#emc | Logs for 2009-10-08

[00:10:16] <andypugh> I was puzzled about the photomodeller comment, so I read back through the archive a bit.
[00:10:59] <andypugh> I am still puzzled, but did see a mention of scrapyard, which in turn reminded me that there is a titanium-only scrap merchant in Sheffield.
[00:11:57] <andypugh> Full of beautiful and immensely expensive lumps of stuff that probably took days to machine, and then got messed up near the end, and scrapped.
[01:55:52] <toastyde1th> toastyde1th is now known as toastydeath
[02:22:18] <spherspline> for the record, hexapods are actually called PARALLEL KINEMATIC MACHINES
[02:22:22] <spherspline> after the fact.
[02:22:28] <spherspline> ive already done most of the simulator code
[03:04:53] <toastydeath> anyone use R
[03:05:20] <toastydeath> the language
[03:24:30] <jymm> Detail sander, Radio, Fan, or Hand Vac
[03:25:26] <jymm> I get one free with the purchase of my 4pc 19.2V cordless tool kit purchase.
[03:27:12] <jymm> I think the cordless hand-vac - pick up spills, lil crap, etc.
[03:28:38] <spherspline> hand vac, you can follow the highway street painter truck and suck up all the extra glas sparkle bead
[03:28:57] <spherspline> rather than paying potters' industries a buck a pound
[03:29:56] <jymm> lol, eh no thanks to either
[03:31:18] <jymm> funyn thing is the fan is $35, and the others are $25
[03:32:07] <jymm> http://www.sears.com/shc/s/p_10153_12605_00911595000P?keyword=11595
[03:36:39] <jymm> Hmmm, I wonder if I get the free fan, then exchange it for this blower and pay the difference? http://www.sears.com/shc/s/p_10153_12605_00911592000P?mv=rr
[03:54:16] <spherspline> this is what i have to deal with http://dixieeuh.ipower.com/page7.html at massive scale
[05:07:39] <spasticteapot> I'm thinking about building a CNC milling machine, and don't really know where to start on the software end of things.
[05:07:46] <spasticteapot> I learned a little CAD in high school, but that's it.
[05:08:03] <spasticteapot> Any suggestions for where I might learn to use some free/cheap (I'm a student) CAD software for Windows or, preferably, Linux?
[05:10:26] <toastydeath> what are you planning on making
[05:10:27] <dgarr> spasticteapot: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Cam
[05:10:41] <toastydeath> because i strongly recommend doing it by hand if you're not doing anything overly complex
[05:10:45] <toastydeath> like tracing pcbs
[05:11:04] <toastydeath> actual milling is often not very complex and can easily be done with some elbow grease
[05:11:37] <spasticteapot> toastydeath: 3D bulgy bits for headphones.
[05:11:49] <spasticteapot> Also, milled-in recesses for speaker drivers.
[05:11:58] <toastydeath> you might be a little disappointed with the free options
[05:12:00] <toastydeath> for anything 3d
[05:12:01] <spasticteapot> And I have no milling experience whatsoever.
[05:12:05] <spasticteapot> :(
[05:12:12] <toastydeath> in that it's essentially nonexistant.
[06:02:24] <jymm> 4pc 19.2V cordless kit w/ 2 batteries http://www.craftsman.com/shc/s/p_10155_12602_00911404000P?keyword=11404&sLevel=0
[06:02:46] <jymm> there's also a rebate for a free cordless tool if purchased by the 10th
[06:03:32] <jymm> I bought the addon 2yr instant replacement warranty and the out the door price was $200
[06:26:23] <alex_joni> jymm: nice
[06:27:24] <jymm> Yeah, a drill + 2 batteries are almost that price alone. the saws are jsut a bonus.
[06:28:00] <eric_unterhausen> if they gave you an extra battery instead, that might be a good deal
[06:28:13] <jymm> it comes with 2 batteries
[06:28:37] <eric_unterhausen> they should give you a coupon for another one not good for a year when those will die
[06:29:05] <jymm> See that's the thing, I bought the warranty. If anything dies in that time, I just take it in and get a new kit
[06:29:14] <eric_unterhausen> how long?
[06:29:21] <jymm> 2 or 3 years
[06:29:28] <jymm> instant replacement at any sears
[06:29:39] <eric_unterhausen> also good at kmart?
[06:29:48] <jymm> yep and OSH too iirc
[06:30:02] <jymm> they have to carry the kit of course
[06:31:20] <jymm> If the flashlight dies in 2yrs, just take it in and the receipt. They'll replace the whole kit
[06:32:33] <jymm> AND, you can even buy the warranty agian on the replacement!
[06:34:59] <jymm> Oh, it comes with a tool bag too
[06:36:32] <jymm> They do have a 12V charger available as well, which I thought was a nice option
[06:36:47] <eric_unterhausen> that is nice
[06:37:24] <jymm> If anyone is gonna get it by the 10th, I have a rebate form for a free tool
[06:37:45] <jymm> Detail Sander, Radio, Fan, or HandVac
[06:37:58] <jymm> (all cordless)
[06:39:38] <jymm> I wanted the Li-Ion, but just a bit out of my budget
[06:42:15] <jymm> LOL click on Jesse James... http://www.t-mobilemytouch.com/
[06:43:06] <mIreland> tap tap is this thing on...
[06:48:29] <mIreland> so 100' of ethernet cable later, and after 170 software updates my lathe is online.
[06:48:45] <eric_unterhausen> that's a lot of ethernet cable
[06:50:36] <mIreland> ircii is too minimal need to try another chat client. any recommendations?
[06:50:51] <jymm> weechat
[06:51:45] <mIreland> thanks i'll try it.
[07:02:09] <Valen> good luck getting it in 2 years
[07:02:13] <Valen> xchat is nice
[07:02:44] <Valen> I have decided, the next drill I get is going to be a 36 volt dewalt with A123 batteries
[07:02:53] <Valen> Li-poly cells all die in a year
[07:03:30] <eric_unterhausen> they wouldn't be lipoly, and they last a lot longer than that
[07:03:40] <eric_unterhausen> depending on how used and stored, of course
[07:05:52] <Valen> most li-ion arent much better than lipo
[07:06:11] <Valen> the A123 cells seem to be the ducks nuts at the moment
[07:06:25] <Valen> we abuse the crap out of them in robots and they are still good after 2 years
[07:09:53] <mIreland> much better.
[07:10:01] <Valen> mIreland theres also xchat if you like things pointyclicky
[07:10:34] <mIreland> this will be more like it i think
[07:11:05] <mIreland> this is terrible I'll never leave my garage again.
[07:13:50] <alex_joni> mIreland: lol
[07:16:09] <mIreland> ubuntu is great. i just sat through 170 softwore updates without so much as a warning
[07:17:31] <mIreland> including hostmot2. wonder what changes there are. seems i was running 2.3.0 before
[07:19:01] <mIreland> now definitely 2.3.3 i am amazed. sorry guys i live under a rock.
[07:20:55] <jymm> mIreland: Just FYI... http://news.bbc.co.uk/onthisday/hi/dates/stories/july/21/newsid_2635000/2635845.stm
[07:22:25] <mIreland> drat how do you open links under weechat? mite try xchat yet~!
[07:22:40] <jymm> right click?
[07:24:17] <jymm> if you're gui, you could just use pidgen
[07:25:29] <mIreland> i like to think i prefer terminal progs but there is of course a learning curve
[07:27:02] <jymm> mIreland: Pigden is already installed, you just need to configure it.
[07:27:24] <alex_joni> mIreland: mostly bugfixes for emc2 2.3.0->2.3.3
[07:27:41] <alex_joni> mIreland: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Released
[07:28:45] <mIreland> thanks aj. i was worried when i noticed so much activity on the drivers i was working with. im still getting up to speed.
[07:33:49] <mIreland> wow, candy.
[07:36:28] <mIreland> think i'll go w pidgin, it is colorful and as gui as it gets. thanx 4 tips
[07:42:55] <mIreland> Mire is still in the room.
[07:57:37] <mIreland> my lathe is nearly functional. next i have to turn my attention to producing g-code. I'd like to stay with free software if it works. any recommendations on that?
[07:59:29] <mIreland> I've written a lot of gcode by hand on a haas mill, so I have some ability already
[08:01:39] <archivist> I use a cam program called insiderearofskull
[08:01:52] <celeron55> there aren't lots of "finished" software for that, but there are a few interesting projects at least, and lots of small scripts
[08:01:56] <celeron55> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Cam
[08:02:33] <celeron55> and yes, i guess writing g-code by hand is underrated 8)
[08:04:51] <mIreland> we wrote a few programs for this HNC on it's original GE control... no canned cycles, just discrete g1 moves for pecking and such. blisterbutton keypad from '78.
[08:05:07] <mIreland> I'd like to hope it can get no worse.
[08:06:55] <archivist> in an editor its reasonable
[08:07:54] <celeron55> that way it's a bit like coding with any scripting language (except that the syntax sucks 8))
[08:10:05] <archivist> named variable syntax is plain awful
[08:12:11] <celeron55> i have also made g-code by making a small python script
[08:12:59] <archivist> I did a gear cutting prog with a php script
[08:13:18] <celeron55> it has only one problem, you have to remember to run the script every time you change it 8)
[08:14:08] <mIreland> what is the advantage of python for our application? i notice it is used very extensively.
[08:14:25] <archivist> mine asks for teeth qty etc and gives you the code to copy paste or download
[08:14:59] <celeron55> i made the script so that it had everything hardcoded in it
[08:15:13] <celeron55> and wrappers for g-code commands
[08:17:01] <celeron55> mIreland: i guess it's just the most popular scripting language, of course any scripting language would do
[08:18:19] <mIreland> marvelous. i have a window open for learning it right now
[08:20:18] <archivist> methinks php is actually more popular, but python is flavour of the month
[08:20:54] <celeron55> php is ugly
[08:21:04] <celeron55> but anyway, here's one python script i made
[08:21:06] <celeron55> http://sooda.dy.fi/paste/?XIM
[08:21:26] <celeron55> it reads a csv file and prints g-code
[08:21:58] <celeron55> actually it's for cutting the traces on a veroboard
[08:24:15] <mIreland> very tidy.
[08:36:11] <mIreland> time for sleep. good tidings all
[08:44:24] <archivist> celeron55, not a lot of difference http://pastebin.ca/1603633
[09:21:24] <celeron55> archivist: well, yeah,
[09:21:27] <celeron55> ops
[09:21:54] <celeron55> i mean yes, simple scripts in php aren't that ugly
[09:22:21] <archivist> even more complex can be clean, its often the programmers fault
[09:22:48] <celeron55> with python itäs hard to do as bad code
[09:24:13] <archivist> I absolutely hate the white space rules in python, I want brackets or something to deliniate a block
[09:25:41] <celeron55> i like the c syntax, too
[09:26:17] <archivist> I like that php is C like in many ways
[09:26:59] <archivist> although that last statement draws flames in ##php :)
[10:11:20] <MattyMatt_> MattyMatt_ is now known as MattyMatt
[10:29:10] <MattyMatt> python could be made to use {}, with the right editor
[10:29:33] <MattyMatt> I'm working on that :)
[10:31:01] <MattyMatt> then when a critical mass of people use it, the python compiler will be forced to accept it directly
[10:50:34] <MattyMatt> eprom programmer arrived
[11:16:53] <piasdom> g'mornin all
[11:29:57] <fenn> python already accepts curly braces and semicolons
[11:38:43] <kanzure> yes it's called perl
[11:42:58] <fenn> "the syntax is just #{ and #} instead of { and },"
[12:10:55] <_hype> hi
[12:11:01] <_hype> any cnc programmers here?
[12:11:37] <Valen> no, we all just pretend
[12:11:42] <Valen> yaknow to get the chicks
[12:11:50] <_hype> ahaha
[12:12:02] <_hype> im on a job now cutting plastc
[12:12:10] <_hype> using an 82 degree c sink
[12:12:17] <_hype> what rpm should i use?
[12:12:39] <Valen> nfi
[12:12:47] <_hype> ?
[12:12:49] <Valen> depends on the plasic as well
[12:12:52] <Valen> no "fine" idea
[12:13:06] <_hype> im thinking about using 1200
[12:13:37] <_hype> i dont have much tolerance
[12:13:39] <Valen> whats the plastic?
[12:13:52] <_hype> and dont want to create heat to open up the hole
[12:14:00] <_hype> its a hard plastic
[12:14:05] <_hype> 1.070 think
[12:14:08] <_hype> thick
[12:14:14] <_hype> 23.500 long
[12:14:20] <_hype> 4.500 width
[12:14:49] <_hype> im drilling the .344 hole first then the .675 82 degree csink
[12:14:56] <_hype> using a drill center first of course
[12:15:51] <_hype> i have to leave .260 to .240 meat at the bottom
[12:16:02] <Valen> its hard to give an accurate RPM without knowing the exact plastic
[12:16:52] <Valen> I'd probably try a fairly low RPM and a high feed rate and try and really cut the plastic off
[12:17:32] <_hype> im keeping it slow
[12:17:42] <_hype> 2 - 4 inches a minute
[12:17:56] <_hype> so paranoid of messing up
[12:18:02] <_hype> the plastic might shift
[12:18:04] <_hype> etc
[12:18:35] <Valen> got a bit to play with?
[12:18:42] <_hype> no
[12:19:08] <Valen> that makes life hard
[12:19:26] <_hype> yea especially when alot of money is envolved
[12:19:38] <_hype> this job goes to lockheed martin
[12:19:45] <Valen> The only problem I've had cutting plastic slow is it deforms rather that cuts
[12:19:57] <_hype> yea it gets reall hot and deforms
[12:20:05] <Valen> then get another bit of the plastic and do a few trials first
[12:20:38] <_hype> im not using vises
[12:20:44] <_hype> i cut a slot for them
[12:20:48] <_hype> using clamps
[12:21:02] <_hype> to hold it down
[12:21:16] <_hype> vises bow it up
[12:22:38] <_hype> what kind of cnc mill do you use?
[12:22:45] <Valen> but like i said, just get some more of the plastic and make some tests
[12:22:51] <_hype> yea
[12:23:30] <Valen> DIY one
[12:23:44] <_hype> i got a fadal
[12:30:06] <Valen> http://www.youtube.com/watch?v=4pAxqw5yeZM
[12:32:17] <_hype> http://www.youtube.com/watch?v=H_4Bd2VeFy8
[12:32:22] <_hype> that is the exact machine i have
[12:33:00] <Valen> running it with EMC?
[12:33:08] <_hype> no
[12:33:16] <_hype> i use bobcad
[12:34:07] <Valen> bobcad doesn't run gcode, it makes it is my understanding
[12:34:16] <_hype> yea
[12:35:40] <_hype> im not familiar with emc i was just looking for a machinist channel
[12:36:19] <Valen> EMC runs gcode
[12:36:29] <Valen> its like mach but cool
[12:41:45] <fenn> _hype: you have to take a thick chip, that means higher feed and slower rpm. otherwise you'll just melt it
[12:41:59] <MarkusBec_away> MarkusBec_away is now known as MarkusBec
[12:42:24] <fenn> also you never said what plastic..
[12:42:42] <_hype> i dont know the exact type of plastic
[12:42:46] <_hype> but its a real hard plastic
[12:43:41] <_hype> im afraid of higher feeds it might shift the part
[12:43:59] <Valen> you also need a really sharp cutter as a rule
[12:44:19] <Valen> on the lathe we made a few of our own, they cut like a knife
[12:47:19] <Valen> you seek to maximise jet power
[12:47:24] <Valen> sorry wrong window
[12:47:32] <fenn> MOAR POWER!
[12:47:59] <Valen> a friend is wanting to put a fire pump on his boat for use as propulsion
[12:48:12] <Valen> to his credit it is going to be a 40Kw pump
[12:48:22] <fenn> how efficient are those?
[12:48:56] <fenn> with a thrust augmentor it just might work
[12:59:16] <jymm> fenn you forgot the UG UG UG
[12:59:27] <fenn> piggly wiggly
[12:59:35] <jymm> Aroo?
[12:59:35] <Valen> ~80%
[13:22:16] <Dave911> Is anyone familiar with GE's old "Mark Century 2000" CNC controls?
[13:22:26] <Dave911> These are vintage 1984... Specifically, does anyone know the specs on these controls - do they use something close to EMC2 Gcode etc?
[13:54:51] <cradek> Dave911: probably the numbers are fixed precision with no decimal point
[13:55:29] <cradek> may not have a G00 (mark century 500 doesn't, you program something like N1G01X00000F8000 for a rapid)
[13:56:06] <cradek> so, not really very close at all
[13:56:59] <jymm> But cradek an whip up an emulator in 4.3239 seconds
[13:57:14] <jymm> while juggling three chainsaws
[13:59:05] <i-pink> hiii all
[14:00:34] <jymm> Is statd necessary? Box is going into datacenter.
[14:02:01] <cradek> not if you're not using nfs
[14:38:04] <cradek> dmesg
[14:38:05] <cradek> oops
[15:22:04] <Dave911> Whoa ... you think they used an implied decimal point? I did a few searches in GE Mark Century 2000 and found that GE has it listed in their product history - the Mark Century control was first brought out in
[15:22:06] <Dave911> 1974!! Yikes. Wasn't the Gcode RS274 standard first setup in 1972?
[15:22:44] <archivist> cnc is older than you think
[15:24:05] <Dave911> I think they called it simply NC way back when right? I worked with an old NC vertical mill back in the 1977 time frame as an operator... The tape was magically created by a guy in the "front office" secret stuff back then
[15:25:24] <archivist> 1960 http://www.collection.archivist.info/showresult.php?srcdata=title&srcprog=searchv10.php&searchv4page=1&errlev=0&searchstr=BK410+microwave&srcdata=subj&Type=BK&Accn_no=410&srcprog=searchv10.php&dir=&file=plate28s.jpg&subject=2844
[15:26:30] <L84Supper> Dave911 : did they tape out with a PDP-11?
[15:31:21] <cradek> probably a teletype, hand-typed
[16:15:14] <i-pink> hii all
[16:16:00] <micges> hi
[16:41:00] <skunkworks> our controll required leading zeros - didn't see the decimal point.
[16:46:09] <tom3p> the original AGie WEDM punched tape code handled lines circles and elliptical arcs, was extremely concise http://www.drahterosion.com/english/programming_e.htm#AGIECODE
[16:50:51] <tom3p> haha, now i remember early copy&paste: 2 teletypes or flexiwriters... advance the 'source' teletype to the position wanted by watching the printout, move the 'source' paper tape to the 'destination' teletype, copy till end of code fragment ;)
[16:53:03] <tom3p> whats the diff between % and 'h' ? answer: ISO vs EIA program start, or ... ASCII vs EBDIC .. or Teletype vs Flexiwritier, pick your weapon and go!
[17:01:35] <gene__> Oh boy, looks awfully quiet today, anybody home?
[17:03:10] <jymm> No, eave a message at the beeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeep
[17:03:26] <gene__> chuckle :)
[17:03:47] <gene__> Got a q re video speeds
[17:04:35] <jymm> speed kills, say no to drugs!
[17:05:08] <gene__> this is the box I may use for the bandsaw, and the only way I can get a good backtrace on the screen ios to exec a g61.1 before running the logo.
[17:05:25] <eric_unterhausen> you have a cnc bandsaw?
[17:05:58] <gene__> even a g64 p-0..1 seems to leave it free to blend acute corners .25" away from the paek.
[17:06:10] <jymm> gene__: in axis?
[17:06:14] <gene__> paek/.peak
[17:06:33] <gene__> yes, axis, freshly updated to 2.33
[17:06:58] <jymm> gene__: ok, out of my realm, probably cradek or jepler
[17:07:10] <gene__> 8.04 LTS install from the cd.
[17:08:11] <gene__> If I slow it down with the feed override, to about 10%, or to under an inch/min, it gets good
[17:08:29] <gene__> and I'm thinking it may be slow video as much as emc
[17:09:52] <tom3p> for study and discussion, make a simple program of 2 intersecting lines at some angle useful to you.
[17:09:54] <tom3p> then the people helping can examine same.
[17:09:54] <tom3p> by 'slow video' do you mean the axis display update?
[17:15:59] <gene__> I'm running ther axis logo. And the screen update seems reasonable, its an ati card, a 9200SE, with whatever driver the installe4r shose at the moment.
[17:16:55] <gene__> I am gettingsome realtime delays, but not storms of them, one per logo run, sometime 2.
[17:18:08] <archivist> latency test!!
[17:18:50] <gene__> You can see it on the first character where the top and bottom right cornerws are bypassed by about a 1/4" radius curve, uless I do a g61.1, then its ok, but obviuously slower, which may give the video time to catch up
[17:19:36] <gene__> latency test says way slower than I'm running, courrently base is 45000.
[17:20:32] <gene__> The machine is a 1ghz athlon, but the memory is not interleaved
[17:21:12] <tom3p> if you see a departure from programmed path when you see the latency error, then there's a conncetion, but i think you see the departure from path at other times,a dn see that it is velocity dependant.
[17:22:07] <gene__> your install of *buntu runs several times faster on thois box than any of the other 4 linux's I've had on this
[17:22:40] <gene__> I call this bos 'goat' cuz its sacrificial :)
[17:22:48] <gene__> bos/box
[17:23:40] <gene__> latency test, be right back, but its over 125k IIRC.
[17:23:47] <tom3p> study the simple program, its not a latency thing you see.
[17:25:27] <gene__> servo thread was about 1.5 milliseconds, base thread 148,xxx
[17:26:20] <gene__> It also averages the base thread at about 25 u-secs
[17:26:51] <gene__> Which figure to I use?
[17:30:16] <Dave911> >L84Supper Dave911 : did they tape out with a PDP-11?
[17:30:17] <Dave911> I don't remember but I believe they had an old green screen terminal so it may have been a PDP....
[17:30:19] <Dave911>
[17:35:13] <gene__> It seems as if the g64 p-.00001 even has no effect, only a g61.1 seems to work??
[17:48:09] <alex_joni> gene__: there's an option in the ini
[17:48:22] <alex_joni> under DISPLAY maybe, called CYCLE_TIME
[17:48:34] <alex_joni> it should be how fast AXIS redraws the screen.. maybe yours is set too low
[17:48:43] <alex_joni> * alex_joni would dig in that direction
[17:49:49] <gene__> ok, but I thinki I already cut that inhalf with no visible effect, brb
[17:53:43] <gene__> I just set that to 30ms, no effect. I hit esc to stop it, the tool pointer stopped instantly, then jumped about 1/5th of the way up the back side of the E to it final resting place
[17:54:48] <gene__> exactly or almost like the video was playing catchup. I'm going down to a 200 millisecond setting for S&G, brb
[17:56:39] <tom3p> whats the name of the file? my 2 rip versions dont come up with the 'letter' program, and i dont recognize the name in any of the nc_file folders
[17:59:14] <gene__> Running axis as the gui, and that is the "E MC2 AXIS" logo it initializes to when you run emc
[17:59:45] <gene__> I don't think it is in the nc_files folder
[18:02:10] <gene__> I just looked around in ~ but its not there, /share maybe?
[18:03:33] <tom3p> hah! this wont work either, from the axis.ini OPEN_FILE = /users/jepler/emc2-src/nc_files/flowsnake.ngc
[18:04:41] <alex_joni> tom3p: that only works on 2.3.x
[18:04:54] <alex_joni> err.. the path is probably wrong too :P
[18:05:00] <tom3p> gotta start the the blue border of the axis window should hav ethe file name in it
[18:05:20] <tom3p> yeh, unless i got some secret password i dont know about
[18:05:33] <alex_joni> gene__: what's your glxgears frame rate?
[18:06:05] <gene__> 563, steady
[18:06:36] <tom3p> gene__: the blue border of the axis window should have the file name in it
[18:06:36] <gene__> pretty good for the radeon driver :)
[18:07:34] <gene__> and my border is brown, and says "axis.ngc"
[18:07:39] <tom3p> thx
[18:07:44] <gene__> no path
[18:07:59] <tom3p> np. will search file system
[18:19:06] <gene__> bah, needs a //usr/share/axis/images/axis.ngc, damned copy/paste
[18:20:31] <tom3p> wth is it doing in /proc/17335/cwd/share/axis/images? in /proc!?!?!
[18:20:45] <frallzor> anyone into solidworks here? :)
[18:21:35] <kanzure> yes, but in a love-hate sort of relationship
[18:21:47] <archivist_attic> yup a bit
[18:22:20] <frallzor> is it possible to take a dxf of a sheetmetaldrawing, and bend the shit out of it?
[18:22:36] <frallzor> or do I have to use solidworks retarded way of making sheetmetalstuff
[18:23:30] <archivist_attic> it can show a flat of the bent sheet, then you can bend it
[18:23:43] <tom3p> gene__ ok, in usr;/share/axis/images , now to fix the default loaded program in some ini...
[18:23:45] <gene__> Humm, ndi, tom3p
[18:23:46] <frallzor> how?
[18:23:57] <frallzor> all i can find is how you do it the retarded way
[18:24:05] <frallzor> http://www.engatech.com/images/techtip7.jpg i want something like this to start with =)
[18:24:13] <frallzor> then just bend the shit out of it
[18:24:15] <archivist_attic> long time since I did it (a few years)
[18:25:01] <archivist_attic> you draw bent then flatten
[18:25:12] <frallzor> thats what i dont want to do
[18:25:21] <frallzor> the retarded way =)
[18:25:41] <archivist_attic> thats not the retarded way
[18:25:41] <frallzor> i have a dxf of the sheet and where to bend, its odd you just cant bend it
[18:27:03] <AchiestDragon> you just need to allow for the radius of the bends on the flat drawing
[18:27:49] <frallzor> say what
[18:28:39] <archivist_attic> its automatic if used correctly
[18:28:51] <frallzor> where is it found?
[18:29:28] <archivist_attic> * archivist_attic waits for it to start
[18:29:54] <frallzor> so i start with a sketch only then?
[18:29:59] <frallzor> or a solid from the sketch?
[18:30:06] <tom3p> gene__: ok, now it loads the file spec'd in the .ini ( as opposed to silently opening nada due to bad file descriptor ).
[18:30:07] <tom3p> can you describe a single thing to look for?
[18:31:34] <tom3p> frallzor: the length of a peice with just one fold, with expected 1" wall and 1" floor is longer than 2". this is the bend radius.
[18:32:12] <gene__> set the homes and run it after blowing the E up to a window full, the look at the outside corners of the E.
[18:32:25] <frallzor> that I know, but that wont do me any good =)
[18:32:30] <archivist_attic> draw a solid in the bent form, insert menu has sheet metal operations
[18:32:30] <tom3p> gene__ ok
[18:33:01] <frallzor> ok... I wont draw a solid in the bent form
[18:33:11] <frallzor> since that is what I dont want to do =)
[18:33:20] <archivist_attic> then no help shall be provided for the idle
[18:33:51] <frallzor> if I wanted to make it bent from scratch I can but i want my sketch of the sheet to bend =)
[18:33:59] <gene__> oh. my. a filosifer yet
[18:34:39] <frallzor> the result of that will be the result of what i have now =)
[18:35:32] <gene__> I can't spell it, so I figured I might as well mangle it :)
[18:36:29] <tom3p> the 1st home finance software i recall was nicknamed mangling your monkey
[18:37:26] <gene__> I don't think I'd want to trust my $ with that ones advice ;)
[18:41:02] <gene__> What sort of rounding do you see when you run the logo?
[18:41:58] <gene__> btw, if you have a machine hooked up, I don't atm, turn off the motor power.
[18:44:00] <tom3p> gene__: like this? http://imagebin.ca/view/umK6swsx.html i think thats normal emc/axis result
[18:44:37] <tom3p> manging your money --> mangling your monkey
[18:44:46] <tom3p> managing
[18:46:35] <gene__> oh. :) That is similar, but shows the screen updates as straight lines, mine are near perfect arcs
[18:47:03] <tom3p> the screen shot is 'out of the box', no mods at all
[18:47:49] <archivist_attic> gene__, have you set very slow acceleration rates
[18:48:14] <tom3p> i doubt ( hope) those jaggy lines are not what will really occur, no reason for a 'knee' to be off path.
[18:48:30] <gene__> look at the lower right corner of the E, where you will probably have short segmernts of straight lines, I have near perfect arcs.
[18:48:46] <alex_joni> frallzor: you need a sheet-metal package for what you want
[18:48:51] <alex_joni> * alex_joni uses alibre for that
[18:49:03] <alex_joni> but I'm not sure you can start from an import and bend it
[18:49:24] <alex_joni> you work the other way around.. you start with a sketch, then extrude it (always to the thickness of the sheet)
[18:49:34] <alex_joni> then select the edge and a feature (bend, etc)
[18:49:37] <gene__> Accel rates are fairly slow since this box is sorta slow. They are in the .35 range, with max speeds of .2
[18:49:52] <alex_joni> gene__: might cause the blending you're seeing
[18:50:01] <tom3p> line 38 to line 39? i got 2 segments off path, trying to be a blend radius
[18:51:05] <gene__> What sortof rates then, bearing in mind I found on the real machine, I could dpuble the speeds if I slowed the accels way down
[18:52:53] <gene__> tom3p, I found thew default ini that it make when you run stepconfig sets up too high maxvels and the accels can't keep up. I've had my .imi file into vim about 100 times I gue4ss
[18:53:28] <gene__> funkity keyboard, or funkity typer...
[18:54:35] <gene__> tom3p, I found the default ini that it make when you run stepconfig sets up too high maxvels and the accels can't keep up. I've had my .ini file into vim about 100 times I guess
[18:55:04] <gene__> no motors hooked up yet, but I already know about what these can do.
[18:55:31] <tom3p> gene__: just chgd ini to .2maxv and .35 acc all axis, re-running axis/emc
[18:57:17] <gene__> That may error too, stepgen maxvel needs 15-20% headroom to play catchup ISTR mine is .4
[19:00:24] <tom3p> gene__: now its a nice curve around the corner, hard to get a scale on it.. but again, looks like normal emc2 action ( me, i'd NEVER program noncontiguous/non-tangential... just me, but i get exactly what i ask for , never need to guess what the machine will do )
[19:03:21] <tom3p> when i go around the corner so fast that my rear slides out... well i'm going to damn fast for me and for my bike and this track!
[19:04:00] <cradek> gene__: don't confuse a jagged preview for bad machine motion - the machine motion is always smooth, and the preview gets updated and drawn when there is free time left over
[19:04:28] <cradek> gene__: the lower your acceleration, the rounder the corners you will get - you can cap the roundness with G64 P...
[19:04:48] <gene__> chuckle, yeah, I just tried to speed mine up and have following errors for joint 1 before it moves more than 1/2" from 0,0,0.
[19:05:05] <gene__> Back to the drawing board, brb
[19:05:13] <cradek> then I bet you've just missed a setting
[19:08:46] <tom3p> gene__ thats too damn slow for any use, it's still cutting the 'M' !
[19:11:13] <gene__> I gotta go see about a problem at the neihghbors, back whenever, sorry
[19:24:40] <L84Supper> anyone know a source for T-Slot framing that is curved to a radius? A few do it by special order
[19:31:23] <dr_head> dr_head is now known as dr_snack
[19:31:49] <dr_snack> dr_snack is now known as turkshead
[20:15:01] <gene__> Like a bad penny, I'm back
[20:15:26] <jymm> Never heard that one before
[20:15:41] <gene__> Had to go oversee some dirt moving, they are building a garage
[20:17:44] <gene__> And now its about tiome to go check what we're doing for dinner, this can wait, I haven't gotten the sled design for the saw set in my mind anyway.
[20:31:50] <jymm> Dallur: FYI... You *REALLY* shouldn't irc as root.
[20:35:33] <jymm> Dallur2: FYI... You *REALLY* shouldn't irc as root.
[20:36:36] <andypugh> But Dallur1 is allowed to?
[20:36:58] <jymm> Dallur1: FYI... You *REALLY* shouldn't irc as root.
[20:38:38] <andypugh> Does anyone know of an electrical component which conducts until 5V@2mA max is applied? (Other than a normally-closed relay with a transistor buffer)
[20:39:36] <andypugh> I want to keep some pins grounded until I output logic-1 on the parallel port.
[20:39:41] <jymm> a fuse?
[20:40:15] <jymm> why dont you use a pull-down resistor?
[20:40:21] <andypugh> A fuse might get annoying
[20:41:08] <andypugh> It's for the amp-enable of my stepper drivers. They need to be pulled down until I want them to work, but the total pull-down current is about 65mA
[20:42:42] <jymm> then use a igger resisot
[20:42:44] <jymm> r
[20:44:15] <andypugh> Then there won't be enough current through the optos to do anything?
[20:44:53] <andypugh> It's basically a fan-out problem. The p-port can drive 1 stepper input, but not 4.
[20:45:01] <jymm> Well, amybe someone else might have some ideas.
[20:47:06] <andypugh> OK, I am going to measure what my p-port lines do during boot.
[20:49:21] <celeron55> http://celer.oni.biz/~celeron55/random/2009-10/ppdisable.png
[20:49:24] <celeron55> something like that?
[20:50:53] <celeron55> hmm, you'd need to invert the input to the transistors i guess...
[20:50:58] <archivist> pport 3ma, opto in 20ma
[20:51:45] <archivist> use a buffer, ls244 or ULN2003
[20:51:45] <celeron55> how about now
[20:51:59] <celeron55> optos work with smaller currents too
[20:52:16] <celeron55> the current needed at the input depends on what current is needed at the output
[20:52:53] <celeron55> one opto i used conducted about the same current at the output what you put through the input side
[20:53:38] <archivist> depends on gain
[20:54:12] <celeron55> of course
[20:55:08] <archivist> I used a ULN2003 on the lathe because the stepper drives need a 12v open collector drive
[21:06:31] <andypugh> Sorry, back
[21:07:54] <andypugh> I looked at using a transistor, but I am not sure if the p-port sinks current when the computer is off?
[21:08:36] <archivist> that would switch the transistor off, which is a good thing
[21:09:07] <archivist> besides one uses a series R
[21:09:40] <andypugh> Not in this case, the drives are on unless there is current flowing in the opto.
[21:10:04] <andypugh> (No it doesn't seem sensible to me either)
[21:10:40] <andypugh> Simplest case is to leave the enable terminals unwired, and the drives are permanently on
[21:13:38] <andypugh> I was pointed at the 74LS244 by Kirk on the mailing list, but I don't see what it does that helps. I might be misreading the datasheet
[21:14:49] <andypugh> I do find myself coming back to using the P-port to operate a normally-closed relay.
[21:20:15] <skunkworks> http://www.cnczone.com/forums/showpost.php?p=659077&postcount=85
[21:23:04] <andypugh> Err, what is it?
[21:23:54] <archivist> vapour ware
[21:24:14] <skunkworks> heh
[21:24:30] <andypugh> Vapour-Hardware or Vapour-Software?
[21:24:38] <skunkworks> both
[21:24:47] <archivist> hehe thats a year old post
[21:28:22] <archivist> hehe less to laugh about on there pages, only one left http://www.safeguardrobotics.com/
[21:37:31] <Dave911> They launched that product way ahead of anything being ready and they sold a number of them - with alpha software.
[21:37:33] <Dave911> At first they were going to use Mach3 as the front end control software for that device, and they really sucked up a lot of info and got a lot of interest from Mach3 users... Then they decided to do their own MMI
[21:37:35] <Dave911> Honestly I don't know who would trust that he will actually come up with something. He might make nice web pages but the "Brain" sucks and a number of people bought a promise and got screwed.
[21:39:39] <spasticteapot> I want to build a CNC mill to produce simple 3d shapes.
[21:39:48] <spasticteapot> I'm told that EMC is not suitable for this.
[21:40:07] <archivist> it is I have made a mill
[21:40:30] <andypugh> If it isn't suitable for that, what is it suitable for?
[21:40:34] <archivist> mine is a 5axis mill with emc controlling it
[21:40:35] <spasticteapot> What software (commercial or otherwise) is recommended for the beginner.
[21:40:46] <andypugh> EMC. :-)
[21:40:51] <oPless> you need a mill first :)
[21:40:59] <spasticteapot> @lawl.
[21:41:13] <oPless> andypugh, EMC2 surely ? :)
[21:41:30] <spasticteapot> I found a purveyour of Chinese ballscrews of acceptable quality - $320 shipped gets me 900mm x 600mm x 320mm.
[21:41:43] <andypugh> I am a beginner, and I am finding EMC2 suite me fine. (Being both Free and free)
[21:42:01] <spasticteapot> I can buy THK rail and bearing cars for not much if I get 'em piece by piece.
[21:42:38] <spasticteapot> And there's a parallel port controller kit for $65.
[21:42:38] <archivist> you can also retrofit a mill
[21:42:39] <andypugh> I would suggest starting with a ready-made mill and converting, I think.
[21:43:12] <andypugh> There is a reason that most machine tools are made from big metal castings.
[21:43:23] <celeron55> depends on if you want to make a mill or if you want to mill something 8)
[21:43:25] <archivist> size and material you want to cut needs to be thought about
[21:43:27] <spasticteapot> Commercially made mills have too much backlash.
[21:43:48] <archivist> get an old cnc mill upgrade
[21:44:09] <andypugh> You should lose the backlash with the ballscrew conversion though.
[21:44:21] <spasticteapot> andypugh: There's also the not insignificant issue of cost and size.
[21:44:33] <archivist> you can also correct for backlash in clever gcode
[21:44:38] <spasticteapot> A 1.5' x 2' vertical mill would be HUGE.
[21:45:07] <andypugh> There are lots of benchtop mills out there.
[21:45:51] <spasticteapot> How hard is it to retrofit an old mill for ballscrews?
[21:46:03] <spasticteapot> Alternately, how much precision can I expect from a converted mill with fiddly gCode?
[21:46:36] <celeron55> how much precision do you need then?
[21:46:44] <andypugh> At least as good as manual milling?
[21:47:27] <archivist> I make clock gears with 9 thou backlash
[21:47:38] <andypugh> If you stick to the same rules with CNC as you would with manual, then the backlash is no more an issue with CNC than manual
[21:47:50] <archivist> gears are withing .1mm
[21:48:03] <spasticteapot> I don't really know much about milling - hence my desire to have the mill do the work for me.
[21:48:07] <spasticteapot> I've used a lathe a bit, though.
[21:48:11] <andypugh> (nice mixing of units Archivist)
[21:48:22] <spasticteapot> archivist: What are you using? .1mm sounds dandy.
[21:48:44] <andypugh> So you know how you compensate backlash on a lathe, the principles are the same
[21:48:51] <spasticteapot> Hmm.
[21:49:09] <archivist> my mill is an edifice from the scrap bin with a couple of new rotary tables
[21:49:15] <andypugh> Though I converted my mill/lathe combo to ballscrews
[21:50:25] <spasticteapot> archivist: Any recommendations for finding candidates for retrofitting?
[21:51:19] <archivist> depends on wallet, I just got a mid 1980's lathe on ebay
[21:52:23] <archivist> seen a few mills on there suitable, but price varies some makes have many chasing
[21:53:29] <andypugh> I couldn't bring myself to buy a proper machine tool and throw half of it away so I bought a cheap chinese lathe/mill combo and threw half of it away.
[21:53:54] <andypugh> I suspect that a proper machine tool might have been a better bet, though.
[21:53:57] <archivist> spasticteapot, desktop http://www.collection.archivist.info/archive/djcpd/PD/2009/2009_08_15/102CANON/IMG_0246.JPG
[21:53:59] <spasticteapot> archivist: What did you get, and for how much?
[21:54:16] <archivist> £270 Scotland
[21:54:29] <spasticteapot> Scotland?
[21:54:34] <archivist> Starturn was the model
[21:55:20] <archivist> I went for that as it had ballscrews from new
[21:55:30] <spasticteapot> I'd buy this if I had a place to put it: http://milwaukee.craigslist.org/tls/1407270953.html
[21:55:42] <andypugh> OK, what's the green thing?
[21:55:59] <archivist> so any Denford mill or lathe had ballscrews it seems
[21:56:18] <archivist> Mikron gear hobbing machine
[21:57:00] <archivist> spasticteapot, are you on the emc user mailing list
[21:58:09] <spasticteapot> archivist: Nope - I'm still trying to figure out what would be involved in buying a CNC milling setup.
[21:58:15] <archivist> look at a mail in the list with the subject "CNC machines in California"
[21:58:26] <archivist> whole pile available
[21:58:27] <andypugh> This is my conversion of a cheap chinese thing: http://www.youtube.com/watch?v=95B8PDlnajk
[22:00:57] <andypugh> http://cgi.ebay.com/Lathe-mill-combo_W0QQitemZ220490067164QQcmdZViewItemQQptZBI_Lathes?hash=item33563b6cdc#ht_500wt_1182
[22:01:02] <andypugh> Might work for you
[22:29:12] <dmess> hi all
[22:30:46] <dmess> pratt & whitney star-turn??
[22:32:56] <archivist> Denford Starturn 4
[22:34:04] <dmess> oh i worked on a p&w star-turn - tape and air driven NC lathe vintage 1967.. the 1st in north america
[22:35:09] <archivist> http://www.collection.archivist.info/archive/djcpd/PD/2009/2009_09_10_Starturn/
[22:39:25] <dmess> nice little lathe... I bought a little hardinge
[22:40:10] <dmess> with removeable cross-slide.. and i have 2 so 1 will be cnc and the other manual
[22:42:40] <andypugh> Indeed, a nice lathe, but that looks like a very old-school toolholder?
[22:43:12] <archivist> that was before I got a quick change toolpost
[22:44:15] <archivist> I had to change that to do repeatable tool changes for production parts
[22:44:55] <andypugh> What size do you have? I only ask as RDG tools got a stock of the 20mm capacity T1 size holders last week.
[22:45:10] <andypugh> (T1 is normally 16mm capacity)
[22:46:17] <archivist> its a 10mm I think
[22:46:55] <andypugh> Pah! you have a woman's quick-change-tool-holder sir!
[22:47:16] <archivist> its a child size lathe
[22:48:26] <andypugh> Yes, you could almost call it cute.
[22:48:50] <andypugh> Though you would only really get away with calling a lathe "cute" in certain circles
[22:59:05] <eric_unterhausen> the machinist at work said my 13" lathe was too small
[22:59:23] <eric_unterhausen> sorta
[23:00:35] <eric_unterhausen> I was trying to figure out why the oiler wasn't working, so of course it sprayed me in the face
[23:00:47] <mozmck> andypugh: have you seen the charge pump circuit Mariss did? http://www.machsupport.com/docs/ChargePumpSafety.pdf
[23:02:23] <mozmck> andypugh: re: your stuff on amp enable...
[23:08:03] <andypugh> No, I hadn;t seen those
[23:11:41] <andypugh> We would typically have a 500Hz charge pump frequency in EMC, yes?
[23:12:07] <archivist> just adjust values to suit
[23:12:22] <andypugh> Just checking my maths
[23:24:33] <mozmck> The charge pump frequency will depend on the period of the thread you attach it to.
[23:26:48] <mozmck> It's basically a high pass filter which then converts the signal to dc to enable the drives.
[23:32:28] <andypugh> I was imagining putting it in the servo thread, as there seems no point having it in the base thread, though it is about 4 lines of C.
[23:33:24] <andypugh> The high-pass filter is the capacitor, I assume, and the two diodes are effectively a rectifier?
[23:34:32] <andypugh> I can't help feeling there should be a curent-limiting resistor before the capacitor to save the p-port from high instantaneous currents.
[23:35:17] <archivist> first diode limits the the reverse bias, second rectifies,
[23:35:30] <archivist> its limited by the R to ground
[23:36:31] <archivist> the r to ground is the series r, the gate to the fet being open effectively
[23:37:10] <andypugh> Average current is, I was thinking in terms of instantaneous at the reversals. Not that electronics is a strong point of mine.
[23:38:53] <andypugh> Perhaps I need to buy a Mosfet then. I have an RS order open in the background (anyone need anything I have a business account to pay no postage)
[23:39:59] <archivist> the instantaneous in that circuit is miniscule, well within
[23:42:07] <archivist> time to go home
[23:44:23] <MattyMatt> parallel ports used to have a sacrificial LS244 or 373, usually in a socket
[23:44:49] <MattyMatt> you could make a buffer board to offer that same level of protection
[23:46:24] <MattyMatt> I'm thinking that a usb->parallel adaptor would be cheaper and more flexible, if your software works with it
[23:48:12] <MattyMatt> ignore me if this isn't a printer port you are talking about
[23:49:10] <MattyMatt> but using a buffer chip as a fuse does work in general
[23:49:56] <andypugh> That's not really my issue.
[23:50:31] <andypugh> I am trying to drive 4 optos from one p-port line, but I want them "On" when the computer is "off" too.
[23:51:17] <andypugh> So a relay seems the way to go, and driving one from the p-port is something I already know how to do.
[23:51:42] <MattyMatt> LS TTL floating inputs go high, especially with a little pull up
[23:52:04] <MattyMatt> so have the buffer at the opto end
[23:52:24] <andypugh> Sorry, I am not clear what you mean by that
[23:53:08] <andypugh> Part of the problem is that the "enable" opto shares a +V connection with the step and direction ones.
[23:53:49] <andypugh> As I need to turn "off" the enable opto to enable the drive, anything I use has to connect to the -V end of the opto.
[23:56:09] <MattyMatt> I scrolled up further. I'm at least #3 to say LS244 :)
[23:56:42] <andypugh> Yes, people keep saying that, and I don't see how it helps.
[23:57:49] <MattyMatt> I'd have to see all your circuits and all the datasheets and everything to see what the problem is, and I just woke up and it's 1am
[23:58:03] <andypugh> Are you suggesting controlling the "Enable" pin with the p-port and running each opto to ground through its own buffer line?
[23:58:03] <MattyMatt> how about pull-down on the opto enable?
[23:58:42] <MattyMatt> I dunno what ground means in an opto isolated circuit tbh
[23:59:04] <andypugh> That's what I intend doing with a relay. Pull-down the -V with a relay, open the relay with the P-Port to enable the drive
[23:59:31] <andypugh> OK, take the opto current to where it wants to go with a channel of the LS244?