#emc | Logs for 2007-02-17

[00:11:37] <CIA-19> 03compile-farm 07BDI-4.51 ( * 10emc2head/: build FAILED ; see http://linuxcnc.org/compile_farm/emc2head_slot6_log.txt
[00:12:24] <skunkworks> what the heck was that? :P
[00:16:55] <jepler> I don't think jmkasunich_'s change is right
[00:28:47] <a-l-p-h-a> fenn, hahaha
[00:39:02] <robin_sz> sigh .. goindustry ... poxy auction company
[00:39:27] <robin_sz> in one auction we bought a banding/strapping machine
[00:40:00] <robin_sz> except the guy on the site decided not to let anyone take anyting and locked the gates
[00:40:21] <robin_sz> the next auction we bought a psace heater, that got stolen before we collected it
[00:40:51] <robin_sz> and this auction we bought a welder, that turned out to be something else when we went to pick it up
[00:41:47] <robin_sz> not great ...
[00:59:16] <a-l-p-h-a> robin_sz, uh. so not cool... did you have to pay for any of that shit then?
[01:03:03] <robin_sz> nope4
[01:03:18] <robin_sz> but when you send you driver 80 miles ....
[01:03:31] <robin_sz> it fscks you off a bit
[01:06:07] <a-l-p-h-a> robin_sz, that is just retarded.
[01:06:20] <a-l-p-h-a> robin_sz, they they mention anything about picking up not available that same day???
[01:07:31] <robin_sz> nah
[01:07:41] <robin_sz> you tunr up and say "where is my ..."
[01:07:47] <robin_sz> and they go "oh dear .."
[01:08:03] <robin_sz> you have to pay witin 48hrs of winning the auction ..
[01:08:06] <a-l-p-h-a> robin_sz, so what are you going to do? Just say "fuck off" to them for everything?
[01:08:14] <robin_sz> refunds are very prompt too
[01:08:25] <robin_sz> well, a bit over 48hrs
[01:08:27] <robin_sz> well ..
[01:08:31] <robin_sz> quite a bit
[01:08:44] <robin_sz> infact 4 weeks later we are still waiting ...
[01:09:14] <a-l-p-h-a> for a REFUND?
[01:09:22] <robin_sz> yeah
[01:09:23] <robin_sz> sucks
[01:09:35] <a-l-p-h-a> call your better business beareu (sp)
[01:09:58] <a-l-p-h-a> was this deposit, on a credit card? Do a charge back.
[01:15:42] <robin_sz> nah, theyu dont take cards ...
[01:15:49] <robin_sz> for that very reason
[01:16:30] <a-l-p-h-a> sue them, for the deposit, as well as court costs. :)
[01:16:41] <robin_sz> remember, sooon as the item is sold, the auction comapany has to pay the seller
[01:16:49] <robin_sz> so they like to get the money from the buyer asap
[01:17:20] <robin_sz> and they wpmnt risk a chargeback, so no cards
[01:17:52] <robin_sz> its not a lot of money, less than 1k total, so nothing bad, but rather annoying
[01:20:26] <CIA-19> 03jmkasunich 07TRUNK * 10emc2/src/hal/drivers/hal_m5i20.c: value is a pin, but invert is not... must remember not to do driver work on a sim box
[01:26:29] <Jymmmmmm> think you could put emc2 + axis on an 8" screen?
[01:26:48] <jmkasunich_> if the pixels are small enough, sure
[01:26:50] <jmkasunich_> ;-)
[01:27:22] <Jymmmmmm> http://search.ebay.com/190080166587
[01:27:38] <Jymmmmmm> 12" maybe?
[01:28:13] <jmkasunich_> 640x480 is pretty low
[01:28:31] <jmkasunich_> 800x600 would probably work, but I seriously doubt 640x480 would fit
[01:29:08] <Jymmmmmm> 1024x768 http://cgi.ebay.com/10-4-TFT-LCD-Touch-Screen-Monitor-for-Car-PC-POS-A5_W0QQitemZ280080943698QQihZ018QQcategoryZ29502QQrdZ1QQcmdZViewItem
[01:29:39] <Jymmmmmm> I'm assuming most touch screens just emulate mouse
[01:29:48] <jmkasunich_> native resolution 800x600
[01:29:49] <SWPadnos> http://www.engadget.com/2006/10/17/sanyo-epson-announce-7-1-inch-1080p-lcd-by-far-the-worlds-smal/
[01:29:57] <jmkasunich_> I would not attempt to run 1024x768 on it
[01:30:55] <Jymmmmmm> here we go... http://cgi.ebay.com/NEW-18-ELO-ENTUITIVE-TOUCH-SCREEN-LCD-MONITOR-BLACK_W0QQitemZ190083114920QQihZ009QQcategoryZ29503QQrdZ1QQcmdZViewItem
[01:31:49] <Jymmmmmm> sucks for brightness
[01:32:03] <CIA-19> 03jmkasunich 07v2_1_branch * 10emc2/src/hal/drivers/hal_m5i20.c: value is a pin, but invert is not.. oops
[01:32:20] <SWPadnos> I have the 17" version of that monitor, and it looks just fine
[01:32:22] <Jymmmmmm> so, would something like this work for emc2 + axis?
[01:32:35] <jmkasunich_> you mean the touch part?
[01:32:35] <Jymmmmmm> SWPadnos does your emulate a mouse?
[01:32:42] <Jymmmmmm> yeah jmkasunich_
[01:33:01] <SWPadnos> yes, the touch one. elo has a Linux driver (I think it's included in X/Linux now)
[01:33:11] <Jymmmmmm> SWPadnos ah cool.
[01:33:26] <SWPadnos> the 17" is also 1280x1024, FWIW
[01:33:33] <Jymmmmmm> oh hell DVI... wonder if it comes with a DVI2VGA adapter
[01:33:55] <SWPadnos> it should have analog as well (mine does)
[01:34:00] <Jymmmmmm> oh, the pic shows a VGA too...
[01:34:07] <SWPadnos> yep
[01:34:17] <Jymmmmmm> SWPadnos does it work well with emc/nix?
[01:34:31] <SWPadnos> if you could get it for $349, that would be a great monitor
[01:34:40] <jmkasunich_> I gotta do something about this compile farm
[01:34:42] <SWPadnos> I haven't used the touch screen on this one yet
[01:35:00] <Jymmmmmm> SWPadnos : Well try it and let me know =)
[01:35:03] <SWPadnos> unfortunately, there's a Windows-only calibration program you need
[01:35:05] <jmkasunich_> 1-1/2 hours to build (because all slots were going at once, and I also had some realtime threads going)
[01:35:06] <SWPadnos> actually, I think it's DOS only
[01:35:24] <Jymmmmmm> SWPadnos DOS works for me, I have a small 10MB partion =)
[01:35:26] <SWPadnos> eek - sequencing the builds would probably be a lifesaver there
[01:35:32] <jmkasunich_> yeah
[01:35:45] <jmkasunich_> didn't want to have to do that, but this is just nutz
[01:35:47] <SWPadnos> I think it may have been DOS, but I don't remember - I've only calibrated my small touchscreen computer
[01:36:01] <SWPadnos> (12" 800x600 - barely adequate for EMC2/axis, plus no other UI panels)
[01:36:54] <SWPadnos> note that the touchscreen interface is serial, not USB
[01:37:25] <Jymmmmmm> SWPadnos Hmmm serial mice ... drivers still around for that?
[01:37:41] <SWPadnos> that's actually a separate module internally, so you could probably get the USB adapter, but it's probably $100
[01:38:01] <SWPadnos> I don't think it's a serial mouse protocol, but as I said, there are drivers
[01:38:32] <Jymmmmmm> oh not serial protocol sucks.... SWPadnos well try yours out and let me know.
[01:38:47] <SWPadnos> one day
[01:38:59] <Jymmmmmm> SWPadnos tomorrow?
[01:39:04] <SWPadnos> probably not that day
[01:39:13] <Jymmmmmm> Sunday?
[01:39:41] <Jymmmmmm> Just a simpe test, nothing fancy...
[01:39:44] <SWPadnos> I'd say that day is also unlikely
[01:40:09] <Jymmmmmm> SWPadnos: Ok, well your sitting on your ass now, so how about now?
[01:40:24] <SWPadnos> sure - I finally get my 22" 3840x2400 monitor working nicely, and you want me to connect up this puny 17" 1280x1024 screen
[01:40:24] <SWPadnos> what a jerk
[01:40:55] <Jymmmmmm> SWPadnos yeah plug it into the VGA connector on the back of or dual input video card =)
[01:41:11] <SWPadnos> how many DVI connections do you suppose 3840x2400 takes?
[01:41:16] <Jymmmmmm> 1
[01:41:25] <SWPadnos> yes, but only at 13 Hz
[01:41:39] <SWPadnos> or two in this case, for a whopping 25-ish Hz
[01:41:46] <SWPadnos> soon, I may fiddle with modelines and get to 33 Hz
[01:41:51] <Jymmmmmm> like you only have ONE machine in the house.... go steal the wifes or something
[01:41:58] <SWPadnos> heh
[01:42:01] <SWPadnos> I
[01:42:02] <Jymmmmmm> or a laptop
[01:42:24] <SWPadnos> I'm too lazy. I think I'll go play a game now :)
[01:42:27] <Jymmmmmm> just wanna know if the touchscreen part is gonna be a pita
[01:42:41] <jmkasunich_> Jymmmmmm: thats easy to answer: yes, it will be a pita
[01:42:44] <jmkasunich_> there, done
[01:42:49] <SWPadnos> from my experience with the other one (also an Elo), it'll be pretty easy
[01:42:56] <jmkasunich_> :-P
[01:43:09] <SWPadnos> it's more of a PITA than a mouse, because there's calibration needed
[01:43:34] <SWPadnos> I haven't tried installing Linux with a touchscreen attached, since BDI 4.30 or so
[01:44:02] <SWPadnos> it may be detected (though serial device detection has rarely worked)
[01:44:08] <SWPadnos> on any OS
[01:49:32] <ds3> most touch screens I've used do not emulate a mouse (in hardware)
[02:00:37] <fenn> you can enter the calibration values by hand, it doesnt take too long to get it right
[02:01:29] <SWPadnos> yep - they're just 4 numbers in the xorg.conf file
[02:01:42] <SWPadnos> though I don't know the acceptable range
[02:02:16] <fenn> the defaults were pretty close
[02:03:57] <SWPadnos> yeah - it's the screen edges that are always the problem
[02:04:17] <SWPadnos> but since Ubuntu (and most any other distro) have a top and/or bottom panel, edges are pretty important
[02:22:34] <tomp> linux touchscreen driver posted today on freshmeat http://www.conan.de/touchscreen/evtouch.html
[02:22:56] <tomp> gonna try with my 'magictouch'
[02:25:11] <tomp> http://www.magictouch.com/addon.html
[03:03:16] <Jymmmmmm> jmkasunich_ I have the molex housing, but can't find a crimper for under $200 yet
[03:04:02] <jmkasunich_> ouch
[03:05:42] <Jymmmmmm> foudn the pins http://www.jameco.com/webapp/wcs/stores/servlet/ProductDisplay?langId=-1&storeId=10001&catalogId=10001&productId=734696
[03:06:35] <Jymmmmmm> the crimper http://www.mouser.com/search/refine.aspx?Msid=53810000&Mkw=538-11-01-0206
[03:10:57] <Jymmmmmm> I think I found a die set for $23!!!
[03:11:06] <Jymmmmmm> err $26
[03:17:35] <tomp> Jymmmmmm: you can use the die set in a vise if careful. btw: ubu 6.06 LTS is almost done cookin ( takes a long time w 48meg )
[03:17:51] <tomp> the $$$ is cuz its from Molex
[03:18:21] <Jymmmmmm> tomp I have a Paladin BNC ratching crimper, as long as the die fits this crimper, I'll be good to go!
[03:18:57] <tomp> cool btw: it just went to boot time ;)
[03:18:58] <Jymmmmmm> Paladin's website, well... sucks. Will have to call.
[03:19:14] <tomp> paladin is not cheap either
[03:19:36] <Jymmmmmm> But I already have the crimper, just need the $26 die
[03:21:07] <tomp> oh, then 26 is not bad investment, cool
[03:23:22] <Jymmmmmm> not at all, IF the die's will work
[03:25:55] <jmkasunich_> jmkasunich_ is now known as jmkasunich
[03:28:36] <tomp> well ubu booted , but using this 'lomem' install method, it drops you to a shell and sez 'you fix it, you're so damn smart' :)
[03:52:03] <CIA-19> 03compile-farm 07BDI-4.51 ( * 10emc2head/: build PASSED
[04:34:15] <jmkasunich> odd... seems like my gecko is only outputing 3.4A or so instead of 6
[04:34:33] <CIA-19> 03compile-farm 07Ubuntu 6.06 LTS (dapper) realtime (2.6.15-magma) * 10emc2.1branch/: build PASSED
[05:06:49] <CIA-19> 03cradek 07TRUNK * 10emc2/src/Makefile: etags
[05:43:22] <jmkasunich> duh... the current is right when I calculate the resistor correctly....
[05:44:26] <cradek> need me to look up ohm's law for you? :-)
[05:44:41] <jmkasunich> its not ohms law that messed me up
[05:44:59] <jmkasunich> the gecko formula is 47K*I/(7-I)
[05:45:15] <jmkasunich> I think I read the first I as a 1
[05:45:51] <jmkasunich> or something
[05:57:42] <jmkasunich> still losing steps....
[06:00:29] <cradek> yuck
[06:00:58] <jmkasunich> I need to find a BNC to gator clips cable, so I can try running it from a signal generator instead of the PC
[06:01:24] <jmkasunich> that will eliminate jitter from the equation, and also the voltage swing issue
[06:01:43] <cradek> oh you're getting steps from the mesa?
[06:01:44] <jmkasunich> the FPGA outputs swing from 3.3 to 0.0 volts
[06:02:06] <jmkasunich> I'm just using the mesa as digital I/O, still running software steps
[06:02:16] <jmkasunich> Pete's new stepper firmware isn't out yet
[06:02:33] <cradek> ah
[06:02:52] <cradek> man I'd use a parport - wouldn't want to fry the mesa experimenting with clip leads
[06:03:05] <jmkasunich> I'm not using clipleads
[06:03:23] <jmkasunich> screwed in wires from breakout board to gecko opto terminals
[06:03:35] <cradek> ah ok
[06:03:40] <jmkasunich> parport lacks the +5V source needed for the geckos
[06:04:00] <jmkasunich> although I might scrounge something up and try the parport to rule out the voltage swing issue
[06:04:36] <cradek> well goodnight - I'll be anxious to see what you come up with
[06:05:01] <jmkasunich> I'm about to call it a night too...
[06:07:02] <jmkasunich> I'm not sure I like the auto-current reduction
[06:07:14] <jmkasunich> it kicks in very quickly (maybe a second after the last step)
[06:07:45] <cradek> yeah I've never been to fond of that idea
[06:08:10] <jmkasunich> of course leaving the current at 6A when the machine is idle isn't so nice either
[06:12:13] <jmkasunich> auto current reduction is controlled by a jumper, so I have options...
[06:12:30] <jmkasunich> it sucks that you have to remove the drive from the heatsink in order to take the cover off and move the jumper
[06:14:24] <jmkasunich> hmm... I could disable auto current reduction, then use a relay to throw another resistor in parallel with the current set resistor to lower the current on demand
[06:36:05] <tomp> you could wire into the jumper, and short it /not sort it externally
[08:12:55] <K`zan> Night folks
[08:14:38] <pier> morning all :)
[08:27:40] <lerneaen_hydra> 'lo all
[08:39:16] <slundell> morning all.
[13:42:49] <Martin_Lundstrom> Hello everyone
[13:43:11] <Martin_Lundstrom> Dallur, are you around? Any tests yet?
[16:18:39] <owhite> hello people.
[16:18:52] <owhite> tomp you available?
[16:40:19] <alex_joni> http://www.youtube.com/watch?v=nEmIAHNkDOo <- if anyone cares about a history lesson :)
[16:51:59] <lerneaen_hydra> alex_joni; is it just me or does romanian (the language) have quite a large latin-based-langauge influence?
[16:58:28] <wb9mjn> Hi All,
[16:59:19] <wb9mjn> Finally got the homing limit disable relay put in. Works great. See the position turn read as it goes past limit (the limt into the EMC
[16:59:31] <wb9mjn> are still active), then come back out ....
[16:59:52] <wb9mjn> The switch only effects the limit inputs to the amplifiers...
[17:01:42] <lerman> Hello, all.
[17:02:02] <wb9mjn> Hi Lerman...
[17:02:08] <lerman> I'm working on GWiz, my GCode Wizard.
[17:02:18] <wb9mjn> Sounds neet....
[17:02:24] <lerman> I could use some opinions..
[17:02:52] <lerman> It will put up a screen (like the mach wizards) showing, for example, a pocket.
[17:03:13] <lerman> It will have labels for center X, Y, depth, finish cut, etc.
[17:04:27] <lerman> The question is: should I put text entry boxes all over the screen next to the labels, or should I have a separate panel with the name of the label and a place for the value (a table with two columns)?
[17:05:02] <wb9mjn> Not an experienced CNC / CAM machinist here....I would like text entry boxes on the drawing, but that may not be
[17:05:06] <wb9mjn> practical..
[17:05:23] <lerman> I did a draft using entry boxes on the graphic and have a concern that it doesn't scale very well. Resizing the graphic causes problems.
[17:05:53] <lerman> (Also, it is harder to create the screens).
[17:06:01] <wb9mjn> An alternative would be a new window that pops up offset from the drawing...My main gripe is that the learning curve is
[17:06:13] <wb9mjn> steep because you have to figure out the asociations....
[17:06:21] <wb9mjn> If its obvious, things are quicker...
[17:06:35] <CIA-19> 03jepler 07v2_1_branch * 10emc2/debian/control.in: depending on libgnomeprint2.2-dev was not enough to get printing support for classicladder
[17:06:52] <lerman> Well, the names would still be on the graphic AND in the table. So, it is pretty obvious.
[17:07:10] <wb9mjn> Once you know where to look....
[17:10:06] <wb9mjn> Anyway....have to go make lunch.....I ll be around off and on today....
[17:10:16] <lerman> In general, you would just look at the table. CenterX and CenterY are pretty obvious. Radial Finish Cut is, also. Percent Stepover is not so obvious. Percent of what. The drawing shows it to help clarify that it is the percent of the tool diameter. I would also have a tool tip that would say that when you rolled over the table entry.
[17:10:25] <lerman> I'm off to lunch, also.
[17:10:33] <lerman> See you.
[17:20:28] <CIA-19> 03jepler 07TRUNK * 10emc2/src/Makefile.inc.in: finish separating the build of html and pdf docs
[17:20:28] <CIA-19> 03jepler 07TRUNK * 10emc2/docs/src/Submakefile: finish separating the build of html and pdf docs
[17:21:04] <CIA-19> 03jepler 07TRUNK * 10emc2/debian/ (control.in rules.in): fix up dependencies: include libgnomeprintui2.2-dev, get rid of latex2html
[17:38:01] <SWPadnos_> SWPadnos_ is now known as SWPadnos
[18:14:26] <owhite> hey people. I have a question.
[18:15:00] <owhite> I have a vcp.hal file that calls load hal_parport cfg="0x9800"
[18:15:13] <owhite> and then I'm driving various pins using that parport.
[18:15:47] <owhite> its the second one that gets activated. the other one is activated using a UCS, or univstep_io.hal
[18:16:25] <owhite> anyone have a suggestion of how I can drive the second parallel port using M64 commands?
[18:20:20] <jepler> motion.digital-out-NN OUT bit
[18:20:22] <jepler> These pins are controlled by the M62 through M65 words.
[18:20:24] <jepler> ^^ the docs say this
[18:20:44] <jepler> so .. just hook these pins to the second parallel port's pins in your hal file?
[18:21:15] <jepler> make sure you also put the parport component's functions in the appropriate thread with addf
[18:21:19] <owhite> oh okay I think I get it.
[18:21:40] <owhite> hal_parport is connected to servo-thread
[18:23:02] <jepler> that should be fine, the signal can't change any more frequently than that
[18:23:25] <owhite> thanks
[18:24:36] <CIA-19> 03jepler 07TRUNK * 10emc2/src/hal/halmodule.cc: add support for RTAPI shared memory segments
[18:45:22] <owhite> jepler: could you point me to the docs for motion.digital?
[18:47:24] <jepler> http://www.linuxcnc.org/docs/html/config/emc2hal/emc2hal.html http://www.linuxcnc.org/docs/html/gcode/main/#sec:M62-to-M65:
[18:50:43] <jepler> looks like M62/M63 may not actually do anything, but M64/M65 do set and reset those pins according to halshow
[18:51:47] <owhite> got it. I'm working on linksp-ing some stuff. its confusing but I'll kep at it.
[18:51:55] <owhite> s/kep/keep/
[18:52:11] <awallin> owhite: what are you controlling with the digital IO ?
[18:52:51] <owhite> lot of stuff. motor power supply, laser power supply, ventilation...
[18:53:24] <CIA-19> 03cradek 07TRUNK * 10emc2/nc_files/flowsnake.py: python implementation of lerman's flowsnake algorithm
[18:55:22] <awallin> I wonder how many digital outputs there are ? the docs say that M64 PX has some implementation related maximum X
[19:02:53] <jepler> awallin: did you look at 'halcmd show'? It answers that question
[19:16:08] <owhite> aw crap! shouldnt this drive my parport pin.? linkpp motion.digital-out-00 => parport.0.pin-06.out
[19:16:29] <owhite> I figured if I ran M64 P0 that would do it
[19:58:07] <alex_joni> lerneaen_hydra: that is correct
[19:58:31] <lerneaen_hydra> alex_joni; what about the pronunciation?
[20:01:08] <owhite> hey alex_joni: how do you link more than on pin to another pin?
[20:03:18] <alex_joni> owhite: halcmd net
[20:05:30] <jmkasunich> I'm liking these steppers.... 2400 RPM with pretty decent torque
[20:05:39] <alex_joni> jmkasunich: hi there
[20:05:44] <jmkasunich> hi
[20:07:19] <jmkasunich> stall point is somewhere between 3000 and 3600 rpm
[20:07:27] <jmkasunich> 100 and 120KHz step rate
[20:09:44] <alex_joni> you're running the 5i20 already?
[20:09:53] <alex_joni> or a signal gen?
[20:09:54] <jmkasunich> no
[20:09:58] <jmkasunich> signal generator
[20:10:01] <alex_joni> right.. nice
[20:10:28] <jmkasunich> I was running software generated steps out thru the 5i20 digital outs yesterday
[20:10:34] <jmkasunich> and saw some lost step
[20:10:43] <jmkasunich> maybe 3-4 steps per rev (2000 steps)
[20:11:24] <owhite> okay people. if I may prevail upon you for some help.
[20:11:45] <owhite> oh hang on...
[20:12:55] <owhite> alright. I'd like to linkpp motion.digital-out-00 parport.0.pin-06-out
[20:13:30] <owhite> and then also control parport.0.pin-06-out in pyvcp.hal...
[20:13:42] <owhite> net bit5 pyvcp.chbutton-sample5 => pyvcp.led-sample5 parport.0.pin-06-out
[20:14:03] <owhite> when i do that, motion.digital-out-00 gets disconnected.
[20:14:10] <jmkasunich> right
[20:14:16] <owhite> any suggestions?
[20:14:20] <jmkasunich> you can't control one thing from two places
[20:14:34] <jmkasunich> either control it from the motion controller, or from the pyvcp
[20:15:08] <owhite> so I cant control something using M64 P0 and pyvcp?
[20:15:50] <jmkasunich> you can't directly do that, just like you can't simultaneously play a CD and the radio thru one set of speakers on a stereo
[20:15:58] <jmkasunich> only one source per output
[20:16:14] <jepler> there are HAL components that combine two inputs, such as 'and2'
[20:16:24] <jmkasunich> you _can_ use a logic gate (AND, OR, etc) to combine two sources if you really want to
[20:16:36] <owhite> logic as in ladder?
[20:16:42] <jmkasunich> that would work to
[20:16:53] <owhite> I dont want to go with ladder.
[20:16:55] <jmkasunich> but and2 is simpler
[20:17:12] <owhite> and2 doesnt appear to be a halcmd
[20:17:20] <jepler> and2 is a HAL component
[20:17:21] <jepler> man and2
[20:17:21] <jmkasunich> and2 says 'if motion.whatever is on AND the pyvcp is on, then the output is on"
[20:17:31] <jmkasunich> there is also an OR function
[20:18:14] <jmkasunich> "loadrt and2" to create the gate, "addf and2.0 servo-thread" to run it, and various link commands to connect it up
[20:18:22] <owhite> is the or function a component?
[20:18:26] <jmkasunich> yep
[20:18:36] <jmkasunich> "loadrt or2" I believe
[20:18:54] <owhite> yep there it is. :-)
[20:19:18] <owhite> alright. back to the drawing board. :-) one more question though...
[20:20:03] <owhite> when I turned on a bunch of appliances with pyvcp, which is good, when I exit from tkemc and it brings down the pyvcp stuff as well, which is good, but it doesnt turn everything off.
[20:20:16] <owhite> is there a way I can make sure pins are set to false on exit?
[20:20:40] <jmkasunich> there are two levels to that
[20:20:45] <owhite> did that make sense?
[20:20:50] <owhite> * owhite listens.
[20:21:04] <jmkasunich> one is estop - if the computer crashes, you probably want everything to turn off - WITHOUT relying on software
[20:21:18] <jmkasunich> two is software, for convenience but not for safety
[20:22:28] <jmkasunich> ideally your hardware has some kind of master enable signal that you can connect to your estop chain
[20:22:42] <jmkasunich> "if this signal goes away, all the relays turn off regardless of what the parport is doing"
[20:22:46] <owhite> right. it does and that works okay.
[20:23:04] <jmkasunich> so then all you really need is to turn off that signal at shutdown
[20:23:10] <owhite> but now I've hooked up things like ventilation, which I think does not need to be connected to e-stop.
[20:24:07] <jmkasunich> well, the state of a parport pins is NOT guaranteed from the time you power the PC on until the time you start running the RT parport driver code
[20:24:21] <jmkasunich> and its also not guaranteed from when you shut down hal until you power down your PC
[20:24:23] <SWPadnos_> SWPadnos_ is now known as SWPadnos
[20:24:27] <owhite> mm-hm.
[20:25:18] <owhite> does that mean the software doesnt provision for a way to specify the states of the parport on shutdown?
[20:25:24] <jmkasunich> right
[20:25:56] <owhite> okay, fair enough.
[20:26:01] <jmkasunich> you might say "we should have such a feature", but that still doesn't do anything about the unknown state at power up
[20:26:12] <jmkasunich> the proper solution is not to rely on the software
[20:26:26] <owhite> well actually at power up is does shut everything to a particular state, I think.
[20:26:34] <owhite> s/shut/change/
[20:26:47] <alex_joni> owhite: only if you're lucky
[20:26:54] <alex_joni> and it's some time of the month
[20:26:54] <owhite> ok.
[20:26:57] <jmkasunich> right
[20:27:25] <jmkasunich> the boot process might go out and touch port, you never can be sure what is going to happen
[20:27:31] <jmkasunich> never ever trust software
[20:27:44] <SWPadnos> there's a danger there with automatic updates (or updates in general)
[20:28:01] <owhite> but doesnt that mean that you're not gauranteeing that the machine could run off in some wild state on startup?
[20:28:08] <SWPadnos> the default behavior of the kernel / OS / drivers may change, and you can't rely on it to do the right thing if you change *anything* in the system
[20:28:34] <jmkasunich> it is YOUR responsibility to make sure that no software fault could possibly result in unsafe behavior
[20:28:40] <jmkasunich> that means estop in hardware
[20:28:41] <SWPadnos> owhite, correct - there's nothing (in software) to guarantee that the machine will be safe as the PC boots, and before EMC runs
[20:28:44] <SWPadnos> that's why you use hardware
[20:28:58] <owhite> fair enough.
[20:29:16] <SWPadnos> anyone who tells you differently is selling something ;)
[20:29:47] <owhite> the expectation is that the user should tie anything that is dangerous to estop. power up, and inspect the system to ensure that important states are set properly.
[20:29:50] <alex_joni> I'm usually selling something.. but that's always HW for safety
[20:30:31] <SWPadnos> and you don't seem to be telling anyone that your software guarantees safe operation (even before it's run :) )
[20:30:47] <jmkasunich> my choice for anything that is the least bit dangerous is some sort of charge pump
[20:30:58] <jmkasunich> where the software has to repeatedly and rapidly (1KHz) toggle a pin
[20:31:18] <jmkasunich> if that pin sits still for more than a few mS, the hardware opens the main estop chain
[20:31:59] <jmkasunich> the main estop chain then turns off everything
[20:35:33] <Guest713> hello friends
[20:35:43] <Guest713> i have one problem
[20:40:51] <owhite> when you create a signame with newsig, is there a way to export that signal to another hal file?
[20:41:08] <owhite> or is that something I also should not be doing?
[20:41:10] <owhite> :-)
[20:41:13] <alex_joni> owhite: the signal remains
[20:41:21] <alex_joni> if you load 2-3 hal files in a certain order
[20:41:32] <alex_joni> then you can use the same name in all files
[20:41:39] <alex_joni> newsig in the first file, linksp in the others
[20:41:42] <SWPadnos> there is no such mechanism - there is only one runnung HAL at any time. all signals are available to all hal files once they are created
[20:42:08] <alex_joni> owhite: once you run them out of order you'll have problems though
[20:42:26] <owhite> right but the order is specified in the .ini file right?
[20:43:21] <owhite> I mean if its a bad practice I wont do it, but it seemed like the reason you're newsig signames is so they are more human-readable. then you'd use them in other hal files.
[20:43:47] <Jymmmmmm> shielded cable to the steppers running parallel to the 120VAC for the spindle. Problems?
[20:43:48] <SWPadnos> it's not bad practice - that's what all the sample configs do
[20:44:14] <Jymmmmmm> OMG... Stanley bought out Jensen Tools. I'll be damn
[20:44:31] <SWPadnos> they load core_stepper or core_servo with one file, then some specific driver with another, then do I/O to motion controller connections
[20:44:37] <alex_joni> owhite: no problems in doing what was described above
[20:52:19] <owhite> how come configs/sim/simulated_hom.hal has or2 commands in it, but it doesnt loadrt or2? it also does halcmd addf or2.
[20:53:28] <alex_joni> loadrt or2 count=1
[20:53:33] <alex_joni> in core_sim.hal
[20:53:58] <owhite> oh sorry.
[20:54:17] <CIA-19> 03jepler 07TRUNK * 10emc2/scripts/emc.in: the [HAL]SHUTDOWN script will be useful for things like setting digital outputs to a known value after a normal shutdown
[20:55:59] <alex_joni> owhite: see.. now you can use a hal file to set some values on shutdown :)
[20:56:49] <owhite> wow. that's really nice man.
[20:56:58] <jmkasunich> in that file, you'll need to unlink the parport pins from whatever is driving them, then setp the pins to the desired state
[20:57:46] <owhite> how did you handle specifying the shutdown script? is that in the .ini file or something?
[20:58:03] <jmkasunich> yeah
[20:58:04] <alex_joni> yup..
[20:58:13] <jmkasunich> in the [HAL] section
[20:58:27] <alex_joni> SHUTDOWN = newfile.hal
[20:58:48] <owhite> that's really nice. 'preciate it.
[20:59:53] <alex_joni> owhite: the bad news is that it's probably be a while till it is in a package
[20:59:59] <jepler> "now" being when emc 2.2 comes out or you compile it yourself :-P
[21:00:04] <alex_joni> unless you run sources from cvs
[21:00:16] <owhite> no problem. I have it all documented how to do cvs -up
[21:00:23] <alex_joni> ok, coo
[21:00:33] <owhite> ...trying to actually _learn_ something :-)
[21:06:19] <skunkworks> is xlite still a jenson product?
[21:06:56] <Jymmmmmm> skunkworks screwdrivers, pliers, or blades?
[21:11:02] <skunkworks> all
[21:11:14] <skunkworks> I have a set - really like it.
[21:11:39] <skunkworks> nut-drivers, blades and pliers.
[21:13:15] <Jymmmmmm> skunkworks Ah, those... not Jensen tools (just a supplier), buy you mean Cooper Tools, the mfg of Cresent, Xacto, and a few others.
[21:13:50] <Jymmmmmm> xcelite
[21:13:56] <skunkworks> ah
[21:14:28] <Jymmmmmm> http://www.cooperhandtools.com/brands/xcelite/index.cfm
[21:14:56] <Jymmmmmm> Now, why in the hell I remembered that from 15 years ago, I'll never know =)
[21:28:28] <owhite> alright folks. trying to run the or2 on some pins...
[21:28:41] <owhite> newsig ventilation bit
[21:28:43] <owhite> linksp ventilation <= motion.digital-out-00
[21:28:43] <skunkworks> http://www.cooperhandtools.com/brands/CF_Files/model_detail.cfm?upc=037103482471
[21:28:46] <skunkworks> that is my set
[21:29:13] <owhite> linkpp pyvcp.chbutton-sample5 => pyvcp.led-sample5
[21:29:13] <owhite> linksp ventilation or2.0.in0
[21:29:13] <owhite> linkpp pyvcp.led-sample5 or2.0.in1
[21:29:13] <owhite> linkpp parport.0.pin-06-out or2.0.out
[21:29:45] <owhite> so that's me trying to or2 between ventilation and pyvcp to the parport pin.
[21:29:57] <owhite> but it doesnt work.
[21:30:09] <SWPadnos> did you add the or2 function to a thread?
[21:30:19] <SWPadnos> and issue the halcmd "start" command?
[21:30:25] <owhite> addf or2.0 servo-thread
[21:30:38] <owhite> I dont use start if its in a .hal file, right?
[21:31:05] <SWPadnos> you do need to start, especially if EMC doesn't do it for you :)
[21:33:17] <owhite> but configs/dallur-thc uses or2 and it doesnt use start.
[21:33:24] <alex_joni> SWPadnos: if it's a hal file started by emc2 then there's no need for start
[21:33:35] <SWPadnos> true
[21:33:39] <alex_joni> the runscript issues start after loading the hal files
[21:34:01] <owhite> alex_joni: any guess where I went wrong with the or2?
[21:34:26] <alex_joni> linkpp ?
[21:34:45] <SWPadnos> and you have the parport.0.write function in the thread as well?
[21:35:16] <alex_joni> owhite: "went wrong" would imply we know what happened
[21:35:25] <alex_joni> does it run but doesn't turn on the pin?
[21:35:28] <alex_joni> or does it barf?
[21:35:37] <owhite> both are attached to servo-thread
[21:35:43] <owhite> it runs but doesnt turn on the pin.
[21:36:00] <SWPadnos> can you post the whole hal file to http://pastebin.ca ?
[21:36:08] <jmkasunich> owhite: word of advice
[21:36:11] <owhite> actually, M64 P0 does turn on parport, but pyvcp
[21:36:12] <jmkasunich> never ever ever use linkpp
[21:36:17] <owhite> oh.
[21:36:40] <alex_joni> http://pastebin.ca/361006
[21:36:51] <jmkasunich> it was written as a shortcut, to save "newsig foo; linksp foo pin1; linksp foo pin2"
[21:36:55] <SWPadnos> does the pin turn on while you hold down the pyvcp button? those aren't toggles, they're momentary
[21:37:02] <jmkasunich> but "net foo pin1 pin2" does the same thing
[21:37:23] <jmkasunich> linkpp pin1 pin2 creates a signal to connect the two pins, but it calls the signal "pin1"
[21:37:32] <jmkasunich> which in my opinion is very confusing
[21:37:55] <owhite> okay. no linkpp
[21:39:04] <owhite> first hal file: http://pastebin.ca/361008
[21:40:14] <owhite> second hal file: http://pastebin.ca/361009
[21:43:54] <SWPadnos> is a pyvcp LED clickable, or is that an input pin on the LED?
[21:44:05] <SWPadnos> it looks like you connecte two inputs with no output to drive them
[21:44:08] <SWPadnos> connected
[21:44:32] <SWPadnos> (that would be pyvcp.led-sample5 and or2.0.in1)
[21:44:34] <owhite> its an input pin to an LED.
[21:45:02] <SWPadnos> ok, so what drives the new signal pyvcp.led-sample5 (created by linkpp(?
[21:45:04] <SWPadnos> )
[21:45:43] <owhite> pyvcp.chbutton-sample5 which is a radio button in in the pyvcp.xml
[21:48:02] <SWPadnos> oh - that :)
[21:49:00] <robin_sz> meep?
[21:50:04] <robin_sz> so
[21:50:14] <robin_sz> how can I build a big ammeter?
[21:50:42] <robin_sz> I suspect measuring the secondary current in my spot welder is too hard
[21:50:48] <robin_sz> currents too high etc
[21:50:58] <robin_sz> measuring primary current will have to do
[21:51:13] <jmkasunich> how many amps?
[21:51:24] <robin_sz> secondary amps?
[21:51:28] <jmkasunich> yeah
[21:51:29] <robin_sz> 20K
[21:51:52] <jmkasunich> yeah, thats a bit high to measure directly (even with a CT)
[21:51:58] <robin_sz> thought so
[21:52:25] <robin_sz> hmmm ... what about using the huge copper connecting strap as a shunt?
[21:52:26] <jmkasunich> the current is AC, right?
[21:52:29] <robin_sz> right
[21:52:40] <jmkasunich> CT wins huge over shunts
[21:52:49] <robin_sz> true
[21:52:54] <skunkworks> CT?
[21:52:58] <robin_sz> R changes etc
[21:53:07] <jmkasunich> current transformer
[21:53:14] <skunkworks> ah
[21:53:35] <robin_sz> theres a couple of BIG flexible straps that connec the head .. like 25mm dia flexible cables
[21:53:39] <jmkasunich> http://www.crmagnetics.com/newprod/ProductView.asp?ProdName=Commercial_Class_Current_Transformers
[21:53:39] <robin_sz> no 30mm
[21:54:30] <robin_sz> so ... primary side current is best bet then?
[21:54:49] <jmkasunich> if you know the turns ratio, and you can tolerate some inaccuracy
[21:54:55] <robin_sz> yeah
[21:55:03] <robin_sz> I dont really want to know the current ...
[21:55:09] <robin_sz> just relative current
[21:55:12] <jmkasunich> the primary current should be pretty representative of the secondary current
[21:55:29] <jmkasunich> although it will include some transformer magnetizing current
[21:55:44] <robin_sz> "it should read at least 5, if it drops below 4, clean your weld tips"
[21:56:16] <robin_sz> so ... I put a CT on the frimary side ...
[21:56:19] <jmkasunich> what is the primary current?
[21:56:20] <robin_sz> and then what?
[21:56:22] <jmkasunich> 10A, 100A?
[21:56:35] <robin_sz> mmm can push closer to 200A
[21:56:41] <jmkasunich> ok, no prob
[21:57:25] <jmkasunich> for instance, CR Magnetics series CR2 has a 28mm dia hole in the middle, and is available with 300:5 ratio
[21:57:30] <robin_sz> so CT ... then what, some sort of moving iron meter?
[21:57:37] <jmkasunich> that means with 300A in the primary, you have 5A in the secondary
[21:57:46] <jmkasunich> connect to any AC ammeter than can read 5A
[21:57:48] <robin_sz> right ...
[21:57:54] <robin_sz> ac ammeter ...
[21:58:25] <jmkasunich> simpson multimeter, fluke digital multimeter, or dedicated ammeter
[21:59:22] <jmkasunich> how long does the current last?
[21:59:35] <robin_sz> .4s
[21:59:36] <jmkasunich> meter settling time and/or response time might be a problem
[21:59:42] <jmkasunich> ah... thats the rub
[22:00:06] <robin_sz> well, 20 to 32 cycles typically
[22:00:36] <jmkasunich> you almost want something that will capture and hold the peak reading
[22:00:47] <jmkasunich> because during the shot, they're gonna be looking at the part, not the meter
[22:01:16] <robin_sz> something to do with a shunt, a diode and a C then
[22:01:50] <jmkasunich> http://www.globalspec.com/FeaturedProducts/Detail/WeschlerInstruments/Columbia_ACDC_ClampOn_Tong_Test_Ammeters/2865/0?fromSpotlight=1
[22:02:19] <jmkasunich> is the goal to make an occaisional measurement, or to build into the machine and check every weld?
[22:02:27] <robin_sz> build in
[22:02:40] <robin_sz> not every weld, but every now and then
[22:02:46] <jmkasunich> go/no-go kind of thing?
[22:03:09] <robin_sz> yeah
[22:03:18] <robin_sz> "at least 3 for 1.2mm steel"
[22:03:40] <jmkasunich> duration is reliabely controlled by other means - no need to measure that?
[22:03:55] <robin_sz> right .. .cycle counters
[22:04:26] <jmkasunich> the problem with diodes is voltage drop
[22:04:47] <jmkasunich> a shunt in the primary would have to be limited to a few tens of millivolts
[22:05:07] <jmkasunich> a shunt on the 5A secondary of a CT could probably tolerate a volt or so
[22:05:12] <robin_sz> ct - shunt .- diode - c
[22:05:27] <jmkasunich> but if you are using diodes and caps, you wnat several volts, so the 0.6v diode drop doesn't mess things up
[22:05:45] <robin_sz> true
[22:06:02] <jmkasunich> hmm
[22:06:30] <jmkasunich> CT, shunt, 120V(or 220V) to 6V transformer wired backwards
[22:06:35] <jmkasunich> diode, cap, meter
[22:06:42] <jmkasunich> the transformer doesn't need to be large
[22:06:59] <jmkasunich> 2-5VA, maybe a 3cm cube
[22:07:19] <jmkasunich> size the shunt for 1/2 to 1 V at 5A
[22:07:23] <jmkasunich> 0.1 ohms for example
[22:07:27] <owhite> alright. I got it. I've been confusing the difference between signames and pins. :-|
[22:07:43] <robin_sz> right ... sounds like a plan!
[22:07:46] <jmkasunich> owhite: linkpp will do that to you because it names signals after pins
[22:08:03] <jmkasunich> 120V->6V transformer in reverse will take 0.5V up to 10
[22:08:10] <owhite> right, but I was also doing it with linksp -- cause I'm not too bright.
[22:09:32] <robin_sz> where did the 0.1ohms come from?
[22:09:48] <robin_sz> secondary resistance of the txfmr?
[22:09:58] <jmkasunich> 5A (ct secondary, 5A at full scale is somewhat of a standard) * 0.1 ohm = 0.5 volts
[22:10:09] <robin_sz> yes ...
[22:10:20] <robin_sz> but where did the the 0.1 come from?
[22:10:36] <jmkasunich> I picked it to get 0.5V at full scake
[22:10:38] <jmkasunich> scale
[22:10:59] <jmkasunich> CTs are designed to transform current - they are most accurate when operated into a short circuit
[22:11:09] <robin_sz> right ...
[22:11:14] <jmkasunich> so you want to keep the secondary voltage low
[22:11:22] <robin_sz> so I dont need a 0.1R anywhere in the circuit ...
[22:11:22] <jmkasunich> hence the 0.5V
[22:11:28] <jmkasunich> yes you do
[22:11:38] <robin_sz> in series with the txfmr 6v winding?
[22:11:45] <robin_sz> across it?
[22:11:51] <jmkasunich> stand by
[22:11:54] <robin_sz> ahh ... got it
[22:12:10] <robin_sz> ct drives 5a into 0.1R
[22:12:26] <robin_sz> txfmr magnifies the volts across the R
[22:12:39] <robin_sz> volts seen on other side of the txfmr ...
[22:12:48] <robin_sz> diode, cap meter ...
[22:13:02] <robin_sz> thanks, understandign has dawned :)
[22:13:51] <jmkasunich> http://www.drawblog.com/images/20070217021332909.jpg
[22:14:16] <jmkasunich> drawing with a mouse can make one look like a spastic retard....
[22:14:23] <robin_sz> nice :)
[22:14:27] <robin_sz> drawblog .. useful!
[22:14:59] <jmkasunich> yeah, ran across it today, this is the first real use for me
[22:16:35] <robin_sz> useful ...
[22:17:02] <jmkasunich> xxx:5 is pretty much standard for CTs, you shouldn't have too hard a time finding 200:5 or whatever makes sense
[22:17:10] <jmkasunich> the rest of the parts are mundane
[22:17:25] <robin_sz> right
[22:17:29] <robin_sz> thanks for that
[22:17:32] <robin_sz> most helpful
[22:17:33] <jmkasunich> sure
[22:18:15] <robin_sz> http://www.drawblog.com/images/20070217021748395.jpg
[22:18:21] <robin_sz> see. just as bad
[22:18:30] <robin_sz> mousing is not easy is it
[22:18:30] <jmkasunich> ;-)
[22:18:42] <jmkasunich> no, its not
[22:18:54] <robin_sz> weird as I CAD all day with a mouse
[22:18:56] <jmkasunich> you get difficulty points for drawing the CT secondary
[22:19:05] <robin_sz> :)
[22:19:12] <jmkasunich> snap grids and such make a world of difference
[22:19:22] <robin_sz> turn1 went bad, but it got better from there on in
[22:19:56] <robin_sz> kewl, it does colours!
[22:19:59] <robin_sz> sorry, colors
[22:31:37] <anonimasu> night
[22:34:35] <Jymmmmmm> * Jymmmmmm lol @ robin_sz.... Silly Brit!
[22:35:01] <alex_joni> night all
[22:35:06] <Jymmmmmm> G'Night anonimasu
[22:35:12] <Jymmmmmm> Jymmmmmm is now known as Jymmm
[22:45:33] <SWPadnos> heh - I actually just had to ask some Mormons to leave my property
[22:45:43] <SWPadnos> maybe I should buy a gun - they might listen faster
[22:46:00] <jmkasunich> they were evangalizing?
[22:46:05] <SWPadnos> yeah
[22:46:18] <SWPadnos> when I said "no thanks", they didn't take the hint
[22:46:23] <jmkasunich> thats annoying
[22:46:32] <SWPadnos> when I said "life is random, there's no higher purpose", they didn't take the hint
[22:46:38] <jmkasunich> tell them you worship satan
[22:46:43] <jmkasunich> or that you're a muslim
[22:46:46] <SWPadnos> when I said "life's what you make it", they didn't take the hint
[22:46:47] <SWPadnos> heh
[22:47:03] <SWPadnos> when tey said "I need to go to church and kneel down and pray", I told them to take their message elsewhere :)
[22:47:14] <SWPadnos> I being me in that sentence
[22:48:29] <jmkasunich> this motor has quite a resonance
[22:48:33] <SWPadnos> I usually like toying with them, but I didn't sleep at all last night and I may be getting a stomach flu, so I'm not in the mood to play with them
[22:48:37] <SWPadnos> the Keling?
[22:48:44] <jmkasunich> yesh
[22:48:48] <jmkasunich> yeah even
[22:49:03] <SWPadnos> been hitting the sauce a little too much tonight? :)
[22:49:06] <jmkasunich> 130 microsteps per second, and again at 2600 uSteps per second
[22:49:19] <jmkasunich> which is 260 full steps/sec
[22:50:00] <SWPadnos> did you see (or already know) Mariss' recent post about the ratio of torque per microstep?
[22:50:27] <jmkasunich> the one that talks about adjusting a pot for smoothest operation?
[22:50:34] <SWPadnos> could be
[22:50:39] <jmkasunich> I did that
[22:50:53] <SWPadnos> the 1200 oz-in stepper that had resonance issues
[22:50:58] <jmkasunich> the manual recommends adjusting at about 1/4 rev/sec (500 uSteps per second)
[22:51:54] <jmkasunich> this one resonates at about 65Hz
[22:52:22] <SWPadnos> could it be 60, or is that a scope-driven number?
[22:52:32] <SWPadnos> scope-measured ...
[22:54:06] <jmkasunich> uSteps at 65Hz will do it, uSteps at double that (130 Hz), and again at 2600 or so, where the current sine wave is 65Hz
[22:54:19] <jmkasunich> frequencies measured with scope and counter
[22:54:47] <SWPadnos> ok
[22:55:07] <SWPadnos> that seems very low. is the motor attached to anything>
[22:55:08] <SWPadnos> ?
[22:55:21] <jmkasunich> nope
[22:55:55] <SWPadnos> wouldn't that tend to make the motor resonate at a low frequency
[22:56:12] <SWPadnos> kind of like having the rotor bounce after hitting the next step position rapidly
[22:56:24] <jmkasunich> could be
[22:56:37] <jmkasunich> spring/mass = magnetic field and rotor inertia
[22:56:46] <SWPadnos> yep
[22:57:08] <SWPadnos> looks like it's movie time - bbl
[22:58:49] <jmkasunich> I bet belt driving a leadscrew will damp it quite a big
[22:58:50] <jmkasunich> bit
[22:59:53] <owhite> hey I used the SHUTDOWN specification in .ini file and its working :-)
[23:00:02] <jmkasunich> cool
[23:00:04] <owhite> ...of course.
[23:05:26] <owhite> I'm off. thanks for you're help people.
[23:58:41] <CIA-19> 03jepler 07TRUNK * 10emc2/docs/src/gui/axis.lyx: document [HAL]SHUTDOWN. improved other things too
[23:58:41] <CIA-19> 03jepler 07TRUNK * 10emc2/docs/src/config/ini_config.lyx: document [HAL]SHUTDOWN. improved other things too
[23:59:35] <CIA-19> 03jepler 07TRUNK * 10emc2/debian/changelog: new item