#emc-devel | Logs for 2009-07-15

[01:21:42] <jmkasunich> cradek: "you can't pull it off the spool without breaking it" heh - what blows my mind is that at some point, they pull it thru a drawplate to make it that size
[01:27:06] <cradek> yeah that borders on ridiculous doesn't it
[01:28:00] <cradek> http://timeguy.com/cradek-files/emc/imag0042.jpg
[01:28:25] <jmkasunich> I read somewhere that they make some crazy fine platinum wire by embedding a moderately fine plat wire inside a thick walled copper tube of sorts, then draw the whole thing down till the copper is small and the plat is damn small, then etch away the copper
[01:28:44] <cradek> ^ this is the motor on my air compressor. it is tripping the breaker sometimes. The one time I'm sure I heard it, I'm pretty sure it started, and then tripped 5 seconds or so later.
[01:29:01] <cradek> how should I debug this problem?
[01:29:21] <jmkasunich> how big is the breaker?
[01:29:25] <cradek> 30
[01:29:58] <jmkasunich> do you have anything that would let you measure the time/current profile during a start?
[01:30:16] <jmkasunich> I guarantee you it draws significantly more than 30A while starting
[01:30:19] <cradek> no unfortunately
[01:30:35] <cradek> ok, I didn't know if 19.50 was peak or steady state or something in between
[01:30:56] <SWPadnos> don't breakers get "weak" once they trip?
[01:31:13] <SWPadnos> (standard consumer electrical panel ones, that is)
[01:31:14] <jmkasunich> 19.5 is the steady state current at rated load
[01:31:15] <cradek> I could just put in a bigger one of course - but I didn't know if there might be an actual problem and that would be a bad idea.
[01:31:45] <jmkasunich> the breaker needs to be (and probably is) sized to protect your wiring - don't increase it unless you know the wiring can take more current
[01:31:49] <cradek> ok, it will definitely be at full load all the time
[01:32:30] <cradek> I wired it - it is big - I will check that but I bet it's fine for more
[01:32:40] <jmkasunich> 30A is 10 AWG
[01:32:48] <jmkasunich> IIRC
[01:33:31] <jmkasunich> can you measure the steady state current at all?
[01:33:41] <jmkasunich> probably not - most multimeters only go to 10A
[01:33:42] <cradek> how would I do that?
[01:34:38] <cradek> maybe I could buy something at HF - can I use some clamp-around thing?
[01:34:45] <jmkasunich> yeah, I'm googling
[01:35:54] <SWPadnos> my radio shack meter goes to 20A, but that's still pretty close to the limit
[01:36:36] <cradek> I don't see how you could get it started and then insert the meter
[01:37:00] <cradek> I guess you'd short the leads to start it
[01:37:06] <SWPadnos> you could do it with a contactor
[01:37:10] <SWPadnos> NO/NC
[01:37:10] <jmkasunich> http://www.harborfreight.com/cpi/ctaf/displayitem.taf?Itemnumber=96308
[01:37:38] <jmkasunich> it isn't clear how high that meter goes
[01:38:03] <jmkasunich> but in general clamp-on's are pretty resistant to overloads, since they don't have a shunt to burn out
[01:38:45] <jmkasunich> if the steady state current is higher than the motor nameplate, then the motor is overloaded
[01:39:09] <jmkasunich> you can reduce the load by putting a smaller pulley on the motor (slows down the compressor, gives a bit less CFM, but eases the load on the motor)
[01:39:36] <cradek> ok, I will get one of these and check it
[01:39:44] <jmkasunich> if the steady state current is within specs, then it is a race between the motor getting started and the time delay in the breaker timing out
[01:39:45] <cradek> a smaller pulley would give an easier start too
[01:39:49] <jmkasunich> yes
[01:40:24] <cradek> when it's starting several times an hour, does the breaker heat up or something?
[01:40:45] <jmkasunich> but if you bought a 5HP compressor because you need a boatload of air, the smaller pulley will set you back
[01:41:03] <jmkasunich> yes, but I can't quantify how much
[01:41:27] <jmkasunich> the breaker isn't the only thing that heats up - repeated starts heat the motor too
[01:41:39] <cradek> thanks, I will do this test
[01:41:41] <jmkasunich> I wouldn't be surprised if starting current is 60A
[01:42:06] <cradek> also I will check the wire size
[01:42:15] <cradek> brb, dessert calls
[01:42:31] <jmkasunich> how much air doyou use? does the compressor run 25% of the time? 50%, 75%, 95%?
[01:43:06] <jmkasunich> if the duty cycle is low, you have too much compressor, and a smaller pulley would be a good thing - reduced cycling and cooler running all around
[01:43:33] <jmkasunich> mmmmm, dessert
[01:49:00] <SWPadnos> hmmm. so any ideas on how to get a tool out of an MT3 taper?
[01:49:23] <SWPadnos> (a Tapmatic 70X that's stuck in an MT3 -> R8 adapter)
[01:49:27] <jmkasunich> does the taper have the cross-slot for a wedge?
[01:49:31] <SWPadnos> yes
[01:49:39] <SWPadnos> so just pound on it until they come apart?
[01:49:41] <jmkasunich> put a wedge in and whack it
[01:49:58] <SWPadnos> ok. I wasn't sure if it was a good idea to hit it that hard
[01:50:09] <jmkasunich> well, how hard is "that hard"?
[01:50:11] <SWPadnos> they sure do hold nicely though - I don't need to worry about that at all
[01:50:37] <SWPadnos> well, I'm not sure. I did a bit of whacking on a screwdriver in the slot, and I couldn't get it to budge
[01:50:49] <jmkasunich> oh - don't use a screwdriver
[01:50:52] <SWPadnos> heh
[01:51:03] <jmkasunich> if you don't have a proper morse taper wedge, set it aside until you can get one
[01:51:10] <SWPadnos> ok
[01:51:27] <SWPadnos> thanks - I figured what I was doing would end up being stupid if I tried harder
[01:52:04] <jmkasunich> http://www.use-enco.com/CGI/INSRIT?PMAKA=292-6230&PMPXNO=940756&PARTPG=INLMK32
[01:52:30] <SWPadnos> oh - thanks
[01:52:32] <jmkasunich> you could make one if you are in a hurry, but $3.40 doesn't buy much of your time
[01:52:54] <SWPadnos> I think I need a jacobs flex collet too (expensive as hell), so maybe it's time to take MSC up on one of their offers
[01:59:19] <jmkasunich> this is cool (or maybe I'm just easy to impress): http://www.gpsvisualizer.com/display/1247623085-12720-
[02:00:10] <SWPadnos> hmm. I wonder if our Garmin can collect the right data for that kind of thing
[02:00:16] <SWPadnos> (older, walking-style Garmin)
[02:00:41] <jmkasunich> this is a walking garmin (obviously, no roads there...)
[02:01:08] <SWPadnos> we're going on a bike tour think this Saturday. it might be interesting to plot our course
[02:01:15] <jmkasunich> if it can record "tracks" it should work
[02:01:17] <SWPadnos> s/think/thing/
[02:01:23] <SWPadnos> I think it can
[02:01:28] <jmkasunich> I use gpsbabel to pull the track out of the GPS
[02:01:37] <SWPadnos> it's been a while. hopefully the batteries haven't exploded in it
[02:01:43] <jmkasunich> mine is newer, it has a USB interface - yours is probably serial, so you get to play with cables
[02:01:59] <jmkasunich> but gpsbabel will almost certainly support it
[02:02:32] <SWPadnos> I think I have the cable. I got it with Windows software
[02:02:50] <SWPadnos> never connected it, but I have the software for downloading maps and whatnot
[02:02:54] <jmkasunich> the image above is tonight's hike - here is last night's sail: http://www.everytrail.com/view_trip.php?trip_id=273890
[02:03:20] <jmkasunich> I think gpsbabel offers a windows build as well as linux
[02:03:41] <SWPadnos> oh - I'd rather use Linux software anyway
[02:04:42] <jmkasunich> http://www.gpsvisualizer.com/ is a very nice website - it uses gpsbabel internally, and merges your file(s) with google maps
[02:10:31] <SWPadnos> ok. it's an eTrex Legend C, and it's USB
[02:10:40] <SWPadnos> and the batteries still work, and have not exploded
[02:10:49] <jmkasunich> cool
[02:10:55] <jepler> I just took a look at git, and the bad G76 I/J/K "fix" was post-2.3.2. that makes me a bit happier.
[02:11:02] <jmkasunich> mine is an Etrex Legend H
[02:11:11] <jepler> there's still the stepconf bug I added, but at least the G76 bug wasn't in 2.3.2
[02:11:13] <jmkasunich> H - high sensitivity reciever (nice for wooded areas)
[02:12:06] <jmkasunich> SWPadnos: one gotcha that I've noticed - when uploading, I always make sure the GPS is displaying it's main menu
[02:12:21] <SWPadnos> oh. good to know
[02:12:36] <jmkasunich> if it is displaying something else (like a map view), I get comm timeouts - I'm guessing that the CPU load of doing the map display slows it down
[02:13:05] <SWPadnos> I wonder if the H model could locate any satellites in my basement
[02:13:11] <jmkasunich> mine can't
[02:13:15] <SWPadnos> err, while in my basement
[02:13:23] <SWPadnos> I don't have any satellites in my basement
[02:13:36] <SWPadnos> ok, neither can my C
[02:13:44] <jmkasunich> when I plug it into the PC and turn it on, it flails, but there is a option to run without the reciever, I turn that on while doing the upload
[02:14:01] <jmkasunich> it also runs off USB power when plugged in, so the batteries last longer
[02:14:16] <SWPadnos> I'm looking at the satellite display, and it shows 7 satellites, all with zero signal
[02:14:28] <SWPadnos> upstairs it could see 3 or so at a time
[02:14:42] <jmkasunich> upstairs indoors? that isn't bad
[02:14:54] <SWPadnos> I think 17 is directly overhead at the moment
[02:15:02] <SWPadnos> and 28
[02:38:09] <CIA-1> EMC: 03jepler 07v2_3_branch * r72639143a136 10/debian/changelog: update
[02:38:16] <CIA-1> EMC: 03jepler 07master * rccca229e389f 10/src/.gitignore: these are now generated
[02:47:09] <jepler> hmph. I got a very broken build after 'git checkout v2_3_branch' that was cured by 'make clean' :(
[02:51:36] <cradek> jepler: oh yay, that's great
[02:51:48] <cradek> the g76 thing, I mean
[02:52:15] <cradek> jmkasunich: (earlier conversation) wild-ass guess is it probably runs 25% of the time when I'm actively using the lathe
[05:27:22] <CIA-1> EMC: 03seb 07master * r7f0b17cbb841 10/src/hal/drivers/mesa-hostmot2/stepgen.c: use jepler's force_precision hack
[05:55:05] <CIA-1> EMC: 03seb 07master * ra0420f7050f1 10/src/hal/drivers/mesa-hostmot2/stepgen.c: fix a stupid compile error
[06:58:23] <CIA-1> EMC: 03cmorley 07master * r172569d0226a 10/src/emc/usr_intf/stepconf/ (stepconf.glade stepconf.py): Make stepconf usable with 800x600 display
[06:58:25] <CIA-1> EMC: 03cmorley 07master * ref2a83092fb9 10/src/emc/usr_intf/stepconf/stepconf.py: change stepconf to close expander widget
[12:10:46] <christel> [Global Notice] Hi all, I need to remove a server from production, it will be somewhat noisy and approximately 5,000 users will be affected. Apologies for the inconvenience and have a good day.
[12:27:38] <jepler> seb: oops, I had noticed that my sample code was broken in the same way when I actually tested it, but didn't mention it to you (the "stupid compile error")
[12:27:44] <jepler> did it fix the problem in your testing?
[12:32:02] <mozmck> jepler: looks like this $(eval echo $datadir)/emc expanded to this ${prefix}/share/emc in emc.tcl
[12:33:38] <mozmck> ...for the EMC2_IMAGE_DIR variable in configure.in
[12:36:31] <jepler> if it says ${prefix} in emc.tcl, I don't think it will work. That's what the 'eval echo' was supposed to fix. :-/
[12:36:38] <jepler> 2.3 has this: EMC2_IMAGE_DIR=$(eval echo $sysconfdir)/emc2/
[12:37:06] <jepler> and emc.tcl ends up with this: variable IMAGEDIR /etc/emc2/
[12:38:13] <mozmck> yeah, ${prefix} doesn't work :) I think the problem is that datadir is defined as ${prefix}/share/emc
[12:38:33] <mozmck> but I can't find where that happens now...
[12:38:52] <jepler> (but is IMAGEDIR actually used?)
[12:39:27] <jepler> (oh yes it is)
[12:42:02] <mozmck> I'll play with it and figure it out...
[12:42:06] <jepler> good luck
[12:42:20] <jepler> I thought I understood this, but now I am less sure :-/
[12:42:28] <jepler> --sysconfdir=/etc
[12:42:42] <jepler> when building the package, it configures overriding sysconfdir so it's not defined in terms of ${prefix}
[12:43:25] <jepler> you might look at whether substituting prefix in emc.tcl helps; tcl accepts the ${name} form of variable reference.
[12:43:26] <mozmck> yeah, $datadir actually is set to $datarootdir which is set to $prefix/share
[12:43:58] <mozmck> maybe it's the extra level that throws it off. I'll try several things and see.
[12:44:24] <jepler> could be; $(eval echo) will only expand the variable one additional 'level'
[12:50:56] <jepler> http://pastebin.ca/1495718
[12:51:12] <jepler> this shell function performs expansion until no '$' characters remain ..
[12:51:26] <jepler> prefix='/usr/local' datarootdir='${prefix}/share' datadir='${datarootdir}'
[12:51:26] <jepler> fe "$datadir"
[12:51:39] <jepler> so this prints /usr/local/share
[12:51:51] <jepler> bbl
[15:39:25] <fenn> how is CIA set up now that you're using git?
[15:39:37] <fenn> does it just make noise when you push to the shared repo?
[15:40:25] <cradek> yes git.linuxcnc.org tells CIA when there is a push
[15:40:43] <cradek> it also sends the email to emc-commit
[17:39:46] <alex_joni> fenn: it's just by using a different commit-hook
[17:40:54] <alex_joni> that works for almost any kind of versioning system where you have access to install a commit-hook
[17:51:51] <seb_kuzminsky> jepler: your force_precision hack seems to have made the float problem go away, at least on the ones i tested, so thanks
[17:55:56] <jepler> seb_kuzminsky: whee
[17:56:55] <cradek> quick! more people should answer the question about autoconf!
[17:56:56] <cradek> haha
[18:00:18] <seb_kuzminsky> ooh, ooh, i want to be the helpful one!
[18:00:46] <cradek> me me me! pick me!
[18:04:45] <SWPadnos> make make auto auto configureconf!
[18:23:09] <SWPadnos> damn. looks like Seb won the competition :)
[19:20:55] <micges> guys I have few questions about joints_axes:
[19:21:16] <micges> 1. what is/was [TRAJ] HOME used for ?
[19:22:45] <micges> it is coord home position, but what for? for saying thaT home coords can be else than 0,0,0 ?
[19:23:39] <SWPadnos> it may not be true that homing the joints puts the machine at 0,0,0[0,0,0,0,0,0]
[19:30:03] <micges> I see
[19:50:35] <SWPadnos> micges, actually, thinking about it a little more, I'm not sure how you could do the equivalent of a HOME_OFFSET without something like [TRAJ]HOME
[19:50:39] <CIA-1> EMC: 03micges 07joints_axes3 * r0817d3a7d49f 10/src/emc/ (5 files in 3 dirs): Rewrite getting home position in cartesian space from inifile
[19:51:19] <SWPadnos> you can't just move each joint to its home position like we do with trivkins, because there may be constraints from other joints
[19:51:43] <SWPadnos> so you'd have to home all the joints, then do a "world mode" coordinated move to the home position
[19:51:57] <SWPadnos> I don't know if that's what it was intended for, but it seems likt it could be useful :)
[19:52:01] <SWPadnos> s/likt/like/
[19:52:25] <micges> I know but I don't have any nontrivkins machine :)
[19:52:35] <SWPadnos> me either. only in sim
[19:54:14] <tissf> hello all
[19:55:43] <micges> hi
[19:57:12] <micges> SWPadnos: each day I do joints_axes3 I want more to classify each part of task and write clean interfaces ;)
[19:58:10] <SWPadnos> OK by me :)
[19:58:19] <micges> hehe
[20:01:26] <tissf> I have a problem with git or ssh someone can help me?
[20:01:50] <tissf> I git pull --rebase
[20:02:25] <tissf> francis@aldi2:~/emc2-dev$ git push
[20:02:27] <tissf> To ssh://tissf@git.linuxcnc.org/git/emc2.git
[20:02:28] <tissf> ! [rejected] master -> master (non-fast forward)
[20:02:30] <tissf> error: failed to push some refs to 'ssh://tissf@git.linuxcnc.org/git/emc2.git'
[20:02:57] <tissf> what have I missed?
[20:06:17] <SWPadnos> sorry - can't help you on that one
[20:07:23] <tissf> ok thanks
[20:08:03] <tissf> the first git push the other day was fine!
[20:08:40] <tissf> I have to wait for jepler or cradek
[20:09:16] <SWPadnos> heh -yep, probably
[20:09:33] <tissf> np
[20:13:11] <seb_kuzminsky> tissf: non-fast forward means your push wants to modify history on the receiving side, not just add new stuff
[20:13:16] <seb_kuzminsky> i think...
[20:14:04] <jepler> please show the output of "git pull --rebsae"
[20:14:08] <jepler> er, "--rebase'
[20:14:09] <jepler> "
[20:14:20] <tissf> ok
[20:16:21] <jepler> (hi tissf)
[20:17:03] <tissf> hi jepler
[20:17:09] <tissf> From ssh://tissf@git.linuxcnc.org/git/emc2
[20:17:11] <tissf> d5119ce..0817d3a joints_axes3 -> origin/joints_axes3
[20:17:12] <tissf> First, rewinding head to replay your work on top of it...
[20:17:14] <tissf> HEAD is now at ef2a830... change stepconf to close expander widget
[20:17:15] <tissf> Applying French translation update
[20:19:20] <tissf> jepler: how rewind head ?
[20:21:59] <jepler> try "git push" again
[20:22:03] <jepler> same error?
[20:22:37] <CIA-1> EMC: 03tissf 07master * r7b3f69d8dd6a 10/src/po/fr.po: French translation update
[20:22:53] <tissf> fine no problem this time
[20:23:22] <tissf> jepler: can you explain please ?
[20:23:26] <jepler> if someone else did a commit between your "git pull --rebase" and "git push", then you might get that error
[20:23:44] <jepler> was it more than 2 hours ago that you did "git pull --rebase"?
[20:24:01] <tissf> ~30mn
[20:24:59] <jepler> I'm not sure, then
[20:25:23] <tissf> ok thank you for help jepler :)
[20:25:51] <jepler> thank you for your work and your patience
[20:26:06] <tissf> np
[20:27:19] <tissf> more to learn with git
[21:23:27] <tissf> bed time here, bye to all