#emc | Logs for 2007-07-26

[00:00:35] <Ziegler> 2.22
[00:00:50] <Ziegler> in vexta language ;-)
[00:01:39] <Ziegler> http://www.orientalmotor.com/products/pdfs/C_VEXTA/StPk29.pdf
[00:01:43] <JymmmEMC> so 59.4 oz in to lb ft is?
[00:01:45] <Ziegler> JymmmEMC: look at that pdf
[00:02:19] <JymmmEMC> Ziegler: NEMA 34?
[00:02:20] <SWPadnos> 16 oz/lb * 12 in/ft = 192 in-oz / ft-lb
[00:02:58] <Ziegler> thought we went over this last night that 23 was going to be difficult to find in your sepcifications
[00:02:59] <SWPadnos> those ratings are in-lb though, not ft-lb
[00:03:08] <SWPadnos> http://catalog.orientalmotor.com/item/stepping-motors--1068/pk-series-stepping-motors/pv264-d2-8aa?&plpver=11&origin=advsrch&by=prod&filter=0
[00:03:24] <Ziegler> I found a couple
[00:03:30] <SWPadnos> 9.4 in-lb, NEMA 23(or as close as I can tell at the vexta site)
[00:03:32] <Ziegler> but its pushing it
[00:03:51] <Ziegler> 59.4/16 = 3.7
[00:05:10] <JymmmEMC> http://catalog.orientalmotor.com/item/all-categories/ac-motor-accessories/pal2p-2?cid=1002&plpver=11&prodid=3001035&itemid=12232&backtoname=Item+%23+PV264-D2.8AA&pane=sb&bc=100|3001048x#
[00:05:50] <JymmmEMC> All these UOM got me lost
[00:06:12] <Ziegler> mounting bracket?
[00:06:42] <JymmmEMC> Ziegler: question?
[00:07:16] <Ziegler> I just thought you were looking for a stepper ;-)
[00:07:42] <JymmmEMC> Ziegler: Yes, a NEMA23 motor, as my machine doens't have nema34 mounting BRACKETS
[00:07:55] <JymmmEMC> hint hint nudge nudge
[00:08:52] <Ziegler> nema 23 may be difficult... large would make it cheaper and easier to find a stepper to your sepcification
[00:08:57] <SWPadnos> is there a belt tension adjustment?
[00:09:05] <SWPadnos> (ie, are the nema 23 bolt holes slots)
[00:09:10] <Ziegler> http://www.orientalmotor.com/products/pdfs/C_VEXTA/StPk26.pdf
[00:09:17] <Ziegler> there you go JymmmEMC
[00:09:18] <JymmmEMC> SWPadnos: did you see the pics I sent you earlier?
[00:09:36] <SWPadnos> yes, but I don't remember them to that level of detaiol
[00:09:41] <SWPadnos> detail
[00:11:13] <JymmmEMC> http://www.k2cnc.com/DuportalPro343Access/images/CNC-Routers-Pully-KG-3925.jpg
[00:11:32] <JymmmEMC> http://www.k2cnc.com/DuportalPro343Access/images/CNC-Routers-KG-3925%205-Backview.jpg
[00:11:46] <JymmmEMC> Y and X respectively
[00:12:21] <SWPadnos> it looks like the answer is yes for Y, and unknown for X
[00:12:34] <SWPadnos> but I'd assume it's yes there as well
[00:12:50] <JymmmEMC> X has the whole rail to slide on from center to edge
[00:12:58] <SWPadnos> what that means is that a NEMA34 bracket should be trivial to make
[00:13:04] <JymmmEMC> but may not have the height clearance
[00:13:20] <JymmmEMC> OR.... buy this one for $17
[00:13:26] <SWPadnos> what kind of belts are those?
[00:13:37] <JymmmEMC> toothed?
[00:13:40] <SWPadnos> duh :)
[00:13:49] <SWPadnos> what tooth pitch?
[00:14:00] <JymmmEMC> no clue
[00:14:00] <SWPadnos> that looks like a very tiny pulley on the Y motor
[00:14:20] <SWPadnos> if that motor is NEMA23, the pulley would have to be ~0.5 inches across
[00:14:32] <JymmmEMC> 1" and 2" (12 tooth and 24 tooth)
[00:14:44] <JymmmEMC> the y pic is nema34 servo motor
[00:14:51] <SWPadnos> ah
[00:21:08] <JymmmEMC> SWPadnos: when you said that half step in the torque chart, so 1/4 step would be one half the torque at given speed?
[00:21:42] <JymmmEMC> ^listed at given speed?
[00:21:43] <Ziegler> narrow down a stepper and driver yet?
[00:22:22] <SWPadnos> ?
[00:22:39] <SWPadnos> microstepping doesn't affect torque, only smoothness
[00:22:49] <JymmmEMC> 16:34:35) SWPadnos: the 2000 pps half step = 1000 pps full-step
[00:23:03] <SWPadnos> the problem I saw was that 2000 pulses/second at half-stepping is equivalent to 1000 pulses/second at full-stepping
[00:23:25] <SWPadnos> so the speeds in the chart look like they're twice what they really are
[00:23:26] <JymmmEMC> SWPadnos: and what is the problem?
[00:23:30] <SWPadnos> so the speeds in the chart look like they're twice what they really are
[00:23:33] <JymmmEMC> ah
[00:23:38] <Ziegler> hehe
[00:23:39] <JymmmEMC> snake oil
[00:23:45] <SWPadnos> yep
[00:24:40] <JymmmEMC> You know what I'm gonna do...
[00:25:20] <SWPadnos> get AC servos and drives?
[00:25:26] <SWPadnos> and a stiffer machine?
[00:25:38] <SWPadnos> and a new touch panel PC for the controller?
[00:25:47] <JymmmEMC> Today is my friday, tomorrow I'm going to toss these applied motion steppers I have P/N: 4023-828D on my machien and see what happens
[00:26:24] <SWPadnos> just for the heck of it, you should make sure you've got the right current setting for the motors you have on the machine
[00:26:42] <JymmmEMC> SWPadnos: already did a long time ago.
[00:26:46] <SWPadnos> ok
[00:27:00] <JymmmEMC> even bought 1% resistors too
[00:27:11] <SWPadnos> have you been doing this testing using the parker drives?
[00:27:20] <JymmmEMC> oh gawd yes
[00:27:21] <SWPadnos> ok
[00:27:34] <SWPadnos> I didn't notice, but I assume they're chopper drives ...
[00:27:38] <JymmmEMC> http://www.applied-motion.com/products/stepper/motors/size23.php
[00:28:23] <JymmmEMC> SWPadnos: no clue
[00:29:05] <SWPadnos> ok. the reason I ask is that chopper drives tend to shift the speed curve to the right with higher voltage, but non-chopper drives don't (I think)
[00:30:30] <Ziegler> you running bipolar or unipolar?
[00:31:05] <Ziegler> series or parallel?
[00:34:28] <JymmmEMC> bi 4 wire
[00:36:58] <JymmmEMC> holey shit
[00:37:10] <JymmmEMC> torque curve for these motors is almost flat
[00:38:06] <JymmmEMC> http://pdf.mouser.com/library/pdf/appliedmotioncatalog.pdf
[00:38:15] <JymmmEMC> pdg pg 129
[00:38:21] <JymmmEMC> PDF pg 129
[00:38:33] <JymmmEMC> bottom left corner
[00:39:44] <JymmmEMC> You know, it's really fscked up when you have to go to a mfg's distributor to get literature because mfg's website has disabled search
[00:40:02] <SWPadnos> it's also screwed up when you can't load a link, but have to click it from the site
[00:40:07] <SWPadnos> like the catalog
[00:40:27] <JymmmEMC> that pdf link didn't come thru?
[00:40:41] <SWPadnos> nope
[00:40:51] <SWPadnos> I had to get to it from Mouser's applied motion "storefront"
[00:41:14] <JymmmEMC> at least they had it, when AM didn't at all
[00:41:15] <Ziegler> link worked for me
[00:41:19] <JymmmEMC> saved locally now.
[00:41:49] <Ziegler> never mind
[00:42:00] <JymmmEMC> Hey, that's not a bad idea... a library of whatever we can find of specs for drives motors etc.
[00:42:17] <SWPadnos> just check the copyrights ...
[00:42:44] <JymmmEMC> fsck the (c)... what they gonna sue me for, turnupts I dont have?
[00:42:56] <SWPadnos> you can host the library then :)
[00:43:16] <SWPadnos> that is a surprising torque curve
[00:43:23] <JymmmEMC> Hell, I'll sell it for $9 - the cost of duplicating the DVD and shipping.
[00:43:26] <SWPadnos> I wonder if "dynamic torque" means something I don't know
[00:43:39] <SWPadnos> that's great for GPL items
[00:43:49] <Ziegler> changes with speed?
[00:43:53] <JymmmEMC> At least I have a REAL (not snake oil) torque curve to go by
[00:44:06] <JymmmEMC> it does say 80V though
[00:44:12] <Ziegler> oriental?
[00:44:34] <SWPadnos> higher voltage increases the corner speed with chopper drives
[00:44:46] <Ziegler> might do yourself a favor and get a driver specifically matched to go with their steppers if you go with them
[00:45:06] <JymmmEMC> SWPadnos: Well, I can do 48VDC * 2 -OR- 49VDC and 11A
[00:45:27] <JymmmEMC> s/49/48/
[00:45:43] <SWPadnos> I didn't notice the voltage rating on the steppers - was it something like 2.8V?
[00:45:50] <SWPadnos> or 2.8A, 3.1V
[00:46:06] <JymmmEMC> label on mine says 1.7V 4.7A
[00:46:12] <SWPadnos> ok
[00:46:13] <JymmmEMC> 0.4 ohms
[00:46:17] <JymmmEMC> 200 SPR
[00:46:31] <Ziegler> I have 3V 3A
[00:46:32] <SWPadnos> the applied motion ones, or the ones on the machine?
[00:46:41] <SWPadnos> I have 150V 10A :)
[00:46:48] <Ziegler> woah
[00:46:49] <SWPadnos> (peak at 39A)
[00:46:50] <JymmmEMC> Ziegler: OEM750 (not OEM750X) is what I have: http://www.compumotor.com/literature/pg170_oem750_series.htm
[00:46:51] <SWPadnos> servos
[00:46:51] <Ziegler> what do you drive them with?
[00:46:54] <Ziegler> oh
[00:46:55] <Ziegler> duhg
[00:47:03] <JymmmEMC> SWPadnos: the ones I got with the oem750's
[00:47:08] <SWPadnos> ok
[00:47:14] <JymmmEMC> SWPadnos: not the POS on my machine
[00:48:07] <JymmmEMC> SWPadnos: and us digital encoders on them as well E2-500-250-IH
[00:48:19] <SWPadnos> cool
[00:48:27] <Ziegler> cool
[00:49:27] <JymmmEMC> at least now I can have real feedback and not have to listen for stalled steps
[00:50:23] <Ziegler> in my case my machine 1 doesnt have limit switches... so I got to watch my machine crash
[00:50:52] <Ziegler> only real damage was a snapped bit
[00:50:56] <JymmmEMC> Ziegler: The drives I have are nice, not fly by night
[00:51:13] <JymmmEMC> Ziegler: => geckos
[00:51:32] <Ziegler> wait... you have geckos?
[00:51:44] <Ziegler> nice
[00:51:55] <JymmmEMC> Ziegler: GAWD..... 17:46:49) JymmmEMC: Ziegler: OEM750 (not OEM750X) is what I have: http://www.compumotor.com/literature/pg170_oem750_series.htm
[00:52:05] <Ziegler> guess what I was getting at, those curves may only be applicable using oriental drivers
[00:52:15] <Ziegler> I saw that
[00:52:18] <JymmmEMC> you're killing me in repeating myself!!!
[00:52:32] <Ziegler> didnt look like geckos website to me
[00:52:37] <JymmmEMC> it was just 10 minutes ago I said that
[00:52:48] <SWPadnos> what was the gecko comment about though?
[00:53:01] <JymmmEMC> SWPadnos: comparison of parker and geckos
[00:53:16] <Ziegler> not my fault you wernt clear
[00:53:22] <JymmmEMC> 17:50:57) JymmmEMC: Ziegler: The drives I have are nice, not fly by night
[00:53:20] <JymmmEMC> (17:51:15) JymmmEMC: Ziegler: => geckos
[00:53:23] <SWPadnos> well, you said "not fly ny night", then mentioned geckos, which are not fly by night ...
[00:53:30] <SWPadnos> s/ny/by/
[00:53:56] <SWPadnos> ok, it looks like an arrow though :)
[00:54:18] <Ziegler> besides JymmmEMC this whole "so tired of repeating myself" thing... you might look at the logs and see how often some have done that for you.
[00:54:18] <SWPadnos> (and I usually do >= for the mathematical symbol :) )
[00:54:42] <JymmmEMC> Ziegler: Sure, but not within 10 minutes (I dount think)
[00:54:51] <Ziegler> you would be suprised
[00:54:51] <SWPadnos> check again
[00:55:01] <JymmmEMC> Unless I need clarification
[00:55:18] <Ziegler> we need it sometimes too ;-P
[00:55:44] <JymmmEMC> you guys have me going to sillions of wesites, when I haven't even had a chance to look/read the first one. I'm not a fast reader
[00:55:45] <Ziegler> I must be a glutton for punishment
[00:56:28] <SWPadnos> you can always ask people to slow down, but remember we're helping you with your machine
[00:56:34] <JymmmEMC> I dont just have the url's to these pics I've been given at my fingertips, I have to look them up,
[00:56:42] <SWPadnos> no problem
[00:56:58] <JymmmEMC> SWPadnos: oh no, I really do appreciate all the help everyone his giving me
[00:57:07] <Ziegler> cool
[00:57:12] <SWPadnos> then quit sounding like an a**hole :)
[00:57:15] <JymmmEMC> Just repeating myself is a huge pet peeve
[00:57:17] <Ziegler> hehe
[00:57:18] <SWPadnos> I mean - you're welcome
[00:57:23] <toast> pew pew
[00:57:28] <JymmmEMC> I'm always an asshole, I don't deny that.
[00:57:30] <Ziegler> LOL
[00:57:32] <SWPadnos> oh, I forgot
[00:57:39] <JymmmEMC> why should now be any different?
[00:57:38] <Ziegler> What is your profession JymmmEMC?
[00:57:50] <SWPadnos> asshole
[00:57:52] <JymmmEMC> Ziegler: agitator
[00:57:54] <SWPadnos> same thing
[00:58:02] <JymmmEMC> what SWPadnos said
[00:58:09] <JymmmEMC> JymmmEMC is now known as MrAsshole
[00:58:16] <Ziegler> agitated a-hole
[00:58:19] <MrAsshole> Ok, the nick is even registered, alright?
[00:58:30] <MrAsshole> been registered for a VERY long time
[00:59:10] <MrAsshole> Ziegler: A burned out SysAdmin for the most part.
[00:59:16] <Ziegler> ahhhhhhh
[00:59:33] <SWPadnos> think BOFH
[00:59:39] <MrAsshole> Ziegler: A SERIOUSLY burned out SysAdmin for the most part.
[00:59:49] <MrAsshole> SWPadnos: No, that's WAY TO KIND
[00:59:54] <toast> *too
[01:00:19] <MrAsshole> wait, I take that back... you havne't met Sam at work... now HE is BOFH to the extreme
[01:00:24] <Ziegler> States or across the pond?
[01:00:34] <Ziegler> BOFH?
[01:00:42] <MrAsshole> Ziegler: Silicon Valley
[01:00:42] <SWPadnos> Bastard Operator From Hell
[01:00:45] <Ziegler> LOL
[01:01:07] <MrAsshole> Ziegler: http://en.wikipedia.org/wiki/BOFH
[01:01:20] <MrAsshole> Ziegler: but in RL too
[01:01:58] <MrAsshole> http://members.iinet.net.au/~bofh/
[01:02:08] <MrAsshole> ok, back in 90
[01:02:35] <Ziegler> I vaguely remember reading some of that way back.
[01:17:02] <toast> there is a big taper in one of the mills at work
[01:17:09] <toast> and i don't know where it's coming from
[01:17:21] <toast> the vise is level, and so is the head
[01:17:35] <toast> it's like 4 thou over 10"
[01:50:57] <maddash> omg omg omg --
[01:51:05] <maddash> RTH| lat min| ovl min| lat avg| lat max| ovl max| overruns
[01:51:05] <maddash> RTD| -350| -365| -94| 1247| 1742| 0
[01:51:27] <maddash> and this is while I'm compiling a kernel on my dual core system
[01:53:43] <maddash> brb let me see what happens when I get rid of 'isolcpus=0'
[01:53:52] <skunkworks> my quantum computer doesn't have any latency.. it is actually in the negative numbers.
[01:54:18] <maddash> skunkworks: jealous, huh?
[01:54:20] <maddash> :P
[01:54:21] <skunkworks> I would show it to you - but once you see it - it stops working.
[01:54:34] <toast> i hear it runs on cats
[01:54:43] <skunkworks> only dead ones. maybe
[01:54:51] <toast> we can't know for sure.
[01:54:56] <skunkworks> right
[01:55:38] <skunkworks> and that concludes what I know about quantum mechanics.
[01:57:28] <toast> lol
[01:58:53] <SWPadnos> maddash, what chipset is in that fine RT computer?
[01:58:57] <maddash> RTH| lat min| ovl min| lat avg| lat max| ovl max| overruns
[01:58:57] <maddash> RTD| -500| -560| 13| 1370| 2960| 0
[01:58:57] <maddash> RTD| -337| -560| 29| 1598| 2960| 0
[01:59:12] <maddash> ^^ that's w/o isolcpus=0
[01:59:21] <SWPadnos> excellent in either case
[02:00:08] <SWPadnos> I'm definitely interested in knowing the CPU and chipset (and maybe manufacturer and model number)
[02:00:28] <maddash> SWPadnos, it's this little beast: http://www.asus.com/products4.aspx?l1=3&l2=11&l3=0&model=1276&modelmenu=1
[02:00:53] <SWPadnos> ok - 865G + ICH5
[02:01:10] <SWPadnos> that's what one of the embedded computers I found is based on, I think
[02:01:09] <maddash> SWPadnos, I feel so ashamed of myself for bugging my dad non-stop to buy another board w/o an onboard graphics chip
[02:01:13] <SWPadnos> what CPU speed?
[02:01:17] <SWPadnos> heh
[02:01:27] <maddash> SWPadnos, 2.8GHz Pentium D
[02:01:36] <SWPadnos> ok, is that the 805 or something?
[02:01:41] <maddash> yep.
[02:02:05] <maddash> I think SWP can do for RT
[02:02:09] <maddash> er, SMP*
[02:02:15] <SWPadnos> me too ;)
[02:02:17] <maddash> not that SWP wouldn't...
[02:03:08] <Skullworks-PGAB> SMP on a dual Piii-600 - would tht be a waste of time?
[02:03:34] <SWPadnos> I don't think so, depending on (a) the workload and (b) your tolerance for high power bills
[02:03:47] <maddash> I think cradek has one similar to that -- he claims to have latency numbers no higher than 5000, I think
[02:04:12] <Skullworks-PGAB> * Skullworks-PGAB heats his home with SCSI RAIDS in winter
[02:04:23] <SWPadnos> ewwww - electric heat
[02:04:29] <SWPadnos> that's so '70s
[02:05:17] <SWPadnos> I wonder if the 965G will be as good as the 865
[02:05:22] <Skullworks-PGAB> yeah - so are the 5.25: Full height Elete server drives...
[02:05:23] <SWPadnos> I guess I'll have to wait and find out
[02:05:35] <SWPadnos> ST4096, baby!
[02:05:45] <SWPadnos> 5.25" full height 80 meg
[02:05:52] <Skullworks-PGAB> not THAT old - but close
[02:05:54] <SWPadnos> 120 if you use an RLL controller
[02:06:07] <SWPadnos> only 15 pounds each or thereabouts
[02:06:14] <Skullworks-PGAB> Fast/Wide SCSI II
[02:06:21] <SWPadnos> ooooohhhhhhh
[02:06:59] <SWPadnos> I have about 12 3.5" fast/wide or ultra-SCSI 4.{3,5} gig drives here
[02:07:13] <SWPadnos> Micropolis, if you can believe it
[02:07:27] <maddash> * maddash has an RT orgasm
[02:07:35] <SWPadnos> TMI
[02:07:53] <Skullworks-PGAB> still use the old RAID in winter - DIVX conversions
[02:08:16] <maddash> Skullworks-PGAB, no gas bill for you, eh?
[02:08:35] <Skullworks-PGAB> sure theres a gas bill
[02:08:44] <SWPadnos> well, I think I'm actually going to do something other than IRC before going to bed. see you all later
[02:08:53] <Skullworks-PGAB> this house was built without any insolation
[02:09:09] <maddash> SWPadnos, ditto. adios.
[02:13:22] <maddash> * maddash forgot a few things
[02:13:38] <maddash> meh, nvm
[02:42:36] <Ziegler> night
[02:47:27] <a-l-p-h-a_> wow. fat nation.
[02:52:00] <a-l-p-h-a_> eight 80 meg
[02:52:00] <a-l-p-h-a_> [22:05:24] <Skullworks-PGAB> not THAT old - but close
[02:52:00] <a-l-p-h-a_> [22:05:28] <SWPadnos> 120 if you use an RLL controller
[02:52:00] <a-l-p-h-a_> [22:05:41] <SWPadnos> only 15 pounds each or thereabouts
[02:52:01] <a-l-p-h-a_> [22:05:48] <Skullworks-PGAB> Fast/Wide SCSI II
[02:52:01] <a-l-p-h-a_> [22:05:53] <SWPadnos> ooooohhhhhhh
[02:52:03] <a-l-p-h-a_> [22:06:33] <SWPadnos> I have about 12 3.5" fast/wide or ultra-SCSI 4.{3,5} gig drives here
[02:52:04] <a-l-p-h-a_> [22:06:47] <SWPadnos> Micropolis, if you can believe it
[02:52:06] <a-l-p-h-a_> [22:07:00] * maddash has an RT orgasm
[02:52:08] <a-l-p-h-a_> [22:07:08] <SWPadnos> TMI
[02:52:10] <a-l-p-h-a_> oops
[02:59:12] <a-l-p-h-a_> http://edition.cnn.com/SPECIALS/2007/fit.nation/obesity.map/ that's what I was talking about
[03:00:02] <Skullworks-PGAB> Excuses - I know you meant FAT tables...
[03:10:14] <Jymmm> better than a virtual one
[04:10:19] <ds2> wow... a seig mini mill only has 4 inches of Y? learn something new everyday
[05:00:40] <Unit41> beware of 0day firefox exploit
[05:00:47] <Unit41> windows only
[05:28:02] <Jymmm> url?
[06:29:01] <Unit41> oops
[06:29:05] <Unit41> forgot about irc for a min
[06:29:22] <Unit41> www.xs-sniper.com
[06:38:49] <Jymmm> lovely
[06:40:13] <Unit41> its not really that bad though
[06:40:23] <Unit41> cause even with js enabled it asks if you want to launch the app
[06:40:33] <Jymmm> Unit41: actually it is
[06:41:26] <Jymmm> the demo links are non evasive, but a batch file cna do anything on the system
[06:45:51] <Unit41> yeah but to run it, you would need to click the launch application button
[06:46:12] <Unit41> firefox has somewhat built in protection
[06:46:36] <Unit41> might be able to fool some people I guess
[07:00:38] <Jymmm> Actually no, I just tried it. in the FF
[07:45:25] <Jymmm> SWPadnos: ping
[07:46:02] <toast> lul
[07:53:40] <alex_joni> Jymmm: pong
[07:54:06] <Jymmm> alex_joni: MITM Attack
[07:54:47] <dimas> Jymmm: :)
[07:57:41] <Jymmm> Generla question... I hit a lot of surplus stores, and see a lot of 5-15 lbs xmfrs of reasonable size, some have voltages on their labels, others don't but not a big big deal there. The thing is usually most don't have a VA rating of any kind. any way of figuring that out?
[08:01:50] <lerman_> lerman_ is now known as lerman
[08:16:54] <archivist> Jymmm mainly by testing
[08:17:27] <archivist> looking for temperature rise, regulation
[08:20:33] <Jymmm> Doens't sound worth it
[08:21:22] <archivist> and looking at wire size at the tags
[08:23:34] <Jymmm> ok, too questionable for *ME*
[08:30:06] <archivist> last one I wanted sized I took to a local transformer company and said "How big is it?"
[08:30:21] <archivist> it was 30 years ago
[08:30:37] <archivist> at least
[08:33:08] <Jymmm> well, ppl knew stuff back then.
[08:34:03] <archivist> he died 25 years ago and then the battle axe wife ran the co till a few years ago
[08:34:31] <archivist> daughter then ran it a few years and shut it
[08:35:04] <archivist> a lot of transformer companies are shutting down and the work going to china
[08:35:13] <Jymmm> sad
[08:35:34] <Jymmm> Ther eIS one here in calif I can get custom made in qty =1 too
[08:35:48] <archivist> we have some of the winding machines from there
[08:37:29] <toast> does anyone think it would be a dumb idea to try and learn scraping in an apartment
[08:37:45] <Jymmm> Got 100HP shop vac?
[08:37:51] <Jymmm> and no carpet
[08:37:58] <toast> scraping doesn't make that big a mess, dude
[08:38:36] <Jymmm> If you think lil pieces of metal are not going to go flying, you need to step away from the crack pipe
[08:38:43] <toast> uh, from scraping?
[08:38:49] <toast> have you seen scraping
[08:39:02] <toast> it's not exactly high speed machining
[08:39:07] <archivist> toast go ahead and try it
[08:39:18] <toast> i'll need a better table
[08:39:22] <lerneaen_hydra> take a picture of the results :D
[08:39:26] <toast> i don't think this one from ikea will hold up
[08:39:26] <lerneaen_hydra> before you clean
[08:39:29] <toast> maybe next semester
[08:39:33] <toast> in school
[08:39:49] <archivist> scrape a bearing
[08:39:59] <toast> i was thinking just a little chunk of cast iron
[08:40:18] <archivist> 3 chunks to get a flat
[08:40:30] <toast> i won't try autogeneration on my first try, dude
[08:40:37] <archivist> hehe
[08:40:38] <toast> i'd probably buy a little AA grade surface plate
[08:40:40] <toast> and just use that
[08:43:35] <toast_> arrgh
[08:59:31] <archivist> I had to generate a .2 module internal gear form a couple of weeks a ago
[09:01:47] <SWPadnos_> SWPadnos_ is now known as SWPadnos
[09:14:53] <Jymmm> SWPadnos: !!!
[11:28:58] <maddash> http://pastebin.ca/634340
[11:29:35] <maddash> ^^ is a patch for src/emc/rs274ngc
[11:29:49] <maddash> /interp_convert.cc
[11:33:35] <alex_joni> maddash: what for?
[11:33:41] <maddash> alex_joni: 'safety'
[11:33:44] <lerneaen_hydra> maddash: fixed your drivers?
[11:33:59] <alex_joni> that's not a safety feature
[11:34:34] <maddash> alex_joni: because on this bridgeport, the Z-axis is always the one carrying the spindle, so it's more 'safe' if the controller program lifted the tool up before traversing in the XY-plane
[11:34:58] <alex_joni> how about 90% of all other machines? which aren't bridgeport?
[11:35:07] <lerneaen_hydra> maddash: what if the tool is a t slot cutter
[11:35:39] <lerneaen_hydra> maddash: that's not something the program should do automatically, that's something that is supposed to be written in the gcode file
[11:35:40] <maddash> lerneaen_hydra: no, I think I pissed the gods off some more -- when i retried "g0x5y5 g0x0y0", after I'd signed off, there was no grinding noise
[11:35:56] <lerneaen_hydra> hmm
[11:36:03] <lerneaen_hydra> even more spooky
[11:36:20] <lerneaen_hydra> I thought of something right after you left
[11:36:29] <lerneaen_hydra> you said you were using the old 1970's stepper drivers
[11:36:37] <maddash> lerneaen_hydra: but "g0x5y5z-5 g0x0y0z0" did a grind
[11:36:44] <lerneaen_hydra> what if they don't like being driven faster than a certain speed?
[11:37:01] <lerneaen_hydra> are the axes seperatly driven?
[11:37:15] <lerneaen_hydra> or is the driver a 3-in-one thing?
[11:37:16] <archivist> the grind sounded like a rattle/resonance to me
[11:37:50] <maddash> alex_joni: the patch was meant to be an illustration, not merged into emc2's official source (if you did, I'd be horrified)
[11:37:54] <lerneaen_hydra> it's still something that the drivers could do, if they pulse current in a certain way
[11:38:03] <maddash> lerneaen_hydra: seperately
[11:38:05] <alex_joni> maddash: no, I didn't
[11:38:14] <alex_joni> maddash: just wasn't sure what you want with it..
[11:38:20] <lerneaen_hydra> maddash: hmm, can't be that then...
[11:38:25] <alex_joni> it seems to me it's easier to writer proper g-code
[11:38:31] <maddash> alex_joni: ditto.
[11:39:04] <lerneaen_hydra> IMO it's a very large safety risk
[11:39:08] <lerneaen_hydra> more than practial/proper
[11:39:24] <lerneaen_hydra> as it's not explicit what the machine will do
[11:40:25] <maddash> how is lifting the tool a risk (on my setup, that is)?
[11:41:10] <fenn> gets you into a bad habit
[11:41:43] <maddash> ...and why can't/shouldn't a t-slot cutter do a g0z0?
[11:42:01] <fenn> it will rip your part to bits
[11:42:23] <maddash> * maddash is trying to dig up reasons to avoid rewriting code
[11:42:35] <fenn> say you cut the slot and want to g0 back to the entrance to pull out of the slot
[11:42:36] <lerneaen_hydra> what if the mill is a horizontal mill
[11:42:46] <lerneaen_hydra> what if it's a 5 axis mill
[11:42:57] <lerneaen_hydra> what if you're doing internal cavity work
[11:42:59] <lerneaen_hydra> etc etc
[11:43:15] <lerneaen_hydra> what is a good amount of safe offset
[11:43:18] <lerneaen_hydra> how can it be changed
[11:43:23] <maddash> fenn: hm, is it proper to zip around the part using g0 ? I always thought that g0 was used when you were done cutting...
[11:43:23] <lerneaen_hydra> and so on and so forth
[11:45:15] <fenn> g0 is when you arent cutting, afaik
[11:46:36] <fenn> in my t-slot example, would you feed back to the beginning instead?
[11:46:55] <maddash> hm, sure, but only after you'd come out the other end of the cut
[11:47:02] <lerneaen_hydra> me? depending on the material I would either g0 or feed out
[11:47:04] <fenn> what if there is no other end
[11:48:23] <maddash> well, being that it's my dad who's insisting on this, he'd probably carefully retrace the entire path backwards w/non-g0 codes
[11:49:00] <maddash> g1x3y3 f120 means that each of the x and y axes are moving @ 120/sqrt(2) inch/min, right? brb.
[11:49:13] <lerneaen_hydra> yes
[11:49:39] <lerneaen_hydra> the tool tip is moving at 120mm/min relative to the table
[11:49:54] <fenn> i think lerneaen_hydra meant no
[11:50:04] <fenn> F is the tool velocity
[11:50:04] <maddash> huh? yes or no?
[11:50:17] <fenn> er, sorry, nevermind
[11:51:36] <fenn> i never understood why physics teachers make a big deal about "speed" and "velocity"
[11:52:25] <maddash> well, speed = abs(vel)
[11:52:52] <fenn> yes but why not just say abs(vel)
[11:52:58] <maddash> it's a big deal because most of the time, the motion is regarded in a >=2-dimensional case
[11:53:27] <fenn> and in that case your +/- number velocity vector representation falls apart anyway
[11:53:56] <maddash> fenn: why not say "thin version of the big computer that you can carry around"?
[11:54:18] <fenn> because you might get it confused with the rack mount version
[11:54:40] <maddash> you can't carry a rack mount version around...
[11:54:45] <fenn> sure you can
[11:55:00] <maddash> 'carry' is defined as being able to use it wherever you go
[11:55:02] <fenn> how do you think they get it on the rack!
[12:39:50] <jepler> rs274ngc specifies that a G0 move is a coordinated linear move from the initial location to the final location -- that's unlikely to ever change
[12:40:41] <maddash> g0 is a traversal move, so how can it be coordinated? isn't that g1's job?
[12:41:23] <jepler> "For rapid linear motion, program G0 X- Y- Z- A- B- C-, where all the axis words are optional, except that at least one must be used. The G0 is optional if the current motion mode is G0. This will produce coordinated linear motion to the destination point at the current traverse rate (or slower if the machine will not go that fast). It is expected that cutting will not take place when a G0 command is executing.
[12:41:25] <lerneaen_hydra> g0 is the same as g1 except that it moves at the maximum speed
[12:41:29] <jepler> "
[12:42:01] <lerneaen_hydra> you could cut with g0 if you felt like it and the machine/tool can take the speed
[12:42:07] <maddash> i don't know how coordinated is used here,
[12:42:31] <lerneaen_hydra> all axes arrive at the end destination simultaneously
[12:42:41] <maddash> lerneaen_hydra: not with g0
[12:42:42] <jepler> maddash: "coordinated linear move" that means that all the points that the tool tip passes through are as near as possible to the ideal line from the initial point to the final point
[12:42:52] <lerneaen_hydra> maddash: yes they are :p
[12:44:31] <maddash> assuming that the x and y axes were identical in every way, g0x2y4 != g1x2y4f[maxvel] in that the g0 move would bring the x to a stop before the y axis
[12:44:56] <jepler> maddash: no, that's simply not true in emc's gcode
[12:44:57] <lerneaen_hydra> um, no, they're identical
[12:45:02] <lerneaen_hydra> there are some systems where that's true
[12:45:11] <lerneaen_hydra> not for emc
[12:45:13] <jepler> that's one behavior that "coordinated" excludes
[12:45:30] <maddash> jepler: that's not how it's supposed to be done
[12:45:49] <maddash> g0 means to go as fast as you can on each axis
[12:45:54] <lerneaen_hydra> most of the commercial NC manufacturers Ive seen do it the EMC way
[12:46:07] <jepler> no, G0 doesn't mean anything except when you refer to the documentation for your machine control.
[12:46:19] <lerneaen_hydra> you won't arrive at the destination faster, Y is your bottleneck regardless of how X moves
[12:46:21] <jepler> then you know what it means
[12:46:51] <maddash> this is some odd shit, then
[12:47:14] <lerneaen_hydra> afaik newer heidenhain likes to go it the non-coord way, older heidenhain is coordinated, seimens is coordinated
[12:47:27] <maddash> because (not to be snobby or the like) bridgeport is virtually synonymous with milling, and this mill does the g0 moves exactly as I specified
[12:47:37] <jepler> lerneaen_hydra gives one good reason that the other common behavior (move axes in non-coordinated fashion, each as fast as possible) doesn't provide a benefit in program execution time
[12:47:51] <maddash> is this some new thing that popped up between rs274 and the ngc version?
[12:48:27] <jepler> since you have the source code, you can of course modify G0 to do whatever you like
[12:49:06] <lerneaen_hydra> IMO there's only bad things from a non-coordnated move, it takes exactly as long to do, and you can end up with situations where you forget that there's something else in the way when doing a non-coord move, whereas a coord move is much easier to see the path
[12:49:11] <maddash> duh, but if this 'coordinated' definition of g0 holds up, i'll have some funny stuff to toss to my dad
[12:49:27] <jepler> if you are *authoring* gcode, you might do well to keep in mind that not all systems act like emc, for instance not assume that since the ideal path between two points doesn't hit remaining material that a G0 "rapid" between them might go ahead and hit the material when run on some other control
[12:49:28] <maddash> lerneaen_hydra: no doubt
[12:49:31] <lerneaen_hydra> additionally different machines with different axis speeds would move differently with the same program!
[12:49:43] <lerneaen_hydra> *very* bad IMO
[12:50:37] <lerneaen_hydra> a machine with a slow Y axis might make a part just fine, whereas with a faster Y axis the tool might crash because it has a different tool path or vice versa
[12:50:52] <jepler> (just like anyone authoring gcode on maddash's system with the "G0 goes to final Z first" patch would do well to keep in mind that other systems will do other things instead)
[12:51:23] <lerneaen_hydra> exactly
[12:51:41] <lerneaen_hydra> it takes one more line of code to do Z first (not even that, just a CR is needed)
[12:52:18] <maddash> jepler: actually, if you try out my patch, it only moves the Z axis partially before starting the other axes.
[12:52:29] <maddash> which is kind of weird
[12:53:30] <lerneaen_hydra> maddash: I don't see the point in moving them non-coord
[12:54:04] <maddash> 07:34:10 <maddash> alex_joni: because on this bridgeport, the Z-axis is always the one carrying the spindle, so it's more 'safe' if the controller program lifted the tool up before traversing in the XY-plane
[12:54:24] <maddash> to avoid collisions between the tool and material
[12:54:47] <maddash> I personally think that this is a feature to implement in the CAM program, not the motion controller or interpreter
[12:54:49] <lerneaen_hydra> I don't like that behavior, it's far too assuming and implicit
[12:54:56] <lerneaen_hydra> yes, exactly!
[12:55:03] <lerneaen_hydra> it's not something for the controller
[12:55:17] <lerneaen_hydra> the controller doesn't have a clue as to where there's material or not
[12:55:29] <lerneaen_hydra> it's up to the coder/cam app to take care of that
[13:13:17] <archivist> specially if the head was rotated
[13:13:41] <alex_joni> or if it's a robot
[13:19:32] <maddash> * maddash wants to rotate someone's head
[13:20:13] <maddash> how appalling -- the interpreter itself uses python!
[13:20:36] <lerneaen_hydra> isn't it done in C?
[13:20:57] <alex_joni> it's actually C++
[13:21:12] <lerneaen_hydra> how often does it run? once per servo period?
[13:21:18] <alex_joni> no, it's not RT
[13:21:23] <lerneaen_hydra> oh
[13:21:32] <alex_joni> it gets run from task
[13:21:37] <alex_joni> which isn't RT either
[13:21:47] <lerneaen_hydra> so if you get starvation a faster cpu would help?
[13:21:52] <lerneaen_hydra> or increasing base_period
[13:21:57] <alex_joni> right
[13:22:01] <maddash> this is an atrocity
[13:22:14] <alex_joni> but it's hard to get starvation from interpreting
[13:22:27] <alex_joni> you get starvation from passing things too slowly between userspace and RT
[13:22:32] <alex_joni> maddash: what is?
[13:22:39] <lerneaen_hydra> alex_joni: oh, I see
[13:22:46] <lerneaen_hydra> how is that handled?
[13:22:52] <lerneaen_hydra> how much bandwidth?
[13:23:09] <maddash> the definition of STRAIGHT_TRAVERSE
[13:23:46] <maddash> lerneaen_hydra: userspace talks with RT using shared mem
[13:24:15] <lerneaen_hydra> oh ok
[13:24:39] <maddash> alex_joni: it makes a python call!
[13:28:40] <fenn> big mean nasty python
[13:35:45] <maddash> it would be nice if we could use a standard port besides the parport for I/O
[13:36:44] <archivist> parports are standard
[13:36:58] <maddash> *besides*
[13:37:41] <maddash> but everything else requires an external microcontroller of some sort
[13:41:44] <fenn> coz a microcontroller is a +real+ general purpose computer, not an appliance
[13:45:07] <maddash> brb
[13:51:13] <jepler> maddash was probably looking at 'gcodemodule', which allows AXIS to preview gcode.
[13:51:28] <jepler> there are several definitions of the function STRAIGHT_TRAVERSE, and that's only one of them
[13:52:34] <CIA-8> 03jepler 07TRUNK * 10emc2/docs/src/gcode/main.lyx: mostly updates for nineaxis
[13:53:23] <cradek> it's always an adventure around here isn't it
[13:53:45] <cradek> jepler: you're my hero for updating documentation
[13:59:30] <alex_joni> * alex_joni agrees
[14:00:45] <archivist> is there a proj to do an html version of the docs
[14:01:13] <cradek> archivist: we already have a mediocre html conversion online
[14:01:38] <cradek> http://www.linuxcnc.org/docs/devel/html/
[14:02:28] <cradek> I think the only big problem is with some of the images not showing up
[14:02:33] <archivist> much faster to get to a page that way
[14:07:55] <fenn> what are y'alls thoughts on tk vs gtk? was tkinter chosen for axis simply because its included with python?
[14:08:22] <fenn> and i assume pyvcp followed from that to keep dependencies down
[14:13:57] <jepler> fenn: I asked anders (awallin) to use tkinter so that the panel could be integrated in the AXIS window
[14:14:47] <cradek> I think for a new app it would be hard to argue that gtk is the wrong choice
[14:15:06] <fenn> ok good that's what i wanted to hear :D
[14:15:20] <cradek> what are you making?
[14:15:30] <fenn> nothing *scurry scurry*
[14:15:55] <jepler> on ubuntu (but probably not kubuntu), python-gtk is already installed so in that respect it would not really represent a new requirement
[14:17:27] <jepler> fenn: I chose tkinter for axis simply because my GUI development experience is almost all tk/tkinter
[14:24:05] <jepler> (one of these days I'll do something about that fact ..
[14:24:05] <jepler> )
[14:26:30] <lerneaen_hydra> is there any practical benefit to QT vs gtk?
[14:27:18] <cradek> having never really programmed either one, they seem about the same to me as a user
[14:27:51] <lerneaen_hydra> that's exactly the same experience I've had too
[14:28:58] <cradek> today it seems like gtk is more popular than kde; I'd prefer gtk for a new app because of that
[14:30:28] <lerneaen_hydra> I assume you mean gnome and not gtk?
[14:30:31] <cradek> yes
[14:30:55] <lerneaen_hydra> I've had the opposite experience actually, finding kde more common
[14:31:00] <lerneaen_hydra> and more to my liking
[14:31:09] <cradek> oh ok
[14:31:27] <lerneaen_hydra> IMO it's irrelevant, as most systems will have an app that uses the other toolkit
[14:31:35] <lerneaen_hydra> so there's no penalty
[14:32:59] <fenn> i prefer kde as a desktop but gtk has wider support and seems simpler
[14:33:22] <fenn> and it all looks more or less the same with gtk-qt-engine
[14:36:12] <fenn> maybe more importantly i know who/why gtk was created in the first place
[14:51:30] <maddash> omg python is embedded into milltask
[14:55:22] <jepler> you're mistaken about that.
[14:56:37] <maddash> src/emc/rs274ngc/ is embedded within milltask, no?
[14:57:09] <maddash> and gcodemodule.cc is compiled along with the rest of src/emc/rs274ngc/, right?
[14:57:20] <jepler> milltask links with librs274 which consists of some of the files from src/emc/rs274ngc
[14:57:27] <jepler> however, gcodemodule.cc is not one of those files
[14:57:30] <maddash> only "some"
[14:57:48] <jepler> gcodemodule is also linked with librs274, and together that produces lib/python/gcode.so which is used by AXIS to show the gcode preview
[14:59:13] <maddash> well, judging from the Submakefile, interp_convert.cc is a part of librs274. convert_straight() inside interp_convert calls STRAIGHT_TRAVERSAL
[14:59:24] <jepler> 08:45:12 -!- maddash [n=bryant@unaffiliated/maddash] has quit []
[14:59:24] <jepler> 08:51:14 <jepler> maddash was probably looking at 'gcodemodule', which allows AXIS to preview gcode.
[14:59:32] <maddash> and STRAIGHT_TRAVERSAL is defined in gcodemodule...
[14:59:37] <jepler> if you continue looking you'll find that there are several definitions of STRAIGHT_TRAVERSAL
[14:59:40] <jepler> you've only found one of them
[15:00:02] <maddash> my, what an embarassing grep mistake
[15:00:12] <jepler> emc/rs274ngc/gcodemodule.cc emc/sai/saicanon.cc emc/task/emccanon.cc
[15:01:09] <maddash> 'sai'?
[15:01:14] <maddash> saigon?
[15:01:19] <jepler> it's a slightly unconventional way for a "library" to work: it assumes there are a bunch of functions defined in the main program (STRAIGHT_TRAVERSE for example) that can be called by the library
[15:01:41] <jepler> sai means "stand-alone interpreter"; it reads gcode and just prints out the arguments to the functions like STRAIGHT_TRAVERSE
[15:02:14] <maddash> hm, this looks new -- it's not in 2.1.6
[15:02:17] <jepler> its associated executable file is bin/rs274
[15:02:21] <jepler> yes that's probably true
[15:04:57] <maddash> brb, repatched my patch
[15:13:47] <lerneaen_hydra> what an interesting conversation
[15:33:28] <cradek> how's the plumbing jmk?
[15:35:28] <SWPadnos> that bad, huh?
[15:35:47] <cradek> how's the net connection jmk?
[15:36:03] <SWPadnos> the plumbers dug up his DSL line :)
[15:39:24] <lerneaen_hydra> heh
[17:01:49] <maddash> do people use stewart platforms for cnc tasks?
[17:02:20] <fenn> not many people
[17:02:58] <fenn> but it sounds promising
[17:03:09] <fenn> any idea why this is happening? Failed to fetch http://www.linuxcnc.org/emc2/dists/dapper/emc2.1/binary-i386/Packages.gz Sub-process gzip returned an error code (1)
[17:03:26] <fenn> i can wget/gunzip it fine
[17:04:58] <maddash> try an aptitude clean
[17:05:12] <maddash> or apt-get clean ; apt-get update
[17:06:03] <fenn> no beans
[17:06:23] <maddash> https://bugs.launchpad.net/ubuntu/+source/apt/+bug/59353
[17:06:30] <lerneaen_hydra> maddash: stwart platforms?
[17:07:04] <maddash> lerneaen_hydra: hexapod
[17:07:29] <lerneaen_hydra> oh, right
[17:07:35] <lerneaen_hydra> they're really cool
[17:07:55] <maddash> hell yeah
[17:08:08] <maddash> but quite costly
[17:08:18] <maddash> you need 6 actuators
[17:08:22] <lerneaen_hydra> yeah, true
[17:09:26] <maddash> fenn: it's most likely a connection problem, or there's something wrong with emc's repo. do you get "gzip: stdin: ..." error?
[17:11:07] <maddash> http://www.mmsonline.com/mag_images/0900rt1a.jpg this looks like a fucking hack
[17:12:14] <lerneaen_hydra> maddash: how so?
[17:12:53] <maddash> lerneaen_hydra: they simply built a new platform on the other side of the bridgeport and flipped the quill around
[17:13:06] <maddash> lerneaen_hydra: there's no elegance in the result
[17:13:10] <fenn> Get:5 http://www.linuxcnc.org dapper/emc2.1 Packages [3421B]
[17:13:17] <fenn> gzip: stdin: not in gzip format
[17:13:32] <lerneaen_hydra> maddash: oh, it's a bridgeport
[17:13:36] <maddash> fenn: figures. there's some mojo going on with the emc2 repo
[17:13:40] <lerneaen_hydra> never seen them before
[17:13:54] <fenn> mojo?
[17:14:05] <maddash> lerneaen_hydra: you've been deprived, my dear friend
[17:14:07] <fenn> that bridgeport thing is a hack
[17:14:37] <lerneaen_hydra> maddash: oh? that good? I've only used sayo and a few other types
[17:14:40] <lerneaen_hydra> mostly sayo
[17:14:42] <lerneaen_hydra> storebro too
[17:15:11] <maddash> lerneaen_hydra: maybe not that great -- i'm new to the cnc world, so I'm just boasting
[17:15:15] <maddash> brb, gotta test out my new tool_change patch
[17:17:07] <anonimasu_get> anonimasu_get is now known as anonimasu
[17:37:03] <maddash> hm, does anyone else find limiting M6 motion to the z-axis useful? http://pastebin.ca/634789
[17:41:10] <maddash> no? you guys must have recycling bins full of broken tools
[17:44:08] <anonimasu> maddash: not really..
[17:44:56] <maddash> you remember to g28 everytime you change your tool?
[17:46:13] <lerneaen_hydra> m6 motion?
[17:49:23] <maddash> load tool
[17:49:28] <maddash> ie, t1me
[17:49:35] <maddash> t1m6*
[17:55:12] <toast_> maddash:
[17:55:20] <toast_> it depends on what kind of tool you are using
[17:55:31] <toast_> er, machine
[17:56:16] <toast_> vertical mills usually only need a Z move to get to the toolchanger position, right?
[17:56:22] <toast_> so it doesn't matter for them
[17:56:51] <organetic> good afternoon
[17:57:03] <toast_> the solution i've seen a lot of things implement is use program number 9999 or some such
[17:57:06] <toast_> as a "toolchange program"
[17:57:11] <toast_> and m6 calls that program
[17:57:16] <maddash> toast_: i think it does, because the current implementation of the toolchange position forces you to move the XY axes as well
[17:57:38] <toast_> whack
[17:57:41] <organetic> I do need some help to develop a custom configuration on linux EMC
[17:57:54] <maddash> whoa you can force m6 to call a g-code routine?
[17:58:00] <toast_> on a lot of controllers, yes
[17:58:16] <organetic> I'm expecting to find someone that can do the configuration for me, and I'm willing to pay for that
[17:58:22] <toast_> and when it sees M6 in THAT program, it does the "default" action to get to the toolchange position
[17:58:39] <organetic> if there's anyone willing to help me, please pvt
[17:58:40] <toast_> but all the crap leading up to the M6 is positioning stuff to make sure the tool is in a safe and clear position
[17:59:19] <toast_> like on our OKK vertical mill
[17:59:27] <toast_> er, horizontal
[17:59:38] <toast_> it goes to Y0, which is the top of the column
[17:59:43] <toast_> on an M6 - no other axis moves
[17:59:51] <toast_> but we have the toolchange program, O9999 or whatever
[17:59:57] <toast_> back the Z axis out all the way
[18:00:39] <maddash> ah, so you call 9999 instead of m6
[18:00:56] <toast_> nope
[18:00:59] <toast_> you call M6
[18:01:11] <toast_> and the machine has a setup option, the toolchange program
[18:01:28] <toast_> and the machine calls o9999 completely hidden to the user
[18:02:05] <toast_> and the machine knows when it hits m6 in THAT program to execute the "default" rapid to Y0
[18:02:59] <toast_> it's a very awesome feature, you can even program macros in
[18:03:39] <lerneaen_hydra> toast_: fanuc?
[18:03:46] <toast_> fanuc and mitsubishi that i have seen
[18:03:51] <toast_> i think there is an option in haas
[18:04:02] <toast_> but i've never seen it used, so that might be a pure fabrication on my part
[18:13:57] <maddash> dad needs to make a part. looks like emc's finally getting the test.
[18:14:35] <anonimasu> maddash, I guess whatever you do you need a qualified operator, even commercial controls will fuck up if you tell them to
[18:23:17] <maddash> the maximum jog speed of an axis is the maximum velocity of that axis, right?
[18:23:35] <maddash> because jogging is done in manual aka free mode
[18:24:16] <toast_> jog or mpg
[18:24:29] <toast_> i suppose "yes" to both of those modes
[18:25:22] <toast_> when i crank the MPG in .1 as fast as i can i get full rapid feed on all our machines
[18:25:31] <toast_> but none of our machines have a full rapid jog
[18:25:40] <toast_> only partial rapids for jog
[18:25:42] <toast_> and feed jog
[18:34:19] <anonimasu> toast_: that's pretty sane
[18:34:38] <anonimasu> I mean you have alot more control with jog wheel.. then you have with a button..
[18:34:49] <anonimasu> you can start slow and then make it go fast..
[18:41:33] <toast_> jar
[18:41:33] <toast_> *yar
[18:41:46] <toast_> i have always noticed that the machine has a little lag
[18:41:48] <toast_> on the jog button
[18:41:51] <toast_> so i am glad it never goes too fast by accident
[18:42:07] <toast_> like, it will go for a bit longer than i push the button
[18:42:21] <toast_> maybe that's just a hallucination, but that is my experience
[18:49:29] <toast_> imho jog buttons are retarded
[18:49:41] <toast_> they always seem to take more effort to use than they're worth
[18:49:50] <toast_> clearly a UI problem
[18:50:17] <lerneaen_hydra> toast_: jogging using the KB?
[18:50:32] <toast_> no, i'm talking about commercial cnc machines
[18:50:36] <toast_> i've never used EMC
[18:50:49] <lerneaen_hydra> oh, they've got noticable latency?
[18:51:00] <toast_> well, to me they do
[18:51:02] <lerneaen_hydra> I've never noticed it
[18:51:07] <toast_> and i've never used emc to say if i see the same thign
[18:51:08] <toast_> *Thing
[18:51:13] <lerneaen_hydra> they don't apply as much deceleration though
[18:51:15] <lerneaen_hydra> could be that
[18:51:21] <lerneaen_hydra> at least the ones I've seen
[18:51:22] <toast_> dunno
[18:51:34] <toast_> it does seem to be worse on the larger machines
[18:51:49] <toast_> our 20"x50" slant bed cnc seems to be the worst
[18:51:56] <lerneaen_hydra> how bad?
[18:51:58] <toast_> dunno
[18:52:08] <toast_> it's just noticible
[18:52:11] <lerneaen_hydra> oh ok
[18:52:15] <toast_> like, "i let the button go but it went on for a bit more"
[18:52:23] <toast_> whereas with the mpg it stops dead on
[18:52:29] <lerneaen_hydra> mpg?
[18:52:36] <toast_> manual pulse generator
[18:52:36] <toast_> the knob
[18:52:42] <lerneaen_hydra> oh jogwheel?
[18:52:43] <toast_> yar
[18:52:49] <lerneaen_hydra> ok
[18:53:53] <toast_> also the jog buttons never seem to be useful for anything other than shuffling the turret about at high speed
[18:53:57] <toast_> or the table, whatever
[18:54:09] <toast_> it would be cool if they had an automatic feed or something
[18:54:19] <toast_> make my job easier =(
[18:54:33] <toast_> hold a jog key and hit start or something
[18:54:55] <lerneaen_hydra> automatic feed?
[18:55:18] <toast_> like, if you could set up some kind of stopping plane/short program
[18:55:21] <toast_> to automatically run
[18:55:31] <toast_> so instead of writing a program in mdi or something
[18:55:39] <toast_> you dial your feed and speed on the control's knobs
[18:55:46] <toast_> and just hit your jog button + feed
[18:55:57] <toast_> and the machine would run that cycle in the direction you asked it
[18:56:07] <lerneaen_hydra> oh, so a sort of shorthand/mini program
[18:56:17] <toast_> yeah
[18:57:00] <toast_> we have two diamond turning machines that have controls like that
[18:57:04] <toast_> but that's all they are
[18:57:12] <toast_> feed direction, speed, and how many times you want to do it
[18:57:20] <toast_> makes single parts very, very snazzy
[18:58:01] <lerneaen_hydra> EMC has a similar function, you can set it to jog a certain distance every time you press the jog button
[18:58:26] <lerneaen_hydra> and control feed with a jogwheel or with the mouse
[18:58:39] <toast_> hmm
[18:58:46] <toast_> that's kind of cool
[18:59:20] <toast_> but alas, i have to go to college and sign up for classes since something is retarded with their prerequisites
[18:59:27] <anonimasu> toast_: heidenhains has a thing where you hit nc start and a jog button..
[18:59:33] <toast_> i can't sign up for any classes, even basic basic math
[18:59:34] <anonimasu> latched traverse :)
[18:59:41] <toast_> that's what i'm talking about, that right there
[18:59:45] <toast_> except maybe fancier
[18:59:47] <toast_> okay bbl
[18:59:54] <lerneaen_hydra> toast_: that's kind of fscked up
[18:59:54] <lerneaen_hydra> later
[18:59:57] <anonimasu> they have incremental jog on the buttons too..
[19:00:00] <awallin> anonimasu: hi
[19:00:04] <anonimasu> hey
[19:01:44] <lerneaen_hydra> awallin: before you replaced everything on your OPTI-BF20, was there anything in particular that you broke/could have been broken easily?
[19:02:54] <awallin> hmm, the circuit board for the original spindle motor did brake
[19:03:07] <awallin> and the gearbox for the original spindle is not that great
[19:03:32] <awallin> but we did use it for 2-years almost stock setup. so it does work like that :)
[19:29:28] <SWPadnos_> SWPadnos_ is now known as SWPadnos
[19:56:19] <toast_> lerneaen_hydra:
[19:56:24] <toast_> how is that kind of fsked up
[19:57:16] <toast_> re: earlier w/ feeds on cnc machines
[19:57:50] <lerneaen_hydra> oh, no about your college
[19:58:02] <lerneaen_hydra> you had to go there in person and fill in some stupid paper?
[19:58:12] <toast_> well like, they've got this big elaborate admissions system
[19:58:14] <toast_> that i skipped
[19:58:24] <lerneaen_hydra> oh, right
[19:58:24] <toast_> and a really restrictive web thing
[19:58:40] <toast_> so now i've got to go back and do the whole admission thing
[19:58:47] <lerneaen_hydra> bleh
[19:59:03] <toast_> yar
[19:59:20] <toast_> but the cool part is the college is small so they fudge stuff all the time
[19:59:23] <toast_> like a class i need isn't going to run
[19:59:33] <lerneaen_hydra> O_o okaay
[19:59:37] <toast_> so, they'll slip me into a different class
[19:59:42] <toast_> and let me do sort of an independant study
[19:59:49] <lerneaen_hydra> haha
[19:59:51] <lerneaen_hydra> how small is it?
[19:59:56] <toast_> i dunno
[19:59:56] <toast_> pretty small?
[20:00:06] <lerneaen_hydra> the college
[20:00:07] <toast_> only a couple thousand students
[20:00:11] <lerneaen_hydra> oh, ok
[20:00:17] <toast_> it looks like a highschool
[20:00:20] <toast_> or more like an office building?
[20:00:22] <lerneaen_hydra> ah, I see
[20:00:43] <lerneaen_hydra> so not a spread out campus with multiple buildings and faculties
[20:00:48] <toast_> tiny enough that each program only has one or two time slots for the "critical" classes
[20:00:51] <toast_> correct
[20:01:13] <toast_> like for my mechanical engineering, the core classes for that major only have one time slot a semester
[20:01:51] <lerneaen_hydra> oh, I see
[20:02:47] <toast_> but the positive side is that it's super hands on
[20:02:48] <toast_> every class is like, a lab
[20:03:09] <lerneaen_hydra> ah, nice
[20:03:47] <toast_> seriously
[20:03:54] <toast_> it makes 8:30 am classes much easier to go to
[20:04:36] <toast_> even when i guess i get up for work at that time anyway
[20:04:38] <lerneaen_hydra> heh, I can imagine
[20:23:12] <SWPadnos_> SWPadnos_ is now known as SWPadnos
[20:28:52] <toast_> ff
[20:39:21] <SWPadnos_> SWPadnos_ is now known as SWPadnos
[20:43:55] <SWPadnos_> SWPadnos_ is now known as SWPadnos
[20:57:27] <SWPadnos_> SWPadnos_ is now known as SWPadnos
[21:01:19] <maddash> "motion> cw 2 0 0 1 0 0 1
[21:01:19] <maddash> error: need to be enabled, in coord mode for circular move"
[21:01:56] <maddash> I already issued beforehand,
[21:01:59] <maddash> "motion> activate 0
[21:01:59] <maddash> motion> activate 1
[21:01:59] <maddash> motion> activate 2
[21:01:59] <maddash> motion> coord
[21:01:59] <maddash> motion> enable
[21:01:59] <maddash> "
[21:02:28] <maddash> ...either usrmot hates me, or i'm blind as a bat
[21:02:49] <SWPadnos> could be both
[21:03:17] <maddash> what am I doing wrong, then?
[21:03:33] <SWPadnos> I have no idea what you're doing, let alone what you're doing wrong. sorry
[21:04:26] <maddash> trying to draw a circle in usrmot
[21:04:54] <SWPadnos> unfortunately, I have never attempted anything with usrmot, so I'm really of no help here
[21:05:01] <maddash> the above should've been equivalent to 'g0x0y0, g2x2y0r1'
[21:05:06] <SWPadnos> except to suggest setting coord mode after enabling
[21:05:16] <maddash> haha parrot
[21:08:43] <maddash> * maddash has something wonderful in mind
[21:09:09] <bill2or3> flan?
[21:09:46] <SWPadnos> Salmon Choo-Chee with coconut sticky rice for dessert
[21:09:48] <maddash> tiene flan?
[21:11:18] <bill2or3> Mmm flan...
[21:30:01] <maddash> wtf wtf wtf usrmot is totally broken
[21:31:09] <jepler> maddash: it may have bitrotted, it's completely unused in typical configurations (see also: milltaskcanon)
[21:31:47] <maddash> jepler: emctaskmain fucks usrmot over with the check for emcStatus->io.aux.estop in the main loop
[21:35:47] <maddash> this stinks.
[21:43:35] <MrAsshole> SWPadnos: salmon for dessert?
[21:43:52] <MrAsshole> MrAsshole is now known as Jymmm
[21:43:53] <SWPadnos> insert comma
[21:44:00] <Jymmm> ah
[21:44:40] <R0BH> Hi all
[21:44:47] <SWPadnos> hi
[21:44:59] <maddash> 'mrasshole'?
[21:45:07] <R0BH> i have a problem with constant surface speed
[21:45:27] <SWPadnos> describe away
[21:45:33] <R0BH> its working backwards, as it moved out on the job it speeds up, and inwards it slows down
[21:45:43] <R0BH> feed per rev is working nicly
[21:45:48] <SWPadnos> "it" being the spindle?
[21:45:55] <R0BH> indeed
[21:46:35] <SWPadnos> well, Id help if I really knew how :)
[21:46:43] <R0BH> running CVS as of tuesday
[21:47:25] <SWPadnos> jepler and cradek did the CSS code, and I haven't had a chance to look at it (so I don't know the specifics of how it works)
[21:47:42] <maddash> 'css'?
[21:47:53] <SWPadnos> constant surface speed
[21:48:01] <SWPadnos> for lathe turning
[21:48:25] <R0BH> ill try and catch ether one when they are about then
[21:49:34] <R0BH> was quite funny to see it speed up when it should of slowed
[21:49:45] <SWPadnos> ok. I suspect it won't be long for one or the other to show up
[21:50:11] <SWPadnos> I imagine it would be
[21:50:11] <R0BH> sure thing, ill hover around for abit,thx
[21:50:17] <maddash> should have*
[21:50:33] <SWPadnos> were you using CSS and FPR at the same time?
[21:50:40] <R0BH> no
[21:50:49] <R0BH> switch out of fpr
[21:50:53] <SWPadnos> ok. I think that was checked at Fest, but I don't remember
[21:50:57] <R0BH> and out css visa versa
[21:51:26] <R0BH> ill go have abit more of a play for now see what i can find
[21:52:20] <Jymmm> SWPadnos: how was the video?
[21:52:28] <SWPadnos> one thing to check - look at motion.spindle-speed-out (I think that's the name), and see if the commanded speed is going up. there could be some weird problem in your setup where very high commanded speeds wrap around or give other unintended behavior
[21:52:47] <SWPadnos> Jymmm, interesting. the audio was out of sync with the video though, so it was kind of funny
[21:53:05] <maddash> wow
[21:53:11] <Jymmm> SWPadnos: I use VLC for ALL video playback, no issues ever
[21:53:21] <maddash> i'm almost omnipresent within the IRC logs
[21:53:21] <SWPadnos> I doin't, and I obviously have issues sometimes:)
[21:53:33] <Jymmm> SWPadnos: win or nix?
[21:53:51] <R0BH> i started at S50 nice low
[21:54:02] <SWPadnos> I think I may be just about ready for a Mozilla restart - it's getting slow and is now using 478M of RAM
[21:54:09] <SWPadnos> Win on this machine
[21:54:11] <maddash> aha, windows?
[21:54:15] <maddash> :P
[21:54:26] <Jymmm> SWPadnos: http://portableapps.com/apps/music_video/vlc_portable
[21:54:30] <maddash> how do you use emc, then?
[21:54:54] <SWPadnos> this is a 6+ year old computer which has survived a hard disk failure and other stuff, so don't knock it
[21:55:02] <SWPadnos> I have other computers
[21:55:14] <Jymmm> ...that I steal from my wife
[21:55:36] <SWPadnos> no, that was just for experimentation :)
[21:55:41] <Jymmm> ah, gotcha =)
[21:56:11] <Jymmm> SWPadnos: try the portable VLC, not harm done to your computer
[21:56:15] <Jymmm> s/not/no/
[21:56:17] <SWPadnos> much better. got rid of all the car ratings windows and their flash crap
[21:56:51] <SWPadnos> now Moz is only using 0-1% of the CPU instead of 97%
[21:56:55] <maddash> SWPadnos: you're married?
[21:57:01] <SWPadnos> last I checked, yes
[21:57:09] <Jymmm> maddash: SWPadnos has 7 wives!
[21:57:12] <maddash> :O
[21:57:18] <SWPadnos> shhh - the other 6 may be listening
[21:57:19] <maddash> * maddash has 72!
[21:57:41] <maddash> SWPadnos: haha, "No swarf rule" for you!
[21:58:18] <SWPadnos> only in the bedroom
[21:58:29] <SWPadnos> and I'm happy with that rule :)
[21:58:29] <maddash> * maddash is shocked
[21:59:45] <Jymmm> SWPadnos: One reason I REALLT like VLC, is it comtains all the CODECS, no need to add any
[22:00:09] <lerneaen_hydra> anyone here know spreadsheet/excel apps any good?
[22:00:24] <maddash> oo.o?
[22:00:34] <Jymmm> SWPadnos: it'll even play FLA/SWF files
[22:00:37] <lerneaen_hydra> yeah, preferably
[22:00:53] <maddash> lerneaen_hydra: OO.org?
[22:01:06] <maddash> lerneaen_hydra: MS Excel? rofl.
[22:01:17] <lerneaen_hydra> what I'm doing is copying a formula around a bit, and I don't want some of the references to change cell automagically when I paste in a new cell
[22:01:23] <lerneaen_hydra> I'm using OO.org atm
[22:01:37] <maddash> what's the problem?
[22:01:40] <Jymmm> I have a DOS spreadsheet if you want, even uses the old 1-2-3 commands =)
[22:01:40] <lerneaen_hydra> seeing as how I'm on a linux box MS excel is sort of out of the question
[22:02:14] <SWPadnos> I think you can do "paste special" and uncheck something
[22:02:53] <lerneaen_hydra> say I have the contents of cell C1 =B1+A1, and I copy C1 to C2, and I want B1 to change to B2 (default behavior), but A1 to be unchanged
[22:02:56] <jepler> lerneaen_hydra: use $ to make a reference absolute. E.g., $A1 to make "A" stay fixed if you copy the cell, A$1 to make "1" stay fixed, and $A$1 to make the whole thing stay fixed
[22:03:05] <lerneaen_hydra> jepler: ah, perfect
[22:03:06] <lerneaen_hydra> thanks
[22:03:15] <SWPadnos> oh - that's too easy :)
[22:03:23] <lerneaen_hydra> yeah
[22:03:26] <lerneaen_hydra> far too easy
[22:10:01] <maddash> my god
[22:10:29] <maddash> whoever wrote interp_read.cc must like pain or something
[22:11:25] <tomp> btw: copy & paste is ok with $A$1 in a formula or reference, while A1 becomes A1,A2,A3...An
[22:12:55] <jepler> R0BH: I did some "testing" of constant surface speed using the sim/lathe configuration. On my system, the speed of the simulated spindle (shown on the right) rises as the diameter decreases, which is what I expect.
[22:14:48] <R0BH> HI jepler back
[22:14:59] <R0BH> mine is doing the opasit
[22:15:07] <R0BH> my hal here, http://innovative-rc.co.uk/EMC/orac-lathe.hal
[22:15:44] <R0BH> is there a list of all the new Gcodes for css etc
[22:17:58] <jepler> we try to keep both the manual and quick reference up to date but I won't guarantee there's nothing missing. When I had to remind myself of the code for CSS, I looked at the quickref: http://linuxcnc.org/docs/devel/html/gcode.html
[22:19:15] <R0BH> thx
[22:20:09] <jepler> R0BH: is your lathe set up so that in the machine coordinate system, X=0 is the center of rotation of the spindle?
[22:20:20] <R0BH> yes
[22:20:33] <R0BH> for the tool/touch off
[22:20:47] <R0BH> machine home no
[22:21:15] <jepler> have you verified the commanded spindle speed is matching the actual spindle speed -- for instance by using halmeter
[22:22:03] <R0BH> in normaly non css etc, ys if i comand s200 i checked and it does do 200rpm etc
[22:22:08] <R0BH> allway to s2000
[22:22:53] <jepler> I'm honestly not sure how CSS plays with G54 offsets, the kind you get with touch off
[22:23:13] <jepler> it's supposed to take into account the tool length offset from G43 / G43.1
[22:23:27] <jepler> but if you are using a G54 offset to make the tool tip be at offset coordinate X=0 I'm not sure what behavior you get
[22:23:58] <lerneaen_hydra> jepler: do you mean you don't know how it's coded or how it should be coded?
[22:24:00] <jepler> lerneaen_hydra: both
[22:24:04] <lerneaen_hydra> oh
[22:24:13] <lerneaen_hydra> shouldn't it take all into account?
[22:24:34] <lerneaen_hydra> x=0 is the middle of the part/center of rotation
[22:24:38] <lerneaen_hydra> I assume
[22:24:56] <R0BH> i have not written a program to drive a offset on and profile something, just trying it in MDI
[22:26:23] <R0BH> i guess if i go past center feeding in, it should change speed in opasite direction as to feeding in
[22:26:30] <R0BH> would tell me if it knows where center is
[22:26:42] <jepler> yes -- when I go past X=0 the speed falls off again
[22:26:48] <jepler> -X or +X behave the same
[22:27:23] <R0BH> if it didt, it would say that it thinks the machine home is x=0 and working from that, as it speeds up as i feed to the out side of the bed
[22:28:11] <R0BH> i will try it and see
[22:30:37] <maddash> why doesn't the rs274 interp use regex? it would be so much smaller, faster, and simpler
[22:30:42] <SWPadnos> I had thought about passing zero - jepler, does the CSS code look at machine zero or zero in the current coordinate system? (or something else)
[22:31:09] <SWPadnos> maddash, write a regex that actually works for G-code, and I'll buy you a donut
[22:31:35] <SWPadnos> remember that spaces are ignored (efefctively removed), so only letters and carriage returns are word breaks
[22:31:59] <maddash> SWPadnos: http://www.codeproject.com/vb/net/CNC_Viewer.asp?df=100&forumid=381127&exp=0&select=2011824
[22:32:33] <maddash> SWPadnos: I like mine glazed with rainbow sprinkles
[22:32:46] <SWPadnos> did you write that? :)
[22:33:07] <jepler> maddash: after skynet became sentient, it read the contents of the IRC logs, went back in time to the 1990s, and wrote emc to have all the design choices you complained about. You see, skynet knew that if emc were instead the flawless jewel it was before the terminator interfered with the timeline, you would instead have been working on the network software that gave rise to skynet, and stopped it before it wiped out the human race.
[22:34:07] <SWPadnos> I think their regex doesn't accept all legal G-code
[22:34:07] <maddash> i'm merely suggesting, not complaining
[22:34:28] <SWPadnos> with G-code, these two lines are identical:
[22:34:38] <SWPadnos> g1x0y23.5Z0
[22:35:01] <SWPadnos> g 1 y 2 3 . 5z0 x 0
[22:35:19] <ds2> I have to ask... is there indent or something equivalent for G code?
[22:35:23] <SWPadnos> actually, those are semantically equivalent, but not identical
[22:35:34] <SWPadnos> g 1 x 0 y 2 3 . 5z0
[22:35:37] <SWPadnos> would be identical
[22:35:53] <SWPadnos> ds2, I don't think so
[22:36:30] <SWPadnos> maddash, I'm not a regex expert, but I did try to help along a syntax hilighting thing, and it never quite worked right
[22:36:44] <cradek> for CSS your machine zero position for X should be at the center of rotation (when there is no tool offset)
[22:37:07] <cradek> if it's going backward I suspect your zero is wrong
[22:37:19] <maddash> SWPadnos: is ultraedit open source? it supports gcode highlighting
[22:37:35] <SWPadnos> kate and gedit do also (as well as emacs and vi)
[22:37:36] <cradek> oh I see you guys discussed that already, sorry
[22:37:51] <SWPadnos> support for hilighting isn't the issue, it's the regexes to determine how to hilight
[22:38:34] <maddash> how the heck does awk handle this kind of stuff?
[22:39:27] <maddash> w/e. anyhow, the interp_read code seems a bit...bloated...
[22:40:20] <maddash> jepler: i'm working on adding support to rs274ngc/* for fanuc's g02.4 tilted arc scheme -- any chance that you'll merge it when I'm done?
[22:40:51] <R0BH> g28g91x0 is sending the z home also not just x
[22:40:57] <cradek> I'd like to have the ability to do arbitrary arcs
[22:41:15] <maddash> cool. i'll keep working.
[22:41:29] <cradek> R0BH: g28 sends all axes home, optionally through a specified point
[22:41:32] <SWPadnos> g28 and g30 home all axes, they just use a waypoint for the specified axes, I think
[22:42:09] <cradek> http://www.linuxcnc.org/handbook/RS274NGC_3/RS274NGC_33a.html#1013246
[22:42:54] <R0BH> how do you set speed limit on CSS
[22:42:55] <cradek> maddash: I have started it a few different times. it's a pretty big project.
[22:43:44] <cradek> G96 Sxxx Dxxx
[22:44:22] <R0BH> same as philips not fanuc then
[22:44:24] <cradek> S is speed in fpm/mpm, D is top rpm
[22:45:11] <lerneaen_hydra> cradek: off the top of my head, wouldn't the simplest approach be two lines of codes, one line Gxx Xa Yb Zc, where a b and c define a vector that's perpendicular to a plane, and then a second line G2/3 that's done just like a normal arc, but in this new, temporary plane instead
[22:45:20] <maddash> gotta go home brb.
[22:45:49] <lerneaen_hydra> it seems to me the math at least would be easy to implement
[22:45:55] <lerneaen_hydra> basic linear algebra
[22:46:06] <cradek> lerneaen_hydra: for the g2/g3, what are ijk?
[22:46:08] <R0BH> jepler, tryed more on CSS and it speeds up still as diamiter increase and when go past x=0 it tryed to go even slower finding
[22:46:30] <cradek> R0BH: if you do G49 G0 G53 X0, where is the tool?
[22:46:35] <lerneaen_hydra> cradek: coordinates to the center of the arc in the new, temporary plane
[22:47:08] <cradek> lerneaen_hydra: that seems good at first glance, I wonder how others do it
[22:47:22] <lerneaen_hydra> it's clean from a math standpoint
[22:47:34] <lerneaen_hydra> as to the practicality when coding, that's another matter
[22:48:17] <lerneaen_hydra> if some other input is decided on it's probably easiest to convert to a system like that and then do math from there
[22:48:27] <lerneaen_hydra> as it's so simple from there
[22:48:59] <lerneaen_hydra> just a change of base matrix P(C<-B)
[22:49:42] <SWPadnos> I think a lot of problems for more advanced features aren't the math, but how a person could write the code (and still have it compatible with G-Code)
[22:49:41] <lerneaen_hydra> essentially G17/18/19 are simplified versions of that, 1,0,0 0,1,0, and 0,0,1
[22:49:53] <lerneaen_hydra> SWPadnos: yes, exactly
[22:50:08] <lerneaen_hydra> this method may be easy to implement, but it would probably be hard to code
[22:50:15] <SWPadnos> like splines, arbitrary arcs, elliptical arcs, etc.
[22:50:42] <R0BH> cradek, G53 puts it at home postion on x
[22:50:44] <SWPadnos> I think someone (jepler or cradek) checked the underlying ellipse code, and it does arbitrary ellipses now
[22:50:56] <lerneaen_hydra> sweet
[22:51:40] <cradek> R0BH: I know what it's supposed to do, I'm asking if your uncompensated (G49) tool tip is at the center of rotation at machine zero (G53 X0)
[22:51:53] <cradek> you will have to set your homing parameters to ensure this
[22:53:03] <R0BH> the full comand moved it to the outter of bed, to the machines home on X
[22:53:32] <cradek> that's why it goes slower when you move in
[22:53:41] <cradek> machine zero needs to be at the center of rotation
[22:54:08] <R0BH> right
[22:54:15] <cradek> when you're moving in, emc thinks the radius is getting bigger, because you're moving away from zero
[22:54:31] <cradek> I think you can fix this by changing just HOME_OFFSET
[22:54:50] <R0BH> ill have a look
[23:00:36] <lerneaen_hydra> /random musing about XYZ arcs/ it seems to me the easist type of arc to imagine mentally is one where you define the star and end coords in a normal coordinate system, and an IJK position. that determines the arc's position completely *except* for the case when you want to make an arc that's exactly half a turn. (also you can't use R to define radius without loosing one degree of freedom). one possible way to fully
[23:00:36] <lerneaen_hydra> define it would be to define a vector that the arc is perpendicular to
[23:01:49] <lerneaen_hydra> but now it's far too late
[23:01:51] <lerneaen_hydra> 'night everyone
[23:01:51] <cradek> so specify all of IJK to locate the center, and the 3 points specify a plane (except for the semicircle or full circle case)
[23:02:11] <lerneaen_hydra> hmm
[23:02:18] <cradek> but ... what about arbitrary helixes?
[23:02:28] <lerneaen_hydra> I think I'll need to draw in a 3d app
[23:02:36] <lerneaen_hydra> oh, that's a lot easier
[23:02:41] <SWPadnos> program no e.g. Z motion, and you'll get an arc in 3D
[23:03:03] <SWPadnos> Z still gives a helix depth, but just perpendicular to the arbitrary plane
[23:03:15] <SWPadnos> or not, since you need XYZ for the endpoint :(
[23:03:33] <cradek> yeah I think you really want to specify the endpoint exactly
[23:03:39] <SWPadnos> how about H :)
[23:03:43] <lerneaen_hydra> you just need a vector to define what the helix center is paralel to, a pitch, a diameter, and (if there's want for the feature) a finish diameter
[23:03:44] <SWPadnos> Helix depth
[23:03:57] <lerneaen_hydra> yes depth/length too
[23:04:26] <lerneaen_hydra> by helix center I mean the helix's centerline
[23:04:42] <cradek> I think H is no good - you don't exactly specify the endpoint
[23:05:14] <jepler> SWPadnos: I don't think it does ellipses. ellipses are not constant velocity
[23:05:14] <lerneaen_hydra> you could specify start coord and how many revolutions
[23:05:24] <lerneaen_hydra> or wait
[23:05:30] <cradek> who said ellipses? ack
[23:05:45] <lerneaen_hydra> you dont want to specify start coord, that's difficult to find exactly
[23:05:46] <SWPadnos> jepler, ok. I recall a screenshot of a non-aligned circle - it may have just looked like an ellipse from perspective
[23:06:06] <SWPadnos> the underlying function is called *ellipse
[23:06:24] <SWPadnos> something_ellipse - and I think it does take major and minor diameters
[23:06:23] <lerneaen_hydra> but alack, I was going to go to bed :p
[23:06:26] <lerneaen_hydra> 'night all
[23:06:29] <SWPadnos> see you KLH
[23:06:31] <SWPadnos> -K
[23:06:32] <cradek> there is no low level code for ellipses
[23:06:36] <cradek> goodnight
[23:06:43] <lerneaen_hydra> *pof*
[23:06:46] <lerneaen_hydra> *poof*
[23:06:56] <SWPadnos> hmmm. I'm pretty sure I saw it before. I'll look for it again
[23:07:23] <R0BH> whats best way to set G53
[23:08:13] <cradek> R0BH: http://www.linuxcnc.org/docs/devel/html/config/ini_homing/index.html
[23:08:27] <SWPadnos> 2late4u
[23:08:31] <cradek> argh
[23:08:38] <SWPadnos> oops - 2l84u
[23:08:55] <cradek> R0BH: http://www.linuxcnc.org/docs/devel/html/config/ini_homing/index.html
[23:09:36] <cradek> R0BH: see especially HOME_OFFSET and HOME
[23:10:23] <R0BH> tryed home_offset but 53 did not offset from 0
[23:11:45] <cradek> which of those 4 homing cases in the picture applies to your machine?
[23:13:07] <R0BH> there is no home switch on this machine
[23:13:08] <CIA-8> 03cradek 07TRUNK * 10emc2/docs/html/gcode.html: fix CSS-related documentation
[23:13:09] <CIA-8> 03cradek 07TRUNK * 10emc2/docs/src/gcode/main.lyx: fix CSS-related documentation
[23:13:21] <cradek> R0BH: how are you homing?
[23:13:30] <R0BH> we have just set a home postion as such
[23:13:51] <cradek> ok that's fine
[23:14:08] <cradek> but you still have to tell emc where it is, even though emc doesn't move when you home
[23:14:16] <cradek> you do that with HOME_OFFSET
[23:15:18] <cradek> do you have encoders with index pulse?
[23:15:44] <R0BH> no just steppers
[23:16:34] <cradek> ok do this: start emc, MDI mode, g49. jog X so it's at the center of rotation. click home.
[23:16:49] <cradek> (assuming HOME_OFFSET=0)
[23:17:04] <R0BH> it is 0 yes
[23:17:39] <cradek> if you have no home switches, you still home by clicking the home button, but emc thinks "oh the switch is right here"
[23:17:53] <cradek> the homing docs are still all relevant, except it doesn't move to find the switch
[23:19:44] <R0BH> ok done home on 0 spindle, and puts zero symbol on x now
[23:20:03] <cradek> ok now try CSS, it will work the right way
[23:20:32] <R0BH> will give it ago, give us sec
[23:38:31] <R0BH> ok set 53 as said, but if u home it with g28, x stays and does not go to where i had machine home near end x travel
[23:39:12] <cradek> yes that's the point! x=0 has to be the center of rotation
[23:39:20] <cradek> but, you can set the g28 location to whatever you want
[23:40:05] <cradek> looks like you set parameter #5161 to change g28's X location
[23:40:12] <cradek> http://www.linuxcnc.org/handbook/RS274NGC_3/RS274NGC_33a.html#1013246
[23:40:28] <Jymmm> Hey, what are most of you lubing your balls with?
[23:41:08] <Jymmm> I degreased the ballscrew compaletely, and jsut now hit it with a shot of wd-40.
[23:42:14] <Jymmm> I noticed that it seems the ballnut has a "bump" (for lack of better words) when vertcial goign A-to-B, but free spins when going from B-toA
[23:42:58] <cradek> rick says to always check the balls (several times each) with a micrometer, sometimes you get oval balls instead of round with cheap screws
[23:43:01] <Jymmm> Like maybe the ball return has friction at some point, because it seems to be every 1.5" or so.
[23:43:28] <cradek> I know squat about it - just repeating what I've heard
[23:43:47] <cradek> hmm, that sure does sound like a bad ball doesn't it
[23:44:10] <Jymmm> cradek: That might make perfect sense.... maybe an oval ball hits the return tube every 1.5"
[23:44:26] <Jymmm> there are 62 balls in the tube iirc
[23:44:30] <cradek> yeah, or maybe one is just the wrong size
[23:44:36] <cradek> that'll keep you busy for a while!
[23:44:51] <Jymmm> Yeah, no shit... even worse getting replacements
[23:45:08] <Jymmm> If I have to replace the balls, might has well pre-load the SOB
[23:45:09] <cradek> if only one is bad, 61 ~= 62 :-)
[23:45:54] <Jymmm> Damn, wish I was in LA... found a ball mfg that will let you pre-load your ballscrews free, just pay for what you use.
[23:46:10] <Jymmm> walk-in only service
[23:46:19] <Jymmm> DIT
[23:46:20] <Jymmm> DIY
[23:46:39] <SWPadnos> do you have to pay for all the balls you drop on their floor?
[23:47:07] <Jymmm> SWPadnos: Heh, they have bins of millions of balls... I somehow doubt it =)
[23:47:44] <SWPadnos> they should be cheap anyway - just get a baggie of them and do it at home :)
[23:47:59] <Jymmm> not cheap via mail order.
[23:48:28] <Jymmm> first you buy a sample kit, then order what you need once you find the right size you need.
[23:49:15] <Jymmm> cradek: These are the cheapest Thomson ball screw, not china imports. you think that still applies?
[23:49:49] <cradek> 18:43:28 < cradek> I know squat about it - just repeating what I've heard
[23:49:56] <cradek> :-)
[23:49:56] <Jymmm> cradek: lol, fair enough.
[23:50:36] <Jymmm> I'll try to find the url to their website again, jsut if you guys ever need it... REALLY REALLY nice guy (owner), talked to him on the phone for a couple hours.
[23:51:02] <Jymmm> Just FYI... their company is in South Central =)
[23:51:14] <SWPadnos> so mail order is better then
[23:51:47] <Jymmm> S. Central aint too bad.... just not a place a white guy wants to be at night.
[23:52:01] <Jymmm> (unarmed)
[23:52:28] <Jymmm> fsck,... dont want to be unarmed durign the day either, but that's another story
[23:52:57] <SWPadnos> so mail ordering lots of unneeded balls looks better and better, doesn't it?
[23:53:12] <Jymmm> So boys and girsl... just check the court docket before heading in person =)
[23:53:17] <SWPadnos> especially when you throw in a drive from SJC (or a cross-country flight :) )
[23:53:39] <Jymmm> They're 10 minutes from LAX
[23:54:34] <ds2> Jymmm: you seen this - http://thms.tedatum.com/Taig/Selecting_the_stepper.pdf
[23:55:29] <Jymmm> thanks ds2
[23:56:04] <Jymmm> Sometimes I miss SoCal... can fidn EVERYTHING there I ever needed. In NoCal, it's order it all from SoCal
[23:56:51] <jmkasunich_> jmkasunich_ is now known as jmkasunich
[23:57:06] <Jymmm> Lot of good surplus stores too... even electro mechanical and industrial surplus ones
[23:58:26] <Jymmm> If I'm so inclined, anyone mind non emc things on the wiki? Like the ball bearing mfg in LA as example?
[23:58:37] <Jymmm> "Resources"
[23:58:48] <ds2> where in SoCal are the good surplus stores?
[23:59:16] <Jymmm> ds2: all over... top of my head... Colorado Blvd, Pasadena
[23:59:16] <R0BH> cradek you still here
[23:59:37] <ds2> Jymmm: that's all developed
[23:59:50] <Jymmm> ds2: so?
[23:59:59] <ds2> so the stores are gone