#emc | Logs for 2008-08-29

[00:00:35] <toastatwork> dushantch: yeah, that's true, but this is a form tap
[00:00:40] <toastatwork> it doesn't cut, it just plows the metal out of the way
[00:00:58] <toastatwork> if it was a cut tap i wouldn't have given it a second glance, but i was expecting a lot more life out of it
[00:01:08] <dushantch> and al changes geometry of tool as it glues itself on its surface, so these two things combined could explain the breakage :)
[00:01:17] <toastatwork> again there's no cutting edge here
[00:01:31] <toastatwork> it's a forming operation =)
[00:01:48] <dushantch> :)
[00:01:58] <dmess> make sure your holes are o/s by enuf or it'll be a problem... did you drill all the holes with 1 drill??
[00:02:05] <toastatwork> dmess: yep, i did
[00:02:10] <toastatwork> the drill hasn't worn, either
[00:02:29] <dmess> hole size is what matters??
[00:02:38] <toastatwork> nope, hole sizes is just called out "Tap drill"
[00:02:42] <toastatwork> *size
[00:02:48] <dmess> 1st and last hole same size
[00:02:52] <toastatwork> yep
[00:03:22] <toastatwork> there's a bad habit around here of putting used taps back in the cases
[00:03:30] <toastatwork> so that's what i am thinking i got
[00:03:49] <dmess> but for a forming tap you usually have an o/s to std. hole prior to tapping.
[00:03:50] <toastatwork> whutevs, the machine is running, if it breaks again i'll look into it more
[00:04:19] <toastatwork> yeah, the hole is o/s
[00:04:35] <toastatwork> the std tap drill is .165, i have a .180 drill in the machine
[00:04:48] <dmess> its so there's room to displace the mat'l into the crest of the thd
[00:04:54] <toastatwork> yep.
[00:05:04] <dmess> that sounds about right..
[00:15:11] <JymmmEMC> http://newspaperads.mercurynews.com/ROP/ads.aspx?advid=32664&adid=6813831&subid=23166801&type=
[00:19:06] <tomp> nice laptop
[00:23:50] <tomp> hah, opened a wodehouse book using tts, 'the head of kays', it began speaking "Mainly about Fenn..."
[00:43:21] <jmkasunich> ok, mechanical question: which will clamp tighter, 1/4-20 or 1/4-28?
[00:43:52] <jmkasunich> (making a holder for a boring bar - drilled and reamed hole, split with a saw cut, three 1/4" SHCS to squeeze the slit shut
[00:44:07] <jmkasunich> I have both threads available
[00:48:15] <fenn> 1/4-28 because you get more leverage
[00:49:04] <jmkasunich> that was my first thought too
[00:49:18] <jmkasunich> but then I started thinking about it in more detail (I should never do that)
[00:49:49] <jmkasunich> it seems the limiting force is friction, which is proportional to the clamp load
[00:50:06] <jmkasunich> so at a given load, both screws will have the same friction force and torque
[00:53:39] <jmkasunich> http://www.cnczone.com/forums/showthread.php?t=56306
[00:53:44] <jmkasunich> looks like fine gets the vote
[00:55:54] <elric> hi is anyone there?
[00:56:09] <fenn> yep
[00:56:18] <elric> hi
[00:56:30] <fenn> what brings you to our fine channel this evening?
[00:56:36] <elric> i need some help to install the emc with the shell files
[00:57:01] <fenn> what seems to be the problem?
[00:57:03] <elric> but dunno if the ubuntu i hav support it or something
[00:57:13] <elric> did this [root@andLinux ~]# uname -r
[00:57:17] <fenn> what ubuntu distro are you using?
[00:57:38] <elric> n got this
[00:58:09] <elric> i read a log of this irc channel where some1 had the same problem
[00:58:29] <fenn> "Port of the Linux kernel that allows it to run cooperatively alongside another operating system on a single machine." ?
[00:58:30] <jmkasunich> fenn asked you what distro you are using
[00:58:33] <elric> im trying to install emc in andlinux
[00:58:42] <elric> dunno if it can be done
[00:58:42] <fenn> are you trying to run this on a virtual machine (inside windows)?
[00:58:52] <elric> yup
[00:59:12] <elric> just for some testing
[00:59:14] <jmkasunich> you will NOT be able to do machine control (or anything requiring realtime) using a virtual machine
[00:59:16] <fenn> that might work for sim, but i dont know anything about colinux
[00:59:44] <elric> hm
[00:59:52] <jmkasunich> the simulator mode should work, but you'll need to compile that - there are no binary packages for sim
[01:00:15] <JymmmEMC> tomp for $400, not bad at all
[01:00:34] <elric> so there´s no easy install for it
[01:00:49] <jmkasunich> easy is in the eyes of the beholder
[01:00:54] <elric> XD
[01:01:13] <elric> u mean the packages that r in the sourgeforge page?
[01:01:13] <fenn> it's actually quite easy to compile if you can follow instructions
[01:01:29] <elric> ic
[01:01:48] <jmkasunich> as long as you aren't afraid of the command line
[01:02:00] <elric> well im just starting with linux
[01:02:07] <elric> i guess i can try it
[01:02:57] <fenn> start here http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Installing_EMC2#On_Ubuntu_6_06_or_8_04_from_source and skip section 2.3
[01:03:37] <elric> nice tnx
[01:10:17] <fenn> jmkasunich: i dont understand the theory behind the "less bite" argument for fine threaded screws
[01:10:59] <fenn> if you have zero thread height (an infinitely fine thread) wouldn't it act just like a weld?
[01:13:05] <fenn> if the hole expands in diameter elastically more than the thread height it could just pop out, but i dont think that's what happens (in metals)
[01:13:48] <jmkasunich> the "less bite" argument?
[01:13:50] <jmkasunich> what's that?
[01:14:13] <fenn> that fine threads would strip because there's less "bite" than a coarse thread
[01:14:32] <fenn> i'm trying to figure out what "bite" is exactly
[01:15:38] <jmkasunich> hard to put into words - but I know for a fact that a coarse screw is far less likely to strip the threads in a soft material
[01:15:51] <jmkasunich> try a wood screw and a fine machine screw in wood sometime
[01:17:40] <toastydeath> that's because wood has less shear strength
[01:18:20] <toastydeath> and so the wood screw is designed to allow a gigantic tooth root in soft materials, providing the area required to resist stripping
[01:19:57] <fenn> but don't you get the same overall cross section along the length of the screw?
[01:20:12] <fenn> basically the surface area of the screw cylinder
[01:20:43] <toastydeath> no, because in a 60 degree thread profile, there's a 50/50 split in the cylindrical area between the bolt and the base material
[01:21:11] <toastydeath> in the wood screw, the bias is like 10-20% cylindrical area going to the screw, and 80-90 going to the base (wood, in this case)
[01:21:28] <fenn> but wood is much more elastic than steel, so most of the load goes on the outer edge (because the wood thread flexes) (i think)
[01:21:47] <toastydeath> doesn't matter where the pressure is, because the shear plane is at the root
[01:21:58] <fenn> uh, i think thats what i just said :)
[01:22:07] <toastydeath> i misread
[01:22:48] <fenn> * fenn looks up "shear plane"
[01:23:46] <toastydeath> http://en.wikipedia.org/wiki/Shearing_(physics)
[01:26:14] <toastydeath> the pitch of the screw doesn't actually affect the strength of the bolt very much, it just makes it harder to tap for any appriciable distance
[01:26:38] <toastydeath> so you don't see 40 tpi threads on big stuff because they'd get damaged really fast, plus it's difficult to do
[01:27:07] <fenn> i always thought screws were coarse thread because it relaxes the tolerances on both screw and nut
[01:27:12] <toastydeath> exactly
[01:27:14] <fenn> (wood screws aside)
[01:28:13] <jmkasunich> * jmkasunich admires nice new 1/4-28 threads
[01:28:59] <JymmmEMC> or woodEN screws =)
[01:30:26] <toastydeath> lol
[01:34:05] <dmess> http://www.youtube.com/watch?v=uVkJKp5O72s&NR=1 chk this out
[01:34:13] <dmess> japanese EMC
[01:35:32] <fenn> with tabs
[01:37:04] <dmess> nice punches... kudos
[01:39:26] <dmess> fine thds have more ultimate surace contact that why they are better
[01:41:00] <dmess> on smaller stuff anyway...
[02:10:22] <antichip> my DRO sold on ebay tonight, sO I am going to order a mesa parallel port card
[02:23:04] <jmkasunich> yay CNC (slitting saw, slow and repetitive)
[02:27:32] <mozmck> what are you slitting?
[02:36:44] <jmkasunich> a steel block for mounting a boring bar
[02:37:41] <jmkasunich> light machine, so I can only go 0.050" deeper each pass (started out doing 0.100, but after a few passes it got stuck)
[02:40:52] <mozmck> ah. I guess it's the slot where you clamp it on the bar right?
[02:41:12] <mozmck> I'd like to build a slitting machine - for cutting fret slots.
[02:43:07] <mozmck> The axis that moves the saw out and back would not really need to be CNC, so I've thought of just using a small dc motor and gear it for the speed I want, and make a simple circuit to reverse it when it hits a switch at the end of the travel.
[02:43:45] <mozmck> it would basically be a one axis machine.
[02:43:52] <jmkasunich> seems the geometry will make for a strange machine - isn't the saw arbor parallel to the neck?
[02:44:14] <mozmck> yes.
[02:44:26] <jmkasunich> that would require a very long arbor, as long as the neck - hard to make it rigid
[02:44:29] <toastydeath> like an old school travelling ram shaper
[02:44:37] <jmkasunich> either that ore a large blade
[02:44:42] <mozmck> the fret slots are usually cut with the fretboard off the neck
[02:44:52] <jmkasunich> still long
[02:45:04] <mozmck> I have a manually operated fret cutter now.
[02:45:28] <toastydeath> how long is this stuff
[02:45:32] <toastydeath> and how accurate need it be
[02:45:34] <mozmck> the saw arbor is mounted on the sliding short axis.
[02:45:52] <toastydeath> 40"? 30?"
[02:46:04] <mozmck> oh, I'd like to make it long enough to do any fret board. 30" maybe or so
[02:46:20] <toastydeath> how much cross stroke
[02:46:23] <mozmck> accuracy would need to be within .001
[02:46:34] <mozmck> cross stoke only about 6" max
[02:46:38] <toastydeath> how do you currently measure to .001"
[02:46:40] <toastydeath> calipers?
[02:46:54] <mozmck> the setup I have uses templates.
[02:47:18] <toastydeath> and the templates are guaranteed to .001" i assume?
[02:47:48] <mozmck> A cnc machine would allow me to calculate fret spacings for any scale and cut them in a few minutes.
[02:47:57] <mozmck> not sure on that :-)
[02:47:58] <jmkasunich> mozmck: it is very unlilely that you are holding 0.001"
[02:48:17] <toastydeath> right but like, machining something to .001" over 30" is not as simple as it sounds
[02:48:46] <jmkasunich> and given that the material is wood, it won't stay in spec long anyway
[02:48:57] <mozmck> I'm probably not holding that tolerance. it bugs me too.
[02:49:13] <mozmck> yeah, wood moves.
[02:49:19] <toastydeath> i'd look into the costs of 35" of ground leadscrew or rack/pinion accurate to .001" in 30"
[02:49:49] <jmkasunich> nah, rolled ballscrews + EMC2's leadscrew error comp
[02:49:51] <mozmck> I have a ballscrew I think is long enough.
[02:50:14] <toastydeath> oh, then awesome
[02:50:42] <mozmck> instrument building is probably the only woodworking craft where we routinely measure in thousandths.
[02:50:53] <jmkasunich> mozmck: what kind of cutters do you use? diameter, thickness, RPM?
[02:51:21] <mozmck> I'm using slitting saws .023 to .025" thick
[02:51:40] <mozmck> the machine I use was built specially for cutting fret slots.
[02:52:18] <mozmck> With accurate templates the repeatability really is quite good. Until stuff starts wearing out...
[02:53:25] <mozmck> I'll try and take a picture of what I have, and see if y'all have ideas on building an arbor. I have a lathe and mill so I can fabricate stuff.
[02:54:08] <jmkasunich> does the existing machine have a very long arbor driven from the end, or a short arbor with a belt or something to drive it?
[02:55:00] <mozmck> There's a short (4 or 5") arbor with a motor mounted on top and a belt to drive it. The motor moves with the arbor. Let me try and get a picture right quick...
[03:02:49] <mozmck> Here's a pic of it. http://imagebin.ca/view/j93Bus.html
[03:04:30] <jmkasunich> slides to/from the operator on those round rods to make the cut?
[03:04:38] <mozmck> yes.
[03:04:54] <jmkasunich> how do you adjust depth of cut?
[03:05:27] <mozmck> I figured on making the arbor part similar to that. but I would have the saw move down the fretboard instead of the table move.
[03:05:43] <mozmck> On this one you loosen some screws and adjust it.
[03:06:01] <jmkasunich> I can't tell how the motor/saw assy is supported
[03:06:30] <mozmck> let me put up another pic from the side...
[03:08:49] <mozmck> I would probably make a way to adjust the height with a screw. Most of the boards I do are the same thickness so I don't need to adjust it often
[03:09:01] <mozmck> http://imagebin.ca/view/zdvq6LuO.html
[03:09:22] <jmkasunich> ok, saw is fixed, and the front half of the table moves
[03:09:32] <mozmck> the mount is just a piece of 1" aluminum screwed to the base
[03:09:56] <mozmck> yes. on the cnc slotter I would like to make the saw move
[03:10:09] <jmkasunich> that would make the machine a lot smaller
[03:10:23] <mozmck> that's my idea.
[03:10:38] <mozmck> I'm not quite sure what to do for ways though.
[03:11:08] <mozmck> I have some round rods with linear bearings on them that might work.
[03:11:08] <jmkasunich> linear bearings - but they are expensive
[03:11:20] <jmkasunich> the problem with round rods is supporting the middle
[03:11:34] <jmkasunich> that kind of length, even a 1.5" rod will flex
[03:11:41] <mozmck> the rods are only abou 1/2" and the bearings go all the way around.
[03:11:47] <jmkasunich> forget it
[03:11:49] <mozmck> yeah that's what I was thinking.
[03:12:49] <jmkasunich> how wide is the widest fretboard you want to cut?
[03:13:04] <mozmck> probably not over 3"
[03:13:29] <jmkasunich> so if you got a 6" wide piece of steel, there would be space front and back for guides/ways/whatever
[03:13:45] <jmkasunich> the cutting forces are light, right?
[03:13:58] <mozmck> that might do it. yes, not much cutting force
[03:14:17] <jmkasunich> mill a V slot in your base, full length
[03:14:30] <jmkasunich> mount your 1/2" rod in the slot
[03:14:53] <jmkasunich> make another V in your carriage to ride on it
[03:15:17] <jmkasunich> better might be two sets of ball bearings, one set at each end, each set consists of two bearings at 45 degree angles
[03:15:53] <jmkasunich> that would locate the carriage in the front/back direction, and control rotation around the vertical and front/back axes
[03:16:01] <mozmck> I see.
[03:16:21] <jmkasunich> to control tilt around the long axis use a single ball bearing riding on the plate in front of the fretboard
[03:16:29] <jmkasunich> like a lathe - flat + V ways
[03:16:36] <jmkasunich> with the workpiece in between
[03:16:50] <mozmck> I was thinking of mounting it with the saw hanging out front, but maybe it would be better to stradle the board?
[03:16:57] <jmkasunich> I think so
[03:17:02] <jmkasunich> that way gravity is your friend
[03:17:05] <mozmck> yes.
[03:17:20] <jmkasunich> if both guides are in the back, the weight of the saw is pulling up on the rear one
[03:17:21] <mozmck> I was just looking at what I had...
[03:17:29] <mozmck> yep.
[03:18:19] <mozmck> yeah, that makes sense. I'll look at doing something like that.
[03:18:47] <mozmck> almost seems a waste to build a whole computer to run one axis...
[03:19:21] <jmkasunich> do you have any other CNC machines?
[03:19:35] <r0ns> wow
[03:19:41] <r0ns> good stuff
[03:20:00] <mozmck> not yet! I should be getting a router next week from the place I work.
[03:20:21] <jmkasunich> CNC router?
[03:20:24] <mozmck> yes
[03:20:42] <jmkasunich> why not figure out how to mount a fret saw "head" on the router
[03:20:51] <mozmck> I've done some trading and we getting a new router there.
[03:21:07] <jmkasunich> remove regular vertical spindle router motor, insert "thing" with horizontal arbor sticking out one side
[03:21:26] <mozmck> hmmm, might could. but I'd like to have a dedicated machine for fret slots.
[03:21:52] <jmkasunich> dedicated is nice - but you said "seems like a waste..."
[03:22:03] <mozmck> Actually, I might could cut the slots with a small mill. I can get them that small.
[03:22:16] <jmkasunich> bet you get a nicer finish with the saw
[03:22:23] <jmkasunich> unless you have really high speeds
[03:23:08] <mozmck> I don't know. I'm not sure how long a mill that small would last before it breaks either.
[03:24:10] <dmess> need a horizontal
[03:24:17] <mozmck> A router will do 25000 to 30000 rpm, should be fast enough.
[03:26:52] <jmkasunich> speed needs to be inverse to diameter - a 0.025 cutter needs to go 10x as fast as a 1/4"
[03:29:01] <mozmck> That would put it at over 240,000 rpm.
[03:29:23] <mozmck> speed in wood cutting is not quite the same as metal.
[03:32:25] <SWPadnos> don't forget that the feeds should also go up by the same factor of 10, to keep the chipload constant
[03:35:19] <jmkasunich> it would take experimenting of course
[03:35:31] <jmkasunich> but I bet you'd be happier with the slots from a saw
[03:35:43] <mozmck> yeah. .025 carbide bits work fine at 30000 rpm in rosewood and ebony. you just adjust your feedrate
[03:36:25] <mozmck> that could be. I do know there is a shop that cuts fret slots for a friend of mine on a cnc router or mill of some sort with a small bit.
[03:36:50] <mozmck> they look great. I need to see if I can go to their shop and see their setup...
[03:40:10] <jmkasunich> I guess another big factor is how deep you need to go
[03:40:54] <jmkasunich> I'm doing a slot that is about 0.040 wide and 0.600 deep, obviously a saw job - but yours might be different
[03:43:44] <mozmck> yeah. I'd really kind of like to build the cnc slotting machine. I think I can do it with spare stuff laying around and it would really be handy.
[07:49:59] <Guest871> hello, does the HAL have a GUI?
[15:23:01] <willburrrr2003> good morning all
[15:23:57] <willburrrr2003> does HAL have a gui?
[15:32:46] <willburrrr2003> anyone awake this morning ?
[15:33:01] <cradek> willburrrr2003: there is no good answer to your question. there are several ways of interacting with hal. ask a more specific question and get a more specific answer.
[15:34:08] <cradek> halscope and halmeter are programs that let you see what's going on at the hal layer
[15:34:17] <cradek> halshow also lets you see what's going on, in a different way
[15:34:59] <willburrrr2003> is there a GUI that lets me see the ladder logic operating in the HAL?
[15:35:13] <cradek> classicladder does have a gui that shows the ladder
[15:35:26] <cradek> HAL lets you connect things to ladder
[15:35:53] <willburrrr2003> ohh, so hal is just the "wiring" ? and clasicladder is the "plc" ?
[15:35:59] <cradek> yes exactly
[15:36:13] <willburrrr2003> :) great
[15:36:26] <cradek> hal also has some functions ("components") itself that are more complex than a plc, for example PID, or step generation
[15:36:34] <cradek> but you wire them up the same way
[15:36:45] <cradek> with pretend wires ("signals") in hal
[15:37:11] <willburrrr2003> where can i find out more on clasicladder?
[15:37:38] <cradek> if you have installed emc2, you can run a sample configuration that has a ladder, and have a look at it
[15:37:49] <cradek> run the demo_sim_cl configuration
[15:38:18] <cradek> pick File/Ladder Editor...
[15:38:26] <willburrrr2003> ok, thanks for the info I will have to check it out
[15:39:10] <cradek> welcome
[15:39:34] <jepler> a documentation page on ladder: http://linuxcnc.org/docs/html/ladder_ladder_intro.html some pages on our wiki about ladder logic: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?search=ladder&dosearch=1 the home page of the project that our ladder is derived from: http://membres.lycos.fr/mavati/classicladder/
[15:40:00] <cradek> I'm just finishing up a retrofit with a lot of logic written in ladder, including tool change, spindle gearing clutches, collet open/close, etc. it is powerful and pretty easy to use.
[15:42:51] <willburrrr2003> nice :)
[15:43:09] <willburrrr2003> i program plc's at work so i think i will enjoy ladder
[15:46:23] <willburrrr2003> what is the difference between the version that emc's ladder was made from and the version that comes with emc?
[15:47:15] <SWPadnos> EMCs version uses HAL for I/O
[15:47:49] <SWPadnos> I think there are some changes to the way modbus is handled (I'm not sure how well it works with the EMC version)
[15:48:04] <jepler> in emc, the modbus support is removed
[15:48:39] <SWPadnos> Chris recently mentioned that it's back in the CVS version
[15:48:47] <SWPadnos> but not in any released EMC2
[15:48:56] <willburrrr2003> oh, i see... possibly because the functions for that had to be used to make it talk with HAL and emc2
[15:49:24] <SWPadnos> right - I/O and realtime hare handled by HAL
[15:49:56] <SWPadnos> there are probably other behind-the-scene changes to make it work with our software build system, etc
[15:50:03] <SWPadnos> scenes
[15:51:13] <willburrrr2003> ok, well have to get back to work I think i will be chatting a lot with you guys, thanks for all the info
[17:34:38] <dmess> hi all
[18:36:28] <willburrrr2003> is classiclader v 7.124 in the new release of emc/ubuntu?
[18:37:23] <jepler> version 2.2.x of emc has classicladder based on version 7.100.
[18:37:52] <jepler> the development version has classicladder based in 7.123 (I think)
[18:38:38] <willburrrr2003> ok so the new changes for emc that are in 7.124 are not in effect in the distribution yet
[18:38:44] <jepler> that is correct
[18:39:08] <willburrrr2003> is there a way to manually update to the new version?
[18:39:46] <jepler> unfortunately, upgrading the classicladder version takes a fair amount of manual work
[18:40:43] <jepler> I was slightly familiar with this at one time (I'm the one who merged 7.100) but now Chris Morley (who is active on the mailing lists but not on IRC) is the one doing the classicladder integration work
[18:41:16] <willburrrr2003> ok, how do i get on the mailing lists?
[18:42:59] <willburrrr2003> I have been reading up on classicladder all morning, can't wait to get home and play with it
[18:43:03] <cradek> http://sourceforge.net/mail/?group_id=6744
[18:45:12] <willburrrr2003> you guys are awsome, can't believe i never noticed this IRC link over the last 2 yrs coming to look at emc
[18:46:52] <willburrrr2003> is it possible to use a PID loop with a stepper motor and a linear position sensor to make emc move the distance commanded (error correction if steps are lost)
[18:47:23] <cradek> theoretically yes, practically no
[18:47:41] <willburrrr2003> so i cannot make a closed loop control?
[18:47:52] <cradek> when PID senses increasing error (falling behind, lost steps) it commands faster motion. with servos that works. with steppers it makes the situation worse.
[18:48:10] <cradek> sure, if you use servos and encoders
[18:48:39] <willburrrr2003> instead of faster steps...can I change the prog to just give more steps until expected value is reached?
[18:49:09] <willburrrr2003> or less steps respectivly
[18:49:11] <cradek> that won't work either, due to the way stalled steppers behave (you must stop and then start again)
[18:49:31] <cradek> the best you could do, and it would be very useful, is to stop and abort when you sense you're not where you expect to be
[18:49:49] <cradek> then the physical cause of the problem can be fixed (machine crash? speed set too high?)
[18:50:22] <cradek> steppers don't randomly drop steps in a correctly configured machine unless something bad is happening
[18:51:17] <BigJohnT> my plasma would just lock up when I went too fast with steppers...
[18:51:33] <cradek> yes steppers just stall when you ask them to go too fast
[18:51:54] <cradek> issuing some extra step pulses can't fix it
[18:52:52] <cradek> IMO, if you go to the trouble of putting encoders or scales on a machine, you really want servos anyway, and then you get a real closed loop cnc machine
[18:53:53] <BigJohnT> * BigJohnT nods
[19:00:55] <willburrrr2003> ok, that makes sense :)
[19:01:10] <willburrrr2003> well off to lunch, have fun all
[19:02:51] <dmess> i'll be making a fully open loop stepper system on my lathe
[19:03:34] <dmess> wish me luck.. may end up with an encoder on the spindle to allow me to thread
[19:04:56] <BigJohnT> so long as you don't overload the steppers your OK AFAIK
[19:06:15] <jepler> a stepper machine is made reliable by picking conservative acceleration and velocity
[19:06:34] <jepler> find the values where your machine never ever stalls, then lower it
[19:08:31] <BigJohnT> and get the highest voltage your drives and steppers can take
[19:08:46] <dmess> my machine will end up having 5" xaxis and 6" z
[19:08:59] <dmess> so speed is not an issue
[19:09:36] <jepler> there's no debate that more voltage helps increase top speed of steppers
[19:10:49] <cradek> dmess: even on a small lathe, speed and acceleration are nice if you want to cut threads
[19:11:00] <BigJohnT> doesn't it also help the torque?
[19:15:56] <jepler> increasing voltage doesn't increase low-speed torque (the constant torque region), but it does extend the constant-torque region to higher RPMs and increase torque at RPMs above the constant-torque region
[19:17:21] <BigJohnT> Ok, I knew it did something but I had slept since I did that research :) for my plasma cutter
[19:19:37] <jepler> now what I don't know is whether stepper machines are more often cutting in the constant-torque region or not
[19:24:14] <BigJohnT> I didn't look into that very much as my plasma cutter does not have any extra cutting loads like a mill or lathe would.
[19:29:51] <BigJohnT> It does seem logical to want to stay in the constant torque region for cutting...
[19:59:23] <tom2> servo 'curves for torque are usually a flat (constant) up to some near max velocity, then they fall off.
[19:59:28] <tom2>  So you have rated torque available for that whole range.
[19:59:29] <tom2>  what do stepper torque curves look like?
[20:00:30] <BigJohnT> http://web4.automationdirect.com/static/specs/surestepmotors.pdf
[20:00:52] <BigJohnT> page 2 tom2
[20:01:56] <tom2> ooh! cut slow if u want lotsa toque!
[20:02:02] <tom2> thx BigJohnT
[20:02:50] <BigJohnT> yep the do best standing still :)
[20:03:01] <BigJohnT> the > they
[20:03:15] <anonimasu_> lol
[20:03:34] <anonimasu_> I like steppers best as paperweights.
[20:04:13] <BigJohnT> they are quite cost effective for some applications...
[20:04:44] <tom2> and the relation to ustepping is worth considering, get smooth fast motion and loose torque, use 1.2 step, get grainy, gain torque!
[20:04:50] <tom2> 1/2
[20:38:22] <jepler> huh I am surprised that torque differs that much depending on microstepping
[21:00:46] <willburrrr2003> do I have to setup pinouts for each parallel port above first one manually...or will stepconf wizard setup multiple ports?
[21:01:09] <jepler> willburrrr2003: stepconf does not do multiple-parallel-port machines
[21:01:30] <willburrrr2003> ok, was hoping but figured it wouldn't
[21:20:24] <willburrrr2003> that means i will be learning alot about hal pin assignments and other data files :) souds like fun
[21:27:25] <archivist_ub> just reading the scroll back, willburrrr2003 never run a stepper too near its resonant frequency either, try to accelerate through
[21:29:50] <willburrrr2003> what is a steppers resonant freq?
[21:29:57] <willburrrr2003> and whay is that bad?
[21:30:33] <SWPadnos> it's the frequency at which the stepper may lose steps
[21:30:40] <SWPadnos> and it's bad because the stepper may lose steps :)
[21:30:48] <willburrrr2003> hehe got it
[21:31:44] <willburrrr2003> so i did right in my tuning...found the settings that were just short of a stall, but noticable had lost steps, then backe dit down a bit until it ran smooth
[21:32:19] <SWPadnos> steppers may stall for (at least) two reasons: one is resonance, the other is that the load is too high for the available torque
[21:32:53] <willburrrr2003> i've only been playing with an actual machine for 2 days, (finally got my z axis wired and built wooohooo)
[21:33:39] <willburrrr2003> ok, so I may have been stalling from torqe requirements due to a p.s. that is just enough to run the system and not resonant freq?
[21:33:50] <SWPadnos> depends on the speed
[21:34:13] <SWPadnos> read this: http://geckodrive.com/upload/Step_motor_basics.pdf
[21:34:27] <willburrrr2003> i belive my max speed in stepconf at 1/2 stepping mode, was 2.5 at acceleration of 30
[21:34:28] <SWPadnos> and you'll know more than I remember at the moment ;)
[21:34:43] <willburrrr2003> ok, great thanks swpadnos
[21:34:55] <SWPadnos> in stepconf, the only thing that "microstepping" does is help calculate the SCALE
[21:34:57] <SWPadnos> sure
[21:35:12] <SWPadnos> it has no other effect as far as EMC is concerned
[21:35:32] <SWPadnos> ok - time to see if I can get this spindle to turn. bbiab
[21:35:41] <willburrrr2003> yes but when it was set wrong it did not match what my card was set for and i moved 2 times as far
[21:36:51] <willburrrr2003> ok, good luck SWP have fun
[21:37:34] <willburrrr2003> off to do some drafting for a bit myself
[22:05:38] <tom2> stuste1 do you want the 3 indicators to read 0 when the ball goes to position? despite angle of a & b?
[22:06:16] <Richmond100> hi all :) potential new user of emc2 here
[22:06:28] <SWPadnos> yay! :)
[22:07:08] <Richmond100> anyone from the uk in here ? or am i the only one ?
[22:07:17] <SWPadnos> there are a few
[22:07:23] <SWPadnos> dunno if they're "listening" though
[22:07:34] <Richmond100> ok thanks.
[22:08:14] <Richmond100> I'm planning on using the 5i20 card from Mesa and converting a Boxford TCL 240
[22:09:52] <Richmond100> IT was a lathe used mainly in schools and already has steppers, PSU and drivers in it. But previously linked to a PC via a dedicated Bradford instruments 1mhz bus.
[22:10:06] <SWPadnos> hmmm. never heard of them
[22:10:43] <tom2> boxford is nice stuff, (lotsa 'fords in uk small machines)
[22:10:55] <Richmond100> Southbend was the US name for the manual version of these lathes
[22:11:25] <tom2> way outclasses the chinese stuff, its shop quality
[22:12:05] <Richmond100> its is good....... I also have a larger Denford lathe based on the Syncro 280VS
[22:12:16] <Richmond100> I plan to use EMC on that too
[22:12:32] <Richmond100> anyway, anyone have experience of the MESA card ?
[22:12:42] <Richmond100> things to look out for, tips etc ?
[22:13:23] <tom2> how well did it work with the previous control? you can judge what you need to replace, and what to keep.
[22:14:19] <Richmond100> only the 1mhz bus and old Z80 processor need to come out. It runs fine in manual mode although the old drivers run hot. I might replace those too, but otherwise will leave eveything else
[22:14:47] <tom2> if the motors, speeds, capabilities were ok, then save the psu & motors, maybe you have limits to save. it sounds like a good conversion candidate
[22:15:16] <archivist_ub> Richmond100, swadlincote east midlands
[22:15:25] <Richmond100> the idea behind using the MESA card was so that all the original control buttons could be used as inputs to the mesa card, and still allow it to run in "manual" mode but via the PC
[22:15:40] <Richmond100> hi there _ub......... I'm in Hull
[22:16:09] <Richmond100> I have exp in CNC conversions, just not using this type of card, or with EMC
[22:16:26] <tom2> run the 'latency tests' on the candiate pc you may use. then decide on the interface (mesa and add on boards or parport ... theres many options, but you already have ideas for the mesa )
[22:16:40] <Richmond100> mainly used mach3....... but can be a #itch to program
[22:17:05] <tom2> oh good, the mesa is just very smart i/o then. you load a file that makes it this way or that or get real clewy and write your own file.
[22:17:41] <tom2> this way or that... = a stepper control and n bits of i/o or a boatload or i/o or...
[22:18:29] <tom2> and being familiar... the i/o level translation and/or opto-isolation is the add on boards
[22:18:49] <archivist_ub> Richmond100, we also have another lurker from the south in here as well, he and I have both done self build mills
[22:19:09] <Richmond100> i have several "PC's" to choose from from a 750mhz P3, upto my 2.5ghz laptop
[22:19:28] <Richmond100> altho laptop wont accept mesa card :)
[22:20:33] <Richmond100> I converted a boxford AUD many years ago, and all was well...... I just need to remember my unix stuff again. but looks like technology in that area has moved on too by look of Ubuntu
[22:21:50] <Richmond100> I too have a small mill with the 4th axis being a rotary table......all running on Mach3, so the par port can cope with "events" etc.
[22:22:24] <Richmond100> but the TCL "needs" to stay "original" on the surface.
[22:23:36] <SWPadnos> about how many buttons and lights are there?
[22:23:43] <SWPadnos> and knobs and stuff :)
[22:24:23] <Richmond100> about 20 including jog, spindle, coolant, air chuck, air guard etc etc etc
[22:24:31] <SWPadnos> hmmm
[22:24:38] <Richmond100> lpus i need to fit limit and home to it as well
[22:25:09] <SWPadnos> when you add in the other stuff that should be there, like limit and home switches, door / safety chain stuff, etc. you may need more I/O than even a (single) Mesa card can provide
[22:25:15] <SWPadnos> right
[22:25:37] <Richmond100> I thought 5i20 gave upto 72 pins of i/o though ?
[22:26:22] <Richmond100> only 6 are used for step / dir on the 2 axes plus 2 for the indexable toolholder
[22:26:32] <Richmond100> sorry 6 in total
[22:27:00] <SWPadnos> does the toolholder have its own PLC?
[22:27:08] <SWPadnos> ... toolchanger ...
[22:27:10] <Richmond100> nope runs from step / dir
[22:27:27] <SWPadnos> so there must be some inputs to tell EMC when it's at the correct position
[22:27:33] <SWPadnos> or in a position at all
[22:28:09] <Richmond100> I was planning to put and index switch on it to know when it was " loocked" or home ( pos 1 )
[22:28:42] <SWPadnos> sure, and you'll likely want to put something on that says "I've reached a tool position", even if all it is is an opto-interrupter
[22:29:04] <SWPadnos> in any case, things to think about as far as I/O count
[22:29:24] <SWPadnos> also, steppers have drive enable, step, dir per channel, so 3 lines per axis
[22:29:33] <SWPadnos> (or they probably should)
[22:29:43] <SWPadnos> actually, drive enable can be a single I/O for all motors
[22:31:34] <SkinnYPup> JMK had a discusson on ladder relating to turret switches recently if it may be of help
[22:31:37] <SkinnYPup> http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2008-08-04.txt
[22:31:51] <Richmond100> the plan was for 6 pins for step / dir on 2 axez and teh tool changer. One for master enable, 1 for Emer Stop, 2 for coolant ( flood and mist ), 4 for air op stuff ) tbc
[22:32:35] <Richmond100> plus about 20 for the "manual" controls, 2 for spindle, 1 input for spindle speed / index
[22:32:45] <Richmond100> thanks skinny I will check it out
[22:33:02] <SkinnYPup> Sure thing
[22:33:04] <SWPadnos> you'll need to use the new "hostmot2" FPGA code and driver
[22:33:27] <SWPadnos> the older one didn't allow you to re-task unused "special functions" as generic I/O (it also didn't have step generators)
[22:34:32] <Richmond100> Well the 5i20 was only ordered today so wont arrive till wed, so I have time to read the online manuals and get upto speed with things
[22:34:55] <SWPadnos> cool. should be fun :)
[22:35:18] <Richmond100> I hope so. I intend writing it up for MEW
[22:36:23] <SWPadnos> MEW?
[22:36:25] <archivist_ub> Model Engineers Workshop, uk magazine
[22:36:28] <SWPadnos> ah
[22:36:29] <Richmond100> ypu
[22:36:54] <Richmond100> yup..... I did the articles on restoring the Myford Super 7 recently
[22:37:23] <archivist_ub> I havnt read any recently so have not seen
[22:38:22] <tom2> the oldest weekly magazine in the world, started being published like just after king arthur. or moses ;)
[22:38:32] <Richmond100> TBH I wanted the write about it, and present the case for EMC..... most CNC in the hobby seems to be going Mach3. A good tool
[22:38:51] <Richmond100> that was ME, not MEW. Started being published aboutt 1990
[22:39:03] <SWPadnos> people are afraid of Linux
[22:39:03] <tom2> ah yes ME , sorry
[22:39:17] <Richmond100> no need to be though, esp now
[22:39:19] <SWPadnos> and there are a few things that Mach3 makes "easier" than EMC2 does
[22:39:22] <SWPadnos> I agree :)
[22:39:27] <tom2> linux... boogie boogie boogie
[22:39:41] <Richmond100> easier ? such as ?
[22:39:42] <SkinnYPup> Heck I am afraid of windows...
[22:39:53] <SWPadnos> "got a boogie on my finger and I can't get it off" - Weird Al Yankovic
[22:39:54] <Richmond100> i'm afraid of the wife!!
[22:39:56] <archivist_ub> I spend more time on linux now than windaz
[22:39:59] <SkinnYPup> lol
[22:40:18] <Richmond100> damn my head hurts now, i just got slapped :)
[22:40:31] <archivist_ub> * archivist_ub avoids Hull
[22:40:56] <Richmond100> he he....... well I dont actually live in Hull, just outside in the "nice" bit lol
[22:41:00] <archivist_ub> Im delicate
[22:41:28] <Richmond100> but she promised to kiss it better later lol
[22:42:13] <Richmond100> thing is...... I am an avid techy, cant get away from it, so EMC seems the way to go.
[22:42:38] <Richmond100> Mach3 easier ? such as ? how ?
[22:42:48] <SkinnYPup> wizzard !
[22:42:53] <archivist_ub> pointy clicky
[22:42:58] <SWPadnos> it's easier to go from your Windows-based CAD model to Mach (no network required)
[22:43:25] <SWPadnos> there are a few G-codes they've added (circular pocket, for instance)
[22:43:30] <Richmond100> can you set up partition to read NTFS files on win though ?
[22:43:48] <archivist_ub> yes
[22:43:48] <Richmond100> Samba it think it is ?
[22:44:04] <SWPadnos> their setup runs within the program, so you can flip to the I/O asignment page then back to the milling screen without re-starting the program
[22:44:05] <archivist_ub> samba is the networking method
[22:44:19] <SWPadnos> NTFS is the native filesystem for Windows :)
[22:44:36] <SWPadnos> you can read and write NTFS on Linux, until it gets changed without notice
[22:44:46] <archivist_ub> hehe
[22:44:47] <SkinnYPup> I prefer using windows in a virtual machine<in linux> when i need mastercam for something
[22:45:31] <Richmond100> lol i know :) I did my MCSE a few years ago but i havent use VMware yet..... must try it, but it is expensive though
[22:45:44] <archivist_ub> I swap hard drives to change operating system atm
[22:45:56] <Richmond100> unless there is a gnu version ?
[22:46:13] <SkinnYPup> Expensive? No vmware server is free
[22:46:20] <SkinnYPup> and there is virtualbox also
[22:48:45] <Richmond100> and XP runs ok ?
[22:55:04] <Richmond100> does it run on UBUNTU, or do XP and ubuntu both run as VM's ?
[22:59:10] <Richmond100> has the channel died ?
[23:02:47] <SWPadnos> no, the channel isn't dead
[23:02:58] <SWPadnos> people just get distracted :)
[23:03:12] <SkinnYPup> crap sorry washin the truck
[23:03:21] <Richmond100> np
[23:03:25] <SkinnYPup> it runs in ubuntu and you can create virtual machines
[23:03:33] <SkinnYPup> xp works great
[23:03:34] <Richmond100> ah, ok
[23:03:50] <Richmond100> just worried about using Sw2008 or ACAD in it
[23:04:17] <SkinnYPup> can run the virtual machine full screen , its released back to host os with ctrl + alt
[23:05:17] <Richmond100> I have a spare PC at present I will have to try it
[23:05:30] <SkinnYPup> I use mastercam and it is just as fast as if windows were an installed and booted os when simulating part and cut
[23:06:00] <Richmond100> seems I can only use ubunto 6 though according to vmware website
[23:06:11] <SkinnYPup> No I have it on 804 also
[23:06:57] <Richmond100> release notes say "Support for 32-bit Ubuntu 6.x as host and guest operating systems"
[23:07:26] <SkinnYPup> You will need to download your appropriate kernel headers for Vmwareserver to build properly
[23:07:52] <Richmond100> sounds like a pain though
[23:07:58] <Richmond100> is it well documented ?
[23:08:20] <SkinnYPup> Oh no not a pain at all
[23:08:41] <Richmond100> so, i need to bite the bullet huh ?
[23:10:16] <SkinnYPup> Virtual box is in synaptic, but I have run across a few things that won't install on a vbox windows machine that work fine on a vmware machine
[23:10:56] <Richmond100> which version should I download, seems to be 3 unix options
[23:11:15] <SkinnYPup> http://register.vmware.com/content/download-107.html
[23:11:33] <SkinnYPup> VMware Server for Linux.
[23:11:33] <SkinnYPup> The core application needed to run VMware Server and interact with it on the local machine. TAR Binary.
[23:11:52] <Richmond100> ok thanks :)
[23:12:04] <SkinnYPup> Make sure to install the kernel headers in synaptic before trying to build
[23:13:15] <Richmond100> ok. So basically, install UBUNTU 8, then install kernel headers, then install VMware. Where are kernel headers
[23:13:58] <SkinnYPup> I don't recall the path to them , you have to specify that in the install.
[23:14:44] <Richmond100> u mean the vmware install ? are the ubuntu headers, or vmware headers ?
[23:14:50] <SkinnYPup> You'll also need build-essential to install
[23:15:13] <SkinnYPup> During the vmware install on ubuntu you are asked the kernel header path
[23:15:47] <Richmond100> so basically i need the ubuntu src files installed too
[23:16:37] <SkinnYPup> Yep
[23:16:52] <Richmond100> ok. slowly getting there lol
[23:16:52] <SkinnYPup> For your current kernel version
[23:16:56] <SkinnYPup> thats ok
[23:17:42] <Richmond100> the last unix system i played with was hp ux about 10 years ago, all tty i/f, no graphics, or x system lol
[23:18:49] <SkinnYPup> I love it b/c I can be in windows, then share files back to the booted ubuntu, too easy
[23:20:44] <Richmond100> hmmmm, seems that the avg joe in the street wanting to play with CNC might have problems with this under linux...... If only there was a way of streamlining it, so that it is all on one CD and install, and tehn play with it on a "live CD", inc vmware etc etc etc
[23:21:29] <SWPadnos> aside from space problems (the livdCD is pretty full), I don't think anyone can distribute VMWare
[23:21:33] <SWPadnos> liveCD
[23:22:46] <Richmond100> ahh i see. are there plans for wizards on EMC ? I also read that G code diffs as well from standards in some areas
[23:23:01] <SWPadnos> and the really free VMs - qemu (and virtualbox?) are probably more trouble than they're worth for a demo (remember, there's usually no swap running from CD)
[23:23:08] <SWPadnos> there are some "wizard-like" things
[23:23:15] <SWPadnos> check the wiki (search for wizard)
[23:23:50] <SWPadnos> new ones can be written in Python - when you try to load a .py file, AXIS will execute it and use the output as the G-code program (if set up to do that)
[23:24:08] <SkinnYPup> Richmond you'll also need to get some serials from vmware, http://register.vmware.com/content/registration.html
[23:24:16] <SkinnYPup> I chose 100 ;o)
[23:24:44] <SWPadnos> there are some additions to the RS274NGC standard, but I think on the whole it's pretty true (EMC was originally written as a proof of concept and reference implementation for the standard)
[23:25:10] <SWPadnos> (actually, it's not a ratified standard, but it plays one on TV)
[23:26:25] <Richmond100> ok. thanks I have some thinking to do then. just looking at virtualbox now and seems to support ubuntu 8.04 str8 off
[23:26:49] <SWPadnos> note that 8.04 has wine installed by default - some stuf fmay just work without a VM
[23:27:10] <SWPadnos> oops - I thought it was default - maybe it isn't
[23:28:10] <SkinnYPup> richmond: I had no problem making virtual machines in vbox, but mastercam just would not install on that virtual machine.. complaining of crc errors
[23:28:58] <Richmond100> ok. changing tack........ what about linux cad / cam packages ? do they exist yet ? how do they perform ?
[23:29:40] <Richmond100> apologies if I am repeating Q's probably heard many times b4.
[23:30:05] <Richmond100> just trying to make an informed choice :)
[23:30:15] <archivist_ub> nothing up to a solidworks yet as far as cad goes
[23:30:29] <SkinnYPup> Reason I still use winderz a lil bit
[23:30:36] <archivist_ub> same here
[23:30:54] <Richmond100> but is there a reasonable 3D cad package ?
[23:31:50] <archivist_ub> Im waiting for a usable gui and some other sense on BRLCAD
[23:32:16] <SkinnYPup> I don't know of anything with bells and whistles , but glad BigJohn mentioned sheetcam
[23:32:46] <SkinnYPup> I gotta bail for dinner, there's also a #vmware channel on freenode should you need it , Good luck !
[23:33:21] <Richmond100> thanks for your help :)
[23:33:37] <SkinnYPup> Sure !
[23:36:32] <Richmond100> seems sheetcam only runs on debian
[23:36:47] <Richmond100> 2 secs got to make a coffee
[23:44:27] <dmess> http://cilqfm.corusradionetwork.com/shared/player/Default.aspx?id=cilqfm
[23:45:02] <JymmmEMC> ESXi FTW!
[23:45:27] <dmess> welcome to Q-stock
[23:46:30] <dmess> can you tune in Jymmmm
[23:47:02] <JymmmEMC> dmess: nfc
[23:47:22] <JymmmEMC> Fusion on the mac
[23:47:39] <dmess> too bad... 4 days of the concert that never was
[23:48:03] <JymmmEMC> oh, missed the url you posted.
[23:48:10] <JymmmEMC> ni audio on this box
[23:48:24] <JymmmEMC> But I got a new laptop yesterday http://newspaperads.mercurynews.com/ROP/ads.aspx?advid=32664&adid=6813831&subid=23166801&type=
[23:48:35] <JymmmEMC> Not sure if I'll keep it or not yet
[23:48:36] <dmess> zz top is up right now
[23:49:40] <Richmond100> wb
[23:50:08] <Richmond100> skinny. just looked at sheetcam, only seems to work on debian tho
[23:50:31] <dmess> quad core??
[23:50:54] <jmkasunich> wow, stuart's cinci must leak a lot of oil
[23:51:08] <SkinnYPup> richmond ubuntu is deb derived
[23:51:12] <dmess> why so??
[23:51:34] <jmkasunich> dmess: talking to me?
[23:51:45] <Richmond100> ahhh. so should work then, except only in alpha. Name is Keith btw
[23:52:03] <SkinnYPup> Works fine for me in 606 and 804 both
[23:52:10] <jmkasunich> dmess: every hole in the table is full, and the test jig is coated: http://www.mpm1.com:8080/cinci/s1030673.jpg
[23:52:28] <dmess> SI jmk
[23:52:29] <SkinnYPup> Sheetcam has a yahoo group also , fairly active
[23:52:49] <Richmond100> I like the idea of pcb milling and drilling....... need to make some interface boards
[23:53:25] <SkinnYPup> Speaking of pcb someone posted this on sheetcam forum today
[23:53:27] <SkinnYPup> http://web.mit.edu/~vona/www-NEW/Visolate/visolate/Visolate-info.html
[23:53:29] <jmkasunich> the test jig is for checking 5-axis kinematics vs. machine alignment - the "ball on a stick" at the top of the pic is in the spindle, if the kins are right, he should be able to tilt A and B all over the place without the ball center moving
[23:55:42] <dmess> even on a 21 " flat screen made in china still looks like shit
[23:56:08] <tom2> jmkasunich: is it custom kins? or could EMC do that already with 5 axis
[23:56:51] <tom2> haha look what visolate did with 'top'
[23:56:53] <jmkasunich> I'm not sure of the details - I think stuart is modifying the "standard" EMC 5-axis kins to suit his machine
[23:57:01] <tom2> ah