#emc | Logs for 2007-08-16

[12:55:47] <skunkworks> SWPadnos: people like it because it is familiar
[12:56:38] <alex_joni> I hate it on machine controls for the same reason
[12:56:45] <alex_joni> because it's crappyly familiar
[12:57:44] <skunkworks> :)
[12:59:08] <jepler> fenn: It's manpage markup, with an extra level of escaping
[12:59:25] <jepler> fenn: it looks ugly in the source code but nice in the generated manpages
[12:59:55] <fenn> what is the name of the software that extracts that?
[12:59:58] <jepler> \fI selects the italic font, \fB selects bold, and \fR selects "roman" (neither bold nor italic)
[13:00:31] <jepler> "man groff_man" for more information about the markup
[13:00:52] <jepler> comp --document extracts all the documentation strings into a manpage with a ".9" extension
[13:01:06] <jepler> then man converts it into a variety of formats
[13:01:22] <fenn> ok thanks
[13:01:44] <skunkworks> we have some vinyl cutters that have xp front ends - easy to use for the most part.
[13:01:58] <skunkworks> * skunkworks doesn't really have to do anything with them
[13:07:09] <jlmjvm> jlmjvm likes amd athlon processors
[13:09:24] <jepler> hm it said something about my post not appearing until it is approved by a moderator, then went back to the forum index
[13:09:42] <fenn> boo hiss
[13:10:06] <jlmjvm> how is delta tau able to use linux,is it emc
[13:10:07] <jepler> I hate web forua
[13:10:20] <fenn> forae?
[13:11:04] <jepler> yeah serves me right for trying to show how "clevar"[sic] I am
[13:11:38] <fenn> all that l33t speek set off their spam filters
[13:11:50] <jlmjvm> has anyone seen the ajax cnc page
[13:12:29] <dimas> url?
[13:12:35] <jlmjvm> looks kinda like the tkemc screen
[13:13:09] <jlmjvm> http://www.ajaxcnc.com/
[13:13:52] <fenn> well.. it's blue
[13:14:01] <jlmjvm> lol
[13:15:46] <jlmjvm> wonder if its emc with some changes
[13:18:03] <fenn> a lot of changes, but yes i bet it is
[13:18:56] <jlmjvm> i bet it is too
[13:19:03] <dimas> looks it has nothing to do with ajax as a web technology :)
[13:19:49] <fenn> wonder what '1MB Standard Program Size' means
[13:20:50] <jlmjvm> where did you see that
[13:21:15] <fenn> http://ajaxcnc.com/features_specs.htm under Operation
[13:21:23] <jlmjvm> never mind,i found it
[13:23:49] <jlmjvm> im pretty sure you have to pay for an upgrade to get unlimited file size
[13:24:33] <SWPadnos> jlmjvm, the Delta Tau unit is not running EMC, it's their own control software / programming language
[13:24:50] <SWPadnos> they do seem to have most or all of the features of EMC, plus a few
[13:25:16] <SWPadnos> fully programmable kinematics, vel, accel and jerk limiting (with kinematics)
[13:25:28] <SWPadnos> sine-curve accel profiles ...
[13:25:49] <fenn> what does delta-tau have to do with ajax?
[13:26:08] <SWPadnos> I was responding to something earlier, while I was AFK
[13:26:39] <jlmjvm> does emc have jerk limiting
[13:26:45] <SWPadnos> nope
[13:27:20] <jlmjvm> darn
[13:27:25] <SWPadnos> yep
[13:28:12] <jlmjvm> i been up all nite running my mill
[13:28:27] <SWPadnos> that sounds dangerous^Wfun
[13:28:57] <jlmjvm> it is,almost fell outta my chair,lol
[13:34:30] <anonimasu> jlmjvm: feel free to implement it
[13:34:42] <anonimasu> ^_^
[13:37:06] <skunkworks> jepler: it went thru. thank you.
[13:45:28] <SWPadnos> jepler, the simple HAL serial driver you wrote (DPP) - was that RT or not?
[13:50:30] <jepler> SWPadnos: I never finished that
[13:50:51] <SWPadnos> err - the serial driver you almost wrote, was that RT or not? :)
[14:16:15] <skunkworks> * skunkworks is going to get everyone to post on cnczone... Buahahahahah...
[14:29:45] <alex_joni> ha
[14:30:52] <skunkworks> Hi alex. Did you get your nice monitor?
[14:46:00] <alex_joni> yeah, looks good
[14:50:10] <skunkworks> ncie
[14:50:11] <skunkworks> nice
[14:50:29] <jepler> SWPadnos: I've kicked around the idea of writing a realtime serial driver but never done it
[14:51:36] <jepler> SWPadnos: more or less what I alluded to in the cnczone post: 10 byte packets, a combination of some small integers or deltas and some bits
[14:53:29] <jepler> clearly it's very simple, that's why there's no need for me to write it
[14:54:45] <alex_joni> heh
[15:01:56] <alex_joni> bbl
[15:24:44] <SWPadnos> jepler, ah. My aim is less for a serial motor controller and more for communication with a PLC, possibly using MODBUS
[15:25:21] <SWPadnos> I'm not sure my app needs RT, but since RT tasks continue to run even if the computer crashes (which improves latency), I'd like to use RT if possible
[15:25:32] <skunkworks> I would think handshaking would be the pain
[15:25:54] <SWPadnos> yes, I'm sure that's at least part of the pain :)
[15:27:44] <SWPadnos> it seems that a fast-thread function to just buffer input/drip-feed output, plus a slow-thread function to interpret the data, should work OK
[15:28:33] <SWPadnos> too bad there's no HAL "stream" data type :)
[15:28:45] <skunkworks> yet..
[16:19:13] <jepler> cradek: for circuit board milling (10/10 at the finest), how much machine backlash do you think would be tolerable using emc's backlash compensation?
[16:19:37] <jepler> .005? .003? .001?
[16:19:46] <cradek> I'm not sure
[16:20:24] <skunkworks> I would think backlash comp would work ok for circuit boards as your not probably going to push the table around with the cutter.
[16:20:44] <cradek> the problem I had with backlash in the machine was not that the circles weren't round (software correction can fix that), but that on long traces the axis would wander to the other side of the backlash, destroying the trace (software correction has no clue)
[16:21:10] <cradek> seems that skunkworks is right, you wouldn't think the axis would move much because of cutting force
[16:22:01] <cradek> maybe more likely is that you fling the axis to the wrong side of the backlash sometimes
[16:22:11] <skunkworks> that could be.
[16:22:26] <cradek> using low accel may help that
[16:22:33] <skunkworks> but that sucks too ;)
[16:23:02] <cradek> yeah cutting features with the same order of magnitude as your backlash is probably not going to work well no matter what
[16:23:35] <cradek> jepler: to answer your question, I'd say .002, since you won't nuke entire features no matter what happens
[16:23:52] <cradek> the only thing you lose is consistency of width, no matter how badly compensation works
[16:24:20] <skunkworks> jepler: descided on a machine yet?
[16:24:37] <cradek> (reliable .002 is probably almost as hard to get as .000 with regular nuts on acme)
[16:25:03] <cradek> I would not hesitate to use springs again for doing pcbs on a sloppy setup
[16:25:16] <jepler> skunkworks: well I'm still thinking about this "zenbot" since it's comparatively inexpensive and nearly ready-to-use
[16:25:26] <skunkworks> cradek: how did your spring loaded setup work (before the dumster nuts)
[16:25:48] <cradek> great actually
[16:26:05] <skunkworks> looked like it should have.. cool.
[16:26:49] <jepler> in addition to unspecified "very low backlash", there's also whatever slop the friction-mounted (?) dremel tool introduces
[16:27:51] <skunkworks> jepler: the only main issue I see with the zenbot is there is absolutly no way to adjust anything on it. (short of re-machining) So your at the mercy of how well they machined it. The feedback from people all seem to be positive though
[16:28:02] <skunkworks> have you emailed them about the backlash?
[16:28:16] <jepler> skunkworks: you mean the ebay feedback, or have you found other stuff people have written about the mill?
[16:28:21] <skunkworks> ebay feedback
[16:28:27] <cradek> plain brass nuts on 1/4-20 rod is going to be a bit sloppy. you'll have to do something.
[16:28:47] <cradek> I highly doubt they are even acme
[16:28:57] <skunkworks> jepler: did you see that there are 2 videos on youtube?
[16:29:45] <jepler> skunkworks: no, I only watched the one in the ebay listing
[16:30:04] <cradek> also, I'm not sure the all-plastic-and-wood construction will give the reproducible depths of .003 you need
[16:30:21] <skunkworks> http://youtube.com/watch?v=4cnrgqTEDh4
[16:30:59] <jepler> skunkworks: that's no the same machine is it?
[16:31:38] <jepler> the spindle mount is totally different for instance
[16:32:37] <skunkworks> right - might be an older version or something - posted by the same guy though
[16:34:38] <cradek> jepler: one of us should have bought the ballscrew XY contraption that ... um, that guy had at workshop
[16:34:46] <cradek> can't remember his name - the one with the puma robot
[16:35:17] <jepler> cradek: yeah maybe so
[16:35:19] <skunkworks> steve something white?
[16:35:27] <cradek> yeah that's him
[16:35:36] <jepler> looks like he has or had another, slightly larger machine: http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=190114389819
[16:35:43] <cradek> he probably still has it
[16:35:50] <cradek> I think I have his email or phone number somewhere
[16:35:52] <skunkworks> I think I have his email
[16:35:53] <skunkworks> heh
[16:37:26] <cradek> I think a fast servo scara robot with dremel would be a neat thing to try for engraving PCBs
[16:37:34] <cradek> not sure if you could make it solid enough but still fast
[16:38:24] <toast> ?
[16:39:11] <jepler> you'll probably find that you need extreme gearing
[16:39:52] <cradek> jepler: that one is bronze on acme, much better
[16:40:54] <jepler> ooh lunchtime -- bbl
[17:14:30] <SWPadnos> I'm not sure that zenbot is really worth the $450 price tag. it has no motors or drives, no spindle, and no controls - it's basically a bunch of HDPE flats bolted together with some 1/4-20 screws thrown in
[17:32:55] <skunkworks> jepler: here you go http://i177.photobucket.com/albums/w231/Spruft/P1010004.jpg
[17:33:11] <skunkworks> yes - that is concrete
[17:33:52] <skunkworks> http://www.elektronikforumet.com/forum/viewtopic.php?t=18483&postdays=0&postorder=asc&start=0
[17:58:05] <jepler> skunkworks: that's crazy
[18:01:41] <cradek> I agree with SWPadnos
[18:03:21] <jepler> the smaller HDPE one does include motors, the larger MDF one doesn't (and isn't currently for sale on ebay)
[18:10:45] <SWPadnos> the one you linked (item # 190114389819) doesn't include motors
[18:11:41] <SWPadnos> ah - the more expensive one does have motors, though they're NEMA 17's
[18:13:40] <jepler> oh -- oops
[18:13:53] <jepler> http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=190141490742
[18:15:47] <SWPadnos> yep - that's the one I'm looking at now
[18:15:58] <SWPadnos> 75 oz-in NEMA 17 motors are included on this one
[18:18:14] <SWPadnos> I'd probably go for the one without motors, and get a set of NEMA-23's from Automation Direct
[18:18:28] <SWPadnos> they have 166 oz-in for $29 and 276 oz-in for $39 each
[18:18:42] <SWPadnos> http://web6.automationdirect.com/adc/Shopping/Catalog/Motion_Control/Stepper_Systems/Motors_-z-_Cables
[18:19:01] <SWPadnos> (or $19 for NEMA-17 83 oz-in)
[18:21:05] <jepler> I don't understand the claim on the ebay auction about NEMA-17 motors having higher RPM than NEMA-23 .. "The small size of these motors enables much higher rotational speed (approximately 1800 RPM vs. 500 RPM with nema-23 motors), while providing more than sufficient torque."
[18:21:10] <jepler> is that correct/accurate?
[18:21:59] <SWPadnos> it's possible that it's accurate - larger/higher torque motors have (a) more inertia, which kills acceleration and (b) higher detent torque, which causese output torque to decrease more rapidly
[18:24:48] <cradek> 1/20 is too fine a pitch really. who needs 1/8000 inch half steps?
[18:26:52] <SWPadnos> me, me
[18:26:59] <SWPadnos> oh wait - no I don't
[18:27:26] <cradek> http://www.mcmaster.com/param/asp/PSearch2.asp?reqTyp=parametric&act=psearch&FAM=acmethread&FT_717=29734&session=desc=Acme%20Lead%20Screws;acmethread;717=29734&sesnextrep=531468682279949&ScreenWidth=1600&McMMainWidth=703
[18:27:30] <cradek> ouch, nice url
[18:30:37] <jepler> 3 turns per inch doesn't seem like very many -- that's less than 4 half-steps for the .003 depth of circuit boards, unless I got my math wrong
[18:31:26] <cradek> I meant 1/4-12
[18:31:42] <cradek> oh, I misunderstood it
[18:31:59] <jepler> you found one with 4 starts for whatever reason
[18:32:29] <skunkworks> our z axis input scale is 10K... 10 threads per inch 1:1 1000 steps per rev.
[18:32:34] <cradek> I guess it goes from 16 turns/inch to 4
[18:35:09] <SWPadnos> wimpy wimpy wimpy. one day, I'll have 40000 counts/inch on all 3 axes
[18:35:15] <SWPadnos> once I have any axes, that is :)
[18:36:47] <skunkworks> well... I will have 60960 counts per inch.. ;)
[18:36:59] <SWPadnos> that is a lot
[18:37:21] <skunkworks> 2540 line encoders 2:1 3 threads per inch.
[18:37:59] <SWPadnos> wow - 2540-line encoders are darned high res
[18:38:26] <SWPadnos> are you using those with the large motors/your big drive
[18:38:34] <SWPadnos> ?
[18:39:16] <jepler> skunkworks: speaking of your 10160-count encoder, did you ever test it on the newest pluto? I forget.
[18:40:06] <skunkworks> I have not tested the index yet.. but I played with it.
[18:40:51] <jepler> ok
[19:57:52] <Vq^> XerOniC!
[19:58:15] <XerOniC> Inte igår =)
[19:58:23] <Vq^> inte direkt :)
[19:58:47] <XerOniC> Allt bra med ADB-mannen? =)
[19:58:55] <Vq^> jadå, jag lever
[19:59:05] <XerOniC> Vad är detta för kanal?
[19:59:13] <XerOniC> Du brukade ju hänga på en annan innan
[19:59:21] <Vq^> #ironpeanut
[19:59:30] <Vq^> http://linuxcnc.org/
[20:09:33] <toast> rigging equipment for small machinery seems to be less expensive than i was anticipating
[21:53:48] <gene_> Hey guys, can you believe a bi-directional 37 ipm for the Z axiz on a micromill?
[21:54:42] <SWPadnos> not at the same time
[21:54:49] <SWPadnos> that's a bit like Schroedinger's cat :)
[21:55:02] <toast> maybe he's talking about two heads
[21:55:09] <gene_> I built some dampers using Jeff at xylotex's ideas. Smoother and quieter
[21:57:40] <gene_> I've often wondered how many lives that cat of Schroedinger's had. I know I've used up the first 9 :)
[21:57:55] <SWPadnos> must be at least 9 then :)
[21:58:29] <gene_> Either that, or he had an in at the pound for an eendless supply...
[21:59:22] <gene_> Only getting about 23 for the other two axis, but the screws are 20 tpi too.
[21:59:48] <jlmjvm> swpadnos:if you use gecko g320 drives and servo motors will emc be closed loop?
[21:59:48] <les_w> evening guys
[22:00:06] <gene_> And I can hang x out off teh right side of the table and bring it back now, couldn't before. Hi Les!
[22:00:10] <SWPadnos> jlmjvm, yes, though the geckos have a loop also
[22:00:16] <SWPadnos> les_w, long time no see!
[22:00:22] <les_w> hi swp
[22:00:36] <les_w> I have been pretty busy
[22:00:49] <jlmjvm> was wonderingif that was possible
[22:00:50] <SWPadnos> yeah - I had the opposite problem for a long time, now I have the overworked problem :
[22:01:27] <les_w> I have had a couple weeks off, but I am building a kitchen so still a lot of work
[22:01:34] <SWPadnos> heh
[22:02:04] <gene_> So am I, and its 90ish in the shop, so I'm headed for the shower and some chinese, take care guys.
[22:02:24] <les_w> bye gene...hot here too
[22:02:27] <SWPadnos> velly good - enjoy
[22:02:33] <jlmjvm> 106 here
[22:02:34] <gene_> hehe
[22:02:43] <les_w> yikes
[22:03:21] <jlmjvm> thats why i ben running my mill at nite
[22:04:30] <jlmjvm> been
[22:04:58] <les_w> Right now I just hang drywall and cement board and do electrics
[22:05:13] <les_w> a code kitchen needs a lot of circuits
[22:08:00] <les_w> I worked a lttle today...just talking on the phone. I have a very high speed data aquisition gig.
[22:17:42] <jlmjvm> swpadnos:im curious about the g320 and closed loop
[22:20:27] <SWPadnos> I have a USC, which acts like a velocity control
[22:20:40] <SWPadnos> you could get the same thing using freqgen instead of stepgen
[22:20:58] <SWPadnos> les_w, how high speed, and what resolution?
[22:21:05] <SWPadnos> (and for how long?)
[22:47:46] <JymmmEMC> http://www.google.com/search?q=weather+san+jose+ca
[22:47:48] <JymmmEMC> heh
[22:48:16] <JymmmEMC> a lil twisting of the knife
[22:48:47] <JymmmEMC> 77F
[23:13:22] <jlmjvm> that must be some really nice weather
[23:14:00] <toast> ff
[23:38:18] <Guest459> Guest459 is now known as skunkworks_
[23:54:10] <jmkasunich__> jmkasunich__ is now known as jmkasunich