#emc-devel | Logs for 2009-01-05

[00:09:07] <BigJohnT> goodnight all
[02:46:00] <jmkasunich> whoever did the logarithmic jog speeds in axis, thanks!
[02:46:24] <jmkasunich> relief groove and chamfer on the end done, came out fine
[02:47:10] <cradek> you get one try to guess who was smart enough to do that
[02:47:24] <jmkasunich> jepler
[02:47:32] <cradek> you win the prize
[02:47:51] <jmkasunich> although you told me about using the <> keys to change the speed, which is the other half of the equation
[03:06:10] <jmkasunich> G75 with no exit move - what does it do? pull X out as fast as possible after stopping Z motion?
[03:06:46] <cradek> g76, yes, rapids out to the drive line
[03:09:34] <jmkasunich> s/5/6 duh
[03:14:45] <cradek> I use g76 with no relief or in/out taper all the time
[03:33:49] <jmkasunich> cradek: another G76 question - the Z position in the G76 command defines the end of the drive line
[03:34:07] <jmkasunich> but does that include the additional Z advance of the 29 degree infeed?
[03:34:46] <cradek> yes, Z travel ends where you want
[03:34:55] <cradek> (why not look at the preview and see?)
[03:35:10] <jmkasunich> lack of landmarks in the preview
[03:35:22] <jmkasunich> although I guess I could put some G0's in there
[03:35:39] <cradek> jog the cone to it and check the number :-)
[03:35:47] <cradek> err, "cursor" haha
[03:36:01] <jmkasunich> yeah
[03:36:20] <jmkasunich> been farting around grinding a tool
[03:37:40] <cradek> threading tool for next to the shoulder?
[03:37:45] <jmkasunich> yeah
[03:38:06] <cradek> I've got a nice one you can borrow
[03:38:18] <cradek> grinding away all the right side of the V took forever
[03:38:23] <jmkasunich> email over, I'll send it back when I'm done
[03:38:30] <cradek> can I fax it?
[03:38:45] <jmkasunich> I just put a 60 degree point on the opposite end of a parting blade
[03:38:52] <jmkasunich> no, I don't have a fax machine
[03:39:13] <cradek> ah, you didn't have to grind much. I made mine out of 3/8" square HSS.
[03:59:52] <cradek> how's your thread?
[04:16:44] <jmkasunich> haven't cut it yet - got distracted
[04:39:12] <jmkasunich> first one came out good - one trail with an intentionally small K, measure pitch dia, set final K value and run - pitch dia came out within a half-thou
[04:39:18] <jmkasunich> second one running now
[04:39:28] <jmkasunich> s/trail/trial
[04:39:43] <cradek> half-thou on diameter is a pretty nice hit
[04:40:57] <cradek> I bet that large a thread has a big pitch dia tolerance
[04:42:43] <jmkasunich> yeah
[04:42:48] <jmkasunich> and its not critical at all
[04:43:10] <cradek> but some are, and you know how to do it
[04:43:11] <jmkasunich> I threadcut the matching hole (so I can cut to fit), and it gets locktited into that hole
[04:43:25] <cradek> ah, haha
[04:43:35] <jmkasunich> 2nd one is done, matches the first to better than a half-thou
[04:43:42] <cradek> that's the definition of not critical!
[04:43:58] <jmkasunich> the other nice thing about large diameter is minimal deflection
[04:44:05] <cradek> yeah
[04:44:38] <jmkasunich> usually when I do one by rough, measure, finish, the 2nd one that does everything in one pass is a different size, even with spring passes
[04:44:53] <jmkasunich> for these, I didn't even bother with spring passes - probably should of, but...
[04:45:14] <cradek> yeah, everything has spring
[04:45:36] <jmkasunich> well, a lot less in a 7/8" diameter than the M6's I did a while back
[04:46:03] <jmkasunich> especially since the M6 were tough steel, these are bronze - less cutting force, less tendency to rub instead of cutting
[04:46:23] <cradek> goodnight, I'm off
[04:46:28] <cradek> gotta go to work soon - for a change
[04:46:29] <jmkasunich> ditto
[09:36:51] <CIA-1> EMC: 03cmorley 07TRUNK * 10emc2/src/emc/usr_intf/stepconf/stepconf.py:
[09:36:51] <CIA-1> EMC: Fix dialog boxes to have international text. Change file paths to use global
[09:36:51] <CIA-1> EMC: variable distdir as it should have been. fix so on a new config we don't try to
[09:36:51] <CIA-1> EMC: copy the custom.clp file to a config directory that doesn't exist yet
[09:40:18] <CIA-1> EMC: 03cmorley 07TRUNK * 10emc2/src/emc/usr_intf/stepconf/stepconf.py: oops shoulda saved the changes before committing...
[13:06:24] <CIA-1> EMC: 03bigjohnt 07TRUNK * 10emc2/src/hal/user_comps/gs2_vfd.c: moved hal_ready to end of inits for SWPadnos
[13:48:47] <christel> [Global Notice] Hi all, I've added a new module to our services installation, that may be of interest to those of you with enforce enabled. It may also cause some adverse interactions with auto-identify scripts, however. Please see http://blog.freenode.net/?p=164 for more information. Thank you for using freenode and have a great day!
[13:57:56] <christel> [Global Notice] Hi all, terribly sorry to bother you *again* but regardless of how much hairdye I go through, I can't get away from being blonde. Needless to say, my previous global resulted in inadvertedly ddosing the machine hosting the blog and website. Please try again later and for now, call me Homer and watch me Doh!
[14:36:02] <jepler> anyone else care to review this change to motion to fix micges' wrong movement speed bug, or shall I just check it in? Also, thoughts on whether it should be fixed on the 2.2 branch or if people there should just use the workaround? http://sourceforge.net/tracker/index.php?func=detail&aid=2478266&group_id=6744&atid=106744 "take2.patch"
[14:43:57] <cradek> it's up to you, but I don't see it as necessary for 2.2 if he is the only one who has noticed it
[14:44:31] <cradek> for this one, the risk of breakage seems higher than the devil we know ... or something
[14:44:59] <jepler> one thing that occurs to me is that stepconf is almost certainly generating incorrect configurations and tickling this bug in 2.2
[14:45:10] <jepler> since servo_period is always 1ms and base_period is any damned thing
[14:48:27] <cradek> hm.
[14:56:43] <jepler> well, bug filed about that
[14:57:29] <cradek> do you know an upper bound for the velocity error?
[14:59:17] <jepler> yes, my guess is 10%
[15:00:02] <cradek> is it 10% too high or too low?
[15:01:05] <jepler> so far I've only seen the error give a real velocity that is too low but I'm not 100% sure of that
[15:06:12] <alex_joni> jepler: did you try to printk in the emcmotSetCycleTime() function?
[15:06:46] <alex_joni> I mean, I wonder if period might get reported with slightly different values, based on timer irregularities, or does it keep the same value forever
[15:08:19] <alex_joni> the second thins is the testing for success on setServoCycleTime and setTrajCycleTime(), I don't remember ever seeing an error like that, but keeping the return value check might be better?
[15:08:45] <cradek> your patch looks fine, but I would probably miss a subtle problem in this area
[15:08:57] <alex_joni> other than that it looks fine to me too
[15:27:27] <jepler> alex_joni: the period passed by hal to the function is the nominal period
[15:27:39] <jepler> it never changes unless the function is moved to a different thread
[15:28:38] <alex_joni> ok, then my concearn is out of place
[15:35:46] <steves_logging> steves_logging is now known as steve_stallings
[15:37:18] <steve_stallings> Is there any way to easily clone the content of the LinuxCNC web site. The Cabin Fever show is coming soon and it might be handy to have the content available without paying for and relying on the spotty WiFi at the convention site.
[15:39:02] <steve_stallings> Dang, I forgot that the search relies on Google....
[15:39:16] <cradek> I think it's all dynamic content too
[15:39:26] <cradek> (same for the wiki)
[15:39:47] <cradek> where is cabin fever?
[15:39:59] <steve_stallings> York, PA
[15:40:06] <cradek> ack.
[15:40:31] <steve_stallings> just an overnight drive from your place, right?
[15:40:48] <cradek> uh yeah
[15:41:37] <steve_stallings> seems a shame that EMC is so little represented at a show almost in NIST's backyard
[15:42:16] <cradek> is there much cnc stuff at these? I thought it was 'model engineering' type stuff, mostly manual machinists
[15:42:27] <steve_stallings> the MACH crew is out in full force this year, mulit-day seminar and two table with 3 running machines on the show floor
[15:43:14] <steve_stallings> we have had a CNC exhibitors area for the last 4 years, typically 5 or 6 guys including me
[15:44:07] <steve_stallings> I managed to drag Matt Shaver there last year for one day of the show.
[15:45:05] <cradek> it sounds like it would be fun, but I can't get there
[15:45:57] <steve_stallings> not even thinking that, though there was a remote possibility that Ray was coming, now not possible
[15:47:00] <steve_stallings> Chris Helgesen, the guy that ended up with my Mesa cards in a Hurco, will be there and we are trying to manage a slide show about his conversion
[15:47:47] <cradek> neat
[15:48:32] <cradek> I think my lathe is a neat conversion (probably only because it's the first one I've done myself) - but the same old story - I was too busy while doing it to take good pictures.
[15:49:19] <steve_stallings> I can come up with the usual demo machine, but what is lacking is a real EMC jockey to demo the stuff and get people excited about all the stuff other than flashy screens
[15:52:39] <jepler> if you need somebody to punch the guy who says "how do I move the estop button to the middle of the screen and change the color of the DRO", I'll do it for the price of a plane ticket.
[15:53:19] <cradek> trolllll
[15:54:30] <steve_stallings> up that offer to actually edit the Axis code live in front of a crowd of onlookers and I'll think about it... 8-)
[15:54:46] <cradek> editing source code is easy
[15:55:13] <cradek> getting it to build or run afterward...
[15:55:59] <steve_stallings> while 3 guys with no clue, one with a real question, and a couple of whiney kids in the background just to make it interesting
[16:02:28] <steve_stallings> dang, you can't get here from there, minimum two stopovers
[16:04:00] <steve_stallings> oops, yes you can, $100 more for one stopover
[17:28:10] <steve_stallings> steve_stallings is now known as steves_logging
[17:42:59] <alex_joni> hey seb_kuzminsky
[17:43:09] <seb_kuzminsky> hi there
[17:44:26] <alex_joni> what's new?
[18:20:37] <skunkworks_> * skunkworks_ is glad he isn't the only one that replies to the wrong email..
[18:44:45] <alex_joni> heh
[18:44:48] <seb_kuzminsky> not much new alex_joni ... been making chips instead of software, which is a fun change of pace :-)
[18:44:54] <alex_joni> sounds good :)
[19:37:40] <cradek> seb_kuzminsky: what are you making?
[19:38:55] <alex_joni> http://felixker.com/wp-content/images/2007/09/potato_chips.jpg
[19:39:49] <cradek> http://www.freefoto.com/images/9905/05/9905_05_31---Chips-with-salt-and-vinegar_web.jpg
[19:41:16] <seb_kuzminsky> cradek: belt-drive spindle conversion for the X2 mini-mill, following the the Hoss design
[19:41:36] <cradek> fun
[19:41:48] <seb_kuzminsky> it's a good starter project for me :-)
[19:42:12] <cradek> it's challenging to bootstrap a machine. it takes some planning.
[19:42:22] <seb_kuzminsky> http://www.hossmachine.info/projects_6.html#belt%20conversion
[19:42:55] <cradek> oh it comes with gears or something?
[19:43:10] <seb_kuzminsky> the mill ships with a two-speed gearbox for the spindle
[19:43:24] <cradek> aha, I bet belts are much better
[19:43:51] <seb_kuzminsky> they're supposed to be quieter and less prone to breaking (the chinese gears are made of nylon)
[19:43:52] <cradek> add a power drawbar while you're at it :-)
[19:43:55] <seb_kuzminsky> heh
[19:44:07] <alex_joni> so .. is this under cnc?
[19:44:22] <seb_kuzminsky> i was trying to think of a way to add a spindle encoder, but it seems tricky with a V-belt, easy with a timing belt
[19:44:29] <seb_kuzminsky> alex_joni: no, the cnc conversion comes later ;-)
[19:44:34] <alex_joni> ah, ok ;)
[19:44:53] <cradek> when you make the pulley, make room for a timing belt below the drive belt?
[19:44:53] <alex_joni> I was wondering.. cause I didn't see any requests for help with a hostmot2 config on the lists :P
[19:44:58] <cradek> haha
[19:45:16] <seb_kuzminsky> yeah, i'm composing it now: YOUR CRAPPY DRIVER DOESNT WORK!!!!11!
[19:45:24] <alex_joni> it's probably hard to figure out which board to use
[19:45:48] <seb_kuzminsky> i'm planning to use a different board for each pwm & each encoder :-P
[19:45:54] <cradek> ha
[19:46:06] <alex_joni> still better than: YOUR CRAPPY DRIVER BROKE MY SPINDLE GEARS !!!11!
[19:46:17] <seb_kuzminsky> heh
[19:47:04] <alex_joni> hmm.. I talked to some ubuntu guys, and it seems we can use ppa for building emc2 packages
[19:47:06] <seb_kuzminsky> i'm dreading that email... YOUR DRIVER WENT CRAZY AND CRASHED MY MACHINE
[19:47:10] <cradek> what is the nub on the front of his motor mount?
[19:47:17] <seb_kuzminsky> spindle lock
[19:47:25] <cradek> ah, neat
[19:47:27] <alex_joni> (including custom kernel image + headers + rtai + whatnot)
[19:47:38] <seb_kuzminsky> alex_joni: wow, cool :-)
[19:47:54] <cradek> what is ppa?
[19:47:55] <alex_joni> seb_kuzminsky: if we ever want to do that :)
[19:48:02] <seb_kuzminsky> cradek: personal package archive
[19:48:06] <seb_kuzminsky> it's part of launchpad
[19:48:08] <alex_joni> cradek: a place on launchpad where you upload sources
[19:48:11] <alex_joni> and they get built
[19:48:14] <cradek> ah
[19:48:16] <alex_joni> and put in a repo iirc
[19:48:21] <seb_kuzminsky> i was planning to add something like that to the buildbot
[19:48:41] <cradek> the launchpad kool-aid is calling us - can you hear it? driinnnkkkk meee!
[19:48:43] <seb_kuzminsky> have a builder make a source package, and trigger a couple other builder to build the sources into .debs under pbuilder
[19:48:55] <seb_kuzminsky> cradek: oh i hear it ;-)
[19:49:02] <alex_joni> kool-aid ?
[19:49:36] <seb_kuzminsky> http://en.wikipedia.org/wiki/Kool-Aid#.22Drinking_the_Kool-Aid.22
[19:49:40] <cradek> ah, thanks
[19:49:46] <cradek> I was at a loss to explain it
[19:49:58] <seb_kuzminsky> that's why jesus invented the wikipedia ;-)
[19:50:08] <alex_joni> Ramen
[19:50:19] <seb_kuzminsky> mmm ramen... bbl lunch
[19:52:03] <cradek> seb_kuzminsky: for those slots, check out useful-subroutines.ngc
[19:52:47] <cradek> if using cnc, I think I'd round all the corners on his design... more friendly that way
[19:52:54] <cradek> (your hands will be up there a lot)
[20:07:23] <fenn_> fenn_ is now known as fenn
[20:57:28] <cradek> http://www.hossmachine.info/images/belt%20conversion%20base%20plate.jpg
[20:58:06] <cradek> seb_kuzminsky: good luck with all those .0001 dimensions
[21:00:03] <jepler> hah
[21:00:33] <jepler> how far is one revolution of the handwheel? 62.5 mil like on the sherline?
[21:00:38] <jepler> or whatever it is
[21:01:05] <cradek> jepler: I think sherline is .05 inch just like any sane machine
[21:01:40] <jepler> what machine is it that one turn of the handwheel isn't any sane number, then?
[21:01:55] <cradek> I think all the harbor frieght stuff is like that
[21:02:03] <jepler> ah that sure could be
[21:03:26] <cradek> ummm
[21:03:29] <cradek> http://www.hossmachine.info/images/IMG_0364_800x600.jpg
[21:03:35] <cradek> this picture is titled "table flatness"
[21:03:42] <cradek> can anyone spot the problem?
[21:04:45] <jepler> I'm still back at the point where I try to figure out why it's labeled in X and Y
[21:05:14] <cradek> well, good question
[21:05:17] <cradek> I have no idea
[21:05:20] <skunkworks_> other than the table is sticking out a mile?
[21:05:30] <cradek> yeah, other than the machine wanting to fall over
[21:05:59] <cradek> ... and off the table
[21:50:22] <fenn> hmm my gingery lathe was .0625/turn until i replaced the screw with a 20-tpi (after i learned a thing or two)
[21:51:17] <fenn> at the time i thought it was a neat idea to attach cordless drills to the handles so i didnt have to crank so much
[21:51:25] <cradek> I bet that was a big pain unless you were working to sixteenths only
[21:51:40] <fenn> i had a 1/64 dial calipers :\
[21:51:49] <cradek> haha
[21:52:07] <cradek> so 4 marks around the dials?
[21:52:21] <cradek> * cradek still doesn't know fractions
[21:52:22] <fenn> no i tried to do .001" markings but gave up
[21:52:39] <fenn> my brain doesn't come with a carry bit
[21:52:44] <cradek> not having an integer number of them would really suck
[22:04:42] <CIA-1> EMC: 03jepler 07TRUNK * 10emc2/src/hal/hal_lib.c: fix the 'thread period less than clock period' error; it was never triggered even when it should have been
[22:12:08] <CIA-1> EMC: 03cradek 07concave_comp2 * 10emc2/src/emc/rs274ngc/interp_convert.cc: error in the unhandled cases. the mission, now, is to handle the important ones. (XXX)
[22:13:33] <skunkworks_> heh
[22:14:07] <cradek> interestingly though, I think this is currently somewhat usable
[22:14:16] <cradek> bbl.
[22:16:00] <cradek> (for XY, not lathe/ZX)
[22:16:19] <cradek> bbl, really now
[22:26:14] <CIA-1> EMC: 03jepler 07TRUNK * 10emc2/src/hal/hal_lib.c: since it's bad to add items after calling ready, treat it as an error
[22:46:38] <seb_kuzminsky> cradek: i have a vague notion to try to make hoss' parts by hand first, and if anything needs redoing when the cnc conversion is done i'll do it then
[22:46:40] <seb_kuzminsky> i'
[22:46:47] <seb_kuzminsky> i'll settle for .005" ;-)
[23:00:03] <cradek> cool. I think you could do most of those parts with a saw and drill press - the slot would be the hard part
[23:00:12] <cradek> milling all the edges will sure make it all prettier, though
[23:01:06] <cradek> on your small machine, drilling most of the material away before trying to mill the slot will help a lot
[23:02:19] <seb_kuzminsky> i'll make sure to do that
[23:02:21] <cradek> * cradek mistakenly thinks seb_kuzminsky asked for advice
[23:02:57] <seb_kuzminsky> well i really appreciate the advice i, though i actually knew that particular trick ;-)
[23:03:04] <cradek> aha :-)
[23:03:21] <seb_kuzminsky> here's a write-up of my first-ever milling project, where i pre-drilled my slots: http://www.instructables.com/id/making_vise_clamps_on_the_milling_machine/
[23:04:00] <cradek> neato
[23:04:07] <cradek> instructables is light on machining projects
[23:04:34] <jepler> yeah, I usually think of it as a fluff site with terrible projects
[23:04:53] <seb_kuzminsky> there's a lot of chaff there, that's for sure
[23:05:04] <seb_kuzminsky> unless you really want to see some kid make a sword out of k'nex
[23:05:08] <jepler> also their "you have to log in for this site to be convenient" bullshit rubs me wrong everyime I visit
[23:05:28] <seb_kuzminsky> any good alternatives? pref with lots of machining projects? :-)
[23:05:35] <jepler> nah
[23:05:45] <jepler> I put stuff on my own blog
[23:05:49] <jepler> I figure google finds it
[23:05:58] <seb_kuzminsky> good idea
[23:07:26] <cradek> http://www.instructables.com/files/deriv/FWY/RX92/FD80X40I/FWYRX92FD80X40I.MEDIUM.jpg
[23:07:45] <cradek> interesting. you can see the transition between 'machine is stout enough' and 'machine is not stout enough'
[23:12:21] <cradek> those are very nice looking clamps
[23:12:39] <cradek> and a nice writeup.
[23:13:07] <cradek> it shows how much work it is to make a "simple" part
[23:13:58] <seb_kuzminsky> thanks cradek :-)
[23:14:40] <seb_kuzminsky> it's a lot of work for sure, but at least it keeps us off the streets
[23:14:48] <cradek> I made a bunch of T nuts for max. had to put each in the vise several times...
[23:15:17] <cradek> aluminum though - because it has an aluminum table
[23:15:31] <seb_kuzminsky> custom size nuts?
[23:15:34] <cradek> interesting to see that you actually can cut steel successfully with this little machine.
[23:15:38] <seb_kuzminsky> or Al as a custom material?
[23:15:57] <cradek> I needed some, the originals were Al too
[23:16:06] <cradek> the originals were low quality and I never had enough of them
[23:17:27] <cradek> looks like a nice little vise.
[23:18:00] <seb_kuzminsky> well i like it, but then it's the only vise i've used :-)
[23:18:58] <cradek> if the base of the jaws is parallel to the table, it's better than mine started out :-)
[23:21:52] <jmkasunich> I like those "screwless" vises
[23:22:02] <jmkasunich> I have a 2.5" one that I've had for years
[23:22:09] <cradek> dinner, bbl
[23:22:14] <seb_kuzminsky> see ya cradek
[23:22:16] <jmkasunich> same here ;-)
[23:22:23] <jepler> dinnertime?
[23:22:28] <jepler> I guess I should get out of the office :-/