#emc | Logs for 2009-01-31

[00:17:41] <JymmmEMC> Skullworks: verify the disk reading issue and pwr mgnt
[00:18:40] <Skullworks> hmm - I don't have a disk
[00:18:58] <JymmmEMC> doesn't matter I believe
[00:19:36] <JymmmEMC> I don't recall the details, but way too many reads
[00:21:26] <Skullworks> I'm running the entire system on a class 6 16GB SDHC card.
[00:24:56] <Skullworks> If anything I need to hack up my own "widescreen" mod for TkEMC.
[00:27:45] <The_Ball> Does anybody know if using HOME_USE_INDEX works when using a mesa 5i23 card? Or if it works at all?
[00:28:52] <Skullworks> laptops and desktop LCD's are trending towards the widescreen formats.
[00:34:24] <alex_joni> good night all
[00:34:43] <alex_joni> The_Ball: a 5i23 can home to index
[00:35:03] <Skullworks> night Alex, thanks for the help
[00:35:09] <alex_joni> but I don't think the standard config (is there one for 5i23?) uses the ini entry HOME_USE_INDEX
[00:35:09] <The_Ball> thanks, must be my wiring
[00:35:22] <The_Ball> I put it in the file
[00:35:23] <alex_joni> The_Ball: ask seb when he's around
[00:35:41] <alex_joni> The_Ball: you need to have the proper stuff in the hal file
[00:35:49] <alex_joni> just adding HOME_USE_INDEX to the ini won't help
[00:35:54] <The_Ball> ah, they need to match up?
[00:36:05] <The_Ball> thanks, I'll dig through the docs
[01:29:37] <Skullworks> ok I guess it only works for non-AXIS based GUI.
[01:40:01] <toastyde1th> toastyde1th is now known as toastydeath
[02:52:28] <topls64> I'm having trouble finding info on using microstepping. Can anyone show me the way to some info?
[02:54:19] <topls64> specifically, problem is each advance of .10" adds .005"
[02:55:29] <eric_unterhausen> .1 gives .105?
[02:56:12] <topls64> 200 step motor, 1000x microstep, 5x in config. yes, .1=.105. secessive moves, no direction change
[02:56:37] <topls64> backlash a different issue for a later day
[02:57:54] <jmkasunich> you are saying that when you command a move of 0.100", you get 0.105" (as measured by a dial indicator or something)?
[02:58:33] <topls64> yes.
[02:58:57] <jmkasunich> when you command a move of 1.000", do you get 1.005"? or 1.050"? or something else?
[02:59:14] <jmkasunich> if it is a scale error, it will be proportional, you'll always get 5% more
[02:59:34] <topls64> I'll have to fire up the machine. I only tried moves of 0.100
[02:59:50] <jmkasunich> fire it up
[03:00:35] <jmkasunich> it is very very difficult to troubleshoot a machine without the machine
[03:01:39] <topls64> ha. true. I remember you. We talked once about using absolute encoders to control a paper tube labeling machine. The project is on hold prolly a few years.
[03:01:46] <topls64> booting...
[03:05:53] <cradek> 1000x microstep?
[03:06:13] <jmkasunich> I think he means 5x microstep, 1000 microsteps/rev
[03:06:41] <cradek> oh, ok
[03:07:02] <topls64> yes. drivers are set to 1000
[03:10:57] <topls64> can I jog the 1.0" in .10 steps, or shall I look up the gcode to move the full inch?
[03:11:15] <jmkasunich> do it in one mive
[03:11:17] <jmkasunich> move
[03:11:27] <topls64> ok
[03:11:37] <jmkasunich> assuming you are at X=0, the mdi command is: G0X1
[03:11:54] <jmkasunich> to eliminate backlash, try this sequence:
[03:12:03] <jmkasunich> G0X-0.1 (get below zero)
[03:12:10] <jmkasunich> G0X0 (move to zero from the left)
[03:12:19] <jmkasunich> G0X1 (move to 1 from the left)
[03:12:30] <jmkasunich> the last move is the one to measure
[03:12:44] <jmkasunich> you can change the X1 to X0.1, or X2, or X10, or whatever
[03:15:50] <topls64> ok. zeroed out. here we go.
[03:19:17] <topls64> my dial is 1" overall. I might have to do a little tweaking here. can I do a 0.900 move?
[03:19:22] <jmkasunich> sure
[03:19:37] <jmkasunich> the idea is just to see if the error is proportional to distance traveled
[03:23:25] <topls64> crap dial indicator shows ~1.00125 on 1" move. So, a quarter thou added per turn?
[03:23:51] <jmkasunich> what does it show for a 0.1" move
[03:24:11] <jmkasunich> I know you said 0.105 before - do it again now, same setup and technique as you used for the 1.000" move
[03:25:44] <jmkasunich> if the dial is set up with zero at X0, and to has 1" of travel, you can do this in MDI:
[03:25:50] <jmkasunich> G0X-0.1 (get clear)
[03:26:10] <jmkasunich> F2 (2 ipm travel, so we don't whack the indicator or shake the setup)
[03:26:24] <jmkasunich> G1X0 (move to zero, verify that indicator reads zero)
[03:26:42] <jmkasunich> G1X0.1 (should move to 0.100, write down reading
[03:26:55] <jmkasunich> G1X0.2 (move to 0.200, write down reading)
[03:27:00] <jmkasunich> G1X0.3
[03:27:05] <jmkasunich> G1X0.4
[03:27:07] <jmkasunich> ....
[03:27:11] <jmkasunich> G1X1
[03:27:27] <jmkasunich> then for grins, go back and do it in a single move:
[03:27:30] <jmkasunich> G0X-0.1
[03:27:38] <jmkasunich> G1X0 (check zero reading)
[03:27:43] <jmkasunich> G1X1 (check 1" reading)
[03:34:54] <topls64> overagers .1=.0006 .2=.0006 .3=.0011 .4=.0008 .5=.0011 .6=.0010 .7=.0020 .8=.0015 .9=.0020 1.0=.0014. Repeatable. Via mdi and manual control
[03:36:23] <topls64> thax for the gcode lesson btw
[03:36:39] <jmkasunich> you're welcome
[03:37:00] <jmkasunich> it looks to me like your machine is configured correctly
[03:37:10] <jmkasunich> you are probably just seeing screw error
[03:37:33] <jmkasunich> what kind of machine is this? ballscrews, acme screws, high-precision or hobbyist?
[03:37:48] <topls64> x3 mill with cncfusion kit
[03:37:58] <topls64> basic kit, not delux
[03:38:06] <jmkasunich> dunno cncfusion, is that a ballscrew kit?
[03:38:13] <topls64> yes
[03:38:14] <jmkasunich> rolled or ground?
[03:39:04] <jmkasunich> preloaded or not?
[03:39:14] <topls64> never been asked or looked into.
[03:39:18] <topls64> not preloaded
[03:39:35] <jmkasunich> you might want to repeat the test going the other way, start at 1.0, or 1.020 if the indicator has enough travel
[03:39:51] <jmkasunich> then G1X1.0, G1X0.9, G1X0.8, etc
[03:39:56] <jmkasunich> without touching the indicator
[03:40:02] <jmkasunich> that will give you an idea of the amount of backlash
[03:40:30] <jmkasunich> new ground screws would cost several hundred dollars
[03:40:37] <cradek> my chinese dial indicator can't be trusted to closer than about .002
[03:40:41] <jmkasunich> new rolled screws maybe 100, or a little less
[03:41:10] <jmkasunich> the readings you are getting are consistent to about a thou, which is pretty good
[03:41:11] <cradek> so don't conclude that something is wrong if that's all you have measured with
[03:41:33] <jmkasunich> ground screws can probably do better, but you need to measure very carefully for that accuracy
[03:42:17] <jmkasunich> if you really want to check an axis to sub-thousandths accuracy, you need to use a DTI and gage blocks
[03:42:45] <jmkasunich> zero DTI against spindle, move table a little more than an inch, insert 1" gage block, move table back to 1.000 and read DTI, etc
[03:42:50] <topls64> I have to see what dial we have at work. I have done some backlash comp on it in the past (not this config). pdf for screws says tols are like .0001-.0002, so...I can see where it may be coming from.
[03:43:25] <topls64> I think I'm trying to get zerobl from a non-zerobl kit :)
[03:46:16] <cradek> an awful lot of stuff has to be perfect to get that last .001
[03:46:41] <topls64> I'm slowly learning that.
[03:46:52] <topls64> And loving every minutes of it
[03:47:36] <topls64> Now to figure out why my z axis likes to "settle in" to the last .100 of travel
[03:56:01] <topls64> thx again for your help
[03:57:51] <jmkasunich> you're welcome
[04:10:44] <topls64> Well, looks like I'll be reloading a ballnut tomorrow
[04:11:05] <topls64> so, why doesn't the estop work in stepconfig?
[04:14:19] <topls64> lesson learned ;)
[05:39:12] <Skullworks> topls64 - the basic cnc Confusion kit doesn't really address thrust backlash very well.
[05:40:26] <Skullworks> just uses the OEM torrington type bearings and simple shaft collars.
[05:41:15] <Skullworks> alot of people traded up to the delux kit for those reasons
[05:47:20] <KimK_> KimK_ is now known as KimK
[10:23:11] <micges> good morning
[10:51:35] <_mr_boo> is this typically the best $500 cnc toys can do? http://img411.imageshack.us/img411/6161/dscn0333qr0.jpg
[10:53:06] <_mr_boo> _mr_boo is now known as mr_boo
[12:25:42] <motioncontrol> good morning.i have the m5i20 card and use the hm2_pci driver.My spindle is analog +/-10 volt and not use the digital signal rev o fw direction, because the drivers is +/-10volt and the spinfle rotation correct with m3 and m4 command.the problem is in the g33.1 z-1 k1 command.when the spindle reverse the z axis go in following error.because?
[12:30:17] <archivist> translate "the problem is in the g33.1 z-1 k1 command.when the spindle reverse the z axis go in following error.because?" to a better question
[12:47:47] <alex_joni> motioncontrol: I suspect your spindle changes direction too fast
[12:47:53] <alex_joni> and the z axis can't keep up
[12:48:07] <alex_joni> what is you z axis max_vel/max_accel ?
[13:19:13] <BigJohnT> morning alex_joni
[13:19:22] <alex_joni> hi BigJohnT
[13:35:50] <motioncontrol> after i prove to change the acceleration value of drives spindle for ceck.the acceleration value now is 0 very fast.thanks at all.
[13:59:14] <Skullworks> Alex - I was wrong about the SIM package working in 8.10 - Something breaks AXIS, all other GUI's do work however, - and I edited the wiki page to note that.
[14:00:42] <Skullworks> I didn't catch it since I was using TkEMC for my tests, but later went thru and tried the others,
[14:36:07] <cradek> strange - today every search result on google says "This site may harm your computer."
[14:36:33] <archivist> Ive not seen that
[14:37:34] <cradek> http://www.google.com/support/websearch/bin/answer.py?answer=45449&topic=360&hl=en&sa=X&oi=malwarewarninglink&resnum=1&ct=help
[14:40:25] <cradek> http://timeguy.com/cradek-files/emc/harmful.png
[14:40:35] <cradek> bizarre
[14:40:37] <archivist> Im seeing it now
[14:40:53] <cradek> oh ok, it's not something strange about me then
[14:40:54] <archivist> and reports are popping up elsewhere
[14:41:28] <archivist> ##php is just starting to talk about it
[14:42:46] <archivist> hehe ite really bad for google.. Im getting a download for answer,py
[14:43:45] <archivist> its a google f up
[14:44:27] <alex_joni> so basicly they disallowed any links
[14:44:39] <alex_joni> now that's a screwup
[14:46:04] <alex_joni> ROFLMAO: http://www.google.com/interstitial?url=http://www.google.com/
[14:49:16] <archivist> the load on the servers will shoot up and make it worse as every one tells their friends
[15:09:48] <DanielFalck> wtf- google....
[15:10:08] <alex_joni> DanielFalck: indeed wtf
[15:10:09] <DanielFalck> anything with linuxcnc in search
[15:10:24] <alex_joni> anything in search
[15:10:39] <alex_joni> even searching for "google" gets the same result
[15:10:51] <archivist> just borked
[15:10:54] <DanielFalck> yes
[15:11:04] <KimK> Hi all. In EMC2 User Manual (v2.2), Chapter 11.6 "G10 Set Coordinate System", in the example given (setting offsets for coordinate system G54 aka No. 1) "Example: G10 L2 P1 x 3.5 y 17.2", what does the L2 do, and what are my other "L" choices there? (P.S. Minor annoyance: Is there a reason why the example is in mixed upper and lower case?)
[15:11:20] <DanielFalck> at least mach3 gets the same results :)
[15:11:44] <alex_joni> well.. in that case google's even right
[15:11:54] <DanielFalck> as for Microsoft....
[15:11:58] <alex_joni> that site surely _may_ harm your computer
[15:14:38] <micges> KimK: L2 means sub function of G10
[15:15:06] <DanielFalck> slashdot has an entry about the google problem
[15:15:27] <micges> KimK: another subfunction will be G10 L1 for setting tool table entry in version EMC 2.3
[15:15:47] <alex_joni> in emc2.2 there is no Lx besides L2 iirc
[15:15:52] <alex_joni> DanielFalck: link?
[15:16:02] <DanielFalck> http://tech.slashdot.org/article.pl?sid=09/01/31/1457221
[15:16:21] <alex_joni> (it's not indexed on google yet.. :P)
[15:18:03] <DanielFalck> https://addons.mozilla.org/en-US/firefox/addon/537
[15:18:22] <KimK> So these G10 sub-functions are incompletely explained in v2.2 because they are being phased in as part of the v2.3 upgrade? OK, thanks very much.
[15:18:25] <DanielFalck> redirect remover
[15:19:34] <micges> KimK: G10 L2 means set offsets
[15:19:47] <alex_joni> KimK: they are incompletely explained because there is only one in 2.2.x
[15:19:58] <alex_joni> so you always have to use L2, not much to explain about that
[15:20:21] <alex_joni> fenn: did you break google?
[15:20:47] <fenn> uh oh..
[15:21:11] <archivist> google is fixed
[15:21:25] <alex_joni> archivist: no it's not
[15:21:36] <alex_joni> try logging in to webmaster tools, or some other google stuff
[15:21:38] <archivist> werked for me
[15:21:43] <DanielFalck> thanks for fixin' it archivist :)
[15:21:46] <alex_joni> "We're sorry...
[15:21:47] <alex_joni> ... but your query looks similar to automated requests from a computer virus or spyware application. To protect our users, we can't process your request right now."
[15:21:55] <fenn> i went downstairs and jiggled some fuses, did that help?
[15:22:17] <archivist> probably in the middle of reverting the cockup
[15:22:35] <alex_joni> right.. search is ok now
[15:22:39] <DanielFalck> I love this one: "Has anyone considered the possibility that as of this morning 95% of the sites on the Internet are infected with malware?"
[15:23:00] <fenn> 95% of the computers accessing the internet, probably not sites though
[15:23:39] <archivist> 95 % of the IIS servers may be
[15:27:45] <fenn> so.. _after_ I turned off the space heater, the circuit it was on stopped working
[15:28:13] <fenn> and now i know that's the same circuit my server was on
[15:28:53] <fenn> doesn't really inspire confidence
[15:28:58] <archivist> switch off arc dropped the earth leakage breaker ?
[15:31:09] <fenn> nah, some thermal expansion thing.. -4F last night and I was using the heater a long time
[15:31:23] <archivist> could be a failing input filter on that line causing imbalance
[15:31:53] <fenn> er.. no
[15:50:15] <mOizt> http://www.youtube.com/watch?v=y9fZpqye8Pc
[16:43:10] <BigJohnT> hmmm I can see my windows machine from each ubuntu machine on the network File Browser but I can't see either Ubuntu machine from the other ???
[16:45:01] <Skullworks> consistently inconsistent
[16:45:22] <BigJohnT> that's what I thought
[16:45:42] <archivist> define see
[16:45:50] <eric_unterhausen> configuration probably
[16:46:32] <BigJohnT> when I open the network browser I can see the windows machine and open files on it
[16:47:14] <BigJohnT> even from my windows machine I can't see the linux machines in network
[16:47:27] <BigJohnT> workgroup maybe?
[16:48:03] <eric_unterhausen> I think so
[16:48:30] <eric_unterhausen> the default workgroup for linux is a little strange
[16:49:05] <eric_unterhausen> but windows computers are constantly advertising their existence
[16:49:33] <BigJohnT> I see host name and domain name and domain name is blank
[17:09:26] <alex_joni> BigJohnT: you need samba running on the linux PC for win pc's to see it
[17:09:39] <BigJohnT> ok thanks
[17:09:41] <alex_joni> the easiest way to do that in Ubuntu, is rightclick on a folder
[17:09:54] <alex_joni> then select share (win network) and it will install/configure samba for you
[17:10:11] <BigJohnT> ok thanks
[17:10:15] <alex_joni> samba has only 30-60 times more config entries that emc2.. so you know what to expect :D
[17:10:54] <BigJohnT> :)
[17:11:21] <alex_joni> BigJohnT: prod me if it doesn't work
[17:12:03] <BigJohnT> ok will do
[17:44:19] <BigJohnT> eric_unterhausen: it seems like all you have to do is share a directory in Ubuntu and then follow the bouncing ball
[17:44:43] <eric_unterhausen> that makes sense, windows usually works the same way, no shared, no see
[17:45:36] <BigJohnT> it was a bit easier on the 8.04 machine than the 6.06 one
[17:47:03] <BigJohnT> now to figure out why alt print screen won't work on my 8.04 machine :/
[17:47:42] <SWPadnos> your alt key is bad?
[17:48:35] <BigJohnT> could be it is a crappy keyboard cause the "heh" key is broken too
[17:48:50] <SWPadnos> does the "bummer" key work?
[17:49:36] <BigJohnT> that's F16 right?
[17:49:50] <SWPadnos> no, that's a fighter jet
[17:52:37] <BigJohnT> how can I test the alt key?
[17:53:54] <SWPadnos> open a program that has a menu (like gedit), and alt should go to the menu bar (but not open any menus)
[17:53:56] <jepler> BigJohnT: use alt-tab to cycle among windows?
[17:54:01] <SWPadnos> or you can hit alt-F for example
[17:54:06] <SWPadnos> oh, that's even easier :)
[17:54:09] <alex_joni> ctrl-alt-delete?
[17:54:10] <BigJohnT> ok
[17:54:13] <alex_joni> that's more obvious
[17:54:17] <BigJohnT> I like that one
[17:54:40] <BigJohnT> it works with ctrl-alt-delete
[17:54:43] <SWPadnos> does the print-screen key work?
[17:54:51] <BigJohnT> yes
[17:55:07] <SWPadnos> ok
[17:55:29] <SWPadnos> I wonder if there's a setting in gscreencaptureorwheteveritscalled to enable/disable the various types of capture
[17:55:30] <jepler> System > Preferences > Keyboard
[17:55:45] <jepler> Shortcuts
[17:55:49] <jepler> System > Preferences > Keyboard Shortcuts
[17:56:00] <alex_joni> isn't alt-PrintScreen also alt-sysrq?
[17:56:31] <SWPadnos> heh, looks like it, doesn't it?
[17:56:39] <jepler> the screenshot actions are under "desktop"; make sure the key you desire is assigned there
[17:56:51] <BigJohnT> it shows take a screen shot of a window as alt print
[17:57:11] <jepler> you could try changing it to something else such as shift-print or control-print
[17:57:22] <jepler> "print" is the name it uses for "printscreen"
[17:57:44] <BigJohnT> ctrl print worked
[17:58:05] <BigJohnT> things that make you go "hmmmm"
[17:58:33] <BigJohnT> it must be the crappy keyboard
[17:58:44] <SWPadnos> probably not
[17:59:19] <SWPadnos> have you tried capturing from different applications?
[18:00:06] <BigJohnT> yes I did and it didn't do anything when I tried to use alt print in keyboard shortcuts
[18:00:23] <mOizt> did You see this? http://www.youtube.com/watch?v=y9fZpqye8Pc
[18:00:41] <mOizt> nice one.. tried it in the wife this morning.. with no apparent sucess
[18:01:51] <mOizt> turn the stereo way up btw ;)
[18:34:23] <JymmmEMC> SWPadnos: SWPLinux http://newspaperads.mercurynews.com/ROP/ads.aspx?advid=32664&adid=7488017&subid=25054042&type=
[18:34:37] <JymmmEMC> WTH can use 8GB ram?!
[18:34:51] <SWPadnos> video processing
[18:34:52] <JymmmEMC> Not that I don't want 8gb ram, but what can use it?
[18:35:31] <eric_unterhausen> I have some matlab scripts that will run you right out of ram
[18:35:44] <SWPadnos> neural nets also
[18:36:00] <JymmmEMC> But many OS'es cna only see/use up to 3.2GB ram, even if you have 4GB
[18:36:05] <SWPadnos> training may take multi-hundred-GB datasets
[18:36:11] <SWPadnos> not 64-bit ones
[18:36:36] <eric_unterhausen> right, you have to go 64bit
[18:37:57] <JymmmEMC> OSX 10.5 is 64bit, but still can't use more than 3.2GB
[18:38:09] <SWPadnos> that's dumb
[18:38:13] <eric_unterhausen> apple, 'nuff said
[18:38:41] <JymmmEMC> Ok, so you guys have 4GB and it see/uses all 4GB?
[18:38:52] <eric_unterhausen> I have 8
[18:39:00] <JymmmEMC> running?
[18:39:07] <eric_unterhausen> we also have 2 comps at work with 128gb
[18:39:12] <eric_unterhausen> yes, current machine
[18:39:13] <alex_joni> it's fixed to the mobo
[18:39:18] <alex_joni> not running
[18:39:28] <alex_joni> eric_unterhausen: I'm sure he meant to ask what OS..
[18:39:29] <SWPadnos> I have a machine with 8GB that's all recognized by Ubuntu 8.10
[18:39:31] <JymmmEMC> running what distro?
[18:39:37] <JymmmEMC> k
[18:39:37] <eric_unterhausen> ubuntu
[18:39:45] <eric_unterhausen> at work, fedora
[18:39:52] <alex_joni> eric_unterhausen: emc2 on it?
[18:40:03] <JymmmEMC> well, big diff between desktop and server hw
[18:40:05] <eric_unterhausen> I ran emc2 on this machine
[18:40:15] <alex_joni> how much did it see?
[18:40:18] <eric_unterhausen> but of course it didn't recognize the 8gig
[18:40:24] <alex_joni> iirc the RT kernel is limited to 4G
[18:41:30] <eric_unterhausen> we aren't going to catch up to the limits of a 64 bit machine for a couple of years yet
[18:41:51] <SWPadnos> they only have 48-bit (or so) physical addressing
[18:42:12] <eric_unterhausen> only
[18:42:36] <JymmmEMC> SWPadnos: Ok, besides it being an HP, what about the rest of the specs? I could care less about BR/HD
[18:42:42] <alex_joni> how much does PAE see on 32bit?
[18:43:01] <SWPadnos> other than Vista, the specs look good
[18:43:04] <eric_unterhausen> seems expensive
[18:43:33] <JymmmEMC> SWPadnos: Heh, well vista was a given. Given that it'll be blown away that is
[18:43:34] <SWPadnos> yeah, with 9800GS graphics and a 750GB hard drive, that's not a great deal
[18:43:46] <eric_unterhausen> alex_joni: what is pae?
[18:43:57] <SWPadnos> Physical Address Extension
[18:44:11] <SWPadnos> it's the way that 32-bit Intel CPUs can address more than 4G
[18:44:23] <SWPadnos> I think that goes up to 64GB, so maybe it's 4 bits
[18:44:28] <eric_unterhausen> are there drivers?
[18:44:28] <JymmmEMC> eric_unterhausen: Yeah, I thought so too, but 4 core + 8gb + mobo + case, etc. sometimes it's better to buy pre-build and just upgrade. 1st 4core I saw on today's ads
[18:44:30] <SWPadnos> (would make sense, with their paging schemes)
[18:45:02] <SWPadnos> JymmmEMC, build one out on NewEgg, it'll be a much better system at the same cost, or an equivalent system for a couple hundred $$ less
[18:45:23] <JymmmEMC> SWPadnos: Sorry, I don't do mail-order in that respect.
[18:45:37] <SWPadnos> ok. then go to Frys or something
[18:45:48] <alex_joni> http://en.wikipedia.org/wiki/Physical_Address_Extension
[18:45:48] <JymmmEMC> hey, there's an idea!
[18:46:09] <SWPadnos> you may not get anything better locally
[18:46:10] <JymmmEMC> SWPadnos: I have nfc on AMD is my biggest issue.
[18:46:16] <eric_unterhausen> screw fry's, they'll just try to sell him some overpriced junk HP or something
[18:46:47] <SWPadnos> the 9850 is a fast CPU, not as fast as the fastest core2 quads, but not exactly a dog
[18:47:08] <JymmmEMC> SWPadnos: I need something I can run a few VM's in.
[18:47:17] <SWPadnos> the core i7 is much faster, but only if you want to spend an extra $1000 on the CPU (itself, not including more expensive motherboards ...)
[18:47:21] <SWPadnos> that'll do it
[18:47:32] <SWPadnos> incidentally, 8GB is good for many VMs as well ;)
[18:47:37] <JymmmEMC> SWPadnos: I'm tossign everything over to VM's now. then just run thin on the lan
[18:47:57] <JymmmEMC> SWPadnos: Yeah, that I know. Will probably run CentOS on it
[18:49:36] <JymmmEMC> who the hell is biostar?! http://newspaperads.mercurynews.com/ROP/ads.aspx?advid=32664&adid=7492425&subid=25066121&type= Sounds like a HazMat cleanup company
[18:50:42] <SWPadnos> biostar is a good company, they've been around for decades
[18:50:49] <JymmmEMC> seriously?
[18:51:02] <SWPadnos> yes
[18:51:30] <JymmmEMC> I have never heard of them
[18:51:32] <SWPadnos> I have a Biostar motherboard in my main machine (the one I've been using since 2001)
[18:51:32] <JymmmEMC> ever
[18:51:56] <JymmmEMC> HT 2G ???
[18:52:07] <SWPadnos> oh wait, this is an aopen. but I've used biostar before :)
[18:52:17] <SWPadnos> what ?
[18:52:26] <JymmmEMC> In the link
[18:52:36] <JymmmEMC> Dual DDR2 800; HT 2G
[18:52:49] <SWPadnos> oh, that's slow
[18:52:56] <SWPadnos> hypertransport
[18:53:05] <SWPadnos> they're up to 3.2 or 3.6G I think
[18:53:17] <JymmmEMC> oh transfer speed?
[18:53:29] <SWPadnos> yes, gigaTransfers/sec
[18:53:36] <SWPadnos> (I think they're 8 or 16 bits wide)
[18:53:49] <SWPadnos> gotta run, bbl
[18:53:55] <JymmmEMC> ah
[18:55:17] <JymmmEMC> http://www.biostar.com.tw/app/en-us/mb/content.php?S_ID=387#
[18:55:47] <JymmmEMC> bastards, max support is 4GB
[18:59:51] <JymmmEMC> Man, this really bites. I need to setup two systems on the backend, one having multiple TV tuner cards in it, and another holding the main VM's + storage. If I could setup a custom LiveCD/USB/CF/SD for the media box then not as much of an issue.
[19:00:44] <JymmmEMC> I'm trying to reduce/consolidate hw, not have more of it! LOL
[19:02:09] <JymmmEMC> Hmmm, there's an idea.... Convert one system to VM, then use it to hold the tuner cards. I have one 64MB disk-on-a-chip I wonder if that would be big enough
[19:03:23] <JymmmEMC> Guess I could use USB if setup as LiveCD, then no writting issues.
[20:18:45] <BigJohnT_> BigJohnT_ is now known as BigJohnT
[20:20:42] <robin_sz> meep
[20:22:41] <skunkworks> robin_sz: !
[20:22:52] <skunkworks> what have you been up to?
[20:23:01] <robin_sz> oh ... running a business :)
[20:23:28] <robin_sz> keeps me kinda busy
[20:23:51] <skunkworks> I bet
[20:24:02] <robin_sz> and you?
[20:24:16] <skunkworks> work work work
[20:24:25] <skunkworks> keeps me out of trouble.
[20:24:25] <robin_sz> yeah, sucks
[20:24:34] <skunkworks> economy effecting you?
[20:24:50] <robin_sz> yeah, cant say it looks great going forward
[20:25:07] <skunkworks> yeck
[20:25:22] <robin_sz> we are bigger, faster and leaner than we used to be ...
[20:25:33] <robin_sz> but orders are thin
[20:25:55] <robin_sz> so, who knows
[20:26:50] <robin_sz> those poxy bankers have stuffed it good and proper
[20:26:58] <skunkworks> yes
[20:27:04] <skunkworks> and still are
[20:27:08] <robin_sz> oh yes
[20:28:10] <robin_sz> and of the billions given to banks, how much has trickled down to small firms?
[20:28:14] <robin_sz> zilch
[20:29:53] <robin_sz> anyway ... apart from all that ... i was actually playing with my EMC driven 2 axis testbed this weekend, so I thought I'd log in :)
[20:38:26] <skunkworks> cool
[20:38:40] <skunkworks> I am still playing with simple amps. http://imagebin.ca/img/703qoFX.jpg
[20:38:43] <skunkworks> getting there.
[20:39:18] <robin_sz> bloody hell, thats a big cap!
[20:42:54] <robin_sz> ive been playing a bit with a friend making switch mode PSU's for big audio amps
[20:43:27] <robin_sz> thats been fun
[20:44:11] <skunkworks> neat
[20:44:39] <robin_sz> bass so loud it distorts your vision :)
[20:45:28] <robin_sz> should I ask what that hug cap is for?
[20:47:02] <skunkworks> it is just across the dc input to the amp. it is what I have a ton of. :)
[20:54:22] <skunkworks> are you liking emc2?
[20:55:23] <skunkworks> robin_sz: steppers or servos?
[21:18:13] <robin_sz> servos
[21:18:45] <robin_sz> dc with hostmot
[21:19:17] <skunkworks> nice - how do you like the now hostmot driver?
[21:19:52] <skunkworks> new
[21:42:55] <Guest865> Hi WOW lots a peps here
[21:50:31] <KimK> Is math allowed in XYZ values? Can G1 X(1.5+0.5) be the same as G1 X2.0 ?
[21:51:05] <alex_joni> yes
[21:51:15] <alex_joni> but you need to use [] not ()
[21:51:30] <alex_joni> hi robin_sz
[21:51:38] <KimK> OK, thanks
[21:55:06] <Guest844> Hi I have a homebuilt router if multiple tools are required I have to stopp and touch next tool. no holders, any tips on how to handle this
[21:58:22] <BigJohnT_> BigJohnT_ is now known as BigJohnT
[21:59:11] <BigJohnT> hi Guest844 what version are you running?
[22:01:28] <Guest844> latest 2.2 I think it is
[22:01:38] <Roguish_> hey all. any one know of a reasonably priced USB to RS485 converter thing?
[22:02:29] <Roguish_> got a laptop w/o serial that i need to speak to an ACTech drive via 485.
[22:02:43] <Roguish_> 2 wire.
[22:04:44] <BigJohnT> Guest844: I think the simplest thing is to split up your g code one for each tool
[22:06:50] <Guest844> ok thanks for the help
[22:07:52] <BigJohnT> Guest844: that way you can use the touch off button for your axis when you locate it
[22:08:09] <Guest844> BTW EMC2 runs way faster than the mach3 I did use. you probaly knew this but I am impressed and want to give props to devemopers
[22:09:33] <BigJohnT> if your cutting lots of small lines and arcs be sure and use G64 Pxx
[22:10:38] <BigJohnT> http://www.linuxcnc.org/docview/html//gcode_main.html#sub:G61,-G61.1,-G64:
[22:10:51] <Guest844> What does that do? and what is the Pxx represent
[22:10:59] <Guest844> OK Ill check it out
[22:11:44] <BigJohnT> also look in the user manual in the important user concepts section for more info on G64
[22:13:40] <BigJohnT> Roguish_ I got mine from Allen Bradley and I think it was $200 or so :(
[22:14:56] <skunkworks> Guest844: really look at a probe switch for touchoff also. I use it for making ciruit boards and it works great. (I physically set the first tool - emc measures it to get length with a simple microswitch - then every tool I put in after that gets measured by the switch and adjusted acordingly
[22:16:06] <skunkworks> Guest844: there is an example in your gcode samples - also here http://cvs.linuxcnc.org/cgi-bin/cvsweb.cgi/emc2/nc_files/tool-length-probe.ngc
[22:16:21] <Guest844> skunkworks in this automated toch offs?
[22:20:55] <Guest158> what is the cause for a joint 0 follow error
[22:21:46] <jmkasunich> the motor (or step generator) isn't keeping up with the motion that EMC is requesting
[22:21:56] <jmkasunich> it is explained in the manual
[22:23:16] <eric_unterhausen> that was fast
[22:24:50] <Guest102> Oops I dropped off, regarding the follow error I searched the wiki and documentation and no results
[22:25:06] <eric_unterhausen> what kind of machine?
[22:25:15] <eric_unterhausen> stepper, servo, closed loop ?
[22:25:27] <Guest102> stepper
[22:26:24] <jmkasunich> google: linuxcnc following error
[22:26:35] <jmkasunich> first hit: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Following_Error
[22:27:00] <Guest102> thanks
[22:27:20] <eric_unterhausen> did you use stepconf?
[22:28:05] <Guest102> yes
[22:29:31] <Guest102> After reading the wiki, I know what the problem is, thanks.
[22:29:40] <eric_unterhausen> what is it?
[22:29:57] <Guest102> one suggestion, the error meassage is follow error, the wiki is titled following error
[22:30:11] <Guest102> probably why my wiki search didn't get a hit
[22:30:35] <jmkasunich> exactly where does this message appear?
[22:30:48] <jmkasunich> I've never heard of it being "follow error" before
[22:30:55] <Guest102> if you put in "follow error" in the wiki search there are 0 results
[22:31:07] <Guest102> joint 0 follow error is a pop up window
[22:31:27] <jmkasunich> are you using a translated version maybe?
[22:31:32] <Guest102> unless I forgot what the error message said 5 minutes ago
[22:31:55] <Guest102> I'll re check
[22:32:00] <eric_unterhausen> I could forget something like that in about 15 seconds
[22:32:23] <Guest102> problem is step gen can't keep up with the commanded position
[22:32:46] <Guest102> I'm exceeding my base period
[22:36:53] <Guest102> my mistake, sorry. The message is "following error" same as the wiki
[22:37:11] <jmkasunich> ok
[22:37:30] <jmkasunich> (if it really was "follow error", that would be a bug and I'd go find and fix it
[22:37:46] <jmkasunich> thanks for checking, so I didn't spend an hour trying to find it ;-)
[22:38:16] <Guest102> appreciate the help, later.......
[22:49:56] <lucasliam> Greets all - I dropped a note into the mailing list earlier on correcting gantry skew. My intent is to rebuild the offending unit and eliminate the skew - in the meantime does EMC have a method for handling this issue?
[22:52:36] <lucasliam> --checks his watch--- must be dinner time
[22:53:53] <jepler> lucasliam: fellow user micges has written a component to do this. I asked him to post it but don't know whether he has yet
[22:54:11] <jepler> lucasliam: ah, here it is: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?ContributedComponents#millkins
[22:55:41] <lucasliam> THanks Jeff - I thought I had seen a discussion on the EMClist last year - that must have been it - I will investigate that method.
[22:56:33] <jepler> lucasliam: after you compile & install millkins, you change 'loadrt trivkins' to 'loadrt millkins' in your hal file, and add a 'setp millkins.skew #.###' to set the amount of correction to apply
[22:56:53] <jepler> I haven't used it myself, and micges may be gone for the night (he's in europe) so I can't do more than make that educated guess about it
[22:57:35] <micges> I'm back
[22:57:36] <jepler> this assumes that your correction is the same at all points on the table .. if not, you'll have to do something more sophisticated
[22:58:01] <jepler> hi micges -- were your ears burning?
[22:58:19] <micges> lucasliam: millkins.c was tested and its working
[22:58:37] <lucasliam> Very noobish on linux so I will have to struggle with the compile part for a bit - regarding the skew - the gantry is pitched so it should be unitorm across the X travel
[22:58:50] <micges> heh I see many micges on talking :)
[22:59:46] <lucasliam> is the #.### a decimal, degree or radian measure?
[23:00:02] <micges> this is scew in mm per mm
[23:00:50] <lucasliam> I ste the machine up in imperial units - will that matter ---cringes---
[23:01:01] <lucasliam> I can certainly make the conversion needed
[23:01:22] <micges> so if you have skew 1mm for 1meter you will set it to 0.001
[23:01:55] <micges> can't convert it to imperial its too late here :)
[23:01:56] <jmkasunich> mm per mm or inch per inch, same thing
[23:02:26] <lucasliam> Straightfoward enough - and true enough on the conversion... units cancel )
[23:02:51] <jmkasunich> for a change ;-)
[23:03:23] <lucasliam> now where to find the noob linux channel for simple compiling....
[23:09:36] <alex_joni> lucasliam: install emc2-dev (sudo apt-get install emc2-dev)
[23:10:11] <micges> good night all
[23:10:14] <lucasliam> oooh.... helps to read the compile instructions just above the file....
[23:11:09] <alex_joni> yup.. sudo comp --install millkins.c should do the trick
[23:13:21] <jepler> don't forget build-essential