#emc | Logs for 2007-01-04

[00:26:01] <jepler> ah, maybe this is a good place to point all those who ask about how to burn the CD: https://help.ubuntu.com/community/BurningIsoHowto
[00:59:22] <cradek> jepler: yay
[01:18:02] <Goslowjimbo> Since things are so quiet, perhaps ypu would like to educate a hobbiest on what NML is. I see the mneumonic ( or TLA - three letter acronymn) quite frequently.
[01:24:12] <Jymmm> In what context have you seen NML ?
[01:24:21] <robin_sz> Neutral Machining Language I think ... its basically a messaging layer within EMC ... different bits of the system use NML message to talk to each other
[01:25:34] <robin_sz> in theory it can pass data between different systems, with different endianness etc
[01:29:24] <A-L-P-H-A> edianness?
[01:29:25] <A-L-P-H-A> :D
[01:29:39] <A-L-P-H-A> hehe... love it when people make up words. :) it makes them sound extra smart. :)
[01:30:08] <robin_sz> sorry, I meant endianness
[01:30:20] <robin_sz> as in big-endian and little-enidian
[01:30:35] <robin_sz> as in big-endian and little-endian
[01:30:46] <robin_sz> see, I just can't spol
[01:30:48] <A-L-P-H-A> "endians"
[01:31:00] <A-L-P-H-A> endianness is made up. :)
[01:31:13] <robin_sz> well, it may be, but its used a lot
[01:31:14] <A-L-P-H-A> robin_sz. sup?
[01:31:22] <robin_sz> north!
[01:31:30] <A-L-P-H-A> peter north?
[01:32:09] <robin_sz> hes always up ;)
[01:32:21] <robin_sz> http://en.wikipedia.org/wiki/Endianness
[01:32:27] <robin_sz> it may be made up, but not by me
[01:32:50] <Goslowjimbo> Is there any documentation on how to use/program with NML? The only thing I found on the net was a NML program which make HTML code.
[01:34:04] <robin_sz> mmm ... there was a lot of stuff on it ... you can control emc directly by using nml messages, for example if you have some sort of allergy to gcode
[01:34:21] <robin_sz> its not something you'd normally want to do ...
[01:34:51] <robin_sz> not for machining anyway
[01:34:58] <jepler> Goslowjimbo: In the context of emc, NML is the 'neutral messaging language' or something like that. It is generally a little like RPC, but different in the specifics.
[01:35:15] <jepler> Goslowjimbo: it is the method used to communicate between some of the userspace parts of emc2
[01:36:54] <robin_sz> iirc, isnt most of the nml stuff, like libnml put together by the Java packages that live in emc ...
[01:37:12] <robin_sz> I think the directory has a notice marked "here be dragons" on it
[01:37:32] <jepler> I think the java source code generator hasn't been used in a long time
[01:37:44] <robin_sz> probably due to the dragons ...
[01:37:48] <jepler> you're right, it's a "here be dragons" part of the system to most people
[01:38:05] <jepler> (including me)
[01:38:27] <robin_sz> there probably a couple of people in NIST who know the ins and outs of that part .. and thats your lot
[01:38:53] <Goslowjimbo> I think for me it should stand for kNow My Limits?
[01:39:00] <A-L-P-H-A> hmm... dragons and knights in European mythology... dragons and samurais in Asian mythology... I rather have Samurais.
[01:39:00] <robin_sz> I suspect the source has now chnaged so much by hand the the generator is redundant
[01:39:10] <A-L-P-H-A> Ninja's are just cooler.
[01:39:28] <robin_sz> A-L-P-H-A, mythology? who said they were myths/
[01:40:00] <A-L-P-H-A> my pet dragon said so.
[01:41:42] <robin_sz> Goslowjimbo, if you had a special application, like, say a 6 axos robot arm, or machine for CNC plucking chickens that you didnt want to use a GCODE front end for, NML lets you get at the task and trajectory bits of EMC without the other bits getting in the way, in theory you can write other controllers than the gcode intperter, like say a robot controller, and NML will be the comms layer .. as I understand it anyway
[01:43:05] <robin_sz> you wouldnt use it to create a program for machining a part, its lower level than that
[01:43:34] <robin_sz> bathtime!
[01:43:39] <robin_sz> I hate bath night ...
[01:43:49] <robin_sz> worst night of the year.
[01:44:15] <A-L-P-H-A> which kid?
[01:44:28] <A-L-P-H-A> aren't they old enough to take their own showers yet?
[01:45:43] <Goslowjimbo> From what I was concluding, NML wound up being the glue between HAL and the GUI? Are you guys a little punchy?
[01:46:03] <A-L-P-H-A> punchy?
[01:47:02] <Goslowjimbo> Maybe you know it as punch drunk. I just didn't want to get into the drinking aspect.
[01:47:43] <A-L-P-H-A> nope
[01:47:50] <A-L-P-H-A> we're normally like this... sober.
[01:47:55] <A-L-P-H-A> well... I am.
[01:48:05] <A-L-P-H-A> robin_sz just pretends to be road runner, for some odd reason.
[01:48:22] <robin_sz> meep?
[01:48:27] <A-L-P-H-A> there you go.
[01:48:35] <robin_sz> with respect, road runner goes "meep meep" ...
[01:48:38] <Goslowjimbo> jepler & robin_sz: How did you print out in red?
[01:48:39] <robin_sz> thats a bird noise
[01:48:50] <robin_sz> Goslowjimbo, we didn;t you did
[01:49:03] <robin_sz> "meep" is more of a hamster or rodent noise
[01:49:17] <Goslowjimbo> I did what?
[01:49:21] <A-L-P-H-A> hamsters hardly ever make a noise.
[01:49:27] <robin_sz> Goslowjimbo, you are the one that printed in red, not us ;)
[01:49:39] <A-L-P-H-A> i had one when I was a kid, I don't rmember it making any noise, other than on the wheel, or chewing on the bars.
[01:49:55] <A-L-P-H-A> 4Is this red?
[01:50:03] <Goslowjimbo> Yes
[01:50:11] <robin_sz> yes
[01:50:11] <A-L-P-H-A> 11Is this cyan?
[01:50:16] <robin_sz> yes
[01:50:20] <Goslowjimbo> yes
[01:50:19] <A-L-P-H-A> <shrug>
[01:50:25] <A-L-P-H-A> we're just lucky.
[01:50:32] <A-L-P-H-A> this maybe red to Goslowjimbo as well.
[01:50:38] <A-L-P-H-A> Goslowjimbo, red?
[01:50:41] <Goslowjimbo> Yes, it is
[01:50:43] <robin_sz> but .. the red Goslowjimbo was simply because the line contained his nick and his client highlighted it for him
[01:50:44] <A-L-P-H-A> :)
[01:50:49] <A-L-P-H-A> robin_sz. BRAVO!
[01:51:02] <Goslowjimbo> Oh
[01:51:14] <A-L-P-H-A> you get a cookie for that stroke of brilliance.
[01:52:03] <Goslowjimbo> Ok, what did you do to get cyan print on my computer?
[01:52:19] <robin_sz> 04some clients will recognise mis-spelings too like Gosloooowjimbo
[01:52:20] <A-L-P-H-A> 7magic brown pooh color
[01:52:38] <A-L-P-H-A> it's a color code.
[01:52:53] <A-L-P-H-A> in mirc, it's control-k
[01:53:00] <A-L-P-H-A> 9green/lime color
[01:53:02] <robin_sz> theres a perl plugin to make each letter a different color somewhere too
[01:53:19] <A-L-P-H-A> robin_sz, yeah, those rainbow scripts make reading really annoying.
[01:53:42] <robin_sz> colors are normally only used on noobie channels anyway
[01:53:58] <A-L-P-H-A> 3we be noobs
[01:54:04] <robin_sz> boobies!
[01:54:05] <A-L-P-H-A> 0you probably can't read this now.
[01:54:08] <A-L-P-H-A> :D
[01:54:21] <Goslowjimbo> Have a nice day
[01:54:21] <robin_sz> with that, I think its bathtime for real
[01:54:42] <A-L-P-H-A> 0if you can read this... we all think r ob in _ sz needs a bath... silly brit. bathing once a year
[01:54:44] <robin_sz> no putting it off now. but, once its done I can relax and forget about it for another year
[01:55:02] <A-L-P-H-A> robin_sz... get a jacuzzi, and go bathe with the wife.
[01:55:23] <robin_sz> nah ... no tthe wife
[01:55:35] <robin_sz> but .. she did mention we might get a secretary for work
[01:55:41] <robin_sz> I think something blonde ...
[01:55:44] <robin_sz> and C cup
[01:55:54] <A-L-P-H-A> nah... too much hassle.
[01:55:59] <A-L-P-H-A> get an Asian, with C cups.
[01:56:07] <robin_sz> they make asians with C cups?
[01:56:11] <A-L-P-H-A> yup...
[01:56:15] <robin_sz> I thought they topped out at AA
[01:56:33] <A-L-P-H-A> first girl I ever slept with was... natural C cups... loads of fun.
[01:56:45] <robin_sz> one of my employees visits bangkok a lot, I'll ask him
[01:56:57] <A-L-P-H-A> robin_sz... no. they don't occur in their native settings.
[01:57:03] <robin_sz> ahh.
[01:57:03] <A-L-P-H-A> only when they drink lots of milk when growing up
[01:57:08] <robin_sz> heh
[01:57:23] <robin_sz> right .. bath calls!
[01:57:25] <A-L-P-H-A> there's a plenty of North American Asians that drank milk growing up
[02:53:46] <fenn> mmm.. rBGH weirdness
[02:58:16] <skunkworks> bovine groth hormone?
[02:58:48] <skunkworks> growth
[03:09:09] <A-L-P-H-A> dunno... it was my sister's observation.
[03:09:17] <A-L-P-H-A> I think it's kinda true.
[03:09:26] <A-L-P-H-A> been looking at girls that like to drink milk more. :)
[03:09:36] <A-L-P-H-A> probably will be an giant increase in breast cancer too. :(
[03:12:13] <SWPadnos> milk isn't the sole determinant
[03:13:17] <A-L-P-H-A> SWPadnos... well... it seems like the common one.
[03:13:42] <A-L-P-H-A> the majority of big eyed girls, drink milk. :)
[03:13:42] <SWPadnos> I have 3 sisters who all drank around the same amount of milk as kids, and they have wildly different breast sizes
[03:13:54] <A-L-P-H-A> nono... within the last 20 years.
[03:13:59] <A-L-P-H-A> not prior.
[03:14:14] <A-L-P-H-A> met many 20 somethings that drink lots of milk, while they ever kids too?
[03:14:32] <A-L-P-H-A> the ones that do have big boobs, generally tend to drink milk.
[03:14:46] <A-L-P-H-A> not saying all milk drinkers get boobs...
[03:14:55] <A-L-P-H-A> not saying all big boobers drink milk.
[03:15:08] <A-L-P-H-A> but a high corrolation that big breasted girls, have and do drink milk.
[03:16:01] <SWPadnos> could be. I haven't done any of that research ...
[03:17:02] <A-L-P-H-A> it was emperical evidence.
[03:17:24] <SWPadnos> wouldn't you like to be the researcher for that one though?
[03:17:40] <A-L-P-H-A> I'll measure... you can write down the #s.
[03:17:46] <SWPadnos> I'll check for consistency
[03:18:34] <A-L-P-H-A> how would we do it as a double blind experiment, and make sure the project isn't flawed? Should we get multiple groups to perform the experiment in tandum?
[03:19:29] <SWPadnos> hmmm - I think I'll leave the experimental design up to you
[03:28:59] <CIA-8> 03compile-farm 07Ubuntu 5.10 (breezy) realtime (2.6.12-magma) * 10emc2head/: build PASSED
[03:29:16] <jmkasunich> thats a good sign
[03:30:37] <jepler> yay whee
[03:37:37] <cradek> has that not been working for a while?
[03:38:34] <jmkasunich> since 12/31 (at least)
[03:38:54] <jmkasunich> its been updating the webpage, but not emailing the list or CIA
[03:50:49] <ejholmgren> wow
[03:51:49] <ejholmgren> did all of you use multivariable calc to arrive at that milk conlcusion ...
[05:24:31] <fenn> A-L-P-H-A: http://www.opinionpower.com/Surveys/157038850.html
[05:25:22] <fenn> wish they would give the results so far
[05:38:35] <A-L-P-H-A> fenn. heh. glad you googled for stuff, and not me.
[05:39:32] <fenn> there's surprisingly little information, at least it was hard to find anything
[05:48:03] <A-L-P-H-A> :)
[05:48:09] <A-L-P-H-A> go with emperical evidence.
[05:48:21] <A-L-P-H-A> give your girl children milk to drink for the future boys of the world.
[09:59:33] <lerneaen_hydra> 'lo
[10:09:34] <fenn> oi
[12:53:53] <lerneaen_hydra> here I was thinking that there may be, for once, a non-flamey discussion with paul_c. looks like I was wrong
[13:03:43] <anonimasu> LOL
[13:04:15] <anonimasu> lerneaen_hydra: let me guess, he kept yelling "EMC2 SUCKS!" like some spoiled kid..
[13:04:28] <lerneaen_hydra> yeah pretty much
[13:05:05] <anonimasu> did you slap him=
[13:05:05] <anonimasu> ?
[13:05:10] <lerneaen_hydra> ...<some stuff about compiling qt3 stuff>... How you integrate the above snippet in to the emc2 build system is up to you and is something I can not help with - The current make structure is fragile, ugly, and based on a false premise.
[13:05:28] <lerneaen_hydra> no, I'll leave that to the real devs
[13:05:33] <anonimasu> omfg.
[13:05:37] <anonimasu> like bdi isnt fragile.
[13:05:57] <anonimasu> or the emc1 build system :)
[13:06:15] <lerneaen_hydra> the funny thing is that the previous posts were actually fine
[13:07:54] <cradek> I've noticed the "real" devs (not agreeing you aren't one) don't waste time on that either...
[13:08:44] <anonimasu> hehe
[13:08:51] <anonimasu> I dont get what's up with paul.
[13:09:26] <lerneaen_hydra> cradek: I'm not sure I'd call myself a dev, more like a knowledgable end user
[13:09:58] <lerneaen_hydra> that's good at suggesting stuff and whining about it ;)
[13:11:01] <cradek> and doing it
[13:11:20] <lerneaen_hydra> doing it?
[13:11:29] <cradek> free software has a nice continuum between user and developer
[13:11:53] <lerneaen_hydra> oh, yeah, that's a very nice thing :)
[13:12:00] <lerneaen_hydra> at least in smaller OSS projects
[13:12:15] <cradek> well crap, I confused you with the other ler* guy again
[13:12:32] <lerneaen_hydra> oh, right
[13:12:47] <cradek> my client cuts off everything after 9 characters, so you're both ler??a?n
[13:12:55] <lerneaen_hydra> IIRC he changed his name to his real name
[13:12:58] <lerneaen_hydra> oh noes
[13:13:13] <lerneaen_hydra> something with f in the beginning
[13:13:34] <cradek> I think he changed to klerman, which will be fine once I get used to it
[13:14:07] <lerneaen_hydra> yeah, thats right
[13:14:22] <lerneaen_hydra> f and k, they're so similar
[13:14:31] <lerneaen_hydra> :p
[13:14:39] <cradek> yeah, close enough
[13:16:36] <cradek> bbl...
[13:23:12] <jepler> morning folks
[13:23:50] <jepler> * jepler is afraid to look at the list
[13:24:52] <lerneaen_hydra> jepler: it could have been worse
[13:25:07] <lerneaen_hydra> so, what's happening?
[13:25:16] <jepler> lerneaen_hydra: umm -- nothing?
[13:25:21] <lerneaen_hydra> ok
[13:25:33] <jepler> lerneaen_hydra: like everyone else I'll spend the morning trying to resist a big "f-u" on the mailing list, apparently
[13:26:06] <lerneaen_hydra> yeah -.-
[13:27:16] <jepler> I wish we could just ban him from our lists like he banned me from the bdi4emc list
[13:28:43] <jepler> (not really -- I don't really want to go down the road of banning people who have a contrary viewpoint or who don't say that emc2 is perfect)
[13:29:24] <lerneaen_hydra> at times like this I find it nice to be able to think, ah, fsck that. and just forget about it
[13:30:26] <lerneaen_hydra> like going to istheshit for instance, and finding stuff like this: http://moods.istheshit.net/
[13:32:25] <lerneaen_hydra> http://hologram.istheshit.net/
[13:34:47] <lerneaen_hydra> http://thedogknowsforsure.istheshit.net/ O.o
[13:35:00] <lerneaen_hydra> that's enough of me spamming for now though
[13:35:11] <fenn> hmmm..
[13:36:04] <alex_joni> morning all
[13:36:44] <lerneaen_hydra> 'lo
[13:38:32] <alex_joni> whaddup?
[13:39:07] <lerneaen_hydra> making my electric bike thing
[13:39:18] <alex_joni> working yet?
[13:39:19] <lerneaen_hydra> almost
[13:39:27] <lerneaen_hydra> not much left now, the final bits
[13:39:27] <alex_joni> nice
[13:39:36] <lerneaen_hydra> I also have to make some type of outer casing
[13:39:48] <alex_joni> lexane
[13:40:05] <lerneaen_hydra> I was thinking 2mm aluminum :p
[13:42:38] <alex_joni> that works too
[13:46:31] <skunkworks> SWPadnos: thanks :)
[13:47:21] <lerneaen_hydra> lerneaen_hydra is now known as hydra_away
[14:34:12] <tomp> good morning
[14:38:41] <Dallur> morning tmop
[14:56:12] <alex_joni> morning
[15:04:26] <skunkworks> Morning
[15:05:22] <alex_joni> got snow?
[15:05:55] <skunkworks> we got an inch last week... but it is gone by now
[15:05:58] <Dallur> got root ?
[15:06:57] <alex_joni> we just got about 50 cm
[15:07:24] <skunkworks> That would be nice.
[15:07:51] <skunkworks> * skunkworks does a quick convert to inches
[15:08:27] <fenn> no snow yet here.. cherry trees blossoming in DC already
[15:08:44] <skunkworks> its an inconvenient truth ;)
[15:08:58] <alex_joni> skunkworks: 1.5 feet I think
[15:09:14] <alex_joni> 20"
[15:09:40] <skunkworks> that is what I got
[15:09:49] <skunkworks> 19.685
[15:10:18] <fenn> units ftw
[15:13:16] <skunkworks> it was in the 40's yesterday.
[16:06:01] <hydra_away> it
[16:06:09] <hydra_away> 's exceednigly warm here too
[16:06:20] <hydra_away> around +5 to +10 degrees
[16:06:27] <hydra_away> when it should be -5 to -10
[16:06:35] <hydra_away> (celcius(
[16:06:36] <hydra_away> *)
[16:09:55] <etla> hi all, what's up?
[16:14:29] <alex_joni> hi
[16:17:41] <etla> did you read the irc logs? about a feed aggregator for linuxcnc.org
[16:18:02] <alex_joni> nope
[16:18:08] <alex_joni> what time?
[16:18:17] <etla> yesterday probably
[16:18:54] <alex_joni> hmm.. no
[16:18:58] <etla> one idea to get more updates for www.linuxcnc.org would be to have a script that lists new stuff that is happening on the developers logs
[16:19:01] <alex_joni> didn't see that
[16:19:03] <etla> blogs i mean
[16:19:15] <alex_joni> hmm.. ok
[16:19:36] <etla> a joomla plugin probably exists for that
[16:19:49] <cradek> I suppose we could just post stuff there too
[16:20:44] <etla> yes, but an automatic script would make it easier. people tend to keep their own sites in good shape, and not worry about a 'community' site...
[16:21:41] <tomp> a meter widget for tcl\tk... how might this be converted to tkinter for use with pyvcp/hal? http://pastebin.ca/304872
[16:21:53] <cradek> in some of the very old irc logs, someone summarized what happened for each day - that was nice
[16:22:23] <etla> tomp: do you have a screenshot ?
[16:22:31] <tomp> will do
[16:22:42] <alex_joni> cradek: I did that for a while
[16:22:54] <alex_joni> not easy
[16:23:00] <cradek> I bet not
[16:23:18] <alex_joni> fun for a while.. then it gets annoying fast
[16:23:51] <tomp> etla: how to do screen/window capture in ubuntu? (tried prtscrn)
[16:24:09] <tomp> got it
[16:24:31] <cradek> in system/keyboard shortcuts, you can bind a key combination to "Take a screenshot of a window" - that's what I like to do
[16:28:07] <tomp> etla: http://imagebin.org/6836
[16:28:50] <tomp> etla: found others too
[16:29:27] <etla> tomp: looks OK, the tcl code can probably be translated pretty easily to python, could you put a link to the code and the screenshot on the pyvcp wiki page?
[16:30:10] <tomp> but links only good for 1 day... so what to do? where to put?
[16:30:39] <etla> the wiki should support uploads
[16:30:42] <tomp> k
[16:30:58] <etla> anyone really good at math?
[16:31:31] <jepler> define "good"
[16:32:10] <etla> can log(x^a -1 ) be simplified to something prettier? x is my variable, and a is a constant
[16:32:51] <etla> an extra parenthesis might help log( (x^a) - 1 )
[16:33:27] <etla> log(x^a) is so easy. but that 1 messes up everything
[16:34:26] <cradek> that voltmeter scares me
[16:35:05] <jepler> http://linux.inet.hr/finally-user-friendly-virtualization-for-linux.html
[16:38:08] <cradek> jepler: I tried kqemu twice (different OSs) and got no benefit or a slight slowdown
[16:38:11] <cradek> did you ever try it?
[16:39:36] <jepler> cradek: yes but I was never sure if it was even being used
[16:39:44] <jepler> cradek: this person obviously saw some benefit from it in his benchmarks
[16:41:26] <cradek> now I just (sadly) use vmware mostly
[16:42:09] <ejholmgren> woo hoo
[16:42:18] <ejholmgren> called in "sick" to work today
[16:42:39] <ejholmgren> putting together my parport relay card is so much more satisfying
[16:42:51] <skunkworks> loser ;)
[16:42:55] <jepler> tsk tsk
[16:42:58] <jepler> I never do that
[16:43:25] <ejholmgren> I rarely do either, but now I have a few weeks of sick time just sitting there
[16:43:59] <ejholmgren> emc can use multiple parports, correct?
[16:44:09] <Dallur> ejholmgren: yup
[16:44:08] <cradek> emc2 can, sure
[16:44:28] <tomp> etla: wiki updated
[16:44:53] <ejholmgren> good, this card only uses 8 of the outputs but wastes an entire port
[16:45:22] <tomp> cradek: why does voltmeter scare you?
[16:45:47] <skunkworks> needs some more divisions ;)
[16:46:30] <tomp> programmable
[16:46:41] <tomp> it's in the code
[16:47:39] <cradek> tomp: http://emergent.unpy.net/files/sandbox/should-be-shot.jpg
[16:48:15] <jepler> the analog voltmeter is hostage to physical and electrical constraints that dictate its form. on the computer there's no need to emulate a needle that sweeps through an angle proportional to the applied voltage. instead, we should be using the "Cracked open dinosaur eggs" graphical user interface paradigm.
[16:48:53] <cradek> (I think the bargraph is better in all ways)
[16:49:25] <ejholmgren> does that mean there are dinosaurs roaming around somewhere on the desktop too?
[16:50:15] <jepler> interesting widgets to add: a meter with a log scale; a meter which maintains peaks for a configurable time; a "strip chart" that shows some history; a meter or readout that changes color depending on the value (e.g., to indicate an out of bounds temperature or spindle load)
[16:50:41] <cradek> strip chart would be useful...
[16:50:55] <ejholmgren> I finally got rid of that ratshack POS soldering iron and got a weller ... should have done that a _long_ time ago
[16:51:28] <cradek> ejholmgren: the weller tips are much better aren't they
[16:51:34] <skunkworks> strip chart would be cool - say for tool velocity or spindle power.
[16:52:29] <hydra_away> strip chart?
[16:52:59] <cradek> hydra_away: like an EKG - shows a little history by scrolling
[16:53:08] <hydra_away> oh, ok
[16:53:13] <cradek> just like xload actually
[16:53:33] <hydra_away> does the background move or does the point sweep slowly across?
[16:53:34] <cradek> in fact starting with the xload code (auto scaling included) might be very smart
[16:53:56] <cradek> hydra_away: I suppose it could be either
[16:54:34] <hydra_away> it seems it would be easier to do it like an osciliscope with a large time scale, so the dot moves and the background is static
[16:54:41] <hydra_away> less cpu used at least
[16:54:50] <hydra_away> probably easier to code
[16:54:59] <cradek> xload gets around the "scroll all the time" problem by scrolling in chunks
[16:57:30] <etla> do add these ideas to the wiki please!
[16:57:32] <cradek> the bargraph could easily be improved to have safe/warning/critical colors with configurable thresholds...
[17:00:18] <tomp> bargraphs not intuitive for me, and limits & bands & setpoints are a natural for the meter widget, maybe people dont use meters anymore ( the digital watch syndrome )
[17:00:42] <tomp> hitchikers guide :)
[17:03:04] <etla> tomp: are you running a 3000x2000 resolution screen? the screenshot of the beter looks BIG
[17:03:16] <etla> meter :)
[17:04:29] <tomp> etla: i wish it autosized when the window was changed, but this tcltk stuff doesnt(easily), while gtk & qt do... another can of worms.
[17:04:49] <jepler> I am not saying there shouldn't be an analog meter lookalike in pyvcp, but my personal preference is for my software to look like other desktop applications, rather than like physical controls -- personally, I wouldn't use an analog meter.
[17:05:20] <jepler> I mean, I wouldn't use an onscreen widget that looks like an analog meter
[17:05:58] <tomp> ok, different strokes
[17:06:44] <etla> jepler: does AXIS use the bwidget notebook ? MDI/jog pages
[17:07:43] <jepler> etla: yes that is the bwidget notebook
[17:08:02] <etla> that would be cool for pyvcp I think
[17:08:29] <jepler> I think it does it directly in tcl, not with bwidget.NoteBook
[17:14:05] <ejholmgren> do SSR's have a lifespan more equivalent to a transistor than a mechanical relay?
[17:15:03] <^eugenics> hello
[17:15:32] <^eugenics> anyone have an idea where I can rent a pallet lifter?
[17:15:59] <ejholmgren> pallet jack or acutal lift?
[17:16:13] <^eugenics> jack i guess ;)
[17:16:24] <SWPadnos> in France?
[17:16:39] <^eugenics> to roll my 350kg pallet away
[17:16:59] <SWPadnos> crowbars and steel rods (plus friends) :)
[17:17:22] <^eugenics> SWPadnos, yes in france, but what type of busuness to search in?
[17:17:32] <^eugenics> :)
[17:17:36] <ejholmgren> rental?
[17:17:57] <^eugenics> something like industrial renting stuff
[17:18:05] <SWPadnos> do you have the equivalent of "rent-all" - a place where you can rent things like - well, almost anything?
[17:18:06] <jepler> ejholmgren: that's my impression -- looking at datasheets on opto22.com I don't see any information about a lifetime switching count for their SSRs
[17:18:37] <SWPadnos> I think SSRs last until you (a) overcurrent or (b) overvolt them, for the most part
[17:18:54] <^eugenics> france is a 3d world country
[17:19:14] <^eugenics> but ill have a look ;)
[17:19:33] <ejholmgren> just burn it in your yard then?
[17:20:05] <^eugenics> I want to keep the machine
[17:20:41] <jepler> "Mechanical relays have a finite cycle life, as their components tend to wear out over thousands to millions of cycles. SSRs do not have this problem; in the proper application, they could be operated almost infinitely." -- http://www.opto22.com/documents/0859_Solid_State_Relays_data_sheet.pdf page 17
[17:20:50] <^eugenics> otherwise probably a common solution for uncivilized people :)
[17:23:02] <ejholmgren> hrmmnn ~775 lbs is pretty heavy
[17:23:20] <SWPadnos> it's the weight of a bandsaw
[17:23:31] <SWPadnos> specifically, the one in my garage :)
[17:23:56] <ejholmgren> too bad it's not round? roll roll roll ;)
[17:24:11] <^eugenics> its a 100 amp plasma cutter from a ship yard industry
[17:24:15] <SWPadnos> one day, I may make the wheeled cart I've planned for it
[17:25:49] <ejholmgren> have a friend that works in a warehouse and a truck?
[17:26:00] <SWPadnos> that works
[17:26:16] <^eugenics> that would been great, but no
[17:26:37] <SWPadnos> I've hired a tow truck for moving machinery - it works pretty well (but only the boom type, not a flatbed :) )
[17:27:12] <^eugenics> maybe it is more easy to look to rent a wheeled motor krane
[17:27:23] <^eugenics> (in this country)
[17:27:27] <SWPadnos> an engine lift?
[17:27:36] <SWPadnos> that would be great for the size item you're talking about
[17:27:38] <^eugenics> yep
[17:37:41] <skunkworks> That is how father got his shaper home - tow truck... that was a long time ago.
[17:38:53] <Dallur> if you have a smooth surface under the pallet (concrete, painted concrete) you can just wet it a bit with water and slide the whole thing along, works really well up to about 600kg :D
[17:41:32] <^eugenics> wow, I should try that
[17:42:42] <^eugenics> I dont think the hole pathway is painted hmmmm
[17:44:08] <Dallur> eugenics: it's fine if it's just bare concrete btw
[17:44:24] <^eugenics> ok
[17:45:05] <^eugenics> Dallur, did you have chance to try your z axis?
[17:49:28] <skunkworks> water pipe and a come-along. (and holes in the concrete for pull points)
[17:50:24] <^eugenics> hmmm
[17:55:06] <skunkworks> that is how we moved the big stuff before the forklift
[17:55:31] <^eugenics> :)
[18:02:21] <^eugenics> gott nytt etla
[18:06:43] <hydra_away> gott nytt år ^eugenics
[18:06:49] <hydra_away> hydra_away is now known as lerneaen_hydra
[18:25:18] <^eugenics> det samma
[18:25:54] <^eugenics> kul att det finns lite svenskar har tycker jag
[18:28:14] <lerneaen_hydra> det är skoj ja
[18:31:39] <lerneaen_hydra> det känns nästan som att antalet svenskar är näst störst här efter amerikaner
[18:34:00] <^eugenics> yep
[18:34:33] <robin_sz> Helan går, sjung hopp fa-de-ral-lan-lal-lan lej!
[18:34:47] <robin_sz> * robin_sz joins in the fun
[18:35:21] <lerneaen_hydra> * lerneaen_hydra whispers with great fever among the other swedes and looks funnily at robin_sz
[18:35:30] <lerneaen_hydra> ;)
[18:36:02] <robin_sz> aww cmon, you know as well as I do, that all swedish is based on drinking songs
[18:36:06] <lerneaen_hydra> hahaha
[18:36:23] <lerneaen_hydra> at my school you may have a point
[18:36:36] <lerneaen_hydra> (btw, is uni referred to as school?)
[18:37:05] <robin_sz> I think thats the only bit of Swedish I know, we learnt it on a sailing trip in Aaland
[18:37:13] <robin_sz> no, uni is called uni
[18:37:21] <robin_sz> school is up to 16
[18:37:35] <robin_sz> after that its college
[18:37:40] <robin_sz> or university
[18:38:07] <robin_sz> "after I left school, I went to university" ...
[18:38:28] <lerneaen_hydra> oh, that's something that's been bothering me, in the US, what is the type of school called that takes place between the age of 15-18 or thereabouts?
[18:38:49] <robin_sz> no idea ...
[18:38:50] <lerneaen_hydra> and what's the type of school called that's 18-25-ish?
[18:39:07] <robin_sz> mmm, thats "college" I think
[18:39:10] <lerneaen_hydra> because I was thinking that it was highschool and uni/collede
[18:39:14] <SWPadnos> 14-18 = high school
[18:39:19] <robin_sz> yeah
[18:39:24] <SWPadnos> 19-22 = college or university
[18:39:36] <lerneaen_hydra> and what's the real difference between college and university?
[18:39:40] <SWPadnos> depends on your birthday though :)
[18:39:43] <robin_sz> IQ
[18:39:50] <lerneaen_hydra> haha
[18:40:00] <SWPadnos> a university generally consists of many colleges - a college is a more specialized school
[18:40:06] <tomp> a university is a collection of colleges
[18:40:06] <lerneaen_hydra> oh, I see
[18:40:10] <robin_sz> we do school to 16 ..
[18:40:31] <robin_sz> thats what we call "secodnary school" ... it has 5 yrs of that from 11 to 16
[18:40:32] <tomp> university of texas:college of architecture
[18:41:08] <lerneaen_hydra> oh, I see
[18:41:13] <robin_sz> after that you can either leave and get a job at 16, or stay on to "6th form"
[18:41:38] <tomp> there's some 'k' terms for lower schools now in the US... k1 thru k12 (form?)
[18:41:42] <SWPadnos> are "forms" years, or do they encompass a certain set of learning?
[18:42:11] <SWPadnos> in the US, it starts at "step-up" or some such, around 2-3 years old (I'm not kidding)
[18:42:17] <robin_sz> you either do the 2 years of 6th form at your school or at a special 6th form college, then you go to university ... you only do 6th fomr if you intend to do uni really
[18:42:17] <SWPadnos> then pre-schol (age 3-4, usually)
[18:42:29] <SWPadnos> kindergarten at age 5
[18:42:30] <robin_sz> forms a re years
[18:42:41] <lerneaen_hydra> SWPadnos: yeah, that I remember from when I was in the US
[18:42:47] <SWPadnos> first grade starts at age 6-ish, and is part of elementary school (1-3 usually)
[18:42:47] <robin_sz> 1st form is the first year of secodnary
[18:43:05] <SWPadnos> grades 4-8 are considered "middle school"
[18:43:12] <lerneaen_hydra> I moved to borkland after 6th grade or so, so that's why my image of the school system is messed up
[18:43:14] <SWPadnos> 9-12 are high school
[18:43:23] <SWPadnos> then there's college (even if it's a university :) )
[18:43:57] <lerneaen_hydra> ok
[18:44:15] <lerneaen_hydra> is there a rule of thumb for how many courses a college has before it's a uni?
[18:44:29] <SWPadnos> (separated in to "undergraduate studies" and "graduate studies" ...)
[18:46:26] <SWPadnos> this is very weird - it's 46 degrees here
[18:47:36] <lerneaen_hydra> maybe this year joe-schmo will start beleiving that global warming is real
[18:49:26] <SWPadnos> nah. as long as there's snow on Kilamanjaro they won't believe
[18:51:16] <robin_sz> lerneaen_hydra, in the uk the difference between college and uni was based on academic standards ...
[18:52:21] <awallin> scripts/realtime stop
[18:52:21] <awallin> scripts/realtime: line 168: rtapi_app: command not found
[18:52:34] <awallin> is that normal? I'm not running an RT kernel btw
[18:52:50] <robin_sz> we used to have polytechnics, teaching practical things like electronics .. and universites doing degrees in electronics .. the uni courses had more theeory and were harder
[18:54:03] <SWPadnos> "polytechnic" would be a "tech school" or "vocational school" here
[18:54:44] <robin_sz> used to be here , but .. about 6 or 8 years ago, the govenrment decided to increase the number of university places here from 10% of school leavers to 50%
[18:55:02] <robin_sz> they did this by:
[18:55:12] <robin_sz> renaming polytechnics as universities
[18:55:14] <robin_sz> and
[18:55:18] <robin_sz> letting anyone in
[18:55:24] <SWPadnos> that was efficient
[18:55:42] <robin_sz> well, you still needed the same school exam grades
[18:55:57] <robin_sz> but, suddenly everyone was getting better grades .. just like that
[18:56:19] <robin_sz> of course, the official line is teaching got better, not exams easier ...
[18:57:11] <SWPadnos> they're employing capitalist tactics to an area where they're not appropriate
[18:57:19] <SWPadnos> (cut corners until you get the profits up)
[18:58:18] <robin_sz> of course the "real" universites kept their high exam standards, but that made it hard for poor grade students to pass the exams, so they had a lot of failures .. lots of fialures makes it difficult to get new students, who wants to go on a course where you fail?
[18:58:30] <robin_sz> no students = no income
[18:59:05] <robin_sz> = course closure or compromise your standards .. do ANYTHIG to get the students through!
[19:00:01] <SWPadnos> I don't have time for you to get me started on education :)
[19:04:13] <jepler> awallin: I don't get that error when running RIP, but I always use emc-environment which might hide the problem (rtapi_app not on $PATH)
[19:05:15] <awallin> jepler: I did run . ./emc-environment (I seem to have to run that every time I boot, can the changes be made permanent?)
[19:07:17] <jepler> awallin: I put it in my ~/.bash_profile which is run from the login shell.
[19:07:43] <awallin> ok
[19:08:50] <awallin> ../emc2.head.dev/scripts/halrun -I pyvcp-dro.hal
[19:08:50] <awallin> Realtime already running
[19:09:00] <awallin> I thought scripts/realtime stop would fix that?
[19:10:50] <CIA-8> 03jepler 07HEAD * 10emc2/scripts/realtime.in: put the emc bindir on the PATH when using run-in-place; fixes error 'rtapi_app: command not found'
[19:12:08] <jepler> awallin: is rtapi_app still running? That's what halrun is checking
[19:12:10] <jepler> I think
[19:13:09] <awallin> [rtapi_app] <defunct>
[19:14:05] <awallin> yay, now realtime stop runs without errors
[19:14:10] <jepler> I bet in that case, pidof thinks rtapi_app is still running
[19:14:10] <awallin> let's try the halrun once more
[19:15:05] <awallin> nope, "Realtime already running"
[19:15:06] <jepler> in unix, a process is "defunct" (or "zombie") from the time it exits to the time its parent process fetches the exit status
[19:15:31] <jepler> you might have a suspended (^Z'd) halcmd left as well?
[19:15:59] <jepler> "ps wauxf" and look for the process that is "above" rtapi_app in the listing
[19:16:42] <awallin> yes, I have a halcmd -f somefile.hal there also. I killed the terminal since it went on printing dot's ......... when it couldn't find pyvcp_dro.xml (jmk's typo)
[19:16:54] <awallin> dots
[19:16:59] <jepler> kill the halcmd
[19:17:46] <awallin> ok, now jmk's dro example runs
[19:17:59] <jepler> wow that was sure a hassle
[19:19:12] <awallin> found one more typo....
[19:20:32] <awallin> hmm. when running pyvcp via halrun, when I write 'exit' at the halcmd prompt pyvcp crashes with a 'KeyboardInterrupt'. any way to gracefully shut down on keyboardinterrupt?
[19:21:02] <jepler> try: mainloop() except KeyboardInterrupt: sys.exit(0)
[19:21:14] <jepler> something like that ^^ split across multiple lines instead of all on one
[19:21:21] <awallin> ok
[19:23:43] <awallin> NameError: global name 'KeybaordInterrupt' is not defined
[19:24:01] <awallin> right. it's misspelled ;)
[19:24:35] <awallin> yep, now it works
[19:25:34] <CIA-8> 03awallin 07HEAD * 10emc2/src/hal/user_comps/pyvcp.py:
[19:25:34] <CIA-8> exit gracefully on KeyboardInterrupt
[19:25:34] <CIA-8> this happens when pyvcp is called from a halrun script
[19:25:34] <CIA-8> and the user types 'exit' at the halcmd prompt
[19:29:06] <CIA-8> 03awallin 07HEAD * 10emc2/lib/python/vcpparse.py:
[19:29:06] <CIA-8> exit gracefully when user tries to run two instances of pyVCP
[19:29:06] <CIA-8> "Error: Only one running pyVCP allowed."
[19:29:40] <SWPadnos> how difficult is it to allow multiple pyvcp instances?
[19:30:03] <jepler> you could add a commandline option to give the component name (and thus the pin prefix) to use
[19:30:26] <SWPadnos> ok - is pin naming the main impediment?
[19:30:59] <awallin> how do we keep track of how many times create_vcp has been called?
[19:31:26] <awallin> now we get to the fun part in naming hal pins, I just love the pyvcp.0.something syntax ;)
[19:32:09] <SWPadnos> well, I'm thinking about things like having some function pop up pyvcp panel, which would be a problem if there's already one :)
[19:32:16] <SWPadnos> I'm not thinking about the programming aspects yet ;)
[19:32:32] <jepler> here's what I have in mind: if you did "pyvcp -c p1 p1.xml" then it would create the compnent p1, and pins like p1.pin-1 etc
[19:32:35] <awallin> swp: yes, it could be useful
[19:32:59] <jepler> or maybe the basename of the xml file should be used instead -- so "pyvcp dro.xml" creates pins like dro.pin-1 etc
[19:32:59] <SWPadnos> "this is left as an exercise for the reader" :)
[19:33:10] <SWPadnos> that sounds like a great idea
[19:33:33] <SWPadnos> (of course it can be broken, but it's still excellent)
[19:35:09] <SWPadnos> I guess I"d still want to provide for a command line option, so you could have multiple instances of the same XML file with different names
[19:35:20] <SWPadnos> I haven't figured out where that may be needed though
[19:35:40] <awallin> jepler: so you just check that argv[1]=='-c' , name the component argv[2], and the file argv[3]
[19:35:55] <SWPadnos> I'm sure there's a python version of getopt ...
[19:35:56] <jepler> awallin: Python provides the 'getopt' module for parsing commandline arguments
[19:36:33] <SWPadnos> and it's waaaaay better than if argv[1]="-c" ...
[19:36:53] <jepler> opts, args = getopt.getopt(sys.argv[1:], "c:")
[19:37:32] <jepler> then for o, a in opts: check if o == "-c": component_name = a
[19:37:59] <jepler> args[0] will be the .xml file
[19:38:40] <jepler> if there's a bad argument, getopt() will raise getopt.GetoptError
[19:39:59] <awallin> I'll try that
[19:40:36] <awallin> it means you probably need to change the way AXIS calls vcpparse too, since pyvcp.py will need to pass the component name as a parameter to vcpparse
[19:41:27] <jepler> ok that's fine
[19:41:47] <jepler> if you want to fix that in axis, be my guest
[19:41:48] <SWPadnos> you can have 3 kinds of naming: default to pyvcp, command-line switch to use XML base name (-x maybe), or specified name (-n "myName")
[19:41:50] <jepler> otherwise I will
[19:42:24] <cradek> awallin: can the user tweak the vcp screen update rate?
[19:42:42] <awallin> cradek: no (not yet?)
[19:42:52] <awallin> it's set to 100ms I think
[19:43:07] <cradek> ok
[19:43:12] <awallin> for all widgets (do we need a different update time for different widgets?)
[19:43:21] <cradek> this sure has generated a lot of interest - thanks for doing this work
[19:43:36] <awallin> :) beginners luck I guess
[19:44:59] <jepler> python and hal are very powerful -- you can do a lot after you've understood their basics
[19:45:43] <jepler> I also hope that robin_sz will soon do a write-up of his work with python and a hardware PLC that he's mentioned here before
[19:49:47] <awallin> hrm... except getopt.GetoptError: doesn't seem to catch invalid arguments...
[19:50:09] <awallin> I have try: opts, args = getopt.getopt(sys.argv[1:], "c:") as suggested
[19:53:17] <jepler> >>> getopt.getopt(["-x", "example"], "c:")
[19:53:19] <jepler> getopt.GetoptError: option -x not recognized
[19:53:27] <jepler> >>> getopt.getopt(["-c"], "c:")
[19:53:29] <jepler> getopt.GetoptError: option -c requires argument
[19:53:42] <jepler> >>> getopt.getopt(["-c", "p", "x.xml"], "c:")
[19:53:42] <jepler> ([('-c', 'p')], ['x.xml'])
[19:54:23] <jepler> it works like I expect it
[19:55:49] <awallin> maybe you can spot the mistake:
[19:55:53] <awallin> try:
[19:55:54] <awallin> opts, args = getopt.getopt(sys.argv[1:], "c:")
[19:55:54] <awallin> except getopt.GetoptError:
[19:55:54] <awallin> print "usage: pyvcp -c hal_component_name myfile.xml"
[19:55:54] <awallin> sys.exit(0)
[19:56:46] <cradek> yes it should print that message to stderr, and exit with an error code >0
[19:56:51] <cradek> :-)
[19:59:04] <awallin> the output from print opts print args is [] and ['pyvcp-dro.xml'] but I still don't get the exception
[19:59:49] <cradek> I don't understand - are you trying to make -c required?
[19:59:52] <jepler> it worked for me; the code with slight modifications here: http://pastebin.ca/305355
[20:00:28] <jepler> oh -- yes, getopt considers all options as, er, optional
[20:01:34] <jepler> component_name = None
[20:01:36] <jepler> for optname, optarg in opts:
[20:01:36] <jepler> if optname == '-c': component_name = optarg
[20:01:36] <jepler> if component_name is None:
[20:01:36] <jepler> print "-c required"
[20:01:37] <jepler> sys.exit(1)
[20:01:47] <jepler> ^^^ you could do something like this (untested) if you want it to be required
[20:05:58] <awallin> if component_name is None:
[20:05:58] <awallin> UnboundLocalError: local variable 'component_name' referenced before assignment
[20:11:22] <awallin> ok, so now it works with the correct command line args: pyvcp -c t pyvcp-dro.xml
[20:11:34] <awallin> but it still crashes when the wrong arguments are given
[20:14:40] <awallin> oh, now it works!
[20:14:48] <awallin> how do I name the tk root window ?
[20:24:16] <awallin> got it.
[20:26:13] <CIA-8> 03awallin 07HEAD * 10emc2/src/hal/user_comps/pyvcp.py:
[20:26:13] <CIA-8> new syntax for calling pyvcp:
[20:26:13] <CIA-8> pycvp -c compname myfile.xml
[20:26:13] <CIA-8> created pins will be named:
[20:26:13] <CIA-8> compname.led compname.number etc.
[20:26:13] <CIA-8> this will allow multiple pyvcp panels to run simultaneously
[20:26:48] <CIA-8> 03awallin 07HEAD * 10emc2/lib/python/vcpparse.py: support the new -c syntax
[20:27:25] <awallin> looks like it didn't even brake the way AXIS calls vcpparse
[20:28:24] <CIA-8> 03jepler 07HEAD * 10emc2/src/hal/halmodule.cc: improve the docstring, including a short working example
[20:36:45] <awallin> hmm, what about if the component name was specified in the XML file ? any comments on that?
[20:37:18] <tomp> jmk: cvs version of vcp_widgets doesnt have the hscale widget. should i stop? no problem ( less problem ! )
[20:40:08] <jepler> awallin: doesn't really work when running from axis, since the component name (axisui) was already set before any of pyvcp is called
[20:41:38] <awallin> jepler: ok, I guess setting the compname on the command line is ok. You can now rename the pyvcp widgets in AXIS to something else with an argument to vcpparse.create_widgets if you want (axis_pyvcp or something)
[20:43:17] <jepler> ok, not sure I'll change anything right now but maybe in the future I'll let the user choose the prefix for himself
[20:48:12] <tomp> jmk: the vscale is there ok
[20:51:57] <paragon36> Hello All
[20:52:07] <cradek> hi
[20:52:27] <paragon36> Hope you had a nice crimbo ... :-)
[20:53:02] <cradek> I don't know what that is
[20:53:50] <paragon36> Could someone please explain the INPUT_SCALE varible I have read the docs and still a little confused!
[20:54:22] <skunkworks> Crimbo is a slang term for Christmas
[20:54:32] <cradek> that scale is how many steps (steppers) or encoder counts (servos) make up a length unit (inch or mm) on your system
[20:55:27] <cradek> skunkworks: aha, thanks
[20:56:07] <paragon36> OK so if I have a servo with a 2000 cpr and I am working in mm I would set it to 2000 0 which would give me 1mm per rev of servo right?
[20:56:28] <cradek> that depends on your leadscrew pitch
[20:56:43] <cradek> it's how many counts make up a mm of TRAVEL on the axis
[20:57:34] <skunkworks> * skunkworks had to look crimbo up on the wikipedia
[20:57:51] <cradek> if your leadscrew moves the axis 1mm per revolution, your scale would in fact be 2000
[20:58:35] <paragon36> OK so if I want a 5 - 1 gearing ... ie 5 motor revs to 1 turn lead screw (1mm) i would set it to 10000 0 !
[20:59:05] <cradek> yes if the axis moves 1mm for every 10000 counts, that's the scale
[21:00:43] <paragon36> The set up I have current for testing servo is that the motor is not connected to anything but I have set a marking point to test acuracy
[21:01:19] <cradek> ok, so you don't really have mm anywhere, you can pretend however you like
[21:01:43] <skunkworks> paragon36: what is your setup? what are you using for drives/interface and such?
[21:01:45] <cradek> you could "imagine" a 1mm per revolution leadscrew, so when you jog 1mm on the screen the motor will turn 1 rev
[21:03:24] <paragon36> Thats the idea or say a 5 - 1 ratio ... the problem is that if I home to X0 then command to X100 I would expect the motor to stop at the same point it started but it out by say quater of a turn or so
[21:04:18] <cradek> is it "stiff" when you turn it to the expected location? maybe it's just terrible tuning
[21:04:23] <paragon36> Then if I go back to X0 this finds its starting point but sometimes its off
[21:04:47] <cradek> or maybe your encoder or some other hardware isn't working right
[21:05:07] <cradek> you could use halmeter to check the operation of the encoder (disable the motor driver and turn it by hand)
[21:05:11] <paragon36> It has a slight movement but then it starts to fight
[21:05:36] <cradek> ok if it's fighting that's good, but your counting is wrong.
[21:05:55] <cradek> are you sure about your encoder resolution? it's very easy to be off by a factor of 4
[21:06:42] <cradek> ok, if you can not reliably go back to 0, it is NOT a scale problem - do you see why I say this?
[21:06:46] <paragon36> well the motor has 500cpr which should produce 2000 counts leading / lagging edges
[21:06:53] <cradek> ok
[21:07:03] <cradek> what hardware are you using to count?
[21:07:21] <paragon36> Etch.ini .... parrallel port
[21:07:43] <paragon36> Its connected directly
[21:07:54] <cradek> you may be losing counts because they are happening too fast then. the parallel port counting is somewhat limited.
[21:08:11] <cradek> you have to do the math to see what your safe top speed is
[21:08:52] <cradek> the parallel port is read every BASE_PERIOD. You must never have more than one edge per BASE_PERIOD or counts will be lost
[21:09:02] <paragon36> Ive noticed that I cannot run the motor at its rated rpm as it runs away with itself before throwing a following error
[21:09:19] <cradek> yes that's a sure sign of losing counts.
[21:09:42] <cradek> it thinks it's falling "behind" so PID commands it to go faster
[21:09:51] <cradek> which makes the problem worse, of course
[21:10:16] <paragon36> Yes I agree.... what cheap hardware solution is available ....
[21:10:28] <cradek> define cheap :-)
[21:10:40] <paragon36> Chips ... lol
[21:11:17] <cradek> ppmc, mesa, motenc all read quadrature at high speed
[21:11:23] <cradek> none are very cheap
[21:11:58] <cradek> if you like to experiment with new stuff, you could try a pluto for about $60
[21:12:09] <paragon36> pluto ?
[21:12:24] <cradek> it's a new FPGA parport device supported by EMC2
[21:12:52] <lerneaen_hydra> paragon36: you're not by any chance the same as parag0n on #gp2x on Efnet?
[21:12:53] <cradek> very new and experimental
[21:13:10] <cradek> http://cvs.linuxcnc.org/cvs/emc2/docs/src/hal/pluto_servo.html?rev=HEAD;content-type=text%2Fhtml
[21:13:16] <cradek> ^^^^ pluto docs
[21:13:32] <paragon36> No lerneaen ... never heard of gp2x !!
[21:13:49] <lerneaen_hydra> ok, that would have been very spooky
[21:14:11] <lerneaen_hydra> (the gp2x is a linux media handheld of sorts)
[21:14:32] <paragon36> Thanks cradek .... taking a look at pluto now!
[21:15:07] <paragon36> Sound interesting lerneaen ...
[21:15:35] <lerneaen_hydra> http://en.wikipedia.org/wiki/Gp2x
[21:16:06] <cradek> paragon36: are you using some kind of PWM and H bridge to run the motor?
[21:16:57] <paragon36> Im using the standard etch_servo connected directly to an H-Bridge
[21:17:11] <cradek> paragon36: with your current setup, try setting your feed override or jog speed slider very low and see if you can maintain position (come back to zero)
[21:17:17] <cradek> ok
[21:17:35] <skunkworks> paragon36: what is your base period set to?
[21:17:36] <cradek> putting the pluto between those will make you very happy then
[21:17:44] <paragon36> h-bridge = L298
[21:18:09] <cradek> paragon36: cool, that's exactly how my lathe works
[21:18:13] <paragon36> BASE_PERIOD = 20000
[21:18:35] <jepler> another option is a microcontroller which functions as a quadrature divider. http://emergent.unpy.net/projects/01149271333
[21:18:39] <paragon36> Computer = AMD 64
[21:19:06] <paragon36> Could the base period be set lower ?
[21:19:15] <skunkworks> paragon36: something to read..
[21:19:14] <skunkworks> http://www.cnczone.com/forums/showthread.php?t=25929
[21:19:17] <paragon36> Thanks jepler ....
[21:19:45] <paragon36> Is there a way to record an IRC session so I can save all this great info?
[21:19:47] <jepler> (10000 per mm) * (10 mm per second) = 100000 per second
[21:20:04] <cradek> paragon36: there are irc logs on linuxcnc.org
[21:20:05] <jepler> paragon36: that depends on your irc client. You might be able to just select all the text and paste it into your favorite text editor
[21:20:09] <jepler> logger_emc: bookmark
[21:20:09] <jepler> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2007-01-04.txt
[21:20:12] <jepler> cradek: good point!
[21:20:16] <cradek> heh
[21:20:28] <cradek> (that url is the log of the current conversation)
[21:20:33] <paragon36> Nice ... :-)
[21:20:48] <skunkworks> paragon36: theoretically you should be able to run the motor at 1500rpm counting at 50khz But they say you should half that. 2000 input scale.
[21:20:56] <paragon36> (10000 per mm) * (10 mm per second) = 100000 per second ... what does this give ... sorry for being dumb Chaps
[21:21:22] <jepler> that gives the number of quadrature edges per second
[21:21:50] <skunkworks> * ment 2000 edges on your encoder
[21:22:14] <jepler> oh I thought I saw something about 5:1 pulleys
[21:22:40] <paragon36> Got it ... Yes you did Jepler ... it's me .... lol
[21:22:59] <paragon36> 5 - 1 == 10000
[21:23:32] <jepler> 10000/second = 10 microsecond, so for that many counts per second you need a BASE_PERIOD no bigger than 10000 (and probably smaller, if the quadrature waveform is not perfectly clean and 90 degrees phase difference)
[21:24:05] <cradek> so what's the safe jog speed (in "mm"/minute) with his current setup?
[21:24:07] <paragon36> You just answered my next question Jepler ... Thank you!
[21:24:31] <jepler> on etch-a-sketch, I think I had more like 7000 edges per second, which can have a much bigger BASE_PERIOD without trouble
[21:24:58] <jepler> INPUT_SCALE=1700, MAX_VELOCITY=4 (inches per second)
[21:26:11] <paragon36> 1700 does not match the encoder count though does it ??
[21:26:47] <jepler> in the etch-a-sketch there's not an exact relationship between rotations and inches or mm
[21:27:23] <jepler> there's some odd-sized pulley inside the body of the etch-a-sketch, which sets the distance per rev
[21:27:35] <SWPadnos> and it probably changes over the "work surface"
[21:27:41] <jepler> it is not precision equipment
[21:27:55] <SWPadnos> oh hm - maybe it isn't winding string onto a pulley ...
[21:28:08] <jepler> no, I think a loop is continuously being pulled over the pulley
[21:28:31] <SWPadnos> well, weird but constant is better than weird but variable :)
[21:28:45] <jepler> my estimate could be +-25% and I'd never care
[21:28:46] <paragon36> Oh sorry jepler I thought you where talking about the emc etch program only and not the complete article ie the linkage / gearing of the etch a sketch
[21:29:14] <jepler> yeah I was referring to the whole setup
[21:30:19] <paragon36> would it be safe to lower my base_period to 10000 from 20000 or could this cause a lock up of my system>
[21:30:31] <SWPadnos> it can cause a lock-up
[21:30:59] <SWPadnos> the fastest anyone (known to me) has gotten a machine was around 7000
[21:31:19] <paragon36> Base period = 7000 ???
[21:31:32] <SWPadnos> yes, but don't try this at home :)
[21:32:11] <paragon36> :-) ... This AMD 64 should be able to handle 10000 you would think ?
[21:32:21] <SWPadnos> not necessarily
[21:32:26] <SWPadnos> CPU speed isn't the main factor
[21:32:34] <paragon36> The pluto looks just the ticket!
[21:32:37] <paragon36> No?
[21:32:39] <skunkworks> ray was getting lower base periods with a pci printer port wasn't he?
[21:33:10] <SWPadnos> yes - PCI parport can be faster (for any device hanging off it - PPMC, pluto, nothing ...)
[21:33:13] <paragon36> Oh ... paraport ... this one is on the motherboard!
[21:33:32] <skunkworks> But yes - if you like experimenting the pluto looks like a great deal.
[21:33:39] <skunkworks> * skunkworks has a pluto
[21:33:43] <paragon36> sure does
[21:33:45] <cradek> yes pci is faster, but before bothering, do the math and figure out if it will be enough help
[21:34:12] <SWPadnos> paragon36, I'm not sure how the AMD64 chips fare in RT-land (does anyone else have info on that?)
[21:34:43] <cradek> to use software counting successfully is a real compromise-juggling act
[21:34:48] <SWPadnos> I don't expect them to be worse than 32-bit chips, but I don't know how the interrupt controllers perform
[21:35:24] <cradek> ideally you figure it all out, then order exactly the right resolution of encoder, and it works great... surplus motors/encoders are a problem.
[21:37:18] <paragon36> I agree cradek ... But I picked up 11 pittman lo-cog motors up off ebay with internal 500 cpr optical encoders 24 volts ... look to be very good quality.
[21:37:36] <cradek> very nice!
[21:38:03] <cradek> one of these days I'll get 3-4 of those for my tabletop mill
[21:38:04] <skunkworks> paragon36: make sure you read through this link http://www.cnczone.com/forums/showthread.php?t=25929
[21:38:15] <paragon36> I do suspect as you suggest the limitation lies with the paraport
[21:38:44] <paragon36> Thanks skunkworks..... Im bookmarking like crazy ... lol
[21:39:16] <SWPadnos> the parport isn't the only thing. it's the amount of time it takes to respond to an interrupt (called latency)
[21:39:35] <skunkworks> paragon36: what machine do you have or plan on building?
[21:39:44] <SWPadnos> modern CPUs are optimized for throughput (ie, processing as much stuff as possible), not latency (ie, responding to an external event quickly)
[21:40:14] <paragon36> A CO2 laser cutter ...
[21:40:26] <skunkworks> oooh - nice
[21:40:33] <skunkworks> how many watts?
[21:40:52] <paragon36> 30 watts should be plenty to cut acrylic
[21:41:13] <skunkworks> right
[21:41:16] <alex_joni> hi guys
[21:41:23] <skunkworks> Hi alex
[21:41:26] <lerneaen_hydra> hi alex
[21:41:32] <paragon36> Hi alex
[21:41:38] <alex_joni> hello
[21:41:41] <skunkworks> alex is the guy that ran a 7000 period ;)
[21:41:52] <paragon36> Arrh right
[21:41:55] <alex_joni> skunkworks: not me :)
[21:42:01] <alex_joni> I was at 7500
[21:42:07] <skunkworks> ah ;)
[21:42:20] <alex_joni> or maybe I don't remember properl:)
[21:42:33] <paragon36> Driving a servo alex?
[21:42:47] <SWPadnos> as I recall, it ended up being 7 ticks of the clock, so it may have been ~6 uS :)
[21:43:52] <paragon36> Be back soon ... just going to read some more of the links you gave me .... Thanks All.
[21:44:04] <jepler> see you paragon36
[21:45:51] <cradek> bye
[21:47:29] <alex_joni> later paragon36
[21:47:33] <alex_joni> SWPadnos: right.. 6500
[21:47:46] <SWPadnos> score! :)
[21:48:05] <paragon36> Skunkworks I love the opening line of the link you sent me "Hi - my name is sam.. I am a cheap bastard. " lol = http://www.cnczone.com/forums/showthread.php?t=25929
[21:48:20] <SWPadnos> well he is :)
[21:48:26] <paragon36> :-)
[21:48:38] <lerneaen_hydra> haha
[21:48:54] <SWPadnos> though he was the second highest purchaser of Mesa gear ...
[21:49:05] <lerneaen_hydra> nice introductory text skunkworks
[21:50:23] <alex_joni> SWPadnos: see.. still proves he's a cheap bastard
[21:50:35] <SWPadnos> heh ;)
[21:50:39] <alex_joni> he would have gotten more expensive cards or directly from mesa otherwise :D
[21:50:53] <alex_joni> or he plans to sell them overpriced later
[21:52:14] <jepler> being cheap doesn't stop you from spending a lot of money :-P
[21:52:22] <jepler> at least, it doesn't stop me
[21:52:30] <SWPadnos> heh
[21:52:37] <skunkworks> jepler: exactly ;)
[21:52:40] <jepler> I start with the thing that costs 1/10 what the guaranteed solution will cost -- if that doesn't work, I try something 1/4 the cost .. and so on
[21:52:53] <jepler> when I'm done I've spent about 2x the cost, but it doesn't stop me from saying I'm a cheap bastard
[21:53:40] <lerneaen_hydra> skunkworks: nice use of a PC's powerled/hddled cable on your encoder ;)
[21:54:10] <lerneaen_hydra> it helps keep up the "I really am cheap" feeling :p
[21:55:01] <SWPadnos> speaking of cheap - I should buy a ferrule crimper that can do 8ga wire ;)
[21:55:11] <SWPadnos> (sometimes I'm reminded of opposites ;) )
[21:56:54] <skunkworks> lerneaen_hydra: thanks for noticing ;) I re-cycle
[21:57:06] <alex_joni> SWPadnos: what's a ferrule crimper?
[21:57:17] <SWPadnos> a tool for crimping wire ferrules
[21:57:33] <alex_joni> what are ferrules then?
[21:57:47] <jepler> neat: http://www.maushammer.com/systems/Watch/Build%20Log/4BDA6BB8-FCA6-4145-A83B-B9CFA21AD76B.html
[21:57:49] <lerneaen_hydra> ferrite thingamabobs?
[21:57:55] <skunkworks> a ferrules is what you use a ferrule crimper for
[21:58:11] <alex_joni> you all are soooo helpfull
[21:58:14] <skunkworks> ;)
[21:58:18] <SWPadnos> things that you crimp onto the end of a wire so none of the strands escape, and to provide a regular surface for screws/other terminals to grab onto
[21:58:32] <SWPadnos> http://cgi.ebay.com/WIRE-FERRULES-ALL-SIZES-LISTED-500PC-BAG-12-L-K_W0QQitemZ220066757821QQihZ012QQcategoryZ73141QQrdZ1QQcmdZViewItem
[21:58:37] <skunkworks> ah - I call it tinning ;)
[21:58:44] <paragon36> my mate had a ferrule he used to go hunting with it ! ;-)
[21:58:51] <skunkworks> so I can get solder migration
[21:58:59] <alex_joni> SWPadnos: oh those
[21:59:00] <SWPadnos> tinning isn't acceptable in most industrial situations
[21:59:00] <jepler> I don't think a ferrule is solder?
[21:59:33] <lerneaen_hydra> oh, those things
[21:59:52] <alex_joni> SWPadnos: I need a crimper for 95 mm^2 wire
[21:59:56] <lerneaen_hydra> definetly not a ferrite
[21:59:59] <SWPadnos> that's big wire :)
[22:00:38] <cradek> * cradek briefly considers trying to get units to tell him what gauge that is
[22:01:19] <SWPadnos> roughly 00
[22:01:41] <SWPadnos> (or 2/0)
[22:01:57] <cradek> ok
[22:02:02] <cradek> that's big
[22:02:03] <alex_joni> yes, it needs to take about 500A
[22:02:10] <alex_joni> at 40V or so
[22:02:36] <cradek> I bought some 0 gauge for battery cables (6 volt car)
[22:02:48] <cradek> modern cables wouldn't start it, too thin
[22:03:32] <SWPadnos> hmmm - 2/0 is only rated for 285A or so (in cabinets, less for transmission)
[22:03:55] <SWPadnos> even 4/0 isn't rated for 400A
[22:04:00] <SWPadnos> let alone 500
[22:04:08] <alex_joni> SWPadnos: might be bigger then 2/0 then
[22:04:09] <skunkworks> we had moved one of the services for our company - the wire was 1 inch in diameter (the actual copper) and we had rented a hydrolic crimper to put ring connectors on
[22:04:26] <skunkworks> hydraulic
[22:04:35] <lerneaen_hydra> O.O
[22:04:45] <SWPadnos> ah - it's closer to 3/0 - I was squaring the number instead of multiplying by pi or thereabouts :)
[22:04:45] <cradek> jeez
[22:04:48] <jepler> ooh he corrects his spelling now
[22:04:48] <lerneaen_hydra> 25mm wire is manly
[22:05:01] <jepler> bbl
[22:05:02] <cradek> and 1" is incrementally manlier
[22:05:19] <lerneaen_hydra> 25.4 then :p
[22:05:56] <alex_joni> skunkworks: something like that
[22:06:04] <SWPadnos> about 6% manlier, I hear
[22:06:10] <alex_joni> lol
[22:06:11] <skunkworks> :)
[22:06:41] <skunkworks> if I was really manly - I would have remembered the actual wire guage
[22:06:58] <skunkworks> 220-221 (mister mom reference)
[22:07:31] <SWPadnos> 4/0 is slightly less than 1/2", so that must have been something like 6/0 (if they even have ratings like that)
[22:07:48] <SWPadnos> actually, maybe 10/0 ;)
[22:08:21] <SWPadnos> http://www.technick.net/public/code/cp_dpage.php?aiocp_dp=guide_awg_to_metric
[22:11:10] <skunkworks> I will have to ask the resident electrition..
[22:11:56] <skunkworks> electrician
[22:11:59] <skunkworks> :o
[22:12:18] <lerneaen_hydra> skunkworks has almost got as bad spelling as I do :D
[22:12:37] <skunkworks> yes - but I don't think english is your first language ;)
[22:12:44] <lerneaen_hydra> um, actually it is
[22:12:45] <alex_joni> he's trying at least
[22:12:48] <lerneaen_hydra> that's the bad thing
[22:12:50] <skunkworks> really?
[22:12:54] <lerneaen_hydra> or at least it was
[22:13:00] <skunkworks> I figured sweedish
[22:13:02] <lerneaen_hydra> I lived in the US till I was 12-ish
[22:13:07] <skunkworks> ah
[22:13:35] <lerneaen_hydra> my english really has suffered recently, after having gone to a swedish highschool for three years and speaking practically no english
[22:17:20] <alex_joni> lerneaen_hydra: wanna hear something funny?
[22:17:30] <alex_joni> while I was in sweden I spok eonly english :-P
[22:17:41] <lerneaen_hydra> ...
[22:17:44] <lerneaen_hydra> and?
[22:17:54] <alex_joni> think you can too if you want :P
[22:18:03] <lerneaen_hydra> oh
[22:18:05] <lerneaen_hydra> that's true
[22:18:32] <lerneaen_hydra> I'm not sure if it will improve, I'll probably pick up an even more distinct swedish dialect :p
[22:21:41] <lerneaen_hydra> btw, anyone know of an app/script that will blank (output black) to a screen?
[22:21:45] <lerneaen_hydra> (linux)
[22:22:01] <lerneaen_hydra> multi-screen aware
[22:22:18] <SWPadnos> hide-my-porn.sourceforge.net
[22:22:20] <SWPadnos> :)
[22:22:26] <lerneaen_hydra> -.-
[22:22:44] <lerneaen_hydra> more like blank a second screen when playing a movie in the primary
[22:22:47] <lerneaen_hydra> err
[22:22:53] <lerneaen_hydra> s/a/the
[22:23:38] <SWPadnos> hmmm. I'm not sure of an app to do that. I know Wolf:ET blanks the second monitor when I'm playing fullscreen on the first
[22:23:57] <SWPadnos> so it's clearly possible for the movie player to do that for you
[22:24:36] <lerneaen_hydra> a system-wide solution would seem to be better
[22:26:03] <SWPadnos> maybe, maybe not. the application (movie player) knows how much screen real estate it needs but an external app may not
[22:26:34] <lerneaen_hydra> hmm, the app could just be defined to blank/darken screen 1 and/or 2
[22:28:15] <paragon36> Cradek ... I was just about to buy the pluto-P card from the us site it sent me to paypal.com and wanted to charge $9.95 for shipping do you know if this is for UK or US shipping?
[22:28:57] <paragon36> http://www.knjn.com/ShopBoards_Parallel.html
[22:30:39] <skunkworks> http://www.knjn.com/ShopInfo.html
[22:33:15] <cradek> paragon36: sorry, no idea
[22:33:21] <anonimasu> bork bork
[22:35:38] <paragon36> Bought it ... :-)
[23:17:24] <lerneaen_hydra> 'night
[23:26:42] <paragon36> cu all .... thanks for your help!
[23:27:10] <skunkworks> bye
[23:41:42] <A-L-P-H-A> anyone kicking around? could you go to [nothing to check, just need to check my logs]
[23:42:22] <A-L-P-H-A> nm... working again.
[23:42:30] <alex_joni> * alex_joni just did
[23:54:51] <A-L-P-H-A> alex_joni, 84.247.5x.xx??
[23:54:56] <A-L-P-H-A> that you?
[23:55:07] <A-L-P-H-A> thanks.
[23:55:13] <alex_joni> I think so :)
[23:55:22] <A-L-P-H-A> ips weren't resolving, but were said to be router itself.
[23:55:30] <alex_joni> yeah, that's me
[23:55:31] <A-L-P-H-A> going to make a LAMP server now, I think.
[23:55:42] <A-L-P-H-A> need a monitor now.
[23:55:42] <alex_joni> just found this site: http://freenet-homepage.de/palmill/index.htm
[23:55:49] <A-L-P-H-A> gonna grab my CNC machine's mon.
[23:56:01] <alex_joni> 2 freeware programs for simulating a mill and a lathe
[23:56:18] <A-L-P-H-A> www.cncsimulator.com
[23:56:20] <A-L-P-H-A> as well.
[23:56:22] <A-L-P-H-A> i used to use that a lot
[23:57:53] <alex_joni> yeah, but cncsim you need those pesky petrol files once in a while
[23:57:57] <A-L-P-H-A> yeah,
[23:57:59] <A-L-P-H-A> but at least it's free
[23:58:34] <jepler> annoyanceware: every 6 months, the software takes another 1% of your soul away
[23:58:38] <jepler> or maybe that's damnationware
[23:58:59] <A-L-P-H-A> you know... this isn't a bad idea... mobo (onboard net) + ram + cpu + flash usb keys (3-4 of them 4gigs each, 16 gigs)... that'd be nice little server.