#emc-devel | Logs for 2010-01-20

[00:43:09] <jepler> seb: autoconf is a new requirement to build master, so it's not in the build-dep of 2.3...
[04:36:47] <cradek> CIA-5: knock knock!
[04:37:32] <SWPadnos> I wonder if there's something weird with the new irc server
[04:37:53] <SWPadnos> I did notice that I couldn't say anything until I had identified, I think that's a change from before
[04:38:22] <cradek> that's because we have channel mode +R set
[04:38:28] <cradek> (because of that url spam thingy)
[04:39:53] <SWPadnos> ok, so it's probably no coincidence that there have been no CIA messages after that mode setting
[04:40:02] <cradek> oh, well duh
[04:40:16] <cradek> CIA-5: hellooo parrot!
[04:40:28] <cradek> hm.
[04:40:36] <SWPadnos> * SWPadnos pets CIA-5
[04:40:45] <SWPadnos> (what were those again? )
[04:40:55] <SWPadnos> CIA-5, help
[04:41:02] <SWPadnos> CIA-5: help
[04:41:04] <SWPadnos> eh
[04:41:19] <cradek> maybe he's mad at us for gagging him
[04:41:40] <cradek> wonder if I should kick him and have him re-join
[04:41:52] <jmkasunich> give him the boot!
[04:42:02] <cradek> I don't want to scare him though
[04:42:12] <cradek> (wonder what makes him come back)
[04:42:24] <jmkasunich> magic?
[04:42:32] <cradek> cradek has kicked CIA-5 from #emc-devel
[04:42:47] <jmkasunich> see? magic
[04:42:50] <cradek> CIA-5: magic
[04:42:58] <SWPadnos> * SWPadnos kicks CIA-5
[04:42:58] <CIA-5> ow
[04:43:01] <SWPadnos> there :)
[04:43:09] <cradek> CIA-5: we missed you
[04:43:25] <jmkasunich> somebody commit something to the sandbox
[04:43:26] <SWPadnos> CIA is not ELIZA I guess
[04:43:36] <jmkasunich> (or don't we have a sandbox anymore?)
[04:44:04] <cradek> too bad I just pushed all mine
[04:44:18] <cradek> look what I did to spiral.ngc :-)
[04:44:59] <jmkasunich> grr - I just downloaded half an iso, then it failed, and when I restarted it started from zero
[04:45:16] <cradek> do you still have the half? maybe wget will do it right.
[04:46:03] <jmkasunich> yeah, there is a *.part file in the directory, with 288M of stuff in it
[04:46:23] <cradek> try copying it to the real name and using wget
[04:47:22] <jmkasunich> I'm looking at this page: http://www.mythbuntu.org/downloads
[04:47:31] <cradek> no, it won't work, according to the manpage
[04:47:39] <jmkasunich> and trying to figure out what file that button actually downloads
[04:48:18] <cradek> you know what's really cool and they should try? hypertext
[04:48:30] <cradek> it's text, and when you click on it, it points to something
[04:48:50] <jmkasunich> that makes way too much sense
[04:49:16] <cradek> http://cdimage.ubuntu.com/mythbuntu/releases/9.10/release/
[04:52:03] <cradek> I think buildbot is still having a bad day
[04:58:51] <jmkasunich> this evening was "interesting".. I went to the computer store and got all the parts for my new computer (woot!), but when I turned onto my street on the way home it was dark - power failure - so I got to sit in the dark surrounded by all my new toys but unable to do anything with them
[05:00:32] <cradek> wow
[05:01:21] <jmkasunich> fortunately it was only down for about 45 mins - but that was longer than the UPS lasted, so uptime is back to zero :-(
[05:01:46] <cradek> I might be weird but I think power outages are fun
[05:02:04] <cradek> I guess that must mean they're very infrequent around here...
[05:09:30] <jmkasunich> they're more fun when it isn't after dark in the winter
[05:11:06] <cradek> arg, I need to get more than 5 hours of sleep tonight
[05:11:14] <cradek> days are too short!
[05:11:17] <cradek> goodnight.
[05:12:34] <jmkasunich> I have that same problem...
[05:12:36] <jmkasunich> goodnight
[07:00:51] <CIA-5> EMC: 03seb 07master * rb369fe35da07 10/src/ (Makefile hal/drivers/mesa-hostmot2/backported-strings.c): supply our own versions of some modern str functions if needed
[07:13:49] <alex_joni> * alex_joni eats CIA-5
[07:13:50] <CIA-5> * CIA-5 tastes crunchy
[07:14:38] <CIA-5> EMC: 03seb 07master * r2545641ce0ff 10/src/hal/drivers/mesa-hostmot2/hostmot2.h: declare the backported string functions, if needed
[07:57:45] <alex_joni> CIA-5: lol, the spiral looks really nice now :D
[07:57:51] <alex_joni> err.. cradek: ^^
[07:57:57] <alex_joni> (stupid autocomplete)
[07:59:16] <micges_work> hi alex_joni
[08:10:44] <alex_joni> hey micges_work
[08:10:48] <alex_joni> how goes it?
[08:13:55] <micges_work> good
[08:14:13] <alex_joni> nice to hear that
[08:14:16] <micges_work> I'm waiting for gantry machine for tests
[08:14:40] <micges_work> and I'm workkiing now on eagle script for making whole hal config
[08:15:16] <micges_work> beside that we have 30-40cm of snow in whole country
[08:15:56] <alex_joni> heh, sounds pretty white
[08:16:03] <micges_work> indeed
[08:16:15] <alex_joni> it's similar in the rest of the country here, but where I am it's still without snow
[08:17:55] <micges_work> transilvania ;)
[08:26:32] <micges_work> some wise guy found problem in gantrykins in ja3 that is related to servo-thread speed, can't wait to see it myself on real machine
[11:41:17] <alex_joni> micges_work: cool
[13:01:02] <Dave911_> Dave911_ is now known as Dave911
[13:18:35] <jthornton> If you are in G91 and have several G1 @.5 ^45 moves is it suposed to form an ever widening spiral?
[13:20:46] <jt-dev> http://imagebin.ca/view/zPO70al3.html
[13:29:49] <jepler> jthornton: I think so, but this is different than the initial version of @^ moves.
[13:34:27] <jthornton> jepler: thanks, I see what you mean now
[13:34:50] <jthornton> * jthornton heads out to work now
[13:38:45] <skunkworks_> skunkworks_ is now known as skunkworks
[13:55:28] <alex_joni> jthornton: yup
[13:55:52] <alex_joni> jthornton: look at spiral.ngc
[14:10:05] <jepler> seb: any concerns about taking out +#warning "kernel is so old it lacks argv_split(), supplying our own..."
[15:05:43] <cradek> JT-Work: polar coords no longer work how I said they would in the list emails :-/
[15:05:58] <cradek> JT-Work: I think they're better now - it was half-assed before
[15:06:16] <JT-Work> LOL ok
[15:06:27] <cradek> the G91 @.5 ^45 you wrote means just what it says: increase radius by .5 and increase angle by 45
[15:06:55] <JT-Work> ah ha that makes sense now
[15:06:55] <cradek> just how G91 X1 Y2 means increase X by 1 and Y by 2
[15:07:27] <JT-Work> it was too early in the morning for thinking I guess :)
[15:07:58] <SWPadnos> so @^ coordinates are always from the current origin?
[15:07:59] <cradek> so to drill along an angled line: G91 G81 ... @1, or to drill along a circle: G91 G81 ... ^30
[15:08:17] <SWPadnos> ie, if I move to X=5 Y=0, it'll be R=5, theta=0
[15:08:19] <cradek> SWPadnos: yes they have the same origin as the XY system
[15:08:26] <cradek> yes
[15:08:54] <SWPadnos> cool
[15:09:03] <cradek> (IMO anything else would be insanity)
[15:09:15] <SWPadnos> yes, I was approaching that insanity, which is why I asked
[15:09:43] <cradek> since we already have plenty of good ways of establishing origins...
[15:09:47] <SWPadnos> I was trying to think about how you'd manage interleaved relative XY and polar moves
[15:09:52] <SWPadnos> yep
[15:09:59] <cradek> what do you mean interleaved?
[15:10:12] <SWPadnos> G91 XY / G91 @^ / G91 XY ...
[15:10:18] <cradek> oh that just works
[15:10:33] <cradek> every point has a particular X,Y,@,^
[15:10:43] <SWPadnos> yeah - if the two sets of coords were separate, it would be mich more insane
[15:10:44] <cradek> you can alter any of them you want, individually
[15:11:17] <SWPadnos> hmmm. can you specify a coordinate system offset in terms of angle and radius?
[15:11:20] <cradek> G0 @1 ^45 (.707,.707) X1 (1,.707)
[15:11:34] <cradek> SWPadnos: hm, I'm not sure
[15:11:48] <cradek> (or G92)
[15:11:52] <SWPadnos> right
[15:12:04] <cradek> please test and get back to us :-)
[15:12:07] <SWPadnos> heh
[15:12:09] <SWPadnos> damn
[15:12:42] <cradek> what's the date when I have to quit adding halfbaked features?
[15:14:31] <jepler> cradek: end of the month, I think
[15:14:34] <cradek> G10 L2 P1 @1 ^0 doesn't seem to do anything...
[15:15:23] <cradek> G92 @1 ^0 is an error
[15:16:00] <cradek> looks like I need to add some error messages
[15:16:18] <cradek> I don't think G10/G92 need to support polar
[15:16:33] <cradek> you can always G0 @1 ^180; G92 X0 Y0
[15:16:58] <cradek> brb
[15:51:56] <CIA-5> EMC: 03cradek 07master * r1f8e3a1f476d 10/src/emc/rs274ngc/ (interp_convert.cc interp_internal.cc): allow polar coordinates for probing, disallow for setup
[16:32:10] <seb_kuzminsky> goodmorning
[16:33:16] <cradek> hi seb
[16:34:45] <jepler> hi seb
[16:35:10] <seb_kuzminsky> hey guys :-)
[16:36:42] <jepler> seb: any concerns about taking out +#warning "kernel is so old it lacks argv_split(), supplying our own..."
[16:38:17] <seb_kuzminsky> no that should probably come out
[16:38:24] <seb_kuzminsky> hey wait! you said you *like* warnings!
[16:40:04] <jepler> wouldn't be the first time I've contradicted myself
[16:41:02] <CIA-5> EMC: 03seb 07master * r774a33ffeb3e 10/src/hal/drivers/mesa-hostmot2/backported-strings.c: remove a superfluous warning
[20:39:24] <skunkworks_> http://www.usbcnc.com/index_products.html#swv3
[20:39:46] <skunkworks_> scary. lets just power through corners...
[20:43:54] <cradek> the boss8 has that option G99. if you add it to a line, the control won't bother decelerating between that line and the next (yes you have to put it on every applicable line)
[20:44:24] <cradek> I'm sure following suffers, but it worked ok especially for tangent stuff
[20:45:43] <cradek> it could only process 3? blocks/sec so you sure couldn't use it to join little lines
[20:46:29] <skunkworks_> yikes
[20:46:53] <skunkworks_> I suppose - you could do the same with emc and high acceleration - large following error... ;)
[20:47:25] <cradek> sure
[20:54:11] <skunkworks_> The k&t only processed one block at a time.
[20:59:07] <cradek> I think boss8 did/does two or even three since it has rudimentary cutter comp
[20:59:33] <skunkworks_> cutter comp? what is that? ;)
[20:59:52] <skunkworks_> our lathes had knobs that you adjusted for each tool number
[20:59:54] <skunkworks_> iirc
[21:00:43] <cradek> for tool offset?
[21:01:02] <skunkworks_> yes
[21:01:36] <skunkworks_> well - I don't know - it might have been a way to change the tool size.. in effect the same thing.. never used it.
[21:03:15] <cradek> huh
[21:03:49] <cradek> the hnc had thumbwheels that went up to 0.3998 or some other crazy number - not long enough to use as full tool offsets, but plenty for wear
[21:03:50] <skunkworks_> a whole pannel of knobs - hundreds of them
[21:04:03] <skunkworks_> that could be
[21:04:24] <cradek> it had **RAM** to store them in! (and a marker board on the front of the control where you were told to also write the numbers)
[21:04:53] <skunkworks_> scary. was it ferrite beads?
[21:05:30] <cradek> no, it's 2101 rams (I think that's the right number)
[21:06:19] <skunkworks_> we had a teacher in school that decided from then on in class ram was going to mean read alterable memory so that it made sense with rom.
[21:07:00] <skunkworks_> he was a little nutty - also had a god complex I think.
[21:10:51] <cradek> who doesn't?
[21:11:17] <skunkworks_> ;)
[21:11:24] <cradek> I like today's dilbert
[21:11:58] <skunkworks_> heh
[21:12:00] <skunkworks_> funny
[21:20:39] <skunkworks_> http://dilbert.com/2010-01-18/
[21:20:55] <skunkworks_> jeeze - I have to get back to work!
[21:24:14] <skunkworks_> I have a question - lets say hypothetically - I wanted to rotate the b axis. but a lot of things have to happen before and after.. I don't think I have seen anything in emc that would be pre-post move pins. So I would geuess I would need to use custom m-codes before and after?
[21:24:41] <cradek> like locking and unlocking the axis?
[21:24:57] <skunkworks_> it actually has to lift it - before it can spin.
[21:26:43] <cradek> hmm
[21:27:16] <skunkworks_> the table is only indexable to 5 degree increments.
[21:41:26] <jepler> sounds like you want M1xx instead of G0/G1
[21:42:23] <skunkworks_> I was thinking that... although I would think m1x1 -> b100 -> m1x2 would work just fine. (
[22:35:08] <alex_joni> skunkworks_: B100 might hold other caveats..
[22:35:25] <alex_joni> like blending, etc
[22:38:18] <jmkasunich> yeah - "indexing" is NOT "moving"
[22:39:25] <alex_joni> especially feed override will byte you
[22:39:53] <skunkworks_> b would never be commanded with other moves..
[22:40:05] <jepler> hah byte
[22:40:50] <alex_joni> jepler: sorry.. left hand only typing
[22:49:26] <jt-plasma> feeding the baby
[22:49:48] <alex_joni> yup
[22:49:58] <alex_joni> done now .. so I got both hands back :D