-01

[08:54:14] <skunkworks> with move-off and reverse run.. the only thing missing is setting the tool lenght...
[08:54:38] <skunkworks> and if you can set the tool lenght acurately - it isn't a problem :)
[09:07:03] <skunkworks> seb_kuzminsky, running the wheezy livecd - I don't get any overruns at startup of the rtai.
[09:07:13] <skunkworks> on the same hardware
[09:15:49] <skunkworks> seb_kuzminsky, http://pastebin.com/qUf44tFg
[10:10:04] <seb_kuzminsky> skunkworks: weird
[10:11:41] <seb_kuzminsky> that's all with the linuxcnc latency-test program, right? can you check it with the rtai latency test, /usr/realtime-3.16.0-9-rtai-686-pae/testsuite/kern/latency
[10:13:34] <skunkworks> that is the rtai latency test.
[10:13:45] <skunkworks> ./run in the latency directory
[10:16:09] <seb_kuzminsky> oh.
[10:16:09] <seb_kuzminsky> hm
[10:16:28] <seb_kuzminsky> ok, let's do this
[10:17:00] <seb_kuzminsky> run "sudo dmesg -c", then "./run", then "dmesg", and email the output of the "run" and the final "dmesg" to the rtai list
[10:17:13] <seb_kuzminsky> we'll see if we can get the rtai gurus' input on it
[10:17:40] <seb_kuzminsky> or you can send it to me and i'll forward it on to the rtai list if you don't want to sign up
[10:18:54] <seb_kuzminsky> or maybe as a first step, pastebin the run & dmesg and i'll take a look at it
[10:22:26] <skunkworks> ok
[10:26:23] <skunkworks_> seb_kuzminsky: http://pastebin.com/xCRqDFb1
[10:27:56] <skunkworks_> every once in a great while the overrun will be 0 when starting rtai
[10:27:56] <seb_kuzminsky> thx
[10:28:06] <skunkworks_> like maybe 1 out of 10
[10:29:10] <seb_kuzminsky> and on the wheezy 3.4-rtai kernel it always starts out with 0 overruns?
[10:29:37] <skunkworks_> yes
[10:29:57] <seb_kuzminsky> ok, i dont see anything goofy in the dmesg, i'm going to ping the rtai list about it
[10:30:11] <skunkworks_> thank you!
[10:37:58] <seb_kuzminsky> thanks for the testing
[10:38:13] <seb_kuzminsky> i'm embarrassed to admit i've not tested carefully on real hardware
[10:38:48] <seb_kuzminsky> i do all my development in virtual machines, and i have only booted that kernel on two real machines, and i haven't paid careful attention like you do
[10:40:27] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/ja9_updates e0d5994 06linuxcnc 10docs/man/man1/halui.1 10docs/man/man9/lineardeltakins.9 10src/emc/rs274ngc/interp_convert.cc 10src/emc/usr_intf/keystick.cc fix some compiler warnings (edit errors) * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=e0d5994
[10:40:27] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/ja9_updates 88a1003 06linuxcnc 10(11 files in 4 dirs) tklinuxcnc: sim configs update * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=88a1003
[10:40:27] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/ja9_updates 2195f83 06linuxcnc 10(6 files in 2 dirs) tklinuxcnc gui: update for joints_axes * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=2195f83
[10:43:10] <seb_kuzminsky> dgarr is a hero
[10:44:29] <skunkworks> did you see his video?
[10:45:32] <skunkworks> https://www.youtube.com/watch?v=TdvcHJeDoEI
[10:55:19] <seb_kuzminsky> wow, very cool
[10:55:38] <seb_kuzminsky> i wonder if incremental jogging works in axis, i noticed he went to hal to do it
[10:55:48] <seb_kuzminsky> in world-mode, i mean
[11:03:28] <cradek> world incremental jogging has worked in ja/axis for a long time
[11:03:37] <cradek> dgarr added world-mode wheel jogging
[11:03:54] <cradek> (which is the awesomest)
[11:04:13] <seb_kuzminsky> oh, right
[11:06:24] <cradek> that video shows some great testing
[11:06:28] <cradek> very sweet
[11:10:11] <skunkworks> http://www.machsupport.com/forum/index.php/topic,31039.msg217674.html#msg217674
[11:10:21] <skunkworks> I don't think the last poster quite understands the problem/
[11:10:51] <skunkworks> So - uvw and reverse run. Thoughts?
[11:37:03] <seb_kuzminsky> Todd Zuercher just reported (i think) that the uvw branch is not quite all the way ready yet
[11:38:36] <skunkworks> really>
[11:39:44] <skunkworks> last I see is rob tweeked the niave cam detector to do uvw and todd was hoping to test some more. It ran slight slower than just z because of it.
[11:41:45] <seb_kuzminsky> http://article.gmane.org/gmane.linux.distributions.emc.user/60934
[11:47:29] <skunkworks> I don't know if that is an issue... But at this point this many axis is getting out of my comfort zone.
[11:48:51] <skunkworks> you have 2 joints that are moving in a way that will make it appear like you are going faster than you actually are.. (2 heads (axis) vs 2 joints making up 1 axis)
[12:36:49] <skunkworks> I say we merge it all. joints/axis, uvw, reverse run!
[12:36:58] <skunkworks> let users sort it out ;)
[12:38:13] <archivist> that may confuse a few......
[12:39:12] <skunkworks> certainly.
[12:39:20] <jepler> skunkworks: better choose some other name for your own project which will never have a stable release version again
[12:39:33] <skunkworks> hey - rob had almost the same answer I had..
[12:39:49] <archivist> I can see how the css for rotaries can be a simple addition soon :)
[12:40:12] <skunkworks> jepler, have faith ;)
[12:41:17] <skunkworks> That's - I am forking. Tuxkit it is!
[12:42:23] <jepler> hmmmm
[12:42:41] <jepler> several tastes that have a different distinctive taste when combined
[12:49:33] <CaptHindsight> any thoughts on a Linuxcnc-fest for next summer? It will be 3 years since the last great fest in Wichita
[12:50:24] <seb_kuzminsky> there was one in texas fall 2014, it was really fun
[12:50:35] <seb_kuzminsky> i'd love to go to a fest soon again
[12:50:46] <archivist> I am waiting for the UK one :)
[12:54:24] <CaptHindsight> http://wiki.linuxcnc.org/cgi-bin/wiki.pl?LinuxCNC_Integrator_Meeting_2014_Germany
[12:55:09] <CaptHindsight> lotsa robots
[13:01:23] <seb_kuzminsky> that's cool
[13:41:00] <lair82> Hey guys, we resolved the problem with my rotary table fiasco, I didn't have the encoder.position signal connected to the pid.a.feedback pin. Kind of important, typo on my part, but it is running now, and I am in the middle of tuning it. So Thank you for all the help this last week,
[13:45:21] <cradek> yay
[15:04:24] <skunkworks_> http://imgur.com/XfjN0KQ running the uvw branch
[15:06:06] <seb_kuzminsky> wow that's great
[15:19:32] <skunkworks_> http://imgur.com/a/0jeQs
[15:20:03] <skunkworks> copying the moves to uvw
[18:50:42] <skunkworks> also - what are the thoughts on m98m99
[19:12:56] <skunkworks> (I think the new TP went really well)
[19:43:55] <jepler> skunkworks: I fear that our gcode interpreter is already hard to maintain. I worry that introducing two entirely different kinds of subroutines will make it worse out of proportion with how helpful it is for certain users with inflexible posts or catalogs of old part programs
[19:44:34] <jepler> that said, I haven't reviewed any m98/m99 patches so I am commenting on the general idea of adding features to the interpreter, not on any specific patch
[19:44:46] <jepler> (do I have the right m98/m99?)
[19:46:25] <jepler> I mean, introducing *second* entirely different kind of subroutine
[19:49:58] <skunkworks> that is what m98/m99 is - yes. The fanuc style of calling a subroutine.
[21:27:07] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/ja9_updates 7300842 06linuxcnc 10docs/man/man1/halui.1 10docs/man/man9/lineardeltakins.9 man page compiler warnings try fix * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=7300842
[21:27:07] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/ja9_updates 9e7e955 06linuxcnc 10(14 files in 3 dirs) sim configs: use [KINS]KINEMATICS from inifile * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=9e7e955
[21:42:09] <seb_kuzminsky> dgarr: if you push that stuff to the main joints_axes9 branch I'd love to test it on my goofy little robot arm