#emc | Logs for 2009-08-10

[00:26:45] <DaViruz> exactly which spindle feedback is used during rigid tapping?
[00:27:36] <DaViruz> (in HAL terms)
[00:55:32] <SWPadnos> DaViruz, according to http://www.linuxcnc.org/docs/html/config_emc2hal.html it's motion.spindle-revs
[01:00:02] <DaViruz> that's what i thought, but it never started the Z feed, it just sat stil at the start coordinate
[01:03:21] <SWPadnos> does your spindle encoder have index?
[01:04:15] <SWPadnos> you need to connect motion.spindle-index-enable to the index-enable pin for the spndle encoder
[01:04:21] <SWPadnos> and index is required
[01:04:42] <DaViruz> i don't quite get that part, where exactly is the index pulse supposed to go?
[01:04:50] <DaViruz> to the enable pin?
[01:05:05] <SWPadnos> in HAL or the physical wire from the encoder?
[01:07:23] <greyscale> * greyscale points and laughs at jymm. Specificly, points at the powerhungry troll. Who acts like a dick to users who just need some help. Ass.
[01:07:56] <jymm> A ban from ##php
[01:08:20] <jymm> sorry
[01:16:59] <DaViruz> SWPadnos: in HAL
[01:24:25] <SWPadnos> DaViruz, ok. assuming that the physical wiring is correct, the index-enable pin is what tells the encoder module that it should reset position at the next detected index pulse
[01:25:29] <SWPadnos> the motion controller needs to know when the spindle crosses the index mark, so it sets the motion.spindle-index-enable pin true, then waits for it to go false again (the encoder module will set the pin false when an index has been detected)
[01:26:14] <SWPadnos> so all you need is something like "net spindle-index motion.spindle-index-enable my.driver.encoder.0.index-enable" in your HAL file
[01:28:39] <DaViruz> hm, ok
[01:29:25] <DaViruz> what's the purpose of that wghen accurate spindle feedback is available?
[01:30:33] <SWPadnos> there are several things you can do with spindle-synchronized motion, such as threading on a lathe
[01:30:57] <SWPadnos> if you don't know the orientation of the spindle, you can't do more than one pass and be sure it will line up correctly
[01:31:13] <SWPadnos> you end up with a multi-start thread
[01:31:31] <DaViruz> well, then the spindle feedback is inaccurate
[01:31:46] <SWPadnos> what is inaccurate about your encoder feedback?
[01:32:11] <SWPadnos> uh, wait. what are you calling "accurate spindle feedback"?
[01:32:12] <DaViruz> nothing, i want to know why it needs double feedback
[01:32:20] <SWPadnos> define double feedback
[01:32:37] <DaViruz> motion.spindle-revs and index
[01:32:41] <jymm> SWPadnos: It's 9 o'Clock - Have you gone out to eat?
[01:32:52] <SWPadnos> jymm, just had a bowl of cereal. thanks :)
[01:33:03] <jymm> SWPadnos: Where's the wifey?
[01:33:08] <SWPadnos> almost home
[01:33:19] <jymm> SWPadnos: you go out then?
[01:33:21] <SWPadnos> she started her day in Cleveland
[01:33:22] <DaViruz> spindle-revs is supposed to be increased by 1 for every revolution
[01:33:26] <SWPadnos> no, it'll be too late
[01:33:40] <jymm> SWPadnos: bday pizza?
[01:33:43] <DaViruz> and then it needs another spindle pulse per revolution.. :)
[01:34:12] <SWPadnos> DaViruz, yes, in theory you could fake an index pulse every time the position crosses an integer boundary
[01:34:47] <SWPadnos> in fact, that might even work OK, but that's not how we do it now ;)
[01:35:27] <DaViruz> so the index is pretty much redundant when spindle-revs is present? :)
[01:35:37] <SWPadnos> there has been some discussion about making a component that would fake an index every N pulses, specifically so cheapskates like you can avoid using a good encoder ;)
[01:35:41] <SWPadnos> no, it's not redundant
[01:35:49] <SWPadnos> it's an index mark
[01:35:59] <SWPadnos> it's independent of someone screwing up the scale
[01:36:22] <SWPadnos> it also resets the position to zero, which is necessary for the current synchronization scheme
[01:36:50] <DaViruz> if someone has inaccurate spindle-revs i don't see the point in using it at all
[01:37:02] <SWPadnos> sure, that would be a bigger problem ;)
[01:37:30] <DaViruz> if index can be faked it's redundant, imo.. :)
[01:37:39] <DaViruz> oh well
[01:38:15] <SWPadnos> that's a bit like saying limit switches aren't necessary since we have soft limits
[01:38:19] <SWPadnos> though not as dire
[01:38:31] <SWPadnos> and you should be able to do without homing too
[01:39:03] <DaViruz> nah, not exactly
[01:39:04] <SWPadnos> consider this - the orientation of the spindle isn't known at startup, so there should be some way to figure it out
[01:39:18] <DaViruz> as far as i can tell the index pulse is just used to start the motion, then spindle-revs is used exclusively
[01:39:26] <DaViruz> for g33.1 that is
[01:40:11] <DaViruz> ok, something like homing the spindle, well i can accept that
[01:40:19] <DaViruz> but i can't see why it would be mandatory
[01:47:18] <DaViruz> oh well, i guess i have to live with it :)
[01:49:13] <Guest297> anyone here?
[01:50:44] <Guest297> can someone help with "reboot into real-time kernel" and when I do, Im not able to connect to any wireless networks so basically i have no internet? Im using EMC2 on ubuntu 8.04
[01:51:06] <SWPadnos> on a laptop?
[01:51:12] <Guest297> no desktop
[01:51:24] <SWPadnos> with wireless only - curious
[01:51:43] <DaViruz> "integ", sounded like a nice hal component to make an integer, but of course it was integral :(
[01:52:03] <SWPadnos> there is a package with extra drivers, called something like "linux-ubuntu-modules"
[01:52:32] <Guest297> i dont have the option to hardwire to try to work around the wireless. okay ill look into that
[01:52:39] <SWPadnos> I believe there's a version for the RTAI kernel we publish, so you might try installing that (you should be able to download it)
[01:53:07] <SWPadnos> err - you should be able to download it with non-RTAI, then install it locally - there shouldn't be any unmet dependencies
[01:53:09] <SWPadnos> bbiab
[01:53:36] <Guest297> thanks
[03:39:23] <tom1> tom1 is now known as tom3p
[03:46:29] <tom3p> http://imagebin.ca/view/f_lk2PEo.html vismach model of 7axis drill
[03:46:52] <tom3p> is there any cgi for vismach? ( any way to subtract one volume from another ) ?
[03:51:01] <tom3p> or a way to make a circular wedge ( pie slice )?
[04:06:16] <geo01005> Is that a dill EDM?
[04:19:23] <fenn> no CSG in vismach that i'm aware of; i think alex was working on importing .stl or .stp files at one point
[04:21:02] <fenn> why is it all fuzzy?
[04:25:10] <tom3p> http://imagebin.ca/view/XpULE6.html i found arcx(x1,x2,r1,r2,a1,a2,step) now rotary counterbalance is more like real one
[04:25:43] <tom3p> fenn: fuzzy... i painted it on velvet ;) no, some gl artifacting at larger scale i guess. real zigzaggy at some zooms.
[04:26:27] <tom3p> ah yes CSG
[04:27:04] <fenn> how did you make the pie slices?
[04:27:10] <fenn> oh nm
[04:29:10] <tom3p> arcX
[04:37:23] <fenn> tom3p: are you going to contribute that drill's vismach file and config/kinematics? (just curious)
[04:49:54] <tom3p> fenn: it doesnt belong to me, so its gonna die soon at imagebin. it doesnt need fancy kins tho, cuz there are no 'synthetic' axis.
[04:50:24] <tom3p> http://www.opencsg.org/
[04:51:16] <tom3p> as for scale, there's 3 steps to get up to the work tank
[04:51:17] <fenn> yeah i just looked at that page yesterday; i've been trying to use opencascade for swept volumes and it's failing miserably
[04:51:59] <fenn> hm. i guess i won't trumpet your screenshots to the entire internet then
[04:52:12] <tom3p> oh, nm
[04:53:04] <fenn> nm meaning i should trumpet them?
[04:53:07] <tom3p> i will animate that vismach eventually, and maybe take the time to figger who owns what here. we all just kinda shook hands.
[04:53:28] <tom3p> nah, i havent had time to think yet
[04:54:29] <tom3p> you still in Austin?
[04:54:29] <tom3p> i got back from taiwan, had 36 hrs at home, been in cincinnati ever since
[04:54:29] <tom3p> and i gotta hit the rack.
[04:54:32] <tom3p> gnite
[04:56:40] <fenn> poof
[06:49:08] <jymm> hey alex_jon1
[06:50:41] <alex_jon1> hi jymm
[06:51:00] <jymm> whats goin on?
[06:51:19] <alex_joni> just got to work
[06:51:24] <alex_joni> restored my net connection
[06:51:31] <alex_joni> need to drop this extra irssi ;)
[06:51:53] <jymm> oh, are you mookie?
[06:52:14] <alex_jon1> nope
[06:52:31] <alex_jon1> * alex_jon1 thinks jymm needs some coffee||sleep ;)
[06:52:38] <alex_joni> i!=1
[06:53:30] <jymm> Well, I didn't know which "extra" connection you were tlaking about, and mookie is the only oe i saw that was an issue.
[06:54:07] <jymm> Besides, couldn't you just shell in and kill it?
[06:54:24] <alex_jon1> sure, need to read the scrollback first though
[06:54:30] <jymm> why?
[06:55:18] <alex_jon1> alex_jon1 is now known as alex_joni
[06:58:44] <micges_work> hi alex
[06:59:06] <alex_joni> hi micges_work
[07:12:56] <jymm> alex_joni: You know in those cheap pencil soldering irons... what do they use to insulate between the layers of the resistance wire?
[07:13:11] <jymm> and why dont the coils short anythign out
[07:18:57] <alex_joni> no idea
[07:19:11] <jymm> you know what I'm alking about?
[07:19:16] <jymm> talking
[09:13:42] <piasdom> g'mornin all
[10:18:32] <Valen1> Valen1 is now known as Valen
[12:36:14] <Vq_> Vq_ is now known as Vq
[13:03:30] <skunkworks823> where was that cool picture of cradeks probing a sphire?
[13:04:14] <Valen> that sounds dirty
[13:04:27] <archivist> geek porn
[13:04:43] <skunkworks823> get your mind out of the gutter
[13:06:47] <jepler> skunkworks823: http://timeguy.com/cradek-files/emc/smartprobe.png
[13:08:17] <archivist> * archivist contemplates real probe diameter and side forces
[13:10:55] <skunkworks823> jepler: thanks!
[14:33:11] <ilya_> Does this guide remind all those `usual' books? www.sherline.com/5300inst.htm
[14:34:09] <ilya_> Oh, no, it must be ordered
[14:41:56] <Valen> http://www.youtube.com/watch?v=aufL76bXLAg&feature=related
[14:48:43] <skunkworks823> heh - people just give you things?
[14:54:38] <ilya_> www.sherline.com/grinding.htm
[15:46:17] <mikegg> what is a reentrant function?
[15:46:54] <mikegg> EMC does not like "addf hm2_5i20.0.read servo-thread"
[15:47:20] <archivist> a function that can safely call itself
[15:52:49] <ilya_> archivist: How do you know programming and electronics?
[15:53:13] <archivist> training
[15:53:14] <ilya_> archivist: you're probably a wise old man ;)
[15:53:43] <jymm> wiseass ancient fart is more like it.
[15:54:33] <jepler> mikegg: in emc terms, a "reentrant" function is one that can be added to a thread multiple times
[15:54:45] <jepler> "reentrant" functions were part of the original hal design, but I don't think any functions are actually designated reentrant
[15:55:31] <skunkworks823> jymm: how long to bans last?
[15:55:44] <archivist> jymm dont be cheeky to your elders
[15:55:48] <jymm> skunkworks823: in where?
[15:55:58] <skunkworks823> php - (last night)
[15:56:15] <jymm> skunkworks823: you mean greyscale?
[15:56:16] <archivist> till some nice op unbans
[15:56:25] <mikegg> that line throws an error "...'hm2_5i20.0.read' is not reentrant"
[15:56:47] <jymm> skunkworks823: Oh, that would be a permban
[15:57:10] <jepler> mikegg: probably some other line in your hal has also added .read to a thread. consider this example I just did interactively: http://pastebin.ca/1524016
[15:57:21] <skunkworks823> jymm: yikes
[15:58:19] <jymm> skunkworks823: He's someone that abused previously was banned, came back as another nick, was warned, contiued, asked not to continue, still did, that just went postal.
[15:58:34] <jymm> s/that/then/
[15:59:16] <mikegg> ah hah. i found it
[16:02:44] <mikegg> thanks guys
[16:08:10] <ilya_> skunkworks you've probably said "html is better than php"
[16:08:29] <ilya_> skunkworks apologize then
[16:08:34] <jymm> ilya_: that doens't make any sense.
[16:09:11] <ilya_> jymm: yeah, I'm suck in programming :) ...
[16:10:42] <ilya_> archivist: How can I train to know programming using a bicycle or a skateboard?
[16:11:03] <archivist> speak sense or dont bother
[16:11:24] <ilya_> ok
[16:41:51] <antichip> good morning everyone
[16:49:08] <antichip> is anyone able or willing to walk me through a 9.04 install with rtia and cvs build of emc2?
[16:49:52] <antichip> or should I just run 8.04
[16:50:27] <jepler> antichip: you should just run 8.04.
[16:51:14] <jepler> cradek: have you put your "force use of software opengl" trick on the wiki?
[16:51:39] <cradek> nope
[16:51:43] <jepler> you should
[16:51:53] <cradek> it's in the list archives somewhere
[16:52:25] <jepler> aha, http://mid.gmane.org/20090705170235.GB14830@timeguy.com
[16:53:59] <antichip> I would still have to get 64 bit support? Will it be easier to use a standard distro and then build emc for the newer cvs or experimental build? am I making sence
[16:55:18] <jepler> antichip: 32-bit ubuntu works fine on 64-bit-capable machines.
[16:59:27] <antichip> I would still have to get 64 bit support? Will it be easier to use a standard distro and then build emc for the newer cvs or experimental build? am I making sence
[16:59:47] <antichip> sorry
[17:00:34] <antichip> so I use 32 bit 8.04, no performance gain with 64 bit modes?
[17:02:19] <jepler> very few users are using a 64-bit OS with emc2, so we don't have a lot of information. But it looks like as far as maximum software stepping rates go there's no huge advantage to systems running in 64-bit mode
[17:03:02] <antichip> Last I knew the cvs emc builds were the only way to get the single pulse spindle encoder. Is this still true?
[17:03:02] <jepler> if your interest is in using emc2 to run a cnc machine, then the shortest path to point B is the 32-bit live cd.
[17:03:26] <jepler> no, that's in 2.3.x which is the current stable release version.
[17:03:52] <antichip> oh well then sh!t
[17:04:14] <jepler> (also, we've been using git for a few months, not cvs: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Git)
[17:04:37] <jepler> bbl, it's lunchtime here
[17:04:59] <antichip> I'll be here ;-)
[17:05:26] <antichip> maybee with progess or questions
[17:05:32] <antichip> thank-you
[17:39:27] <frallzor> wooow is ilya at it today too =)
[19:31:06] <micges> http://yro.slashdot.org/story/09/08/06/2322209/Microsoft-Patents-XML-Word-Processing-Documents?from=rss
[19:47:52] <mikegg> arg. anyone know what causes a "hm2/hm2_5i20.0: encoder.00 dT <=0, how can this be?" error?
[19:52:48] <jerk> What if I assign a relatively low feedrate for, say, a drilling -- while I could use some quicker one, as for less rigid steel?
[19:54:51] <archivist_attic> you must apply engineering knowledge to drilling speed
[19:55:41] <jerk> I surely should.
[19:57:54] <jerk> I think there wouldn't be a surface enforcement has a place only If I assign unreasonable low feedrate, when the cutter actually strikes the same coordinate and doesn't actually advance.
[19:58:25] <cradek> mikegg: that is fixed in later emc2 versions. maybe you need an update. what version are you using?
[20:00:52] <jerk> s/be//
[20:01:11] <mikegg> 2.2.8
[20:01:27] <jerk> i meant it would. But if I don't, it wouldn't.
[20:01:59] <cradek> mikegg: I know for sure that's fixed in 2.3.3
[20:02:21] <mikegg> ok cool. thanks!
[20:02:58] <cradek> here is the list of fixed things (but, I don't see it there, which is very odd) http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Released
[20:03:12] <cradek> here are instructions for updating from 2.2 series to 2.3 series: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?UpdatingTo2.3
[20:04:18] <cradek> I guess several things say "lots of hostmot2 improvements" etc
[20:05:07] <mikegg> ah, excellent. was just fumbling through the synaptic package manager...heh
[20:06:16] <jerk> cradek: when a lamer asks you about the drilling and milling feedrates taken from those books for huge machines with frames made of ingot steel, would you agree with those advices on feedrates, or oyu would say that the rigidity is a problem form the most of the cheap machines?
[20:06:53] <cradek> the books often give numbers not appropriate for hobby size machines
[20:06:59] <jerk> *you and *for
[20:09:24] <jerk> cradek: I think I should start from really low feedrates then, and write down my comments about the material, thickness,diameters, and chosen feedrates -- to finally choose appropriately quick feedrates?
[20:09:39] <cradek> experimenting is always good
[20:10:02] <cradek> although sometimes going too slow is bad - keep that in mind
[20:10:13] <jerk> cradek: will the machinery's manual of the particular machine contain such tables or information?
[20:10:23] <cradek> doubtful
[20:10:25] <jerk> cradek: why is it bad?
[20:10:49] <cradek> some materials harden if the cutter rubs across them instead of cutting. also, rubbing wears a cutter faster than cutting does
[20:10:51] <jerk> Do they contain doubtful (i.e. useless) information?
[20:13:40] <jerk> cradek: saying about the steel (and i drilled many holes with manual drill), only a very low feedrate whch strikes the hole or the side could harden the steel -- numbers like 0.02 mm/revolution already isn't a hardening.
[20:14:08] <cradek> yes regular steel does not really work-harden
[20:14:19] <jerk> Hardening of metal relies on pressure.
[20:14:38] <jerk> cradek: then which steel should I be awared of?
[20:14:56] <jerk> jerk is now known as good_kid
[20:14:58] <cradek> stainless and titanium like to work-harden
[20:15:17] <cradek> I'm sure not a materials expert
[20:15:35] <good_kid> cradek: I have a book about hardening.
[20:17:36] <good_kid> Well, there different means are described, but it seems hardening needs some force applied, while the feedrate, even the small, one, sounds like a guaranteed advance and ....
[20:18:06] <good_kid> cradek: Anyway, have you ever had problems with hardening in your work?
[20:18:16] <cradek> brb
[20:18:21] <good_kid> ok
[20:20:57] <good_kid> ...and if there's a feedrate, it is never a hardening.
[20:26:32] <good_kid> But some pressure and relatively long (in comparison with an advance by means of any assigned feedrate) influence is always a surface hardening in a normal direction.
[20:34:38] <good_kid> For example, some roll with a stiff spring (where both the roll size and the corresponding stiffness are ´╗┐empirically standadrdised); or somple reducing the microrelief by means of a diamond stroking, where the instrument or the appliance is rather applied or acts then actually guaranteed to move by some *external* to this interation in this approach force.
[20:36:14] <good_kid> *or simple reducing of the micro-relief by means of the ...
[20:38:28] <good_kid> *... than, ...interaction
[20:38:41] <good_kid> o.m.g.
[20:44:07] <good_kid> Ok, certain materials have certain thickness for surface hardening.
[20:58:04] <fenn> good_kid: you should probably get "machinery's handbook"
[20:58:59] <good_kid> for a particular machine with some common tables and advices? From where?
[21:11:26] <good_kid> brb a.s.c.a.w.r -- after some coffee and washing rourine
[21:20:01] <toastydeath> i just brought home a 10 flute, 3" diamter, 10" long endmill
[21:20:10] <toastydeath> with an intergal NMTB 50 taper
[21:20:47] <toastydeath> roughing, no less
[21:21:17] <archivist> next bring home a mill to use it
[21:22:18] <toastydeath> haha i know, right
[21:22:23] <archivist> a truck and crane may be required
[21:23:30] <good_kid> toastydeath: are you also a toastatwork?
[21:25:18] <frallzor> why the new nick ilya?
[21:26:31] <good_kid> frallzor: because my other nicks (ilya_, ilya__, jerk, dirty_underware, and new_underware) aren't as good as this one.
[21:27:06] <archivist> we dont need two new nicks a day
[21:27:25] <good_kid> my #lexical_sexuality for discussing single female lawyers in skirts with no pants is a suck, too.
[21:30:45] <good_kid> ´╗┐/msg ChanServ DROP #lexical_sexuality 5c9c0c82:a0903916
[21:30:49] <good_kid> omg
[21:32:06] <Xteven> fail
[21:32:21] <good_kid> they add some invisible characters
[21:35:06] <toastydeath> good_kid: yes
[21:36:50] <good_kid> * good_kid puts down the butter and his coffee cup. -- "It's OK then... There are too much of tasty bakery around this summer..."
[22:53:08] <antichip> what and or where are the different branches of emc?
[22:56:06] <good_kid> Can I update my offline logs with, for example, `wget -r -c http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2009-08-10.txt" each next day after the GMT+0/
[23:02:10] <toastydeath> has anyone taken apart an AC unit before
[23:03:01] <skunkworks_> central air or window?
[23:03:06] <skunkworks_> what is the issue
[23:09:49] <toastydeath> window
[23:09:58] <skunkworks_> really not much you can do. the compressor is hermetically(sp) in sealed in both units
[23:09:58] <toastydeath> i want to take an ac unit apart and put the radiator in the bathtub
[23:10:25] <skunkworks_> ? the hot side or cold side?
[23:10:30] <toastydeath> hot side.
[23:10:33] <toastydeath> in a cold tub.
[23:10:37] <skunkworks_> in the house?
[23:10:43] <toastydeath> ...yes?
[23:10:51] <toastydeath> put the cold side outside the bathroom.
[23:11:05] <skunkworks_> are you looking for efficency?
[23:11:18] <toastydeath> only partially, water is free for me, power is not
[23:11:20] <skunkworks_> or to heat the tub water for a bath?
[23:11:33] <toastydeath> no, i was planning on letting it leak out
[23:11:44] <toastydeath> with a bit of a drip from the faucet to refill
[23:11:48] <skunkworks_> unless you are constantly running the water - the heat is going to make it back into your house..
[23:11:52] <toastydeath> yep
[23:11:54] <toastydeath> that's the plan.
[23:12:36] <toastydeath> not sure what else to do, i'm on the top floor of an apt building with horrible insulation
[23:12:48] <toastydeath> buying two or three window AC units would be prohibitive
[23:13:18] <toastydeath> i need to look and see if I can force them to replace the unit, i suppose
[23:13:37] <skunkworks_> yeck
[23:15:17] <toastydeath> i can't find anything about what the landlord is responsible for
[23:15:20] <toastydeath> insofar as upkeep