#emc | Logs for 2006-07-24

[03:09:05] <A-L-P-H-4> A-L-P-H-4 is now known as A-L-P-H-A
[04:26:49] <CIA-12> 03cradek 07HEAD * 10emc2/src/Makefile: remove debug output
[04:41:03] <Jymmmm> A-L-P-H-A not like here, where it's all rolled into one =)
[04:41:26] <A-L-P-H-A> heh
[04:41:53] <Jymmmm> A-L-P-H-A but there's not 400 ppl in here either.
[04:41:59] <A-L-P-H-A> true
[04:42:12] <A-L-P-H-A> how do I build a rocket ship to fly to moon, with DIY components?
[04:42:35] <Jymmmm> sit on a box of dynamite?
[04:44:21] <A-L-P-H-A> see, that doesn't have the explosive power needed to send anything into orbit... or if so, the amount sould be so great, it's unfeasable.
[04:44:48] <Jymmmm> mfg plant?
[04:45:20] <A-L-P-H-A> some rich guy, toy maker, wanted to build one out of hydrogen peroxide.
[04:45:33] <A-L-P-H-A> distilled hydrogen peroxide.
[04:45:38] <A-L-P-H-A> cause the strong stuff, you can't buy.
[04:45:47] <A-L-P-H-A> but makes awesome rocket fuel, he says.
[04:52:47] <jmkasunich> you aren't talking about John Carmack are you?
[04:52:59] <A-L-P-H-A> no
[04:53:04] <A-L-P-H-A> he doesn't make toys, he makes games.
[04:53:28] <jmkasunich> but he is working on DIY rockets, and for a while he was using hydrogen peroxide
[04:53:49] <jmkasunich> I think he's moved up to LOX/Alcohol now
[04:54:25] <A-L-P-H-A> no, it was some other eccentric money bags
[05:09:46] <Jymmmm> The company that fuels NASCAR lost their permit for high-perfomance fuel, so NASCAR next season will be using regular fuel. They can only make high perf fuel for the military.
[05:38:55] <jmkasunich> sounds like they have their priorities right for a change...
[05:38:59] <jmkasunich> goodnight all
[05:39:12] <Jymmmm> G'Noght john
[06:21:00] <anonimas1> morning
[06:24:50] <A-L-P-H-A> hey anonimas1.
[06:24:55] <A-L-P-H-A> haven't seen you around in a while
[06:45:30] <anonimas1> anonimas1 is now known as anonimasu
[06:45:37] <anonimasu> yeah not actively
[06:46:57] <anonimasu> got to go to work now
[06:49:12] <NickServ> This nickname is owned by someone else
[06:49:12] <NickServ> If this is your nickname, type /msg NickServ IDENTIFY <password>
[06:54:43] <NickServ> This nickname is owned by someone else
[06:54:43] <NickServ> If this is your nickname, type /msg NickServ IDENTIFY <password>
[07:04:29] <NickServ> This nickname is owned by someone else
[07:04:29] <NickServ> If this is your nickname, type /msg NickServ IDENTIFY <password>
[07:12:14] <NickServ> This nickname is owned by someone else
[07:12:14] <NickServ> If this is your nickname, type /msg NickServ IDENTIFY <password>
[09:32:13] <root> hi there
[09:32:53] <root> root is now known as EvertL
[09:34:01] <EvertL> i'm trying to run emc2 with rtai 3.4 (vulcano). it all goes well up untill shutdown, where we get a:
[09:35:35] <EvertL> ERROR: Can't remove rtai modules, kill the following process(es) first
[09:36:21] <EvertL> /dev/rtai_shm root 4078 ....m io
[09:36:56] <EvertL> i see a similar error on an irc log about classicladder:
[09:37:16] <EvertL> does anybody know how to handle this?
[09:40:07] <EvertL> we can't manually stop the process
[09:42:09] <EvertL> have to reboot again........ :(
[09:45:37] <root> root is now known as EvertL
[09:52:46] <EvertL> still the same...
[09:56:34] <EvertL> and this also happens, strangly enough not every time: Loading Real Time OS, RTAPI, and HAL_LIB modules
[09:56:34] <EvertL> Starting EMC IO program: io
[09:56:35] <EvertL> iocontrol: machine: 'EMC-HAL' version '1.2'
[09:56:35] <EvertL> Running HAL config file /root/emc2/configs/motenc//core_servo.hal
[09:56:36] <EvertL> Running HAL config file /root/emc2/configs/motenc//motenc_motion.hal
[09:56:36] <EvertL> Running HAL config file /root/emc2/configs/motenc//motenc_io.hal
[09:56:37] <EvertL> Starting EMC TASK program: milltask
[09:56:39] <EvertL> task: machine: 'EMC-HAL' version '1.2'
[09:56:41] <EvertL> Starting EMC DISPLAY program: tkemc
[09:56:43] <EvertL> emc/task/iotaskintf.cc 162: Command to IO level (EMC_LUBE_OFF:+1505,+12, +0,) timed out waiting for last command done.
[09:56:46] <EvertL> emc/task/iotaskintf.cc 165: emcIoStatus->echo_serial_number=1, emcIoCommandSerialNumber=2, emcIoStatus->status=1
[09:56:48] <EvertL> emc/task/iotaskintf.cc 170: Last command sent to IO level was (EMC_AUX_ESTOP_ON:+1206,+12, +2,)
[09:56:54] <EvertL> the last three lines keep repeating themselves
[10:06:33] <EvertL> the taskinterface seems to have some problems
[10:06:49] <EvertL> iotaskintf
[13:39:24] <Lerneaen_Hydra> 'lo
[13:39:46] <ValarQ> ho
[13:55:36] <CIA-12> 03jepler 07HEAD * 10axis/extensions/emcmodule.cc: improvements for nontrivial kinematics
[13:55:39] <CIA-12> 03jepler 07HEAD * 10axis/scripts/axis.py: improvements for nontrivial kinematics
[13:55:39] <CIA-12> 03jepler 07HEAD * 10axis/tcl/axis.tcl: improvements for nontrivial kinematics
[14:27:07] <CIA-12> 03jepler 07HEAD * 10emc2/src/emc/kinematics/tripodkins.c: remove unused code
[15:07:03] <CIA-12> 03jepler 07HEAD * 10axis/extensions/emcmodule.cc: backplot colors for probing and toolchange; range check the color number when making the backplot
[15:07:04] <CIA-12> 03jepler 07HEAD * 10axis/rs274/options.py: backplot colors for probing and toolchange; range check the color number when making the backplot
[15:07:04] <CIA-12> 03jepler 07HEAD * 10axis/scripts/axis.py: backplot colors for probing and toolchange; range check the color number when making the backplot
[15:28:27] <jepler> Someone suggested that there should be a way to use the mouse wheel as a jog wheel. Does anyone have a suggestion about how that should be activated?
[15:29:03] <jepler> e.g., middle (wheel) click on the axis button?
[15:34:39] <CIA-12> 03jepler 07HEAD * 10axis/scripts/axis.py: remove debugging message
[15:35:34] <CIA-12> 03jepler 07HEAD * 10axis/thirdparty/rb.tcl: remove debugging message
[15:49:53] <Lerneaen_Hydra> jepler: it would be good to be either able to control one axis or two simultaneusly
[15:50:14] <Lerneaen_Hydra> possibly a third with a scroll wheel?
[15:51:23] <Lerneaen_Hydra> as for how to enable/disable the different modes of jogwheel-ness, I don't have an immediate idea as to what would be good
[15:51:45] <jepler> I don't own any mice with more than one wheel
[15:52:22] <Lerneaen_Hydra> It would seem though, that it may be beneficial to be able to remap axises without restarting emc & changing a hal config, so a button of some sort would be a starting point
[15:52:23] <jepler> oh, you mean a mode where moving the mouse moves two axes, and moving the wheel moves a third? that's not quite what I had in mind
[15:52:29] <Lerneaen_Hydra> yes
[15:52:35] <Lerneaen_Hydra> or rather
[15:53:10] <Lerneaen_Hydra> you can configure it like that, or you can configure it so that the X axis on the mouse moves Z on the machine, and Y and scroll don't do anything
[15:53:26] <jepler> do you think that the mouse provides fine enough control to be useful?
[15:53:36] <Lerneaen_Hydra> personally I would think so
[15:54:34] <Lerneaen_Hydra> I can easily move a cursor to a 5 pixel line that has nearly 1300 pixels quickly and effectivly
[15:55:04] <Lerneaen_Hydra> If one thinks of the machine as a screen with a pointer then I think the resolution would be enough
[15:56:11] <jepler> I have been surprised time after time by poor mouse control, even among programmers
[15:57:07] <Lerneaen_Hydra> hmm. How easy would it be to hack a rudimentary system together to test if it's plausible or not?
[15:57:11] <jepler> for example, I wrote some code that differentiated between a drag and an incomplete double click when the motion left a 5-pixel window around the first click location. Lots of people couldn't double-click anymore!
[15:57:47] <Lerneaen_Hydra> oh. that doesn't bode too well..
[15:59:31] <jepler> like I said, I was surprised
[15:59:41] <jepler> now I just assume no user has good motor control
[16:00:46] <cradek> even someone with good motor control can be clumsy sometimes
[16:01:08] <jepler> that's true, and the price to pay for an accidental mouse motion would be pretty high if it's moving a CNC machine
[16:01:37] <Lerneaen_Hydra> that would be the same if they are clumsy with a dedicated jogwheel
[16:01:49] <jepler> with a dedicated jogwheel, the detents help
[16:02:04] <jepler> (same would be true of a mouse wheel, especially since they have a small number of detents per revolution)
[16:02:13] <Lerneaen_Hydra> granted, it's easier for some other person to see: ooh! a computer! (moves mouse) BLAM!
[16:02:30] <Lerneaen_Hydra> jogwheels usually have detents?
[16:02:45] <Lerneaen_Hydra> The only one I've seen didn't...
[16:02:49] <jepler> the ones I've seen have
[16:02:58] <jepler> (which is a grand total of .. 3?)
[16:09:27] <Lerneaen_Hydra> so... we can discuss whether it should work or should not work, but since no one seems to know definetly whether it will or not, is someone going to hack a rudimentary test system together?
[16:44:08] <Mess> hello all
[16:45:09] <ValarQ> hello you
[16:50:08] <Mess> how are things these days... has alpha been around??
[16:58:21] <ValarQ> hmm
[18:09:33] <jepler> Lerneaen_Hydra: http://emergent.unpy.net/files/sandbox/mousejog http://emergent.unpy.net/files/sandbox/mousejog.hal
[18:09:59] <Lerneaen_Hydra> jepler: sorry, I can't test it today, I'll test it tomorrow though
[18:10:03] <jepler> no hurry
[18:10:33] <Lerneaen_Hydra> have you tested yet?
[18:11:16] <Lerneaen_Hydra> jepler: How is the mouse grabbed? via a button in axis?
[18:13:24] <cradek> there's a window and you click in it to cause the grab
[18:15:39] <Lerneaen_Hydra> oh, ok
[18:30:35] <Mess> Hi all .. again
[20:05:47] <jepler> Lerneaen_Hydra: I don't have any real hardware, all I did was observe that it worked on the simulator
[20:06:01] <Lerneaen_Hydra> jepler: oh, ok
[20:06:26] <Lerneaen_Hydra> jepler: so is loading that hal file enough?
[20:06:57] <jepler> you need a newish AXIS for the 'hal' module, and put the 'mousejog' program in bin/
[20:07:03] <jepler> and then you load the hal file
[20:07:36] <Lerneaen_Hydra> oh, there were two files
[20:07:50] <Lerneaen_Hydra> emc2/bin?
[20:08:12] <jepler> yes
[20:08:15] <CIA-12> 03jepler 07HEAD * 10axis/scripts/axis.py: entering teleop mode seems to break jogwheels, so don't automatically do it for KINEMATICS_IDENTITY
[20:39:21] <Lerneaen_Hydra> g'night all
[20:39:27] <jepler> see you Lerneaen_Hydra
[20:55:08] <Jymmm> http://code-browser.sourceforge.net/features.html
[21:00:34] <Jymmm> It's a unique concept, but I like it.
[21:13:31] <fenn> wow it's almost as good as "kate"
[21:13:38] <fenn> :D
[21:15:11] <Mess> back later we're out for dinner on PEI again.. my wife and daughter are with me
[21:26:02] <Jymmm> Mess Got Lobster?
[21:27:03] <Jymmm> and it can be run from a floppy too (m4)
[21:27:07] <Jymmm> (M$)
[22:13:58] <Jymmm> fenn oh man, even editing the UI is live changes
[22:17:53] <skunkworks> join #emc-devel
[22:18:00] <skunkworks> oops
[22:23:36] <anonimasu> 3
[22:24:49] <Jymmm> anonimasu how was holiday?
[22:24:57] <anonimasu> work and more work :)
[22:25:07] <Jymmm> thought you were gone for 2wks?
[22:25:25] <Jymmm> on holiday that is
[22:30:38] <anonimasu> oh that's a long time ago
[22:30:55] <Jymmm> last week is a long time ago?
[22:31:01] <anonimasu> oh I didnt have vacation
[22:31:03] <anonimasu> just a day
[22:31:28] <Jymmm> so you lied to us saying you'll be back in two weeks huh?!
[22:40:14] <anonimasu> that was about ½ a year ago :9
[22:40:39] <anonimasu> I think
[22:40:51] <anonimasu> while my gf still talked to me
[22:49:25] <Jymmm> lol
[22:49:39] <anonimasu> :)
[23:27:11] <danex> jmkasunich, just read your e-mail in the list about the ballscrew comp
[23:27:29] <jmkasunich> hi
[23:27:39] <jmkasunich> any thoughts?
[23:29:01] <danex> On Mazak Machines it was called pitch error comp, from what I can remember it was set up by first dividing the ball screw into x number of sections
[23:29:34] <danex> then giving a value of plus or minus to each section
[23:29:55] <jmkasunich> do you remember roughly how many sections?
[23:29:59] <Mess> hi all
[23:30:06] <jmkasunich> hi Mess
[23:30:37] <danex> The number of sections varied by the condition of the ball screw
[23:30:52] <jmkasunich> a few, or a few hundred?
[23:31:02] <Mess> variable ball screw mapping??
[23:31:14] <danex> A few hundred
[23:31:36] <jmkasunich> Mess: yes
[23:32:29] <Mess> the higher the resolution (more sections).. the better you can make a BAD machine look good
[23:32:37] <danex> One other Item, a parameter could be set so that the axis would only position from one direction, this for the backlash error correction
[23:32:37] <jmkasunich> yeah
[23:32:39] <anonimasu> ´hehe
[23:33:09] <danex> I made several bad machines look good this way
[23:33:10] <jmkasunich> only usefull for some types of work...
[23:33:22] <jmkasunich> hard to mill pockets from only one side
[23:33:40] <danex> True oh How True
[23:33:50] <Mess> that would be unidirectional position and only good for jig boring type operations
[23:35:22] <jmkasunich> is that actually a machine thing? or do you just write the part program to always approach from the same side?
[23:35:26] <anonimasu> night
[23:35:33] <jmkasunich> goodnight
[23:35:55] <danex> It was a machine parameter
[23:36:07] <Mess> some machines will do it... add hockey sticks and j-hooks...
[23:36:09] <danex> goodnight anonimasu
[23:36:20] <jmkasunich> interesting
[23:36:42] <jmkasunich> IMO, I'd rather do that with a postprocessor - g-code is perfectly capable of describing the neccessary motion
[23:38:01] <danex> That is the best way, Mazak tried to control all because of the conversational control
[23:38:38] <Mess> im running mazak's this week.. full iso code,,
[23:38:53] <Mess> on PEI again
[23:40:18] <danex> which models?
[23:41:20] <Mess> they have 17 mazaks...im on a 1 M vertical and a 500mm horizontal-4 axis
[23:41:39] <Mess> 2 very nice integrex
[23:42:03] <Mess> numerous QT's
[23:42:44] <danex> jmkasunich, I will look at the pitcherror comp in a couple of machines and let you know how many sections and ballscrew travel
[23:43:17] <danex> mess, are the integrex the new style head ?
[23:43:40] <jmkasunich> danex: thanks
[23:45:58] <danex> The older style (pre Fusion Control) were much more rigid
[23:58:17] <danex> jmkasunich, On my current project, 2 Motenc- lite cards , I have all six axis encoders reading, still have a few problems with the m-codes
[23:58:37] <jmkasunich> six axis? what kind of a machine is it?
[23:59:06] <danex> Beveling Metal
[23:59:18] <jmkasunich> like weld prep?
[23:59:28] <danex> Yes
[23:59:41] <jmkasunich> what M-codes are you trying to use?
[23:59:51] <danex> Two Heads connected to one drive axis