#emc | Logs for 2007-06-05

[00:16:37] <Ziegler> what var in the ini changes the direction the stepper...
[00:16:51] <Ziegler> to fix the fact that I have it wired backwards
[00:16:54] <jepler> Ziegler: negate "scale"
[00:17:00] <Ziegler> ah :)
[00:17:03] <jepler> INPUT_SCALE to be more precise
[00:17:04] <Ziegler> easy enough
[00:17:09] <Ziegler> danke
[00:20:52] <jepler> bitte
[00:58:24] <Twingy> I've got working gerber to gcode now
[01:24:42] <Ziegler> gerber to gcode?
[01:24:47] <Ziegler> isnt gerber gcode?
[01:33:10] <jtr> jepler: ever get an answer on your square wave question?
[01:41:13] <jepler> jtr: no
[01:41:22] <jtr> jepler: a universal counter/timer is (was) the usual test gear for measuring the period of a pulse - don't remember if they can easily be set to watch a pulse train for a long pulse.
[01:41:53] <jtr> You have microcontrollers, yes?
[01:42:09] <jepler> yes -- I can imagine constructing a microcontroller program to do this
[01:42:20] <jepler> but I was more wondering if there is a special kind of device to do this that would be in a well stocked lab
[01:44:24] <jtr> I haven't been in a well stocked lab in a long time - the tools around today have advanced considerably since then.
[01:44:54] <jepler> while in my case, I've never been in a proper electronics lab
[01:45:19] <ds2> couldn't a basic 555 setup as a missing pulse detector do it?
[01:49:44] <jtr> ds2: that would be one way of doing it, but it may be faster to program a micro devel system than to build a circuit.
[01:50:40] <jepler> probably depends on the time spans you want to measure
[01:52:49] <jepler> I imagined a microcontroller that would measure with 1us granularity and maintain a readout of 4 numbers: longest/shortest and high/low
[01:53:29] <Twingy> Ziegler, no
[01:53:51] <Ziegler> whats the difference?
[01:53:54] <Twingy> Ziegler, gerber is a union of traces, g-code is an outline of the traces, very different things
[01:53:59] <jtr> I worked in a factory maintaining automatic test equipment up until '88, in the next job I built one to test a product I designed. The test eqip was off the shelf stuff
[01:54:15] <Ziegler> hmmm
[01:54:22] <Ziegler> interesting... ill have to read a bit more into it
[01:54:25] <Twingy> gerber is very easy cause things can overlap (hence unions)
[01:54:55] <Twingy> where as g-code has to have the outline of all these unions, which requires multiple passes on the gerber data
[01:55:22] <Twingy> I've been working on this for the past 2 months in my spare time
[01:55:54] <Twingy> ~1k lines of code
[01:55:54] <jepler> * jepler wanders off
[01:55:55] <Ziegler> I feel dumb... I always thought that the g in g-code stood for gerber
[01:56:46] <Twingy> if you made circuit boards you'd know about all of this
[01:57:30] <Unit41> lier
[01:57:37] <Unit41> its been longer than 2 months :P
[01:58:12] <Unit41> shit it seems like forever ago
[01:58:50] <Twingy> well, last gcam release was 04.09
[01:58:59] <Twingy> so not quite 2 months
[01:59:17] <Twingy> I will probably make the next release on the 10th
[01:59:46] <Twingy> releasing gerber->g-code as beta for the first iteration
[02:00:09] <Unit41> me slams a CANADIAN for celebration
[02:00:30] <Twingy> I'd like to see gcam in the next EMC Ubuntu LiveCD distro
[02:00:44] <Twingy> or have a debian package made
[02:01:04] <martin_lundstrom> Hello, is anyone aware of the jogging problem with the latest trunk?
[02:01:52] <jepler> martin_lundstrom: no -- perhaps you should describe it
[02:02:06] <Unit41> Ide never give ubuntu a second chance
[02:02:13] <Unit41> or most other distros for that matter
[02:02:19] <martin_lundstrom> I can only jog once, then I get a lock situation
[02:02:20] <Ziegler> ahh.. I see gerber files are specifically for doing printed circuit boards
[02:02:39] <jepler> martin_lundstrom: which GUI? How are you jogging (e.g., keys, mouse click, etc)?
[02:03:05] <Twingy> martin_lundstrom, try adjusting your max velocity and accelerations
[02:03:14] <martin_lundstrom> jepler: axis
[02:03:31] <Twingy> does the program freeze?
[02:03:34] <Unit41> suse works great for me and if I were going to switch Ide do it next release of opensolaris
[02:03:37] <martin_lundstrom> Twingy: no
[02:03:46] <martin_lundstrom> only the jogging
[02:03:47] <Unit41> open solaris 11
[02:03:55] <Ziegler> Twingy: isnt gerber and g-code the exact same code?
[02:04:04] <Ziegler> command wise?
[02:04:09] <Twingy> no
[02:04:15] <Twingy> very differenct
[02:04:18] <Twingy> *different
[02:04:20] <Ziegler> minus tooling details?
[02:04:26] <Twingy> different format/syntax
[02:04:33] <Twingy> different commands
[02:04:42] <Twingy> very different
[02:04:53] <Twingy> only thing they have in common is they are ASCII
[02:05:18] <Twingy> however a drill file is g-code or very close to it usually
[02:05:26] <Ziegler> so no G1 codes or G0 codes?
[02:05:28] <jepler> martin_lundstrom: how are you jogging? by pressing a key (which key?) or by clicking?
[02:05:32] <Twingy> Ziegler, correct
[02:06:15] <martin_lundstrom> jepler: by cllicking
[02:07:02] <martin_lundstrom> jepler: key jogging works fine
[02:07:06] <martin_lundstrom> :)
[02:07:44] <Ziegler> ahh difference being gerber is RS274x
[02:07:50] <Ziegler> g-code is RS274d
[02:08:14] <jepler> martin_lundstrom: OK -- thanks for your report. that one must have slipped through my testing the last time I changed that stuff
[02:08:25] <jepler> and I never click those buttons except when I am testing changes..
[02:09:25] <Twingy> jepler, who worked on the calibration interface in axis last?
[02:10:50] <jepler> Twingy: I'm not sure -- probably petev
[02:10:56] <martin_lundstrom> I have a problem with...
[02:10:59] <martin_lundstrom> [19482.423888] 116779: ERROR: Unexpected realtime delay: check dmesg for details.
[02:10:59] <martin_lundstrom> [19482.423906]
[02:10:59] <martin_lundstrom> [19482.423907] In recent history there were
[02:10:59] <martin_lundstrom> [19482.423908] 198946, 232978, 163862, 199156, and 197255
[02:10:59] <martin_lundstrom> [19482.423909] elapsed clocks between calls to the motion controller.
[02:11:00] <martin_lundstrom> [19482.423913] This time, there were 206880 which is so anomalously
[02:11:02] <martin_lundstrom> [19482.423914] large that it probably signifies a problem with your
[02:11:04] <martin_lundstrom> [19482.423915] realtime configuration. For the rest of this run of
[02:11:06] <martin_lundstrom> [19482.423917] EMC, this message will be suppressed.
[02:11:12] <Twingy> jepler, ever used turbocnc axis configuration interface?
[02:11:12] <martin_lundstrom> what can I do?
[02:11:13] <jepler> Twingy: no
[02:12:08] <Twingy> it's alot easier to use than axis right now, i.e. all of the fields have units, and it says things like, at 18000Hz max velocity and 24000Hz acceleration your max velocity is 18.75ipm achieved in 3.79 seconds
[02:12:18] <martin_lundstrom> and my base period is not very fast
[02:12:23] <jepler> martin_lundstrom: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?TroubleShooting#Unexpected_realtime_delay_check_dmesg_for_details
[02:12:42] <cradek> Hz of acceleration?
[02:12:41] <martin_lundstrom> ill have a look
[02:13:25] <Twingy> cradek, I forget the units for acceleration in TurboCNC, but the big thing is they all have units and each field makes sense
[02:13:52] <Twingy> in EMC they make sense if you read through the manual, in turbocnc you know immediately what they are without the manual
[02:14:51] <Twingy> like when I first ran it there is a little calculator built into the configuration menu that factors in threads per inch, steps per revolution, and microstepping
[02:15:11] <Twingy> and figures out the linear movement equal to one step etc
[02:15:20] <Twingy> 0.00003125" etc
[02:15:34] <Twingy> just lots of little things like that which make it a pleasure to use
[02:15:51] <Twingy> other than that I prefer EMC in just about every other aspect
[02:16:38] <Ziegler> lol
[02:16:39] <cradek> Twingy: we'd happily consider anything you'd like to contribute to fix up those loose ends
[02:16:43] <Twingy> not everyone may agree with me, but friends I have talked to that switched from TurboCNC to EMC preferred the turbocnc interface for configuring axis a little more
[02:16:51] <Twingy> cradek, yea, it's gtk right?
[02:17:12] <cradek> some of the programs are gtk
[02:17:24] <Twingy> cause like I'm knee deep in writing code for GCAM right now
[02:17:40] <Twingy> but when I'm done this gerber stuff I wouldn't mind improving the calibration interface
[02:18:06] <Twingy> well, no calibration really
[02:18:09] <Twingy> should really be called setup
[02:18:16] <Twingy> or configure
[02:18:31] <Twingy> since you are configuring every aspect of your axis (currently done with a text editor and EMC manual)
[02:18:40] <Twingy> that traditional unix way
[02:18:55] <Twingy> but since you're running under ubuntu and ubuntu's moto is linux for humans, it deserves and interface
[02:19:14] <Twingy> s/and/an
[02:19:38] <Twingy> anyways, I'll poke around at it
[02:19:42] <cradek> we'd love to have you work on pulling the configuration together into a gui
[02:19:54] <Twingy> ok, that should be easy
[02:20:10] <Twingy> the hard part is learning what the heck MAX_STEPGEN and MAX_VELOCITY etc "mean"
[02:20:22] <Twingy> I just need to go through the manual
[02:20:22] <cradek> no, the hard part is making a gui that is as flexible as emc is
[02:20:42] <Twingy> that shouldn't be all that hard, can be done with a tabbed notebook for each axis
[02:21:01] <cradek> a program that only knows step/dir and a parport is pretty simple to configure
[02:21:15] <cradek> however, emc is nothing like that program
[02:21:51] <martin_lundstrom> I put up my BASE_PERIOD to 60000 and seemed to do the tric for me
[02:22:07] <Twingy> cradek, have you never used turbocnc?
[02:22:12] <cradek> no
[02:22:16] <Twingy> s/never/ever
[02:22:29] <Ziegler> dont taint him LOL
[02:22:41] <Twingy> well, tainting one to get a better interface out of it is not such a bad idea
[02:23:07] <cradek> to be blunt, I'm not interested in writing a configuration gui - that will have to come from someone else, probably a new volunteer
[02:23:27] <Twingy> that's fine, that's the essence of open source etc...
[02:23:29] <cradek> I'm interested in working on the capabilities of the machine controller itself
[02:23:33] <cradek> definitely
[02:23:49] <cradek> everyone brings something, everyone adds the part they want added
[02:23:59] <Twingy> as long as it doesn't suck ass
[02:24:04] <Jymmm> cradek: Brewery automation module?????
[02:24:07] <cradek> ?
[02:24:29] <Twingy> i.e. I don't let shit code get into gcam
[02:24:31] <Jymmm> cradek: New module for emc
[02:25:30] <cradek> Twingy: it's a little harder to keep a larger team in line, but we seem to do ok
[02:27:38] <Jymmm> Twingy: If you pay for TCNC, you do get the source code for it. Then you coudl always write your own interface.
[02:28:02] <Twingy> I don't want to change the turbocnc interface...
[02:28:08] <Twingy> I want to improve the one in EMC :)
[02:28:13] <Twingy> for axis setup etc
[02:28:17] <Twingy> very minor thing
[02:28:58] <Jymmm> Twingy: Use the source luke
[02:29:10] <Twingy> yep yep, I'll talk more about in on weeks to come
[02:29:21] <Jymmm> talk i scheap
[02:29:23] <Jymmm> cheap too
[02:29:39] <Jymmm> Nike.... JUST DO IT!
[02:29:42] <Jymmm> lol
[02:29:44] <Twingy> exactly, I should stop talking and work on gcam
[02:31:53] <Jymmm> Does everyone know what I mean id I said "pot metal"?
[02:32:06] <Jymmm> Does everyone know what I mean when I say "pot metal"?
[02:32:37] <cradek> pot metal means zinc
[02:32:50] <Jymmm> the yellowish stuff?
[02:33:00] <Jymmm> you might find it on old telco stuff
[02:33:10] <Ziegler> pot metal
[02:33:15] <cradek> it has a certain look, not sure if it's yellow
[02:33:21] <Ziegler> is a zinc aluminum alloy
[02:33:29] <Ziegler> its grey / silver in color
[02:33:43] <Ziegler> zamak and za-12 are common examples
[02:33:48] <Jymmm> No, this is mroe yellowish and slight other colors in it
[02:34:11] <Ziegler> pot metal has slight quantities of zinc in it... but not that you would notice it
[02:34:43] <Ziegler> yellowish.. brass copper alloy? whats it used in again?
[02:34:58] <Jymmm> this sounds right http://en.wikipedia.org/wiki/Pot_metal
[02:35:38] <Jymmm> I've only seen it in cheap mass production stuff
[02:35:54] <Ziegler> hehe... if seen it in high end stuff too
[02:35:58] <Jymmm> I just never knew what it's REALLY called, always knwon it as "pot metal"
[02:36:22] <Ziegler> za-27 is used quite a bit for die-casting
[02:36:36] <Ziegler> has tensile stength of cast iron
[02:36:38] <Jymmm> kinda like this... http://dieselgiant.com/mbodo30.JPG
[02:37:13] <Ziegler> I have an old 9x20 craftman lathe... the gears are made from potmetal
[02:37:27] <Ziegler> that gear looks like brass
[02:37:57] <Jymmm> it's not, just the light hitting it funny
[02:38:07] <Ziegler> ah
[02:38:33] <Ziegler> blow toarch will turn potmetal into a puddle
[02:39:18] <cradek> 'The primary component of pot metal is zinc'
[02:39:53] <Ziegler> za-12 12% aluminum 2% cu, about smidge of mag, and all the rest zinc
[02:40:34] <Ziegler> pot metal can take plating too
[02:40:54] <Ziegler> http://www.enigma-replica.com/StatusPhotos/Zamak%20Cast.JPG
[02:41:56] <Ziegler> http://www.eazall.com/images/ingots-on-feet.jpg << pot metal ingots
[02:42:16] <Ziegler> * Ziegler drools
[02:42:36] <Jymmm> Ok this is the CLOSEST I could find... see the base http://www.techeblog.com/files/portable_rotary_phone.jpg
[02:42:58] <Ziegler> the base?
[02:43:02] <Ziegler> ah
[02:43:47] <Ziegler> hmm
[02:43:54] <Ziegler> trying to think what else it might be.
[02:44:01] <Ziegler> if not potmetal
[02:44:06] <martin_lundstrom> night all
[02:44:09] <Ziegler> seems to be a likley piece for it
[02:45:34] <Jymmm> Well shit... I was trying to find out how this stuff machines =(
[02:45:52] <Jymmm> I have a case I need to modify
[02:45:59] <Jymmm> and it's made of that stuff.
[02:46:13] <Ziegler> well potmetal is a bit gummy
[02:47:32] <Ziegler> better than aluminum tho
[02:48:03] <Ziegler> That put it in the ball park?
[02:48:17] <Jymmm> yeah, thanks
[02:48:35] <Ziegler> hehe... instead of machining it... melt it down and cast a new case >:)
[02:48:48] <Twingy> * Twingy casts aluminum in his back yard
[02:49:08] <Ziegler> * Ziegler cast iron in his back yard :P
[02:49:09] <Jymmm> I think I'm screwed either way... The case is U shaped, will be hard to even drill a hole on the drillrpress
[02:49:32] <Twingy> tap it with an awl before you start drilling
[02:49:51] <Twingy> start with a 1/16" and work up
[02:49:53] <The_Ball> when casting alu, is the process sensitive to the type of alloy mix being used?
[02:50:06] <Twingy> sensitive? not really
[02:50:07] <Ziegler> yes
[02:50:12] <The_Ball> clear as mud
[02:50:15] <Twingy> I usually add a little zinc
[02:50:19] <Jymmm> No, I mean U shaped... 6" tall |_______14" wide___________|
[02:50:30] <Twingy> but not really, you heat it up, limit exposure to air, and pour it
[02:50:30] <Jymmm> |_________
[02:50:42] <Ziegler> casting with an a356 alloy compared to an 60xx serries is much easier to cast
[02:50:51] <Ziegler> has less solidification shrinkage... etc
[02:51:11] <The_Ball> i see
[02:51:12] <Ziegler> that said... I pour 6061 quite a bit
[02:51:14] <Twingy> I haven't noticed much difference between 2xxx and 6xxx castings
[02:51:31] <Twingy> I try to make my stuff as close to 7xxx as possible though
[02:51:34] <Ziegler> get your hands on a casting alloy
[02:51:44] <ds2> Silicon in the alloy supposely helps a lot
[02:51:47] <The_Ball> i tried to machine some "grating" plate alu in my lathe, bloody impossible, don't know what alloy it was though
[02:51:52] <Ziegler> is 7 serries and 6 serries forged alloys?
[02:52:04] <Ziegler> The_Ball: extruded
[02:52:06] <Twingy> 7 is tempered
[02:52:16] <Ziegler> as gummy as it gets before you get to pure aluminum
[02:52:34] <The_Ball> Ziegler, probably yes. the kind they use for toolboxes and the like.
[02:53:03] <Ziegler> yeah... I spent my first 6 months casting using extrusions for my scrap
[02:53:13] <Ziegler> then I start purchaing ingots of a356
[02:53:45] <Ziegler> a356... think cast aluminum pieces "mag" wheels, engine heads and pistons... etc
[02:53:53] <Twingy> I make my own ingots from soda cans
[02:54:02] <Ziegler> hehe been there
[02:54:09] <Twingy> very tricky though
[02:54:19] <Twingy> soda cans generate alot of slag
[02:54:20] <Ziegler> yeah.. lots of loss
[02:54:28] <The_Ball> i have a ninja motorcycle front wheel i can melt, but don't know if the ally is good for it
[02:54:35] <Twingy> cause you don't have the proper CO2/Argon atmosphere
[02:55:10] <CIA-2> 03petev 07TRUNK * 10emc2/src/emc/ini/emcIniFile.hh: -Added default value for num parameter to several FindXXX() methods.
[02:55:11] <CIA-2> 03petev 07TRUNK * 10emc2/src/libnml/inifile/inifile.hh: -Added default value for num parameter to several FindXXX() methods.
[02:55:15] <Ziegler> Twingy: for the price of a proper argon setup... you might as well purchase certified ingots
[02:55:27] <Twingy> precicely
[02:55:31] <Twingy> that's why I tend not to melt the cans
[02:55:49] <Twingy> only once in a blue moon I'll melt down 100 cans and pour brick
[02:55:58] <Ziegler> wow.. hard core man
[02:56:11] <Twingy> then I'll cut the brick up on my band saw
[02:56:15] <Twingy> and alloy it with 6061
[02:56:14] <Ziegler> what sort of cast do you do
[02:56:20] <Twingy> sand cast
[02:56:33] <Ziegler> for what purpose though?
[02:56:38] <Twingy> you know that pink fomular[tm] stuff from home depot? I use that too
[02:56:50] <Twingy> I'm working on an aluminum rocket motor
[02:56:54] <Ziegler> lost foam then to ya?
[02:56:56] <Twingy> mostly aluminum
[02:57:02] <Ziegler> interesting
[02:57:17] <Ziegler> I cast prototype gun parts for a gun smith
[02:57:22] <Twingy> which is why I bought my cnc mill and wrote gcam
[02:57:24] <Ziegler> a bit of art now and then
[02:57:46] <Ziegler> oh... and when I am done with my router machine
[02:57:56] <Twingy> everyone and their uncle wants $50 or $100 for their crappy CAM software
[02:57:58] <Ziegler> I will be making patterns to cast a new machine :)
[02:58:19] <Ziegler> hehe BobCAD wanted even more
[02:58:38] <Twingy> yea so I was like forget it, I'll just write something basic
[02:58:49] <Twingy> for my needs
[02:59:00] <Ziegler> ive only written a image-gcode progam
[02:59:10] <Ziegler> and I stopped working on it
[02:59:19] <Ziegler> when I saw one was included with EMC2
[02:59:20] <Ziegler> hehe
[02:59:33] <jepler> I think that one is almost usable by now
[02:59:33] <Twingy> I think gcam should get rolled into the EMC distribution
[02:59:36] <jepler> goodnight
[02:59:41] <Ziegler> the only thing unique about it was that web bassed
[02:59:49] <Jymmm> laters jepler
[02:59:56] <Ziegler> yes Twingy you have mentioned that. ;-P
[03:00:05] <Ziegler> hehe
[03:01:09] <Ziegler> any of you ever stop by backyardmetalcasting.com?
[03:01:40] <Ziegler> decent forum there for metalcasting
[03:02:01] <The_Ball> Twingy, good work with gcam!
[03:02:02] <Twingy> I really need to get my hands on that find silicate powder sand
[03:02:08] <Twingy> *fine
[03:02:11] <Twingy> thx The_Ball
[03:02:26] <Ziegler> Where are you at Twingy?
[03:02:45] <Ziegler> what mesh sand do you have now?
[03:02:50] <Twingy> Ziegler, my basement
[03:03:04] <Twingy> Ziegler, I am just using generic beach sand
[03:03:11] <Jymmm> Twingy: I think he went WHERE on the planet
[03:03:16] <Ziegler> I guess I know of a few location through the us that US Silica supplies to
[03:03:33] <The_Ball> i received a pittman dc servo motor today... didn't do my research properly, it's dc alright, but still brushless
[03:03:39] <Twingy> Jymmm, he could always just whois me... maryland
[03:03:42] <Ziegler> they have a nice foundry gradation that works well
[03:03:56] <Ziegler> for aluminum to iron
[03:04:02] <Jymmm> Twingy: you think so, huh? try whois'ing me =)
[03:04:08] <Ziegler> google US Silica
[03:04:48] <Ziegler> the other option... is to look into petrobond
[03:05:03] <Ziegler> or call a local foundry in maryland and ask them for some
[03:05:10] <Ziegler> fellow metal casters can be very generous
[03:05:43] <Ziegler> should be a number of foundries in your area
[03:11:37] <Ziegler> alright... off to bed
[03:12:05] <Ziegler> let me know how the sand hunt goes Twingy
[03:44:36] <Unit41> Twingy: powder sandblasting ?
[03:44:49] <Twingy> huh?
[03:44:53] <Unit41> whats it fer
[03:45:09] <Twingy> sand casting
[03:45:16] <Unit41> green sand ?
[03:45:32] <Unit41> or better ?
[03:45:46] <Unit41> they use ceramic for engine parts
[03:46:25] <toastydeath> fff
[03:48:41] <toastydeath> you can get a very detailed sand cast with green sand
[03:49:08] <Unit41> beachsand sounds fine enough too
[03:49:12] <Unit41> that white stuff
[03:49:22] <ds2> how close can you get with green sand compared to investment casting?
[03:49:45] <toastydeath> investment casting can support thinner parts, higher dimensional accuracy, and better surface finish
[03:50:09] <toastydeath> it can get in the ballpark now with a very fine sand that is activated instead of green
[03:50:23] <toastydeath> activated by pumping co2 through it/chemically bonded
[03:50:31] <ds2> wonder what process they use to make those "alloy wheels" that seems so common thesedays... even unmachined surfaces look very clean
[03:50:32] <ds2> Hmm
[03:50:43] <toastydeath> they're pressure forged
[03:50:46] <toastydeath> it's a type of casting
[03:50:49] <Unit41> dang
[03:50:53] <toastydeath> casting/forging hybrid
[03:50:56] <ds2> Oh
[03:51:04] <ds2> thought they were investment casted
[03:51:12] <toastydeath> some are
[03:51:23] <toastydeath> it would be unfair to say wheels are made by xyz process
[03:51:27] <ds2> but that is a LOT of wax given how many they must make
[03:51:30] <toastydeath> since there's a bunch of acceptable ways to make a wheel
[03:51:36] <toastydeath> well, you can re-use the wax
[03:51:46] <toastydeath> in large scale investment casting
[03:52:06] <Unit41> foam sounds better than wax
[03:52:10] <Unit41> for home
[03:52:10] <toastydeath> so it's actually competative with die casting
[03:52:25] <ds2> wow
[03:52:31] <toastydeath> the wax is baked out of the mold in an oven
[03:52:35] <toastydeath> and you're left with the ceramic
[03:52:45] <toastydeath> instead of just vaporizing the wax out during the shot
[03:52:57] <ds2> I see.
[03:53:13] <toastydeath> that's only for the high production stuff though
[03:53:23] <toastydeath> because they have injection molds to make the wax positives
[03:53:29] <ds2> ah
[03:53:37] <toastydeath> the melt goes right back into the machines for the next batch
[03:53:42] <ds2> very different from what a jeweler does making rings
[03:53:46] <toastydeath> yes
[03:56:22] <ds2> do they still do that sort of stuff in the US?
[03:56:36] <toastydeath> yes
[03:56:44] <toastydeath> we still do moldmaking
[03:56:53] <ds2> no, the casting part
[03:56:56] <toastydeath> yes
[03:57:00] <ds2> what part of the US?
[03:57:03] <toastydeath> it's less common, but we still do it
[03:57:03] <toastydeath> all over
[03:57:07] <ds2> oh
[03:57:13] <toastydeath> midwest primarily
[03:57:20] <ds2> would love to get a tour seeing it done
[03:57:20] <Unit41> wonder what kind of certification rims need to go through
[03:57:36] <Unit41> for road wortheyness
[03:57:38] <toastydeath> but california does it, and the northeast has always had manufacturing
[03:57:41] <toastydeath> none
[03:58:00] <ds2> california...must be down in LA then
[03:58:00] <toastydeath> you can make rims at home and use them, and they would be okay in most states
[03:58:04] <Unit41> I can get scrap alu by the truckload
[03:58:29] <Unit41> its 3x more money than steel to buy
[03:58:32] <toastydeath> you could most certainly cast your own wheel
[03:58:44] <toastydeath> but aluminum wheels need more expensive casting/forging
[03:58:49] <toastydeath> to be very lightweight
[03:59:04] <toastydeath> just make sure you've got enough bulk to take impacts
[03:59:16] <toastydeath> not using low profile tires helps, but that's another topic
[03:59:34] <toastydeath> re: molding/casting
[03:59:39] <toastydeath> small-scale stuff is pretty much a global market
[04:00:05] <toastydeath> and the united states is still known for very high quality dies and part runs
[04:00:29] <toastydeath> so mass-produced parts that need to meet high standards in manufacturing accuracy
[04:00:32] <toastydeath> are still done here
[04:02:53] <ds2> bah cast... machine it out of billet ;)
[04:03:46] <Unit41> http://www.lowpressurecasting.com/BP3.html
[04:04:26] <ds2> think they did that on american chopper
[04:04:55] <toastydeath> yeah, it works out of billet too
[04:05:02] <toastydeath> but billet wheels have to be thicker and heavier
[04:05:07] <toastydeath> than forged wheels
[04:05:43] <ds2> but billet wheels are like telling people you light cigars with $100 bills ;)
[04:06:18] <toastydeath> you could cast then machine it?
[04:06:35] <toastydeath> that is what i would do, personally
[04:07:49] <ds2> depends on your facilities
[04:22:32] <CIA-2> 03cradek 07TRUNK * 10emc2/docs/html/gcode.html: one typo, one obsolete thing
[04:23:20] <CIA-2> 03cradek 07TRUNK * 10emc2/docs/html/gcode.html: typo
[04:25:27] <toastydeath> indeed it does!
[04:25:38] <toastydeath> but i imagine for the price of a wheel-sized billet
[04:25:46] <toastydeath> i could easily afford the stuff to do small-time casting
[04:26:40] <ds2> i think i will buy my wheels =)
[04:51:07] <toastydeath> probably much cheaper
[04:57:29] <eric_u> finally have an up to date flash install for linux
[04:57:41] <eric_u> I'm probably months behind
[04:59:53] <chr0n1c> hey so i got a new mobo to test for realtime when the ram gets here
[05:00:07] <eric_u> google street view is very cool, hopefully it never goes realtime
[05:00:23] <chr0n1c> it's an amd64 sempron... MSI - K9MM-V
[05:00:44] <chr0n1c> you guys talking about ggole earth?
[05:00:48] <chr0n1c> google earth*
[05:01:01] <chr0n1c> or have you seen it?
[05:01:31] <eric_u> if you go to http://www.google.com/maps?f=q&hl=en&q=436+Judah+St,+San+Francisco,+CA+94109,+USA&sll=37.796133,-122.427376&sspn=0.071486,0.16016&ie=UTF8&ll=37.766678,-122.465694&spn=0.008939,0.027058&z=16&om=0&layer=c&cbll=37.762171,-122.46665&cbp=2,201.779239864865,0.535192208928214,0
[05:01:43] <eric_u> select street view
[05:02:07] <chr0n1c> i can see my house and my old truck in google earth
[05:02:18] <chr0n1c> my old badass truck i shoulda never sold it :(
[05:02:44] <eric_u> that's cool, google never bought decent views of State College
[05:02:59] <chr0n1c> the city i live.. dayton is all fuzzy
[05:02:59] <eric_u> you can tell there is a city there, that's it
[05:03:21] <chr0n1c> then ten minutes down the road you can see people in thier backyard pools and bushes in the yard
[05:03:37] <chr0n1c> someimes almost even blades of grass.. but dayto, ohio is fuzzy
[05:03:40] <chr0n1c> dayton
[05:04:10] <eric_u> Penn State has a nice site, you can tell I didn't fertilize my yard very well, there are stripes of green/brown/green/brown
[05:04:20] <chr0n1c> lol
[05:05:07] <eric_u> that link I gave shows a billboard advertising microsoft's version of google earth
[05:05:21] <chr0n1c> terraserver?
[05:05:25] <chr0n1c> it's been around for years
[05:05:39] <chr0n1c> i don't remember how i found it
[05:05:47] <chr0n1c> terraserver.microsoft.com?
[05:06:12] <eric_u> maps.live.com
[05:06:55] <chr0n1c> the terraserver url is still up
[05:07:04] <chr0n1c> *just checked it
[05:07:56] <chr0n1c> blah my flash player isn't up to ddate, it won't lemme see
[05:08:35] <eric_u> microsoft has better pics of state college anyway
[05:08:48] <chr0n1c> they have better pics of dayton also
[05:09:04] <eric_u> no street view though
[05:09:04] <chr0n1c> except WPAFB is blacked out last time i checked
[05:09:23] <eric_u> that's pretty silly
[05:09:37] <chr0n1c> sort of
[05:09:46] <chr0n1c> but it makes sense in a paranoid world
[05:09:59] <eric_u> I noticed that they no longer black out the roofs in D.C.
[05:10:06] <eric_u> I thought that was really silly
[05:10:47] <chr0n1c> some of those places you can't fly over.. so the google earth's and clones would be the only way to see them
[05:12:12] <eric_u> except they would only be useful to someone with an air force or ICBM fleet, and those people have damn good pictures of WPAFB
[05:12:31] <eric_u> if it was on google maps, maybe you could find your way around
[05:12:45] <eric_u> there are three of every number on base, I've been very lost more than once
[05:13:35] <eric_u> actually, there can be up to 4 of any given building number, areas a, b, c and kittyhawk
[05:14:11] <eric_u> two of those areas run together, don't remember which twol
[05:17:40] <chr0n1c> amd64 sempron, msi-k9mm-v, 1 gig of ram... and a 2 gigabyte hd from 1995. sounds like a winnner!
[05:18:03] <chr0n1c> ok.. i'm putting a 40 gig ide in it...
[05:18:26] <eric_u> why did you get amd?
[05:18:45] <chr0n1c> well, it was cheap!
[05:18:59] <chr0n1c> i have bought amd for the last 4 boxes i built
[05:19:12] <eric_u> I thought about going back to intel this time
[05:19:22] <eric_u> they are also quite cheap
[05:19:42] <chr0n1c> i wanted this mobo, and it came with the cpu in a bundle
[05:20:01] <eric_u> newegg?
[05:20:07] <chr0n1c> i had a case and a power supply already.. and ram is in the mail
[05:20:11] <chr0n1c> compusa
[05:20:33] <eric_u> how cheap was it?
[05:20:33] <chr0n1c> i was there sunday helping a friend pick out a mobo to replace his.. the sata controller went bad in his game box
[05:20:55] <chr0n1c> not cheap enough..
[05:21:09] <eric_u> wbat speed?
[05:21:34] <chr0n1c> 3000+ which i just read was 1.8 ghz?
[05:21:57] <chr0n1c> good enough for me, i'm not tracking space shuttles or anything
[05:22:59] <eric_u> I need ram
[05:23:01] <chr0n1c> this box is getting candy paint (buy a pro car painter here in town i know) and it's going on ebay ;)
[05:24:01] <eric_u> I'm trying to solve some very large problems, keep running out of ram
[05:24:04] <chr0n1c> so i'm not even going to use it except i was gonna do the realtime test on it before it goes out
[05:24:14] <chr0n1c> how much you got?
[05:24:26] <eric_u> learning about sparse matrices, not fun for some reason
[05:24:32] <eric_u> I have 2 gig
[05:24:40] <chr0n1c> i saw 1 gig ddr2 chips for 32 bux
[05:24:41] <eric_u> we just got a computer at work that has 16gig
[05:24:44] <chr0n1c> when i bought mine last night
[05:24:55] <eric_u> that's not bad
[05:24:59] <chr0n1c> yeah at compusa i saw the mobo.. it could hold 16 gig
[05:25:02] <eric_u> how many slots can you have?
[05:25:03] <chr0n1c> i was like oh shit
[05:25:11] <chr0n1c> and my buddy was like .. what's the big deal
[05:25:14] <chr0n1c> :O
[05:25:33] <chr0n1c> the mobo i jsut got has 2 slots
[05:25:36] <eric_u> need 64 bit os to use it
[05:28:06] <eric_u> haven't used a 64 bit os since I left wpafb in '89
[05:28:13] <eric_u> before that actually
[05:34:32] <Jymmm> Well, I replaced the fuse holder with a 15A resetable circuit breaker, now to try and fit EVERYTHING else in the case! LOL
[05:36:16] <chr0n1c> :D has anyone found my http://nofingclue.com website yet? lol!
[05:36:40] <chr0n1c> and nofingclue related stories to add to it, anyone?
[05:36:43] <Jymmm> it's slow
[05:36:50] <chr0n1c> yeah a tad
[05:36:54] <Jymmm> a lot
[05:36:59] <chr0n1c> i need to turn of statistics
[05:37:23] <jmkasunich> any O-word experts here?
[05:37:41] <chr0n1c> i can do the O-face ;)
[05:37:41] <Unit41> only one way to find out... WOOT OFF
[05:37:46] <Jymmm> Only the obvious... Oh Fsck!
[05:37:50] <Unit41> woot
[05:38:03] <jmkasunich> you guys are very helpfull
[05:38:33] <Jymmm> jmkasunich: Come on... i'm still learning what G0 and G1 do
[05:38:43] <jmkasunich> its 1:30 in the morning, and the g-code reference cards go to the printer at sunrise
[05:38:56] <chr0n1c> G0 is easy it even spells what it does
[05:38:58] <chr0n1c> it go's
[05:39:05] <jmkasunich> I'd like to have some info about O words (subroutines, branching, etc) on them
[05:39:23] <chr0n1c> one thing i have never attempted was subroutines
[05:39:38] <Jymmm> jmkasunich: I foudn this... http://www.imsrv.com/deskcnc/DeskCNC%20G-code.pdf
[05:39:42] <jmkasunich> I think I'm gonna have this for operators: http://jmkasunich.dyndns.org/pics/operators.txt
[05:39:51] <Unit41> brb gotta release the Eanis
[05:39:58] <jmkasunich> so far got this for owords: http://jmkasunich.dyndns.org/pics/owords.txt
[05:40:06] <jmkasunich> I could add about 8-10 more lines
[05:41:19] <Unit41> it only takes 500 grams of fantastic plastic to make a complete mill
[05:41:29] <Unit41> minus the bed frame
[05:41:30] <Jymmm> jmkasunich: do you have an ieee account?
[05:41:36] <Unit41> and other parts
[05:41:58] <jmkasunich> Jymmm: no
[05:42:10] <chr0n1c> unit41: what about welding and drill/tap/screws?
[05:42:13] <Jymmm> jmkasunich: Damn... http://ieeexplore.ieee.org/Xplore/login.jsp?url=/iel5/9397/29813/01359546.pdf
[05:42:22] <jmkasunich> Unit41: even less - about 10 grams of plastic will do
[05:42:37] <Unit41> deffinatly need it though
[05:42:49] <chr0n1c> i built my mill without, lol
[05:42:51] <jmkasunich> as long as its 2-1/8 x 3-3/8 inches
[05:42:56] <chr0n1c> not even any jb weld
[05:43:14] <Unit41> I never welded one spot yet
[05:43:17] <jmkasunich> and about 0.040 thick, with embossed numbers on top
[05:43:24] <Unit41> just got it back
[05:44:17] <Unit41> if you use good l brackets you can get away with no welds
[05:44:18] <Unit41> L
[05:44:20] <Unit41> bracke
[05:44:52] <Unit41> im going to post diagrams on cnczone soon
[05:46:32] <Jymmm> jmkasunich: Ironically, guess where all the refernces are pointed to... http://www.google.com/search?q=o+word+gcode
[05:47:04] <jmkasunich> of course
[05:47:57] <Unit41> nifty
[05:48:29] <Jymmm> I never realized that '#n' was a variable.
[05:56:21] <robin_sz> meep?
[05:57:31] <Jymmm> you up early robin_
[06:15:10] <CIA-2> 03jmelson 07v2_1_branch * 10emc2/configs/ppmc/ppmc.ini: update ini file
[06:16:04] <CIA-2> 03jmelson 07v2_1_branch * 10emc2/configs/ppmc/ppmc_io.hal: update config files
[06:16:54] <CIA-2> 03jmelson 07v2_1_branch * 10emc2/configs/ppmc/ (ppmc_load.hal ppmc_motion.hal ppmc_servo.hal): update config files
[06:39:11] <Unit41> disambiguation
[06:50:25] <jmkasunich> http://jmkasunich.dyndns.org/pics/gcode-ref-back-single.pdf
[07:19:02] <Jymmm> jmkasunich: Maybe many know already, but I like to see examples personally. It helps to make sure one understands the syntax. Like passing/receiving variables as example.
[07:19:33] <jmkasunich> I put a few examples in there
[07:19:46] <jmkasunich> this isn't to replace the manual, you can only do so much with one sheet of paper
[07:19:58] <jmkasunich> more like a quick reference / memory jogger
[07:20:10] <jmkasunich> http://jmkasunich.dyndns.org/pics/gcode-ref-front-single.pdf
[07:20:25] <jmkasunich> at this point, what you see is what you're gonna get, it goes to the printer in the morning
[07:21:45] <Jymmm> jmkasunich: I suspect this is for fest?
[07:21:51] <jmkasunich> yeah
[07:22:04] <jmkasunich> we're gonna have a bunch of them printed up and laminated, as handouts
[07:23:32] <Jymmm> jmkasunich: I guess I would have handed out CD's instead. Adding the Refernce charts as files on it - Especially since everyone will have a computer.
[07:23:53] <jmkasunich> we'll have CDs too, burn on site, or maybe some pre-burnt
[07:24:41] <Jymmm> jmkasunich: you guys have a project for this year?
[07:24:52] <jmkasunich> no single thing
[07:24:57] <jmkasunich> lots of stuff
[07:25:01] <Jymmm> jmkasunich: heh
[07:25:53] <jmkasunich> regarding the cards - everybody might have a computer, not everybody has a nice printer, and fewer yet have a laminator to keep greasy fingerprints off of it
[07:26:05] <jmkasunich> I think people will like them
[07:26:24] <Jymmm> Yeah, I like my laminated size chart
[07:26:38] <Jymmm> letter sized?
[07:26:50] <jmkasunich> half of an 8x11 sheet
[07:26:58] <Jymmm> k
[07:27:03] <jmkasunich> http://jmkasunich.dyndns.org/pics/gcode-ref-front-double.pdf
[07:27:10] <Jymmm> 404
[07:27:14] <jmkasunich> thats what we'll print for the back
[07:27:19] <jmkasunich> oops
[07:27:32] <jmkasunich> http://jmkasunich.dyndns.org/pics/gcode-ref-back-double.pdf
[07:27:35] <jmkasunich> try that
[07:27:49] <jmkasunich> I don't have a two up pdf for the front
[07:28:11] <Jymmm> so comp is a big thing for most folks?
[07:28:25] <jmkasunich> no bigger than lots of other things
[07:28:26] <The_Ball_> jmkasunich, handy, thanks
[07:28:32] <jmkasunich> but its one where a pic is worth 1000 words
[07:28:47] <jmkasunich> The_Ball_: you're welcome
[07:28:57] <Jymmm> Yeah, the illustration is good.
[07:29:10] <The_Ball_> * The_Ball_ is just starting with emc/cnc
[07:29:13] <Jymmm> it took me a bit to see where it was on/off
[07:30:21] <Jymmm> emc needs a logo =)
[07:32:11] <jmkasunich> I need sleep
[07:32:14] <jmkasunich> 3:30 am
[07:32:15] <Jymmm> Do geckos need airflow?
[07:32:44] <jmkasunich> they need heatsinking, above a certain current
[07:32:52] <The_Ball_> mine got really hot, but when i mounted them on a heat sink they are cool
[07:32:53] <Jymmm> I'm intending to use a1/8" al plate to mount them on
[07:33:05] <jmkasunich> read the fine manual
[07:33:11] <Jymmm> a EAL heatsink, or a metal plate to disapate heat
[07:33:18] <jmkasunich> its on the gecko site, you don't have to buy the drive first
[07:33:25] <Jymmm> jmkasunich: what page? 1, 2, or 3?
[07:33:25] <jmkasunich> I used a sink, but I over-do everything
[07:33:33] <jmkasunich> read it all
[07:33:35] <jmkasunich> goodnight
[07:33:40] <Jymmm> night
[07:34:19] <The_Ball_> i also over did it, two drives on a 10x4" long and 2" deep fins
[07:34:33] <Jymmm> btw I have read it a few times. But I wasn't sure of a metal plate was enough, or the heatsink needed fins as well.
[07:35:20] <Jymmm> The_Ball where did you get the heatsink from?
[07:35:37] <The_Ball_> jaycar
[07:35:52] <jmkasunich> depends on how much load you'll be putting on the drive, how much air movement you have, what your supply voltage is, and how big the plate/sink is
[07:35:52] <The_Ball_> it's a .au chain
[07:35:52] <Jymmm> j who?
[07:36:02] <The_Ball_> www.jaycar.com.au
[07:36:08] <Jymmm> oh
[07:40:12] <The_Ball_> my "to inches" doesn't work properly (i have the flu) but that's the one http://www.jaycar.com.au/productView.asp?ID=HH8546&CATID=&keywords=heatsink&SPECIAL=&form=KEYWORD&ProdCodeOnly=&Keyword1=&Keyword2=&pageNumber=&priceMin=&priceMax=&SUBCATID=
[07:43:19] <Jymmm> I just found this too... http://www.timgoldstein.com/cnc/heatsink.asp
[11:29:07] <kwajstabo> hello
[11:29:40] <kwajstabo> why free frograms can i use for generating g code?
[11:29:48] <kwajstabo> what free frograms can i use for generating g code?
[11:30:18] <archivist> vi
[11:30:42] <cradek> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Cam
[11:31:03] <cradek> but archivist is also right - a lot of parts are practical to program by hand
[11:31:44] <cradek> bbl
[12:22:03] <kwajstabo> what would be the best free tool for g code generating for 4 axis hot wire machine for styrofoam cutting (similar to EDM machine)? Maybe edm360?
[12:23:06] <kwajstabo> maybe apt360?
[12:46:09] <skunkworks> I Know this was recently descussed on the emc user email list.
[12:48:17] <skunkworks> I think the main problem is normally machines with parallel axiss program the second set in relative coordinates to the first.
[12:49:15] <skunkworks> emc does not do this yet.
[12:51:42] <jepler> it does not seem terribly difficult to write a kinematics module to do this
[12:52:05] <jepler> (this is not an offer to write it, though)
[12:52:24] <skunkworks> :)
[12:52:48] <maddash> :O what happens when I attempt to access a hal_bit_t from component A that was created in component B?
[12:53:32] <maddash> answer: the call in comp A doesn't return!
[12:53:42] <skunkworks> jepler: I would think that would be great for a wire edm.
[12:54:45] <jepler> maddash: in the HAL model, each component only reads or writes its own pins. Connections between pins are established through signals which are generally created and connected by halcmd.
[12:57:03] <maddash> jepler: yeah, it took me a while to realize the "only its own pins" part
[13:02:00] <maddash> jepler: but if the hal_bit_t is part of a shmem segment, shouldn't it be shared as well?
[13:04:00] <jepler> maddash: even if it worked, it would still be incorrect use of HAL. Given that fact, I'm not too interested in exactly why it turns out to make your program segfaul or whatever
[13:04:08] <jepler> +t
[13:07:40] <maddash> jepler: heh.
[13:11:11] <maddash> jepler: so hal_bit_t should never reside in shmem?
[13:22:40] <maddash> odd, i'm getting "HAL: ERROR: data_ptr_addr not in shared memory," almost as if HAL wanted the hal_bit_t to reside in shmem
[13:24:02] <maddash> aha,, got it.
[14:13:27] <lojz> hello
[14:14:21] <lojz> is there some kind of log system that records all the posts in this irc channel?
[14:14:23] <jepler> welcome lojz, Robbo
[14:14:27] <jepler> logger_emc: bookmark
[14:14:27] <jepler> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2007-06-05.txt
[14:14:49] <jepler> lojz: yes, most of the time, logger_emc does that
[14:37:41] <lojz> i have one question about skunkworks's post about wire edm machine: "I think the main problem is normally machines with parallel axiss program the second set in relative coordinates to the first". How should i write the kinematics module to do this? Wouldn't be easier to make a script which would edit/change g code for parallel axes coordinates?
[14:38:05] <jepler> that's certainly another way to do it
[14:44:44] <lojz> is there a free software capable of g code generating for 4 axis wire machine? I checked the wiki's list, but havent found one...except the APT seems to be universal. Has anyone try to use it for this?
[14:50:07] <jepler> I don't know of any, but I have never used such a machine either
[14:54:50] <lojz> * lojz slaps logger_emc around a bit with a large trout
[15:37:53] <martin_lundstrom> Hello everyone
[15:38:13] <jepler> hi martin_lundstrom
[15:39:48] <martin_lundstrom> did anyone see Dallur lately?
[15:40:29] <martin_lundstrom> I have some ffedback for his THC config
[15:42:13] <jepler> martin_lundstrom: on july 1, he said "<Dallur> l8r, see you guys in 2 weeks" -- http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2007-06-01.txt
[15:48:04] <SWPadnos> July is in 4 weeks
[15:48:20] <jepler> oh june 1, he said ...
[15:48:24] <jepler> man you'd think I'd know one month from another
[15:48:47] <SWPadnos> I can usually tell them apart in real life, but the numbers always get screwed up
[15:48:51] <archivist> could have been last july
[15:49:29] <skunkworks> Hi archivist. What do you think of these goof balls?
[15:49:54] <skunkworks> Well - you don't have to answer that.
[15:50:01] <archivist> heh /me reserves judgement
[15:50:08] <archivist> teehee
[15:50:47] <archivist> mostly normal when that jymmmmmm is not here
[15:51:18] <skunkworks> he is not here - we can talk about him ;)
[15:51:41] <archivist> i did check first
[15:52:14] <archivist> a little birdy told me there is another clockwork person in here
[15:52:41] <SWPadnos> do yo umean the time guy? :)
[15:53:06] <archivist> possibly
[15:53:32] <skunkworks> cradek is the timeguy
[15:53:40] <cradek> that's me
[15:54:27] <archivist> clock is maaa day job
[15:54:35] <cradek> I don't do it much anymore, but I did clean an Accutron last night
[15:54:45] <cradek> I saw your website - some very neat clock stuff there
[15:55:07] <archivist> Im playing with a crap german clock self correcting mech atm
[15:55:29] <cradek> what is self correcting?
[15:56:05] <archivist> it gets the striking in time with the hour hand
[15:56:14] <skunkworks> * skunkworks is a little too shakey to work on anything smaller than a grandfather clock.
[15:57:03] <cradek> I have not done much clock stuff - my tools are all watch size
[15:57:17] <archivist> skunkworks, you should try hand turning under a stereo zoom microscope
[15:58:25] <archivist> we dont do much watch size but do go up to tower clock size
[15:58:43] <skunkworks> they where playing a HD video of a clock maker/repairer at the local best buy. It was pretty cool.
[15:58:55] <cradek> did you see the E Howard tower clock in this month's Bulletin? It's really amazing
[15:59:10] <cradek> Howard's first tower clock (1847 iirc) in perfect condition
[15:59:24] <archivist> nah I dont get it
[15:59:39] <cradek> well this month it's "clock porn" haha
[15:59:49] <archivist> hehe
[16:00:59] <cradek> were you the one who had a source for very fine wire?
[16:01:05] <archivist> yup
[16:01:29] <cradek> I'm looking for some #54 or #55 and it's VERY hard to get, do you do anything near that?
[16:01:44] <archivist> the size that solder can disolve
[16:01:56] <cradek> the size that a sharp look can dissolve
[16:02:15] <archivist> will have to measure/check what we have lying around
[16:02:29] <cradek> are you in the US?
[16:02:34] <archivist> UK
[16:02:58] <cradek> cool, maybe you have some different sources there
[16:03:04] <skunkworks> he has a bbc remote camera truck - I think that is cool :)
[16:03:17] <cradek> I'd really like to get some of that, if you have some, please let me know
[16:04:20] <cradek> I think it's way too fine to respool, and that makes it hard to buy less than a full spool (significant part of a pound)
[16:04:55] <skunkworks> cradek: what is that used for?
[16:04:56] <skunkworks> in a clock
[16:05:16] <cradek> I want it to try to rewind coils in Accutrons
[16:05:42] <cradek> a lot of them are failing - they'll all be gone in 50 years at this rate
[16:05:53] <skunkworks> The coil opens up?
[16:05:56] <cradek> yep
[16:06:05] <archivist> THE coil person is no longer here but I see no real problem re spooling a bit
[16:06:30] <cradek> archivist: that's great
[16:06:59] <archivist> hmm just getting a req for a 2mm od pinion
[16:07:17] <archivist> I get the small jobs
[16:07:48] <cradek> that's not clock size...
[16:08:12] <archivist> most time is machine setup for that sort , its for a musical box
[16:08:26] <cradek> ah
[16:08:35] <archivist> other clockmakers use us for the hard stuff
[16:09:25] <archivist> I need a cnc gear miller with auto centering to make it easy
[16:10:13] <cradek> how do you polish a pinion that size? I've seen a pear (?) wood wheel setup but for bigger stuff
[16:10:42] <archivist> I tend to cheat a bit, dremel
[16:11:02] <archivist> and let the miller cycle a few extra times
[16:11:11] <cradek> well I guess it's the results that count
[16:11:24] <cradek> I would have never thought to try that :-)
[16:11:47] <archivist> we have got a wood polishing machine
[16:12:09] <archivist> but a pain to set up for one offs
[16:12:17] <cradek> any photos of these kinds of jobs on your site?
[16:13:50] <archivist> I must put some links up but www.archivist.info/clock/ for a start
[16:15:33] <archivist> www.archivist.info/cnc/ for machinery
[16:17:09] <archivist> top right hand for small pinions
[16:17:30] <archivist> circa 1956 machine
[16:17:45] <cradek> ah - brand new!
[16:18:12] <archivist> bottom right 30's
[16:18:17] <skunkworks> Now we have to ask why your not running emc for your little mill....
[16:18:25] <cradek> wow, 3 fusees on that crazy clock
[16:18:47] <cradek> is it english or did others use fusees in clocks too?
[16:19:05] <cradek> I guess english stuff isn't so uncommon for you :-)
[16:19:06] <archivist> skunkworks, it needs doing the software in that cnc is crap
[16:20:02] <cradek> easy retrofit I bet
[16:20:11] <cradek> it's just one axis of motion plus indexing right?
[16:20:42] <archivist> x y and z
[16:20:59] <archivist> indexing was extra
[16:21:14] <archivist> thats do it yourself
[16:22:59] <archivist> and it has no idea of a zero position
[16:23:30] <archivist> ie some random place in space after switch on
[16:24:05] <archivist> and the comms locks up if the pc is too fast
[16:24:05] <cradek> heh many people are liking to have home switches
[16:24:25] <skunkworks> they come in handy
[16:25:35] <skunkworks> archivist: what kind of interface is it to the computer? do you know?
[16:25:41] <archivist> rs232
[16:25:47] <skunkworks> eww
[16:26:11] <archivist> crap in large doses
[16:26:33] <cradek> that just means you have to replace a few more parts!
[16:26:44] <cradek> is it steppers?
[16:26:48] <skunkworks> are they steppers?
[16:26:50] <skunkworks> ;)
[16:27:22] <archivist> hehe from the stepper drivers (ST Bricks) back
[16:28:06] <archivist> but I dont own it yet!
[16:28:14] <skunkworks> ah
[16:28:41] <archivist> only me used it for the last few years
[16:30:10] <archivist> The owner is 80 odd just got to wait a bit!!
[16:30:49] <archivist> but his dad lasted well into his 90's
[16:32:34] <skunkworks> st bricks?
[16:33:22] <archivist> what was SGS Thomson semiconductors
[16:34:39] <skunkworks> you would think you would be able to get into there and maybe be able to use step and direction.. or maybe phase drive.
[16:36:21] <archivist> step abd dir
[16:44:42] <ytf> hi. I am having trouble downloading the LiveCD of EMC2. Any idea why a download to a Windows machine would be 300 bytes bigger (and have a different md5) than what other people report?
[16:45:02] <The_Ball_> ytf, download with what?
[16:45:10] <ytf> I've tried Firefox and Opera just in case it was due to a browser bug.
[16:45:16] <ytf> same result each time.
[16:45:28] <ytf> also tried EU mirror
[16:45:37] <SWPadnos> virus on your machine
[16:45:42] <ytf> heh
[16:45:47] <The_Ball_> do you have proxy configured?
[16:45:49] <SWPadnos> I'm serious
[16:46:00] <SWPadnos> I just responded on the users list
[16:46:10] <ytf> OK, not inconceivable, SWP.
[16:46:34] <SWPadnos> I have Windows 2000 here, and the file size and md5sum are correct
[16:46:45] <ytf> Ball: hmm. not that I am aware of. I'm running CA antivirus
[16:47:02] <ytf> thanks for the thoughts
[16:47:26] <SWPadnos> did you say you have a working ubuntu 7.04 CD?
[16:47:32] <ytf> yes
[16:47:48] <ytf> two flavors: the LiveCD and the special install (Xubuntu)
[16:47:53] <SWPadnos> boot from that, download the file to a USB stick, then reboot into Windows to burn it
[16:48:11] <ytf> SWP: interesting approach.
[16:48:13] <SWPadnos> (unless you have 2 CD drives, and can burn with the liveCD still in a drive)
[16:48:19] <ytf> *nod*
[16:49:10] <ytf> OK, I'll give that a try.
[16:49:13] <SWPadnos> also, while you're at it, download one of those linux virus scanner/system repair CDs (like Ultimate Boot CD) :)
[16:50:22] <ytf> hmm. Well, I'm running Windows on this machine I'm doing the downloads on. i thought you were mentioning Windows-as-virus at first, thus the "heh".
[16:50:56] <SWPadnos> yep - figured ;)
[16:52:55] <ytf> correction. I'm running CA _firewall_, not AV. I wonder if the firewall is goosing up the download somehow.
[16:53:05] <ytf> proxy weirdness
[16:53:38] <SWPadnos> it shouldn't be, if that were the case, you'd probably get extra crap on web pages and stuff
[16:55:03] <ytf> my AV stuff is Grisoft AVGfree. I'll do a scan, maybe I'll be lucky enough to see something if I haven't been rootkitted :\ :)
[16:55:54] <ytf> OK, I'll storm the castle now. Thanks, folks!
[16:55:57] <ytf> bye
[16:55:59] <SWPadnos> it's always best to boot from a CD for virus scanning - it prevents some attacks (some viruses know what the AV programs are)
[16:56:03] <SWPadnos> have fun storming the castle
[17:20:05] <martin_lundstrom> jepler: thanks for the hint with Dallur
[17:42:05] <maddash> what's the difference betwen "AXIS_* -> MAX_VELOCITY" and "TRAJ -> MAX_VELOCITY"? the user manual states that the former is the max vel. "while coordinated motion is in effect." what does this mean?
[17:42:32] <maddash> "coordinated motion" == auto mode?
[17:43:14] <SWPadnos> coordinated motion means that emc is moving more than one axis at a time fora single move
[17:43:33] <SWPadnos> like G0 x1 y1 z1 is coordinated, whether it's issued from MDI or auto mode
[17:44:32] <maddash> ok, so the entry under [traj] only holds true if it is < the entry under [axis] and/or emc is in manual mode?
[17:44:44] <SWPadnos> no
[17:45:09] <SWPadnos> the [TRAJ] limit is the max the trajectory planner will attempt, and that operates on the vector sum of axis motions
[17:45:40] <SWPadnos> if any [AXIS] limit is lower than [TRAJ] fora particular move, the entire move (all axes) will be slowed down to accomodate the lower limit
[17:46:20] <SWPadnos> as an example, if you have X/Y/Z axis limits at 1, then the machine can actually move at sqrt(3) velocity, for a diagoonal move
[17:47:12] <maddash> sqrt(3)?
[17:47:29] <maddash> nvm.
[17:47:43] <SWPadnos> sqrt (x^2 + y^2 + z^2) (velocity limits)
[17:48:30] <SWPadnos> in free mode (when homing or jogging, I think) the TRAJ limit isn't used, since each axis basically has an independent planner in that mode
[17:49:04] <maddash> what if the axis limit == traj limit? then the maximum velocity (diagonal move or not) is the traj limit itself?
[17:49:24] <maddash> also, what if the seek_velocity > traj limit?
[17:49:40] <SWPadnos> what's seek_velocity?
[17:50:07] <SWPadnos> actually, it doesn't matter what it is. here's how it works (in coordinated mode):
[17:50:10] <maddash> sorry, I meant HOME-SEARCH-VEL
[17:50:21] <SWPadnos> ok, that's free mode, afaik
[17:50:53] <SWPadnos> basically, wither of the motion planners will never exceed the limit for that axis
[17:51:01] <SWPadnos> s/wither/either/
[17:51:19] <SWPadnos> the coordinated planner will additionally not exceed the TRAJ limits
[17:51:35] <SWPadnos> (even if it would be possible from the AXIS* limits alone)
[17:52:33] <SWPadnos> gotta run. hopefully that clears it up a little
[17:54:30] <maddash> thanks.
[18:14:04] <archivist> hmm cradek finest wire ive found is on the fat side at .073 mm including insulation/laquer
[18:16:11] <cradek> thanks for looking, this would be more like .015mm
[18:16:43] <archivist> dunno if we have smaller yet
[18:17:23] <cradek> wow! 330km per pound!
[18:18:03] <anonimasu> hehe
[18:18:07] <cradek> this is why I don't want to buy a 1/3 to 1/2 pound spool :-)
[18:19:23] <archivist> I wonder how many miles I have on the bench atm 44 swg and about 4-5 lbs
[18:20:15] <cradek> ~ 60
[18:20:20] <archivist> bah more than 6 the scales error out
[18:20:20] <ds2> are you making motors nows? =)
[18:20:47] <anonimasu> lol
[18:20:49] <anonimasu> lol
[18:20:56] <anonimasu> err sorry im lagged
[18:20:59] <anonimasu> cellphone lagged.
[18:28:10] <skunkworks> cradek: no solenoids or such with that small of wire?
[18:29:12] <cradek> skunkworks: see the coils? http://members.iinet.com.au/~fotoplot/acctech214.htm
[18:29:57] <skunkworks> cradek: oh. that is a wrist watch?
[18:29:59] <cradek> if I can find the wire, *maybe* I can wind them under control of HAL
[18:30:10] <cradek> yes
[18:30:34] <cradek> the wire has to be wound with about 1 gram of tension
[18:30:46] <skunkworks> wow. (I don't think we have anything that small :))
[18:30:58] <cradek> nobody does, it's extremely hard to find
[18:31:10] <cradek> #52 is not too hard to find, but it would be too big
[18:31:56] <archivist> reduce the turns so it still fits
[18:32:13] <archivist> may effect battery life a bit
[18:32:48] <cradek> archivist: I think that would affect the circuit's operation and the tuning fork amplitude too
[18:32:57] <cradek> the amplitude is very critical
[18:33:37] <archivist> Im assuming the resonance of the fork should control that to an extent
[18:34:06] <cradek> you mean control the amplitude? I don't think so
[18:34:32] <cradek> you can easily see the difference between mercury (1.35) and silver oxide (1.55) batteries
[18:35:23] <archivist> 1.55 being better?
[18:35:43] <cradek> actually no, the higher amplitude causes problems in some watches
[18:35:52] <cradek> they were of course designed for the mercury batteries
[18:36:36] <archivist> ok so thicker wire if it detunes it may be cured by a resistor change or battery change
[18:37:29] <jepler> amplitude, not frequency
[18:37:32] <archivist> so try it if you have a dead one
[18:38:37] <cradek> right, the frequency is the fork resonance - the amplitude, current draw, battery life etc are the things it would mess up
[18:39:37] <cradek> I'm sure they didn't use #55 wire "for fun" considering the HUGE price increases once you get below about #51
[18:41:13] <skunkworks> I am trying to think what I have fixed that had such small wire that I just stuck it into the pool of solder.
[18:41:19] <archivist> probably to increase battery life mainly, I wonder if the modern batteries have more capacity than the ones they were designing for
[18:42:10] <cradek> not really, they guaranteed a year when new and we get no more than 14 months with modern batteries
[18:43:27] <cradek> so, they're probably about the same
[18:45:18] <jepler> clearly you're going to have to manufacture the wire yourself
[18:45:24] <jepler> how hard can it be?
[18:45:29] <cradek> haha
[18:45:51] <cradek> easy - first, drill a .001 inch hole, then, pull a mile of wire through it
[18:46:01] <cradek> then, simply insulate it
[18:46:11] <jepler> oh that's easy -- just use hair spray
[18:46:28] <cradek> http://members.iinet.com.au/~fotoplot/acctech221.htm
[18:46:49] <cradek> this model (mostly used in ladies watches) has smaller coils
[18:47:11] <cradek> the whole gear train is in the lower right corner there - it's worm gears - one is in sideways
[18:47:33] <cradek> they're *very* challenging to reassemble
[18:47:47] <ds2> drilling 0.001 holes is easy ;)
[18:47:49] <ds2> * ds2 ducks
[18:48:57] <jepler> cradek: can't you just turn thicker wire down to .001 on your sherline? none of this pushing wire through holes, that's asinine.
[18:49:12] <cradek> sure, with a follower rest, why didn't I think of it?
[18:49:41] <skunkworks> and a barstock feeder.
[18:49:53] <jepler> skunkworks is on to somethin
[18:49:52] <jepler> g
[18:49:58] <cradek> yep
[18:50:16] <skunkworks> I see a project for the cnc workshop
[18:50:40] <ds2> will this use a 3J chuck ?
[18:50:54] <cradek> making the coil winder (controlled servos) would be fun - but without the wire available, why bother
[18:52:03] <cradek> sorry to bore you all with my pipe dream project :-)
[18:52:08] <skunkworks> cradek: is there anyone out there that rebuilds these watches now?
[18:52:50] <cradek> there are about four people that I know of, I'm one
[18:53:12] <skunkworks> have you rewound the coils by hand?
[18:53:13] <cradek> the guy who belongs to that website is in australia and is good
[18:53:27] <cradek> no, nobody fixes the coils as far as I know.
[18:53:31] <skunkworks> ah - ok
[18:53:36] <cradek> it would be totally impossible to do by hand
[18:54:08] <skunkworks> aww - where is your sense of adventure?
[18:54:09] <cradek> we all mix and match coil sets to sacrifice some watches for others - eventually they will all be dead
[18:54:48] <cradek> ... which is why it would be nice to be able to rewind them!
[18:55:16] <archivist> hmm just been for a nother wire dig and found 46 swg
[18:55:18] <cradek> (but it's probably not practical to think I can do it in my basement with simple homemade equipment)
[18:56:02] <cradek> archivist: this wire is probably $6000/lb, I'd be awfully surprised if you had any so small without knowing for sure where it is :-)
[18:56:30] <archivist> it came as a job lot from a company that shut down
[18:56:42] <cradek> ahhh
[18:56:52] <cradek> then forget that I told you the price!
[18:57:05] <archivist> * archivist searches harder
[18:57:16] <cradek> hahaha
[18:57:34] <archivist> * archivist starts lifting floor boards
[18:58:00] <jepler> $6000/lb? That's nothing compared to the street price of drugs in drug busts.
[18:58:06] <skunkworks> This is a bunch of bull -> <cradek> (but it's probably not practical to think I can do it in my basement with simple homemade equipment)
[18:58:37] <archivist> when I see our ex worker next I shall ask him
[18:58:40] <cradek> skunkworks: you can be sure I'll try if I can find the wire
[18:58:48] <cradek> (for a price I'm willing to pay)
[18:59:04] <cradek> archivist: thanks a lot for trying
[18:59:13] <skunkworks> Hell - I refine my own plutonium in my basement.
[18:59:19] <skunkworks> oops
[18:59:46] <jepler> yeah cradek -- after you re-wind accutrons, you need to work on a nyoo-kyoo-lurr wristwatch
[19:00:21] <cradek> isn't mercury bad enough?
[19:00:26] <cradek> or radium
[19:00:49] <skunkworks> I ment I refine my own radon in my basement.
[19:01:03] <cradek> that I can believe...
[19:01:20] <archivist> we have the odd amount of meercury around
[19:01:31] <cradek> yeah I think we all do
[19:01:44] <cradek> err, I mean skunkworks does
[19:01:49] <cradek> * cradek glances around
[19:02:07] <archivist> we made a mercury pendulum last year
[19:02:32] <cradek> cool
[19:02:38] <cradek> I always thought that was a clever hack
[19:02:50] <cradek> also the steel/brass schemes
[19:03:06] <archivist> www.archivist.info/pendulum
[19:03:17] <cradek> but "pour some more in the jar" and "dump some out of the jar" to adjust the compensation is hard to beat
[19:03:46] <archivist> fun was proving it in solidworks
[19:04:13] <cradek> what's the material here?
[19:04:22] <archivist> by the way that shiny metal is cast iron
[19:04:27] <cradek> it's a nice finish
[19:04:57] <cradek> is there a glass jar inside? (isn't the mercury in glass?)
[19:05:04] <cradek> <- not a clock guy
[19:05:06] <archivist> no
[19:06:35] <archivist> I just hope it doesnt leak through the irom pores
[19:06:42] <skunkworks> couldn't you use anything (oil or such) why mercury?
[19:06:52] <skunkworks> just the weight?
[19:07:08] <cradek> the mercury is used for temperature compensation
[19:07:27] <cradek> as the pendulum gets longer (warmer), the mercury expands "up" to compensate
[19:07:41] <skunkworks> wow.
[19:08:10] <skunkworks> my grandpa make wood clocks. Nothing that accurite.
[19:08:14] <skunkworks> :)
[19:08:25] <archivist> the user wanted a steel rod but that did not calculate out for me so invar it has to be
[19:08:37] <cradek> other schemes use metals that expand differently (brass and steel I think) in mechanical configurations that balance out
[19:09:05] <ds2> isn't easier thesedays to build a temp control chamber? ;)
[19:09:16] <archivist> they tend to have stiction problems with the rods
[19:09:19] <cradek> now we'd just use metal that doesn't change size
[19:10:26] <cradek> archivist: invar rod with mercury pendulum is a bit of a clash of technologies isn't it?
[19:10:37] <archivist> true
[19:10:37] <skunkworks> the clock just needs to be in a temp controlled chamber :)
[19:10:49] <cradek> yes, ideally in a vacuum too
[19:11:06] <cradek> ... on granite buried deep into the earth
[19:11:26] <skunkworks> at that point - an atomic clock might be the answer.
[19:11:41] <cradek> but that's science - this is art!
[19:12:01] <archivist> we have an artcle from a bloke who used a concrete block
[19:12:02] <cradek> actually timekeeping is a neat clash of science and art - I think it's why I like it
[19:12:35] <ds2> doesn't the value of G change ever so slightly when you go deep enough?
[19:12:49] <alex_joni> ds2: or up
[19:13:03] <archivist> and with the moon
[19:13:09] <ds2> so buried deep in the earth would make it in accurate ;)
[19:13:23] <alex_joni> ds2: not too deep, or you'll end up on the other side with -g
[19:13:31] <archivist> they become G meters
[19:13:32] <skunkworks> well - you would calibrate it at that depth - of cource
[19:13:33] <cradek> G is constant - m & r change I think
[19:13:44] <ds2> but if you go deep enough, G should be zero ;)
[19:14:00] <cradek> G, the gravitational constant, is ... constant
[19:14:10] <ds2> oops 'g'
[19:14:12] <ds2> =)
[19:14:25] <ds2> the figure that is about 9.8m/s =)
[19:14:34] <ds2> or 32ft/s
[19:14:38] <archivist> G is an approximate constant
[19:15:04] <skunkworks> ouch
[19:15:07] <skunkworks> http://www.cnczone.com/forums/showthread.php?p=305468#post305468
[19:16:20] <skunkworks> that may reqire a "patches are welcome" post
[19:16:26] <kwajpol> I am following this instructions: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?AptProgrammingForEMC and i can't install libf2c, dxflib and pyopengl. I get "couldn't find package" error. Any idea hat could be wrong?
[19:16:30] <cradek> maybe we should send him the source ... oh, wait
[19:17:31] <jepler> kwajpol: my only guess is that those names are not debian package names, and that you can use "apt-cache search", google, and a little bit of guessing to find the actual package names.
[19:17:36] <cradek> skunkworks: he could use 'run from line', or put each tool's gcode in its own file
[19:18:00] <cradek> skunkworks: but if he just wants to whine, I suggest you don't bother trying to help
[19:18:06] <jepler> (if there are debian packages of all those things in the first place)
[19:18:12] <skunkworks> :)
[19:19:19] <jepler> kwajpol: for instance, I happen to know that the ubuntu 6.06 package name for "pyopengl" is "python2.4-opengl"
[19:22:26] <jepler> kwajpol: if you're able to determine the proper package names, please improve the wiki by adding them.
[19:23:11] <alex_joni> cradek: you should have seen he's using advanced psychology leverages to try to convince us change the software to suit him
[19:23:37] <jepler> alex_joni: all cradek can see is that if that's what he had in mind, it didn'twork
[19:24:03] <alex_joni> I'd invite him to go to austria
[19:24:12] <alex_joni> specifically to this village: http://en.wikipedia.org/wiki/Fucking
[19:25:21] <skunkworks> :) The frequently stolen traffic sign
[19:25:31] <cradek> eh I don't want to be that way
[19:25:39] <cradek> he should use whatever software he wants
[19:26:09] <cradek> if he picks free software, he can improve it to suit his tastes if he wants, he holds all the cards
[19:26:15] <skunkworks> http://www.cnczone.com/forums/showthread.php?p=305473#post305473
[19:26:33] <alex_joni> yeah, but it seems he chose to pay for nonfree sw .. so it's his choice
[19:26:53] <cradek> when you do that, complaining is your only choice
[19:27:13] <alex_joni> skunkworks: you failed to say I mentioned austria
[19:27:30] <skunkworks> :) I could not bring myself to do that.
[19:28:01] <alex_joni> I was relying on that
[19:28:11] <cradek> I sympathize with his need, and machines like that are why I did the probing/tool length stuff
[19:28:20] <skunkworks> kill them with kindness
[19:28:21] <cradek> (that's a MUCH better solution than jogging)
[19:28:31] <skunkworks> cradek: exactly
[19:28:38] <cradek> moving the machine while the program is paused is dangerous at best
[19:29:05] <cradek> it's a bad solution
[19:36:49] <archivist> hmm any cnc lathe subs in the east midlands
[19:37:14] <archivist> somebody has just walked in with a job
[20:13:02] <jepler> maybe I'll work on a digital I/O only firmware for pluto during fest .. shouldn't be too much work
[20:19:24] <jepler> hopefully fast enough for "fast thread" use: it would be about 4 I/O instructions (10 epp bus cycles) to read and write everything
[20:20:28] <cradek> what kind of app do you have in mind?
[20:21:25] <JymmmmEMC> jmkasunich: ping
[20:21:41] <jepler> cradek: nothing more specific than "more digital I/Os from a regular parport"
[20:22:30] <cradek> I wonder if any of jmk's stepgen code could be reused to make a (even rudimentary) step generator
[20:24:04] <jepler> we've talked about stepgen for pluto
[20:24:48] <jepler> I think you are unlikely to end up using code from jmk's stepgen effort on pluto, since from the very beginning you're carefully counting bits
[20:25:26] <jepler> (flipflops in this case)
[20:25:34] <JymmmmEMC> jmkasunich: It looks like I wasn't the only one to think of using a pulse signal through the ESTOP circuit (I found this after I had thought of the idea) http://www.pilz.com.au/pulse.htm
[20:25:38] <cradek> I see
[20:25:47] <cradek> I'm impressed with what the current firmware does - it has everything I needed
[20:26:02] <cradek> except maybe watchdog - we should test that
[20:26:10] <jepler> oh yeah that's a much better thing to work on -- watchdog is important
[20:26:43] <cradek> is all that stuff on your laptop?
[20:26:44] <jepler> it looks like there's room for the watchdog counter and "tristate all outputs", but getting the watchdog counter operations right is tricky
[20:26:47] <jepler> yes
[20:26:50] <JymmmmEMC> I thought you guys were playing with mesa board, Now pluto too?
[20:26:52] <cradek> (assuming you don't intend to take another machine)
[20:27:14] <jepler> cradek: I thought about bringing my amd64 but swp will have one of those he said
[20:27:26] <cradek> good, more room for my ridiculous machine
[20:27:47] <JymmmmEMC> ridiculous machine?
[20:28:02] <jepler> JymmmmEMC: server-sized tower case
[20:28:03] <jepler> weighs about 250 lbs or so
[20:28:20] <JymmmmEMC> 250LBS ?! WTH
[20:28:29] <cradek> it's more like 450lbs
[20:28:31] <JymmmmEMC> what's in it?
[20:28:35] <cradek> 700 with the monitor
[20:28:42] <jepler> JymmmmEMC: "Pluto_servo is an emc2 software driver and associated firmware that allow the Pluto-P board to be used to control a servo-based CNC machine." http://linuxcnc.org/docs/2.1/html/hal/pluto_servo.html
[20:28:46] <cradek> pentium 166
[20:29:09] <JymmmmEMC> cradek: oh, heh.... you mean it's just dman heavy being so old?
[20:29:10] <jepler> I got a pluto before I got a mesa
[20:29:54] <JymmmmEMC> jepler: it's just for servo though?
[20:30:05] <jepler> JymmmmEMC: yes, it is PWM out and Quadrature in
[20:31:08] <JymmmmEMC> Hmmm I see Pluto, Pluto 2, and pluto 3, is this thing break pass the 2us delat of pport?
[20:31:35] <JymmmmEMC> delay
[20:32:47] <jepler> JymmmmEMC: it is based on offloading work from the PC to the external board---the FPGA is in charge of counting at 40MHz and turning the PWM on and off. The PC just sends the desired duty cycle about every 1ms (the "servo-thread" speed)
[20:33:05] <jepler> JymmmmEMC: just like with mesa and jon elson's board there is no "fast thread" at all
[20:33:40] <jepler> so no worrying about what BASE_PERIOD you need to properly receive the quadrature feedback from the stepper motors
[20:34:12] <JymmmmEMC> jepler: Ah, thought it be nice for steppers, with the extra IO and such
[20:34:13] <jepler> the PC just retrieves the number of quadrature counts seen every 1ms
[20:34:52] <jepler> JymmmmEMC: it would probably be nice for high-speed stepper waveforms too, but somebody has to write the firmware to do it first
[20:35:52] <JymmmmEMC> jepler: Heh, I barely understand the FPGA acronym, much less coding for it! LOL
[20:36:04] <jepler> JymmmmEMC: it's a bit different than programming for computers or microcontrollers
[20:36:44] <JymmmmEMC> Yeah, I understand you are actually "writing" your own soft IC (so to speak).
[20:37:05] <JymmmmEMC> Boing 737, 30000 feet and climbing =)
[20:38:07] <jepler> JymmmmEMC: I have no doubt that anyone who can learn the fundamentals of microcontroller programming can learn the fundamentals of FPGA programming
[20:38:43] <JymmmmEMC> jepler: I haven't written C (much less ASM) since 1984
[20:39:57] <JymmmmEMC> $60 isn't too shabby either.
[20:40:38] <cradek> jepler shoehorned enough into it to make a nice little $60 servo controller
[20:40:51] <JymmmmEMC> sweet
[20:41:22] <jepler> if you decided to buy one of those knjn.com boards to learn about FPGAs, you would probably be better off with a USB one than the parallel port one
[20:41:24] <JymmmmEMC> It be nice to be able to use ANY computer and have smooth waveform for stepper control.
[20:41:53] <JymmmmEMC> jepler: Sadly enough, I only have ONE computer (laptop) that has USB2, the rest are 1.1
[20:43:10] <JymmmmEMC> Hell, my laptop is unique in that it has USB2, Firewire, DVDRW, built in FDD, SD slot, serial port and parallel port all built in.
[20:43:42] <JymmmmEMC> P4 3GHz
[20:44:22] <SWPadnos> that's just like my laptop
[20:44:33] <SWPadnos> except for the USB2 and P4 3GHz
[20:44:40] <SWPadnos> and DVDRW
[20:44:44] <SWPadnos> and SD slot(s)
[20:44:46] <JymmmmEMC> SWPadnos: heh
[20:44:55] <SWPadnos> ok, so it has a floppy and serial/parallel ports :)
[20:45:05] <SWPadnos> and FireWire!
[20:45:15] <JymmmmEMC> Yeah, there are only two mfg's that still have paraport's... Fujitsu and Panasonic Toughbooks.
[20:47:05] <JymmmmEMC> and before anyone says anything... No, I can't use USB2SERIAL adapters. The polling mucks up RT data collection from GPS/Radio.
[20:48:54] <JymmmmEMC> SWPadnos: floppy and FW huh? I'm surprised at that combination. What brand?
[20:49:22] <SWPadnos> M-Tech
[20:49:28] <SWPadnos> it's 5 years old
[20:49:30] <SWPadnos> oops, 6
[20:49:37] <JymmmmEMC> Ah, never heard of it
[20:50:13] <SWPadnos> http://www.mtechlaptops.com
[20:50:22] <SWPadnos> I think they're just about next door to you, come to think of it
[20:50:39] <SWPadnos> it's a Clevo / Kapok computer, like most of the clones and many of the name brands
[20:52:38] <JymmmmEMC> gotcha
[20:56:06] <SWPadnos> here you go: http://www.mtechlaptops.com/specifications/mtechd900k.htm
[20:56:48] <SWPadnos> no floppy though (except USB)
[21:01:45] <anonimasu> SWPadnos: seems like a awfully expensive laptop
[21:05:18] <JymmmmEMC> Is it common to have writeback cache on a floppy?
[21:05:37] <anonimasu> err actually no it's not expensive
[21:07:31] <SWPadnos> it's a big fast powerful laptop, so expensive just goes along with the territory ;)
[21:07:49] <SWPadnos> it's cheap for those specs, but still expensive relative to the entry level
[21:07:57] <anonimasu> heh
[21:08:45] <SWPadnos> try this one: http://www.mtechlaptops.com/specifications/mtechm550n.htm
[21:08:56] <SWPadnos> it's ~$800
[21:09:52] <anonimasu> my little laptop costs like 2300$
[21:09:56] <anonimasu> the work one..
[21:10:02] <anonimasu> ultraportable..
[21:10:08] <anonimasu> so I guess I shouldnt whine :D
[21:10:58] <JymmmmEMC> I'm not impressed with Dual CPU/Core systems
[21:11:10] <JymmmmEMC> Lack of real support for them
[21:11:34] <anonimasu> well, it's not spectacular on any OS.. :)
[21:11:38] <anonimasu> except maybe solaris or something
[21:12:20] <JymmmmEMC> Yeah, but SMP is something *I* feel that should be handled by the OS in it's entirety, not the application(s).
[21:12:33] <anonimasu> agreed..
[21:12:42] <cradek> eh?
[21:13:19] <cradek> smp IS handled by the OS
[21:13:25] <SWPadnos> JymmmmEMC, I thought I told you that was impossible
[21:13:35] <SWPadnos> do I need to up your meds?
[21:13:40] <cradek> ha
[21:13:46] <cradek> ok I won't try to help here
[21:13:52] <SWPadnos> heh
[21:13:53] <JymmmmEMC> SWPadnos: READ MY LIPS.... I said I FEEL that is the way it should be damnit!
[21:14:07] <SWPadnos> he's talking about having the OS auto-SMP-ify non-threaded apps ...
[21:14:08] <SWPadnos> JymmmmEMC, heh
[21:14:22] <cradek> um
[21:14:24] <cradek> ok
[21:15:27] <JymmmmEMC> SWPadnos: Apps can only go as fast as the CPU/OS they're running on. Why couldn't (in theory) the OS just simpely tell the app "I have super powers, so bring it on!" and let the OS distribute the load
[21:15:43] <SWPadnos> a single app can't be "distributed" by the OS
[21:15:49] <JymmmmEMC> SWPadnos: bullshit
[21:15:52] <SWPadnos> multiple apps can be (and are) balanced on multiple cores
[21:15:57] <SWPadnos> JymmmmEMC, not bullshit
[21:15:58] <cradek> you said you did assembler in '84
[21:16:08] <JymmmmEMC> cradek: no, just C
[21:16:11] <cradek> remember how you'd do an operation, and in a later instruction check the flags that resulted?
[21:16:15] <cradek> oh
[21:16:21] <SWPadnos> same thing
[21:16:24] <SWPadnos> i++;
[21:16:33] <SWPadnos> if (i<10) blah
[21:16:54] <SWPadnos> what if the core that gets the if statement executes it before the core that gets the i++ statement?
[21:17:15] <cradek> things have to happen sequentially, programs are written that way
[21:17:31] <cradek> if things can be separate and concurrent, programmers use threads
[21:17:50] <archivist> bring back the transputer
[21:17:50] <JymmmmEMC> This is just a pet peeve of mine is all.... You have to have an OS that supports multiple cpu's, why couldn't it take whatever is thrown at it, and let the OS split the load across CPUs. Kinda like a conductor of sorts.
[21:17:58] <cradek> it does
[21:18:05] <SWPadnos> it does, for various chiunk sizes of code
[21:18:12] <SWPadnos> one chunk is a program
[21:18:15] <SWPadnos> another is a thread
[21:18:15] <alex_joni> JymmmmEMC: it does, but you need to give it more than one process
[21:18:26] <cradek> if you want to make a PBJ and 2 slices of bread and one jar of peanut butter, adding people won't make it get done faster
[21:18:37] <cradek> but, if you have an entire loaf of bread and case of PB, it *will*
[21:18:40] <Guest829> does the lath work in diamiter programming, or are you stuck with radius programming?
[21:18:42] <SWPadnos> what the OS can't do is take something that was written to be sequential (an individual app or individual thread), and make it work in parallel
[21:18:51] <jepler> Guest829: radius programming only
[21:18:58] <anonimasu> SWPadnos: yep..
[21:19:07] <cradek> Guest829: diameter is shown on the screen now (not released yet)
[21:19:08] <JymmmmEMC> Take for example some fancy video cards that you can use multiple cards and "bind" them together to get better performance.
[21:19:09] <alex_joni> SWPadnos: actually the processor does that
[21:19:24] <alex_joni> for some limited parts of it
[21:19:45] <SWPadnos> JymmmmEMC, in that case, the software (driver) is written to support multiple CPUs (the individual processing units)
[21:19:58] <Guest829> does constent surface speed work yet? or is it to be done
[21:19:59] <alex_joni> JymmmmEMC: try running deltaforce (cpu graphics only) on those cards
[21:20:03] <alex_joni> would it be better?
[21:20:07] <alex_joni> Guest829: to be done
[21:20:39] <SWPadnos> alex_joni, yes, the CPUs have some support for out-of-order execution, but only on a small sequential set of instructions, and the amount of work (chip area) needed to synchronize that is almost as much as the processing units themselves
[21:20:46] <Guest829> how do you use tool comp?
[21:20:51] <Guest829> for lath
[21:20:56] <alex_joni> SWPadnos: I know, I did study this for about a year
[21:21:02] <SWPadnos> heh :)
[21:21:06] <alex_joni> Guest829: in the wiki there's a nice picture
[21:21:25] <cradek> Guest829: you have to set up a lathe format tool table, then use G40/41/42 as usual
[21:21:31] <alex_joni> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Lathe_Advanced_Features
[21:21:37] <cradek> but be in G18 plane first
[21:21:41] <Guest829> but that is tool nose radius comp
[21:21:49] <alex_joni> SWPadnos: I used to be familiar with henessy & patterson, tannenbaum, etc
[21:21:57] <alex_joni> the books, not the people
[21:22:01] <SWPadnos> I probably am, but not by name ;)
[21:22:01] <Guest829> tool length comp is what need
[21:22:23] <Guest829> i have tool tbl setup
[21:22:27] <JymmmmEMC> SWPadnos: I honestly wouldn't mind having a 3rd CPU doing all the logistics if I'd get that kind of performance on dual cpu. *ie twice as fast processing)
[21:22:54] <JymmmmEMC> SWPadnos: (crypto stuff specifically)
[21:23:04] <SWPadnos> JymmmmEMC, I'm not sure it's possible for arbitrary programs
[21:23:26] <alex_joni> SWPadnos:
[21:23:28] <SWPadnos> a CPU can do this because it knows exactly which instructions depend on what data and more importantly, what results from other instructions
[21:23:40] <SWPadnos> alex_joni, my client isn't opera - now what? ;)
[21:23:50] <alex_joni> get a proper client :P
[21:23:51] <SWPadnos> heh
[21:23:54] <JymmmmEMC> SWPadnos: Yeah, I understnad, just bugs me. I even looked into distributed clusters too. no luck.
[21:23:55] <cradek> Guest829: then just issue G43
[21:24:27] <cradek> Guest829: after loading the tool, like T1 M6, G43 H1
[21:24:31] <SWPadnos> JymmmmEMC, no luck because it may not be a solvable problem for the general case (some specific cases, sure)
[21:24:35] <cradek> Guest829: then the next move will be compensated
[21:24:35] <SWPadnos> though it is still annoying
[21:24:43] <alex_joni> SWPadnos: http://www.amazon.com/Computer-Architecture-Quantitative-Approach-Kaufmann/dp/1558605967
[21:24:53] <JymmmmEMC> SWPadnos: Yeah, the app HAS to be written to support SMP
[21:24:54] <Guest829> so mill G43 H1 is same for lathe then?
[21:24:59] <cradek> yes
[21:25:08] <cradek> except it gives X and Z offsets, not just Z
[21:26:11] <JymmmmEMC> Well, this is my last hope... reflashed the BIOS to see if this resolves the issue of loosing CMOS settings, but retains RTC setting.
[21:27:10] <JymmmmEMC> Tyan thought the mobo might be shorting to the case, but I benched it and that's not it.
[21:27:25] <Guest829> does G49 cancel X and Z as in milling from G43
[21:28:51] <cradek> yes
[21:30:20] <JymmmmEMC> No FPGA's are static are they? Menaing they HAVE to be reloaded each time (from PC or flash chip)?
[21:30:26] <alex_joni> SWPadnos: that guy tanenbaum .. is the guy that wrote minix
[21:30:37] <SWPadnos> yep
[21:31:07] <alex_joni> and he has a quite famous thread of emails with another guy called linus
[21:32:53] <SWPadnos> JymmmmEMC, there are lots of flash FPGAs
[21:33:26] <SWPadnos> there are also some that have the cells directly changed when they're programmed, so they're live at power-up