#emc | Logs for 2004-12-11

[00:02:19] <paul_c> well.. EMC won't trap the slot smaller than tool error
[00:02:35] <paul_c> nor can it do adaptive tool offsets (yet)
[00:02:44] <SpeedBump> even simple things like a 'v' cut into a straight edge can lead to problems. old CNC controls (like a burney) will just radius from one edge to another. in the 'v' case this gouges a big chunk out of the material...I have seen this happen. That's why I ask...
[00:05:27] <paul_c> best thing to do is grab a copy of BDI-Live and run some evaluations - It saves the hassle of doing an initial install.
[00:05:39] <SpeedBump> * SpeedBump nods
[00:06:45] <paul_c> If you find the code is usable (or adaptable) for your needs, the next release will have a bunch of stuff that might be of interest..
[00:06:45] <SpeedBump> I'll look into it. Is there a design document somewhere outlining the structure of EMC2?
[00:07:23] <paul_c> only the emc2_source_code_notes
[00:07:34] <SpeedBump> in CVS somewhere?
[00:07:38] <paul_c> http://www.linuxcnc.org/EMC2_Code_Notes.pdf
[00:07:47] <SpeedBump> thanks
[00:10:57] <SpeedBump> does this channel log it's data? can I see the discussions from the sunday developer meetings somewhere?
[00:11:47] <paul_c> http://cia.navi.cx/stats/project/emc
[00:12:04] <paul_c> wrong link...
[00:13:44] <paul_c> http://www.linuxcnc.org/EMC_news_history/IRC-logs/
[00:16:47] <paul_c> That should keep you busy for a few hours...
[00:18:24] <paul_c> * paul_c heads off to bed.
[00:23:09] <robin_sz> * robin_sz needs more curry
[01:00:20] <asdfqwega> * asdfqwega loves lurking
[01:00:29] <asdfqwega> <_<
[01:00:33] <asdfqwega> >_>
[01:00:42] <asdfqwega> o_O
[01:17:19] <Imperator_away> Imperator_away is now known as Imperator_
[02:08:56] <ottos> good day gents...
[12:35:29] <paul_c> * paul_c is away: Making chips
[15:26:10] <rayh> Need just a bit of boot advice.
[16:30:44] <danfalck> * danfalck is away: going out to the shop...
[16:39:39] <dave-e> dan...going shopping is always a good idea
[16:41:11] <paul_c> eeeks... Not this time of year.
[16:47:31] <dave-e> no paul...not that kind....to the shop...to the shop
[16:48:36] <tbl> hey paul
[16:48:52] <tbl> good morning! sorry for ditching you the other day i had to leave to take care of somethings
[16:49:14] <dave-e> hi mike...what are you doing these days
[16:49:17] <paul_c> Hi Mike
[16:49:36] <tbl> unix consulting, competing with sherline
[16:49:42] <tbl> going pretty well actually!
[16:49:48] <dave-e> good
[16:49:58] <paul_c> did you find that ISO ?
[16:50:03] <tbl> I did
[16:50:43] <paul_c> please don't put it on a public server - It's still under testing
[16:50:53] <tbl> -rw-r--r-- 1 root root 604M Dec 10 16:40 bdi-4.02.iso
[16:50:57] <tbl> well
[16:51:05] <tbl> sherline seems to think it is good for public use!
[16:51:10] <tbl> what is it based off of?
[16:51:41] <tbl> like
[16:51:42] <paul_c> I need to add a post-config script and clean up a few things.
[16:52:05] <tbl> they have an open PO with me, I am just waiting for them to supply me with a disk image
[16:53:31] <paul_c> The CD is Debian based, and running a 2.6.9 kernel.
[16:53:32] <dave-e> morning les
[16:53:39] <les> morning
[16:53:47] <tbl> 2.6!!!
[16:53:49] <tbl> you are crazy
[16:53:50] <dave-e> how are the cherry chips
[16:54:05] <tbl> what is the advantage of running 2.6?
[16:54:13] <les> growing soon
[16:54:22] <dave-e> good
[16:54:37] <les> laying out the tool paths
[16:54:43] <dave-e> more fun
[16:54:47] <paul_c> RTAI development is concentrating on 2.6 kernels.
[16:55:34] <tbl> is the performance increase significant enough to justify the risk of an unstable kernel?
[16:56:02] <dave-e> I think paul likes the bleeding edge
[16:56:19] <paul_c> Is 2.6.9 any more unstable than the pre18 2.4 series ?
[16:57:09] <tbl> i don't know! asking you =)
[16:57:20] <dave-e> to paul's horror I'm also going shopping
[16:57:37] <tbl> but every time i have worked with -latest linux kernel i have been very disapointed
[16:57:49] <tbl> i built a huge SQL server a few weeks ago
[17:02:09] <paul_c> Been running a 2.6.9 kernel continously for the last month - No problems so far.
[17:02:12] <tbl> i had to buy redhat enterprise for $1500 because it was x86-64
[17:03:43] <paul_c> Debian has 64 Bit packages in their repository.
[17:39:52] <les> hmm I need to use tool length compensation
[17:40:13] <les> does not seem to be well documented in the handbook
[17:40:49] <les> which is the tool file and what is the format?
[17:46:41] <les> ah...I see it...tool.tbl
[18:34:32] <robin_sz> 'ello
[18:35:29] <les> hi
[18:35:59] <les> just working on tool paths here
[18:36:31] <robin_sz> getting smoother?
[18:36:46] <robin_sz> is segmotqueue working fine for you now?
[18:36:53] <les> heh...I hope
[18:37:15] <les> no, sadly some more problems crept up
[18:37:21] <robin_sz> oh.
[18:37:26] <les> just have to run low accel
[18:37:32] <robin_sz> oh.
[18:38:21] <les> segmentqueue still stops cold every once in a while
[18:38:32] <robin_sz> beginning to sound like its terminal :(
[18:38:42] <les> works, but not reliable enough for production
[18:38:54] <robin_sz> if it cant be fixed in over a year, then perhaps its time to start afresh?
[18:39:39] <les> yes...the original author is on to other things...and just can't do a lot of free programming
[18:39:48] <robin_sz> I liked that 'segment assembly' mehtod I saw in another system ...
[18:40:13] <robin_sz> run down the path in the forwards directio, acceleratring out of the corners
[18:40:20] <robin_sz> turn it around and run backwards
[18:40:47] <les> well...just a plain old splining of the path would do
[18:40:48] <robin_sz> couple of iterations and you have the perfect path of little segments
[18:40:54] <les> with corner slowing
[18:41:02] <robin_sz> oops, tea time, back in ten
[18:41:25] <les> funny thing is emc does have cubic "subinterpolation"
[18:41:43] <les> ok...back to tool paths for me...and lunch
[19:20:21] <dave-e> well I've always thought cnc packages needed a way to get a toolpath off a function
[19:21:15] <dave-e> at least 2nd order polynomial and maybe a generalized function give the function axis, limits and watch it go
[19:22:02] <dave-e> I'm still thinking 2D but....
[19:39:19] <picnet> * picnet Tips hat.
[20:46:56] <dave-e> certainly didn't get much response off that suggestion did I. ;-)
[20:51:06] <paul_c> * paul_c is back
[22:04:52] <robin_sz> * robin_sz returns from tea
[22:36:33] <CIA-6> 03Zathras 07BDI build system * 10Babylon Cluster/comps.xml: File changed. New revision:1.19