#emc-devel | Logs for 2009-01-31

[00:06:17] <jmkasunich> alex_joni: got? more like "is getting"
[00:06:20] <jmkasunich> or did he give up?
[00:11:49] <alex_joni> nope, he's pretty persistent
[00:12:06] <alex_joni> jmkasunich: did you see micges's patch for changing motion controller params to pins?
[00:12:15] <jmkasunich> assuming that he is in the US, the timing of his mails makes me think maybe he is doing this for work
[00:12:19] <jmkasunich> no
[00:12:29] <alex_joni> that was on the devel list iirc
[00:12:34] <jmkasunich> I saw it go by, didn't look at it
[00:13:07] <alex_joni> ah, ok.. was wondering if you would
[00:13:15] <alex_joni> last time I poked it went tits up
[00:13:23] <jmkasunich> I've been neglecting EMC lately
[00:14:53] <alex_joni> busy with that robot ?
[00:15:15] <jmkasunich> yeah, and a maching project that it is too cold to work on
[00:15:24] <jmkasunich> both projects have deadlines
[00:15:32] <jmkasunich> so I'm a bit frazzled
[00:15:41] <jmkasunich> s/maching/machining
[00:17:12] <alex_joni> heh, I even read that as machining
[03:10:25] <cradek> one more day for features...
[03:10:40] <jmkasunich> heh
[03:11:01] <jmkasunich> I'm writing a weird logging library
[03:13:17] <jmkasunich> I want to be able to log images
[03:13:43] <jmkasunich> I decided the easiest way to view images in a log is to write the log in HTML, and use a browser to view it
[04:35:02] <cradek> jmkasunich: what is this? http://www.hgrindustrialsurplus.com/sub/product_detail.aspx?id=12-403-428&searchtable=2&sortExpression=&SortASC=&pageSize=50&currentPageIndex=0&searchNAP=
[04:35:48] <jmkasunich> no idea
[04:35:56] <jmkasunich> tapping head?
[04:36:01] <jmkasunich> coolant inducer?
[04:36:20] <jmkasunich> does google say anything about su-matic?
[04:36:25] <cradek> it's a company
[04:36:32] <jmkasunich> no shit
[04:36:35] <jmkasunich> what do they make
[04:36:39] <jmkasunich> ;-)
[04:37:05] <cradek> Multiply the capacity
[04:37:06] <cradek> of your machine without any big investments
[04:37:15] <jmkasunich> ah, they make spam
[04:38:00] <cradek> I wonder if it's a speed multiplier thing
[04:38:14] <cradek> it has a stationary part
[04:38:49] <jmkasunich> could be
[04:39:04] <cradek> seems like they make 'spindle heads'
[04:39:55] <cradek> sure looks like qc30 to me
[04:39:56] <jmkasunich> that might be an increaser
[04:40:03] <cradek> tomorrow's not hgr day is it?
[04:40:09] <jmkasunich> every saturday is
[04:40:18] <jmkasunich> I might go, haven't decided yet
[04:40:21] <cradek> oh right!
[04:40:27] <cradek> would you look if you do, please?
[04:40:47] <cradek> I'd kill for one of those speed increasing dealies
[04:40:47] <jmkasunich> ok
[04:42:59] <jmkasunich> http://jmkasunich.com/pics/mach-vision/mazelog-000.html
[04:43:29] <jmkasunich> this is gonna be very handy as I get deeper into my imaging pipeline
[04:43:43] <cradek> wow
[04:44:24] <jmkasunich> various things can be printed or not, using a mask
[04:44:25] <cradek> that mirror above is a great design
[04:45:19] <SWPadnos> huh: http://www.hgrindustrialsurplus.com/sub/product_detail.aspx?id=12-354-271&searchtable=2&sortExpression=&SortASC=&pageSize=50&currentPageIndex=1&searchNAP=
[04:45:35] <jmkasunich> tomorrow I should start doing some real processing - I have a mask image, I need to read it in and apply it
[04:46:02] <jmkasunich> SWPadnos: a deal, if all the hardware is there, and you need such a thing
[04:46:05] <cradek> what is a mask image?
[04:46:08] <SWPadnos> heh
[04:46:08] <jmkasunich> bet its $30
[04:46:25] <SWPadnos> not from Steelcase, but maybe something similar
[04:46:31] <jmkasunich> cradek: I took the raw image into gimp, colored everything that isn't floor black
[04:46:45] <cradek> ah, the supports and stuff
[04:46:46] <jmkasunich> so it will mask out the camera, board, support struts, etc
[04:46:51] <cradek> neat
[04:47:04] <SWPadnos> hmmm
[04:47:42] <SWPadnos> unless your maze wall detection algorithm can handle voids in the edge lines, you may want to "reconstruct" lines that get broken by the mast (specifically where the struts are)
[04:48:07] <jmkasunich> the algorithms work by matching image pixels to "ought to be there" map pixels
[04:48:26] <jmkasunich> pixels in either image that are masked don't contribute
[04:48:40] <jmkasunich> ones that match contribute positive, ones that don't match contribute negatuve
[04:48:42] <jmkasunich> ive
[04:48:53] <jmkasunich> a positive peak indicates a match
[04:49:18] <jmkasunich> I'll probably have to come up with some heurestics (sp?) to decide what consitiutes a "peak"
[04:49:33] <SWPadnos> ok, a masked convolution of two images (more or less)
[04:49:39] <jmkasunich> yes
[04:49:56] <jmkasunich> the images will be converted from normal grayscale to signed
[04:50:08] <jmkasunich> black = -127, white = +127, mask = 0
[04:50:23] <jmkasunich> the do a pixel-by-pixel multiply and sum
[04:50:36] <jmkasunich> then
[04:55:04] <jmkasunich> SWPadnos: was there any reason you called my attention to that desk part?
[04:55:23] <SWPadnos> well, if you go, can you take a look at it for me?
[04:55:39] <jmkasunich> sure - you got a keyboard you need to mount?
[04:56:15] <SWPadnos> that looks like something I saw at Verizon Wireless - a nice way of storing a keyboard out of the way until you need it
[04:56:28] <SWPadnos> so I think my mother, wife, or machines could use it, yes :)
[04:56:34] <jmkasunich> it doesn't seem to have the slide-y part
[04:56:41] <jmkasunich> unless that is in the lower half of the box
[04:56:49] <SWPadnos> it's more of an articulated arm, not a drawer thing
[04:56:59] <jmkasunich> oh
[04:57:12] <jmkasunich> I guess I'll have to look at it, and then call you to make sure it is what you have in mind
[04:57:21] <jmkasunich> shipping it will cost more than buying it
[04:57:43] <SWPadnos> it shouldn't be toooooooo heavy, so UPS ground may be bearable
[04:57:49] <SWPadnos> but it's sure to be more than $5
[04:58:16] <jmkasunich> I should get some sleep - I need to take the cat to the vet again tomorrow, and I didn't make an appt
[04:58:29] <jmkasunich> that will be the determining factor on HGR - when I get up, and when I get an appt
[04:58:39] <SWPadnos> ugh. good luck
[04:58:50] <jmkasunich> good night
[05:47:20] <KimK_> KimK_ is now known as KimK
[06:10:50] <CIA-2> EMC: 03seb 07TRUNK * 10emc2/src/hal/drivers/mesa-hostmot2/ (TODO encoder.c hostmot2.h): I think this fixes the encoder velocity bug
[10:31:27] <micges> alex_joni: joints "home_state" can't be change to pin of some reason, typecasting it in control.c has no effect
[10:31:49] <micges> have no idea why
[10:41:00] <micges> bbl
[10:46:47] <micges> max_velocity slider isn't working after last fixing
[10:47:33] <micges> in sim_mm when you set it to 33 then its working, and when you set it to 58 it's moving to max
[10:47:58] <micges> (trunk)
[10:50:13] <micges> now it's setting max_vel on moving mouse over it (strange)
[11:59:12] <CIA-2> EMC: 03alex_joni 07TRUNK * 10emc2/configs/SherlineLathe/.cvsignore: silencing
[11:59:47] <CIA-2> EMC: 03alex_joni 07TRUNK * 10emc2/docs/man/.cvsignore: silencing
[12:07:06] <alex_joni> micges: I can confirm there is something screwy with max_velocity
[12:07:24] <alex_joni> running sim-axis-mm here
[12:07:55] <alex_joni> I start the splash g-code, then move the max_velocity to 58 (as a value), leave the slider area and the slider jumps to max right, and it reads 57.9
[12:08:23] <alex_joni> (actually it happens for any speed)
[12:08:51] <alex_joni> same effect on sim/axis
[12:09:55] <micges> yes that is what I saw too
[12:13:12] <BigJohnT> seb_kuzminsky: you awake?
[12:23:29] <micges> bbl
[13:46:06] <BigJohnT> how do I get the "pre-2.3 CVS HEAD" to go away so I can take some screen shots of AXIS?
[13:48:10] <alex_joni> BigJohnT: change emc2/VERSION
[13:48:23] <BigJohnT> ok thanks
[13:49:51] <alex_joni> you need to rerun ./configure though
[13:49:58] <BigJohnT> :)
[13:50:11] <alex_joni> maybe even regenerate ./configure (using autoconf)
[13:50:34] <BigJohnT> let you know in a minute
[13:51:24] <BigJohnT> or three...
[13:52:06] <BigJohnT> nope running configure and make didn't do the trick
[13:54:31] <alex_joni> autoconf first (that regenerates ./configure)
[13:54:34] <alex_joni> then run ./configure
[13:54:36] <alex_joni> then make
[13:54:55] <BigJohnT> do I just do autoconf from my /emc2-trunk directory
[13:55:24] <alex_joni> you need to be in src
[13:55:40] <alex_joni> and need to install autoconf, if it's not there
[13:55:40] <BigJohnT> ok
[13:56:06] <BigJohnT> it is there I was reading the man page and now it is done
[14:00:26] <BigJohnT> thanks alex_joni
[14:04:29] <alex_joni> is it working?
[14:08:07] <BigJohnT> yes
[14:24:35] <cradek> yay, seb fixed the encoder thing!
[14:26:25] <alex_joni> yay 2.3.x
[14:27:11] <cradek> yeah 2.3 is going to be great
[14:37:45] <seb_kuzminsky> morning guys
[14:37:54] <seb_kuzminsky> i *think* the encoder stuff is fixed
[14:38:11] <BigJohnT> morning
[14:38:19] <seb_kuzminsky> i've got some docs on the encoder algorithm that i wanna clean up and check in
[14:38:25] <seb_kuzminsky> hi BigJohnT :-)
[14:39:32] <BigJohnT> seb_kuzminsky: on the mailing list you mention a PIN file... where is it located or is it created when you load hostmot2?
[14:40:30] <alex_joni> BigJohnT: no, it's supplied - not created
[14:41:00] <alex_joni> look in src/hal/drivers/mesa-hostmot2/firmware/*/
[14:41:02] <seb_kuzminsky> it's next to the .BIT files in firmwares/<BoardType>/...
[14:41:08] <seb_kuzminsky> heh what alex said
[14:41:16] <seb_kuzminsky> bbl, breakfast
[14:41:36] <alex_joni> (for a source dir.. it gets installed in the firmware folder ..)
[14:42:21] <BigJohnT> ok thanks
[14:42:37] <alex_joni> np
[14:42:40] <BigJohnT> bbl gotta take one of the cats to the vet
[14:52:06] <micges> alex_joni: is it hard to implement current tool number as a interpreter variable ?
[14:54:09] <alex_joni> do you want to read it? or write it?
[14:55:23] <micges> write it
[14:55:32] <alex_joni> yes, it's hard
[14:56:33] <micges> I've looked how to current origin system was read/write as a paramter and I've lost
[14:57:01] <alex_joni> micges: the problem with interpreter variables is that it's very problematic because of lookahead
[14:57:13] <alex_joni> so if you write #xxxx = 12 in your program
[14:57:24] <alex_joni> then you can do T#xxxx M6 to use it
[14:57:36] <alex_joni> but the interpreter is way ahead of the current execution
[14:57:50] <alex_joni> so when #xxx holds 12, your current tool might be some other value
[14:59:52] <micges> oh
[15:00:15] <micges> I forgot about 1000 blocks lookahead
[15:03:32] <micges> as you said this it would be very usefull have acces to canon variables from g-code (like current tool, current prepared tool, x/y/z min and max limit)
[15:03:44] <alex_joni> yes, but what means current?
[15:04:01] <alex_joni> that's why I stopped implementing this.. it's not easy to get it right..
[15:04:50] <micges> current problem:
[15:05:03] <micges> I have tool changer with 8 position of tools
[15:05:42] <micges> can't handle this from iocontrol so then I've hardcoded all tool change to gcode
[15:06:35] <micges> and numbers of tool_in_spindle are available thru M66 and so
[15:06:50] <micges> and all tool change is like this:
[15:08:11] <micges> IF TOOL_IN_SPINDLE != NEXT_TOOL:
[15:08:11] <micges> IF TOOL_IN_SPINDLE != 0:
[15:08:11] <micges> PUT DOWN CUREENT_TOOL
[15:08:11] <micges> GET NEXT_TOOL
[15:08:11] <micges>
[15:08:38] <micges> this is whole idea (more less )
[15:08:49] <alex_joni> right
[15:13:39] <CIA-2> EMC: 03alex_joni 07TRUNK * 10emc2/src/emc/kinematics/tripodkins.c: fix writing to variables only after they have been allocated
[15:14:25] <CIA-2> EMC: 03alex_joni 07TRUNK * 10emc2/configs/sim/tripodsim.hal: remove SHMEM_KEY fromt he hal file
[15:32:25] <CIA-2> EMC: 03alex_joni 07TRUNK * 10emc2/src/emc/kinematics/scarakins.c: fix writing to variables only after they have been allocated
[15:37:26] <cradek> tool change really should be done outside the interpreter. with the current architecture of emc2 I think it's a mistake to write a tool change in the gcode
[15:38:11] <alex_joni> cradek: might be the only option for rack-type toolchangers atm
[15:38:36] <cradek> well personally I'd code that in C in CHANGE_TOOL at the canon level
[15:39:09] <alex_joni> yeah, I was talking about a non-recompiling method
[15:39:38] <cradek> yeah that would not be my first concern
[15:41:29] <micges> cradek: that is good idea to do in canon
[15:42:32] <micges> I've done it in a way that I'm sure its works
[15:45:27] <cradek> bbl
[15:48:48] <CIA-2> EMC: 03alex_joni 07TRUNK * 10emc2/configs/demo_mazak/demo_mazak.hal: remove reference to SHMEM_KEY, not used anymore
[15:50:01] <CIA-2> EMC: 03alex_joni 07TRUNK * 10emc2/configs/plasma-thc/kinematics.hal: remove reference to SHMEM_KEY, not used anymore
[15:51:04] <CIA-2> EMC: 03alex_joni 07TRUNK * 10emc2/configs/plasma-thc-sim/kinematics.hal: remove reference to SHMEM_KEY, not used anymore
[15:54:49] <CIA-2> EMC: 03alex_joni 07TRUNK * 10emc2/configs/Sherline4Axis/stepper_xyza.hal: remove reference to SHMEM_KEY, not used anymore
[15:55:36] <CIA-2> EMC: 03alex_joni 07TRUNK * 10emc2/configs/sim/sim-xyyz.hal: remove reference to SHMEM_KEY, not used anymore
[15:56:24] <CIA-2> EMC: 03alex_joni 07TRUNK * 10emc2/configs/stepper-gantry/kinematics.hal: remove reference to SHMEM_KEY, not used anymore
[15:58:12] <CIA-2> EMC: 03alex_joni 07TRUNK * 10emc2/configs/classicladder/cl-estop/ (cl-estop.hal cl-estop.ini): remove reference to SHMEM_KEY, not used anymore
[16:02:07] <jepler> oops, I must not have done a very thorough job on that
[16:02:42] <alex_joni> ah, don't sweat it..
[16:02:56] <alex_joni> those were some custom hal files, probably very few people use them
[16:32:47] <mOizt> What is generally considered a max stepper rpm?
[16:33:12] <BigJohnT> depends on the stepper and the power supply and the driver circuit
[16:33:34] <BigJohnT> and the step generator
[16:34:01] <mOizt> but i.e with 34 size steppers..
[16:34:12] <mOizt> with general "low-cost" drivers
[16:37:36] <BigJohnT> I'd guess about 600 -800
[16:37:48] <BigJohnT> or less
[16:38:17] <mOizt> ok..
[16:40:41] <mOizt> Know anyone who made coil winding machine using emc?
[16:41:51] <BigJohnT> http://web3.automationdirect.com/static/specs/surestepmotors.pdf
[17:13:43] <CIA-2> EMC: 03cradek 07TRUNK * 10emc2/share/axis/tcl/axis.tcl: this made it worse; reverting
[17:16:33] <BigJohnT> cvs needs an "awwshit" button
[17:19:22] <alex_joni> right next to the "heh" key on my keyboard
[17:20:11] <cradek> is that next to "jepler would you fix it please"?
[17:20:16] <BigJohnT> :)
[17:20:50] <BigJohnT> you guys have all the cool keys
[17:21:58] <alex_joni> cradek: on my keyboard it's next to "jepler would you .... oh, n/m, you already did"
[17:23:07] <jepler> cradek: grngrngnrg
[17:25:52] <KimK> Is there a checklist of common errors that can "blank" AXIS? I'm writing (fiddling with?) a short g-code program in sim mode, and I can't see anything in the AXIS window.
[17:28:25] <alex_joni> the preview?
[17:28:40] <alex_joni> KimK: try running glxgears and see if that displays anything
[17:29:15] <BigJohnT> alex_joni: We have ignition! I can see one of the ubuntu machines from this ubuntu machine
[17:29:29] <KimK> glxgears works OK here
[17:31:16] <KimK> Yes, I don't see any AXIS preview, or anything while AXIS is running. I can see it running through the g-code (g-code window), but nothing in the preview window, except the XYZ numbers are changing.
[17:32:24] <jepler> KimK: there are options in the "View" menu that allow most elements of the display to be turned off
[17:32:40] <KimK> OK, thanks, let me check...
[17:36:23] <SWPadnos> On the "CNC Toolkit" group, there's a discussion of motion of rotary axes
[17:37:15] <KimK> Most are checked (purple), except for "Distance to Go". I played with commanded/actual and machine/relative, no change. I wonder what I did to cause this?
[17:37:20] <SWPadnos> apparently, at the fest at NAMES 2003, there was a discussion about a method suggested by Rab Gordon for setting rotary movement speeds (summary here, search for "Gordon": http://www.linuxcnc.org/content/view/16/12/lang,en/)
[17:37:45] <SWPadnos> KimK, does the "splash sscreen" display?
[17:37:52] <SWPadnos> -a
[17:37:54] <SWPadnos> -s
[17:38:39] <KimK> Good timing, SWP, I was thinking I should restart EMC2/AXIS anyway, brb...
[17:39:22] <SWPadnos> rab asked me if his suggestion had been implemented in EMC2, which I don't think is the case
[17:39:28] <CIA-2> EMC: 03cmorley 07TRUNK * 10emc2/src/hal/classicladder/calc.c: Change old timer to use new control pin. Ladder programs that used the old timers need to connect the 'E' and new 'C' pins together so timer works as before
[17:39:31] <jepler> cradek: on further reflection, I think this was what I intended before:
[17:39:31] <jepler> - if {$vel == $max_speed} { return 1 }
[17:39:34] <jepler> + if {$vel >= 60*$max_speed} { return 1 }
[17:39:42] <SWPadnos> (using correct kins would be the EMC2 solution I think)
[17:40:02] <jepler> hm no
[17:41:28] <jepler> SWPadnos: no, it's not in emc2
[17:41:42] <jepler> we do what the documentation says, and the documentation says something different than that
[17:41:43] <SWPadnos> thanks. that's what I thought :)
[17:41:47] <SWPadnos> heh
[17:44:39] <KimK> OK, I Restarted EMC2/AXIS and now the problem is gone, AXIS is displaying normally. Thanks to all for helping, I'll let you know if it comes back.
[17:45:05] <SWPadnos> is the g-code that was used to cut the knobs at CNC workshop (in the MAX 5-axis youtube video) available online?
[17:45:38] <SWPadnos> KimK, is it possible that you were zoomed in so far or panned so far that there was just nothing visible in the viewport?
[17:45:56] <SWPadnos> (or zoomed out for that matter)
[17:47:40] <KimK> SWPadnos: Yes, it is certainly possible, but I tried zooming out (maybe not enough?). Maybe there should be an "auto-center-and zoom" button on AXIS? (A new feature! And whaddya know, it's still only Jan 31st ;)
[17:47:51] <SWPadnos> there is
[17:48:06] <CIA-2> EMC: 03jepler 07TRUNK * 10emc2/src/emc/usr_intf/axis/scripts/axis.py: initial maxvel setting was off by a factor of 60
[17:48:09] <jepler> I use "v" for that kind of thing
[17:48:13] <SWPadnos> hit any of the view buttons and it should re-denter ans set scaling so the gcode fills the window
[17:48:14] <jepler> but it also cycles the view
[17:48:16] <KimK> Really? Where is it?
[17:48:23] <SWPadnos> those Z Y X P buttons
[17:48:28] <SWPadnos> (P=perspective)
[17:48:37] <SWPadnos> above the 3D view
[17:49:09] <KimK> Oh, I thought those just set the view, I didn't notice they were auto-center-and-zoom too. Nice! (Boy, you guys are fast, LOL)
[17:49:14] <CIA-2> EMC: 03jepler 07TRUNK * 10emc2/share/axis/tcl/axis.tcl: vels above max_speed should be clipped
[17:49:23] <jepler> cradek: let me know if I've fixed this enough times yet
[17:49:41] <SWPadnos> if you pressed any of those when you had the problem of not seeing anything, then zoom/pan wasn't the problem :)
[17:50:23] <KimK> At this point, I'm not sure what I did. Sorry.
[17:53:05] <KimK> Now back to my original problem of getting the cutter comp to "back off" (follow the edge) for making a pocket.
[17:59:20] <KimK> Do errors found ahead by AXIS affect current cutting? If I have handwritten g-code in-progress "foo_1, foo_2, foo_3, stop, foo_4_bad, stop", and AXIS says error in foo_4_bad, can I still run the first three parts OK?
[17:59:46] <SWPadnos> gouging error?
[17:59:53] <KimK> I think it was
[18:00:05] <SWPadnos> well, try it :)
[18:05:47] <KimK> Sorry, someone came in. OK, brb...
[18:05:52] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/share/axis/images/axis.ngc: fix splash so it does not violate machine limits of standard config
[18:06:02] <mOizt> mistress?
[18:11:56] <KimK> SWPadnos: I had an M0 instead of M30, fixed that, now AXIS doesn't see my error (unfinished code?) in the next section. I'm still working on the cutter comp, though. It looks like the cutter comp is coming off right at the end as it should, but it never came on through the first 98% in the first place? Not sure, still looking.
[18:12:54] <SWPadnos> err. ok
[18:13:18] <SWPadnos> well, just remember - no need to ask questions about what will happen when you're using a simulator to look at stuff. :)
[18:13:29] <SWPadnos> it's not like you'll break anything
[18:13:37] <CIA-2> EMC: 03jepler 07TRUNK * 10emc2/docs/man/man9/motion.9:
[18:13:37] <CIA-2> EMC: convert most 'axis.#.*' and 'motion.*' params to pins
[18:13:37] <CIA-2> EMC: based on a patch by Micha?? Geszkiewicz
[18:13:39] <CIA-2> EMC: 03jepler 07TRUNK * 10emc2/src/emc/motion/ (control.c mot_priv.h motion.c):
[18:13:39] <CIA-2> EMC: convert most 'axis.#.*' and 'motion.*' params to pins
[18:13:40] <CIA-2> EMC: based on a patch by Micha?? Geszkiewicz
[18:14:37] <jepler> BigJohnT: your change to axis.ngc is wrong. stop making it.
[18:15:16] <BigJohnT> ok
[18:15:47] <KimK> SWPadnos: You bet. I'll be sure to report any simulator breakage, though.
[18:15:50] <CIA-2> EMC: 03jepler 07TRUNK * 10emc2/share/axis/images/axis.ngc: revert to rev 1.7
[18:15:56] <SWPadnos> that's a good thing :)
[18:16:04] <BigJohnT> how can it be changed so it is not in error on startup with a standard config?
[18:16:39] <SWPadnos> it's out in Z, right?
[18:16:45] <BigJohnT> and Y
[18:16:46] <jepler> BigJohnT: you can't.
[18:16:54] <BigJohnT> :(
[18:17:15] <SWPadnos> the thing to do would be to increase the limits on whichever you think is the "standard config"
[18:17:27] <jepler> if you think there's a specific problem with it, that's one thing. but "It loads partially outside the working volume" is not a problem with axis.ngc.
[18:18:04] <BigJohnT> I was using the standard stepconf generated config
[18:18:09] <SWPadnos> I'm not sure there's any volume that's available to all sample configs
[18:18:26] <SWPadnos> some Z will range from -$something to 0, others from 0 to +$something
[18:18:42] <BigJohnT> which seems to hang up newbees because they get an error right from the get go
[18:18:48] <SWPadnos> there isn't a standard stepconf config. you enter the travel limits into stepconf
[18:18:59] <jepler> SWPadnos: no, but there are defaults (which no one should take)
[18:19:08] <SWPadnos> yep, about to say that
[18:19:14] <SWPadnos> heh
[18:19:40] <BigJohnT> but everyone does take the defaults when they are trying out EMC
[18:19:54] <SWPadnos> how difficult do you think it would be to have no defaults (blanks for the edit boxes) and disable the "next" button until they're all filled in?
[18:20:00] <jepler> BigJohnT: your change makes it so that the clueful user who knows how to touch off will get burned, because your g91 offset moves Z around arbitrarily
[18:20:06] <jepler> SWPadnos: you do it and let me know
[18:20:10] <SWPadnos> heh
[18:20:21] <SWPadnos> ok, I've determined that it's "too hard" :)
[18:20:23] <jepler> SWPadnos: but that'll trip up the novice user, because he won't be able to get past that page
[18:20:37] <jepler> "emc's no good, the configurator's next button doesn't even work"
[18:20:41] <BigJohnT> yep
[18:20:43] <jepler> I'd better stop typing now, because my blood is boiling.
[18:20:44] <jepler> bbl
[18:20:49] <SWPadnos> yeah, there would have to be messages and crap telling them which things need to be filled in before they can go on
[18:20:55] <BigJohnT> have a cold one on me jepler
[18:22:49] <SWPadnos> BigJohnT, the defaults for stepconf could be changed relatively easily
[18:23:08] <SWPadnos> I think the AXIS splash goes from -0.01 to +something
[18:23:25] <SWPadnos> just make the defaults slightly bigger than that
[18:23:40] <BigJohnT> It didn't take much to bring it out of the error condition
[18:23:57] <SWPadnos> .001 or .0001 IIRC
[18:24:04] <SWPadnos> err, .01 or .001 IIRC
[18:24:46] <BigJohnT> I must have missed the revert by cradek and thought I forgot to do it but I really did...
[18:25:22] <SWPadnos> yeah. messing with G92 is a bad idea - some people actually use it, and that's an effect that lasts across runs of EMC2
[18:26:51] <BigJohnT> I guess if they start the file then stop it it doesn't get to the G92.1 ...
[18:31:46] <CIA-2> EMC: 03cmorley 07TRUNK * 10emc2/configs/demo_step_cl/demo_step_cl.clp: change ladder timers to use new control pin
[18:31:50] <CIA-2> EMC: 03cmorley 07TRUNK * 10emc2/configs/dallur-thc/dallur-advanced.clp: change ladder timers to use new control pin
[18:31:50] <CIA-2> EMC: 03cmorley 07TRUNK * 10emc2/configs/common/configurable_options/ladder/estop.clp: change ladder timers to use new control pin
[18:31:51] <CIA-2> EMC: 03cmorley 07TRUNK * 10emc2/configs/demo_sim_cl/demo_sim_cl.clp: change ladder timers to use new control pin
[19:03:35] <KimK> SWPadnos: I got it (me?) figured out. In order to see what I "expected to see" (maybe not the way AXIS was intended to work?) I have to run the g-code once with G40 to draw the part itself, *don't erase*, and then run again with G42 to show the desired path. Maybe there should be a checkbox for "also show uncompensated toolpath (part edge)"? Anyway, thanks to you and everyone who helped me.
[19:07:21] <KimK> Hmmm, I haven't used optional block yet, I could double the size of my g-code (cut everything "twice") and get the desired effect...
[19:19:46] <cradek> KimK: I like to put /G42 and then I can toggle between the two paths with the block delete
[19:20:06] <cradek> unfortunately it is not possible to show both paths with the current architecture
[19:22:45] <KimK> cradek: I see, I might do it the other way, have G42 as "real" and option out the part edge (just for show in AXIS anyway). And I found that the "skip lines with /" checkbox removes lines from execution, but not from the preview, so that works out nicely too.
[19:24:55] <cradek> in 2.3, it reloads the preview for you - maybe 2.2 didn't do that
[19:26:22] <KimK> Reloads the preview? You mean trading option blocks for "cool display mode" won't work in 2.3?
[19:26:54] <cradek> sorry, you lost me
[19:27:08] <cradek> when you change the block delete setting, the preview changes to match
[19:27:24] <cradek> you can always keep an old backplot (red lines) to compare it to
[19:28:31] <KimK> OK, so in 2.3 the "skip lines with /" removes g-code from both execution and preview?
[19:28:48] <cradek> yes
[19:29:05] <KimK> OK. Well, it was a hack anyway, LOL
[19:29:15] <KimK> Forget I brought it up
[19:30:38] <KimK> We can still get it to show the part edge if needed so that's the important thing.
[19:36:08] <fenn> i think some day there will be a solid remaining preview
[19:36:41] <KimK> that will be great
[19:36:57] <KimK> * KimK will be glad when 2.3 comes out and the 2.2 vs 2.3 differences go away
[19:37:14] <alex_joni> KimK: check /topic
[19:37:19] <fenn> i guess the poll should be updated
[19:37:25] <alex_joni> it's coming soon
[19:37:29] <alex_joni> fenn: what poll?
[19:37:37] <fenn> on linuxcnc.org
[19:37:42] <alex_joni> hold your horses
[19:37:46] <alex_joni> still 2 months away :)
[19:37:56] <fenn> 2 months?
[19:38:12] <alex_joni> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Emc2.3Status
[19:38:25] <fenn> gah nevermind.. 2.2 is at the bottom of the poll
[19:38:41] <alex_joni> yeah, didn't found a way to reorder it :/
[19:45:18] <KimK> cradek: I was just thinking that the v2.2 non-refreshed display (always showing optional blocks) is kind of nice. It shows up in the preview, but doesn't get executed. Is there a way to keep that by having one of those "old behavior = true" settings? Or "always preview / lines = true"?
[20:18:45] <BigJohnT_> BigJohnT_ is now known as BigJohnT
[20:22:55] <BigJohnT> would there be any problems if I changed each line in axis.ngc from G00 X [.5+1536.00*#3] Y [2556.00*#3] to G00 X [.5+1536.00*#3] Y [.5+2556.00*#3]?
[20:23:51] <BigJohnT> and similar for the Z
[20:30:14] <BigJohnT> http://pastebin.ca/1323841
[20:37:17] <jepler> BigJohnT: I don't think you're doing the right thing by "fixing" this "problem", because there is no problem
[20:40:49] <BigJohnT> I agree it is not a problem but it is confusing to a newbee who is just trying EMC out...
[21:05:04] <BigJohnT> from the Axis user manual "When the coordinates displayed are relative, the display will include a cyan “machine origin”
[21:05:06] <BigJohnT> marker ( )."
[21:05:33] <alex_joni> roundy thingie with a cross in it
[21:05:41] <BigJohnT> is this marker still displayed somewhere? I can't find it
[21:05:52] <alex_joni> it's in the preview
[21:06:03] <alex_joni> try to touch off, then you'll see it
[21:06:08] <BigJohnT> ok
[21:06:52] <BigJohnT> got it thanks alex_joni
[21:07:04] <alex_joni> does it work?
[21:07:28] <BigJohnT> yep, I just never noticed it and I was updating the manual...
[21:07:50] <BigJohnT> and trying everything to make sure it works
[21:08:03] <alex_joni> coo
[21:08:10] <BigJohnT> and agrees with the manual
[21:09:48] <micges> will be created branch for new things tomorrow or we will "wait" 2 month for it ? (and forgot)
[21:10:45] <micges> (new things discovered while testing 2.3 beta)
[21:17:56] <alex_joni> micges: if new things are bugfixes, they go on TRUNK
[21:18:08] <alex_joni> if they are new things, no reason they couldn't go on a branch
[21:40:40] <micges> thanks just make sure
[21:42:58] <jepler> I will discourage creation of branches during the time we're stabalizing for 2.3.
[21:43:16] <jepler> to encourage work on 2.3 is why I specifically do NOT call for 2.3 branching from TRUNK right away
[21:44:21] <jmkasunich> micges: if you have new features that didn't make the deadline for 2.3, file a feature request at sourceforge so they don't get forgotten
[21:46:43] <micges> I'm just wondering becaouse I was not following cvs while 2.1 => 2.2 and wonder how things will be done
[21:47:00] <alex_joni> micges: we try to keep 2.3 in TRUNK for as long as possible
[21:47:05] <alex_joni> and branch at the last moment
[21:47:14] <alex_joni> that way all 2.3 fixes will be in TRUNK aswell
[21:48:24] <micges> ok
[21:49:37] <alex_joni> jepler: isn't #2187024 fixed in TRUNK? I haven't watched stepconf changes closely lately..
[21:58:22] <BigJohnT_> BigJohnT_ is now known as BigJohnT
[21:59:34] <jepler> alex_joni: I don't know
[21:59:40] <jepler> I should test it when I test stepconf, which I should do soon.
[22:02:18] <alex_joni> ok, was just looking over the open bugs
[22:02:28] <alex_joni> I should probably try to concentrate on the task ones tomorrow
[22:02:49] <alex_joni> good night all
[22:04:43] <micges> good night alex
[22:06:49] <jepler> micges: BRIAN GLACKIN posts asking for table skew correction. can you add your kinematics component to http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?ContributedComponents#Kinematics for him?
[22:08:53] <micges> yes I can
[22:09:37] <jepler> his post was on the users mailing list.
[22:11:04] <jepler> hm, I was at first excited to see this http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?ContributedComponents#EtherCAT_Digital_I_O but on closer inspection it's not a proper HAL realtime component, because it uses kernel spinlock API calls inside HAL functions: spin_lock(&master_lock);
[22:11:21] <micges> jepler: for binary install "sudo comp --install millkins.c" will works ?
[22:11:33] <jepler> which raises the question: what to do about buggy components on that page
[22:11:56] <jepler> micges: yes, when emc2-dev is installed
[22:28:08] <invite|MkS> hi
[22:28:23] <invite|MkS> I'm willing to write a nesting prog
[22:28:47] <invite|MkS> to create sheet metal for plasma cutting
[22:28:53] <invite|MkS> or laser
[22:29:12] <invite|MkS> and I'm currently retaining DXF as an input format
[22:29:21] <invite|MkS> IS that a bad choice ?
[22:29:51] <invite|MkS> the exported code will be used under AXIS using the DXF lib from ribbonsoft
[22:32:25] <jepler> I think there are people who would like to see a dxf-to-gcode conversion program
[22:32:39] <jmkasunich> with or without nesting
[22:34:14] <invite|MkS> so the DXF choice isn't finaly a bad one
[22:34:39] <invite|MkS> although pydxf is working for simple pieces
[22:34:56] <invite|MkS> it becomes buggy when dealing with complicated shapes
[22:35:21] <invite|MkS> I have made some tests and the dxflib from rubbonsoft is
[22:35:35] <invite|MkS> better, although it may need some tuning
[22:35:56] <invite|MkS> but I'm gonna concentrate on the nesting part
[22:36:14] <invite|MkS> as this is the part of the project where I'll do the maths
[22:36:25] <jmkasunich> is the ribbonsoft code GPL?
[22:36:30] <invite|MkS> and I need it in my school project
[22:36:57] <invite|MkS> yep it's GPL v2
[22:39:25] <jepler> cool
[22:39:55] <jepler> I think micges is someone who might be interested in this
[22:40:47] <invite|MkS> I'll also probably code it under Qt (which is now LGPL)
[22:40:54] <invite|MkS> and use CGAL for the math
[22:40:59] <jepler> look out for CGAL
[22:41:03] <jepler> important parts of it are *NOT* gpl
[22:41:43] <invite|MkS> yes that's the main problem
[22:41:44] <jepler> (for example, the part that's relevant for computing offset polygons, which I was interested in)
[22:43:03] <CIA-2> EMC: 03cmorley 07TRUNK * 10emc2/src/hal/classicladder/classicladder.h: Lower bit and word defaults - no need to be that high - makes the symbol window less crowded
[22:44:00] <invite|MkS> the minkowski sum package is QPL
[22:44:07] <invite|MkS> and I'll mainly use it
[22:44:14] <invite|MkS> for nesting
[22:45:09] <invite|MkS> I'll probably use it for testing and maybe try to implement my own/search for a
[22:45:24] <invite|MkS> another candidate later
[22:46:09] <CIA-2> EMC: 03cmorley 07TRUNK * 10emc2/src/hal/classicladder/ (emc_mods.h emc_mods.c files.c): Add auto assignment of variables to the symbol window - It just adds the variable name as the symbol name unless the program already loaded a symbol name
[22:50:50] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/ladder/classic_ladder.lyx: update default table
[23:00:57] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/gui/images/ (4 files): add imagesX
[23:04:43] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/gui/axis.lyx: updating to match 2.3
[23:06:00] <alex_joni> invite|MkS: do you plan to do auto nesting?
[23:06:05] <alex_joni> or only manual for starters?
[23:44:39] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/common/ (User_Concepts.lyx machining_center.lyx): minor edits