#emc | Logs for 2009-09-25

[00:04:51] <MarkusBec> MarkusBec is now known as MarkusBec_away
[00:54:33] <frallzor> is there a certain MPG that emc has native support for or is it samesame?
[01:37:09] <__PyNewbie__> i was working on a reply for him.
[01:52:42] <eric_unterhausen> what was the gist of your reply?
[01:55:23] <__PyNewbie__> very new to hal, but i would think for any pulse gen you would need to add an encoder - loadrt encoder2 num_chan=1, addf encoder2.update-counters base-thread
[01:55:58] <__PyNewbie__> addf encoder2..capture-position servo-thread
[01:58:14] <__PyNewbie__> or is maybe the original (loadrt encoder num_chan=1) is all, and then use encoder.1.scale, mode, etc. this is where im still soft with hal config.
[02:01:22] <__PyNewbie__> then you need some 'net mpg-phase-a encoder.1.phase-A' or something.
[02:02:20] <eric_unterhausen> or the short answer is that all mpgs require at least some minor hal script programming at this time.
[02:05:05] <__PyNewbie__> sure, if its not in stepconf then you gotta dig in, thats where i am at now.
[02:20:59] <tomp> tomp is now known as tom3p
[02:44:08] <__PyNewbie__> after a some testing, 1 pulse per rev threading works, but i sure want more pulses. any tips on encoder pulse spacing? keeping the on/off time the same makes sense (ie, .125hole, .125space).
[03:50:29] <__PyNewbie__> cya folks, thanks again
[08:34:21] <cnc_valen> question for you gcode wizards
[08:34:50] <cnc_valen> cutting a bearing housing atm with just the x and y tables running
[08:34:58] <cnc_valen> Z is manual
[08:35:13] <cnc_valen> (its a bering housing for zthat isbeing cut)
[08:35:37] <cnc_valen> using cutter compensation to do the cut and its working well
[08:36:06] <cnc_valen> only problem is we are driving back to the center after doing the circle
[08:36:18] <cnc_valen> and then back out to the circle to do the cut
[08:36:45] <cnc_valen> with the cutter compensation it leaves a little wedge where it starts and stops the circle
[08:37:41] <cnc_valen> is there an easy/correct way to get it to complete the circle with cutter comp on before returning to center?
[08:38:14] <cnc_valen> currently we have this
[08:38:24] <cnc_valen> G41
[08:38:28] <archivist> follow it with an arc then center
[08:38:35] <cnc_valen> G1 X0 Y-13
[08:38:45] <cnc_valen> G3 X0 Y-13 I 0 J -19
[08:39:43] <cnc_valen> hmm any easy way to work out the end of the arc?
[08:40:14] <archivist> maffs :((
[08:40:58] <archivist> this is where do me an extra 2 degrees is whats needed
[08:40:59] <cnc_valen> i'm only a rocket scientist i suck at maths
[08:41:20] <cnc_valen> (my degree is space science ;->)
[08:41:43] <cnc_valen> so i am certified to say what is and is not infact rocket science ;->
[08:41:47] <archivist> the leadin and leadout to a job is important
[08:43:49] <archivist> its the blending , you could try the exact mode so it does not blend the two moves together
[08:44:06] <cnc_valen> we are trying to use the cutter compensation
[08:44:16] <cnc_valen> its mainly a learning exericise
[08:44:48] <cnc_valen> trying to learn the right way to do things before we start cheating ;->
[09:11:20] <Mire> does it make sense for an encoder to have only 2 outputs?
[09:19:44] <Mire> I'm attempting to interface my existing spindle encoder to a 7i33. if I find only 2 outputs does it mean +A, +B, and wire the enable low?
[09:21:27] <cnc_valen> something like that
[09:21:41] <cnc_valen> problem is emc uses the index pulse on spindle operations
[09:21:59] <cnc_valen> or at least i believe it does
[09:22:14] <cnc_valen> check with somebody who knows more before acting ;->
[09:33:35] <Mire> Is anybody near with experience with Pico systems resolver converters?
[09:39:52] <alex_joni> cnc_valen: depends what you need to do
[09:40:03] <alex_joni> for axis feedback you don't need index
[09:40:23] <alex_joni> you need index on a spindle if you want threading or rigid tapping
[09:40:38] <alex_joni> you need index on an axis feedback if you want to home on index
[09:40:51] <alex_joni> Mire: depends what you need to know
[09:45:09] <cnc_valen> alexi_joni he said it was a spindle encoder
[09:45:32] <Mire> resolver converter docs hint at a way to set excitation freq. jumpers seem only to correspond to amplitude. Does it matter? machine is Hardinge HNC.
[09:49:02] <Mire> spindle has resolver and encoder both, but if there is index pulse i havent found it. Hope I can just use it as is...
[09:58:37] <cnc_valen> oh just to add to the general knowledge here we have blown the fuses in our motor wires several times now with no ill effects
[09:58:57] <cnc_valen> I'll put a pic up of my board mods tonight
[09:59:24] <MarkusBec_away> MarkusBec_away is now known as MarkusBec
[10:09:43] <alex_joni> Mire: maybe wait for a while
[10:09:53] <alex_joni> there are some US guys that run a hardinge HNC
[10:10:13] <alex_joni> regarding the resolver converter, just email Jon Elson, I'm sure he'll reply quickly
[10:20:29] <Mire> I love how personal all this is. of course Jon can tell me all i need, I'll ask him.
[10:38:45] <alex_joni> well.. we're just a bunch of guys.. it's surely personal ;)
[14:02:47] <mozmck_work1> mozmck_work1 is now known as mozmck_work
[15:43:06] <cradek> urgh, renishaw wants $368 for a BT40 holder
[15:44:07] <archivist> they had to lay off a bunch recently
[15:45:00] <cradek> I might try to cut up a $50 shell mill holder instead - a pain, but not rocket science
[15:46:25] <archivist> Im learning how easy it is to blow fuses on the cnc lathe
[15:46:32] <cradek> uh-oh
[15:46:46] <cradek> spindle?
[15:47:10] <archivist> I think the originals are set close for the educational market and not an engineers use
[15:47:45] <archivist> take a good bite and the thing stalls and pops a fuse
[15:48:00] <archivist> safe I suppose
[15:48:07] <cradek> need more spindle!
[15:48:29] <archivist> magnetic circuit breaker next
[15:56:36] <skunkworks_> cradek: Do you have a probing project?
[15:57:01] <cradek> I have a probe, or most of one, that I want to use on the jr
[15:57:21] <cradek> I'll probably just use it for locating edges and holes - nothing fancy
[15:57:55] <skunkworks_> because it is there! :)
[15:59:53] <cradek> I think the mill is good enough to make a decent CMM
[16:00:09] <cradek> I will have to measure some gage blocks
[16:00:26] <skunkworks_> I bet.
[16:08:11] <archivist> I have a bt50 mounted probe its just too effing big to use on my little mill
[16:10:03] <skunkworks_> cradek: is the probe wireless?
[16:10:17] <cradek> skunkworks_: yes
[16:10:45] <skunkworks_> what is the speed of it?
[16:10:54] <archivist> mine has flashy leds
[16:10:56] <cradek> but I can't let the tool changer move it without paying $368 first
[16:11:23] <cradek> skunkworks_: 9.8 m/s^2 if I drop it
[16:11:34] <cradek> (I don't know what its response time is)
[16:11:34] <skunkworks_> heh
[16:11:51] <skunkworks_> ah - I didn't know if my question was too vauge ;)
[16:12:03] <archivist> 1.0333 secs to expensive
[16:12:19] <skunkworks_> I would be afraid of crashing it :)
[16:12:30] <cradek> skunkworks_: way under a msec I bet
[16:17:24] <skunkworks_> yes
[16:19:13] <cradek> skunkworks_: I think a small crash (snapping the stylus off) is < $100 to recover
[16:19:40] <cradek> if I turn maxvel down, I think I can manage to not break it
[16:21:22] <skunkworks_> http://www.youtube.com/watch?v=gtcF5BxciCI&feature=PlayList&p=57A30635D42102D5&playnext=1&playnext_from=PL&index=66
[16:22:08] <skunkworks_> boy the thing kids have now.... http://www.youtube.com/watch?v=rEp7xjhnDYo&feature=PlayList&p=57A30635D42102D5&index=67&playnext=2&playnext_from=PL
[16:23:18] <cradek> wow that looks weak - wonder what it cuts
[16:29:21] <skunkworks_> foam I would guess
[16:31:14] <cradek> ah, that could work
[17:38:14] <mboehm> hello, is it possible to change hal connections after homing sequence?
[18:14:17] <jepler_> mboehm: perhaps you want to use mux2/mux4 to switch based on halui.xxxx.is-homed
[18:14:25] <jepler_> usually hal connections aren't changed after startup
[18:14:37] <jepler_> it's also possible to home an axis more than once per session
[18:15:03] <jepler_> (this is useful for example in a stepper system for recovering after a stepper has stalled for whatever reason)
[18:15:17] <cradek> mboehm: can you say what you're trying to do, in a more general sense?
[18:19:56] <mboehm> i want to use linear motors and do the commutation via hal
[18:20:11] <mboehm> when i start up i have no position information
[18:20:45] <mboehm> so i use the motor like a stepper with microsteps
[18:21:11] <mboehm> after homin i have the position an can calculate the commutation
[18:21:58] <cradek> I don't know enough about this topic to make smart suggestions
[18:24:16] <mboehm> so i must reconfigure hal after homing
[18:25:16] <cradek> are "homing" the motors (orienting the commutation signal with the drive signal) and homing the joint related? it seems like two different tasks
[18:26:06] <cradek> I thought AC motors would orient themselves, possibly with a little position jump, when started up - are linear motors different or am I mistaken?
[18:31:54] <robh_> heres a nice EMC project for UK guys http://cgi.ebay.co.uk/XYZ-Vulcan-810-Vertical-Machining-Centre_W0QQitemZ190335752818QQcmdZViewItemQQptZLH_DefaultDomain_3?hash=item2c50e51a72&_trksid=p3286.c0.m14
[18:33:47] <cradek> isn't there supposed to be a Y axis motor there...?
[18:35:07] <cradek> that's sure going to go cheap I bet
[18:35:09] <mboehm> the jump is up to 20mm and for proper operation i need a constant distanc between the magnetic poles of the "rotor" an the stator
[18:35:34] <robh_> cradek, its probly on the back of the machine
[18:36:27] <cradek> huh, those ballscrews are tiny compared to the ones on my (smaller) machine
[18:36:46] <cradek> robh_: bet you are right
[18:37:21] <mboehm> its like oprating a stepper motor with microsteps in torque mode and the motor has very big steps
[18:37:56] <robh_> it is a X Y Z so not very upmarket range ;)
[18:40:15] <robh_> archivist, i did post a bit on mailing list but not been many takers yet for a UK meet/get to gether
[18:43:17] <archivist> Im watching it
[18:44:04] <archivist> I should answer it to show interest and so should pjm :)
[18:44:38] <robh_> yes ture, maybe it would spur the others 2 wake up
[18:45:24] <archivist> I can nag ian in sheffield as well
[18:47:37] <robh_> grate so thats 4 of us then
[18:47:39] <archivist> I was thinking as its impromptu meeting one of us should print a logo and sit in the cafe area
[18:49:40] <robh_> holding it up high lol always get some dolly bird to walk around with it that pull a crowd
[18:50:16] <archivist> hmm wanted one renta tart
[18:50:57] <robh_> any one through of writting some probing cycles
[18:55:35] <alex_joni> mboehm: you should be able to reconfigure HAL after homing
[18:55:48] <alex_joni> but if that causes bad things to happen, you get to fix it ;)
[18:56:05] <alex_joni> I would do like jeff suggested, use a mux and switch based on is-homed
[18:56:55] <jymm> Hey, i think we need to move EMC from ubuntu to this... http://hannahmontana.sourceforge.net/Site/Home.html
[18:57:57] <micges> ouch..
[18:58:02] <jepler_> for those of us who are afraid to load that page, can you tell us a bit more?
[18:58:25] <jymm> jepler_: load the page you wimp!
[18:58:35] <mboehm> lol
[18:58:37] <jepler_> what if it affects my system with mallware?
[18:59:00] <jymm> jepler_: Then stop using windows and install it!!!
[18:59:38] <cradek> uh is that for real?
[18:59:56] <jymm> Sadly, yes.
[19:00:11] <cradek> I even read the FAQ
[19:00:18] <cradek> and "is this for real?" is not question #1
[19:00:44] <cradek> hahaha http://hannahmontana.sourceforge.net/Site/Song.html
[19:01:18] <jymm> You know s/he made it for their daughter
[19:03:19] <jymm> So, when will you convert emc to it? huh? huh? huh? Then you can bring the youngns into CNC'in !!!
[19:03:44] <jepler_> jymm: you must have us confused with people who are interested in wide adoption of emc
[19:04:03] <cradek> I'm sure you can already build emc on hannahmontanalinux
[19:04:39] <alex_joni> it's kubuntu
[19:04:47] <alex_joni> so I'm sure our packages work without changes
[19:05:19] <jymm> But, AXIS needs to be altered to keep with the HM theme
[19:05:31] <jepler_> one of axis's features is that it is absolutely unthemable
[19:05:32] <jymm> maybe a bouncing head instead of a cone
[19:06:18] <jymm> jepler_: Oh speaking of that, how's the modularity of AXIS coming along?
[19:07:11] <jepler_> jymm: the git branch I created for that purpose is out there
[19:07:51] <jepler_> as far as I know, nobody intends to build anything on it
[19:07:54] <jymm> jepler_: No, I was jsut asking how it was going is all.
[19:08:06] <jepler_> jepler_ is now known as jepler
[19:08:15] <jepler> I don't plan at the moment to do anything more with it
[19:08:37] <jymm> jepler: ah, ok.
[19:11:04] <micges> jepler: nobody is a bad word ;)
[19:12:06] <micges> in a few days I be able to include gremlin code to my project
[19:12:29] <jepler> I stand corrected.
[19:13:53] <mozmck_work> so you don't have to disable the parport_pc module now? does that mean I can use my parport for something else if emc is not using it at the moment?
[19:14:24] <jepler> mozmck_work: if the only version of emc you want to run is 'master', yes.
[19:15:32] <mozmck_work> that's what I'm running for testing on my main computer at home. I also have a parport based programmer that I need to use so this will be nice.
[19:24:30] <micges> ``
[19:37:14] <pjm> archivist that VMC on ebay looks nice, its only down the road from me too!!!
[19:43:12] <alex_joni> http://www.theregister.co.uk/2009/09/25/open_source_code_audit/
[19:51:44] <jymm> Sears is selling Cannibal bbq's! HAWT DAMN!!! I'll take two please! http://www.theregister.co.uk/2009/08/21/sears_baby_roaster/
[19:55:01] <ZeroFlex> brppz0rs
[19:56:50] <archivist> pjm which one?
[19:58:52] <pjm> the vulcan 810
[20:02:53] <robh_> not sure what drives that has in it but should be easy enuth to interface if ucan find the manuals
[20:04:30] <archivist> tempting toy
[20:07:24] <robh_> ifi t had had random tool changing then it would be sweet little toy at 24m/min
[20:08:32] <robh_> what gets me is how much ppl spend on orac or denford compared to something like that vmc lo
[20:09:14] <archivist> yup they only want a "small" machine by a name theu know
[20:28:08] <MarkusBec> MarkusBec is now known as MarkusBec_away
[20:30:48] <MarkusBec_away> MarkusBec_away is now known as MarkusBec
[22:31:27] <MarkusBec> MarkusBec is now known as MarkusBec_away
[22:58:25] <MarkusBec_away> MarkusBec_away is now known as MarkusBec
[23:14:52] <MarkusBec> MarkusBec is now known as MarkusBec_away
[23:38:52] <MarkusBec_away> MarkusBec_away is now known as MarkusBec