#emc | Logs for 2008-04-15

[01:04:04] <JymmmEMC> SWPLinux: ping
[01:08:24] <JymmmEMC> BigJohnT: coleco or atari 2600?
[01:15:46] <LawrenceG> prang
[01:19:53] <Tekrad> sweet...I fixed the infinitely variable speed change motor on the hardinge hlv-h :)
[01:29:16] <JymmmEMC> BigJohnT: ping
[01:45:01] <BigJohnT> watching MASH
[01:46:24] <BigJohnT> hoyt clagwell
[01:47:16] <JymmmEMC> BigJohnT: you mean M*A*S*H ?
[02:44:52] <Hugomatic> Hi guys... I'm starting to wish my Sherline was bigger. Is a CNC router like big mill or will it only cut plywood?
[02:45:36] <Hugomatic> Do you think Torchmate is a good brand?
[02:50:34] <jmkasunich> Torchmade sounds like a plasma cutter
[02:51:06] <Hugomatic> jmkasunich: right, but they also have router heads. What do you have?
[02:51:54] <Hugomatic> I mean, do you have a CNC machine at home or work?
[02:52:06] <jmkasunich> I have a Shoptask 3-in-1, converted to CNC
[02:52:20] <jmkasunich> just finished running this: http://jmkasunich.com/pics/port-fine-machined-2019.jpg
[02:52:33] <Hugomatic> nice... looks big and beautiful
[02:52:57] <jmkasunich> not that big - about 1-3/8" between screw holes
[02:53:37] <Hugomatic> jmkasunich: I was talking about the machine, not the part...
[02:53:58] <jmkasunich> work envelope is about 12 x 10 x 3.3, not really that big
[02:54:38] <jmkasunich> plasma cutters don't have to deal with cutting forces - even if they have a router head, a machine based on a plasma cutter design is probably only rigid enough to cut wood and such
[02:54:45] <Hugomatic> jmkasunich: you're right, not much bigger than the Sherline after all. But I bet you can take bigger cuts
[02:55:20] <jmkasunich> much - roughing for that part is a 1/4" endmill, cutting 0.075 deep, 0.100 wide, and 12 ipm
[02:55:52] <jmkasunich> actually, the envelope is probably more like 16 or 18 x 10 x 3.3
[02:56:30] <Hugomatic> Do you use a lot of CAM or do you mostly write gcode directly?
[02:56:49] <jmkasunich> I mostly write g-code directly, often using O-word loops
[02:57:05] <jmkasunich> the part in the photo uses something that jepler has been working on
[02:57:28] <jmkasunich> its very much alpha quality software, and you have to edit C++ source code to enter your part profile, etc
[02:57:33] <Hugomatic> I mostly write python instead of the o-word cause I 'think' they are emc specific
[02:57:48] <jmkasunich> O-words are EMC specific
[02:58:02] <jmkasunich> I operate in my basement, so I don't care about other controls
[02:58:17] <jmkasunich> and my real job has nothing to do with machining
[02:58:29] <cradek> I think writing .py scripts that print gcode and using python as an input filter in emc is a very good scheme
[02:58:30] <Hugomatic> What do you do?
[02:59:36] <cradek> Hugomatic: what do you want to make with a bigger machine?
[03:00:40] <jmkasunich> I'm an electrical engineer - I design large VFDs
[03:00:43] <Hugomatic> cradek: I don't know really. Maybe furniture or something. It just seems that I always try to make parts that are too big
[03:01:00] <Hugomatic> I'm a softy.
[03:01:03] <cradek> I know the feeling
[03:01:29] <cradek> wood is a totally different world from metal
[03:01:45] <Hugomatic> cradek: and metal is out of the question for routers?
[03:01:45] <cradek> a very different machine (especially spindle) is needed
[03:02:20] <cradek> except for cutting aluminum with very small tools, yes - the speed is way too high and the machine usually too floppy
[03:02:36] <jmkasunich> cutting forces are probably an order of magnitude different between wood and metal
[03:02:37] <cradek> I'm sure some machines are made to do both - with a swappable spindle maybe?
[03:02:53] <jmkasunich> its more than just the spindle
[03:02:54] <cradek> yes, and so are the spindle speeds
[03:02:58] <cradek> right
[03:03:02] <Hugomatic> cradek: what machine do you have?
[03:03:19] <cradek> I have a couple desktop machines but also a full size bridgeport
[03:03:47] <jmkasunich> the typical router gantry would probably deflect 0.020 if you put a couple hundred pounds on it - for metal you want deflections much smaller than that
[03:04:13] <Hugomatic> Aside from Sherline, what other machines come with EMC? I'd like to standardise on the software, at least
[03:04:34] <jmkasunich> its not so much that machines come with EMC - people put EMC on machines
[03:04:38] <cradek> EMC can run most any machine. Few come with it
[03:05:12] <jmkasunich> this machine is running EMC: http://jmkasunich.com/pics/g-and-l-yaxis-1856.jpg
[03:05:42] <cradek> if you don't need super precise new stuff, you can sometimes find old good cnc machines with dead controls, just waiting for EMC
[03:05:58] <jmkasunich> this is my machine: http://jmkasunich.com/pics/monitor-mounted.jpg
[03:06:00] <Hugomatic> jmkasunich: too big for me... my landlord would kill me ;-)
[03:06:58] <jmkasunich> better pic of my machine: http://jmkasunich.com/pics/pc-working-1761.jpg
[03:07:34] <Hugomatic> cradek: thats a good tip. But even a Bridgeport mill won't handle a full plywood sheet. But I guess that's what I would like
[03:08:02] <cradek> a bridgeport would be very bad at cutting wood
[03:08:28] <cradek> (and I'd know, I've done it)
[03:08:29] <Hugomatic> jmkasunich: nice machine. I envy you.
[03:08:34] <jmkasunich> right - the ideal wood-and-metal machine doesn't exist
[03:08:51] <cradek> one of my machines running emc: sherline lathe (servo retrofit) http://timeguy.com/cradek/cnc/lathe
[03:08:55] <fenn> *hexapod cough splort*
[03:09:03] <jmkasunich> wood you are usually looking for big work envelope, high speed spindle - rigidity is less important, and you can afford a big machine (cost and weight)
[03:09:27] <jmkasunich> for metal you need rigidity almost above all else - that means lots of weight, so the work envelope is smaller
[03:09:45] <jmkasunich> metal capable machines with a 4' x 8' work envelope are quite a few tons
[03:09:58] <fenn> rigidity is a matter of geometry, the mass is for vibration damping
[03:10:07] <Hugomatic> cradek: yes, I've seen your machine and your site before, you're the 'timeguy'!
[03:10:14] <cradek> Hugomatic: http://timeguy.com/cradek/
[03:10:29] <cradek> did you see the video of my little 5 axis mill?
[03:10:36] <jmkasunich> goodnight all
[03:10:49] <Hugomatic> creadek: So you don't think Sherlines are toys
[03:10:50] <fenn> g'night
[03:10:51] <Hugomatic> ciao
[03:11:25] <cradek> Hugomatic: the lathe is surprisingly good for stuff, but I think the mill really can't cut much metal
[03:12:10] <cradek> I think the sherline stuff is good at what it says it does.
[03:12:17] <Hugomatic> cradek: that's encouraging, cause I've been trying really hard.
[03:13:34] <Hugomatic> Are you guys prepping a new EMC? Is there a possible Ubuntu upgrade on the radar?
[03:13:34] <cradek> as with any machine, it helps to learn about speeds and feeds, and actually run the numbers, until you get a very good feel for it
[03:14:05] <cradek> Hugomatic: we'll have a hardy-based install ready some time after hardy is officially released
[03:14:35] <cradek> but that's just the OS; EMC itself doesn't care much about that, and work in EMC continues regardless of the Ubuntu front
[03:14:59] <Hugomatic> cradek: I've been sloppy. I always run at max speed, and vary the feed
[03:15:16] <cradek> Hugomatic: that can be exactly the right thing to do in some materials with some tool sizes
[03:15:39] <cradek> in others you'll just burn up your tools
[03:17:02] <Hugomatic> And I also like your python scripts (image to gcode)... I have written a few routines to write my own.
[03:17:49] <cradek> thanks for the kind words
[03:18:07] <Hugomatic> thanks for the free tools
[03:18:08] <cradek> I think I'm off to bed - goodnight and good luck finding what you are seeking
[03:19:04] <Hugomatic> ciao... I'll think about what you said. I hoped a router was just a big mill.
[06:35:23] <sweeper> http://jamphat.com/rap/
[06:41:32] <micges> hi
[06:41:36] <sweeper> ho
[06:47:36] <micges> hi
[06:53:16] <K`zan> Night Folks.
[11:36:11] <micges> I saw today one small error: when running program step by step and step into G4 code then after that interpreter is in reading state, not paused that is should be
[11:36:57] <micges> when step by step program without G4 then interpreter is always in paused state
[11:41:20] <BigJohnT> how do you tell that?
[11:43:03] <micges> BigJohnT: I do not understand
[11:44:00] <BigJohnT> how do you know it is in the reading state? just wondering
[11:44:58] <micges> pause button bacame unpressed and It is shown in EMCSTATUS window
[11:47:59] <alex_joni> micges: that's not so simple
[11:48:19] <alex_joni> interpreter reading / pause is also dependent on the length of the program
[11:48:54] <alex_joni> so if you have a very short program with no G4, then you'll probably have interp paused (because it reached the end of the file, and paused interpreting)
[11:49:14] <alex_joni> if you have G4 then it doesn't read further (if I remember it correctly)
[11:49:16] <micges> oh
[11:49:34] <alex_joni> micges: I spent quite some time fixing pause/step/pause
[11:49:41] <alex_joni> and this was the best I could come up with..
[11:49:51] <alex_joni> anything else had other problems
[11:50:14] <alex_joni> (btw, I noticed that in AXIS if you start a program with step, pause is not in the desired state..)
[11:50:23] <alex_joni> but other than that I don't think there are any issues
[11:50:43] <alex_joni> micges: what version are you running?
[11:50:50] <micges> 2.2.2
[11:51:27] <micges> sorry 2.2.4
[11:52:04] <alex_joni> yup, I changed it quite a bit for 2.3.x
[11:52:21] <alex_joni> now the AXIS pause button doesn't get it's info from the same place as in 2.2.x
[11:52:44] <micges> I saw this becouse I'm writing own interface and must program logic for start/step/pause//stop
[11:53:10] <BigJohnT> bbl
[11:53:18] <micges> where you write the changes ? axis or interpreter ?
[11:53:35] <alex_joni> date: 2007/11/12 22:23:54; author: alex_joni; state: Exp; lines: +3 -1
[11:53:35] <alex_joni> * add a new task status variable
[11:53:35] <alex_joni> this one keeps track of the task paused state, and GUIs should update the state
[11:53:35] <alex_joni> of the pause button accordingly
[11:53:35] <alex_joni> * update axis to use the new task_paused variable
[11:53:43] <alex_joni> micges: all over ;)
[11:54:18] <micges> this is in trunk ?
[11:54:22] <micges> I cant find it
[11:54:32] <alex_joni> http://cvs.linuxcnc.org/cvs/emc2/src/emc/task/emctaskmain.cc.diff?r1=1.99;r2=1.100
[11:55:07] <alex_joni> 1.100 is 2 revisions after v2.2 was branched
[11:55:08] <micges> thanks
[11:55:11] <alex_joni> http://cvs.linuxcnc.org/cvs/emc2/src/emc/task/emctaskmain.cc?graph=1
[11:58:56] <micges> alex_joni: can you simply explain me trunk/branch/release ?
[11:59:21] <micges> release I know - version that was released as packets
[12:00:40] <sweeper> TOO MUCH COOL STUFF TO PLAY WITH
[12:00:41] <sweeper> like the idea of inkjet + cnc
[12:00:46] <sweeper> and the ngw100
[12:03:06] <micges> ok I understand now
[12:06:42] <sweeper> anyone manage to run emc on the NGW100? it looks custom made for the job!
[12:06:50] <sweeper> and only $80 too...
[12:09:04] <Guest220> Guest220 is now known as skunkworks_
[12:21:44] <jepler> sweeper: you'd have to first port adeos ipipe/rtai to avr32. next you'd find that avr32 doesn't have a floating-point unit, and emc2 freely uses floating point for everything, so either you extensively modify emc2 to use fixed-point arithmetic everywhere or you get performance 10-100x lower than you hoped for by using a software floating-point emulation.
[12:22:13] <jepler> except for the size, you're better off in every way when you pick up a cast-off Pentium III system for emc
[12:26:59] <skunkworks_> good morning jepler
[12:27:12] <jepler> hi skunkworks_
[12:29:17] <sweeper> jepler: roger! :D
[12:29:49] <alex_joni> jepler: there was some ARM support from RTAI
[12:29:51] <sweeper> jepler: well, one benefit the avr32 has over the p3 is the 40ish GPIO lines ;)
[12:29:57] <alex_joni> but I (personally) wouldn't go there
[12:30:27] <jepler> alex_joni: AVR32 is not ARM, is it?
[12:31:13] <skunkworks_> how much does the avr32 cost?
[12:31:35] <sweeper> skunkworks_: the ngw100 board costs $80
[12:31:51] <sweeper> http://www.avrfreaks.net/wiki/index.php/Documentation:NGW/NGW100_Hardware_reference <-- that one
[12:32:32] <sweeper> actually, it's closer to 60 GPIO lines
[12:34:16] <sweeper> http://www.sparkfun.com/commerce/product_info.php?products_id=8607 <-- mmm, this is pretty, but they don't say what the resolution is on this particular version is
[12:35:04] <sweeper> they mention cnc, but I'm gonna go ahead and doubt them, since an actuator small enough to give proper resolution would be tricky
[12:36:52] <sweeper> bah, only 5 inches long
[12:36:58] <sweeper> maybe 4 inches of actual sensor
[12:37:33] <jepler> bbl
[12:37:38] <jepler> * jepler drives to the office
[12:38:05] <sweeper> is there a better place to discuss this sort of thing? this is really the only cnc-related channel I know of :D
[12:38:13] <archivist> resolusion is the a to d reading the device, accuracy, well your gues is the best you will get
[12:39:08] <sweeper> yea...
[12:39:20] <sweeper> I wonder how temp sensitive it is
[12:39:24] <alex_joni> sweeper: I used a SBC (geode CPU on it) with emc2
[12:39:35] <alex_joni> it was around 2-300$, but it featured some GPIO
[12:39:41] <alex_joni> which I never got around to use/code for
[12:40:05] <alex_joni> one 2 x parport card is 10-20$, so it's not interesting to do it differently
[12:41:30] <archivist> sweeper, specs on spec pdf, would stop me thinking of that for any serious work
[12:41:54] <sweeper> archivist: yea, +/- 20% is... D:
[12:42:01] <archivist> 5% linearity and 20% resistance
[12:42:14] <archivist> linearity is the bad one
[12:42:25] <sweeper> oh, gotcha, yea
[12:42:39] <sweeper> you could calibrate resistance with stops, I guess
[12:42:56] <archivist> nah calibrated voltage sourcs
[12:43:13] <archivist> or from the ref that the a/d uses
[12:44:46] <sweeper> hmm. would 1% linearity be enough? does 1% translate to 1mm over 100mm?
[12:44:58] <archivist> ish
[12:45:57] <archivist> and toys like that are worst at the ends
[12:46:27] <sweeper> still not good enough. blah. ALTHOUGH I suppose someone with a really accurate cnc bot could map the resistance curve and sell the pot paired with the graph
[12:47:19] <archivist> you could calibrate
[12:48:39] <sweeper> yea
[12:49:04] <sweeper> * sweeper squirrels it away in his "TBD" bookmark folder
[13:27:24] <ALS> Hey I'm writing a program to insert and remove Line# and was wondering if there are other charaters that would lead a line '(' '/' letters ?
[13:32:37] <ALS> brb
[13:35:10] <Vq^> ?
[13:35:15] <cradek> ?
[13:36:04] <ALS> block skip,comments,?
[13:36:10] <cradek> http://www.linuxcnc.org/docview/html//gcode_main.html#r2_1
[13:36:16] <ALS> others?
[13:36:45] <ALS> '[' this one?
[13:38:15] <ALS> just wpondering if i'm missing any?
[13:38:22] <cradek> did you look at that link?
[13:38:44] <cradek> it describes what is allowed in great detail
[13:39:04] <ALS> ya i've there there
[13:39:22] <ALS> been there
[13:55:20] <ALS> what about _a,_b,_c "The global parameters _a, _b, _c, ... _z have been reserved for special use. In the future"?
[13:55:39] <ALS> could they lead a line?
[14:00:01] <cradek> do you mean you want to add remove line numbers only (like N001)?
[14:00:29] <cradek> if so, it seems like you wouldn't have to parse the whole line
[14:00:58] <ALS> why
[14:01:40] <cradek> because they can only be at the beginning of a line (possibly preceded by /) and they have a fixed format (N followed by up to 5 numbers 0-9)
[14:02:01] <cradek> so if you don't see the line starts with N or /N, it doesn't have a line number
[14:02:14] <ALS> compare for 'N'
[14:05:28] <ALS> i'm looking for the first instance 0f letters after the first letter the N and or other characters
[14:06:08] <cradek> that sounds unnecessarily hard
[14:07:03] <ALS> how would you remove a line#?
[14:08:04] <cradek> if there is a N in the first or second column, remove the N and any numbers 0-9 that immediately follow it
[14:08:20] <cradek> also remember you have to ignore any whitespace
[14:08:53] <cradek> so you have to decide what to do with a line like "/ N 12 3 M 2"
[14:10:13] <cradek> in my example you know to stop removing things (and you can stop parsing the line) when you see the M
[14:12:00] <ALS> the '/ would remane N.../ would take out the line #
[14:12:43] <cradek> brb
[14:12:53] <ALS> yep
[14:17:50] <archivist_emc> hmm /me hugs an evil perl script building some xml (sucking cpu) and irc and EMC cutting a gear
[14:18:04] <archivist_emc> and vnc
[14:22:21] <ALS> N00020(1/4 end mill) or N00001#<_endmill_rad> = [#<_endmill_dia>/2.0] looking for these also?
[14:24:27] <ALS> cradek: ^^ ^^
[14:25:58] <ALS> brb
[15:44:12] <ALS> Thanks BBL
[16:10:33] <tomp2> join #cam
[16:11:06] <tomp2> :)
[16:13:35] <archivist> #cam? a competing channel
[16:19:07] <tomp2> discusses APT, can output gcode forr emc2
[18:13:36] <cradek> huh, 15% off enco orders until Friday
[18:21:51] <skunkworks_> cool
[18:22:19] <cradek> I don't need anything though...
[18:22:59] <skunkworks_> When has that stopped you before ;)
[18:23:18] <cradek> yeah, their free shipping always gets me too
[18:23:42] <archivist> get some stock!!
[18:24:06] <cradek> true. it's not getting any cheaper is it
[18:24:45] <archivist> heh prices over here are getting silly, especially fuel
[18:24:55] <alex_joni> cradek: until next time
[18:25:10] <cradek> you too? I figured it was just the crashing dollar at work
[18:26:36] <archivist> methinks a world
[18:26:53] <archivist> wide correction is happening
[18:28:12] <skunkworks_> damn americans
[18:28:33] <fenn> its those wisconsonian bankers
[18:28:48] <skunkworks_> exactly
[18:29:00] <skunkworks_> giving house loans to every tom dick and harry
[18:29:01] <archivist> a wunch of bankers
[18:29:22] <fenn> archivist: the plural is 'grunch'
[18:29:50] <archivist> swap leading letters, spelling was deliberate
[18:47:59] <skunkworks_> http://www.capsteps.com/sounds/lirty-spitzer.mp3
[18:59:18] <SWPLinux> kinda like their old "lirty dies" stuff it seems
[19:19:07] <fenn> anyone do tapping of pipe threads? is it worth getting an interrupted thread tap?
[19:20:14] <cradek> fenn: I am using pex for a project and I recommend it... it works great
[19:20:31] <cradek> never again the teflon tape
[19:21:33] <skunkworks_> I love pex.
[19:22:44] <skunkworks_> http://www.electronicsam.com/images/house/garinfloor.JPG
[19:22:56] <skunkworks_> It is just like running electrical wire
[19:23:01] <cradek> skunkworks_: I didn't believe it would not leak, but it seems to work great
[19:23:17] <cradek> (but we'll have to wait to see how it's working in 50 years)
[19:26:23] <skunkworks_> Scary huh. The house I built I used pex. the only leaks I had was a few pipe thread connections.
[19:29:45] <tomp2> god i forgot i clicked on that capitol steps link, the laptop came to a crawl, google wouldnt search, file finder was sluggish, dang thats sloooow
[19:29:57] <SWPLinux> but it's funny!
[19:30:28] <tomp2> it was , i'll listen someday, its like the 'sugly blister' fairy tale
[19:30:52] <SWPLinux> they do a segment at basically every show which they call "lirty dies"
[19:31:27] <tomp2> didnt know they were still around, I'll go look for firesign theatre on lunch now
[19:31:52] <SWPLinux> they have a few shows on NPR every year
[19:32:13] <SWPLinux> April 1, some Christmas or other holiday thing
[19:32:38] <tomp2> go<sh>, they have podcasts
[19:32:42] <SWPLinux> I saw them when they were in Burlington around 10 years ago
[19:33:27] <SWPLinux> (just as Ricky Martins second album was doing poorly. I only know because they sang "Livin' La Vida Broke-a" :) )
[19:36:51] <tomp2> :) find 'livin large in my malibu' (times are tight, rent is due, livin large in my malibu') should be an mp3 on some torrent site
[19:42:40] <fenn> pex is like a pvc pipe replacement? i'm trying to tap air line fittings in aluminum castings
[19:44:34] <tomp2> yeah, i was thinking pex wont overcome water lines in a mold, but whats an interrupted thread? aka 'skip tooth'?
[19:45:17] <tomp2> skip tooth is handy if you DONT have pipe tap reamers
[19:45:37] <fenn> yes skip tooth
[19:46:03] <fenn> pipe tap reamer makes a tapered hole first?
[19:46:26] <tomp2> yes, and the tapping is easy, without, it's hell, too much stock
[19:46:46] <tomp2> find one & try it, the decide how to spend money
[19:46:53] <tomp2> then decide
[19:47:08] <fenn> uh.. if i knew anyone with these kinds of toys i probably wouldnt be buying anything
[19:47:59] <skunkworks_> heh
[19:48:41] <fenn> well, i can always make a reamer
[19:49:15] <christel> [Global Notice] Hi all! As you are aware we're preparing to move to a new Services package -- thanks to the hard work from the Atheme guys, in particular jilles (and nenolod) we are now ready for you all to come help us make sure they're approaching ready for production! Please join us over at the testnet. Connect to irc.freenode.net on port 9001. Thank you and have a great day!
[19:52:06] <tomp2> common tool for moldmakers, it'd be hard to make one for 20$ stock & labor see http://www.industrialboys.com/columnar.aspx?cat_id=2450&session_id=827393843&category_site=STARTOOL
[19:53:05] <fenn> its like a d-bit but easier since the diameter doesnt have to be perfect
[19:53:42] <tomp2> oh, and use it with a dead center in a chuck, like in a drill press or BP. do the reaming by hand with oil
[19:54:04] <tomp2> dunno d-bit, and reamers are damn symmetric
[19:54:29] <tomp2> else they dont ream
[19:54:31] <fenn> d-bit is like a simple straight one-flute drill bit
[19:54:45] <tomp2> ok
[19:55:06] <tomp2> oh, pipe tap reamer is umpteen flutes, like at least 8
[19:55:13] <fenn> yeah i dont need all those flutes
[20:41:49] <SWPLinux> jymm: the gamepad works on my 7.10/CVS EMC2. I don't know which is the important part
[20:41:58] <SWPLinux> (OS or hal_input)
[20:43:14] <SWPLinux> the names of the buttons/analog axes are quite different though, so definitely get the basics working before doing the HAL file work
[20:43:31] <awallin> is this ubuntu 8 that was mentioned on the mailing list going to be an lts version, i.e. around for a long time?
[20:46:12] <skunkworks_> Yes
[20:46:15] <alex_joni> awallin: yes
[20:46:46] <awallin> anything useful emc-wise which is new in it?
[20:47:10] <alex_joni> support for new PCs
[20:47:21] <alex_joni> but emc-wise that's about it :)
[20:47:42] <alex_joni> (and the fact that you'll get security fixes for another 3-5 years)
[20:51:28] <jepler> if you don't need any of the new hardware support in 8.04, there's no reason to switch from Dapper before at least June 2009 (when they stop bugfixes for "desktop" software) or maybe 2011 (when they stop releasing any updates or security fixes). On the other hand, I'd keep using a working system in the shop until it died; maybe I'd unplug it from the internet after security fixes are no longer made by canonical
[20:52:08] <awallin> ok. hopefully emc2.3 and beyond won't require the new OS then...?
[20:52:17] <alex_joni> awallin: surely not
[20:52:33] <alex_joni> emc2.2 as it is now should run on quite a variety of OS-es
[20:52:43] <alex_joni> (as long as they are RT-patched linuxes)
[20:53:20] <alex_joni> jepler: one can still have a box behind a firewall, and it should be pretty ok
[20:53:35] <jepler> at some point, we'll stop packaging new emc versions for old distributions -- for instance, if emc 2.4 or 2.5 is released after June 2009, I'd probably see little reason to package it, particularly if it turns out to be difficult
[20:56:22] <alex_joni> otoh we stopped packaging stuff for breezy
[21:32:18] <jymm> SWPLinux Well, figure it out already.
[21:41:43] <SWPLinux> jymm: no way man, that's not my job
[21:41:48] <SWPLinux> I'm soldering ;)
[23:23:13] <JymmmEMC> SWPLinux: Actually, it is... Else you'll be demoing TurboCNC, and not emc =)
[23:24:20] <JymmmEMC> SWPLinux: Can the non-cvs work with it?
[23:25:07] <gene> Question on feed rates, when mixed. I need A to do 720 degrees a minute, while x goes to-.290" at about .007" minute
[23:25:47] <JymmmEMC> gene: The answer is 42
[23:26:03] <gene> Yes Douglas
[23:26:27] <alex_joni> x goes to -.290 is not a distance
[23:26:49] <alex_joni> gene: if you say x goes from 0 to -0.290 .. then it might be solvable
[23:27:41] <gene> yes, its zeroed ATM
[23:29:28] <gene> I can zero A too if the helps
[23:29:36] <gene> I can zero A too if that helps
[23:29:57] <jmkasunich> gene: use inverse time
[23:30:08] <alex_joni> hey jmkasunich
[23:30:14] <jmkasunich> G93
[23:30:22] <jmkasunich> hey alex_joni
[23:31:05] <gene> as in divide the -.290 by the 7 thou a minute?, or vise versa?
[23:31:14] <jmkasunich> http://www.linuxcnc.org/docview/html//gcode_main.html#sub:G93,-G94:-Set
[23:31:38] <jmkasunich> basically, figure out how long you want the move to take, then enter 1.0/that time
[23:33:05] <jmkasunich> if you want to go 0.290 inches at 0.007 a minute, that is 41.42 minutes, so I think you would enter F0.0241 (reciprocal of 41.42)
[23:33:05] <gene> chuckle, right now, but don't have the machinery to do that
[23:33:58] <jmkasunich> if you want A to be doing 720 degrees a minute the whole time, then your A command has to be 41.42 * 720 = 29822 degrees
[23:34:05] <jmkasunich> so something like this should work (I think):
[23:34:14] <jmkasunich> G0X0A0 (initial position)
[23:34:22] <jmkasunich> G93F0.0241
[23:34:31] <jmkasunich> G1X0.290A29822
[23:35:18] <jmkasunich> oops, I didn't read the fine print
[23:35:25] <jmkasunich> the F word needs to be on the G1 line
[23:35:33] <jmkasunich> G0X0A0 (initial position)
[23:35:41] <jmkasunich> G93 (select inverse time mode)
[23:35:59] <jmkasunich> G1X0.290A29822F0.0241 (make move, at specified rate)
[23:36:07] <jmkasunich> G94 (back to regular mode)
[23:36:13] <gene> humm, round figures gave me nearly the same. x=-.290, ahh, now that makes more sense, BRB
[23:38:07] <gene> a too faqst, lost synch quickly
[23:38:13] <gene> a too fast, lost synch quickly
[23:39:50] <jmkasunich> whats too fast? the A axis?
[23:40:07] <alex_joni> isn't 29822 too far away?
[23:40:13] <jmkasunich> I have no clue
[23:40:36] <gene> A is way to fast, stalls in 2 degrees
[23:40:39] <jmkasunich> but I suppose that might be an issue - it is many revolutions of the A axis
[23:40:42] <alex_joni> gene: you're sure you need 720 deg/minute?
[23:40:56] <jmkasunich> 'splain to us what you are trying to do
[23:41:10] <gene> trying to get a smooth cut
[23:41:23] <jmkasunich> on what?
[23:42:24] <jmkasunich> so far, you've described spinning a rotary table 80 some times around while traveling X just over a quarter inch - that's not making a whole lot of sense from here
[23:42:25] <gene> I need to trim the end of a quite hard piece of shafting so there is a .750 diameter stud projecting .540"
[23:43:12] <jmkasunich> no lathe?
[23:43:25] <gene> my lathe is tied up with an unfinished bullet swage die in it, so i can't use it to do this
[23:43:37] <jmkasunich> ok, so you are using the mill as a pseudo-lathe
[23:43:46] <jmkasunich> hence the many revs of the work
[23:43:50] <gene> the general idea, yes
[23:43:58] <jmkasunich> ok, first things first
[23:44:21] <jmkasunich> prove to yourself that your A axis settings are correct - velocity and accel
[23:44:35] <gene> i can do it by hand, standing here hlding the keyboard byttons
[23:44:51] <jmkasunich> can you make A do a rev or two at 720 degrees per minute, using ordinary g-code?
[23:45:05] <jmkasunich> G0A0; G1A720F720
[23:46:06] <alex_joni> I would just start a XYZ config, with A hooked to a siggen
[23:46:15] <alex_joni> and let it spin at the needed speed forever
[23:46:28] <jmkasunich> alex_joni: I might do that too, but that involves HAL hackery
[23:46:54] <jmkasunich> btw, I wouldn't even use siggen - just stepgen in velocity mode
[23:47:29] <alex_joni> yeah, that
[23:47:38] <gene> G1A720F720 spins a about 720 degrees a second i'd guess, inhstant stall
[23:47:59] <jmkasunich> how far does it turn?
[23:48:16] <jmkasunich> (before it stalls)
[23:48:18] <gene> 1/2 degree, maybe
[23:48:31] <jmkasunich> so how do you know its trying to do 720 a second?
[23:48:43] <gene> i don't :)
[23:48:51] <gene> thats called a swag
[23:48:53] <jmkasunich> so why did you say that?
[23:49:24] <jmkasunich> either EMC has a gross bug in A-only feedrates (unlikely), or your A axis is misconfigured
[23:49:30] <gene> You've never heard of a swag, Scientific Wild assed Guess?
[23:49:36] <jmkasunich> of course I have
[23:49:46] <jmkasunich> but I'm trying to solve a problem here - guesses don't help
[23:49:48] <gene> I thought as much :)
[23:50:12] <jmkasunich> try G1A30F30 and see if it will do 30 degrees per minute
[23:50:51] <jmkasunich> make sure you are in G94 mode - if you still have G93 (inverse time) on, then F720 means "complete the move in 1/720th of a minute)
[23:53:07] <SWPLinux> JymmmEMC: I don't know if its the OS difference or the EMC2 difference that makes the difference
[23:53:18] <SWPLinux> you should upgrade EMC2 to the latest to check
[23:53:39] <JymmmEMC> SWPLinux: what is the latest NON- cvs
[23:53:53] <alex_joni> 2.2.4
[23:53:54] <gene> Ok, G1A360F5 will run, and shows 1800 degrees in the velocity
[23:54:07] <jmkasunich> hmm
[23:54:15] <alex_joni> * alex_joni points JymmmEMC at /topic #emc
[23:54:48] <jmkasunich> gene: I'm somewhat flabbergasted
[23:55:24] <jmkasunich> gene: are you talking to us from the EMC PC?
[23:55:33] <alex_joni> 1800 sounds like F30
[23:55:45] <jmkasunich> or he's still in inverse time mode
[23:55:49] <alex_joni> (if the display is in seconds..)
[23:55:56] <gene> yes, heres the axis 3 stanza
[23:55:59] <gene> [AXIS_3]
[23:56:03] <gene> TYPE = ANGULAR
[23:56:07] <gene> HOME = 0.0
[23:56:11] <gene> MAX_VELOCITY = 72.0
[23:56:15] <gene> MAX_ACCELERATION = 1200.0
[23:56:19] <gene> STEPGEN_MAXACCEL = 1260.0
[23:56:23] <gene> SCALE = 160.0
[23:56:27] <gene> FERROR = 1
[23:56:31] <gene> MIN_FERROR = .25
[23:56:32] <gene> MIN_LIMIT = -99999.0
[23:56:32] <gene> MAX_LIMIT = 99999.0
[23:56:32] <gene> HOME_OFFSET = 0.0
[23:56:32] <gene> This was set by stepconf
[23:56:59] <jmkasunich> can your rotary really do 72 degrees per second? and can it really go from zero to top speed in 1/16th of a second?
[23:57:13] <gene> motor is direct on screw, screw is one turn per 10 degrees
[23:57:53] <jmkasunich> did you use stepconf's test mode to actually run it at that speed and accel?
[23:58:09] <gene> 72 pushes it, 55-60 would be a beter setting
[23:58:15] <gene> yes
[23:58:18] <jmkasunich> ok
[23:58:24] <jmkasunich> now back to the g-code
[23:58:29] <jmkasunich> I assume EMC is running
[23:58:34] <gene> yes
[23:58:36] <jmkasunich> are you using AXIS?
[23:58:39] <gene> yes
[23:58:48] <jmkasunich> go to the MDI screen (F5)
[23:58:59] <jmkasunich> what is the A axis at right now?
[23:59:06] <jmkasunich> (on the DRO)
[23:59:06] <gene> its there, just on another screen
[23:59:56] <jmkasunich> what is the A axis at right now?