#emc | Logs for 2007-01-11

[00:12:16] <skunkworks> mmmm chocolate covered cherries.. I think it is the best part about the christmas season.
[00:15:12] <fenn> soaked in booze too
[00:15:21] <EldonB46> It is snowing around Seattle - lots of fun! :)
[00:15:32] <SWPadnos> I see your cherries and raise you some Bailey's Chocolate Truffles
[00:24:33] <skunkworks> :)
[00:54:00] <skunkworks> fenn: did they give a copper weight for the box of clad you got?
[00:55:17] <fenn> it was mixed
[00:55:45] <fenn> anything from 1/2 to 4 oz
[00:56:02] <fenn> i think if i weigh it i can calc the thickness
[00:56:47] <fenn> or peel off a tiny bit
[00:57:02] <skunkworks> right
[00:57:02] <fenn> or (least favorite option) etch and measure the difference
[00:57:09] <skunkworks> :)
[00:57:31] <skunkworks> double and sigle sided?
[00:57:40] <fenn> mostly double
[00:58:21] <fenn> the seller (abcfab) is on vacation for 6 months
[00:58:34] <skunkworks> by choice? ;)
[00:58:41] <fenn> dunno\
[00:59:49] <ds3> is there a table of copper oz to thickness of foil?
[01:01:17] <fenn> 1 oz = 35micron?
[01:02:14] <fenn> == .00138"
[01:02:23] <SWPadnos> it's ounces/square foot, so you can get thickness if you know that copper has a density of 559 lb/cu ft
[01:02:39] <fenn> hmm thats a lot of units swapping
[01:02:42] <SWPadnos> or 0.324 lb/in^3
[01:02:44] <SWPadnos> heh :)
[01:03:30] <skunkworks> * skunkworks googles for a chart ;)
[01:04:25] <fenn> units '(oz/ft^2)/(559lb/ft^3)'
[01:04:26] <fenn> 0.0013416816 inches
[01:04:46] <SWPadnos> see - it worked ;)
[01:05:24] <fenn> * fenn gives SWPadnos a bailey's irish truffle
[01:05:33] <SWPadnos> mmmmmmm - no thanks, I just had one :)
[01:05:49] <fenn> ah the ingratitude in this channel is sickening
[01:06:22] <SWPadnos> ack. I hate it when I accidentally click on the launcher for a really big program that I haven't run in years
[01:06:54] <SWPadnos> (like Metrowerks Codewarrior for Motorola DSPs)
[01:09:07] <skunkworks> I always click on the print button when I am trying to refresh the web page.
[01:09:51] <SWPadnos> I should order the launch bar buttons by program size, so an accident doesn't take much longer than I'm expecting the program I want to take
[01:15:41] <Rugludallur> nigh guys
[01:15:42] <ds3> Or just use a CLI ;)
[01:16:01] <SWPadnos> nah - I'd probably start mis-tabbing ;)
[01:18:04] <ds3> toggle switches then!
[02:17:44] <robin_sz> * robin_sz got his robot working :)
[02:19:42] <skunkworks> robin_sz: Great - how many coppies of the software did you end up getting?
[02:26:39] <robin_sz> 1 so far
[02:26:46] <robin_sz> in italian
[02:35:45] <tomp> borco vatuna, solo una! does the robot wave it's arms a lot?
[02:37:42] <robin_sz> well, so far it has refused to work late, taken several long breaks and stopped for 2 hours for lunch
[02:38:56] <tomp> :)
[02:59:13] <A-L-P-H-A> shit... jymm doesn't have out here enough, for me to ask him Qs.
[03:00:01] <tomp> jepler: new cvs update "make: *** No rule to make target `../configs/lathe-pluto/emc.nml', needed by `configs'. Stop."
[03:02:03] <cradek> tomp: you must not have got a full update
[03:02:16] <cradek> tomp: in the toplevel directory (above src) cvs update -dP?
[03:03:11] <tomp> ok did cvs -z5 update will chg now thanks
[03:03:29] <cradek> -z5 is fine, but you also need -dP
[03:03:47] <cradek> cvs -z5 update -dP
[03:07:39] <tomp> cradek: worked fine
[03:07:42] <tomp> thanks
[03:08:46] <tomp> now i can play with awallin's new pyvcp widgets ( the new phone books are here ! the new phone books are here !)
[03:09:06] <SWPadnos> Navin R. Johnson ...
[03:46:09] <lerman_______> lerman_______ is now known as lerman
[04:16:36] <jmkasunich_> jmkasunich_ is now known as jmkasunich
[04:22:12] <Jymmm> I picked up an IR non-contact thermometer, does anyone know a way to see how accurate it is?
[04:22:38] <skunkworks> boil some water?
[04:22:54] <Jymmm> how do you know when it's at 212 deg?
[04:23:07] <skunkworks> its boiling?
[04:23:15] <Jymmm> and when is that?
[04:24:06] <skunkworks> the water temp is not going to go over 212 until it is gone
[04:24:23] <Jymmm> are you sure?
[04:24:23] <skunkworks> or what ever offset you are from sea-level
[04:24:43] <Jymmm> is that distilled water? tap water?
[04:25:53] <skunkworks> http://www.biggreenegg.com/boilingPoint.htm
[04:26:39] <Jymmm> ah cool... and my gps is calirbated to the airport too
[04:27:12] <A-L-P-H-A> Jymmm, I was looking for you earlier... damn you for not being here earlier.
[04:27:55] <A-L-P-H-A> what is 100°C in °F?
[04:28:39] <A-L-P-H-A> 100°C that's boiling at 101.2kPa as 20°C room temp. If my STP numbers are remembered correctly
[04:28:42] <A-L-P-H-A> SI of course! :P
[04:29:04] <A-L-P-H-A> stupid americants using some old system...
[04:31:24] <skunkworks> bed time - night
[04:31:32] <Jymmm> G'Night skunkworks
[04:53:32] <lerman_______> lerman_______ is now known as lerman
[08:59:33] <Dallur> morning+
[09:02:53] <alex_joni> morning-
[09:18:25] <anonimasu> morning
[09:24:46] <alex_joni> well.. lunch soon :)
[09:36:03] <Dallur> Funny about that 3d printer team, turns out they are using EMC to mill the parts for the 3d printer :P
[09:37:14] <alex_joni> yeah :)
[09:37:22] <alex_joni> well.. I liked the idea initially
[09:38:50] <Dallur> so did I, but I would have thought some sort of wax would be more practical, have a sort of "hot was gun" squirting it, that way you can re-use it
[09:38:59] <alex_joni> heh :)
[09:39:15] <alex_joni> well.. I liked it untill I saw they plan to use java (yuck-tm)
[09:42:20] <anonimasu> didnt they use hot glue?
[09:43:15] <Dallur> I'm not the biggest java fan but I recognize it's strong points and short failings, rapid development, ease of maintenance, modularity and code re-usability are all excellent, on the other hand OS integration, optimization, hardware integration and control is where it is really lacking ... turns out this project depends mostly on the second group
[09:43:49] <Dallur> anonimasu: I thought once it solidified there was no going back, let me check
[09:45:51] <anonimasu> hot glue is re-heatable..
[09:46:08] <anonimasu> :)
[09:46:43] <alex_joni> Dallur: ever seen RT-Java ?
[09:48:41] <Dallur> alex_joni: never ever.. it's just silly imho to have RT stuff on a virtual machine
[09:48:55] <alex_joni> well.. seen people who wrote that
[09:49:01] <anonimasu> :D
[09:49:06] <alex_joni> they even made a RT-garbage disposer
[09:49:10] <anonimasu> omfg.
[09:49:13] <alex_joni> so it's deterministic
[09:49:22] <anonimasu> * anonimasu is not a java fan
[09:49:23] <alex_joni> sick sick people :P
[09:49:24] <anonimasu> it's a scary language.
[09:49:53] <anonimasu> sure, garbage collection :D
[09:57:26] <Dallur> Is there any reason why one could not use a standard emc2 mill setup with a "glue gun" as a 3d printer ?
[09:58:37] <Dallur> I would think the only part you need is something to generate the gcode
[09:58:58] <alex_joni> the problem is the nozzle I'd say
[09:59:11] <alex_joni> a standard glue gun outputs too much, without control
[09:59:29] <anonimasu> hm, well, a needle and a dispenser mechanism..
[09:59:39] <anonimasu> that loads yop a glue pellet :)
[10:02:23] <alex_joni> depends on the resolution you want
[10:02:32] <alex_joni> if you aim for 600dpi it gets tricky fast D
[10:02:33] <alex_joni> :D
[10:12:07] <alex_joni> or 4800dpi :D
[10:17:14] <anonimasu> 4800 dpi is just 0.005
[10:17:36] <anonimasu> 25.4/4800
[10:17:55] <anonimasu> though getting it fast enough is hard..
[10:18:19] <anonimasu> alex_joni: laser's probably a good way to go
[10:26:13] <Dallur> anonimasu: the tough part with laser is to get the photons to stick (chuckle)
[10:32:57] <lerneaen_hydra> 'lo
[10:33:18] <Dallur> hey lerneaen_hydra
[10:34:35] <lerneaen_hydra> what's happening
[10:36:04] <Dallur> just pondering about 3d printers
[10:36:29] <lerneaen_hydra> ah, cool
[10:39:09] <anonimasu> Dallur: lol
[10:39:43] <anonimasu> hm..
[10:39:49] <anonimasu> I've got a production run of 10 parts soon..
[10:40:00] <anonimasu> I hate fixtures.
[10:40:12] <anonimasu> im pondering a way to zero my tools easily..
[10:40:32] <anonimasu> let's see I have inductive sensors.
[10:41:20] <Dallur> make a homing sequence which homes to an inductive sensor
[10:41:20] <anonimasu> or microswitches..
[10:41:28] <anonimasu> I wonder how repeatable they are..
[10:41:35] <anonimasu> I just need it for setting up Z on my toold..
[10:41:39] <anonimasu> need a toolchange..
[10:41:51] <anonimasu> I need a limit switch on a axis.
[10:41:54] <anonimasu> really so I can offset my fixture
[10:41:59] <Dallur> anonimasu: I wired everything up with microswitches, had to throw them all away because of HF
[10:42:20] <anonimasu> this is at the mill..
[10:42:21] <lerneaen_hydra> Dallur: HF?
[10:42:27] <Dallur> lh: high frequency
[10:42:34] <anonimasu> the switches are integrated into the prefab stuff..
[10:42:36] <lerneaen_hydra> noise?
[10:42:59] <Dallur> lh: yup, just having a HF arc going off <1m away would trigger them
[10:43:21] <Dallur> lh: shielding issue probably
[10:43:44] <lerneaen_hydra> yeah
[10:44:53] <Dallur> anonimasu: regarding the repeatability of inductive sensors, I have no idea, probably depends somewhat on the sensor but I'm hoping it will be good since I will be using them for all homing/limits
[10:45:30] <alex_joni> Dallur: inductive sensor + index pulse :P
[10:45:53] <lerneaen_hydra> haha
[10:46:28] <Dallur> alex_joni: index pulse ?
[10:46:32] <alex_joni> encoder :D
[10:48:23] <Dallur> alex_joni: gotcha, I was thinking in terms of inductive "switches" not sensors so the threshold is built into the sensor body
[10:49:03] <alex_joni> there are some very precise optical switches
[10:49:24] <Dallur> alex_joni: diffr. pattern stuff ?
[10:49:39] <alex_joni> sorry?
[10:49:56] <Dallur> alex_joni: sorry misunderstood, :P
[10:50:26] <Dallur> alex_joini: disregard that, I thought you were talking about distance measuring instead of "breaking" the beam
[10:52:21] <A-L-P-H-A> ?
[10:52:26] <A-L-P-H-A> wussup?
[10:55:03] <alex_joni> Dallur: no, simple breaking the beam
[10:55:10] <alex_joni> or even distance treshold
[10:55:19] <alex_joni> simple/cheapish :D
[11:00:54] <lerneaen_hydra> A-L-P-H-A!
[11:00:56] <lerneaen_hydra> 'day
[11:02:03] <anonimasu> analog inductive are super precise.
[11:02:03] <anonimasu> :
[11:02:04] <anonimasu> :)
[11:02:49] <anonimasu> but they cost 70 eur
[11:02:51] <alex_joni> it's all relative :)
[11:03:01] <alex_joni> even saying that is cheap is relative :P
[11:03:11] <anonimasu> lol
[11:03:11] <anonimasu> yeah
[11:04:40] <lerman_______> lerman_______ is now known as lerman
[11:05:52] <lerneaen_hydra> no, cheap is well defined, that's what I buy
[11:06:37] <A-L-P-H-A> hi
[11:06:45] <A-L-P-H-A> still haven't fixed my sleep cycle.
[11:06:47] <A-L-P-H-A> :(
[11:07:01] <lerneaen_hydra> oho, what have you been busy with eh?
[11:18:21] <A-L-P-H-A> not that girl.
[11:18:35] <A-L-P-H-A> oh wow... it's thursday already?
[11:18:44] <A-L-P-H-A> I was supposed to go have dinner with her tonight.
[11:18:52] <A-L-P-H-A> 'was'... wonder should I.
[11:23:20] <A-L-P-H-A> lerneaen_hydra... so you?
[11:23:20] <A-L-P-H-A> what's up?
[11:32:15] <alex_joni> burning backups here
[11:33:15] <A-L-P-H-A> sounds like fun... want to play kdice?
[11:33:20] <A-L-P-H-A> http://kdice.com/
[11:33:35] <alex_joni> nah.. too lazy
[11:33:44] <A-L-P-H-A> nothing even to download. :)
[11:33:45] <A-L-P-H-A> lazy arse.
[11:33:49] <A-L-P-H-A> alex_joni, you even play games?
[11:33:59] <alex_joni> not ever :)
[11:34:07] <alex_joni> well.. sometimes I do, but very seldom
[11:34:34] <alex_joni> I still play DF with some friends once / year maybe
[11:35:05] <A-L-P-H-A> DF?
[11:35:08] <A-L-P-H-A> Delta Force?
[11:35:16] <alex_joni> yeah :)
[11:35:33] <A-L-P-H-A> 1?
[11:35:37] <alex_joni> yup
[11:35:41] <A-L-P-H-A> wow. old school.
[11:36:02] <alex_joni> yeah .. played it a lot in college
[11:36:05] <lerneaen_hydra> oh come on now A-L-P-H-A, no point in being shy
[11:36:07] <alex_joni> a couple of years ago :D
[11:37:19] <alex_joni> A-L-P-H-A: delta force / Ice Flows - team deathmatch :D
[11:37:33] <lerneaen_hydra> alex_joni: nonono, ut2004, that's where it's at
[11:37:43] <alex_joni> nah, don't really like UT
[11:37:58] <alex_joni> tried CS for a while, didn't really like that either
[11:38:02] <lerneaen_hydra> why not?
[11:38:05] <alex_joni> dunno..
[11:38:25] <alex_joni> no particular thing that bothers me
[11:38:50] <alex_joni> didn't like newer DF's either
[11:39:08] <lerneaen_hydra> hmm, I see
[11:39:23] <alex_joni> graphics is better.. but the feeling is worse..
[11:39:26] <alex_joni> at least for me :)
[11:40:00] <alex_joni> guess I used to play too much DF1 :D
[11:40:14] <alex_joni> back then it was the only choice (no accel needed, TCP, ...)
[11:40:30] <lerneaen_hydra> oh,right
[11:40:41] <lerneaen_hydra> just a matter of what you're used to
[11:40:49] <lerneaen_hydra> does DF run native on linux?
[11:41:04] <alex_joni> no idea :)
[11:41:12] <alex_joni> I liked HL1 too :)
[11:41:28] <lerneaen_hydra> yeah
[11:41:35] <lerneaen_hydra> HL1 was good, HL2 as well
[11:42:08] <alex_joni> I grew out of gaming before HL2 appeared
[11:42:41] <lerneaen_hydra> oh, I see
[11:59:53] <A-L-P-H-A> I love UT2k4, I also play UT99
[12:07:08] <lerneaen_hydra> yay!
[12:07:34] <anonimasu> iab
[12:07:39] <lerneaen_hydra> wb
[12:14:32] <anonimasu> ,thanks
[12:16:55] <anonimasu> *yawn*
[12:17:43] <lerneaen_hydra> * lerneaen_hydra wants death note to air ep 14
[12:18:32] <A-L-P-H-A> ?
[12:18:38] <A-L-P-H-A> who's got UT99, or UT2k4 installed?
[12:18:55] <lerneaen_hydra> I do
[12:18:54] <A-L-P-H-A> lerneaen_hydra, where are you?
[12:19:01] <A-L-P-H-A> lermann, want to play?
[12:19:02] <lerneaen_hydra> sweden, borkland
[12:19:24] <A-L-P-H-A> what version do you want to play, and got a server you want to frag at?
[12:19:39] <A-L-P-H-A> I find ut99 faster game play. :)
[12:19:40] <lerneaen_hydra> A-L-P-H-A: err, I was just about to go and finish my bike up
[12:19:48] <A-L-P-H-A> no!
[12:19:48] <lerneaen_hydra> how long you going to be online?
[12:19:50] <A-L-P-H-A> no bike...
[12:19:56] <A-L-P-H-A> come play ut99.
[12:19:59] <A-L-P-H-A> at this ip... sec.
[12:20:04] <lerneaen_hydra> I don't even have ut99
[12:20:16] <A-L-P-H-A> bah... okay.
[12:20:25] <lerneaen_hydra> gonna stay here for a few more hours?
[12:20:25] <A-L-P-H-A> want to play ut2k4 somewhere then? you got a fav server?
[12:20:37] <A-L-P-H-A> honestly, probably not, cause I haven't slept yet.
[12:20:42] <lerneaen_hydra> ah
[12:20:46] <A-L-P-H-A> I'm in Toronto... so it's 7:20am.
[12:20:46] <lerneaen_hydra> what's the time?
[12:20:49] <lerneaen_hydra> O.o
[12:20:52] <lerneaen_hydra> dude, sleep
[12:20:57] <lerneaen_hydra> now
[12:21:03] <A-L-P-H-A> need to frag some pixels.
[12:21:19] <lerneaen_hydra> need to frag sheep
[12:21:37] <A-L-P-H-A> I wonder is there a mod, to frag sheep.
[12:21:45] <lerneaen_hydra> doubtlessly
[12:22:35] <anonimasu> hehe
[12:22:39] <anonimasu> hehe
[12:31:41] <alex_joni> A-L-P-H-A: no sleep today
[12:31:46] <alex_joni> the only way to get back on track :)
[12:37:41] <anonimasu> :)
[12:37:52] <A-L-P-H-A> I'm trying
[12:37:54] <A-L-P-H-A> but I'm weak. :)
[12:38:11] <alex_joni> heh
[12:38:15] <A-L-P-H-A> I was supposed to have dinner with this girl today, but I think I'm going to pass... she isn't going to put out... but she is fun.
[12:38:27] <alex_joni> at least stay up one hour later than yesterday
[12:38:44] <A-L-P-H-A> I woke up 2 hrs earlier than I would have... but thentook a nap to make up the time diff.
[12:38:59] <A-L-P-H-A> so happy that I figured out how to fix the noise from my computre.
[12:39:05] <A-L-P-H-A> oiled the stupid video card car.
[12:39:07] <A-L-P-H-A> car=fan
[12:39:35] <A-L-P-H-A> maybe I'll read instead.
[12:45:25] <lerneaen_hydra> A-L-P-H-A: what type of oil?
[12:45:38] <lerneaen_hydra> it will get noisy quite fast again
[12:45:42] <lerneaen_hydra> a few weeks or so
[12:45:57] <A-L-P-H-A> 3in1
[12:46:48] <A-L-P-H-A> "Masters Lubricating Oiler with pull-out spout". "A preimium quality, all purpose lubricating oil for use on bearings and moving parts of pumps, motors, fans, compressors, etc."
[12:47:45] <alex_joni> ha
[12:47:56] <A-L-P-H-A> ha as well.
[12:47:59] <A-L-P-H-A> ha to what?
[12:48:11] <alex_joni> to PR BS
[12:48:18] <A-L-P-H-A> alex_joni. build me a femme-bot !
[12:48:24] <alex_joni> heh
[12:48:39] <alex_joni> it would be a bit cold to the touch
[12:48:40] <alex_joni> :D
[12:49:12] <alex_joni> darn, not so eager about backups anymore :(
[12:50:02] <lerneaen_hydra> huh wth?
[12:50:03] <A-L-P-H-A> nah... get water pipes, running through silicone, and it shuold be very lifelike. :)
[12:50:15] <A-L-P-H-A> lerneaen_hydra, are you working on your motorized bike?
[12:50:28] <lerneaen_hydra> I needed some info on the net
[12:50:31] <A-L-P-H-A> the convert, from bicycle, to motorized bicycle?
[12:50:36] <lerneaen_hydra> yeah
[12:50:37] <A-L-P-H-A> lerneaen_hydra, you needed to play UT.
[12:50:38] <A-L-P-H-A> 2k4.
[12:50:46] <lerneaen_hydra> sshh
[12:50:47] <A-L-P-H-A> what time is it there?
[12:50:48] <lerneaen_hydra> sleep
[12:50:51] <lerneaen_hydra> 1:50pm
[12:50:56] <A-L-P-H-A> I just played a hyper game... can't sleep
[12:50:58] <alex_joni> 2:50pm here
[12:51:06] <A-L-P-H-A> don't you have work?
[12:51:10] <alex_joni> me?
[12:51:16] <A-L-P-H-A> both of you!
[12:51:19] <lerneaen_hydra> I've got winter vacation
[12:51:23] <alex_joni> I am at work.. keep postponing what I should really do
[12:51:29] <lerneaen_hydra> and I need to finish the bike asap
[12:51:31] <lerneaen_hydra> :/
[12:51:37] <alex_joni> so I'm burning some dvd backups instead
[12:51:43] <A-L-P-H-A> lerneaen_hydra? Winter vacation, for how long?
[12:51:50] <alex_joni> A-L-P-H-A: till summer
[12:51:51] <alex_joni> :D
[12:52:06] <lerneaen_hydra> becuase I'm so good and don't have to redo any tests, till the 22nd
[12:52:15] <A-L-P-H-A> Sweden, is it so cold, that everything freezes over, so none of you work?
[12:52:16] <lerneaen_hydra> or rather, I'm too lazy to redo test
[12:52:19] <lerneaen_hydra> *tests
[12:52:34] <A-L-P-H-A> what tests?
[12:52:38] <lerneaen_hydra> haha, actually this year it's been about +8 degrees C all winter
[12:52:41] <lerneaen_hydra> math and electricity
[12:52:46] <lerneaen_hydra> I'm doing EE atm
[12:52:50] <A-L-P-H-A> oooooooooh, you're in school.
[12:52:54] <A-L-P-H-A> forgot.
[12:52:56] <lerneaen_hydra> indeed
[12:53:06] <lerneaen_hydra> I'm still wet behind the ears, as it were
[12:53:07] <A-L-P-H-A> and you said you used to be an americant.
[12:53:14] <lerneaen_hydra> indeed
[12:53:21] <A-L-P-H-A> how come you're in sweeden then?
[12:53:30] <lerneaen_hydra> parents work and stuff
[12:53:51] <A-L-P-H-A> should you be trying to picking up hot sweedish girls?
[12:54:04] <A-L-P-H-A> named, Heidy or something.
[12:54:05] <lerneaen_hydra> already am man, already am
[12:54:09] <lerneaen_hydra> no, that's german
[12:54:20] <alex_joni> gudrun
[12:54:29] <lerneaen_hydra> stereotypical swedish name is probably anna or something
[12:54:37] <A-L-P-H-A> lerneaen_hydra, so picked up?
[12:54:38] <alex_joni> bah.. anna
[12:54:42] <A-L-P-H-A> Anna?
[12:54:50] <A-L-P-H-A> Anna, isn't very Swedish.
[12:54:51] <lerneaen_hydra> gudrun makes me think of a fugly 50year old militant-feminist
[12:54:52] <alex_joni> not a very exciting name
[12:54:59] <A-L-P-H-A> Anna is just... plan banana.
[12:55:00] <alex_joni> lerneaen_hydra: yeah :D
[12:55:17] <lerneaen_hydra> A-L-P-H-A: exactly
[12:55:24] <A-L-P-H-A> Gurtrude.
[12:55:27] <A-L-P-H-A> Helga
[12:55:30] <lerneaen_hydra> haha
[12:55:40] <A-L-P-H-A> Penelope (hate that name too)
[12:55:50] <A-L-P-H-A> Lena is Austrian, isn't it?
[12:55:56] <alex_joni> sorta
[12:56:05] <lerneaen_hydra> could be, not too uncommon here
[12:56:09] <lerneaen_hydra> ah, bbl
[12:56:13] <A-L-P-H-A> exGF was austrian, named Lena... Chinese too... :( kinda miss her.
[12:56:15] <lerneaen_hydra> I get stuck here too easily
[12:56:25] <A-L-P-H-A> lerneaen_hydra, cause it's cold outside.
[12:56:30] <A-L-P-H-A> and more fun in here to just chat.
[12:56:36] <A-L-P-H-A> sit, enjoy your stay.
[12:56:39] <A-L-P-H-A> your milage will very.
[12:56:53] <A-L-P-H-A> How do you say little girl in Swedish?
[12:57:08] <A-L-P-H-A> Cause I was goint to say lolita, but that doesn't work for Swedish girls.
[12:57:35] <alex_joni> no, that's french
[12:58:06] <A-L-P-H-A> "sui mui mui", mean little girl, if pronouced in english... [cantonese]
[12:58:41] <A-L-P-H-A> crap I'm bored.
[12:58:47] <A-L-P-H-A> lerneaen_hydra! you didn't stick around did you.
[13:00:08] <A-L-P-H-A> haha... lerneaen_hydra, you suck. http://www.unreal.ie/utstats/chatlog.php?match=703
[13:01:49] <A-L-P-H-A> ohhhhhhhhhhh... lerneaen_hydra has a lair...
[13:01:53] <A-L-P-H-A> http://lerneaenhydra.shacknet.nu/ hahaha
[13:03:57] <alex_joni> http://www.unreal.ie/utstats/matchstats.php?match=703
[13:04:05] <A-L-P-H-A> does he get better?
[13:04:39] <A-L-P-H-A> nope. :( http://www.unreal.ie/utstats/matchplayer.php?match=703&player=8
[13:05:42] <A-L-P-H-A> he only played on that server for like 3 games.
[13:08:31] <A-L-P-H-A> oh freak'n wow. http://ut2004stats.epicgames.com/playerstats.php?player=1266135 here's his global scoring...
[13:08:46] <A-L-P-H-A> wow... he's nearly suicided as many times as he's killed.
[13:09:10] <alex_joni> yeah, just before getting killed it's better to suicide
[13:09:19] <alex_joni> worse stats for the opponents
[13:09:20] <alex_joni> lol
[13:09:26] <A-L-P-H-A> uh, yeah. whatever. :)
[13:10:41] <A-L-P-H-A> he last played yesterday
[13:11:14] <A-L-P-H-A> for a total of about an 1:24.2 minutes. :)
[13:14:39] <A-L-P-H-A> lerneaen_hydra, are you japanese?
[13:28:20] <alex_joni> A-L-P-H-A: what's the command to search for a IRC channel?
[13:28:37] <A-L-P-H-A> ?
[13:28:54] <A-L-P-H-A> /msg chanserv help list
[13:28:57] <A-L-P-H-A> that
[13:29:17] <A-L-P-H-A> alex_joni, that help?
[13:29:25] <alex_joni> not quite, but I found what I was looking for :)
[13:29:38] <A-L-P-H-A> what where you looking for , and how?
[13:29:47] <alex_joni> http://searchirc.com/network/Undernet
[13:30:16] <A-L-P-H-A> oh.
[13:30:17] <A-L-P-H-A> okay
[13:30:19] <A-L-P-H-A> sure.
[13:32:04] <A-L-P-H-A> buahaha. http://www.ex-parrot.com/~pete/upside-down-ternet.html
[13:35:07] <skunkworks> :) thats cool
[13:36:06] <alex_joni> yeah, older but nice
[13:44:27] <anonimasu> that's so hillarious
[14:02:53] <alex_joni> hi jeff
[14:03:45] <jepler> hello alex
[14:10:19] <awallin> jepler: do you know how much the performance difference between python and C is typically? for normal algorithms like sorting searching etc.
[14:11:04] <A-L-P-H-A> umm... optimized for both... C is faster.
[14:11:16] <A-L-P-H-A> C is compiled. isn't python interpretted?
[14:11:18] <jepler> awallin: it varies greatly. If you are able to use the parts of Python that are coded in "C", then they are very fast--for instance if you have a list of all strings or all numbers, then l.sort() runs very fast.
[14:11:47] <jepler> awallin: if you do math with 1000+ elements at a time using one of the numeric extensions, it will run at fortran-like speeds
[14:12:18] <jepler> awallin: on the other hand, you can write very slow code in Python -- the overhead for one Python function calling another Python function may be 1000x the overhead of one C function calling another, for instance.
[14:13:01] <awallin> jepler: how hard it is to re-code certain algorithms in C once you have a working python implementation? i.e. make the "hard" parts in C functions that are callable from python
[14:15:26] <jepler> awallin: once you isolate the "slow" stuff inside a single function call, you have to re-write that code in C (or C++), and then write the interface between C and Python, doing stuff like managing the reference counts of Python objects at that boundary.
[14:16:01] <jepler> awallin: there is a learning curve to the "C" api for extending Python, and you won't master it in a day
[14:17:27] <awallin> ok, I just find it a lot quicker to prototype in python than for example C, so I guess I'll go ahead with python and worry about speed later
[14:17:29] <jepler> for image-to-gcode I found that the speed was tolerable just using numarray
[14:17:48] <jepler> but it's near the lower end of tolerable
[14:19:35] <jepler> maybe once you run into a situation where it's too slow, we can talk about this but with a specific goal in mind
[14:19:40] <jepler> bbl
[14:22:01] <awallin> ok
[14:31:23] <lerneaen_hydra> A-L-P-H-A: O.o
[14:31:31] <lerneaen_hydra> I'm not japanese
[14:32:37] <lerneaen_hydra> suicides are often from ons levels that have very narrow paths and an infinite fall down
[14:33:06] <lerneaen_hydra> oh, wait
[14:33:09] <lerneaen_hydra> they're seperated
[14:33:17] <lerneaen_hydra> wtf, why do I have so many suicide
[14:34:19] <lerneaen_hydra> funny
[14:34:29] <lerneaen_hydra> anyway, my suckiness is relative
[14:35:04] <lerneaen_hydra> if you arrange the players in a typical game in point ranking I'm usually in the middle somewhere
[14:35:19] <lerneaen_hydra> there are always lots of players though that have tons of points
[14:35:24] <lerneaen_hydra> and some that really suck
[14:35:39] <lerneaen_hydra> also usually I tend to go for mission objectives rather than frags
[14:35:54] <lerneaen_hydra> so power nodes and objectives
[14:37:06] <lerneaen_hydra> on instagib deathmatches I really suck though
[14:37:14] <lerneaen_hydra> that may actually be a good thing
[14:37:41] <lerneaen_hydra> I don't really want to meet the people that aim to 1px accuracy and hit 90% of the time
[14:40:20] <alex_joni> excuses, excuses
[14:42:35] <awallin> what game is this? if I may ask...
[14:46:01] <alex_joni> UT2k4
[14:46:33] <lerneaen_hydra> ok, ok, I suck ;)
[14:46:57] <skunkworks> doom isn't hip anymore? ;)
[14:47:15] <alex_joni> lerneaen_hydra: http://www.joecartoon.com/pages/superfly_anim
[14:47:39] <lerneaen_hydra> supafly!
[14:48:53] <alex_joni> you do suck :))
[14:49:15] <lerneaen_hydra> huh?
[14:49:56] <alex_joni> it's a quote from that flash
[14:50:04] <lerneaen_hydra> oh, right
[14:51:14] <alex_joni> not sure from which one.. there are 6 I think
[14:52:57] <lerneaen_hydra> bbl
[15:29:18] <skunkworks> alex_joni: thanks again for fixing the irc java client. working great
[15:31:25] <alex_joni> no problem :)
[15:47:42] <alex_joni> hi rayh
[15:51:52] <alex_joni> bbl
[16:05:24] <CIA-8> 03awallin 07HEAD * 10emc2/docs/src/Master_Integrator.lyx: include vcp documentation in the integrator manual
[16:28:56] <CIA-8> 03awallin 07HEAD * 10emc2/docs/src/hal/vcp.lyx: start to document pyvcp
[16:29:09] <awalli1> how do I build only the docs ?
[16:29:20] <awalli1> 'make' inside /docs ?
[16:30:56] <jepler> in docs/src, 'make', or in src/ 'make docs'
[16:31:03] <cradek> no, "make docs" in the usual place
[16:31:16] <awalli1> ok, thanks
[16:31:31] <cradek> you have to configure --enable-build-documentation or however it's spelled
[16:32:57] <awalli1> make docs is a _little_ verbose...
[16:33:48] <cradek> heh just a little
[16:40:41] <awalli1> can I cross-reference from the integrator manual to the user manual?
[16:41:22] <awalli1> or how do I create a reference to another section of the manual anyway?
[16:42:38] <cradek> I think only alex and maybe ray really know how to use lyx
[16:42:42] <jepler> you can create references from one .lyx file to another, but it doesn't make sense to do so if they don't end up in the same pdf file
[16:43:02] <cradek> oh and maybe jepler :-)
[16:43:05] <awalli1> ok, but I see the HAL tutorial is also in the integrator manual, so maybe that's ok
[16:43:26] <jepler> in the same lyx window, load the document you want to refer to, then the document where you want to add the reference
[16:43:37] <awalli1> thanks, I got it.
[16:44:18] <jepler> then in Insert > Cross Reference, you can choose the "other" lyx file from the drop-down menu
[16:44:48] <awalli1> are the docs now built automatically also each night?
[16:45:14] <jepler> the 2.1 and HEAD docs are automatically built and uploaded to the website several times a day
[16:45:26] <awalli1> cool
[16:45:44] <awalli1> soon I will be answering all questions with RTFM!
[16:47:01] <skunkworks> :)
[16:47:16] <skunkworks> I am just glad I am in the inner circle....
[16:47:39] <skunkworks> right?
[16:48:09] <skunkworks> hello?
[16:48:24] <awalli1> hi
[16:48:26] <Dallur> there is no circle
[16:48:33] <jepler> .msg cradek can't we get rid of that skunkworks fellow?
[16:48:39] <jepler> .msg he's talking again
[16:49:34] <cradek> /msg jepler I can't figure out how to ban him
[16:50:39] <cradek> /msg jepler I tried to break the java client he uses but alex fixed it for him
[16:54:04] <skunkworks> :)
[16:54:28] <skunkworks> there is no circle - only short line segments
[17:00:44] <Dallur> skunkworks: There is circle but instead of variable radius circle there are only short line segments
[17:02:18] <awalli1> jepler: do you know how images (screencaps) work with lyx?
[17:02:50] <alex_joni> awalli1: how do you mean?
[17:03:14] <awalli1> alex_joni: where do I place the file, is .png ok?, how do I link to it from lyx.qt
[17:03:19] <awalli1> lyx-qt
[17:03:33] <alex_joni> usually we place the file into the same folder as the lyx
[17:03:37] <alex_joni> the image I mean
[17:03:52] <alex_joni> then for inserting you need to define a floating thingie, and insert the image into it
[17:04:09] <alex_joni> I find it usually easier to copy one that's already defined and simply edit that one
[17:04:45] <alex_joni> awalli1: in lyx-qt you have a menu option insert->image or the like
[17:06:59] <awalli1> trying...
[17:07:04] <alex_joni> awalli1: keep bugging me if it's not 100% clear :)
[17:08:08] <awalli1> oh, I need to experiment with the size % to get the image to fit on one page...
[17:09:21] <alex_joni> you have page-size in %
[17:09:27] <alex_joni> there are 2 different things
[17:09:31] <alex_joni> one for lyx preview
[17:09:38] <alex_joni> the other for output (aka pdf or html)
[17:10:13] <CIA-8> 03awallin 07HEAD * 10emc2/docs/src/hal/pyvcp_AXIS.png: screenshot of sim/lathe, i.e. AXIS with pyVCP
[17:13:00] <CIA-8> 03awallin 07HEAD * 10emc2/docs/src/hal/vcp.lyx: continue with pyVCP documentation
[17:13:14] <awalli1> alex_joni: how long before the online pdf/html is updated ?
[17:13:27] <alex_joni> that's jepler's doing.. usually 1-2 hours max
[17:13:35] <awalli1> ok, great!
[17:13:37] <alex_joni> but if you have the build docs, it updates on your system too
[17:13:46] <alex_joni> so you can check the result
[17:13:47] <awalli1> sure, I tried that
[17:13:51] <alex_joni> ok
[17:13:53] <awalli1> I need to go now, bbl
[17:13:59] <alex_joni> l8er awalli1
[17:14:37] <awallin> me too ;)
[17:19:59] <alex_joni> heh.. seems we have an emc2 user who wants teach-in :)
[17:21:21] <cradek> alex_joni: machine off, move the stuff, read the numbers on the screen, write gcode
[17:21:45] <alex_joni> yeah, but doing that 200 times is nerve-wrecking
[17:22:01] <alex_joni> cradek: it's for a robot
[17:22:19] <cradek> then use the hal python module to read the encoders and write the gcode line when you push a button
[17:22:38] <cradek> SMOP?
[17:22:40] <skunkworks> couldn't you modify the probing code?
[17:22:47] <skunkworks> have a teach button.
[17:23:08] <alex_joni> skunkworks: I thought about that too.. but it's not good
[17:23:21] <cradek> yeah, same idea, but you don't really need to use emc at all
[17:23:25] <alex_joni> because when you teach-in you don't know in advance in what direction the user will move
[17:23:35] <alex_joni> cradek's approach is a bit better
[17:23:51] <alex_joni> cradek: pyvcp & hal python module might do the trick
[17:24:03] <skunkworks> ah - I was thinking points - you guys want to capture movements?
[17:24:13] <cradek> I think all the tools are there, and pretty straightforward, for a cluser
[17:24:22] <alex_joni> cluser?
[17:24:26] <cradek> skunkworks: I was thinking points
[17:24:29] <alex_joni> skunkworks: no, points
[17:24:36] <alex_joni> but probing works with movement
[17:24:37] <cradek> it's a new word that's a combination of "clueful" and "user"
[17:24:54] <alex_joni> might work for clueless luser too
[17:24:56] <cradek> I think skunkworks came up with it
[17:25:05] <skunkworks> jepler I think
[17:25:20] <cradek> alex_joni: hmm, you're right it's a bit ambiguous
[17:25:27] <cradek> I still like it anyway
[17:25:30] <skunkworks> I am not that whitty ;)
[17:26:08] <alex_joni> yeah, it's nice
[17:27:32] <cradek> also usrmot can directly issue points (from an ascii file) by sending SET_LINE directly to realtime
[17:27:44] <cradek> no gcode writing needed
[17:27:55] <cradek> (not that the format is much simpler than G1)
[17:30:01] <skunkworks> informed user - infuser?
[17:30:23] <cradek> haha
[17:30:27] <alex_joni> skunkworks: leave the whitty things to others :P
[17:30:29] <cradek> I think that word is taken already
[17:30:38] <skunkworks> I would guess ;)
[17:32:01] <cradek> recently a coworker coined the term "extrabstraction" to be used where you really mean "another layer of crap on the code"
[17:32:35] <alex_joni> heh
[17:46:33] <tomp> the probe trips a switch, say the switch was in your hand, you watch the motion and press the button when robot is where you want, the position is capture by std probe code
[17:52:04] <alex_joni> tomp: how do you tell it what way to go for it to reach the position you want?
[17:54:05] <tomp> mdi
[17:55:17] <tomp> or program you give it a silly distance (too far) but stop it with switch ( like most probes) and collect the data
[17:55:30] <tomp> unreasonable?
[17:56:22] <tomp> i've done similar in polar on 5 axis machines
[17:57:20] <alex_joni> the problem is direction
[17:57:28] <alex_joni> remember this is 6 axes
[17:58:03] <alex_joni> I wouldn't be too confident to send it to X21 Y3 Z4.5 A17 B22 C31 and stopping it in the exact spot I need
[17:58:59] <tomp> is the problem the kinematics (motion along a vector) or stopping exactly or other
[18:00:23] <tomp> is teaching that exact ( what i saw the robot guys doing was all eyeball and 'close enug to grab it' )
[18:00:32] <tomp> enuf
[18:04:38] <alex_joni> you move one axis, then the other, so on
[18:04:47] <alex_joni> don't see how you can do that with probing calls
[18:05:32] <alex_joni> SWPadnos: err.. here
[18:05:53] <alex_joni> odd way to name a festival: http://www.sperm.cz/
[18:06:39] <SWPadnos> heh
[18:10:49] <tomp> ok, so emc probing is single axis: that wont work
[18:15:16] <mdynac> good afternoon, all
[18:16:17] <lerneaen_hydra> alex_joni: almost like www.pungkuk.com
[18:17:54] <lerneaen_hydra> pungkuk is basically scrotum-dick in swedish
[18:23:46] <tomp> where is probing documented?
[18:24:26] <lerneaen_hydra> a german porno site?
[18:24:32] <lerneaen_hydra> :p
[18:24:38] <lerneaen_hydra> sorry
[18:40:56] <tomp> EMC2_user_manual.pdf says g38.2 straight probing isnt implemented in emc2... it doesnt exist (reference http://linuxcnc.org/EMC2_User_Manual.pdf linked from http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?EmcKnowledgeBase)
[18:41:25] <tomp> has probing been implented in some other way in emc2?
[18:43:29] <lerneaen_hydra> * lerneaen_hydra pokes alex_joni
[18:43:40] <lerneaen_hydra> if you're lucky alex will know ;)
[18:48:21] <CIA-8> 03rayh 07HEAD * 10emc2/configs/m5i20/m5i20_io.hal: comment out spindle speed up and down
[18:55:45] <slundell> Hi all!
[19:05:46] <lerneaen_hydra> slundell: you're not swedish by any chance?
[19:05:46] <lerneaen_hydra> (your IP and name seem to hint at that)
[19:06:08] <slundell> yes, im the artist formally known as simon78
[19:06:16] <slundell> regged my nick
[19:09:02] <lerneaen_hydra> god kväll god kväll
[19:09:15] <lerneaen_hydra> en hel del svenskar här, en till kommer tom nyu
[19:09:18] <lerneaen_hydra> *nu
[19:09:23] <^eugenics> hej
[19:09:34] <lerneaen_hydra> ^eugenics: another one to our cause
[19:09:50] <lerneaen_hydra> not long before we can conquer #emc and control it
[19:09:52] <^eugenics> wats up?
[19:10:17] <^eugenics> :) i think only enhansing
[19:11:35] <lerneaen_hydra> another swede here ;)
[19:12:26] <awallin> any idea of how to get two columns in lyx?
[19:13:00] <awallin> or, alternatively, how to insert multiple lines of text/graphics into a table cell
[19:13:01] <^eugenics> another xterm ;)
[19:13:10] <jepler> awallin: beats me
[19:14:59] <maddash> SWPadnos , are you here?
[19:15:04] <SWPadnos> nope
[19:15:06] <SWPadnos> err -yewp
[19:15:08] <maddash> ok.
[19:15:08] <SWPadnos> yep
[19:15:12] <rayh> Hi maddash. Could you describe what you are trying to do.
[19:15:18] <maddash> ok
[19:15:28] <rayh> and I'll read for a bit.
[19:15:59] <maddash> my aim is to give out the signal pulses at a fast rate. I assume that the internal processing of the gcode (linerate) is directly related to the speed of the pulses (pulserate). am I correct?
[19:16:24] <SWPadnos> yes and no
[19:16:35] <SWPadnos> that's why I suggested using sim instead of stepper_*
[19:16:44] <maddash> elucidate, if you don't mind
[19:17:03] <SWPadnos> one sec - kepp on going with what you want to do so I don't sidetrack you
[19:17:23] <rayh> pulse rate is feedrate/60 * input_scale
[19:18:20] <CIA-8> 03awallin 07HEAD * 10emc2/docs/src/hal/pyvcp_mypanel.png: first pyvcp panel screenshot
[19:20:13] <maddash> so when I used mini, I assumed that the backplot and the cmd box below it were updated in the same intervals. the "smoothness" of the display then varies on the polling time (pollrate). so if I have a high linerate (=pulserate?) and a low pollrate, then the backplot and cmd box would both be "jumpy", right?
[19:20:42] <maddash> ie, the cmd box would jump from line 1 to, say, line 15.
[19:20:51] <SWPadnos> yep, prettry much
[19:20:56] <SWPadnos> -r
[19:20:59] <cradek> at each poll, the program listing will show the line that's currently causing the motion
[19:21:06] <maddash> right.
[19:21:23] <SWPadnos> or the line that was causing the motion at the last time the status struct was passed to userspace (which may be behind a little)
[19:21:37] <cradek> note that this means it MAY skip lines even if the poll is very fast. Many lines don't cause motion at all.
[19:22:32] <rayh> Right. Mini's display is driven by what is happening in EMC each time around mini's loop.
[19:22:55] <maddash> so I judge the pulserate/linerate via the position of the "jumps" (or skips, as cradek says). and I could also calculate an approximation of the linerate: linerate = (jump_1 - jump_0)/(time_1 - time_0)
[19:23:16] <cradek> I don't understand what you mean by linerate
[19:23:25] <maddash> jump_1 may not necessarily = jump_0, b/c the display is "jumpy"
[19:23:31] <cradek> a particluar line can take anywhere from zero to "a long" time to run
[19:23:52] <maddash> cradek: speed of internal processing of a given gcode line.
[19:24:02] <awallin> * awallin is rebooting... bbl
[19:24:18] <cradek> the highlight line does not follow the gcode interpretation - it follows the machine's motion, which is smooth
[19:24:34] <maddash> "highlight line"?
[19:24:36] <cradek> the gcode interpretations runs (far) ahead of the machine
[19:25:04] <cradek> yeah, however mini shows it, it highlights the line that's responsible for the current motion
[19:25:25] <maddash> cradek, let me rephrase: linerate = the rate at which the machine takes the motion pulses of a corresponding gcode line
[19:25:54] <cradek> I still don't know what you're asking
[19:26:01] <SWPadnos> here's what will limit the execution speed of a single line of G-code (which causes motion):
[19:26:02] <jepler> do you mean the feed rate (The rate at which the tool tip moves)?
[19:26:05] <maddash> how could I increase the linerate?
[19:26:31] <maddash> jepler: not feedrate. linerate... b/c right now, feedrate >> linerate.
[19:26:44] <SWPadnos> 1) if you're actually outputting pulses to a parallel port, then the rate at which you can generate pulses will be a limit
[19:27:00] <maddash> jepler, ie, the tool tip could move faster, but it's not getting its cmds fast enough from emc2.
[19:27:08] <cradek> why do you think that's the case?
[19:27:40] <maddash> cradek, you mean, why feedrate >> linerate?
[19:27:45] <cradek> you're simulating a machine with mass, maximum velocities, and maximum accelerations.
[19:27:53] <jepler> so the units of what you are calling "linerate" is distance per second?
[19:27:55] <cradek> and in your case, step generation
[19:28:05] <SWPadnos> emc2 can process thousands of lines of G-code per second - the interpretation time is miniscule compared to the time it takes an actual machine to move
[19:28:14] <cradek> right, what SWPadnos says
[19:28:30] <maddash> jepler, linerate = (lines of gcode converted to pulses and sent to the machine) per sec
[19:28:50] <cradek> I doubt that is your limitation unless your gcode is very perverse.
[19:28:57] <SWPadnos> I loaded a 66000-line (roughly) G-code file into axis, and the preview was up on screen within 15 seconds
[19:29:07] <SWPadnos> that includes generating all the openGL line segments
[19:29:33] <SWPadnos> "converting to pulses and sending to the machine" is what takes the time
[19:30:04] <SWPadnos> actually, "secnding pulses to the machine" is what causes slowdowns, there is no "conversion to pulses" per se
[19:30:44] <cradek> if you want to see the program preview, use AXIS. If you really want to simulate an imaginary machine that runs at light speed, use sim -- not stepper
[19:31:35] <rayh> but if what you want to do is plan for a real machine running steppers, then what you are doing is about right.
[19:32:08] <cradek> if you're planning for a real machine, "line rate" means nothing
[19:32:11] <rayh> and we just need to get your ini setup for number of steps per inch and max velocity and accel.
[19:32:35] <SWPadnos> you can speed things up a lot by changing the INPUT_SCALE to 100 or something
[19:32:56] <cradek> I still don't understand the goal.
[19:33:07] <SWPadnos> that gives 0.010 resolution, but outputs fewer pulses per unit of motion
[19:33:32] <SWPadnos> (it won't be accurate though, since small moves would cause no output at this resolution, but might cause output at a higher resolution)
[19:33:44] <cradek> that's why to use sim
[19:33:57] <maddash> ok, I'll try sim.
[19:34:13] <SWPadnos> right, unless you're hooking the thing up to something (like LEDs) to see how it interfaces to the world
[19:34:51] <rayh> If all blocks of g-code were the same length then the speed of lines going by would be related to the feedrate.
[19:34:59] <maddash> erm, looks like sim depends on minigl as well; how do I compile it?
[19:35:11] <cradek> brb
[19:35:21] <maddash> actually, *where* do I compile it?
[19:35:22] <SWPadnos> apt-get build-dep emc 2emc2-axis ;)
[19:35:34] <SWPadnos> oops: apt-get build-dep emc2 emc2-axis
[19:35:39] <jepler> maddash: sim/tkemc and sim/mini do not use axis
[19:36:05] <SWPadnos> I assume that this 3GHz Pendium D isn't resource constrained ...
[19:37:58] <maddash> SWPadnos, hardly, it's running a barebones linux-2.6.17-rtai.
[19:39:23] <maddash> ok, so most of my program right now belongs to a gigantic "G2 x... y... r..." chunk.
[19:39:42] <maddash> but even in sim/mini, it's going from line to line rather slowly
[19:40:00] <maddash> line to line of "x[blah] y[blah] r[blah]", that is.
[19:40:06] <cradek> that's because it's running the (pretend) machine according to its acceleration and velocity constraints.
[19:40:12] <SWPadnos> that is a function of programmed feedrate and ini limits
[19:40:21] <cradek> yes, and the feed rate
[19:40:25] <cradek> whichever is least
[19:40:26] <maddash> so setting max_vel and max_accel should do it, then
[19:40:36] <SWPadnos> change the ini limits (MAX_*) to 100x their current values, and make MAX_OVERRIDE 100 or so
[19:40:57] <SWPadnos> then you'll be able to get a simulated machine that can accelerate at 10G's
[19:41:02] <SWPadnos> and run 100 MPH
[19:41:34] <SWPadnos> you will, however, see almost nothing if you try to view a backplot
[19:41:36] <cradek> can you say again what your goal is in doing that? it's a very bad simulation of a machine then...
[19:41:40] <SWPadnos> or at least nothing useful
[19:41:52] <maddash> do I have to chg the max_* of all the axes, or just the 2 (x,y) I'll use?
[19:41:56] <cradek> I'm not trying to be difficult, but I don't understand what you're trying to accomplish
[19:42:03] <jepler> good g-code will have segments long enough to have a "cruise phase": starting from rest, the machine can reach the requested velocity and stop again, while respecting the acceleration limit.
[19:42:06] <SWPadnos> the two you'll use plus in [TRAJ]
[19:42:19] <cradek> (it's hard to give the right advice then)
[19:42:52] <maddash> cradek, I'm trying to configure sim/mini so that the max_* values are actually *close* to the machine's limits. right now, they're (max_* values) far below...
[19:43:12] <cradek> ok now we're getting somewhere
[19:43:16] <jepler> with any real machine, the time taken for such a line will be much longer than the maximum count of primitive motions per second (the so-called "traj period" is the limit here)
[19:43:21] <cradek> do you know the max velocity and acceleration for your machine?
[19:43:30] <SWPadnos> configuring it so that it matches a physical machine will cause the G-code to execute as slolwy as if there were a machine attached ...
[19:45:28] <rayh> maddash: How did you produce the g-code program you are trying to view?
[19:45:56] <maddash> rayh, autocad.
[19:46:28] <maddash> rayh, but autocad outputs G2 in "xyz ijk" format, so I had to convert to "xyz r" format
[19:46:42] <cradek> xyz ijk is preferable, emc takes both
[19:46:44] <jepler> emc2 accepts "xyz ijk" format
[19:46:52] <maddash> cradek, ah ok.
[19:47:23] <cradek> what gcode generator are you using in autocad?
[19:47:27] <cradek> just curious
[19:47:41] <rayh> me to -- curious
[19:50:27] <paragon36> Hello Everyone... :-)
[19:50:40] <lerneaen_hydra> g'day
[19:51:46] <eholmgren> mate
[19:52:32] <lerneaen_hydra> crickey!
[19:53:37] <tomp> awallin: <Shift-Button-4> and <Control-Button-4>could be bound as speed ranges for jogwheel (and similarly for Button-5) (! or for x y z hmmm)
[19:54:20] <awallin> tomp: a good idea, I'm working on documentation now... but add any wishes to the wiki page
[19:54:29] <tomp> k
[19:56:39] <rayh> awallin: I like the tinyvcp pic.
[19:57:49] <awallin> which one ?
[20:01:03] <rayh> The one you just put in docs.
[20:01:55] <awallin> ha :) well it's a simple example...
[20:03:47] <rayh> Great stuff you're doing with the system.
[20:04:18] <awallin> thanks, it's mostly cosmetic you know, but sometimes that's needed too
[20:05:52] <robin_sz> meep?
[20:07:13] <lerneaen_hydra> meep
[20:15:57] <anonimasu> m33p
[20:21:46] <eholmgren> m00p
[20:26:23] <maddash> cradek, rayh: "<cradek> what gcode generator are you using in autocad?
[20:26:23] <maddash> <cradek> just curious" --scratch that - I don't know what I was thinking - I use powerstation
[20:26:45] <cradek> oh ok
[20:28:17] <maddash> cradek: just curious how long does it take your sim/mini to execute http://pastebin.ca/313243 ?
[20:28:45] <maddash> cradek, I had to crank up the max_* numbers to 1000x their original values just to get 10sec.
[20:29:11] <cradek> the program is only F10
[20:30:24] <anonimasu> maddash: why dont you run axis?
[20:30:36] <anonimasu> maddash: The preview is 9000 times better..
[20:30:43] <anonimasu> and you dont need to run the program to get a plot
[20:31:14] <maddash> anonimasu, wasn't tryng to preview the code itself, was trying to preview the machine performance
[20:31:30] <anonimasu> preview?
[20:31:37] <maddash> anonimasu, to rephrase: wanted to simulate the process, not the result.
[20:32:38] <anonimasu> well, im not pushing you to axis or anything.. but it's really a very nice way to preview/simulate/whatever
[20:32:54] <maddash> anonimasu, well, axis also uses more resources...
[20:33:15] <maddash> anonimasu, but that only matters when I use a 500mhz w/192 ram, or if I'm in OCD mode...
[20:34:20] <lerneaen_hydra> hey, neat effect, if an IRC user can't be reached for a while, but then later comes back all the messages will nicely be added
[20:34:48] <anonimasu> maddash: well, have fun ^_^
[20:35:23] <cradek> maddash: this file has 80 inches of cutting
[20:35:51] <anonimasu> hm 80/25 = time
[20:35:53] <anonimasu> right?
[20:35:58] <anonimasu> 3.2 minutes
[20:36:24] <cradek> feed is 10 inch/min, so it will take about 8 minutes
[20:36:26] <jepler> F10, not F25
[20:36:31] <anonimasu> ah
[20:37:19] <maddash> cradek, ah, forgot about that
[20:37:46] <maddash> cradek, no wait, I didn't - I cranked up to F10
[20:37:55] <maddash> cradek, F500*
[20:37:57] <cradek> I doubt your machine will go 80 inches in 10 seconds
[20:38:10] <maddash> it's gotta do it in 7 sec.
[20:38:14] <anonimasu> 203mm/s
[20:38:19] <anonimasu> not going to happen
[20:38:32] <cradek> what kind of machine?
[20:38:38] <maddash> well, for starters, I'm not milling, or cutting anything
[20:38:50] <anonimasu> maddash: just curious
[20:38:55] <anonimasu> what spindle will you use for cutting?
[20:39:13] <cradek> that's 700 inches/minute
[20:39:16] <anonimasu> when you will be cutting..
[20:39:32] <cradek> and the acceleration will have to be very high because it's not a straight line
[20:39:41] <maddash> anonimasu, no spindle...
[20:39:44] <tomp> maddash: you're just moving
[20:39:58] <anonimasu> maddash: in the future..
[20:40:10] <maddash> indeed. the motors will be positioning a liquid-dispensing nozzle
[20:40:13] <anonimasu> ?
[20:40:17] <anonimasu> ah..
[20:40:37] <lerneaen_hydra> so you're decorating cakes at a furious pace?
[20:40:47] <maddash> LOL
[20:40:57] <maddash> no, not really...
[20:42:41] <maddash> oh god, I posted up the wrong gcode.
[20:42:46] <anonimasu> maddash: how many pulses/sec is that?
[20:43:03] <lerneaen_hydra> you're going to need very interesting hardware
[20:43:13] <anonimasu> yeah..
[20:43:16] <lerneaen_hydra> 2m/s, and what resolution?
[20:43:16] <anonimasu> that's where I were going
[20:43:37] <lerneaen_hydra> if it's coase (1mm or so) I guess it wouldn't be that hard
[20:43:41] <eholmgren> rapid prototyping?
[20:43:42] <lerneaen_hydra> *coarse
[20:43:49] <cradek> I can get it to run in 2-3 secs if I turn up the numbers to "silly"
[20:44:09] <maddash> nice way of putting it
[20:44:13] <cradek> I could try "sillier" if you want
[20:44:31] <anonimasu> cradek: did you turn on the blending?
[20:44:46] <maddash> cradek, just curious, how'd you find the length?
[20:45:00] <cradek> File/Properties in AXIS
[20:46:05] <jepler> it also provides an estimated time but the estimation algorithm is very naive (distance / requested feed rate)
[20:46:32] <lerneaen_hydra> jepler: does it take G0 feedrate into account (maxvel)
[20:46:48] <CIA-8> 03awallin 07HEAD * 10emc2/docs/src/hal/ (4 files): pyvcp widget screenshots
[20:47:04] <cradek> maddash: http://timeguy.com/cradek-files/emc/maddash.png
[20:47:23] <cradek> (I removed the first and last G0 lines which were silly)
[20:47:26] <CIA-8> 03awallin 07HEAD * 10emc2/docs/src/hal/vcp.lyx: document LED, Button, Checkbutton, Radiobutton
[20:48:45] <cradek> I changed it to F700 which gives approximately 7sec runtime like we calculated
[20:49:18] <jepler> http://emergent.unpy.net/files/sandbox/velocity-plot.png http://emergent.unpy.net/files/sandbox/maddash.ini
[20:49:23] <maddash> holy crap.
[20:49:26] <awallin> cradek: with 7 seconds of run time, you would get 7000 traj periods? so the geometry is sampled with 7000 points ?
[20:49:39] <cradek> yes something like that
[20:49:56] <jepler> here's the velocity vs time plot when running with that .ini file and maddash's code with modified feed rate
[20:49:57] <lerneaen_hydra> aliasing nastyness in other words
[20:50:36] <awallin> lerneaen_hydra: depending on the geometry, more or less aliasing I guess...
[20:50:36] <cradek> * cradek shrugs
[20:50:47] <cradek> a millisecond is a long time at 700 inches/min
[20:51:16] <maddash> thank god for pentium d
[20:51:19] <cradek> typically you have machine mass to help you out
[20:51:24] <awallin> is it possible to stream motor-pos-cmd to a file or memory buffer, and look at it afterwards ?
[20:51:31] <SWPadnos> yes
[20:51:33] <maddash> yes
[20:51:36] <SWPadnos> halstreamer should do that
[20:51:36] <cradek> that's what halscope is for
[20:51:39] <jepler> awallin: mad halsampler
[20:51:43] <jepler> er, man halsampler
[20:51:53] <SWPadnos> right - sampler
[20:51:56] <jepler> You have: 80 inches * (4000 steps / inch) / 7 seconds
[20:51:54] <jepler> You want: steps / second
[20:51:54] <jepler> * 45714.286
[20:52:12] <jepler> even with a modest steps per inch, it looks like the step rate is tough to achieve with software stepgen
[20:52:45] <jepler> BASE_PERIOD of about 10uS
[20:52:58] <awallin> halsampler data would be interesting to plot on top of the original geometry, for those crazy vel/acc settings
[20:54:12] <cradek> the acceleration on these arcs must be about 300 inch/sec^2 or .8G
[20:54:23] <maddash> erm. I uploaded the wrong gcode source - the one from pastebin is from an old project...the real source is about 14 inches in total
[20:54:35] <awallin> I read somewhere that HSM is defined as starting at >1g
[20:54:58] <cradek> well EMC can command it, but getting it into a machine isn't so simple
[20:56:57] <awallin> jepler: what part of the manual is in HTML?
[20:57:06] <awallin> I don't see everything in HTML
[20:58:43] <maddash> jepler, what're the emergent properties of meat?
[20:58:48] <cradek> bbl
[21:07:16] <anonimasu> maddash: that's really insanely fast.
[21:07:48] <anonimasu> maddash: I'd love to see/hear about your hardware
[21:08:30] <maddash> classified information. what's your access code? lol.
[21:08:58] <anonimasu> heh, in other words you have no idea?
[21:09:14] <jepler> awallin: all the lyx stuff should be converted to html too
[21:09:40] <tomp> maddash: do you need that pattern: or might a spiral be ok ( faster, like threading and 2 axis X A )
[21:09:42] <jepler> awallin: I have to add some rules whenever a new lyx file is done (well, someone has to)
[21:10:00] <maddash> anonimasu, exactly.
[21:10:02] <jepler> by "rules" I mean lines in the docs/src/Submakefile
[21:10:47] <maddash> tomp, I don't understand the phrase in the parentheses.
[21:11:02] <tomp> like the groove on a record
[21:11:44] <jepler> maddash: "Emergence is a term used in Philosophy, Systems Theory and the Sciences to describe the development of complex organized systems." (wikipedia). I am an emergent property of the meat that constitutes my body
[21:12:17] <cradek> yuck, you're made of meat?
[21:12:24] <awallin> jepler: ok, I've now added vcp.lyx to the integrator handbook, so you might generate HTML from vcp.lyx also
[21:12:42] <jepler> awallin: OK I'll do that shortly
[21:13:25] <tomp> madddash: to emc, it could be like threading, because it could become an interpolation of a rotary and a linear motion
[21:14:04] <maddash> jepler, I know what emergence is (wikipedia fails to include sociology!), I'm just trying to figure out a) what kind of meat and b) what kind of props
[21:14:18] <tomp> maddash: this would evenly disperse the liquid over a disk
[21:14:35] <maddash> tomp, huh?
[21:14:35] <cradek> maddash: I suspect jepler is made of nearly the same kind of meat you are
[21:15:28] <maddash> cradek, didn't realize "meat" referred to "human flesh"...in fact, seeing the bird, I was thinking of 2 things: SARS and bird meat...
[21:15:47] <skunkworks> <mib>meat sack - monkey boy</mib>
[21:16:08] <jepler> maddash: http://baetzler.de/humor/meat_beings.html
[21:16:18] <cradek> I'm made of gleaming stainless steel, but I think the rest of you are definitely made of meat
[21:16:38] <cradek> for meat, you show very complex behaviors
[21:16:48] <CIA-8> 03jepler 07HEAD * 10emc2/docs/src/ (Submakefile index.tmpl): add vcp to list of documentation to be converted to html
[21:16:59] <maddash> tomp, linear and arc motion can't happen simultaneously. and what makes you think that project dispersed liquid over a disk?
[21:17:20] <tomp> if the great juju hadn't meant us to eat people, he wouldn't have made us of meat ( Flanders & Swan )
[21:17:43] <maddash> "great juju"?
[21:18:06] <tomp> maddash:you said " indeed. the motors will be positioning a liquid-dispensing nozzl"
[21:18:34] <tomp> flanders and swan, 1950's 'at the drop of a hat' on broadway for 15 years
[21:18:53] <maddash> tomp, yeah, I don't know what I was thinking there - just quoted the writeup for that project verbatim.
[21:18:55] <jepler> awallin: http://linuxcnc.org/docs/devel/html/hal/vcp/index.html
[21:19:16] <Jymmmm> Where's the irc iface on the website?
[21:19:53] <maddash> jepler, those aliens are horrifically dumb. how'd they ever make it to space?
[21:20:09] <tomp> http://www.members.optusnet.com.au/pennywyatt/Interests/FlandersSwann/DropOfaHat/At%20the%20Drop%20of%20a%20Hat09.html
[21:20:47] <jepler> maddash: I take it you now regret asking why my blog is called that.
[21:21:34] <maddash> jepler, almost. fortunately, you didn't give a precise definition of "meat," so almost anything can be "meat."
[21:25:19] <skunkworks> jepler: I figured it had something to do with you being a vegitarian and left it at that..
[21:25:32] <skunkworks> guess I was wrong :)
[21:25:44] <Jymmmm> jepler is a vegan??? Oh, that explains EVERYTHING now.
[21:25:51] <cradek> so are you all ashamed to be made of meat, or didn't you realize it until now?
[21:26:06] <cradek> I mean come on
[21:26:21] <Jymmmm> <cradek> I meat come on
[21:26:27] <Jymmmm> =)
[21:28:29] <alex_joni> * alex_joni just came back
[21:28:37] <alex_joni> only to find disturbing things in here
[21:28:38] <alex_joni> talking meat
[21:28:42] <Jymmmm> * Jymmmm invites alex_joni over for dinner
[21:29:01] <skunkworks> * skunkworks goes to timeguy.com to get to jeffs site ;)
[21:29:48] <lerneaen_hydra> wtf? I leave this channel for one minute and look what happens
[21:29:52] <alex_joni> emergent.unpy.net ?
[21:30:04] <Jymmmm> lerneaen_hydra then don't leave... DUH!
[21:33:24] <Jymmmm> * Jymmmm ponders if 2" conduit will work for an antenna mast 20' tall
[21:33:44] <Jymmmm> w/o a welder that is
[21:36:46] <alex_joni> awallin: small typo in the docs for pyvcp
[21:36:51] <alex_joni> http://www.linuxcnc.org/docs/devel/html/hal/vcp/index.html
[21:36:58] <alex_joni> <button>
[21:36:59] <alex_joni> [ <halpin>''my-button'' [ <text>''ON'' </button>
[21:37:14] <alex_joni> the picture of the button actually reads 'OK' not 'ON'
[21:37:45] <Jymmmm> Where's the irc iface on the website?
[21:38:17] <cradek> 'where to ask for help'
[21:38:19] <alex_joni> http://www.linuxcnc.org/index.php?option=com_mospjirc&Itemid=8&lang=en
[21:38:20] <EldonB46> Jymmmm: Do you have two 10' sections of 2" conduit, on one 20' section and what is the wind load of the antenna?
[21:38:21] <Jymmmm> Jymmmm is now known as Jymmm
[21:38:56] <alex_joni> awallin: I also notice some/most of the examples are quoted as {} statements, not XML markup
[21:39:12] <Jymmm> EldonB46 2 10' pieces (going to buy today), The wind load is minimal, it's a discone antenna
[21:39:31] <Jymmm> EldonB46 Just not sure if I trust the couplers
[21:39:47] <Jymmm> as they're cast
[21:39:50] <EldonB46> I wouldn't
[21:40:09] <Jymmm> What I'd REALLY liek to do is a pushup using condiut
[21:40:09] <skunkworks> uncle welds up his own rone(sp) style towers.
[21:40:32] <alex_joni> awallin: sorry.. didn't see that half the docs refer to VCP and half to pyVCP, that's confusing
[21:41:18] <lerneaen_hydra> um, what is the main difference between VCP and pyVCP? python addons?
[21:41:20] <Jymmm> skunkworks: Yeah, no welder here. And never have welded, especially galvanized
[21:42:37] <alex_joni> lerneaen_hydra: VCP is gtk, pyVCP is python, VCP has 3-4 widgets, pyVCP has a 10
[21:42:46] <lerneaen_hydra> ah, I see
[21:42:52] <alex_joni> pyVCP can be run standalone or integrated with AXIS, VCP only alone
[21:43:09] <lerneaen_hydra> does python have its own gui toolkit?
[21:43:19] <jepler> lerneaen_hydra: you have your choice of many
[21:43:21] <Jymmm> EldonB46 You think 2" + 1.5" conduit would be ok, or could it buckle?
[21:43:31] <Jymmm> telescoping that is
[21:43:32] <jepler> lerneaen_hydra: tk, gtk, qt, wx, and probably two dozen others you've never heard of
[21:43:41] <lerneaen_hydra> jepler: oh, so probably gtk, qt, wx and several more?
[21:43:44] <jepler> java swing for all I know
[21:43:47] <lerneaen_hydra> right ;)
[21:44:04] <Jymmm> jepler are you vegan, or were they just teasing you bout that?
[21:44:41] <jepler> Jymmm: I'm vegetarian, but not vegan (I eat dairy and eggs)
[21:44:42] <alex_joni> jepler: jpype all the way
[21:44:49] <Jymmm> jepler Ah, ok.
[21:44:49] <EldonB46> Jymmm: use at least a 2' overlap, and some guy wire at the 10' and say 16' level?
[21:44:59] <lerneaen_hydra> * lerneaen_hydra gasp
[21:45:07] <lerneaen_hydra> jepler is plantflesh!
[21:45:26] <Jymmm> EldonB46 The roofline is 13' so I'm just trying to get maybe 6' above that.
[21:45:52] <robin_sz> lerneaen_hydra, careful now, there are more vegetarians around than you might think! ;)
[21:46:09] <lerneaen_hydra> it's a conspiracy! the plants are trying to take over the world!
[21:46:12] <lerneaen_hydra> * lerneaen_hydra dies
[21:46:17] <Jymmm> * Jymmm makes reservations for House of Prime Rib!!!
[21:46:28] <Jymmm> SWPadnos COME ONE LET's GO!!!
[21:46:33] <Jymmm> -e
[21:46:36] <skunkworks> seeds of doom - dr. who circa late 70's
[21:46:39] <lerneaen_hydra> lerneaen_hydra is now known as lerneaen_hydra_p
[21:46:49] <lerneaen_hydra_p> lerneaen_hydra_p is now known as hydra_posthumous
[21:47:06] <hydra_posthumous> * hydra_posthumous comes with Jymmm and SWPadnos
[21:47:28] <alex_joni> where 'ya goin' ?
[21:47:30] <Jymmm> hydra_posthumous http://houseofprimerib.citysearch.com/
[21:47:43] <EldonB46> Jymmm: what is the Diskcone tuned to (or how long are the cone elements) This is going to relate to CNC ans you "may" need to machine a slip joint for the two conduits ;)
[21:47:50] <hydra_posthumous> mmm... yummy!
[21:48:06] <hydra_posthumous> alex_joni: to that place to get some life into me ;)
[21:48:17] <Jymmm> http://houseofprimerib.citysearch.com/page/o2oz/The_Prime_Rib_Cart.html
[21:48:23] <alex_joni> * alex_joni will have some King Henry VIII Cut
[21:48:33] <Jymmm> what alex_joni said
[21:48:41] <maddash> ugh. food.
[21:49:00] <hydra_posthumous> we all know maddash prefers cake
[21:49:02] <Jymmm> EldonB46 http://www.eham.net/reviews/detail/2777
[21:49:28] <hydra_posthumous> from his cake-decorating machine that decorates cakes faster than you can blink
[21:50:35] <maddash> hydra_posthumous, grrrr it's not cake!!!
[21:50:34] <robin_sz> there ar 5 houses in the courtyard where I live, and purely by chance, there are more vegetarians than meat eaters here
[21:50:44] <Jymmm> EldonB46: I'll take some pics later when I start assembling the antenna. I ahvne't seen very many details pics on the web.
[21:50:54] <robin_sz> robin_sz is now known as lerneaen_hydra
[21:50:55] <hydra_posthumous> maddash: oh shush, we know all about your plans. ;)
[21:50:57] <alex_joni> robin_sz: thought you lived at a farm
[21:51:04] <hydra_posthumous> * hydra_posthumous gasp
[21:51:05] <alex_joni> alex_joni is now known as robin_sz
[21:51:05] <lerneaen_hydra> ooh ..l I feel all weird ...
[21:51:14] <robin_sz> meep?
[21:51:14] <maddash> hydra_posthumous, yes, to take over the world, one cake at a time. oops, did I just type that?
[21:51:22] <hydra_posthumous> hydra_posthumous is now known as alex_joni
[21:51:29] <Jymmm> lol
[21:51:31] <maddash> er...
[21:51:34] <maddash> lol
[21:51:40] <maddash> LOL
[21:51:39] <robin_sz> :P
[21:51:44] <robin_sz> robin_sz is now known as alex_joni
[21:51:49] <maddash> seriously, how'd that happen?
[21:51:56] <lerneaen_hydra> what?
[21:52:07] <Jymmm> Jymmm is now known as Jymmmmmmm
[21:52:07] <alex_joni> maddash: I have an identified nick
[21:52:13] <alex_joni> lerneaen_hydra__: :P
[21:52:22] <maddash> boo, tyrannical meat
[21:52:28] <alex_joni> flapping meat
[21:52:47] <lerneaen_hydra> alex_joni, I think I saw a video about flapping meat once
[21:52:51] <maddash> bah, I get my energy straight from a 110V AC socket.
[21:52:53] <alex_joni> there's a story
[21:52:59] <lerneaen_hydra__> aw, alecx had to spoil the fun
[21:53:00] <lerneaen_hydra__> hmm, this is interesting, is this alex?
[21:53:00] <maddash> food is for cavemen.
[21:53:02] <lerneaen_hydra__> or someone else?
[21:53:03] <Jymmmmmmm> maddash wuss
[21:53:06] <lerneaen_hydra__> ack, this is very interesting
[21:53:07] <Jymmmmmmm> Jymmmmmmm is now known as Jymmm
[21:53:11] <lerneaen_hydra__> cakes however...
[21:53:18] <alex_joni> lerneaen_hydra__: you'll never know :)
[21:53:25] <lerneaen_hydra__> * lerneaen_hydra__ gasp
[21:53:25] <lerneaen_hydra> lerneaen_hydra is now known as Jymmmm
[21:53:29] <lerneaen_hydra__> oh now!
[21:53:31] <lerneaen_hydra__> lerneaen_hydra__ is now known as lerneaen_hydra
[21:53:48] <maddash> what's a learning hydra/
[21:53:56] <alex_joni> ROFL
[21:54:07] <alex_joni> a not_learning_hydra
[21:54:09] <Jymmmm> is this like chess ? spot an unused nick, jump onto it?
[21:54:13] <lerneaen_hydra> a lerarning hydra is a hydra that's learning stuff
[21:54:16] <lerneaen_hydra> err
[21:54:16] <maddash> well, I was thinking phonetically
[21:54:20] <lerneaen_hydra> *learning
[21:54:26] <alex_joni> sounds german
[21:54:41] <EldonB46> Jymmm: Maybe, consider using three section to get to 20' plus, I would suggest packing (or slip joint) for between each section and guy wires. discones are great antennas. - Have fun!
[21:54:47] <Jymmmm> maddash, phonetically? something to do with phones?
[21:55:12] <maddash> Jymmm, what's a phone?
[21:55:17] <lerneaen_hydra> maddash: roman mythological creature
[21:55:18] <alex_joni> handsfree
[21:56:01] <Jymmmm> a discone is a special antennae that gives a poor match at all frequencies
[21:56:02] <maddash> maddash is now known as lernaean_hydra
[21:56:20] <lerneaen_hydra> aha, you imposter!
[21:56:20] <lernaean_hydra> :P that's the way to spell it
[21:56:37] <lernaean_hydra> hah! says you, my mispelled counterpart!
[21:56:51] <lerneaen_hydra> I know, I introduced the misspelling so when searching on the name I'd get my results rather than 500 000 pages about the real thing ;)
[21:56:56] <alex_joni> all I'm reading is learning something
[21:57:08] <lernaean_hydra> lernaean_hydra is now known as maddash
[21:57:12] <lerneaen_hydra> ... wth?
[21:57:17] <maddash> maddash is now known as sudo_maddash
[21:57:38] <sudo_maddash> suddenly, I feel...more privileged...
[21:57:39] <alex_joni> ooh.. supercow powers
[21:57:40] <Jymmmm> Jymmmm is now known as maddassh_
[21:57:55] <sudo_maddash> haha aptitude has super cow powers
[21:57:57] <lerneaen_hydra> oh, noes, what have I done to #emc?
[21:58:10] <sudo_maddash> whoa.
[21:58:11] <lerneaen_hydra> * lerneaen_hydra dies
[21:58:16] <alex_joni> sudo_maddash: too late
[21:58:17] <alex_joni> :P
[21:58:20] <alex_joni> you had your chance
[21:58:21] <lerneaen_hydra> lerneaen_hydra is now known as hydra_posthumous
[21:58:21] <sudo_maddash> haha.
[21:58:31] <maddassh_> maddassh_ is now known as robin_sz
[21:58:35] <hydra_posthumous> oh no, dead again
[21:59:09] <hydra_posthumous> I find it interesting to note that cradek and jepler have been very very quiet since the start of this... insanity
[21:59:22] <hydra_posthumous> (they're probably hunting wabbits!)
[21:59:25] <hydra_posthumous> so shush!
[21:59:31] <skunkworks> skunkworks is now known as credak
[21:59:38] <credak> :)
[21:59:42] <hydra_posthumous> * hydra_posthumous *snerk*
[21:59:43] <sudo_maddash> it infuriates me when I type a 100-letter `sudo echo` cmd and then have bash tell me "echo: permission denied".
[21:59:49] <hydra_posthumous> nice misspelling
[22:00:06] <credak> it is his klingon name.
[22:00:28] <hydra_posthumous> hydra_posthumous is now known as jeplar
[22:00:32] <jeplar> why hello
[22:00:52] <credak> it is always fun until someone looses an eye
[22:00:57] <sudo_maddash> veggie-eating meat. how's that for irony?
[22:01:07] <jeplar> * jeplar uh oh
[22:01:28] <alex_joni> you gonna behave?
[22:01:36] <anonimasu> no
[22:01:38] <anonimasu> :D
[22:01:39] <sudo_maddash> * sudo_maddash ducks for cover
[22:01:44] <jeplar> jeplar is now known as lerneaen_hydra
[22:01:46] <credak> credak is now known as skunkworks
[22:01:52] <sudo_maddash> ROFL
[22:01:59] <alex_joni> keep going
[22:02:02] <alex_joni> :)
[22:02:16] <lerneaen_hydra> * lerneaen_hydra is midly disconcerted
[22:03:15] <sudo_maddash> sudo_maddash is now known as maddash
[22:04:17] <lerneaen_hydra> * lerneaen_hydra deathly silence
[22:04:56] <eholmgren> holy nick changing batman
[22:05:07] <Jymmm> lol
[22:05:19] <lerneaen_hydra> wth has batman got anything to do with it?
[22:05:38] <Jymmm> lerneaen_hydra Then take off the wonder woman outfit already!
[22:05:43] <alex_joni> maddash: latency is only partly because of frequency
[22:05:48] <alex_joni> or clock speed
[22:05:53] <lerneaen_hydra> * lerneaen_hydra mumbles something about privacy
[22:05:57] <alex_joni> the main thing for latency is determinism
[22:06:14] <maddash> alex_joni, I was considering adding smp and acpi to the kernel sources
[22:06:15] <alex_joni> and a 3ghz pentium also has a lot of "features" which make it work faster usually
[22:06:40] <alex_joni> let me explain about RT for a while, then decide for yourself if you want to do that or not
[22:06:40] <Jymmm> lerneaen_hydra it's not a pretty sight seeing you in WW outfit with hairy boobs sticking out of the top! LOL
[22:06:55] <alex_joni> maddash: do you know how RT works?
[22:06:55] <lerneaen_hydra> hairy? hairy!?
[22:07:04] <lerneaen_hydra> ha ha ha, I think not
[22:07:22] <alex_joni> maddash: btw, you don't need to leave #emc-devel.. just keep general things in here, and devel related stuff in #emc-devel
[22:07:30] <maddash> alex_joni, isn't it just an issue of task scheduling?
[22:07:34] <alex_joni> maddash: yes and no
[22:07:48] <alex_joni> for it to work in realtime, the RT stuff is above linux
[22:07:55] <maddash> alex_joni, well, I was thinking of recompiling the kernel from src...so I thought it was devel related.
[22:08:06] <alex_joni> not emc-devel related :)
[22:08:18] <alex_joni> but lets stick to the topic for a while
[22:08:43] <alex_joni> there is a layering system called ADEOS (now it's called IPIPE) which allows more than one OS to coexist
[22:09:20] <alex_joni> so during the time when the RT is running, all other of linux is "turned off"
[22:09:21] <maddash> alex_joni, isn't that just an optimized case of software virtualization?
[22:09:27] <alex_joni> smething like that
[22:09:46] <alex_joni> but the RT stuff works on periodic interrupts from some int controller
[22:10:02] <alex_joni> during the RT task (which is quite short) NOTHING else is allowed to happen
[22:10:05] <maddash> alex_joni, if that were so, then how could I abort while mini is processing a gcode file?
[22:10:19] <alex_joni> no hardware interrupts, no memory transfers which the CPU needs to take care of, etc
[22:10:28] <maddash> alex_joni, ah, ok, so uniform intervals of interrupts.
[22:10:37] <alex_joni> maddash: we are talking about tasks running at microsecond intervals here
[22:10:57] <alex_joni> users (humans) live in seconds, and tens of seconds worlds (perceptions)
[22:11:25] <alex_joni> so.. if something happens during the RT task, it's bad
[22:11:54] <alex_joni> it means the RT task doesn't fullfill it's purpose, or even worse it extends the time for it to execute so much that the whole machine locks up
[22:11:57] <lerneaen_hydra> alex_joni: something happens?
[22:12:15] <Jymmm> EldonB46: Interesting... http://www.mfjenterprises.com/products.php?prodid=MFJ-1910
[22:12:37] <alex_joni> lerneaen_hydra: an interrupt which isn't disabled
[22:12:46] <alex_joni> lerneaen_hydra: a cache miss, which results in swapping
[22:12:57] <lerneaen_hydra> hmm, that sounds nasty
[22:13:03] <maddash> alex_joni, what would be a typical "RT task"? b/c milling, let's say , often takes longer time than "microsecond intervals"
[22:13:06] <alex_joni> lerneaen_hydra: power management which decides to power down the HDD
[22:13:16] <alex_joni> maddash: emc has a number of tasks
[22:13:40] <alex_joni> there are a few high level, user space (which run only in normal linux): like g-code interpreter, user interface, etc
[22:13:47] <alex_joni> those don't matter so much
[22:14:01] <alex_joni> I mean it doesn't hurt for them to be stopped for a couple of milliseconds
[22:14:07] <alex_joni> even a second or so is not harmfull
[22:14:31] <alex_joni> then there are tasks which run faster: for example parts of the motion controller: trajectory calculations
[22:15:03] <alex_joni> these tasks run usually once every millisecond (1000/second), and they decide where the machine should move to, still obeying accel and speed limitations
[22:15:36] <alex_joni> usually that part is called TP (trajectory planner), and it gets commands from above (emc, interpreter, etc) about where it should go
[22:16:00] <alex_joni> then it takes a move (say a linear X move for 10") and samples it in 1millisecond intervals, and generates positions
[22:16:42] <alex_joni> * alex_joni catches a breath
[22:16:47] <alex_joni> ok.. next
[22:16:56] <alex_joni> there is a task which runs even faster
[22:17:18] <alex_joni> running once every few microseconds, it's responsible for outputting the actual step signals on the paralel port
[22:17:30] <alex_joni> these need to be very precise for stepper motors
[22:17:59] <alex_joni> a motor which should run at 20000 pulses / second, won't appreciate if the distance between pulses is not equal
[22:18:13] <alex_joni> say you have a pulse now, one 20 us later, one 2 us later, etc
[22:18:19] <alex_joni> this will cause the motor to stall
[22:18:42] <alex_joni> which is bad
[22:19:05] <alex_joni> so these tasks need to run _exactly_ when planned, not sooner, not later, nor should they take more time than planned
[22:19:51] <maddash> ah. now I see why acpi is shunned in rt kernels.
[22:19:59] <maddash> but what about smp?
[22:20:02] <alex_joni> now.. imagine something happens during one of that tasks (say powermanagement decides to scale the CPU frequency, because it determined that CPU is only 12% loaded)
[22:20:27] <anonimasu> ...) .... .15. ....... (704x396 DivX611 120f .).avi
[22:20:29] <anonimasu> hm..
[22:20:30] <anonimasu> damn.
[22:20:32] <alex_joni> smp is simply disabled because the people who write the RTAI add-on weren't able to make it work so far
[22:20:41] <maddash> alex_joni, yeah, then you'd get uneven pulses
[22:20:42] <alex_joni> at least not reliable on all machines
[22:20:44] <CIA-8> 03awallin 07HEAD * 10emc2/docs/src/hal/vcp.lyx: remove VCP documentation, we only have pyVCP documentation now.
[22:20:58] <anonimasu> vlc dosent love japanese charachters.
[22:21:02] <alex_joni> maddash: in the best case you get uneven pulses, and the machine stops(freezes)
[22:21:09] <alex_joni> in the worst case your PC locks up
[22:21:42] <maddash> alex_joni, well, wouldn't it be cool to have one core/processor just feed signals, while the other does peripheral work, ie gcode processing, trajectory calc,etc.
[22:21:58] <alex_joni> maddash: that would be the best
[22:22:06] <alex_joni> unfortunately it doesn't work.. so it's disabled
[22:22:19] <maddash> alex_joni, is there a source branch for smp?
[22:22:22] <alex_joni> you can be my guest and get your try at it..
[22:22:35] <lerneaen_hydra> maddash: you can use one computer for RT and one for GUI stuff afaik
[22:22:40] <alex_joni> maddash: no, you simply need to enable it in the linux source config, and in the RTAI config
[22:22:46] <lerneaen_hydra> * lerneaen_hydra looks to alex_joni
[22:22:58] <alex_joni> lerneaen_hydra: that is correct
[22:24:15] <maddash> alex_joni, yes, but you still need the emc source to be altered, ie, to position its modules between cores..
[22:24:15] <alex_joni> but I didn't see the need so far :)
[22:24:41] <alex_joni> maddash: I think/hope you can tell RTAI to put all RT tasks on one processor
[22:24:47] <alex_joni> and the normal tasks on the other
[22:24:55] <alex_joni> that way you need not touch emc
[22:25:03] <maddash> aha
[22:25:36] <alex_joni> unfortunately I don't think that's possible with the current RTAI.. I have no idea honestly
[22:26:02] <alex_joni> I know some people asked about SMP on the rtai mailing list in the past, and the developer said send me a board I might work on it
[22:26:27] <maddash> ROFL "go buy your own"
[22:26:46] <alex_joni> go write your own RT system
[22:26:47] <alex_joni> :P
[22:27:42] <lerneaen_hydra> alex_joni: was that before or after dual-core was available for cheap?
[22:28:41] <alex_joni> lerneaen_hydra: before, but it hasn't changed afaik
[22:28:48] <lerneaen_hydra> oh, ok
[22:40:50] <alex_joni> maddash: still want to know why modern processors aren't that good for RT performance?
[22:41:36] <maddash> alex_joni, well, hyperthreading's probably not good for rt
[22:41:56] <maddash> alex_joni, and a lot of modern processors attempt to use out of order execution
[22:42:00] <alex_joni> hyperthreading is generally bad
[22:42:09] <alex_joni> maddash: pipelines are very bad
[22:42:13] <alex_joni> for determinism
[22:42:27] <alex_joni> because they work very well in most cases, and very bad in some cases
[22:42:28] <maddash> powerpc has a short pipline
[22:42:41] <alex_joni> well.. you'd have to port emc to powerpc :D
[22:42:52] <alex_joni> caches are also badish
[22:43:29] <alex_joni> a hard realtime system works without caches, or any other gimmicks which improve throughput
[22:43:51] <alex_joni> RTAI isn't very hard realtime, so some things are allowed
[22:44:01] <alex_joni> but APM, ACPI, etc are a big no-no :)
[22:44:32] <alex_joni> you can get lucky and some things work, but getting it right for every platform out there is not possible, that's why most RT kernels have ACPI switched off by default
[22:45:34] <maddash> crap. so pentium d is a nono?
[22:46:01] <alex_joni> I didn't say that
[22:46:06] <lerneaen_hydra> less bad than a P4 or something with the netburst architechture at least
[22:46:12] <alex_joni> otoh modern processors have lots of rude power :D
[22:46:19] <alex_joni> so it kinda compensates
[22:46:42] <alex_joni> but you can't say a 3GHz processor is better for pulses than a 800MHz PIII
[22:46:52] <alex_joni> sometimes it's quite the opposite
[22:47:01] <maddash> which times?
[22:47:16] <alex_joni> depends on the mobo, chipset, memory, processor
[22:47:31] <alex_joni> lunar phase, humidity, aliens in the vecinity, etc
[22:48:16] <alex_joni> maddash: too many factors are involved for someone to be able to say what's generally better ...
[22:48:30] <alex_joni> I have an older Athlon XP which works GREAT
[22:48:46] <maddash> the cpu's the only thing doing any processing, so there wouldn't be any non-uniform latencies, right?
[22:49:20] <alex_joni> it's only 1.2GHz (Athlon XP 1400+), but I was able to run emc2 with a base period of 6.5 usec
[22:49:24] <alex_joni> maddash: yes and no
[22:49:36] <alex_joni> non-uniform latencies come from all sorts of things
[22:49:41] <alex_joni> like cache misses/hits
[22:50:08] <maddash> ie, cpu
[22:50:15] <alex_joni> imagine it needs some data, in one case it's in the L1 cache, in the next case it's in L2, in the next case it's in the main memory, in the next it's on the HDD
[22:50:27] <alex_joni> not necessarely CPU
[22:50:36] <alex_joni> HDD is not CPU :)
[22:50:42] <alex_joni> but that's extreme
[22:50:52] <alex_joni> even memory speed might make a difference
[22:51:20] <maddash> so...bus speed and cpu...
[22:51:41] <alex_joni> and external interrupts..
[22:51:51] <alex_joni> but the biggest killer is shared memory graphics
[22:51:59] <maddash> omg...
[22:52:43] <alex_joni> imagine the GPU taking over the bus, while the processor needs something and just sits there waiting for the onboard to fetch it's data
[22:53:04] <maddash> I just realized that I have integrated graphics
[22:53:09] <alex_joni> again.. it depends how it's implemented on the mobo
[22:53:19] <alex_joni> maddash: in some cases it might even work :)
[22:53:21] <maddash> I'll just stay in console mode
[22:53:34] <alex_joni> maddash: might make a difference, must not
[22:53:46] <alex_joni> you were running emc2.0.5 right?
[22:53:52] <maddash> no, 2.1a
[22:53:57] <alex_joni> even better
[22:54:09] <alex_joni> did you configure for sim? --enable-simulator?
[22:54:16] <maddash> you gave the src link yesterday
[22:54:21] <maddash> no
[22:54:27] <alex_joni> ok, good
[22:54:40] <maddash> why?
[22:54:44] <alex_joni> did you ever see a popup dialog which says "Unexpected realtime delay!" ?
[22:55:15] <alex_joni> emc2 after 2.0.4 tells you if there's something fishy with the time it took to execute some RT tasks
[22:55:16] <lerneaen_hydra> 'night all
[22:55:34] <alex_joni> night lh
[22:56:41] <alex_joni> huh?
[22:57:07] <alex_joni> maddash: if you never saw that dialog chances are your system is immune to the onboard graphics
[22:57:39] <alex_joni> but just to be sure, I would check following this procedure: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?TroubleShooting#Unexpected_realtime_delay_check_dmesg_for_details
[22:58:13] <alex_joni> start the RTAI latency test as described there, and let it run for a couple of minutes, while starting programs, minimizing/maximizing programs, etc
[22:59:53] <mdynac> anyone available to help with an emc problem?
[23:00:05] <alex_joni> maddash: I read back what you discussed earlier with the guys in here, and I also think you need to approach the problem otherwise
[23:00:11] <alex_joni> mdynac: sure.. shoot
[23:01:14] <alex_joni> sudo_maddash: did you disconnect?
[23:01:21] <sudo_maddash> I think so...
[23:01:28] <mdynac> well when i turn the machine on in emc both axis fly away...
[23:01:30] <sudo_maddash> but ping yahoo.com worked fine...
[23:01:34] <alex_joni> sudo_maddash: darn.. how much did you get?
[23:01:37] <sudo_maddash> funny.
[23:01:42] <alex_joni> mdynac: describe the setup a bit
[23:01:53] <alex_joni> mdynac: motors, drivers, way of driving
[23:01:53] <sudo_maddash> alex_joni, darn, can you c/p it over?
[23:02:02] <alex_joni> sudo_maddash: tell me the last you got
[23:02:08] <sudo_maddash> <lerneaen_hydra> 'night all
[23:02:08] <sudo_maddash> <maddash> alex_joni, no...that actually happened with the old 500mhz cpu...
[23:02:28] <mdynac> servos, driven by motenc lite, the dac out voltage goes to 10 volts when i do a "machine on"
[23:02:57] <alex_joni> do the encoders work?
[23:03:06] <mdynac> yep
[23:03:25] <alex_joni> mdynac: the only thing I can imagine of to saturate PID like that is if encoders turn the other way
[23:03:44] <alex_joni> doesn't it ferror?
[23:04:08] <mdynac> hmmm, i haven't changed anything except the harddrive crashed in the pc, so i had to re-install
[23:04:26] <mdynac> yes it ferrors
[23:04:46] <mdynac> joint following error....
[23:04:52] <sudo_maddash> alex_joni, should I dmsg|grep -i realtime?
[23:05:25] <alex_joni> sudo_maddash: what for?
[23:05:38] <alex_joni> mdynac: try inverting the scale in the ini
[23:06:05] <sudo_maddash> realtime delay msgs...
[23:06:05] <alex_joni> mdynac: I very much suspect the motor turns one way and the encoder is meant to turn the other way
[23:06:21] <alex_joni> sudo_maddash: nope, you would have gotten a message
[23:06:23] <mdynac> okay, but why would emc put out 10 volts with no axis command given?
[23:07:01] <mdynac> everything worked fine last week
[23:08:07] <alex_joni> mdynac: when switched on a servo hunts for the zero position
[23:08:16] <alex_joni> that means it moves from one encoder tick to another
[23:08:26] <mdynac> i c
[23:08:27] <alex_joni> back and forth slowly/fastly depending on the pid
[23:08:31] <cradek> sounds like you didn't get the right config restored
[23:09:05] <alex_joni> if it commands the motor to move one way but sees the feedback go the other way pid will only start to command harder in the right way (which is obviously wrong:)
[23:09:16] <alex_joni> not sure I am able to express clearly what I mean
[23:09:48] <mdynac> which scale in the ini?
[23:10:06] <alex_joni> INPUT_SCALE
[23:10:15] <alex_joni> but it's just a wild guess
[23:10:27] <mdynac> okay
[23:10:52] <alex_joni> it would be easier to analyse the system in person (I would: 1. move the encoder by hand and note the direction it goes, also check if the GUI numbers change)
[23:11:06] <mdynac> it's set to 40000 right now, i never set it to -40000, ever....
[23:11:19] <alex_joni> 2. I would command (in HAL) a tiny voltage to make the motors move one direction say .1 and check which way they go
[23:11:45] <alex_joni> maybe you had a setp -1 somewhere.. would do the same
[23:11:52] <alex_joni> did you set up this config?
[23:11:58] <mdynac> yes
[23:12:02] <mdynac> i set it up
[23:12:25] <sudo_maddash> sudo_maddash is now known as maddash
[23:12:30] <mdynac> i'll check it line by line tomorrow
[23:12:48] <alex_joni> mdynac: it might help if you put it up at pastebin.ca
[23:13:00] <alex_joni> guys in here might take a look at it, and maybe spot what's wrong :)
[23:13:09] <mdynac> did that once a couple weeks ago
[23:13:48] <alex_joni> did it help?
[23:14:38] <mdynac> it was suggested that my ferror and minferror were too large.....
[23:15:36] <mdynac> but that was another issue...
[23:15:48] <alex_joni> so it helped.. chances are it might help again :D
[23:16:55] <maddash> hmm I'm getting "joint 1 error" in mini when I really crank up the max_* values in my ini
[23:17:07] <mdynac> i'll just check my configs when i get to the machine tomorrow, and i will copy them and post them if need be....
[23:17:12] <alex_joni> maddash: sim/mini.ini ?
[23:17:38] <maddash> no, regular mini
[23:17:43] <alex_joni> maddash: doesn't sounds like it, it sounds like you are running a certain stepper config with mini as the interface
[23:17:57] <alex_joni> this means you are actually trying to output pulses on the parport
[23:17:56] <maddash> alex_joni, wait - yes, I am.
[23:18:35] <alex_joni> maddash: you can only output a certain amount of pulses per second, if you try to get more you'll get a joint * following error
[23:18:57] <maddash> alex_joni, how do I know what the max is?
[23:19:08] <alex_joni> maddash: you calculate it
[23:19:18] <alex_joni> it's not that complicated
[23:19:38] <alex_joni> lets define a pulse: it's a signal starting from 0 going to 1 then back to 0, right?
[23:20:33] <maddash> ok
[23:20:52] <alex_joni> that means emc needs to change parport twice for a pulse to happen
[23:21:08] <alex_joni> now, remember I told you there is some code in emc which runs very fast to output pulses?
[23:21:29] <alex_joni> it actually runs at a certain frequency (or period between code calls)
[23:21:31] <maddash> rtc. yep.
[23:21:43] <maddash> so I'm going too fast for it?
[23:21:44] <alex_joni> the period is defined in the ini as BASE_PERIOD
[23:21:50] <alex_joni> wait a bit :)
[23:22:39] <alex_joni> the default value for BASE_PERIOD is 50000
[23:22:47] <alex_joni> a rather large & conservative value
[23:22:58] <alex_joni> 50000 is in nanoseconds, meaning 50 microseconds
[23:22:58] <mdynac> http://pastebin.ca/313570
[23:23:06] <mdynac> there is my ini file
[23:23:19] <alex_joni> that's a frequency of 20kHz
[23:23:34] <alex_joni> needing 2 changes / pulse, means a max pulse rate of 10kHz
[23:23:48] <alex_joni> mdynac: will need the hal too
[23:23:55] <mdynac> k
[23:25:05] <alex_joni> maddash: now.. 10k pulses / second is pretty little
[23:25:20] <alex_joni> we can convert that to actual machine speed by looking at INPUT_SCALE
[23:25:33] <alex_joni> INPUT_SCALE tells emc how many pulses/inch of travel your machine has
[23:25:39] <mdynac> here it 'tis http://pastebin.ca/313581
[23:25:42] <alex_joni> say 2000 pulses/inch
[23:26:59] <alex_joni> mdynac: I notice your second axis has output_scale=10
[23:27:09] <alex_joni> the first one has OUTPUT_SCALE=1
[23:27:27] <mdynac> should be a 1 sorry
[23:27:41] <alex_joni> mdynac: that would cause it to run away :D
[23:27:55] <mdynac> both axis?
[23:28:10] <alex_joni> no.. second one
[23:28:28] <alex_joni> both run away?
[23:28:36] <mdynac> well the x takes off right away then the y starts
[23:28:54] <mdynac> until i get a joint following error
[23:29:13] <alex_joni> odd
[23:29:43] <alex_joni> ferror is quite high.. but that's ok atm
[23:30:02] <mdynac> i monitor the output to the drives and the x gets 10volt right away,then the y gets 10volts a few seconds later....
[23:30:42] <jepler> if you suspect it's because of a change we made between your last working checkout and this one, you can use "cvs up -Ddate" to get a copy of the source code as it was on date
[23:31:11] <mdynac> well i was in the process of fixing the ferror and pid tuning last week, but the hard drive crashed and now this......
[23:31:31] <jepler> e.g.: cvs up -D"2005/12/1"
[23:32:06] <alex_joni> mdynac: are you SURE the encoders work?
[23:32:08] <jepler> later you'd use some combination of -A and -r to get back to the current time and the development branch you were using (e.g., v2_1_branch)
[23:32:21] <maddash> hmmm, input_scale = 4000
[23:32:50] <alex_joni> I see 40000
[23:33:25] <mdynac> i will check them again tomorrow, it sure acts like they are screwed up somehow, mabe i lost the 5 volts to power them....
[23:34:07] <alex_joni> mdynac: can you turn axes by hand?
[23:34:48] <mdynac> well i can't check that right now cause they run away too quick....
[23:36:10] <alex_joni> mdynac: start emc, switch to estop reset, DON'T turn on
[23:36:18] <mdynac> k
[23:36:20] <alex_joni> move axes by hand, check if readout changes
[23:36:34] <alex_joni> if it does, then the encoders are ok, if not.. you found your problem
[23:36:34] <mdynac> will do
[23:46:10] <alex_joni> mdynac: I need to head to bed soon
[23:46:15] <alex_joni> getting close to 2am :)
[23:47:18] <mdynac> np, i will check everything tomorrow and get back with you guys, i'm sure it is software config problem, but i will re-check the encoders also....thanks guys
[23:48:23] <alex_joni> mdynac: from what I know around this time guys from over the pond start to get active in #emc
[23:48:41] <alex_joni> so keep asking in here, I'm sure someone else will answer :)
[23:48:50] <mdynac> k thanx
[23:49:44] <alex_joni> no problem
[23:49:50] <alex_joni> good night all