#linuxcnc-devel | Logs for 2016-09-26

[00:08:58] <cradek> > error getting parameter 5400
[00:09:03] <cradek> these test failures are super weird
[00:19:22] <linuxcnc-build_> build #4532 of 0000.checkin is complete: Failure [4failed] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/0000.checkin/builds/4532 blamelist: John Morris <john@zultron.com>, Chris Radek <chris@timeguy.com>
[00:33:31] <linuxcnc-build_> build #4529 of 1201.rip-lucid-rtai-i386 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1201.rip-lucid-rtai-i386/builds/4529 blamelist: Robert W. Ellenberg <rwe24g@gmail.com>
[00:36:37] <linuxcnc-build_> build #3728 of 1301.rip-precise-rtai-i386 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1301.rip-precise-rtai-i386/builds/3728 blamelist: Robert W. Ellenberg <rwe24g@gmail.com>
[00:37:50] <linuxcnc-build_> build #2343 of 1401.rip-wheezy-rtai-i386 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1401.rip-wheezy-rtai-i386/builds/2343 blamelist: Robert W. Ellenberg <rwe24g@gmail.com>
[01:15:58] <linuxcnc-build> build #4533 of 0000.checkin is complete: Failure [4failed] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/0000.checkin/builds/4533 blamelist: Robert W. Ellenberg <rwe24g@gmail.com>
[07:01:50] <skunkworks> zlog
[07:29:06] <jepler> These are probably because min and max are macros in kernelspace: /home/buildslave/emc2-buildbot/wheezy-rtai-i386/rip-wheezy-rtai-i386/build/src/emc/tp/math_util.h:4:13: error: expected identifier or ‘(’ before ‘{’ token
[07:29:10] <jepler> some of those other tests sure look bogus
[09:09:35] <skunkworks> I really need to look at the DRO issues and start a bug report
[09:11:34] <skunkworks> I saw this weekend - Homed the K&T - > Ran the machine to x0y0 (which was previously set). The DRO seemed to display machine coordinates. (x and y wasn't zero and the active gcodes displayed G54)
[09:12:18] <skunkworks> The second I touched off Z - the axis dro's displayed 0 for x and y and .5 for Z (what I touched off to)
[09:13:32] <skunkworks> This is 2.7.6 - but I don't see anything in the change log of 2.7.7 that may address that.
[09:25:09] <cradek_> > Interp: fix startup regression regarding coordinate systems and more
[09:25:28] <cradek_> skunkworks: this 2.7.7 change fixes a lot of weirdness about offsets disappearing
[09:25:42] <skunkworks> ah - ok
[09:26:44] <cradek> 2.7.6 is really quite broken :-(
[09:28:54] <skunkworks> I have been making jig to press a 2 cycle engine crank apart.. Surprisingly a lot harder to press than you would think.. started turning a bearing puller inside out in the press.
[09:29:45] <skunkworks> required 2 jigs out of 1 inch thick steel. My ears rang a bit after the last one let loose.
[09:30:39] <cradek> what's pressed together on a 2 cycle crankshaft?
[09:30:59] <skunkworks> everything
[09:31:07] <archivist> the crank pin into the webs
[09:31:22] <skunkworks> let me find a picture
[09:31:26] <cradek> cool
[09:31:31] <cradek> I had no idea
[09:31:44] <archivist> us old bikers :)
[09:32:10] <skunkworks> The lightbulb went on as a test bed for stirling engines because I can press it back together 90deg instead of the 180deg it is normally set to
[09:32:15] <archivist> fun part is getting them inline after pressing back together
[09:32:21] <skunkworks> right
[09:33:12] <cradek> so it was a 2 cylinder 2 stroke?
[09:33:21] <skunkworks> if you look at this - notice ball bearings in the center and the connecting rods don't have end caps.
[09:33:22] <skunkworks> http://www.ebay.com/itm/Ski-Doo-MXZ-Adrenaline-Mach-Z-1000-Crankshaft-Crank-Twin-Snowmobile-Engine-/371551100684
[09:33:32] <skunkworks> cradek, yes
[09:33:48] <cradek> funky
[09:34:31] <cradek> so much mass it's a built-in flywheel
[09:34:52] <archivist> also seen in 4 smoke engines
[09:35:12] <skunkworks> I have not seen that - but I don't do motorcycles and such
[09:35:53] <skunkworks> youtube videos just shows you hammering the crank to 'teak' the alignment ;)
[09:36:20] <cradek> well to be honest, I'd probably try that first too
[09:37:06] <skunkworks> this engine - the 'lobes' are 100mm diameter - I should be able to make a sleeve to line things up pretty close
[09:38:17] <archivist> or the mass of the crank flywheel half against an immovable object
[09:40:03] <archivist> anyway, alignment within a thou is your aim
[09:51:30] <skunkworks> sure
[09:52:12] <skunkworks> the only issue is the bearings I need seem to only be in the netherlands... K24X30X13
[10:52:31] <archivist> skunkworks, are you making a high pressure stirling?
[10:52:51] <skunkworks> yes
[10:53:12] <skunkworks> high as in maybe 100psi or a bit more
[10:53:48] <skunkworks> probably helium. (I can't bring my self to use hydrogen)
[10:54:49] <archivist> 100atm for lols :)
[10:55:26] <archivist> I used to read the hot air engine competition articles in model engineer
[11:33:41] <seb_kuzminsky> wow you folks have been busy this weekend, awesome
[11:35:56] <jepler> hi seb_kuzminsky
[12:30:26] <seb_kuzminsky> i want to add NIST's document "The NIST RS274NGC Interpreter - Version 3" to our docs, maybe in the "Glossary, Copyright, History & Overview" section, any objections?
[13:20:00] <JT-Shop> sounds like a good idea to me, I'm sure many would like to know the history of EMC
[13:25:34] <cradek> you mean a link to it, right?
[13:26:11] <cradek> suggest a link here: http://www.linuxcnc.org/docs/html/gcode/rs274ngc.html
[13:43:21] <seb_kuzminsky> cradek: i meant we should have a copy on wlo, and link to that one from our docs
[13:43:37] <cradek> oh sure
[13:44:46] <seb_kuzminsky> the place you just suggested is probably a better place than down in History
[13:46:39] <jepler> somebody with a tuit should make a new Showcase entry..
[14:37:47] <Tom_L> could somebody with ops please do something about dioz in the main channel
[14:38:47] <mozmck> zlog
[15:06:15] <skunkworks> has anyone made a probing routine that probes around a profile?
[15:06:46] <cradek> I did over (smartprobe.ngc), but not around
[15:06:57] <skunkworks> right - I have played with that one
[15:27:20] <JT-Shop> like a smart probe around or a defined size?
[15:27:51] <skunkworks> right
[15:27:58] <skunkworks> sorry
[15:28:35] <skunkworks> an example would be a representation of a bell housing. The perimeter that mounts to say the engine.
[15:29:04] <skunkworks> so it would just start walking around a shape
[15:29:26] <skunkworks> If I cannot find something - I could try to write one I guess
[15:29:52] <JT-Shop> so if you wanted 360 measurements around an object you rotate the coordinate system after each probe?
[15:30:09] <skunkworks> hmm - that would be one way...
[15:30:24] <skunkworks> I was thinking something similar to smart probe logic though
[15:30:32] <JT-Shop> and have a backoff that will clear the largest change
[15:31:03] <skunkworks> that might be a cheap and dirty way to try it though.
[15:31:07] <JT-Shop> just thinking out loud while putting this part through my CAM software
[15:31:26] <skunkworks> Thank you - that is a start
[15:31:29] <skunkworks> bbl
[16:44:18] <cradek> if the shape is convexish, you could just circumscribe it and step around and probe toward the center. it'd work ok-ish and be very easy to write
[17:26:57] <cradek> personally I wouldn't rotate the coordinate system, because I'd be afraid of finding bugs I didn't want to know about
[17:29:04] <JT-Shop> hmm, the documents are not specific about what coordinates are stored in 5061-5069
[17:31:11] <cradek> I think "controlled point" is defined ... somewhere
[17:32:13] <cradek> http://linuxcnc.org/docs/html/gcode/machining-center.html#_controlled_point
[17:32:20] <cradek> not specific enough
[17:32:21] * JT-Shop likes it now you can search the whole pdf at once
[17:32:30] <cradek> I guess I agree the docs don't say
[17:32:52] <cradek> I think the intent of the "controlled point" language is that it's including all offsets
[17:33:11] <JT-Shop> that makes sense
[17:33:22] <cradek> it would be really strange otherwise
[17:33:56] <cradek> as it is, you get some values, and you can use those immediately in the next probe moves/gcodes
[17:34:13] <cradek> can't really imagine how it would work otherwise
[17:35:36] <cradek> on the other hand lots of people have requested g53 g38.x probes which would do the other thing
[17:47:46] <mozmck> g53 g38.x probe would be very nice.
[17:49:50] <JT-Shop> the controlled point is not very clear... and I think using "controlled point" in the G38 description is confusing
[17:51:01] <JT-Shop> needs to the location of the axes including any offsets... or something like that
[17:52:29] <Tom_L> (probe tip)
[17:53:33] <JT-Shop> yea that would be the location of the axes however a note could be added to remind to subtract the radius of the tip
[17:53:53] <Tom_L> (which should be at the center of the probe ball)
[17:54:11] <JT-Shop> should be if your ball is centered
[17:55:08] <JT-Shop> the cool thing is I was able to search the entire PDF at one time for a search term...
[17:55:42] <cradek> you might set the controlled point to be the tip (bottom) of the ball, or the center of the ball, depending on your offsets
[17:55:58] <cradek> I think it's wrong for the docs to assume one kind of setup
[17:56:07] <JT-Shop> good point
[17:56:38] <Tom_L> even when you use a ball in space on a setup you probe for the center of the ball
[17:56:48] <Tom_L> not sure i agree completely
[17:56:56] * JT-Shop tries to visualize how that might work
[17:57:04] <cradek> I can imagine other setups. some probes even have more than one ball
[17:57:15] <Tom_L> especially on a 5 axis etc
[17:57:30] <Tom_L> right, that's why to use the center
[17:57:31] <mozmck> Our probe is the tip of a plasma torch
[17:57:42] <mozmck> So "center" would be meaningless
[17:57:44] <Tom_L> because the center will be the center always, the probe length could vary
[17:58:18] <cradek> maybe instead of removing talk about controlled point from the probe docs, defining controlled point better is the way to go
[17:58:39] <cradek> and linking to it where appropriate?
[17:58:44] <cradek> but I defer to the docs guy :-)
[17:58:51] <JT-Shop> I was not thinking of removing but noting that it includes any offsets
[17:59:05] <JT-Shop> and a link to control point is good too
[17:59:20] * JT-Shop tries to remember till the morning...
[23:54:06] <KGB-linuxcnc> 03Dewey Garrett 05master a055cb1 06linuxcnc 10src/emc/usr_intf/pncconf/build_INI.py pncconf build_INI.py bugfix for JA handling lathe * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=a055cb1