#emc-devel | Logs for 2008-09-05

[00:51:51] <CIA-29> EMC: 03rayh 07TRUNK * 10emc2/docs/src/common/Glossary.lyx: Chapter headings to ease appendex setups.
[00:57:35] <CIA-29> EMC: 03rayh 07TRUNK * 10emc2/docs/src/Master_User.lyx: Fix appendex a bit in TOC.
[00:58:11] <CIA-29> EMC: 03rayh 07TRUNK * 10emc2/docs/src/Master_Integrator.lyx: Fix appendex a bit in TOC.
[01:00:42] <jepler> rayh: thanks for your work on the docs
[01:01:32] <rayh> Glad that I found that bit of stuff.
[01:02:22] <rayh> Hope that I didn't break any of the html.
[01:02:47] <rayh> BTW I notice that make clean does not remove any of the man docs. Should it?
[01:07:17] <jepler> yes probably
[01:07:42] <rayh> I'll see if I can fix that. I did something in that makefile a while back.
[01:07:47] <jepler> there is a "docclean" target in docs/src/Submakefile, and when docs are enabled that should be executed when clean is
[01:07:50] <jepler> clean: docclean pythonclean
[01:07:53] <jepler> if it's not working I'm not sure why
[01:08:04] <jepler> if you don't get it, remind me and I'll have a look -- I'm distracted by other stuff at the moment
[01:08:12] <rayh> mp
[01:08:18] <rayh> catch you later.
[01:44:30] <jmkasunich> jepler: other stuff = acme leadscrews?
[02:09:48] <jepler> jmkasunich: not yet -- I have the screw but not the new nuts
[02:10:12] <jmkasunich> two different sources?
[02:10:25] <jepler> yeah. screws from mcmaster, nuts from dumpstercnc
[02:10:34] <jepler> though you're right that I was playing with the machine
[02:11:00] <jmkasunich> I've been looking at PCB mills on youtube and getting ideas
[02:11:09] <jmkasunich> like I need another project
[02:11:24] <jmkasunich> what does the zenbot use for ways?
[02:12:40] <jepler> for each axis it has two 1/2 inch aluminum tubes or rods, and the plastic rides directly on that. http://axis.unpy.net/files/01188441458/img_7050.jpg
[02:16:53] <jepler> anyway, my machine had started stalling, particularly during the last step of homing. I scratched my head about that, then decided to just increase the driver current (I had it at 50% of rated for a long time, and increased it to 70%). Then I started playing to see how much higher that let me set the velocity
[02:17:48] <jepler> .. and I ended up getting an 80% speed increase, from 1.0 in/sec to 1.8in/sec, to my great surprise
[02:18:14] <jmkasunich> those rods are aluminum? they don't look it
[02:18:31] <jmkasunich> and I'd think aluminum would suck for that, unless anodized or hardcoated
[02:18:49] <jepler> I guess I'm not sure what metal they are
[02:19:46] <jmkasunich> is it supposed to run dry (no oil)?
[02:20:01] <jepler> apparently
[02:21:24] <jepler> hm .. this old e-mail from the seller suggested using "a light oil like tri-flow or even wd-40" on them if I thought it was rough -- at certain speeds there is a "honk" sound, which I think comes from the motion between the rails and plastic
[02:21:44] <jmkasunich> sticktion
[02:22:02] <jepler> I tried WD-40 and a silicone spray lubricant but it didn't seem to make a difference
[02:22:38] <jmkasunich> dip it in Vactra #2 ;-)
[02:27:44] <cradek> remind me to give you some
[02:27:55] <jmkasunich> who, jeff?
[02:28:03] <jmkasunich> I already have a gallon
[02:28:23] <jepler> hi cradek
[02:29:32] <cradek> hi
[02:29:34] <cradek> yes, jeff
[02:31:03] <jmkasunich> http://www.youtube.com/watch?v=-gzjw0svMA8&feature=related
[02:54:53] <jepler> 'night folks
[06:51:55] <Lerman_> Lerman_ is now known as Lerman
[09:01:12] <Lerman_> Lerman_ is now known as Lerman
[10:54:31] <micges> while studing changing soft limits in runtime, I saw that It will be very effective to have different soft limit to program and jog
[10:55:22] <micges> while there is different tp to program and jog this is possible
[11:05:20] <alex_joni> interesting
[11:08:30] <BigJohnT> hi alex_joni
[11:15:03] <CIA-29> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/gui/ (axis.lyx tkemc.lyx): fix broken reference
[11:49:49] <alex_joni> BigJohnT: you didn't mail me :)
[11:51:53] <BigJohnT> no but I did create an account on the emc page
[11:58:09] <BigJohnT> bbl time for work
[12:03:58] <rayh> SWPadnos, Got your coffee yet?
[12:09:29] <alex_joni> if I'm not around.. would you guys let BigJohnT know, that I updated his account to "Publisher"
[12:13:19] <rayh> Hi alex_joni
[12:14:12] <alex_joni> hi ray
[12:14:22] <alex_joni> and let him read http://www.scribd.com/doc/2300190/Joomla-v-1-5-Webmaster-and-Publisher-Tutorial
[12:14:30] <alex_joni> how's it going ray?
[12:14:39] <rayh> Good.
[12:14:58] <rayh> You?
[12:18:59] <jepler> micges: I don't immediately understand the value of that. Do you want to jog to locations that it's not OK for the program to go? or the reverse?
[12:21:57] <micges> jepler: I mean program have larger area that jog has
[12:22:46] <jepler> micges: can you tell me why that's desirable?
[12:25:37] <micges> 1: machine have definable "safe" area above the material (can't destroy tool)
[12:27:28] <micges> 2: when have table with specyfic areas (tool change area, anything else change area) all of that must be automatic without chance to jog into them
[12:28:25] <jepler> I see
[12:28:48] <micges> our current machine have this
[12:30:02] <micges> we will do it ourself but it will be very cool that ability of make such of thing will be in future version of emc
[12:30:53] <jepler> if I recall correctly, the NML message is in 2.2. TRUNK only exposes it to Python. I still have no plans to create a user interface for it.
[12:31:29] <jepler> the use I had in mind was to modify the soft limits on a lathe to reflect the distance of the tailstock along "Z"
[12:32:16] <micges> I see
[12:33:32] <micges> safe move area is cool thing in mill that have large speeds
[12:33:39] <rayh> Chuck barriers.
[12:33:55] <rayh> Vice avoidance.
[12:34:02] <micges> rayh: what it is ?
[12:34:22] <rayh> So you don't run a tool into the chuck jaws.
[12:34:23] <jepler> yeah, something more sophisticated than "inside this box OK, outside not OK" would really be much more powerful
[12:34:37] <rayh> Yes it really would.
[12:34:38] <micges> yes
[12:35:39] <rayh> Chuck barriers can include the ability to run a drill or boring bar into the center as well as avoiding the jaws.
[12:38:06] <rayh> And they should include include tool length both in x and z.
[12:38:55] <rayh> But as micges understands, we've got a start at it.
[12:39:07] <alex_joni> as long as you don't define it as boxes :)
[12:39:10] <jepler> there always was an nml message to modify soft limits
[12:39:20] <jepler> I didn't do much, mostly allowed it to be sent from Python code
[12:39:22] <alex_joni> this is also usefull to limits around singularities of kins
[12:39:35] <alex_joni> this = this idea
[12:40:42] <jepler> in that case, one of you guys get to work
[12:40:49] <jepler> your assignment is due next friday
[12:42:57] <rayh> There you go. I could have a drawing of the theory by then ;)
[12:43:59] <jepler> keywords: octree, CSG
[12:44:03] <alex_joni> sounds like a plan
[12:44:10] <alex_joni> rayh does the drawing of the theory
[12:44:18] <jepler> ooh my nuts shipped last night
[12:44:18] <alex_joni> SWPadnos does the pseudocode
[12:44:32] <alex_joni> jepler (with his new nuts) can do bugfixing
[12:44:44] <alex_joni> cojones
[12:44:45] <jepler> nah, I'll just whine that whatever the user interface is, I don't like it
[12:47:09] <alex_joni> hmm.. I bought alibre 3d-cad, and it features a very nice Machinist Toolbox
[12:53:04] <alex_joni> if anyone needs some info regarding machining speeds, etc.. I suspect it's in here
[13:11:59] <fenn> i can't picture CSG describing singularities since they are 6d (they move around depending on rotation)
[13:20:17] <rayh> singularities are about the worst case
[13:34:46] <Lerman> I'm still having a problem with TRUNK.
[13:34:56] <Lerman> Anyone available to talk?
[13:40:39] <cradek> I would very happily use a negative Z limit based on the loaded tool. It wouldn't have to be any more complex than that.
[13:41:19] <jepler> Lerman: the limits error?
[13:41:28] <jepler> Lerman: you pastebinned the gcode, maybe we should take a look at your inifile too
[13:42:15] <Lerman> Yup. I just ran a one line program (well, a few more) that failed. The debug msg: Issuing emc traj linear move displays the move that causes the error.
[13:42:39] <Lerman> Are the coords in that message absolute (machine) or relative coords?
[13:42:46] <cradek> machine
[13:43:13] <Lerman> Well, then the problem is that the g54 offsets are not being applied.
[13:43:36] <Lerman> What might cause that?
[13:43:48] <cradek> when you touch off, do the program preview and axes move?
[13:43:52] <Lerman> I seem to remember something about the "var" file.
[13:44:07] <cradek> yes, offsets won't work if the var file isn't writable
[13:44:15] <Lerman> They did (now they are set so they won't move again).
[13:44:30] <Lerman> -rw-r--r-- 1 lerman lerman 1526 2008-09-05 09:31 univpwm.var
[13:44:59] <cradek> when you touch off, do the program preview and axes move?
[13:45:33] <Lerman> They did move. I'll go try again with a different offset. I'll be back in a minute.
[13:46:52] <jepler> it might be useful to bring that config to a net connected computer and change the hal files so that it runs as sim..
[13:47:09] <jepler> since this particular problem seems like it would demonstrate perfectly well without the hardware around
[13:47:47] <Lerman> Yes. Setting the touch off to 1.0 moved. As did setting it back to zero.
[13:49:59] <cradek> this is very puzzling because I *know* trunk works. I cut several things last night by moving G54 around.
[13:50:22] <cradek> so we don't have many clues about what makes your setup different. You have to get us those clues.
[13:50:41] <cradek> at least a tar of your entire config, preferably modified so it still loads without the ppmc hardware
[13:50:47] <Lerman> Yup. What would be most useful.
[13:51:16] <cradek> we already have the gcode but it doesn't sound like that's even relevant
[13:51:17] <SWPadnos> do you have a PPMC cage or a USC/UPC?
[13:51:30] <Lerman> usc/upc
[13:51:51] <SWPadnos> ok. I can hook up mine and test the complete config, if it comes to that
[13:52:04] <Lerman> %
[13:52:06] <Lerman> g1 f5.0 y1 x-.1
[13:52:08] <Lerman> %
[13:52:20] <Lerman> That is the simplest program that fails.
[13:54:26] <Lerman> What should I see in the univpwm.var file? Everything is zero except for 5220 which is 1.000...
[13:54:47] <cradek> then there are no g54 offsets. they should be in there.
[13:55:12] <cradek> [RS274NGC]
[13:55:12] <cradek> PARAMETER_FILE = emc.var
[13:55:21] <cradek> check this setting to see if it's using the var you think it is
[13:55:46] <SWPadnos> they're only guaranteed to be there if you've exited from EMC with G54 active though, I think
[13:56:03] <SWPadnos> err - with G54 offsets set anyway
[13:56:05] <cradek> that is incorrect
[13:56:11] <Lerman> Which vars contain g54 offsets? It is using univpwm.var
[13:56:11] <cradek> ok that's more correct
[13:58:24] <Lerman> If 5221... contains the G54 offsets then they are set to zero.
[13:58:32] <Lerman> In the file I'm using.
[13:59:24] <Lerman> That file does seem to get touched at least, so that means I'm using the correct univpwm.var (not one in another directory).
[13:59:26] <cradek> strange, I can't find that answer in the docs, even though I know it's there
[13:59:34] <SWPadnos> http://www.linuxcnc.org/docs/devel/html/common_machining_center.html#cap:Parameters-Used-by
[14:00:07] <cradek> well there you go.
[14:00:15] <Lerman> Also page 59 of the user manual.
[14:00:19] <SWPadnos> just found it myself :)
[14:00:41] <cradek> Lerman: can you tar up your entire config directory and get it to us?
[14:01:41] <Lerman> Yes. Working.
[14:01:53] <SWPadnos> huh: http://forums.reprap.org/read.php?1,12143,page=2
[14:02:08] <SWPadnos> AXIS running an arduino board
[14:02:53] <SWPadnos> (for something - I haven't read the whole thread)
[14:04:06] <Lerman> http://www.se-ltd.com/~lerman/files/ -- see the file univpwm.tar
[14:10:53] <jepler> I don't see the problem :(
[14:11:17] <Lerman> But do you see the solution? :-)
[14:11:36] <jepler> no, I mean it works for me
[14:11:57] <jepler> I copied core_sim.hal from configs/sim and modified the inifile like this: http://pastebin.ca/1194853
[14:12:13] <jepler> I put your 3-line program with %-signs in a gcode file
[14:12:26] <jepler> start emc, hit F1 F2 X HOME Y HOME Z HOME to home all axes
[14:12:45] <jepler> while at machine origin I touch off X -1 Y -1 Z 1
[14:12:48] <jepler> load program, hit run
[14:13:16] <jepler> no warnings from axis or emc, though because Z is not specified in the program the preview and the backplot don't match up
[14:13:42] <Lerman> Well, I'm touching off by homing. Then jogging to where I want to be the origin. Then touching off zero.
[14:13:57] <cradek> what is touching off by homing?
[14:14:17] <cradek> explain what you do, please
[14:14:34] <Lerman> Sorry. First I home. Then I jog to where I want zero to be. Then I touch off with the value zero in the screen.
[14:14:47] <cradek> oh, I see what you mean
[14:15:04] <cradek> jepler: the var file gets written for you, right after you do the touch off?
[14:15:04] <Lerman> When done with that, X, Y, Z show as 0,0,0.
[14:15:34] <jepler> cradek: yes
[14:15:36] <jepler> $ grep -v '0\.00000' univpwm.var
[14:15:36] <jepler> 52201.000000
[14:15:36] <jepler> 52211.000000
[14:15:36] <jepler> 52221.000000
[14:15:38] <jepler> 5223-1.000000
[14:15:59] <cradek> Lerman: check for that behavior on yours?
[14:16:05] <Lerman> So, the values I'm giving are all zeros.
[14:16:23] <cradek> ok, that shouldn't matter, but jepler could try it that way too for grins
[14:16:26] <Lerman> OK. I'll go touch off 1.0 1.0 1.0 and see what I get.
[14:16:38] <cradek> no hang on.
[14:16:50] <cradek> jepler did touch off -1,-1,-1, so he got +1 in the var file
[14:17:06] <cradek> if you jog to +1 and touch off 0, you'll also get +1
[14:17:28] <cradek> shouldn't matter
[14:17:38] <cradek> but I still suspect the var file is not being written
[14:19:47] <Lerman> OK. I touched off in the same position with X of 0 and Y of 0. The .var file has been changed so that 5221 shows -1.000... and 5222 also shows -1.000...
[14:20:22] <Lerman> Note that these values are negative. Also note that the table is about 18" from the X zero position.
[14:20:52] <Lerman> Ooops correction above. I touched off with X of +1.0 and Y of +1.0.
[14:22:05] <cradek> so you home X, move to machine's X18, touch off X +1, and you get 5221 = -1?
[14:22:09] <Lerman> So that I would think the 5221 (X) value that should be stored would be -18.0 (or possibly +18.0) depending on conventions.
[14:22:20] <Lerman> Yes.
[14:22:28] <cradek> yeah, or -17 or -19 or +17 or +19 or something like that
[14:23:30] <cradek> turn on task issue debug again, then when you do the touch off, you can see the MDI command that AXIS is sending
[14:23:43] <cradek> also the TRAJ_SET_ORIGIN nml message
[14:23:44] <Lerman> BRB.
[14:23:49] <cradek> see which is wrong
[14:24:28] <jepler> Issuing EMC_TASK_PLAN_EXECUTE -- (+509,+268, +59,G10\032L2\032P1\032x[0.907059996639-[1.000000*[0.0]]]\n,)
[14:24:31] <jepler> Issuing EMC_TRAJ_SET_ORIGIN -- (+224,+84, +0,0.907060,0.825733,-0.874133,0.000000,0.000000,0.000000,0.000000,0.000000,0.000000,)
[14:24:37] <jepler> ...
[14:24:39] <jepler> Issuing EMC_TRAJ_LINEAR_MOVE -- (+220,+116, +0,0.807060,1.825733,-0.874133,0.000000,0.000000,0.000000,0.000000,0.000000,0.000000, +2,0.083333,1.500000,6.000000, +0,)
[14:25:05] <jepler> (this is using the X HOME Y HOME Z HOME jogjogjog X END 0 Y END 0 Z END 0 method)
[14:26:24] <cradek> when I move to machine 5 (g0g53x5) I get the touch-off mdi command x[5.000000000000-[1.000000*[0.0]]]
[14:26:46] <cradek> then Issuing EMC_TRAJ_SET_ORIGIN -- (+224,+84, +0,5.000000,0.000000,0.000000,0.000000,0.000000,0.000000,0.000000,0.000000,0.000000,)
[14:27:02] <cradek> wonder what is different over there.
[14:29:07] <Lerman> OK. traj_set_origin ... -1.234 (I used +1.234)
[14:29:50] <Lerman> task_plan_execute... G10 L2 P1[0.00 - [1.0000 * [1.234]]]
[14:30:36] <Lerman> (with some interspersed \032's and some addition insignificant zero digits)
[14:31:47] <cradek> the first 0.00 should be the current axis position in machine coordinates
[14:32:33] <Lerman> And clearly it is NOT.
[14:33:26] <cradek> but with jeff running your config it is...
[14:34:04] <Lerman> I'm going back out to tell axis to display the current machine coords instead of relative coords to see what it says. Also, I'll try jogging in a negative cirection and see if that works.
[14:34:24] <cradek> I would say next that you're not running the same code, but this hasn't changed for moons and moons
[14:34:38] <cradek> good idea.
[14:38:38] <Lerman> Axis was displaying zero for the machine coordinates. But I was able to jog in a negative direction. (Does jogging obey soft limits?)
[14:38:56] <jepler> after homing, jogging obeys soft limits
[14:39:10] <Lerman> Then I homed the X axis again. Jogged back about 15 inches. And everything seems fine for that axis.
[14:39:31] <jepler> is it a transcription error that your task_plan_execute line didn't have an axis letter after P1?
[14:39:39] <Lerman> Well, the axis had been homed. Hold on BRB.
[14:40:15] <cradek> what did you mean "seems fine"? did it show correct machine coordinates after you homed again?
[14:40:40] <cradek> because a zero machine coordinate after jogging around does not sound right
[14:41:40] <Lerman> Sorry about wasting your time. Apparently, I had NOT homed the axis. So, it took the position when I started the machine as zero, zero, zero.
[14:42:15] <Lerman> Sorry, sorry, sorry. I owe you guys a slice of pie at the next fest. Maybe even a slice each. Or a whole pie.
[14:43:12] <cradek> I'm still recovering from the 7 pies you bought at the last one
[14:43:24] <cradek> mmm pie
[14:43:53] <Lerman> That's how legends are created. Two pies becomes seven. We couldn't drink the tap water, so we drank...
[14:43:55] <jepler> cradek: yeah, I was thinking that for "lunch" we need to eat another bagel at 11:30, then head to stauffer's and order just pie and coffee
[14:44:19] <cradek> Lerman: was it only 7? I remember 9
[14:53:07] <cradek> jepler: is friday apple dumpling day?
[14:53:16] <jepler> cradek: it often is
[14:53:20] <jepler> I'm not sure it's as systematic as that
[14:53:42] <cradek> oh maybe I was a fool for looking for a pattern
[14:55:22] <cradek> on a lathe, you (I?) always want touch-off of X to change the tool table. I never want a g54 X offset
[14:56:45] <jepler> you mean you want a tool table editor?
[14:56:48] <cradek> or maybe not touch off, but editing the X offset in the tool table is something you do constantly
[14:56:59] <cradek> yes maybe that's what I mean
[14:57:08] <cradek> I'm not sure what I mean
[14:57:24] <cradek> I'm a "user story"
[14:57:33] <cradek> I should have a funny name
[14:57:41] <SWPadnos> you do
[14:57:45] <cradek> oh right
[14:57:56] <SWPadnos> welcome to the club
[14:58:31] <SWPadnos> heh - everyone here has a "funny name"
[14:58:48] <SWPadnos> (except maybe Ray)
[14:59:01] <SWPadnos> bbiab
[15:04:57] <cradek> but I don't really want to edit the tool table, just like I don't want to edit a list of offsets. I know tool 3 cut the diameter 0.0014 oversize, so I need to make the tool [longer|shorter] by [.0014|.0007], which is tedious and easy to screw up
[15:05:39] <cradek> since it is currently 1.3426 and I can't do math
[15:33:11] <rayh> Some of the later controls on the Hardinge had two pages of overlapping offsets.
[15:33:27] <rayh> One the measured offset the other an offset from that offset.
[15:36:33] <cradek> rayh: is that what they call a tool wear offset?
[15:39:09] <rayh> That would be one way to name it.
[15:59:21] <cradek> so, in summary, someone should come up with a coherent design for what I want, and then implement it
[15:59:49] <SWPadnos> s/should/could/
[15:59:59] <SWPadnos> :)
[16:07:51] <cradek> uh
[18:48:57] <CIA-29> EMC: 03cradek 07TRUNK * 10emc2/docs/html/gcode.html: make this more useful for canned cycles
[19:08:01] <CIA-29> EMC: 03cradek 07TRUNK * 10emc2/docs/html/gcode.html: g28.1, g30.1 are new in trunk
[19:17:52] <micges> what are they do ?
[19:18:06] <micges> (g28.1)?
[19:18:15] <cradek> set the reference points
[21:02:53] <jepler> hey seb_kuzminsky
[21:03:28] <jepler> seb_kuzminsky: aren't you somewhere in the denver/boulder area? I am going to be travelling to denver at the end of the month.
[21:05:51] <seb_kuzminsky> hi jeff, yep i'm in boulder, we should get together for a beer or something
[21:06:04] <jepler> yeah I was thinking something like that would be nice
[21:06:14] <seb_kuzminsky> what days are you free?
[21:07:32] <jepler> the driving days are the 25th and 28th (thursday and sunday), and I don't have specific plans for the other two days -- my girlfriend and I are just going to visit some friends in the area
[21:08:21] <seb_kuzminsky> i'm free that whole window, so whenever you've got time let's get together for a bit
[21:08:24] <jepler> ok
[21:08:39] <seb_kuzminsky> you're going to be in denver? boulder's about 30-60 minutes north of there, depending on where in denver you are
[21:08:57] <jepler> yeah -- we'll have a car though
[21:09:16] <jepler> one set of friends is in a denver suburb, the other lives in the mountains 15 minutes outside of boulder
[21:09:34] <seb_kuzminsky> do you want to go for a hike in the foothills? there's some gorgeous spots just above boulder
[21:09:42] <seb_kuzminsky> or drinking's always fine by me too :-)
[21:10:06] <jepler> one followed by the other is even better
[21:10:08] <jepler> it depends how much time you have
[21:10:11] <BigJohnT> I'll drive the beer cart :)
[21:11:25] <seb_kuzminsky> i'm always free in the evenings after about 8, and i can probably swing an afternoon + evening if we plan it out
[21:11:36] <jepler> ok -- I wanna get out of the office at the moment, we can set the details up later
[21:12:12] <jepler> if I had to guess, I'd say saturday for boulder, but there are too many things up in the air at the moment
[21:13:07] <seb_kuzminsky> saturday would be great, let's figure out the details via email
[21:13:28] <jepler> ok, talk to you later
[21:15:16] <seb_kuzminsky> later
[21:16:09] <seb_kuzminsky> BigJohnT: you'll drive the sagwagon? ;-)
[21:18:33] <BigJohnT> if that means the beer cart yes :)
[21:19:02] <seb_kuzminsky> where do you live?
[21:19:21] <BigJohnT> LOL, Poplar Bluff MO
[21:19:36] <seb_kuzminsky> better start driving :-)
[21:20:07] <BigJohnT> I am but it is to the house and the easy chair :)
[21:20:45] <BigJohnT> * BigJohnT looked up sagwagon :)
[21:22:29] <BigJohnT> see you guys later