#emc | Logs for 2007-02-22

[00:08:33] <LawrenceG> Jymmmm: whats up?
[00:09:43] <LawrenceG> -.-. --.- -.. . ...- . --... .. -
[00:11:59] <SWPadnos> CQDE3E7IT?
[00:12:44] <LawrenceG> close ... cq de ve7it
[00:13:23] <LawrenceG> kind of quiet in here today
[00:13:45] <SWPadnos> yep
[00:21:32] <Jymmmm> LawrenceG hey , not a whole lot, SSDD =)
[00:22:12] <LawrenceG> did you get your crimper?
[00:22:36] <Jymmmm> they culdn't find them, calling the mgr as we speak
[00:23:25] <Jymmmm> busy, guess 85,000 others wanted them too =)
[00:24:36] <Jymmmm> SWPadnos Why are you coming out, conference?
[00:24:45] <Jymmmm> (or just for the Prime Rib =)
[00:25:05] <SWPadnos> yep, same one as last year
[00:25:17] <SWPadnos> the conference. hopefully they'll have fresh prime rib ;)
[00:25:39] <Jymmmm> they say aged 14 days
[00:25:47] <SWPadnos> that's fresh enough for me
[00:26:05] <Jymmmm> lol, woudl you rather have it still moo'ing?
[00:28:51] <jtr> Jymmmm: disk grinder option - http://www.abana.org/ronreil/Forge1.shtml#Grinder
[00:30:09] <jtr> I'm chicken - running a 1725 RPM motor on mine.
[00:30:55] <Jymmmm> heh
[00:32:02] <jtr> Does spin very smoothly. Got lucky on a tilting work table for it at Cabin Fever in 2005. $10
[00:33:47] <Jymmmm> jtr your welcome to ship it to me, "I promise I'll give it back (in 30 or 40 years)
[00:35:09] <jtr> heh
[00:35:45] <Jymmmm> holy shit 750GB Seagate PATA $300
[00:38:11] <ejholmgren> booyah
[00:39:38] <Jymmmm> looking at mobo's too...most of them only have 2 dimm slots, eeeesh.
[00:41:44] <jtr> Yesterday I could have given you 10 PCs with, oh, 7 or 8 slots. 'course they were all ISA...
[00:42:13] <ejholmgren> just buy a legacy sun system ...
[00:42:30] <ejholmgren> 8 slots with 128MB in each :)
[00:42:51] <Jymmmm> Can't, need good video card, only reason I'm even looking.
[00:42:52] <ejholmgren> (the _hard_ way to get 1GB)
[00:43:06] <Jymmmm> oh, my desktop has 1.5GB ram right now.
[00:43:20] <jtr> Sparc stations were in the pile. I did save out some Deeco touch terminals - emulated a VT100
[00:43:27] <Jymmmm> Dual P3 800 cpu w/ 1.5GB ram
[00:43:56] <ejholmgren> someone was selling an AGP ethernet adapter on craigslist the other day ...
[00:44:04] <Jymmmm> lol
[00:44:09] <ejholmgren> 'not PCI, this is definitely AGP"
[00:44:14] <ejholmgren> hesus
[00:44:40] <ejholmgren> it was an old 3com ISA card with bnc to boot
[00:44:59] <Jymmmm> 3C509TX
[00:45:12] <Jymmmm> I have 4 of them
[00:45:23] <Jymmmm> Lifetime warranty on all 3Com nic's
[00:46:01] <Jymmmm> I'm waiting till the run out of ISA cards to replace, thn have them replace them with gigbit cards =)
[00:46:51] <ejholmgren> http://minneapolis.craigslist.org/sys/282375516.html ...
[00:46:59] <ejholmgren> looks like someone told them they were an idoit
[00:47:05] <ejholmgren> har, idiot
[00:47:27] <Jymmmm> Rare Vintage Collection
[00:47:29] <Jymmmm> wth
[00:47:36] <ejholmgren> no kidding ...
[00:48:23] <Jymmmm> 24 port switch $99, bastards! it's 10/100
[00:49:30] <Jymmmm> I need gigabit!
[00:50:32] <Jymmmm> You have to ALWAYS carry at least $100 to work, you never knew when they're seeling equipment =)
[00:52:27] <Jymmmm> and I want an IP/KVM =)
[01:23:51] <CIA-6> 03cradek 07TRUNK * 10emc2/src/emc/rs274ngc/interp_convert.cc: correctly notice missing r word
[01:56:33] <ejholmgren> what are you going to do with 24 ports of gig-e?
[02:01:54] <skunkworks> I think I have a box full of those 3com III
[02:09:12] <skunkworks> http://www.youtube.com/watch?v=QsmiIeAkE-o&NR
[02:12:11] <crepincdotcom> jmkasunich: thanks, I'll give it a read
[02:16:46] <skunkworks> holy crap - that video is cool. makes the water jacket and everthing.
[02:24:50] <ejholmgren> now _that_ is definitely tool porn
[02:24:57] <ejholmgren> just watch that block gyrate
[02:25:27] <skunkworks> that thing has the moves to please.
[02:26:15] <skunkworks> did you see that it single point bored the cylinders?
[02:26:24] <skunkworks> with a ball end mill
[02:26:27] <skunkworks> I think
[02:27:47] <K`zan> Just not sure at all about this here flycutting stuff: http://wrlabs.shacknet.nu/~vw/MyMachineShop/tmp/flycut1/images.html
[02:29:29] <skunkworks> not sure about what?
[02:34:02] <K`zan> Much of anything to do with it :). Either it is right or the tool still isn't, this right hand tool does pretty much the same thing I guessed at before.
[02:34:20] <K`zan> Maybe I need a better radius on that...
[02:34:43] <K`zan> Maybe I need more angle, damnifiknow, info on fly cutting bits seem to be hard to find.
[02:36:58] <skunkworks> how sharp is it?
[02:38:22] <K`zan> It will cut me...
[02:39:25] <SWPadnos> it looks like the bit is chattering. how fast are you running the spindle?
[02:44:29] <K`zan> best guess about 2K rpm
[02:44:36] <K`zan> mabe 1500
[02:44:46] <SWPadnos> cucutting aluminum?
[02:44:49] <SWPadnos> -cu
[02:44:54] <skunkworks> what diameter?
[02:44:55] <K`zan> Yes 6061t6
[02:45:00] <a-l-p-h-a> hahahahhaa
[02:45:07] <K`zan> 1" across the top
[02:45:08] <a-l-p-h-a> wayyy too fast for something that isn't balanced.
[02:45:13] <SWPadnos> yeah
[02:45:34] <SWPadnos> the radiuswe need is the distance from center of shank to end of cutting tip
[02:46:04] <skunkworks> what is the effective diameter of the fly cutter
[02:46:14] <skunkworks> what swpadnos said
[02:46:16] <SWPadnos> heh
[02:46:22] <SWPadnos> plus a space, for good grammar
[02:46:27] <SWPadnos> and spelling
[02:46:32] <K`zan> about 1-3/8"
[02:47:29] <SWPadnos> ok. 2 * pi * r = ~0.7 feet per revolution
[02:47:36] <SWPadnos> of cutter travel
[02:47:59] <SWPadnos> that means that at 2000 RPM, you're doing ~1420 SFM, which is probably 5x what it should be
[02:48:17] <SWPadnos> cut it down to the 300-500 RPM range and see what happens
[02:48:23] <SWPadnos> maybe even slower
[02:48:34] <SWPadnos> hmm - closer to 1440 SFM
[02:49:18] <K`zan> OK will do - dinner is burning bbiaf - THANKS MUCH
[02:49:26] <SWPadnos> hope it works ;)
[02:50:13] <skunkworks> you need one of the hady-dandy sfm/rpm/whatever slide rules. - they come in handy
[02:50:34] <SWPadnos> actually, you need the chart that tells you the cutting speeds for various materials
[02:50:34] <skunkworks> cut per tooth
[02:50:40] <SWPadnos> the calcs are pretty easy :)
[02:51:23] <skunkworks> that is on the slide rule I use. has a bunch of materials and sfm for carbide and hhs
[02:51:39] <SWPadnos> ah - the database info is useful
[02:51:54] <skunkworks> gets you close
[02:54:49] <skunkworks> although somethimes I just guess
[02:55:09] <CIA-6> 03cradek 07TRUNK * 10emc2/src/emc/rs274ngc/interp_convert.cc:
[02:55:09] <CIA-6> make the entry lengths all the same, so the spindle is in the right
[02:55:09] <CIA-6> position when each pass starts cutting
[02:56:58] <lerman_> lerman_ is now known as lerman
[03:18:29] <CIA-6> 03cradek 07TRUNK * 10emc2/src/emc/rs274ngc/interp_convert.cc: cleanup only - no logic changes
[03:25:30] <CIA-6> 03cradek 07v2_1_branch * 10emc2/src/emc/rs274ngc/ (4 files): backport angled entry/exit for canned threading cycle
[03:31:39] <CIA-6> 03cradek 07v2_1_branch * 10emc2/nc_files/g76.ngc: document all g76 features
[03:33:27] <CIA-6> 03cradek 07v2_1_branch * 10emc2/debian/changelog: note angled entry/exit backport
[03:34:43] <CIA-6> 03cradek 07v2_1_branch * 10emc2/debian/changelog: oops
[03:36:39] <Jymmm> WooHoo I found a crimper for the molex/amp pins at Fry's for $13, as well as the pins and connectors
[03:37:33] <SWPadnos> cool. is that normally in stock?
[03:38:00] <Jymmm> It appears it is, as they had 30-24ga and a 24-14ga version
[03:38:15] <Jymmm> maybe 30 to 18ga
[03:38:37] <Jymmm> But I guess it depends on which fry's you hit that they actually have it in stock.
[03:38:51] <SWPadnos> ok - Fry's will definitely have to be on the list then :)
[03:39:00] <SWPadnos> I guess it would be anyway though
[03:39:07] <CIA-6> 03cradek 07TRUNK * 10emc2/bin/.cvsignore: new program
[03:39:24] <Jymmm> You've been to fry's before, no?
[03:39:53] <SWPadnos> yes, a couple of times (the one near NASA Houston
[03:39:59] <Jymmm> oh, ok.
[03:40:02] <SWPadnos> we went to the big one there just before they closed, if you remember
[03:40:12] <Jymmm> Well Sunnyvale and Fremont fry's are the best.
[03:40:19] <SWPadnos> ie, they weren't sure why we came in when they were trying to close ;)
[03:40:25] <Jymmm> Brokaw is ok, but older.
[03:40:50] <Jymmm> ah, no I dind't.
[03:41:09] <Jymmm> They were just pissed is all as you notice NONE of them said anything =)
[03:41:11] <SWPadnos> I think you were looking for something for your GF
[03:41:27] <Jymmm> If it didn't happen in the last 4 hours, I don't remember
[03:41:33] <Jymmm> s/hours/seconds/
[03:41:53] <SWPadnos> actually, I think the old guard guy at the door asked us why we wanted to come in. I think we said we were looking for someone or something
[03:42:05] <Jymmm> heh
[03:42:30] <Jymmm> Well, HF and Frys and both in Fremont.
[03:43:42] <cradek> http://www.mikromart.com/includes/en/pr_man04_photo.htm
[03:43:51] <cradek> neat laser? work
[03:44:22] <SWPadnos> whichever. I haven't bought a ticket yet, so I'm not sure when I'm arriving. it should be March 31 though
[03:46:35] <SWPadnos> a friend of my wife's does bent metal sculptures like that, but not so technical
[03:46:57] <SWPadnos> it's even an Aeron-like chair ;)
[03:47:46] <cradek> many more at 'back to shop'
[03:48:09] <Jymmm> see, I need a damn laser!!!
[03:49:31] <ejholmgren> Jymmm: doesn't everyone?!
[03:49:50] <Jymmm> not sure bout that.
[03:49:57] <SWPadnos> hmm. I may buy one of those garden cubes for my wife's b-day
[03:50:08] <cradek> I also think the garden is a cool one
[03:50:10] <Jymmm> garden cubes?
[03:50:22] <SWPadnos> http://www.mikromart.com/includes/en/pr_cube01_photo.htm
[03:50:41] <SWPadnos> she's more into the artistic, so the guy at a desk or whatever wouldn't be great for her
[03:50:53] <SWPadnos> the person on the bicycle would be good though - she likes biking
[03:51:05] <Jymmm> unless you piss her ff and it has a dart board attchment for the missing head.
[03:51:13] <SWPadnos> as long as it's not mine
[03:51:29] <Jymmm> is that SS ?
[03:51:46] <ds3> ewwwwwwww fremont frys Yuck
[03:52:05] <SWPadnos> holy crap - they come flat. you have to fold them yourself
[03:52:11] <tomp> alex_joni: the cooltool pdf shows GoSloJimbo's dro ( or similar ), we just fixed that for him acouple days ago
[03:52:32] <SWPadnos> he had it working in 2.0.5 though
[03:52:37] <Jymmm> SWPadnos, smart aren't they =)
[03:52:41] <SWPadnos> we had to fix it for 2.1, I think
[03:53:04] <SWPadnos> Jymmm, yeah "insert tab A into slot B"
[03:53:48] <Jymmm> SWPadnos that works, shit, maybe I should look into a laser attachemnt for my machine
[03:57:39] <steve_stallings> steve_stallings is now known as steves_logging
[04:02:11] <Jymmm> Chemcial Milled Stainless Steel
[04:02:20] <Jymmm> 95x35mm
[04:02:27] <Jymmm> 150 microns
[04:02:42] <K`zan> SWPadnos: skunkworks: Thanks. Dinner was salvageable :-). Going to try another pass at 3-500 rpm and look for a sfm calculator :-).
[04:02:44] <SWPadnos> hmm - where did you find that?
[04:02:57] <Jymmm> the flahs navigation stuff
[04:03:08] <SWPadnos> oh. I hated that so I gave up early
[04:03:21] <SWPadnos> it looked nice, but I perfer a somewhat less artsy interface
[04:04:03] <Jymmm> that's what he his, an artist. Sam Buxton
[04:05:26] <Jymmm> goto Tourist Info
[04:05:49] <SWPadnos> yeah - I saw that in the read me section
[04:52:01] <Jymmm> artsy is ok, but gimme at least linkable url's
[04:52:35] <SWPadnos> yes. that's one reason why I hate flash
[04:53:28] <Jymmm> I like flash, but like anything when in the wrong hands........
[04:54:06] <SWPadnos> well, you'd need separate pages (with or without flash) to get anything linkable
[04:54:22] <SWPadnos> unless flash can do something different depending on an anchor in the URL
[04:54:32] <Jymmm> Yeah, I understand.
[06:40:39] <sed> is anone awake?
[06:42:58] <sed> I am running some parts for a friend while he is gone, I rebooted his EMC machine while the Z axis was not home. Is that hard to reset?
[07:37:06] <K`zan> Night all
[10:09:23] <jlmjvm> Good Morning All
[10:11:37] <jlmjvm> Has anyone used ubuntu and emc2.1 on a computer with a pci express video card
[10:31:19] <alex_joni> jlmjvm: not really, any issues?
[10:38:57] <jlmjvm> the computer has onboard vid right now,and wont run it.But it will run on everything i have that does have a video card,but they are all agp
[10:39:49] <jlmjvm> was hoping it would run if i put a card in it
[10:42:50] <glyndwr> hello everybody. I am collaborating in the starting of a plan for building an EDM machine for micro-nanomanufacturing for reasearch purposes only. Can somebody give me some suggestions/information?
[10:43:18] <jlmjvm> is there a preffered gpu,nvidia or radeon
[10:50:23] <alex_joni> jlmjvm: nvidia and ati have issues with RT if you want to use the accelerated driver
[10:50:31] <alex_joni> so I would suggest the cheapest card you get :D
[10:50:40] <alex_joni> glyndwr: there are some people working on emc2 and EDM
[10:50:51] <alex_joni> I suggest you drop an email to the emc-users mailing list
[10:51:20] <alex_joni> glyndwr: also, have a look at this: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?ElectricalDischargeMachining
[10:54:16] <jlmjvm> alex:so just get the basic card,not the latest gaming card,
[10:54:42] <alex_joni> jlmjvm: right
[10:55:00] <alex_joni> maybe you can drag the LiveCD to your local HW coputershop
[10:55:04] <jlmjvm> why did i have that smiley face?
[10:55:07] <alex_joni> and try it out before buying
[10:57:38] <jlmjvm> i know ubuntu wont even boot up with the new nvidia 6100 onboard video,wont load x,but the fedora core can run it
[11:00:24] <jlmjvm> what im looking at is most cards are going pci express and want to find one that works with this software
[11:05:51] <alex_joni> then my advice is to take the liveCD and try it out on such a hardware before buying
[11:07:48] <glyndwr> thank you alex_joni I am going to read that page. The focus is on micro-nano EDM. which demands actuators as accurate as possible. Also what is seen as critical is the spead of reactions of such actuators when short circuit are detected. I am considering buying such things from www.pi.ws. But what exactly to buy (piezo stages, linear motors ...)? 4 axes are requested. the working space as a very maximum of a few millimitres Are these components compatible
[11:09:00] <jlmjvm> alex:if i used 2 parallel ports with breakout boards could i close the loop on a 3 axis mill with 3 gecko drives?
[11:09:49] <alex_joni> glyndwr: I am sorry, this is where my knowledge stops :)
[11:10:01] <alex_joni> jlmjvm: close the loop?
[11:11:08] <jlmjvm> alex:read the us digital encoder position back in
[11:11:18] <alex_joni> jlmjvm: why is that?
[11:11:57] <alex_joni> jlmjvm: the gecko drives do the PID tuning and all
[11:12:08] <alex_joni> so you drive them (from teh PC's side) open loop
[11:12:19] <alex_joni> if you want a closed loop system, then some other amps might be better
[11:12:48] <alex_joni> otherwise you'll end up having 2 PID loops (one in the gecko, one in the PC over that), and that will lead to instability issues (e.g. tuning nightmare)
[11:13:07] <alex_joni> otoh, the parport is limited in counting high frequency encoder counts
[11:15:09] <jlmjvm> alex:thought i could use encoders on a gecko setup and be able to home to an encoder index
[11:18:57] <jlmjvm> alex:actually if they just functioned as a dro with in inposition tolerance that would be great
[11:21:07] <jlmjvm> alex:prolly be better off to get elsons universal stepper board
[11:22:34] <alex_joni> jlmjvm: depends what you want to do.. but if you can afford it that sounds like a good idea
[11:31:00] <jlmjvm> alex:I just thought you could add encoders to a stepper setup and home to the index mark through a second parallel pot
[11:32:40] <glyndwr> ok. thank you very much alex_joni. I have read your link and I also noticed that you are there :-). I'll check the links in that page. BTW, I know one of them: the Scott Ananian'web pages. It will be great if I could find someone interested in it. As I said, is a merely research project involving two universities and other four partecipants. So it has to be completely open. That's why EMC would be the perfect match.Unfortunaly, I 'd like going on the moon
[11:33:32] <jlmjvm> alex:would that be possible?
[11:39:47] <alex_joni> jlmjvm: anything is possible
[11:39:47] <alex_joni> :D
[11:40:02] <alex_joni> the problem is how much time/work you want to spend on it
[11:40:25] <alex_joni> you can use the feedback form a gecko for homing (index pulse) and for feedabck
[11:40:38] <alex_joni> but I would do that without a PID on the PC side..
[11:40:52] <alex_joni> I guess HAL is flexible enough to let you set it up any way you want
[11:43:48] <alex_joni> so to answer your question, theoretically it's possible, but I have nfc how well it would work
[11:44:06] <jlmjvm> alex:sounds like more than i thought
[11:44:16] <alex_joni> more
[11:44:16] <alex_joni> ?
[11:44:27] <jlmjvm> more work
[11:44:33] <alex_joni> not much more..
[11:46:09] <jlmjvm> cant you read as encoder as a position read out on a stepper machine?
[11:46:45] <jlmjvm> and just use the numbers as a comparison
[11:46:59] <alex_joni> jlmjvm: you can
[11:47:08] <alex_joni> the issue is what do you do with the numbers?
[11:47:22] <alex_joni> compare to commanded position and issue ferror if too far away?
[11:47:36] <jlmjvm> that would be cool
[11:48:58] <jlmjvm> if you could do just that and use the z index in the homing cycle it would be an awesome setup for a stepper mill
[11:50:33] <jlmjvm> dont think it would be real closed loop,but would be a heck of an upgrade
[11:52:02] <jlmjvm> a position monitor and very accurate homing
[13:22:01] <alex_joni> hi anders
[13:24:14] <jepler> good morning!
[13:25:29] <awallin> hi
[13:25:49] <awallin> still battling the filtern-in-PID-loop problem I described yesterday...
[13:25:56] <awallin> filter
[13:26:48] <awallin> I have a digital PID loop that runs at 200kHz and seems to work OK, it corrects for disturbances of all frequencies (theoretically up to 100kHz)
[13:27:30] <awallin> but how could I filter either the process variable or the error signal so that only high(low) frequencies are comensated, and low(high) frequency distrurbances are not corrected ?
[13:28:28] <awallin> I've tried simple low-pass or high-pass filters, but these have so much phase lag that there are always several frequencies at which the lag is 180 degrees, which means positive feedback! (think about pointing microphones at loudspeakers)
[13:31:01] <SWPadnos> you wanted a 1 KHz cutoff, right?
[13:31:34] <SWPadnos> that's actually a fairly high Q for a filter - it's only 1% of the total bandwidth
[13:32:46] <SWPadnos> I think if you just average over N samples, where N is sample rate / (2*cutoff frequency), you'll get a reasonable low pass filter
[13:33:04] <SWPadnos> for high pass, subtract the low pass filter output from the input signal
[14:02:30] <xemet> hi
[14:02:38] <jepler> hi xemet
[14:02:50] <xemet> problem:
[14:03:02] <xemet> I hal component using comp
[14:03:08] <xemet> I wrote...
[14:03:23] <xemet> this component has 1 input pin float
[14:03:32] <xemet> and an output pin integer
[14:03:56] <xemet> my intention is to use it to generate counts for the joypad
[14:04:00] <alex_joni> xemet: so where's the problem?
[14:04:09] <xemet> the component work...for 1 axis
[14:04:22] <xemet> but if I use loadrt joycounts count=3
[14:04:28] <xemet> for X Y Z
[14:04:50] <xemet> the joycounts.0.counts is always the same also for 1 and 2
[14:05:03] <alex_joni> can you put the component on pastebin.ca ?
[14:05:10] <xemet> of course
[14:05:46] <xemet> http://www.pastebin.ca/367637
[14:05:46] <alex_joni> did you look at knob2float.comp ?
[14:05:58] <alex_joni> that's the exact opposite of what you want I think
[14:05:59] <xemet> uhm not...
[14:06:03] <jepler> float fcounts=0;
[14:06:16] <xemet> alex: where can I find a complete list of ALL hal components available?
[14:06:18] <jepler> this "float" is shared by all the instances
[14:06:44] <jepler> the easiest thing to do is make 'fcounts' a parameter, so each instance of joycounts has its own
[14:06:44] <alex_joni> xemet: http://cvs.linuxcnc.org/cgi-bin/cvsweb.cgi/emc2/src/hal/components/http://cvs.linuxcnc.org/cgi-bin/cvsweb.cgi/emc2/src/hal/components/
[14:07:15] <SWPadnos> that's quite a link
[14:07:23] <SWPadnos> or two
[14:07:31] <xemet> Error: emc2/src/hal/components/http:/cvs.linuxcnc.org/cgi-bin/cvsweb.cgi/emc2/src/hal/components: no such file or directory
[14:07:41] <alex_joni> http://cvs.linuxcnc.org/cgi-bin/cvsweb.cgi/emc2/src/hal/components/
[14:07:52] <alex_joni> sorry .. my rightclick acted on its own :P
[14:07:53] <jepler> xemet: http://www.pastebin.ca/367639
[14:08:23] <xemet> thanks
[14:08:52] <xemet> just for curiosity, is there other ways to do that?
[14:09:10] <alex_joni> you can just use a float<->int converter
[14:09:38] <alex_joni> scale the float signal first, then use the converter
[14:09:45] <xemet> I mean to seperate the instances
[14:09:56] <jepler> xemet: you can also use 'option data', which is described in the documentation
[14:10:03] <xemet> when I load more components of the same type
[14:10:57] <jepler> xemet: ddt.comp is an example which uses 'option data'
[14:11:03] <xemet> thanks
[14:11:46] <xemet> so if I type "option data float", will it work?
[14:12:05] <alex_joni> option data ddt_data;
[14:12:10] <alex_joni> typedef struct { float old; } ddt_data;
[14:12:12] <jepler> yes, but then you'd refer to 'data' instead of 'fcounts'
[14:12:25] <xemet> ah ok
[14:12:48] <jepler> I think all our code uses a 'struct' even if it contains just one item, because then you can still give that variable a useful name
[14:12:59] <jepler> "fcounts" or "old" is (a little) more explanatory than "data"
[14:13:17] <xemet> ok
[14:13:34] <jepler> of course, you actually have to write "data.fcounts"
[14:14:13] <xemet> will try now to change the module with the parameters
[14:14:48] <xemet> this is the first thing I write in C...so I don't know anything about it
[14:15:22] <alex_joni> this is not really C :)
[14:15:38] <alex_joni> but close enough
[14:15:52] <xemet> ok
[14:15:52] <jepler> the part after ";;" is just C with some extra surprises
[14:16:14] <xemet> what are the surprises?
[14:17:31] <jepler> there is a preprocessor macro for each pin or parameter name which allows you to refer to its value in the current instance with just the pin's name
[14:18:12] <jepler> otherwise, to refer to a pin 'p' you have to write '*(inst->p)'
[14:18:34] <xemet> yes I've seen this in the documentation
[14:18:48] <xemet> like the FUNCTION() to call a realtime function
[14:20:23] <xemet> now the module works
[14:20:28] <xemet> thank you
[14:21:06] <jepler> hooray
[14:21:09] <jepler> I'm glad I could help
[14:23:32] <xemet> I return to my work. I will modify the joypad wiki page adding that later
[14:43:28] <alex_joni> xemet: nice
[14:51:05] <biglibigli> i want to access my emc (commands and information) from serial port, how can i do this
[14:52:54] <biglibigli> some body help me
[14:54:30] <biglibigli> anyone in this room?
[14:55:10] <alex_joni> you can set up a serial port as a linux terminal
[14:55:25] <alex_joni> and run keystick as the EMC user interface on that terminal
[14:55:32] <alex_joni> is that what you want?
[14:59:08] <biglibigli> no i want ems send commands like (stepper clock..) and information like (x,y,z) to a micro controller and receive command like stop or ... from microcontroller
[15:00:39] <alex_joni> biglibigli: then you need to write some software to do that
[15:00:51] <alex_joni> gotta run now, sorry I can't help you more
[15:01:11] <biglibigli> tanx
[15:01:23] <alex_joni> (you probably want to look at emc2/src/emc/usr_intf/halui.cc, and modify that to send/receive serial information instead of setting/clearing hal pins)
[15:02:47] <biglibigli> can i redirect hall pins to file system?
[15:03:20] <biglibigli> to a file.
[15:06:25] <jepler> you can write userspace compoent using the HAL APIs (documented in the manual pages in section 3hal) as well as the standard library file I/O.
[15:06:50] <jepler> so if you can define what it means to "redirect HAL pins to a file", then you can write the software that does it
[15:10:06] <SWPadnos> maybe he meant halsampler/halstreamer functionality?
[15:10:18] <SWPadnos> oh well, I guess we'll never know
[15:12:29] <jepler> halhamster
[15:21:23] <skunkworks> logger_emc: bookmark
[15:21:24] <skunkworks> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2007-02-22.txt
[15:23:44] <skunkworks> well they removed the house inspection contingency from their off :)
[15:23:50] <skunkworks> offer
[15:49:04] <pier> hi all... I was wondering if it would be possible to use emc with my hot wire foam cutting machine...
[15:50:10] <pier> and get rid of the dos program I use
[15:58:44] <jepler> I don't know of anyone using emc on an "xyuv" machine, which I understand is the typical setup for hot wire
[16:00:02] <skunkworks> how is the code generated - if it is normal xyuv code then I would think that it would work - 2 perpendicular axises parrallel to one another. If I am understanding it
[16:01:25] <jepler> skunkworks: emc assumes at most 2 linear and 2 rotary axes. You can try commanding "XYAB" moves for your 4 linear axes but I don't know how well that will work in practice
[16:01:52] <jepler> you can't command arcs on AB
[16:02:51] <pier> sorry brb
[16:03:25] <skunkworks> I figured it would be all short line segments.
[16:03:41] <skunkworks> xy - uv
[16:04:09] <cradek> might be fun to try using hal to make some more axes relative to XY
[16:04:10] <skunkworks> I am not really sure I know how a foam cutter works I guess
[16:04:12] <skunkworks> * skunkworks googles
[16:04:17] <cradek> I haven't thought through the details of course
[16:04:29] <cradek> probably wouldn't want to try it as a first emc project
[16:05:20] <skunkworks> i figured 2 sinced perpindicular axises xy and uv - you would just be giving them posistions in each line of where you wanted both to end.
[16:05:34] <skunkworks> well that didn't make sense
[16:06:22] <cradek> I think on a wire machine, UV is relative to XY, so if U=0,V=0 the axes track together
[16:06:29] <skunkworks> 2 axises perpendiclar to them selves parallel to 2 more axises perpendicluar to itself
[16:06:38] <skunkworks> ah
[16:06:40] <cradek> UV are nonzero when you want the wire tilted by a certain amount
[16:06:51] <skunkworks> ok - that makes sense
[16:07:09] <skunkworks> Hi ray.
[16:08:08] <skunkworks> but I would think it would be just as easy to have u and v absolute. for the diyer
[16:08:21] <skunkworks> or whatever you want to name the axises
[16:08:40] <cradek> that would be easy for emc, but you'd have to change your programs and write more complex gcode all the time
[16:09:13] <skunkworks> right.
[16:10:28] <SWPadnos> or sum/difference the commands in HAL, outside the motion controller
[16:10:30] <skunkworks> I could se a simple utility that could convert the u,v to absolute. but if emc could handle it automatically it would be cool. I would think it would have to be higher up than hal to take into acount axis limits.
[16:10:49] <SWPadnos> add X+U for the motor, and subtract in the position feedback path
[16:11:21] <skunkworks> * skunkworks may just be talking out of his ass again
[16:12:15] <SWPadnos> no - you're right that limits could pose a problem
[16:12:32] <SWPadnos> also accel/vel limits
[16:13:06] <xemet> hallo, I've updated the joypad wiki page I had posted yesterday with a new page that explain how to write a new hal component to eliminate sim-encoder-encoder components
[16:13:15] <xemet> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Here
[16:13:43] <xemet> very simple hal component...it is my first. And it works :)!!
[16:15:05] <xemet> I think I should change the name of this page. "Here" is not a good name
[16:15:18] <SWPadnos> gotta love comp - it makes that kind of component about 1000 times easier to write :)
[16:15:39] <xemet> It is great!!
[16:21:10] <jmkasunich> xemet: thanks for writing the wiki pages
[16:21:57] <xemet> welcome, I hope there are not too many errors :)
[16:22:07] <xemet> I'm not so good at english
[16:23:28] <skunkworks> that is cool.
[16:23:30] <skunkworks> :)
[16:23:34] <jepler> xemet: you should probably just assign the 'counts' output every time: counts = floor(fcounts);
[16:23:47] <xemet> it is my first wiki page, so if you have suggestions or if you find that something could be done better pleaselet me know
[16:23:50] <skunkworks> jmkasunich: vacation day?
[16:23:58] <xemet> I've tried
[16:24:14] <xemet> but this doesn't work very well
[16:24:41] <jepler> xemet: output pins should be assigned every time, because if they are not, they temporarily lose their value after first being 'link'ed in halcmd until the next assignment
[16:25:42] <xemet> ah ok
[16:25:45] <eholmgren> what was the name of that crazy bar surface that responed to objects placed on it?
[16:25:49] <xemet> so I remove the if statement
[16:25:52] <jmkasunich> skunkworks: yeah, sort of - we have a carpenter working here, I stayed home to keep the dog out of his way
[16:26:21] <xemet> now that I think about it efectivly it seems very useless
[16:26:23] <skunkworks> and an eye on the carpender ;)
[16:27:16] <skunkworks> SWPadnos: that is what I ment by limits.. I forgot about 'limits' ;)
[16:29:54] <jmkasunich> skunkworks: I'm not worried about the carpenter - he's very good at what he does, gets all his business by word of mouth and reputation - definitely not fly-by-night
[16:30:17] <jmkasunich> (finish carpenter - he's doing stair and upstairs hallway railings)
[16:30:34] <skunkworks> jmkasunich: that is good to find.
[16:30:58] <skunkworks> * skunkworks hates finishing stairs and railings.
[16:31:34] <jmkasunich> I'll do deck railings, thats rough work - but the finish stuff takes tools and skills I don't have
[16:33:50] <skunkworks> jmkasunich: selling/buying of the house is going well. Might actually have time to tinker in a month or two. Have not even unwrapped the mesa. Probably play more with the pluto first though.
[16:35:21] <jmkasunich> peter wallace sent me a first stab at 5i20 stepgen yesterday, I'm gonna try to hack together a crude driver for it today
[16:35:36] <skunkworks> we where able to get the old controller to pull the pallet off of the table - so we can start tearing into that also. (figuring out how to install the servos.)
[16:35:44] <skunkworks> Cool
[16:38:06] <skunkworks> jmkasunich: did you see this? http://www.youtube.com/watch?v=QsmiIeAkE-o&NR
[16:38:29] <jmkasunich> yeah, amazing
[16:39:58] <skunkworks> I can't believe how far the cut the waterjacket in.
[16:42:40] <jepler> xemet: after you piqued my curiousity, I bought a spaceball off of ebay. I'll let you know if I can get mine to work.
[16:43:16] <xemet> jepler: I've removed the totally useless if statement
[16:43:17] <xemet> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?HowToWriteNewHalComponent
[16:43:18] <jepler> curiosity, that is
[16:43:27] <xemet> AH, wich spaceball is?
[16:43:35] <xemet> 4000 FLX?
[16:44:47] <jepler> xemet: yes
[16:44:58] <xemet> serial?
[16:45:22] <jepler> yes
[16:46:33] <xemet> well...I'm sure you will succeed in getting it work
[16:47:44] <xemet> let me know if you find the problem
[16:48:54] <cradek> that seems like a crazy way to hone cylinder walls
[16:49:06] <jepler> xemet: I will
[16:49:44] <xemet> jepler: if you're interested, here is a small application that uses the original spaceball driver (the one provided by 3dconnexion). It opens a little window that shows the axes values
[16:49:46] <xemet> http://www.pastebin.ca/367821
[16:50:00] <jepler> xemet: does that work with your device?
[16:50:11] <xemet> you can download it from 3dconnexion website
[16:50:51] <xemet> I've not tried *this* but with the driver there is an already compiled application that does the same
[16:51:10] <xemet> the one provided with the driver works
[16:51:28] <xemet> and shows all axes values correctly, so I think the spaceball works very well...
[16:52:15] <xemet> Before discovering the possibility of input attach I thought that could be possible to write a components like hal_joystick to read axes values...
[16:52:40] <xemet> based on that application
[16:53:28] <jepler> yes -- if you have a library that works, while the joystick driver doesn't, writing a userspace HAL component to use the library sounds like a good option
[16:56:25] <xemet> well...I've to study C to do that I think...
[16:56:39] <xemet> I've a book but never have the time to read it...
[16:57:12] <xemet> I don't know if I have the library...the only think I have is that application I sent to you
[16:57:34] <xemet> and it is the only thing the 3dconnexion support forum has given to me
[16:57:56] <xemet> and they told me that the axes values are sent using XeventMessage
[16:58:07] <jepler> did you add lines to your X server configuration so that this program would work
[16:58:08] <jepler> ?
[16:58:42] <xemet> I've not used this program
[16:58:48] <jepler> oh
[16:58:53] <xemet> I've used the one provided with the driver
[16:58:59] <xemet> it is already compiled
[16:59:01] <xemet> and works
[16:59:09] <xemet> without adding nothing
[16:59:42] <jepler> you didn't have to configure anything?
[16:59:53] <xemet> no...
[17:00:02] <xemet> simply click on an icon :)
[17:00:25] <jepler> I wonder if their driver interferes with the other driver
[17:01:01] <xemet> http://www.3dconnexion.com/support/download_driver.php?product_id=13&os_id=1&driver_id=52&action.x=126&action.y=19
[17:01:12] <xemet> here is the driver and the compiled demo apps
[17:01:53] <jepler> if it is easy to do, I suggest you try uninstalling their driver and then see if the linux joystick driver works
[17:02:13] <xemet> but their driver is not always running
[17:02:17] <xemet> I've to launch it
[17:02:41] <xemet> and if I launch it the joystick driver says me that it is impossible to initialize the device
[17:02:49] <jepler> oh, hm
[17:02:57] <jepler> darn, I thought that could be the reason
[17:04:43] <xemet> well...the joystick driver author Vojtech Pavlik told me that he will try to see if can reproduce the problem, he had a two or three spaceballs...
[17:05:06] <xemet> but I don't think he have the time to do that...
[18:07:09] <eholmgren> * eholmgren meeps
[18:07:47] <Vq^> *chirp*
[18:37:02] <tomp> glyndwr: hello, you interested in micro-edm with emc?
[18:46:27] <tomp> glyndwr: Scott Ananian's Garden of EDM design is likely too coarse(on/off) and too powerful for your needs. I say this because you mention using PI actuators, which are piezo and move in nanometers.
[18:47:09] <tomp> glyndwr: if you need that motion increment, then your generator must be less powerful and finer units of energy
[18:49:59] <tomp> glyndwr: for motion control emc can drive those actuators but it designed for velocity & position control, not process driven control ( eg: gap signal controling position and velocity ). A wire EDM was controlled by using an analog of the gap to proportion the velocity, reducing velocity until gap sparking was stabile.
[18:50:49] <tomp> glyndwr: This method searched for a stabile process but could not backup. Some ideas have been proposed to allow backup, but none have been documented.
[18:53:31] <tomp> glyndwr: I suggest you study the TP ( trajectory planner) to see what it is capable of doing. EMC's TP will always try to be in position and on time, while EDM wants to be on path while maintaining the process. If you have the facilities, you may be able to substitute a new TP, but thats many man years of work so far.
[18:56:04] <tomp> glyndwr: I'd be very interested in your work, and hope you can achieve what you need. EMC is constantly evolving and may have a solution for you, or may get a solution from you.
[18:58:56] <skunkworks> being able to control emc with a jog wheel to run foward and reverse thru your gcode.
[18:59:02] <skunkworks> would be cool :)
[19:04:03] <tomp> glyndwr: i imagine your work is near this precision, or even finer http://www.gaisertool.com/products_wedges.php?menu=products For this work, the generator was irrelevant, we used DC and the interface of the carbide tools and the silver electrodes created their own oscillation, and i merely adjusted the current to create the desired finish/rate. Machine tools were built of granite with 1mm pitch leadscrews and 1mm/min max velocity.
[20:47:04] <crepincdotcom> gah the interwebs are cryig
[20:47:06] <crepincdotcom> *rying
[20:47:10] <crepincdotcom> crying
[20:47:14] <crepincdotcom> and i am failing at typing.
[20:49:59] <CIA-6> 03cradek 07v2_1_branch * 10emc2/debian/changelog: document various backported fixes
[20:59:12] <CIA-6> 03cradek 07v2_1_branch * 10emc2/debian/changelog: remove noise
[21:14:31] <CIA-6> 03paul_c 07bdi-4 * 10emc2/src/libnml/cms/tcp_srv.cc: Resolve typecast issues.
[21:54:03] <CIA-6> 03jmkasunich 07TRUNK * 10emc2/src/hal/utils/m5i20cfg.c: fix usage message
[22:56:23] <Jymmmm> jmkasunich you alive?
[23:04:56] <jmkasunich> yeah
[23:05:21] <Jymmmm> jmkasunich do you have/use AMP connectors by chance?
[23:05:39] <Jymmmm> AMP 60619-1 and 60620-1
[23:05:54] <jmkasunich> use at work, yes
[23:05:55] <jmkasunich> have, no
[23:05:59] <jmkasunich> have catalogs, etc, no
[23:06:09] <Jymmmm> ok, thanks.
[23:06:44] <Jymmmm> Found the crimper for $13, but only have the molex pins - wrong diameter (.093 instead of .084)
[23:07:08] <Jymmmm> * Jymmmm sighs*
[23:07:11] <jmkasunich> so get molex connectors?
[23:07:25] <jmkasunich> seems like the crimper is the key item
[23:07:38] <jmkasunich> or do you already have the pins and such
[23:08:03] <Jymmmm> PC PS are the amp mat n lock, since I have the extractors already, just need the 0.084" pins
[23:08:49] <Jymmmm> and like you, I can harvest the nylon shells all over the place.
[23:09:04] <Jymmmm> Dumpster Diving 101 =)
[23:11:58] <Guest282> hello
[23:15:09] <crepincdotcom> hey
[23:17:36] <Guest282> i have a little problem witch emc2
[23:17:59] <crepincdotcom> whats the issue
[23:18:12] <crepincdotcom> (note: i am certainly not amazing with EMC yet)
[23:19:51] <Guest282> as i make to read two marcks them that they g
[23:19:52] <twice2> anyone have xylotex 425 oz with 28800 input scale doing over 15 ipm?
[23:20:31] <Guest282> arrive from rwo swicth to gcode?
[23:21:27] <twice2> i didn't get it
[23:21:47] <crepincdotcom> twice2: check your BASE_PERIOD to see how fast youre polling the port. BASE_PERIOD/marks-per-rev=max rps
[23:22:23] <Guest282> Sorry my for my bad inglais
[23:22:29] <crepincdotcom> Guest282: i'm having trouble understanding what youre asking...
[23:24:04] <twice2> my bp is 14000, not sure what you mean marks tho
[23:24:24] <crepincdotcom> i, how many pulses per revolution your encoders give
[23:24:28] <crepincdotcom> *ie
[23:24:44] <twice2> heh steppers :)
[23:24:47] <crepincdotcom> ahh im sorry, i think i missinterpretted your question
[23:25:02] <crepincdotcom> yes, i definatly did
[23:25:02] <crepincdotcom> sorry then
[23:25:07] <twice2> np
[23:26:15] <twice2> i've been trying to do it oo calc sheet i.e. on paper to see what is realistic
[23:26:53] <crepincdotcom> its a trade off between top speed and torque
[23:27:57] <twice2> yep, just wondering if ne1 else has similar experience
[23:29:28] <fabiotv> hello
[23:29:54] <crepincdotcom> hey
[23:31:32] <fabiotv> speck italian3
[23:32:01] <crepincdotcom> nope
[23:32:54] <fabiotv> nope is nothing
[23:33:05] <crepincdotcom> nope == no
[23:33:13] <crepincdotcom> no i do not speak italian, i am sorry
[23:34:51] <fabiotv> there is someone that speaks it and knows well emc2
[23:35:14] <twice2> maximum step rate is one step per two BASE_PERIODs for step and direction, i got that but
[23:36:16] <crepincdotcom> fabiotv: i don't think so
[23:36:50] <twice2> the product of MAX_VELOCITY and its INPUT_SCALE for max request is where i'm in error
[23:39:11] <twice2> so 28800 (input scale) * 0.30 (Max_Velocity) is not 42300
[23:42:34] <jepler> twice2: are you talking about the error "The requested maximum velocity of [number] steps per second is not attainable"?
[23:44:13] <twice2> no, trying to calc input_scale * max_velocity like emc2 does so i can tune these steppers to top
[23:46:57] <jepler> twice2: where did you get the number 42300 from?
[23:47:00] <twice2> i've got them running, but only to 15ipm. If max possible is less than the requested step rate emc will recalc max_vel
[23:47:11] <twice2> dmesg
[23:48:18] <twice2> but it may not be accurate, it's otmh
[23:51:35] <jepler> the emc 2.1 "requested velocity .. is not attainable" message is based on STEPGEN_MAXVEL, not MAX_VELOCITY
[23:51:48] <twice2> so in calc i have my bp as 14000ns that is 35714 possible steps per second
[23:52:10] <twice2> it's not exactly what emc give but very close
[23:52:57] <jepler> emc does not give the exact base period you request, but a "close by" period. (It is restricted to the precision of a certain hardware timer)
[23:52:59] <twice2> really? hmm
[23:53:06] <jepler> the exact period is shown in dmesg
[23:54:02] <jepler> or in 'halcmd show thread'
[23:54:04] <twice2> so how would you calc max requested (input_scale * max_vel) right?
[23:54:30] <jepler> when I request 14000ns I get 13410ns
[23:55:32] <twice2> mk, what is your input_scale, i you don't mind telling me :)
[23:56:51] <jepler> I don't have a machine
[23:56:54] <jepler> I'm just a developer :-P
[23:57:11] <twice2> rsg i know
[23:57:51] <twice2> i'm not smart, just reading paragraph one basic config :)
[23:58:27] <twice2> requested steps per second