#emc | Logs for 2007-04-13

[00:13:46] <tomp> Rugludallur: woohoo! looks like it works,, you try & see if hal agrees with the values pleez. the pyvcp_widgets.py file http://pastebin.ca/437355, a test xml file http://pastebin.ca/437359
[00:14:27] <tomp> Rugludallur: keep it isolated for now
[00:14:42] <Rugludallur> tomp: on it :D
[00:18:34] <tomp> Rugludallur: i'm making these graphic components for gEDA, and wrote some text files describing gui widgets for pyvcp.. there needs to be one tool somehow, where we have a canvas and widgets, and wire up 'schematics'... if not one tool, then a single data format... i got c, python, dot, xml and this gEDA format... its doable but others wont use such disparate tools...
[00:19:32] <tomp> meanwhile, i'll add this patch to my private pyvcp_widgets.py ;)
[00:19:56] <Rugludallur> tomp: mind if I add a min/max check and some ?
[00:20:55] <tomp> Rugludallur: this is open source, but juried ( approved by peers ) so, du vas zu vilt :)
[00:24:17] <Rugludallur> tomp: http://pastebin.ca/437390
[00:25:42] <tomp> range checking? ( make sure it's 'in between' )
[00:25:59] <Rugludallur> tomp: just want to make sure that initvalue is within min/max
[00:26:30] <tomp> good thinking, i always write user-hostile code ( f 'em if they cant take a joke ;)
[00:26:31] <Rugludallur> tomp: btw there is a type in that pastebin post
[00:26:47] <tomp> ?
[00:27:06] <Rugludallur> tomp: my pastebin D:
[00:27:30] <tomp> ok, run it & then paste it, we'll make a good one
[00:28:02] <Rugludallur> tomp: did you see anything on that problem with updating spin by typing ?
[00:28:12] <Rugludallur> tomp: looks like it was never implemented from what I can tell
[00:28:46] <robin_sz> meep?
[00:29:11] <tomp> didnt try to type in a value... just did chgd 1234.0 to 123.0 and it incs decs ok for me
[00:29:50] <Rugludallur> tomp: check the signal behind
[00:29:56] <tomp> but i just looked at gui, not at what hal sees
[00:29:57] <Rugludallur> tomp: or the pin, it's not updated
[00:30:35] <tomp> ok, is the pin updated ok without typing into the field?
[00:30:55] <Rugludallur> tomp: yup, increment/decrement updates fine, even after entering a number when you increment it works
[00:31:14] <Rugludallur> tomp: probably just need to hook up a refresh for when focus leaves the button
[00:32:29] <tomp> Rugludallur: update() is called cyclicly, and happens when old value != value_now .... should not need special code, need for it to notice that it has been changed tho
[00:33:33] <tomp> Rugludallur: are you inspecting self.value?
[00:33:39] <Rugludallur> tomp: I seem to remember something similar when I did the spinbutton for the old vcp
[00:34:08] <Rugludallur> tomp: looking at the pin created (and specified) in the xml
[00:35:38] <tomp> def update(self,pycomp):
[00:35:48] <tomp> pycomp[self.halpin] = self.value
[00:35:57] <tomp> if self.value != self.oldvalue:
[00:36:04] <tomp> self.v.set( str( self.format % {'b':self.value} ) )
[00:36:06] <tomp> self.oldvalue=self.value
[00:36:46] <Rugludallur> tomp: yes but I'm betting update is not called periodically, probably called from the widget
[00:36:57] <SWPadnos> err - self.value.set(...) ??
[00:37:31] <tomp> no, there's special 'tracking vaslues' like an event almost, and it's name is v and its method is set
[00:37:33] <SWPadnos> or is 'v' the "displayed value"?
[00:37:43] <SWPadnos> ok
[00:39:17] <tomp> Rugludallur: dunno, all widgets got an 'update()' and all widgets's updates get called cyclicly ( or none do, there's no individual action )
[00:40:15] <tomp> Rugludallur: so if any updates get called then all updates get called, prob might be in this update code tho
[00:40:46] <SWPadnos> without understanding quite what's happening, it looks to me like some interaction with 'v' is the problem
[00:41:27] <SWPadnos> but then again, that's ludictous :)
[00:41:28] <SWPadnos> ludicrous
[00:41:47] <tomp> hokey smokes, the c code has special code for the spin button value....
[00:42:04] <tomp> it does the limit test
[00:43:08] <Rugludallur> tomp: but only for values set by the button right ?
[00:43:38] <Rugludallur> tomp: it does not protect you from setting it as an inital value on the hal pin right ?
[00:43:47] <tomp> dunno yet
[00:44:29] <tomp> did you run this from a terminal? any 'GTK errors of hal export pin errors?
[00:44:38] <Rugludallur> tomp: nothing of the sort
[00:44:45] <tomp> or hall pin export errors
[00:44:53] <Rugludallur> tomp: nope
[00:45:10] <tomp> ctfs ( cant type fer sh&t)
[00:45:44] <tomp> ok, slow down & isolate the problem, no patches to offer for now
[00:47:18] <Rugludallur> bahh it's called spinbox ,, and here I was stumped at why I found no def of spinbutton :P
[00:47:41] <tomp> note bona , the special 'tracking variable' in gtk is called an 'adjustment', acts like an event ( change it and a message occurs )
[00:49:38] <tomp> the 'adjustment' belongs to the widget ( the hierarchy is important, the value may not propagate to the storage structure )
[00:49:56] <SWPadnos> ok, basic question: in the __init__ function, what is init_ (vs. initval) ?
[00:50:15] <SWPadnos> or is that the typo? (I thought it was the quad quotes instead of triple quotes above)
[00:50:19] <Rugludallur> SWPadnos: typeo
[00:50:25] <Rugludallur> SWPadnos: let me repost
[00:50:30] <SWPadnos> ok, thanks
[00:50:41] <SWPadnos> I don't know enough python to spot typos yet :)
[00:50:53] <Rugludallur> http://pastebin.ca/437421
[00:51:26] <Rugludallur> SWPadnos: python comes naturally, I find c code hard to read most of the time but python just ,, flows
[00:51:33] <SWPadnos> heh
[00:51:37] <tomp> __init__ is the name of the instantiator, initval is an argument
[00:51:44] <SWPadnos> there's still a typo in pyvcp_jogwheel
[00:51:55] <SWPadnos> there's an extra " in the initial comment
[00:52:53] <SWPadnos> fixed here: http://pastebin.ca/437423
[00:53:27] <SWPadnos> urm - line 686 looks the same to me - still initval vs. init_
[00:53:43] <SWPadnos> err - nevermind
[00:53:52] <SWPadnos> * SWPadnos needs more protein in his diet
[00:54:02] <SWPadnos> or something
[00:54:02] <tomp> SWPadnos, may be a typo but that code runs somehow
[00:54:40] <SWPadnos> I don't see how. the """" should be a one or two quote string
[00:54:49] <SWPadnos> or an error
[00:55:59] <tomp> in vcp_widgets.c the adjustment gets tied to the spin button, this should connect a message maker to the widget
[00:56:30] <tomp> then the widget is made sensitive to changes with 'gtk_signal_connect(GTK_OBJECT(gwp), "changed", GTK_SIGNAL_FUNC(spin_button_changed), wp);'
[00:57:25] <tomp> and the hal pin to the adjustmnet's value (GTK_SPIN_BUTTON(gwp)); *(hd->pin) = pd->pin_state;
[00:57:55] <tomp> so the error must be tracked in those few lines, but with a debugger ( or lotsa print statements )
[00:58:02] <tomp> :)
[01:02:04] <Rugludallur> tomp: I think that might be my code :P
[01:10:12] <tomp> Rugludallur: lemme get clear on this.... when the spinbutton is created >without< an initial value, then hal sees the correct value?
[01:10:13] <tomp> and when the spin button is created >with< an initial value, hal does >not< see the correct value?
[01:12:30] <Rugludallur> tomp no ,, sorry I should have been more clear
[01:12:33] <Rugludallur> inital values work fine
[01:12:44] <tomp> i got a halmeter on it and it tracks the pin 'emd99.OnTome uS' fine
[01:12:59] <tomp> ok
[01:13:00] <Rugludallur> ok, try putting your mouse cursor into the spinbox
[01:13:04] <Rugludallur> change the value
[01:13:14] <tomp> edit it like a text field?
[01:13:16] <Rugludallur> take focus out of the spinbox (enter or click other)
[01:13:19] <Rugludallur> tomp: yes
[01:13:27] <Rugludallur> tomp: put in a different number
[01:13:54] <Rugludallur> tomp: and then notice that the number change is shown but not "commited" to hal
[01:14:29] <tomp> ok, it doesn't immediately reflect the new value when i hit return, only after i inc or dec it... ok thats the symptom...
[01:14:36] <Rugludallur> tomp: right
[01:14:51] <tomp> i think thats a gtk bug, out of our hands
[01:15:25] <Rugludallur> tomp: might be, I don't know if it should notify of the change or what
[01:16:02] <tomp> because that connection is lower than pyvcp, its up to gtk to notify the app of the change in adjustment... best to ask it gtk land :(
[01:16:04] <Rugludallur> tomp: I don't know how the interaction python/tk
[01:16:35] <Rugludallur> tomp: anyway, im tired (1:23 over here) and I have to be up at 6:30
[01:16:42] <tomp> if its in emc, its in the .c file
[01:16:51] <tomp> good nite; sleep fast :)
[01:17:04] <Rugludallur> tomp: might be different ways to get that refresh/update signal from the widget :D
[01:17:09] <Rugludallur> night, thanks again :D
[01:21:18] <cradek> Infernal Revenue Service
[01:21:37] <SWPadnos> heh
[01:31:34] <ds3> how well would a joypad (like the ones on games systems) work as a jogging control?
[01:37:19] <cradek> I hear analog joysticks work great
[01:37:30] <cradek> I think a joypad isn't analog, so you'd only get one speed
[01:37:54] <ds3> oh hmmmm
[01:38:13] <ds3> reason I asked is I found a bunch of joypads that plug in and can emulate up to 32 keys
[01:38:26] <ds3> they were going for $1 each!
[01:38:34] <cradek> to work nicely with emc, you need a device that linux recognizes
[01:38:44] <cradek> seems like it's good at recognizing joystick type stuff
[01:38:51] <cradek> (usb joystick)
[01:38:55] <ds3> this one plugs in as a keyboard and emulates a keyboard!
[01:39:05] <ds3> so linux/dos/etc is not an issue
[01:39:49] <ds3> more trying to decide if I should keep it as is (a joy pad) or rip it apart and wire it up with buttons for a panel
[01:39:50] <cradek> that sounds like it must depend on a driver to assign the keys
[01:40:01] <tomp> xemets notes on usb joypads for jogging http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?HowToWriteNewHalComponent http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Using_A_Joypad_To_Move_Your_CNC_Machine
[01:40:04] <ds3> no drivers; the keypad is programmed ahead of time w/o other software
[01:40:11] <cradek> brb
[01:45:44] <tomp> re spinbuttons: c examples from gtk 1.2 work, python examples from pygtk work, so the binding to the hal pin is suspect
[01:47:07] <tomp> meaning the value in the text field can be edited or incremented or decremented
[01:52:35] <Jymmm> * Jymmm is back from surgery with vikoden script i hand! WOOHOO!
[01:53:52] <Jymmm> I wonder when they expire???
[01:56:31] <Jymmm> ds3 where at?
[01:56:54] <ejholmgren> jigga wha?
[01:57:13] <Jymmm> ejholmgren say what?
[01:58:08] <ejholmgren> typing under the influence of perscription strength pain-killers
[01:58:17] <ejholmgren> woo hoo
[01:58:48] <Jymmm> ejholmgren Ah. Nah, He just wrote me the script, I doubt I'll fill it before it expires.
[01:59:08] <Jymmm> I already have some here anyway if I need them.
[02:01:32] <ds3> Jymmm: excess sol'n
[02:01:42] <Jymmm> ds3 milpitas?
[02:01:44] <ds3> yep
[02:01:47] <Jymmm> ah
[02:02:05] <ds3> got 5 of them to tear up
[02:02:08] <Jymmm> I havent been to them since they moved from brokaw
[02:02:32] <ds3> the new place is bigger but less of a temptation to visit
[02:02:40] <Jymmm> bummer
[02:02:59] <ds3> that's cuz they are not next to frys ;)
[02:03:08] <Jymmm> you got it
[02:22:03] <tomp> re vcp spin btn, the code that clips the value to the min/max range is >after< the widget is refreshed and repainted
[02:23:36] <tomp> gtk_widget_show(gwp); should be last thing before return (this is not the bug yet )
[02:28:58] <tomp> the gtk docs say that editing the field does not emit any signal only using the arrows will do that ( as written now)
[02:29:00] <tomp> "The "change-value" signal is emitted when the spinbutton value is changed by keyboard action using the Up Arrow, Down Arrow, Page Up, Page Down, Control+Page Up or Control+Page Down keys."
[02:29:20] <tomp> not when the text field is edited
[02:32:06] <tomp> "value-changed" might do it, that may reflect an edited field, if so, the 2 signals must be or'ed to get the function we need ( damn symantics if ["change-value" or "value-changed" ] sheesh! )
[02:44:40] <chron1c> ehhh... howdy
[02:45:58] <ejholmgren> word
[02:46:37] <chron1c> i am already feeling the urge to build a bigger machine.. this one isn't even done yet
[02:46:40] <chron1c> :)
[02:46:51] <chron1c> i need to set up limit switches so i stop crashing
[02:47:00] <chron1c> good thing my table is mdf?
[02:47:16] <cradek> chron1c: you can home by eyeball, then use soft limits
[02:47:20] <cradek> no hardware required
[02:47:22] <cradek> (that's what I do)
[02:47:33] <chron1c> i am finding the homing is a bit erratic?
[02:47:43] <cradek> oh you have home switches? erratic how?
[02:48:01] <chron1c> i have switches... just not wired or mounted yet
[02:48:11] <cradek> what do you mean erratic then?
[02:48:19] <chron1c> erratic like not always setting sero when i tell it too
[02:48:25] <chron1c> zero*
[02:48:42] <chron1c> using axisaxis
[02:48:50] <chron1c> i like the homing in the mini
[02:49:00] <chron1c> but everything else about axis
[02:49:24] <cradek> explain what you do, and how the results differ from what you expect
[02:49:57] <chron1c> well i was thinking it may be because of the g-code setting tool offsets and work planes?
[02:50:12] <chron1c> what i really don't get is the blue dot in axis?
[02:50:14] <cradek> definitely could be
[02:50:16] <chron1c> is that true home?
[02:50:25] <ejholmgren> ef ... I finally burn a debian 3.1_r5 iso, and 4.0 comes out
[02:50:28] <cradek> the blue dot is the machine home. if you g0 g53 x0 y0 z0, the cone goes to the blue dot
[02:50:52] <chron1c> ok, how do i change the blue dot (machine home in axis?
[02:50:53] <cradek> the coordinate axes marked XYZ are the current coordinate system offsets, which might include G5x (G10 L2 Px) and/or G92
[02:51:04] <chron1c> yes i know coordinates
[02:51:10] <chron1c> and axes'
[02:51:16] <chron1c> and some g-code
[02:51:24] <cradek> the blue dot is your machine's 0,0,0
[02:51:38] <cradek> assuming your machine is homed
[02:51:56] <cradek> if it's not home, the blue dot is the location of the tool when you started emc
[02:52:00] <cradek> not homeD
[02:52:15] <chron1c> if i close axis
[02:52:21] <chron1c> then restart it
[02:52:28] <chron1c> the dot always stays the same place...
[02:52:45] <cradek> that's because the offsets (G5x) are permanent until you clear them
[02:52:48] <chron1c> unless i load mini then zero all, close it then open axis
[02:52:54] <chron1c> ohh
[02:53:07] <chron1c> hmm
[02:53:08] <cradek> mini might zero G5x offsets, I don't know
[02:53:14] <chron1c> to clear g54?
[02:53:25] <chron1c> i use mdi and ...?
[02:53:32] <cradek> g10 l2 p1 x0 y0 z0
[02:53:36] <chron1c> i looked but couldn't figure it
[02:53:50] <cradek> do you home your machine?
[02:54:06] <chron1c> i don't have the switches?
[02:54:08] <chron1c> wired*
[02:54:09] <cradek> if not, there's no reason to use the offsets (touch off button)
[02:54:26] <cradek> just jog where you want the axis's zero to be, and click home
[02:54:35] <chron1c> oh...
[02:54:44] <chron1c> i should have read up on axis
[02:54:51] <chron1c> i thought that would move the machine
[02:54:58] <cradek> you CAN home without switches - mark a home position for each axis with a sharpie, and jog there to home
[02:55:02] <chron1c> to the set already home pos.
[02:55:17] <cradek> without switches and homing velocities set up, clicking Home just zeroes the axis without any motion
[02:55:25] <chron1c> ahh
[02:55:49] <chron1c> when i wire the switches in... do i put them at the part loading positions for each axis?
[02:55:51] <cradek> the default stepper file has homing velocities = 0
[02:56:05] <cradek> no, you should put home switches at one extreme of travel
[02:56:19] <cradek> otherwise when the homing motion starts, it can't know which way to move to find the switch
[02:56:20] <chron1c> well, that's what i meant when i say part loading
[02:56:52] <cradek> ideally the home switch should stay tripped as you move all the way into the limit switch
[02:57:11] <tomp> re vcp spinbutton: the signal "changed" is connected to the function 'spin_button_changed()", but "changed" doesnt seem to be a valid message. however "change-value" and "value-changed' are in the gtk docs. i suspect its just not a valid message but why do values ever get propagated to the hal-pins then? becuz it's polled AND sensitized to messages ( the msg never happens, the poll does )
[02:57:12] <cradek> so you can't ever get on the "wrong" side of the home switch
[02:57:17] <chron1c> if i only have three limit inputs on my driver card... do i wire them in series and emc will know which end is which?
[02:57:27] <chron1c> i guess i don't have home and limit
[02:57:29] <chron1c> only limit
[02:57:30] <toastydeath> i got on the "wrong" side of the home switches today
[02:57:33] <toastydeath> it was lol
[02:57:41] <cradek> toastydeath: ouch
[02:57:41] <toastydeath> "oops!"
[02:57:45] <toastydeath> nah it
[02:57:51] <toastydeath> 's a commercial machine
[02:57:52] <cradek> chron1c: you have lots of options to use your inputs
[02:57:55] <toastydeath> there are limit switches behind it
[02:58:03] <cradek> toastydeath: good!
[02:58:46] <cradek> chron1c: I'd wire together both limits of each axis
[02:59:06] <cradek> chron1c: then you can home all three at the same time (home on one of the limit switches)
[02:59:12] <chron1c> can emc tell a difference in voltage for the inputs? like home for full voltage and a resistor on another switch to the same pin for limit?
[02:59:39] <jmkasunich> no, parport pins are digital, they understand 1, and 0
[02:59:41] <cradek> chron1c: that's a hardware issue, and no, most input pins are digital
[02:59:41] <jmkasunich> not 0.5 ;-)
[02:59:47] <chron1c> oh..
[02:59:48] <cradek> right
[03:00:06] <cradek> but you can wire X+ and X- limit switches together, and hook those to one input
[03:00:38] <chron1c> yes, i read that in my driver info, but how do i tell emc which is + and _
[03:00:41] <chron1c> -?
[03:00:41] <cradek> then as long as you're not ON the "other" limit switch when you start homing, it will work
[03:01:01] <cradek> if you're sitting on the "wrong" switch, it will crash when you try to home
[03:01:04] <jmkasunich> how many pins do you have available for switches?
[03:01:08] <chron1c> it will always home in the same direction?
[03:01:08] <cradek> he said 3
[03:01:15] <chron1c> i have 3 pins
[03:01:42] <jmkasunich> how many switches per axis? 1 (at one end), 2 (one at each end) or 3 (limits at each end, plus home)
[03:01:42] <chron1c> hmm..
[03:01:53] <cradek> jmkasunich: can you start a home when you're already sitting on a limit?
[03:02:00] <chron1c> i bought 6 lever microswitches from jameco
[03:02:07] <chron1c> i haven't mounted them yet
[03:02:11] <jmkasunich> cradek: good question
[03:02:19] <cradek> jmkasunich: if not, my suggestion is safe
[03:02:22] <jmkasunich> chron1c: so, 1 at each end
[03:02:27] <chron1c> yes...
[03:02:40] <cradek> jmkasunich: I bet you can't (you shouldn't be able to anyway)
[03:02:44] <jmkasunich> why not wire all + limits to one pin, all - inputs to the other
[03:02:56] <jmkasunich> and use home sequencing to handle the home all case?
[03:03:01] <cradek> jmkasunich: that is safer, but you can only home one at a time then
[03:03:16] <jmkasunich> so, how often do you home?
[03:03:18] <chron1c> interesing...
[03:03:47] <jmkasunich> with the + and - limits independent, you if you hit a limit you can back off without overriding
[03:03:54] <cradek> jmkasunich: not true
[03:04:00] <jmkasunich> ?
[03:04:30] <cradek> I don't think emc does that, although I agree it should be possible
[03:04:31] <jmkasunich> maybe my memory is rusty, but I thought the + limit only stopped further + movement
[03:04:51] <cradek> jmkasunich: no, I think it turns machine off, and you can't turn it back on without overriding
[03:04:59] <chron1c> jmk: that's how my milltronics machine worked
[03:05:10] <chron1c> i could move off of + limit in neg dir
[03:05:26] <cradek> I could be wrong, but I'm fairly sure I'm not
[03:05:38] <chron1c> but this is jsut some stuff i rigged together, it's no milltronics
[03:06:28] <chron1c> i tried to run the prototrack software on my home computer.. they had it look for a dongle thing though :(
[03:06:54] <chron1c> those pro's are smart like that
[03:06:55] <chron1c> lol
[03:06:56] <jmkasunich> cradek: you're right (just looked at the code)
[03:07:11] <cradek> cheater
[03:07:12] <jmkasunich> fault handling in that area could stand to be improved
[03:07:16] <chron1c> so you guys are the ones that work on the code?
[03:07:30] <jmkasunich> yes
[03:07:38] <cradek> yes, but not the only
[03:07:50] <jmkasunich> cradek is the trajectory planner dude, I do mostly HAL stuff, and some of the motion controller
[03:08:04] <chron1c> awesome
[03:08:12] <cradek> I'm mostly a bugswatter
[03:08:17] <chron1c> it's nice to get a chance to chat with you folks
[03:08:23] <jmkasunich> cradek and jepler did axis, alex_joni does lots of different stuff, and others contribute as well
[03:08:50] <chron1c> i have had my eye on emc for about 4 years
[03:08:54] <jmkasunich> lerman does interpreter stuff (like o-words)
[03:09:05] <chron1c> i i just now got around to using it a few months ago
[03:09:47] <chron1c> it took me three years to collect parts and get the motivation to finally build the machine
[03:10:38] <chron1c> the last part of the machine i needed was the 8 inch drill press frame/head that i dtripped out so i could mount the slides on it
[03:10:45] <chron1c> stripped*
[03:10:59] <cradek> I'm off to bed early tonight - nice to meet you chron1c
[03:11:07] <chron1c> g'night cradek
[03:11:15] <jmkasunich> goodnight chris
[03:11:35] <cradek> * cradek waves
[03:18:14] <SWPadnos> hey jmkasunich, have you tried to run easycad under wine?
[03:18:27] <jmkasunich> not lately
[03:18:45] <SWPadnos> I just got wine installed (on my 64-bit system, no less), and it seems quite good
[03:18:46] <jmkasunich> I don't recall if I ever did or not - I use a VM to run it
[03:19:00] <jmkasunich> have you tried easycad on it?
[03:19:13] <SWPadnos> nope. I haven't gotten to any major programs yet
[03:19:22] <chron1c> i just took some pics of my machine, should i post a link when i upload them?
[03:19:25] <SWPadnos> I may try installing TurboCAD on it
[03:19:36] <jmkasunich> easycad is far better than turbocad
[03:19:40] <SWPadnos> chron1c, sure. I think there's a page on the wiki for that kind of thing
[03:19:40] <chron1c> imight need to redo the pics with a better lamp
[03:20:04] <jmkasunich> download it and try it (I think theres a trial mode or something, check the website)
[03:20:04] <SWPadnos> well, I have TurboCadCAM handy (from the review)
[03:20:31] <chron1c> i have a copy of turbocad for mac... anyone need that? or.. can i run that under linux somehow?
[03:20:44] <SWPadnos> I can also see if CadMax will work (though I think they're out of business, and I would need a new activation key)
[03:21:08] <SWPadnos> if you can run it under Linux (x86), it's probably still faster to draw on paper
[03:22:13] <CIA-19> 03jmkasunich 07TRUNK * 10emc2/src/hal/drivers/mesa_5i2x/firmware/Makefile: moving things around before its too late
[03:22:27] <SWPadnos> heh
[03:22:36] <SWPadnos> is the VHDL committed yet, or just the makefile?
[03:22:43] <jmkasunich> just the makefile
[03:22:50] <SWPadnos> ok, I thought I was missing something
[03:22:53] <jmkasunich> I'm about to commit a few vhdl files, hence the "too late"
[03:22:57] <SWPadnos> heh
[03:23:28] <jmkasunich> I was originally gonna have one dir for common files, one for 5i20 specific stuff, and one for (future) 5i22 stuff
[03:23:34] <jmkasunich> decided that was overly complex
[03:23:53] <SWPadnos> well, a library directory would be good, since the components can be used for pluto and anything else FPGA as well
[03:24:10] <SWPadnos> even if the driver needs to be different, and the bus connection is different ...
[03:24:26] <jmkasunich> and tools and makefiles are different...
[03:24:41] <SWPadnos> right - I guess I look it as an include directory
[03:25:23] <jmkasunich> isn't the pluto stuff in verilog?
[03:25:50] <SWPadnos> it is, but there's no reason that someone can't write VHDL for pluto (or verilog for Mesa)
[03:26:22] <jmkasunich> true, if they don't mind starting at square 1
[03:26:46] <SWPadnos> the pin (UCF) files are actually more important than the VHDL or verilog source
[03:26:58] <SWPadnos> and you can mix and match verilog and VHDL
[03:28:04] <jmkasunich> only if you want to pull your hair out
[03:28:19] <SWPadnos> well, you can't start losing hair too early ..
[03:28:53] <CIA-19> 03jmkasunich 07TRUNK * 10emc2/src/hal/drivers/mesa_5i2x/firmware/ (4 files): constraints files for 5i20
[03:29:54] <toastydeath> i shave my head!
[03:29:55] <toastydeath> bring it on!
[03:30:06] <chron1c> lol, i shave my head also
[03:30:10] <toastydeath> a+!
[03:30:23] <chron1c> easir than combing it and stuff
[03:31:53] <toastydeath> agreed
[03:32:06] <toastydeath> plus, hey, free air conditioning
[03:32:12] <toastydeath> get cold, wear a hat!
[03:32:14] <SWPadnos> not so useful here
[03:32:26] <SWPadnos> I use my hair so I don't have to buy a hat :)
[03:32:33] <SWPadnos> it saves the environment
[03:33:24] <chron1c> i just took some beautiful pics of my machine... but... i left my pipe on the bench? muahahahaha... guess i better move it and take more before the whole world sees them
[03:33:27] <jmkasunich> SWPadnos doesn't need to buy a coat either
[03:33:38] <SWPadnos> hey - it's not that long
[03:33:46] <jmkasunich> lol
[03:34:00] <SWPadnos> hey - have they gotten more salt for the roads there?
[03:34:07] <SWPadnos> (or did it all melt)
[03:34:10] <jmkasunich> its melting
[03:34:11] <chron1c> where you at swp?
[03:34:43] <SWPadnos> http://www.google.com/maps?q=Essex+Junction,+VT+05452,+USA&sa=X&oi=map&ct=title
[03:34:58] <chron1c> vermont/
[03:34:59] <chron1c> ?
[03:35:18] <SWPadnos> yes :)
[03:35:49] <chron1c> ahh
[03:38:07] <CIA-19> 03jmkasunich 07TRUNK * 10emc2/src/hal/drivers/mesa_5i2x/firmware/ (oneshot-test.vhd oneshot.vhd): oneshot package and testbed
[03:44:22] <CIA-19> 03jmkasunich 07TRUNK * 10emc2/src/hal/drivers/mesa_5i2x/firmware/ (stepgen-core-test.vhd stepgen.vhd): step generator package and preliminary testbed
[03:44:58] <ejholmgren_> ejholmgren_ is now known as ejholmgren
[03:45:45] <jmkasunich> well thats cool... cvsweb can do syntax highlighting on VHDL source
[03:47:16] <SWPadnos> ah - I should update and try simulating some stuff
[03:47:29] <SWPadnos> to do simulation builds, do you need the Xilinx tools?
[03:47:34] <jmkasunich> nope
[03:47:45] <jmkasunich> just ghdl and gtkwave, ubuntu has packages for both
[03:47:56] <SWPadnos> ok. I think the makefile barfs if you don't have a valid XILINX environment variable
[03:47:57] <jmkasunich> do "make oneshot-test.wave"
[03:48:13] <jmkasunich> I commented out that barface
[03:48:18] <jmkasunich> age
[03:48:19] <SWPadnos> ok
[03:48:38] <jmkasunich> later I'll make it barf if you try synthesis
[03:48:52] <SWPadnos> heh - I'm sure something will fail in that case :)
[03:48:57] <chron1c> brb
[03:49:13] <jmkasunich> yeah - but the intentional fail prints a usefull message
[03:49:28] <SWPadnos> details, details
[03:49:34] <jmkasunich> in fact, I should add one that says "you need GHDL and GtkWave to do this" for simulation
[03:49:44] <chr0n1c> ahh a real irc client
[03:49:48] <chr0n1c> i was using trillian
[03:49:56] <chr0n1c> it doesn't work so good
[03:52:01] <chr0n1c> if anyone is interested here is a link to the cnc page on my website... http://ohiopctech.com/cnc/ (it's jsut pictures for right now until i add some text.
[03:52:24] <chr0n1c> i just took some pics of my semi finished machine
[03:52:45] <chr0n1c> it cuts but i want to work on it some more and make a better mount for the dremel
[03:52:55] <chr0n1c> and add the limit switches
[03:53:56] <chr0n1c> let me know what you think of it if you take a peek
[03:57:24] <toastydeath> i think it is a row of pictures!
[03:57:27] <toastydeath> and i would be right!
[04:00:53] <CIA-19> 03jmkasunich 07TRUNK * 10emc2/src/hal/drivers/mesa_5i2x/firmware/ (Makefile oneshot.vhd): comment fixes, makefile bugfix
[04:01:13] <K`zan> http://arstechnica.com/news.ars/post/20070412-drm-lock-ins-and-piracy-all-red-herrings-for-a-music-industry-in-trouble.html
[04:02:32] <chr0n1c> nice observation toasty
[04:02:40] <toastydeath> ty
[04:02:41] <chr0n1c> i meant of the machine
[04:03:16] <chr0n1c> i've already got a paying job for it when i get the z-axis aligned and mounted so i can cut steel
[04:03:28] <chr0n1c> ;)
[04:03:32] <chr0n1c> woohoo
[04:04:26] <chr0n1c> i may switch to a trim router instead of a dremel?
[04:12:32] <ds3> how well does the dremel work?
[04:14:04] <chr0n1c> it works nice
[04:14:13] <chr0n1c> it's of kilter though
[04:14:19] <chr0n1c> i need to make a nice mount for it
[04:14:24] <CIA-19> 03swpadnos 07TRUNK * 10emc2/src/emc/usr_intf/axis/scripts/axis.py: Pass the ini file along to the POSTGUI_HALFILE
[04:14:32] <chr0n1c> but for cutting steel i don't know
[04:14:35] <ds3> is there play on up and down the spindle on yours?
[04:14:36] <chr0n1c> i haven't tried yet
[04:14:44] <chr0n1c> i have only done wood and plastic
[04:15:00] <chr0n1c> this one doesn't have much play at all
[04:15:00] <ds3> but all you have done is engraving, right?
[04:15:14] <chr0n1c> the play is in my ghetto wire tie mounting :D
[04:15:37] <chr0n1c> well, i did use it to chop out a part from the plastic i engraved
[04:15:46] <ds3> Hmmm sounds like a nice 30K spindle for engraving
[04:15:49] <chr0n1c> mill out*
[04:16:05] <chr0n1c> it's variable speed also
[04:16:21] <ds3> is this the one they are calling the "Dremel Pro"?
[04:16:57] <chr0n1c> i bought it from sams club, the brand name is worksmith
[04:17:14] <chr0n1c> it was 30 bux in a big kit
[04:22:54] <ds3> nice
[04:23:11] <ds3> is the slanted handle where the LCD for the tach is located?
[04:24:10] <ds3> do you have a DTI for run out measurements on the spindle?
[04:25:16] <CIA-19> 03jmkasunich 07TRUNK * 10emc2/src/hal/drivers/mesa_5i2x/firmware/ (bus-test.vhd businterface.vhd): bus interface logic and preliminary testbed for it
[04:26:10] <CIA-19> 03jmkasunich 07TRUNK * 10emc2/src/hal/drivers/mesa_5i2x/firmware/ (size-nobus.vhd size.vhd): utility files - usefull during development, not part of final design
[04:49:11] <CIA-19> 03jmkasunich 07TRUNK * 10emc2/src/hal/drivers/mesa_5i2x/firmware/5i20-nobus.ucf: use correct name for clock
[04:55:12] <chr0n1c> dti?
[04:55:25] <jmkasunich> dial test indicator
[04:55:27] <ds3> Dial Test Indicator
[04:55:30] <chr0n1c> i jsut bought an indicator
[04:55:36] <chr0n1c> i need to build a stand for it
[04:55:49] <chr0n1c> it's a 0-1" travel dial
[04:56:34] <chr0n1c> i need to get a nice 0-.100 indicator for home
[04:57:18] <chr0n1c> i was thinking of building a mini indicol holder for the dremel spindle, lol
[04:57:38] <chr0n1c> it would be nice to have
[04:57:55] <ds3> to tram it?
[04:57:58] <chr0n1c> yes
[04:58:05] <ds3> heh
[04:58:18] <chr0n1c> or for tramming parts
[04:58:25] <chr0n1c> on the table
[04:59:12] <toastydeath> you're tramming with a .001" indicator?
[04:59:16] <chr0n1c> i probally don't need that kind precision at home for hobby stuff, but it wouldn't hurt anything
[04:59:21] <toastydeath> oh, good point
[04:59:40] <ds3> I am very curious what the run/alignment is like on a dremel
[04:59:44] <chr0n1c> if i build the mini indicol holder i would be
[04:59:58] <ds3> it seems to be priced just right for use as 30K spindle for engraving
[05:00:03] <chr0n1c> i will post results if i ever get around to doing it
[05:00:08] <chr0n1c> yes
[05:00:19] <chr0n1c> i'd go name brand dremel maybe?
[05:00:31] <ds3> Oh yours is not a Dremel brand?
[05:00:35] <chr0n1c> lots of bits and things for it
[05:00:40] <chr0n1c> no it's a knockoff
[05:00:52] <ds3> I have a very old dremel is there is noticeable play on the axis
[05:00:52] <chr0n1c> it is sturdy though i have to say that
[05:01:04] <chr0n1c> go get a new one...
[05:01:13] <chr0n1c> i remember using my dads old dremel...
[05:01:22] <chr0n1c> theyhave advanced a bit since then
[05:01:29] <ds3> unless I know what I am getting it isn't worth it, the dremel does what it does just fine; don't need another one
[05:02:01] <chr0n1c> this one is a "worksmith" from sams club
[05:09:54] <chr0n1c> i jsut updated my cnc page with a description of the macin in case anyone wants to link to it http://ohiopctech.com/cnc/
[05:10:17] <chr0n1c> it should be worthy of at least a mention ;)
[05:13:56] <chr0n1c> how difficult is the mechanics for a haxapod?
[05:14:12] <chr0n1c> and will it do the work of a 4/5 axis mill?
[05:14:41] <chr0n1c> hexapod
[06:05:48] <tomp> geda components for hal updated, wiki has new pix & tbz2, hyput, integ & knob2float added
[07:30:26] <Dallur> morning
[07:31:05] <Dallur> tomp: so you found the bug, nice :)
[12:14:42] <skunkwork> skunkwork is now known as skunkworks
[12:28:00] <tomp> Dallur: have no fix :( the change is noticed by GTK but not by the halpin.
[12:28:09] <tomp> I do not know understand how the pin is connected to the data.
[12:28:16] <tomp> The data is not mentioned in the pin instantiation.
[12:28:18] <tomp> The pin and the data are never lval/rval pairs in an equate or assignment.
[12:40:30] <tomp> pycomp.newpin(halpin, HAL_FLOAT, HAL_OUT) does not mention the variable self.v, so I dont see how it ever works. After i get that bit, then i might be able to see why it does not work in this one case.
[13:07:16] <tomp> Dallur: i see the sequence now, and it comes down to our expectations of 'command(self)' in pyvcp_widgets.py.
[13:07:24] <tomp> we expect the halpin should immediately reflect changes to 'self.value' by the call to self.value = self.v.get() .
[13:07:38] <tomp> I am not sure what gtk thinks is a needed to mark a change, we assume it is 'enter' or 'return', and gtk does not.
[13:07:48] <tomp> we know that inc & dec are recognized by gtk.
[13:07:59] <tomp> i think another field that reflects the same value will show us how gtk works.
[13:08:16] <tomp> if the 2nd field changes with an edit and an 'enter' then the bug is in emc code, else it is on gtk or our expectations of the gtk widget set. ( this is in a non emc context, just gtk & c )
[13:48:32] <lerneaen_hydra> 'lo
[13:51:40] <alex_joni> 'lo
[13:57:03] <lerneaen_hydra> what's happening?
[14:12:27] <Martin_Lundstrom> Hello folks
[14:12:35] <Martin_Lundstrom> Dallur, are you around?
[14:12:53] <Martin_Lundstrom> lerman, are you around?
[14:13:08] <Martin_Lundstrom> sorry lerneaen_hydra !!
[14:15:37] <Martin_Lundstrom> Quiet here today
[14:21:24] <lerneaen_hydra> 'lo
[14:21:31] <lerneaen_hydra> I'm sort of afk
[14:21:40] <lerneaen_hydra> every now and then I come past the computer
[14:25:14] <Dallur> I'm around now :D
[14:28:42] <lerman> Martin_Lundstrom: I'm around, now.
[15:17:09] <skunkwork> cradek: welcome back ;)
[15:17:19] <cradek> yargh!
[15:17:29] <cradek> oops
[15:19:45] <skunkwork> and again ;)
[15:20:10] <cradek> now to stay, I hope
[15:20:11] <SWPadnos> darned dodgy clients ;)
[15:20:52] <cradek> machine crashed last night - nothing in the logs. maybe a power problem - hard to tell as I'm not there
[15:20:58] <SWPadnos> bummer
[15:22:18] <Martin_Lundstrom> Dallur, how is the config going?
[15:22:22] <cradek> yay, SWPadnos fixed an AXIS bug
[15:22:27] <SWPadnos> hah - yay!
[15:22:39] <SWPadnos> I guess I should backport that one
[15:22:53] <cradek> [HAL]SHUTODWN isn't in the branch at all
[15:22:54] <lerman> Martin_Lundstrom: were you looking for me, before?
[15:23:09] <cradek> (I htink)
[15:23:21] <Martin_Lundstrom> lerneaen_hydra, jag skulle ev behova lite hjalp med att svarva en axel, vad tycker du om det?
[15:23:22] <SWPadnos> that was POSTGUI_HALFILE
[15:23:39] <Martin_Lundstrom> lerman, no, sorry my mistake
[15:24:02] <lerman> Ahhh. You typed ler and got lerman instead of lernean...
[15:24:12] <Martin_Lundstrom> yep
[15:25:36] <lerneaen_hydra> Martin_Lundstrom: aha, ok, hur stor ytterdiameter?
[15:29:18] <cradek> SWPadnos: oh, right
[15:29:27] <cradek> SWPadnos: well, thanks
[15:29:29] <CIA-19> 03swpadnos 07v2_1_branch * 10emc2/src/emc/usr_intf/axis/scripts/axis.py: Backport POSTGUI_HALFILE + ini fix
[15:29:33] <SWPadnos> no problem :)
[15:37:51] <SWPadnos> so, I was thinking about making a servo tuning program
[15:38:03] <Martin_Lundstrom> lerneaen_hydra, ca 2.5cm
[15:38:25] <SWPadnos> the (first) problem is that there are a zillion ways to connect to servos due to the different hardware options, etc.
[15:39:16] <SWPadnos> I guess it might make sense to ask for an ini file, then try running the hal files from it, but without EMC / motion
[15:39:41] <SWPadnos> hmmm. actually, I can make a component that provides all the axis.* pins, which would get me most of the way there
[15:42:13] <lerneaen_hydra> Martin_Lundstrom: hmm, I can't fit that through the bore hole in my chuck, it's only 16mm or so, but I do have a tailstock so if they're not too long it shouldn't be much of an issue
[15:42:24] <lerneaen_hydra> where are you now, btw, in sweden or france?
[15:43:50] <lerneaen_hydra> SWPadnos: it would definetly need to read the config from a working setup (just bad PID values), and just ignore the values it's going to adjust
[15:44:20] <SWPadnos> no, I want to make a program that will exercise the axes, and calculate values based on that
[15:44:27] <SWPadnos> so I can't have EMC in the middle
[15:44:54] <lerneaen_hydra> no, but you can still read the INI for data about hardware and stuff
[15:45:27] <SWPadnos> yes. the HAL files making the connections will fail though, since they'll try to connect to motion (which I don't want to load)
[15:45:37] <Dallur> Martin_Lundstrom: really well, everything is moving fine and homing well with gantrykins
[15:45:44] <SWPadnos> though I guess I can load everything, then disconnect motion and reconnect to my own exerciser
[15:45:54] <Dallur> Martin_Lundstrom: Tonight or tomorrow there will be some cutting :D
[15:46:14] <lerneaen_hydra> SWPadnos: exactly, what I mean is you'll need to load the information in the hal file
[15:46:46] <SWPadnos> yep - I figured that ;)
[15:53:09] <Dallur> bbl going home
[15:53:10] <lerman> SWPadnos: Consider splitting the program into two parts. One part could drive the tuning and be written in gcode. The second part could gather the realtime data (impulse response, etc.)
[15:53:52] <SWPadnos> lerman, I was thikning of using halstreamer/halsampler to gather the data
[15:54:27] <SWPadnos> I actually want to do step response though, so I can't use the TP (since that already does accel/vel profiles)
[15:54:28] <lerman> Then you just need a way to get that data into the interpreter for the purpose of analysis.
[15:55:26] <SWPadnos> I'm envisioning a dialog that has some selections for the axis you want to tune, an entry for the excursion length, and the PIDFF values (maybe no FF)
[15:55:58] <SWPadnos> you hit the "tune" button, and it starts knocking the selected axis back and forth by the excursion length, measuring the response as it goes
[15:56:28] <SWPadnos> possibly using an FFT or some other analysis to detect oscillation, then using some easy formula like Z-N or something
[15:57:12] <skunkwork> SWPadnos: that would be cool
[15:58:03] <chr0n1c> ehh i think i just pissed my boss off... it's the first time he's ever been short with me when i called in for the day... guess i better start going in on time next week
[15:58:26] <chr0n1c> but.. this machine is so addictive i stayed up waaaay to late messing with it
[15:58:42] <chr0n1c> hopefully the novelty wears off and i can get back to my life ;)
[15:58:53] <lerneaen_hydra> let's not go into what you did with the machine
[15:59:00] <chr0n1c> lol
[15:59:02] <lerneaen_hydra> could invoke some nasty mental images
[15:59:04] <chr0n1c> i only engraved with it
[15:59:22] <lerneaen_hydra> yeah, of course, that's what it started with
[15:59:25] <skunkwork> wait until you have to come in even when your sick because no one can fix the problem.
[15:59:29] <lerneaen_hydra> but one thing leads to another...
[15:59:56] <chr0n1c> one time.. i wrote this limmerick while i was waiting on the lathe part cycle.. it started with... "if the machines only had breasts..."
[16:00:49] <chr0n1c> the guy who was working in the lathe dept. with me burned the paper i had written it on... he said i was mentally disturbed
[16:02:34] <chr0n1c> haha... that guy used to fall off his stool when the part cycle buzzer went off everytime.. because he never turned the buzzer on.. and while he was in the bathroom or turned away i would flip the buzzer back on
[16:03:11] <chr0n1c> then he finally figured out that he could hardcode the part buzzer into his g-code.. ruined that fun for me
[16:03:36] <tomp> Dallur: i wrote pygtk code to test wether gtk is at fault, it isnt, pygtk can immediately update a label when text field is edited, changes need to be made in pyvcp stuff http://pastebin.ca/438353
[16:03:52] <tomp> Dallur: now to find out where and what to chg
[16:06:03] <lerneaen_hydra> chr0n1c: haha, being disturbed is fun
[16:06:14] <chr0n1c> yes... yes it is!
[16:06:20] <lerneaen_hydra> more insanity to the people!
[16:08:37] <chr0n1c> if.. i change the "G0" to "G1 F10" will that make all rapid moves in the program i am posting only do rapid moves at 100 ipm?
[16:08:53] <chr0n1c> the change would be in my mastercam post**
[16:09:09] <chr0n1c> i mean 10 i[pm*
[16:11:40] <chr0n1c> lemme re-phrase that... if i change the mastercam post processor from "g0" to "g1 f10" will my machine never try to do a rapid move and only do the rapid moves at 10 ipm?
[16:13:46] <cradek> yes but why would you want that?
[16:13:59] <cradek> it might screw up the feed rate for subsequent (real) g1 moves
[16:14:05] <chr0n1c> i am trying to massage the generic mpfan post to work with emc so i don't have to hand edit the code everytime i have a new g-code program
[16:14:34] <chr0n1c> well, i will never really need to go more than 10 ipm? my machine is tiny and i loose steps when i go faster
[16:15:18] <chr0n1c> i think the post as it is adds a f number to all g1 moves anyways...
[16:15:29] <chr0n1c> so i may be ok there...
[16:15:33] <cradek> then configure emc so a rapid is 10ipm
[16:15:43] <lerneaen_hydra> um, shouldn't you be fixing the real problem instead? make it so that g0 doesn't go over 10 ipm
[16:15:45] <jepler> chr0n1c: in that case you should change your .ini file to make the maximum speed on each axis low enough that you do not stall the motors or lose steps
[16:15:54] <cradek> haha
[16:15:58] <chr0n1c> ahh, i see
[16:16:05] <lerneaen_hydra> what a slumbering beast he has awakened
[16:16:06] <cradek> when three people give the same advice you can be pretty sure it's right
[16:16:12] <chr0n1c> i didn't think to do that
[16:16:38] <jepler> as described in the manual the .ini file units are per second, not per minute -- so 10ipm is .166 or something like that.
[16:16:52] <chr0n1c> so the rapid moves is in the stepper ini? i haven't got it all dechipered yet
[16:16:58] <jepler> yes
[16:17:10] <chr0n1c> ok, thanks!
[16:17:20] <jepler> [AXIS_n] MAX_VELOCITY I think
[16:18:07] <chr0n1c> and it is in seconds... so 10ipm/60
[16:18:35] <cradek> yes
[16:19:00] <chr0n1c> that seems a lot easier than editing a post that i know nothing about ... lol
[16:22:38] <chr0n1c> i made a page for my machine pics and description last night it's @ http://ohiopctech.com/cnc/ i glanced throught he wiki, but i didn't see how to add my url to the list
[16:22:55] <chr0n1c> do i need to make an account there?
[16:22:56] <lerneaen_hydra> any sanely set up machine should be able to do everything that you can do from code/interface without breaking stuff
[16:22:57] <jepler> on the front page there is a link called BasicSteps
[16:23:09] <jepler> read that page, it gives the secret to modifying the wiki
[16:23:15] <chr0n1c> ok, ty
[16:23:42] <chr0n1c> and i have permission to add my link and such? don't wanna upset anyone..
[16:23:52] <SWPadnos> we can delete it if it's too upsetting :)
[16:23:57] <chr0n1c> lol, ok
[16:25:12] <jepler> whee somebody's actually milled the axis "splash"
[16:25:37] <lerneaen_hydra> yay
[16:26:05] <lerneaen_hydra> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?IRCquotes <-- I want to see that page expanded greatly!
[16:26:12] <chr0n1c> i engraved it in my table and on some plastic...
[16:26:53] <chr0n1c> it was a perfect test program for my machine since i only ave 3.5" x 11.5" of travel on my x/y config
[16:27:15] <jepler> it's too big for any of the machines I have access to, actually
[16:27:27] <chr0n1c> hmm
[16:28:29] <SWPadnos> I think I've determined that I don't have the slightest understanding of how the whole pyvcp thing works
[16:28:48] <SWPadnos> there are too many levels to it for me to grasp before having (at least) a second cup of coffee
[16:28:57] <skunkwork> jepler: where are you now?
[16:29:00] <jepler> SWPadnos: indeed, I felt that way too
[16:29:04] <jepler> skunkwork: same city
[16:29:14] <jepler> tomorrow is frankfurt (for the second time)
[16:29:29] <jepler> (second time on this trip I mean)
[16:30:11] <SWPadnos> hmmm. how's the eating for a vegetarian over there? I saw lots of sausages, sauerbraten, herring, pig knuckles ..., but not much veggie food
[16:31:56] <SWPadnos> I guess the breads are actually nutritious, which helps
[16:32:24] <jepler> SWPadnos: we get by
[16:32:32] <jepler> today was indian
[16:32:40] <SWPadnos> mmmm. nice
[16:32:58] <lerneaen_hydra> oh noes! emc is being overrun by sentient plant lifeforms!
[16:32:59] <jepler> but yeah the best choices are not the local foods
[16:33:02] <lerneaen_hydra> ;)
[16:33:07] <SWPadnos> I guess my sister is mostly vegetarian, and she's survived the last 12 years r so there ;)
[16:34:36] <chr0n1c> if they don't have mcdonalds in frankfurt i would starve.. lol.. i hate cooking
[16:35:05] <SWPadnos> they have lots of other stuff. just don't try to find anything open on a Sunday
[16:35:35] <jepler> turns out there are 3 days around easter when everything's closed: friday, sunday, monday
[16:35:34] <chr0n1c> hey does anyone use artcam?
[16:36:12] <chr0n1c> i am curious as to what everyone uses for designing thier programs...
[16:36:22] <chr0n1c> g-code programs
[16:36:29] <jepler> but if you eat pigs and cows you can eat in germany -- even on holidays. look around and you'll find some restaurant open.
[16:36:43] <cradek> I use a text editor, or a scripting/programming langauge
[16:37:12] <cradek> hmm sometimes I also use autocad/REALIZE
[16:38:05] <chr0n1c> well, i was playing with artcam last night and got it to successfully post a program that needed no editing afterwards to run in emc
[16:38:53] <chr0n1c> ithe picture istarted with was my website banner... then i created the toolpaths and posted it with the generic g-code (*.tap) post.. .it worked great
[16:39:09] <cradek> great
[16:39:24] <chr0n1c> i cut it into a block of packing foam, it looks neat... not much detail cause i used too big of a cutter though
[16:41:08] <skunkwork> cradek: just got dilbert wwwd (what would wally do) book. :)
[16:41:26] <chr0n1c> dilbert is my fav comic
[16:42:00] <jepler> with emc 2.1 there is image-to-gcode -- you might give it a try
[16:42:14] <jepler> not many people are using it yet, it's a fairly new piece of softwary
[16:42:16] <jepler> software
[16:42:27] <cradek> jepler: did I tell you I finally cut something with it?
[16:42:32] <jepler> cradek: no you didn't
[16:42:35] <jepler> how'd it go?
[16:42:39] <cradek> I finally cut something with it
[16:42:46] <cradek> fine
[16:42:57] <chr0n1c> what kind of machin cradek?
[16:43:01] <chr0n1c> e*
[16:43:15] <cradek> there's unexpected behavior at the edges of the image (you should have at least a tool diameter of empty space I guess)
[16:43:18] <jepler> cradek: anything interesting?
[16:43:35] <lerneaen_hydra> http://www.qdb.us/45116
[16:43:39] <jepler> cradek: hmm I always used images that were "uninteresting" at the edge
[16:44:35] <chr0n1c> how do i get to the image to g-code part of emc?
[16:45:04] <cradek> in AXIS, click open, pick image-to-gcode.py
[16:45:14] <jepler> chr0n1c: there are some magic lines to put in your ini, then you get it when you choose a .png file
[16:45:19] <jepler> it's somewhere in the docs
[16:45:31] <cradek> oh man my instructions were really wrong
[16:45:32] <cradek> ignore me
[16:45:58] <cradek> correct instructions: you can run sim/axis to see it work
[16:46:21] <cradek> then you can copy the right lines from sim/axis.ini into your ini to enable it for your config
[16:46:31] <jepler> chr0n1c: http://linuxcnc.org/docs/2.1/html/gui/image-to-gcode/index.html
[16:46:33] <cradek> ^^ also correct instruction
[16:46:57] <chr0n1c> i was wondering.. how hard would it be to add some sort of runtime clock to the verify and actual program runtime?
[16:47:01] <jepler> argh it doesn't say what section of the ini to add those lines to
[16:47:05] <chr0n1c> ty jepler
[16:47:14] <cradek> chr0n1c: File/Properties
[16:47:27] <jepler> look at the sim/axis.ini file to find out what section to put it in
[16:48:12] <jepler> chr0n1c: as cradek says, axis has an *estimate* built in. It is only an estimate because it neglects accel/decel/blend times.
[16:48:26] <jepler> nowhere does it report the actual time after milling
[16:48:46] <jepler> though I've wished for that too actually
[16:49:12] <chr0n1c> ok... how about a timer that starts when i hit run and pauses when i pause or stop the program.. resets everytime i hit run or start a new program?
[16:49:36] <jepler> yeah that doesn't exist yet
[16:50:06] <cradek> I don't understand why that would be useful - can you explain
[16:50:25] <cradek> (remaining time, I can see)
[16:50:43] <jepler> to me it's mostly a gimmick, not necessarily "useful"
[16:50:52] <chr0n1c> well, it was a feature i liked from the milltronic machine i used to run.. i could verify a program and know how long i could go do something else while the machine was running
[16:51:13] <chr0n1c> like i would set up parts, then actually leave the shop.. some programs take hours to engrave
[16:51:34] <jepler> so you want to know "how much time left now"
[16:51:39] <cradek> I understand, but I don't see why a (forward) running clock helps with that
[16:52:05] <chr0n1c> jep: yes
[16:52:11] <jepler> say it is milling line 1000 of 2000, you want to see the estimate for lines 1000 to 2000 shown somewhere
[16:52:21] <cradek> ok I can see that
[16:52:40] <jepler> one way to get that is to know the total time and the time so far .. but it would be nicer if the computer is the one to compute it
[16:52:50] <chr0n1c> it's really just a fancy pants gimmick.. but it does have some use
[16:52:53] <cradek> bbl, lunch
[16:53:45] <jepler> see you chris
[16:53:45] <chr0n1c> when i verify in mastercam it will tell me the total runtime.. sometimes it's wrong though.. because of feedrate over-rides and such at the machine
[16:54:17] <jepler> the axis estimate also ignores feed override
[16:54:22] <chr0n1c> so i have a basic idea of things i draw in mastercam but any other g-code i have no idea
[16:54:49] <jepler> look at file > properties in axis
[16:54:52] <jepler> it shows an estimate too
[16:55:03] <chr0n1c> i will look into the axis estimate feature.. i have only had emc2/ubuntu installed for about a week now, i am still learning. .. ;)
[16:59:58] <chr0n1c> http://www.qdb.us/leetness LOL!! i have jsut been reading that page
[17:00:06] <chr0n1c> that is some funny stuff
[17:22:44] <Rugludallur> Tomp: hmm so it's in the c code ?
[17:25:05] <tomp> Rugludallur: i havent found it yet, but now am sure that pygtk can update something immediately, wether we can get it to update the halpin immediately is unknown. the way the data was tied to the pins is... weird. the code i pasted is pretty direct.
[17:25:59] <Rugludallur> tomp: i see
[17:26:13] <tomp> did you try the pasted code?
[17:26:41] <Rugludallur> tomp: nope, just came home
[17:31:58] <chr0n1c> if you guys hear of anyone from the dayton ohio area that is going to the cnc party coming up in IL... let me know!
[17:32:14] <chr0n1c> i would really like to go, but my car isn't worthy of driving that far
[17:42:32] <Rugludallur> hmm is that the fest ?
[17:43:03] <Rugludallur> when/where is it ?
[17:43:36] <SWPadnos> June 11-17, Galesburg, IL
[17:43:50] <SWPadnos> http://www.cnc-workshop.com
[17:44:15] <skunkwork> http://www.cnc-workshop.com/
[17:44:18] <skunkwork> oops
[17:44:23] <skunkwork> little slow :)
[17:44:31] <SWPadnos> o k
[17:44:32] <SWPadnos> :)
[17:54:56] <lerneaen_hydra> why can never webpages be coded nicely? the CNC-workshop page isn't exactly easy on the eyes
[17:55:11] <SWPadnos> it's made by machinists
[17:55:28] <lerneaen_hydra> and therefore...?
[17:55:40] <lerneaen_hydra> I assume that they're not complete monkeys
[17:56:01] <SWPadnos> artists are generally better at making eye-pleasing things
[17:56:40] <lerneaen_hydra> it doesn't take much of an artist to see that that page could use some work (and not a lot at that)
[17:57:31] <lerneaen_hydra> like, less garish colors (lower saturation), not a tiny tiled bg, and less mixing of colors (prefferably not all content centered and in the same column, tables or css come to mind)
[17:58:06] <SWPadnos> less flashing ...
[17:58:13] <lerneaen_hydra> flashing?
[17:58:32] <SWPadnos> blinking text
[17:58:50] <SWPadnos> in the ":site directory"
[17:59:04] <lerneaen_hydra> oh, I've only got to the first page
[17:59:14] <lerneaen_hydra> argh!
[18:00:21] <lerneaen_hydra> random q.; is rigid tapping finished?
[18:00:32] <SWPadnos> random a: I don't know
[18:00:35] <SWPadnos> :)
[18:00:39] <lerneaen_hydra> seeing as how thread cutting in lathes is it would seem that it should be done
[18:00:43] <lerneaen_hydra> as it's nearly the same thing
[18:01:09] <SWPadnos> well, the main difference is reversing the spindle, rather than moving out to the drive line and going back to the start of the cut
[18:01:41] <lerneaen_hydra> yeah, but that's nothing difficult to code
[18:01:51] <lerneaen_hydra> not like syncing the motors to the spindle encoder
[18:02:06] <SWPadnos> it either involves multiple motion segments (easy), or "negative time" (hard)
[18:02:28] <lerneaen_hydra> multiple segments seems like the sanest way to implement it
[18:02:31] <SWPadnos> is that the G76 cycle (or is that lathe threading?)
[18:03:19] <lerneaen_hydra> hmm, that's beyond my knowledge, I never muck about in Gcode
[18:10:13] <skunkwork> I like mucking
[18:27:41] <Rugludallur> hmmf has anyone here tried opening up a halmeter, hitting select, it opens the "Select Item to Probe" window, then tried hitting hal meter Select again ?
[18:28:38] <Rugludallur> For me it freezes emc up solid, can't even shut down and clean up with a term signal
[18:28:38] <SWPadnos> only you, I imagine ;)
[18:29:15] <Rugludallur> SWPadnos: This is what happens when you let a stupid user mess around with software
[18:30:02] <Rugludallur> SWPadnos: we start doing things which were not supposed to do
[18:30:20] <skullworks-GAB2> hmm sounds like an item to add to HAL docs under things Not to do.
[18:30:41] <Rugludallur> skullworks-GAB2: ill report it on sourceforge for now
[18:30:50] <SWPadnos> I see the problem in meter.c
[18:31:10] <Rugludallur> SWPadnos: that button should probably stay down while the selection window is open
[18:31:18] <SWPadnos> the second attempt to pop up the select window tries to grab the HAL mutex, which it can't do since it's already held by the meter process
[18:31:36] <SWPadnos> I'll add code to make sure you only get one probe window (once I figure out how to do that ;) )
[18:32:51] <Rugludallur> SWPadnos: ill hold off the bug report for now
[18:33:48] <SWPadnos> ok.
[18:33:53] <cradek> SWPadnos: holding any kind of lock while the user is doing something is normally bad - are you sure that's not the bug?
[18:34:16] <SWPadnos> that may not be the problem, but it did leap out at me
[18:34:31] <cradek> what exactly does the mutex lock?
[18:34:37] <SWPadnos> it probably holds the lock only while getting the pin list
[18:34:53] <SWPadnos> all the HAL data (but not from RT access, so it's safe to hold it all day)
[18:35:11] <SWPadnos> unless you then try to grab it again in the same process ...
[18:35:39] <SWPadnos> I'm adding a flag to say that the window is already up, and refocus it instead of trying to recreate it ...
[18:36:23] <cradek> even with that fix, couldn't a second halmeter have the same problem?
[18:37:04] <SWPadnos> if the mutex is the problem, then it's still a different process, so once you release the lock in one halmeter, the other would be able to run
[18:37:29] <SWPadnos> if locking is the issue, then it's because one process blocks on itself
[18:37:52] <cradek> the lock is so the list of hal pins stays consistent while you traverse it, right?
[18:38:15] <cradek> sorry I should look at the code
[18:38:29] <skullworks-GAB2> I'm looking on the Wiki index - where would be a good spot to post a M code proposal?
[18:39:37] <SWPadnos> cradek, yes, that's why it's there. though the function that uses the lock only builds the lists and shows the window. it doesn't wait around for the user to make a selection
[18:41:01] <SWPadnos> so it's some other issue, which would also be fixed by making sure that the popup only pops up one at a time
[18:41:04] <cradek> so it's released as soon as the list is built for that dialog?
[18:41:17] <SWPadnos> yes
[18:41:28] <cradek> ok, I totally misunderstood then
[18:41:41] <SWPadnos> me too - I needed to get to the bottom of the function to see the release ;)
[18:42:19] <SWPadnos> Rugludallur, can you try clicking select, selecting something (and click the OK button), then click select in the meter window twice?
[18:42:51] <SWPadnos> (I only have sim right now, so I'm sure the locking works differently)
[18:44:27] <cradek> actually I think that locking might be the same...?
[18:46:03] <SWPadnos> could be - I guess I could check ;)
[18:46:15] <Rugludallur> SWPadnos: even worse than before
[18:46:42] <Rugludallur> SWPadnos: select, selected a pin, hit select again and then again and now it locks up in a loop refreshing the pin list
[18:47:35] <SWPadnos> that's not quite what I asled for :)
[18:47:39] <SWPadnos> asked
[18:47:46] <SWPadnos> wait a bit - it may finish eventually
[18:47:56] <cradek> mine does finish after a while
[18:48:03] <SWPadnos> (on my Opteron, it takes several seconds to refresh the lists
[18:48:05] <SWPadnos> )
[18:48:33] <Rugludallur> SWPadnos: been more than a minute here and I can see the list being refreshed over and over and over
[18:48:33] <cradek> it's in popup_probe_window/gtk_clist_append
[18:48:35] <SWPadnos> I don't get a lockup with sim. in fact, I can click exit in the main window, and the selection window goes away as well
[18:48:50] <SWPadnos> it gets refreshed for each item that gets added, I suspect
[18:48:50] <cradek> it takes a few seconds here
[18:49:09] <cradek> ahh and maybe it's slower because it's already mapped
[18:49:28] <SWPadnos> I'm using a slightly modified demo-step-cl config, so I have a fair number of pins/params to refresh
[18:49:46] <Rugludallur> SWPadnos: I actually did what you asked, just forgot to add the hit OK button in between
[18:50:11] <SWPadnos> but as a comparison, it takes about 1/4 as much time as a full recompile of emc ...
[18:50:16] <SWPadnos> ok
[18:50:28] <Rugludallur> SWPadnos: hmm my base period is 15000, that might be a factor
[18:50:30] <SWPadnos> per button press
[18:50:56] <SWPadnos> yeah, if you don't have a lot of spare CPU for user tasks, it'll take a while
[18:51:44] <cradek> if (probe->window == NULL) {
[18:51:43] <cradek> create_probe_window(probe);
[18:51:43] <cradek> - }
[18:51:45] <cradek> + } else return;
[18:54:17] <cradek> raising the window might be nicer, but this fixes the bug
[18:56:23] <SWPadnos> does window->destroy reset that pointer to null?
[18:58:37] <cradek> the docs I found do not say, but it works when I test it
[19:01:37] <Jymmm> cradek what do you use to cut pcb's with? Would a paper cutter work?
[19:02:03] <SWPadnos> well, it appears the pointer is NULLed when the window is destroyed
[19:02:28] <SWPadnos> Jymmm, probably not. glass epoxy is quite a bit harder than paper - you'd run through cutters like toilet papre
[19:02:32] <SWPadnos> papaer
[19:02:33] <SWPadnos> paper
[19:02:36] <cradek> Jymmm: carbide
[19:02:54] <skullworks-GAB2> try a micro100 engraving tool
[19:02:54] <cradek> Jymmm: anything else is immediately destroyed
[19:03:13] <Jymmm> cradek score and snap? like you do with glass?
[19:03:24] <SWPadnos> doesn't work too well
[19:03:28] <cradek> no I cut with an end mill
[19:03:53] <SWPadnos> V-scoring is available from board houses, but they route a groove, and you need to be very careful during the "snap" phase, or SMT components will break
[19:04:04] <Jymmm> Ah. I was thinking of buying some pcb's at RS, then cutting into strips
[19:04:17] <SWPadnos> or the contacts will break off, even if the component itself doesn't break
[19:04:42] <Jymmm> SWPadnos not a good thing (TM)
[19:04:54] <SWPadnos> nope. and hard to debug, too
[19:05:04] <cradek> RS pcbs are not FR4 and they cut easily
[19:05:30] <Jymmm> cradek what are RS pcb's ?
[19:05:35] <cradek> crap
[19:05:45] <Jymmm> crapping is good
[19:05:46] <cradek> something soft, I don't know what it is
[19:05:52] <Jymmm> else your full of shit =)
[19:06:02] <Jymmm> <rimshot>
[19:06:19] <skullworks-GAB2> * skullworks-GAB2 runs for cover...
[19:06:36] <SWPadnos> they seem to be just the binder / epoxy stuff, no fiberglass in there
[19:06:53] <SWPadnos> but it's been a while since I've seen an RS PCB
[19:07:26] <Jymmm> Not like I'm going to have a lot of stress on them
[19:07:37] <SWPadnos> you will when you try to break them apart
[19:07:43] <Jymmm> lol, true
[19:07:49] <SWPadnos> they tend to flex a bit before breaking
[19:08:15] <SWPadnos> you basically can't put any components closer than 1/4" or so from a score line
[19:08:29] <Jymmm> I have a table saw with a 1/16" blade on it.
[19:08:32] <SWPadnos> which means the minimum dimension is 1/2 inch or more
[19:08:37] <skunkwork> just don't use the old bakalite or what ever it is called copper clad
[19:08:54] <SWPadnos> that may work, but will probably still take out 1/8 inch - the teeth should be pitched left/right of the band
[19:08:55] <skunkwork> m:
[19:08:58] <skunkwork> oops
[19:09:10] <Jymmm> skunkwork stinky
[19:09:35] <Jymmm> SWPadnos it's just a 4" table saw
[19:09:36] <skunkwork> and the traces come off if you look at them funny ;)
[19:09:43] <Jymmm> no offset teeth
[19:09:49] <Rugludallur> I'm not to good with my gcode but running a sim program and entering m3 followed by a g1 x10y10 should turn motion.spindle-on = true right ?
[19:10:11] <SWPadnos> only if spindle speed is nonzero, I think (?)
[19:10:19] <skullworks-GAB2> M3 might require a Sxxxx ?
[19:10:22] <Rugludallur> swpadnos: set f10 before that
[19:10:28] <SWPadnos> that's feed, not spindle speed
[19:10:30] <skunkwork> s word
[19:10:42] <Rugludallur> sorry :I
[19:11:06] <Rugludallur> skullworks-GAB2: right on :D
[19:11:49] <skullworks-GAB2> speaking of M3...
[19:12:11] <skullworks-GAB2> I have a proposal for M41, M42
[19:12:33] <skullworks-GAB2> not sure where to put it on the wiki
[19:13:00] <skullworks-GAB2> I see 2 issues it could resolve
[19:13:13] <skullworks-GAB2> one is a problem I will have
[19:13:42] <skullworks-GAB2> the other would be for users of a converted Bridgeport step pulley head
[19:15:34] <skullworks-GAB2> M41 = calls (or defines) low gear and M42 is high gear.
[19:16:01] <cradek> on the mazak, gear changes happen automatically to get the specified S speed
[19:16:42] <skullworks-GAB2> with a relay controlled spindle on a bridgeport M3 in low = M4 in high
[19:16:56] <cradek> it's done in ladder (switching gears is complex)
[19:17:06] <skullworks-GAB2> because the backgear reverses spindle rotation
[19:17:18] <cradek> yow
[19:17:27] <cradek> all the better to do in hal :-)
[19:17:35] <skullworks-GAB2> yeah - but my proposal is for semi-manual machines
[19:19:01] <cradek> if you'd put a sensor on your gear change lever, you could feed that signal into HAL so it could compensate for the negative
[19:19:24] <cradek> it could also then send the right voltage (scaled) to the vfd if you have one
[19:19:53] <skullworks-GAB2> yeah - thats the ideal - but for those without the fancy hardware...
[19:20:19] <cradek> hardcoding the details of a particular machine (two gears, one is reversed) into the core of emc is not going to be accepted - that kind of thing is what HAL lets you handle
[19:21:19] <skullworks-GAB2> What I was going to propose is being able to define M41 Px to allow prompting
[19:21:41] <skullworks-GAB2> function would be thru HAL
[19:21:54] <cradek> maybe I should let you write it without interrupting, sorry
[19:22:32] <skullworks-GAB2> no - interupting is fine - keeps things on track
[19:24:13] <chr0n1c> ok, i'm taking a break from building my sturdy, tram-able z-axis mount ;)
[19:24:19] <skullworks-GAB2> likewise my problem will be that with the speed control for the seig X3 there is no software method of spindle direction
[19:24:31] <SWPadnos> there's a "hal manual toolchange" component, a similar one should be doable for spindle speed setting (and gearchanging)
[19:24:58] <skullworks-GAB2> you have to manually switch a polarty output switch on the motor leads
[19:25:13] <SWPadnos> hmmm. though I don't think we have a spindle speed interlock like the toolchange interlock
[19:25:30] <chr0n1c> i forgot how much a pain in the arse cutting steel with a hacksaw could be... since i don't have and power saw at home other than a jigsaw...
[19:25:33] <chr0n1c> whew!
[19:26:02] <Jymmm> chr0n1c: jig saw + hacksaw blade
[19:26:12] <chr0n1c> is there a general CNC channel... not specifally for emc
[19:26:23] <alex_joni> chr0n1c: this is the closest to that
[19:26:26] <chr0n1c> jymm! you are a genius
[19:26:28] <skullworks-GAB2> what will happen if I use a spindle encoder is that the step/dir commands for the spindle motor will have different values for hi/lo
[19:26:39] <Jymmm> chr0n1c: http://power-tools.hardwarestore.com/54-367-jig-saw-blade-sets/bosch-high-carbon-steel-t-shank-jigsaw-blade-284117.aspx
[19:26:54] <chr0n1c> i have a few extra blades in the garage
[19:27:02] <chr0n1c> i could dremel one to work
[19:27:11] <chr0n1c> maybe put two of them side by side
[19:27:13] <alex_joni> oh-oh.. you need to get it very concentric
[19:27:15] <Jymmm> chr0n1c they make jigsaw blades for steel
[19:27:32] <chr0n1c> yeah...
[19:27:32] <skullworks-GAB2> I will need to use HAL to redefine step/dir valuse based on state of M41|M42
[19:27:38] <chr0n1c> i didn't wanna drive to the store
[19:27:44] <chr0n1c> so i am going with what i have here
[19:27:50] <chr0n1c> lol
[19:28:00] <chr0n1c> it's exercise anyways... the hacksa that is
[19:28:06] <chr0n1c> hacksaw*
[19:28:19] <SWPadnos> skullworks-GAB2, what are M41 and M42 normally? (or are you using some otherwise unused M codes?)
[19:28:43] <alex_joni> skullworks-GAB2: you can easily use M101 M102
[19:28:50] <alex_joni> those are custom M-codes definable by the user
[19:29:00] <Jymmm> chr0n1c: Is a $1.73 blade worth the trouble? http://www.amazon.com/Black-Decker-Metal-Cutting-Jigsaw/dp/B000LNKE4S/ref=sr_1_21/104-7333755-7854304?ie=UTF8&s=hi&qid=1176492485&sr=1-21
[19:29:01] <skullworks-GAB2> normally M41 calls a shift to low gear - M42 + high
[19:29:05] <SWPadnos> interesting: http://www.machinemate.com/FullListCodes.htm
[19:29:11] <chr0n1c> lol.. but i want it now!
[19:29:34] <Jymmm> chr0n1c: Then get your ass off the computer, pickup the car keys and GET ER DONE!
[19:29:36] <chr0n1c> i had the motivation to do the z-axis, i wanted to get started before i lost my train of thought
[19:29:44] <chr0n1c> i just like bitching i guess
[19:29:56] <SWPadnos> interesting redefinition of G20/G21 there
[19:30:00] <skullworks-GAB2> likewise most machines will shift as needed based on the M3/M4 Sxxxx value
[19:30:04] <chr0n1c> driving to the store would have definately killed the motivation
[19:30:26] <skullworks-GAB2> but there are times when you want to override the auto gear selection
[19:30:47] <chr0n1c> lol, $1.73
[19:31:58] <skullworks-GAB2> <SWPadnos> interesting redefinition of G20/G21 - they use the G71
[19:32:09] <skullworks-GAB2> which is also common
[19:32:29] <SWPadnos> yeah - apparently it's in the DIN 66025 standard
[19:32:39] <skullworks-GAB2> some machines accept either
[19:32:40] <SWPadnos> (though I haven't seen that document)
[19:39:49] <Rugludallur> phahh standards, the only good thing about them is that there are so many to choose from
[19:40:09] <skullworks-GAB2> * skullworks-GAB2 takes that as a joke...
[19:40:38] <Rugludallur> * Rugludallur :D
[19:44:56] <skullworks-GAB2> cradek : do you have a programmable speed control on your mini lathe?
[19:45:23] <cradek> on my lathe there's just a knob to turn. the nist lathe has full speed control from emc
[19:46:30] <skullworks-GAB2> have you looked at the CNC4PC C6 speed controller board - its a cheap otion.
[19:48:00] <skullworks-GAB2> Only problem is the SEIG spindle board does not have a way to reverse the spindle other than the polarity switch to the motor.
[19:49:28] <skullworks-GAB2> I was thinking about using a DPDT relay and a FAT diode (to protect the board from back EMF)
[19:51:37] <feoc> moo
[20:00:00] <alex_joni> apt-get moo
[20:00:08] <feoc> heh
[20:00:21] <alex_joni> have you mooed today?
[20:00:32] <feoc> :)
[20:01:56] <Rugludallur> package moo is blocked by package bull
[20:02:14] <feoc> think i gonna have to buy new axis and spindle controls for my lathe
[20:02:26] <feoc> and i dont think they will be cheap :(
[20:03:33] <feoc> the ones on it are ancient siemens ones
[20:03:39] <feoc> cant find any info about em
[20:03:43] <feoc> and there all in german
[20:07:09] <chr0n1c> use babelfish to translatethe german?
[20:10:44] <feoc> well the writing on the boards is in german
[20:10:58] <feoc> tbh im probaly better off just getting new controls
[20:11:09] <feoc> no idea if they work or not
[20:11:16] <feoc> and even less clue to how to set em up
[20:11:28] <chr0n1c> i wish i wish i had a small lathe... even if it was manual
[20:11:54] <feoc> i just wished my big lathe worked
[20:12:04] <feoc> stupid siemens computers
[20:12:17] <chr0n1c> ;) can you use it manual?
[20:12:29] <feoc> nope
[20:12:34] <chr0n1c> hmm...
[20:12:50] <chr0n1c> sonofabiotch
[20:12:59] <feoc> getting a speed and feed overide error
[20:18:07] <SWPadnos> Rugludallur, can you try applying the patch at http://pastebin.ca/438713 to meter.c, and tell me if the select problem goes away
[20:18:24] <Rugludallur> SWPadnos: just a sec
[20:19:21] <SWPadnos> I thikn the screen updates were what took so long, the patch just prevents updates until all the list manipulations are done
[20:20:04] <SWPadnos> the advantage of this over cradek's fix (and the one I had thought of) is that you can leave the selection window open and refresh the lists whenever you want (ie, if you load/unload some comps)
[20:23:53] <feoc> how can you load axis so it loads up as full screen ?
[20:25:03] <chr0n1c> i would like to know the answer to that as well
[20:25:16] <chr0n1c> full screen axis startup?
[20:25:40] <SWPadnos> there are two ways I can think of (other than changing AXIS itself)
[20:26:01] <SWPadnos> 1) there may be some size/position entries that AXIS will use in the X resource database
[20:26:38] <SWPadnos> 2) you can proabbly change the emc script to provide a --geometry=1024x768@0x0 (or similar) argument to the DISPLAY program
[20:27:06] <SWPadnos> you'd need to look at the specifics of either of those methods
[20:27:16] <feoc> k
[20:28:34] <feoc> just makes it easyer for me if other people who might use the machine dont even see linux
[20:28:45] <feoc> other than just logging in and emc loading up
[20:33:03] <SWPadnos> you can actually have that happen automatically as well
[20:33:15] <feoc> oh?
[20:33:58] <SWPadnos> go to System->Administration->Login Window
[20:34:13] <SWPadnos> on the Security tab, there's an option to auto-login a user
[20:34:36] <SWPadnos> I'm not exactly sure how you get EMC to run automatically, but I'm sure it's possible
[20:35:40] <skullworks-GAB2> this sounds like a prime wiki topic... Autostart EMC
[20:36:58] <robin_sz> meep?
[20:37:24] <CIA-19> 03swpadnos 07TRUNK * 10emc2/tcl/bin/pickconfig.tcl: Make keypad enter button act the same as the return key (press ok)
[20:38:02] <alex_joni> SWPadnos: you can just have gdm start up emc/AXIS instead of gnome
[20:38:42] <SWPadnos> that's one option :)
[20:39:51] <Rugludallur> SWPadnos: works great now :D
[20:39:54] <chr0n1c> so when gdm starts emc... the machine would not do anything else?
[20:40:15] <CIA-19> 03swpadnos 07v2_1_branch * 10emc2/tcl/bin/pickconfig.tcl: Backport fix so keypad enter presses OK button
[20:40:22] <chr0n1c> would samba and the likes still work?
[20:40:29] <SWPadnos> Rugludallur, great. I'll check that one in :)
[20:40:59] <robin_sz> * robin_sz considers a cnc control for his spot welder
[20:41:10] <alex_joni> chr0n1c: there would be no other interface available
[20:41:13] <robin_sz> actually, a plain old PLC will be fine
[20:42:29] <chr0n1c> i found the perfect box for my step driver and power supply... an old three bay external hot swap drive case...
[20:42:38] <chr0n1c> it's made of aluminum and some plastic
[20:42:42] <chr0n1c> very sturdy
[20:42:52] <chr0n1c> has a fan port and power plugs already in the back
[20:43:23] <chr0n1c> there is already a 12/5 volt power supply in it
[20:43:30] <chr0n1c> and some blinky leds ;)
[20:44:21] <CIA-19> 03swpadnos 07TRUNK * 10emc2/src/hal/utils/meter.c:
[20:44:21] <CIA-19> Prevent screen updates to probe selection lists while scanning HAL.
[20:44:21] <CIA-19> This fixes a problem reported by Rugludallur where halmeter would appear to
[20:44:21] <CIA-19> freeze if "select" was pressed while the selection dialog was visible.
[20:44:23] <chr0n1c> so i have two 5 1/4" drive bays... and the 3 bay space behind them for things to be stuffed in there
[20:46:51] <CIA-19> 03swpadnos 07v2_1_branch * 10emc2/src/hal/utils/meter.c: Backport probe selection window redraw fix
[20:53:00] <feoc> can anyone spare a few to talk me through using halscope?
[20:53:08] <feoc> im lost with it
[20:54:15] <cradek> did you read the halscope section in the manual?
[20:55:17] <feoc> erm
[20:55:19] <feoc> i tryed
[20:55:30] <alex_joni> it's just soo many words
[20:55:35] <feoc> :(
[20:55:42] <alex_joni> feoc: just kidding
[20:55:53] <alex_joni> feoc: did you select the channels you want to scope?
[20:56:00] <feoc> tis a lot to have to get my head around
[20:56:02] <feoc> yeah i did
[20:56:06] <feoc> but nothings apearing
[20:56:11] <alex_joni> hang on
[20:56:20] <alex_joni> for something to appear you need to trigger
[20:56:39] <alex_joni> first you select the trigger channel
[20:57:05] <alex_joni> then trigger mode (upslope, downslope, etc)
[20:57:42] <feoc> i did the select source
[20:57:49] <feoc> set it to the pos cmd bit
[20:58:05] <alex_joni> ok.. now you need to figure out what you want
[20:58:14] <alex_joni> single shot or continuous
[20:58:38] <alex_joni> under Run Mode you have A. Normal, B. Single, C. Stop
[20:58:46] <feoc> basicly im trying to do what cradek was showing me yesterday in this pic
[20:58:46] <feoc> http://timeguy.com/cradek-files/emc/tuned.png
[20:59:29] <alex_joni> you probably want single for that
[21:00:24] <feoc> ok
[21:01:58] <feoc> ah ok i seem to have some lines up now
[21:02:45] <feoc> nothing usefull im just trying to figure out halscope on my home comp using the live cd
[21:08:16] <feoc> iv never spent so much time reading about any program in my life
[21:08:29] <alex_joni> feoc: ever used a real scope?
[21:08:48] <feoc> not reall
[21:09:10] <alex_joni> well.. you wouldn't have needed to read up on it then..
[21:09:11] <feoc> altho i was refering to reading about emc in general lol
[21:09:25] <alex_joni> well.. I'm still reading about it :P
[21:09:32] <feoc> heh
[21:09:36] <alex_joni> and it's been 2 years + now
[21:09:45] <feoc> fair enough
[21:10:17] <feoc> gotta admit emc is a good program
[21:10:26] <alex_joni> it's come a long way
[21:10:41] <feoc> just makes things look deceptivly easy till you actually try to set things up
[21:10:57] <feoc> i assume stepper motors are a lot simpler to get going ?
[21:11:47] <alex_joni> yeah a bit
[21:11:55] <alex_joni> but servos aren't that hard either
[21:12:02] <alex_joni> (the 5th or 6th time :P)
[21:12:09] <feoc> lol
[21:12:23] <feoc> ill let you know in 6 axis's time
[21:15:00] <feoc> so this pid tuning stuff
[21:15:11] <feoc> whats the best/easyest way to go about it ?
[21:15:25] <feoc> the simple or Z-N method/
[21:15:48] <alex_joni> I'd do the Z-N
[21:17:21] <feoc> how do i find out the Kc and Pc from halscope?
[21:20:59] <alex_joni> feoc: you count lines
[21:21:07] <alex_joni> then you look at lines / division
[21:21:31] <feoc> in what way?
[21:21:57] <feoc> what bit do i need to count
[21:25:15] <skunkwork> you need to figure the period of oscillation
[21:26:05] <feoc> is that from counting the waves per division ?
[21:26:29] <feoc> also on what bit do i need to count?
[21:26:41] <feoc> feedback or f-error ?
[21:28:00] <Jymmm> * Jymmm just can't win... replaced laptop with pc + crt. pc+crt+drivers powered up is 3a, now I need to plugin a 8port switch and it draws almost 1a. *sigh*
[21:29:25] <skullworks-GAB2> so what is the problem (unless running off generator...)
[21:30:12] <Jymmm> spindle + shopvac draws 11a, not including the lights in the garage
[21:30:57] <skullworks-GAB2> on a 15A breaker?
[21:31:02] <Jymmm> yep
[21:31:59] <skullworks-GAB2> does switch have to be in the garage?
[21:32:38] <Jymmm> no other place for it, unless I could split the one cat5 cable for two connections.
[21:32:53] <skullworks-GAB2> why 2?
[21:33:17] <skullworks-GAB2> you need 2 at same time?
[21:33:32] <skullworks-GAB2> can't just swap cable/
[21:33:43] <Jymmm> The livingroom shares the common wall with the garage, and the TV + TVPC share the same ciruit too.
[21:34:07] <alex_joni> good night all
[21:34:10] <Jymmm> nite
[21:34:15] <skullworks-GAB2> night alex
[21:34:19] <feoc> im off too
[21:34:21] <feoc> gnight
[21:34:35] <skullworks-GAB2> nite
[21:36:02] <Rugludallur> Jymm: you can "split" a cat5 cable just fine
[21:36:30] <Jymmm> Rugludallur: But a PITA in terminating.
[21:36:32] <Rugludallur> Jymmm: Ethernet only uses 4 wires so you can in fact run two
[21:36:50] <Rugludallur> Jymmm: most proper network thops sell a small multiplexor which you can use to hook it up
[21:37:09] <Rugludallur> Jymmm: s/thops/shops/
[21:37:38] <Jymmm> I might have one, will have to look around. will need two of them.
[21:47:49] <CIA-19> 03jmkasunich 07TRUNK * 10emc2/src/hal/drivers/mesa_5i2x/firmware/Makefile: fix naming of intermediate files
[21:58:27] <SWPadnos_> SWPadnos_ is now known as SWPadnos
[22:05:38] <Jymmm> Great.... I only have one cat5 splitter.... *sigh*
[22:20:16] <chr0n1c> woohoo... i got my max_velocity set and figured out the homing stuff... after i setup the sof limits i will be in mini-cnc haven :D
[22:20:36] <skullworks-GAB2> WHAT MACHINE TYPE?
[22:20:38] <chr0n1c> i already have sambe going so i can save my work from my windows box right to the machine's g-code folder
[22:20:42] <skullworks-GAB2> oops
[22:20:52] <skullworks-GAB2> caplock...
[22:21:21] <chr0n1c> http://ohiopctech.com/cnc
[22:21:35] <chr0n1c> i re-did the z-axis today.. it is a little better
[22:21:39] <chr0n1c> no pics yet...
[22:22:13] <skullworks-GAB2> what old GTW box?
[22:23:16] <chr0n1c> it's a p2 233mhz mmx
[22:23:37] <chr0n1c> 333 mb ram
[22:23:42] <skullworks-GAB2> hmm sounds like a 400286 mbd
[22:23:54] <chr0n1c> huh?
[22:23:59] <chr0n1c> the model #
[22:24:00] <chr0n1c> ?
[22:24:07] <chr0n1c> of the mb
[22:24:40] <skullworks-GAB2> * skullworks-GAB2 worked warranty repair and tech support for GTW for about 3 yrs
[22:24:56] <chr0n1c> built in video card/network/usb
[22:24:56] <skullworks-GAB2> motherboard
[22:25:17] <skullworks-GAB2> ok - not a 4000286
[22:25:18] <chr0n1c> ooops i was wrong no built in video
[22:25:30] <skullworks-GAB2> that used a AGP card
[22:25:40] <chr0n1c> i think it has a rage vid card
[22:26:02] <chr0n1c> but it runs everything just fine... a tad bit sluggish
[22:26:07] <skullworks-GAB2> I love there slot 1 P3 boards - nearly bullet proof
[22:26:08] <chr0n1c> nothing unusable so far
[22:26:41] <chr0n1c> this machine is coming together so smoothly
[22:26:51] <chr0n1c> i am so happy with emc right now
[22:27:06] <skullworks-GAB2> weak link in that sys will be the powersupply...
[22:27:09] <chr0n1c> as long as it only gets better and no major bug i think you all are on a great path
[22:27:33] <skullworks-GAB2> agreed
[22:27:33] <chr0n1c> i already changed out the power supply it died a few months back when i was using the box as a webserver
[22:27:54] <skullworks-GAB2> figures
[22:28:41] <skullworks-GAB2> they underpowered most of there systems - got FED tax credits for higher "Energy Star" ratings.
[23:19:13] <Rugludallur> hmm I'm trying to make a parameter available as a pin and untill now I was under the impression that "constant" did exactly that but I'm starting to think it's not capable of that, can anyone shed light on this ?
[23:20:24] <Rugludallur> From what it looks like all constant can do is make a parameter available which sets a pin, but I can't see how that is usefull when you can use a signal instead or directly set a pin