#emc | Logs for 2009-04-01

[00:01:23] <leito> hello everybody, i was trying the 2.3 beta version and i have problem i don't know if anyone had the same problem, but the thing is that when i load a g code with a WHILE conditional the loading waits and waits and never finish... is this bug a common problem? thanks in advance for your help
[00:07:23] <jepler> usually it means you accidentally wrote a while loop that never terminates because the condition of the while loop is wrong.
[00:11:18] <leito> ok thanks jeff for your help now i will see that and i tell you if the problem persists or not... and another question that is not so important... in hardy when i use the live cd option i can create shortcuts with emc but when i install it on the hard drive ... i cannot . and when i use dapper this doesn't happens
[00:14:17] <leito> i was wondering if this is easy to solve because i will use it for a production machine so the idea is to get into the config as fast as linux can
[00:16:29] <jepler> I'm not familiar with this bug. I don't use the standard desktop so I can't test whether creating a desktop shortcut works ..
[00:16:39] <jepler> if it is important to you, please report it on our bug tracker.
[00:17:49] <leito> ok jeff, thanks for your help, i tried to load some example g code programs and it was ok so the problem was on my g code..
[00:18:35] <leito> sorry for all the questions
[00:23:22] <jepler> I wish I had all the answers for you
[00:25:45] <leito> no problem jeff, we are preparing the machine so we have time to explore the last doubts about the program.. but you are very helpful so there is no problem :)
[00:30:17] <leito> i hope soon when the machine is finished i can upload videos and pictures... we are changing a cylindrical grinder to make not only cylinders, the idea is to use it to make excentric shapes like crankshafts
[00:30:29] <jmkasunich> question - when using tool offsets from the tool table, the diameter units are always machine units, regardless of whether G20 or G21 is in effect
[00:30:41] <jmkasunich> what about when using G41.1 transient tool offsets?
[00:30:53] <jmkasunich> is the D word in machine units, or currently selected units?
[00:36:14] <jepler> jmkasunich: good question. I can probably find out what the implementation does in a minute or two..
[00:36:27] <cradek> g20/g21 units
[00:37:01] <jmkasunich> cradek: thanks
[00:37:15] <jmkasunich> it isn't in the manual (at least, not the current release manual)
[00:37:50] <jmkasunich> concave comp is in 2.3.0 beta, but not in 2.2.8, right?
[00:38:01] <cradek> right
[00:38:08] <jmkasunich> * jmkasunich debates installing the beta
[00:38:09] <cradek> if you are testing it, yay
[00:38:15] <cradek> do it do it do it
[00:38:22] <jmkasunich> "not for production"
[00:38:25] <cradek> bah
[00:38:37] <jmkasunich> I'm cutting openings in the top of multi-hundred $$ IGBT modules
[00:38:39] <cradek> revert if something big is wrong -- but it isn't
[00:39:02] <cradek> really? wild
[00:39:11] <cradek> what kind of packaging?
[00:39:54] <jmkasunich> http://www.infineon.com/cms/en/product/promopages/primepack/index.html
[00:40:45] <cradek> interesting - why?
[00:40:57] <jmkasunich> so the infrared camera can see how hot the chips are
[00:41:08] <cradek> neat
[00:41:30] <jmkasunich> I'm not cutting on the modules themselves, I just have lids
[00:41:39] <jmkasunich> once the cutting is done the lid and module will get joined
[00:42:00] <jmkasunich> but I have a limited supply of lids, no screwups allowed
[00:42:31] <jmkasunich> are there dapper packages for 2.3.0beta?
[00:42:55] <jmkasunich> actually, I'm probably better off compiling the 2.3 branch and doing RIP, that will co-exist better with what I have
[00:43:31] <jepler> jmkasunich: nobody's built dapper packages yet.
[00:43:40] <cradek> oh I forgot about that problem
[00:43:42] <jmkasunich> or, I say screw it and code corner arcs... I'd rather not be at this all night
[00:44:36] <jmkasunich> is there a list of config changes needed?
[00:45:10] <jepler> on the wiki
[00:46:00] <jepler> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?UpdatingConfigurationsForDevelopmentVersions
[00:46:07] <jmkasunich> yeah, thanks
[00:46:19] <jmkasunich> (was just digging thru my mail for the release notice, the links are there...)
[00:46:28] <jepler> looking at that list, you may not need to do anything
[00:46:45] <jepler> the stuff mentioned by dallur is to conform to the 2.3 style, not required changes
[00:47:57] <jmkasunich> have to do the NO_FORCE_HOMING thing, I only have switches on Y and Z, normally don't home X
[00:48:29] <jepler> lots of newbie users don't know about home-by-eye and soft limits
[00:48:36] <cradek> or you could just poke home - wouldn't matter
[00:48:55] <cradek> jepler: that's the one thing I wish everyone could know
[00:49:05] <cradek> how to use homing and work offsets and tool offsets
[00:49:17] <jmkasunich> I'm getting cold feet - I've already got the machine zeroed on the part, and it is so tempting to just do it
[00:49:31] <jmkasunich> "it" = cut the holes
[00:49:42] <cradek> isn't it super easy to just offset the square (?) yourself
[00:50:10] <jmkasunich> pretty much - its not exactly square, since I'm working around some terminals and stuff
[00:50:16] <cradek> I usually just didn't use cutter comp - wasn't worth the trouble
[00:50:17] <cradek> ah
[00:50:36] <jmkasunich> but I think thats what I'm gonna do
[00:50:42] <jmkasunich> testing will have to wait for another day
[00:51:09] <tomp> testing does not involve expensive parts ;)
[00:51:24] <cradek> that's true I bet
[00:52:55] <jmkasunich> doing the offsets by hand is a bit of a nuisance, since the part is dimensioned in mm and I have a 1/4" cutter
[00:53:05] <cradek> oogh
[00:53:27] <jmkasunich> (hence the G20/21 question)... although I'm writing the program by measuring stuff with dial calipers, I can just do it all in inches
[00:53:28] <cradek> bet I could do it for you in a few seconds
[00:53:49] <jmkasunich> you probably could, if I had a nicely drawn path
[00:53:57] <jmkasunich> but I'm doing this incrementally (at least on the first part)
[00:54:06] <jmkasunich> rough it out, measure, tweak, etc
[00:54:16] <jmkasunich> this is a complex piece of plastic
[00:55:19] <jmkasunich> my favorite technique for stuff like this is O100 if [1 EQ 1] some-stuff O100 endif O101 if [0 EQ 1] more-stuff O101 endif, etc
[00:55:48] <jmkasunich> then on the next part, I make all the zeros into ones and it just runs
[00:57:58] <tomp> does IF[0] evaluate true?
[00:58:12] <tomp> less typing than 1eq0
[00:58:14] <cradek> surely note
[00:58:16] <cradek> not
[00:58:32] <cradek> well that might be an error - no operator
[01:04:47] <jepler> experiment and use the preview plot. something like this will help you see what IF[0] does, and what [0 EQ 1] means .. http://pastebin.ca/1378501
[01:05:28] <jepler> IF [0] and IF [1] are accepted juts fine
[01:05:41] <jepler> s/uts/ust/
[01:15:14] <tomp> thx
[01:15:29] <tomp> 'just ask the machine' ;)
[01:18:05] <jmkasunich> hmm, do I dare to climb mill...
[01:18:22] <jmkasunich> plastic - good
[01:18:29] <jmkasunich> large stickout - bad
[01:19:04] <cradek> wonder if it's sticky melty plastic or good plastic
[01:19:26] <jmkasunich> good
[01:19:31] <jmkasunich> glass filled
[01:19:37] <jmkasunich> I've already milled a slot
[01:19:40] <cradek> ah
[01:19:50] <jmkasunich> cutting a slot, climb vs. conventional doesn't exist
[01:19:53] <cradek> carbide?
[01:19:56] <jmkasunich> yes
[01:20:06] <jmkasunich> 1/4" diameter, 2" stickout
[01:20:11] <jmkasunich> damn short Z travel again
[01:20:31] <jmkasunich> the glass would eat HSS for lunch
[01:20:32] <cradek> wow 8x stickout...
[01:20:37] <cradek> right
[01:20:41] <jmkasunich> 0.1" DOC so far
[01:20:51] <jmkasunich> 4 flute, 0.001 per flute
[01:20:56] <cradek> speed cranked way up I assume?
[01:21:01] <jmkasunich> 1340 RPM
[01:21:15] <jmkasunich> the machine could go faster, but I don't want to risk meltage
[01:21:57] <cradek> I bet you won't break it even if you unluckily get an occasional .004
[01:22:12] <jmkasunich> the machine has about 0.010 of lash
[01:22:28] <jmkasunich> it's compensated, but that won't help if the cut drags the table around
[01:22:28] <cradek> hmm, ouch
[01:22:40] <tomp> snug the gibs, cut one direction
[01:22:40] <cradek> yeah, and 2" is awfully long
[01:22:56] <jmkasunich> wish I could reduce it, but its tough
[01:23:13] <tomp> cnc classic ladder gib snuggers
[01:24:33] <tomp> off to (hopefully) last visa extension
[01:24:36] <tomp> bye
[01:33:54] <jmkasunich> man, naive cam detector can do some strange things
[01:34:13] <jmkasunich> when I was originally planning to do things in mm, I set G64P0.1 (= 0.004")
[01:34:23] <jmkasunich> later I changed to inches, and didn't fix the P
[01:34:45] <jmkasunich> then I commanded G0X0Y0; G1Y0.0875; G1 X4.0
[01:34:51] <jmkasunich> it merged the two G1 moves
[01:35:18] <jmkasunich> the preview shows the two moves
[01:37:18] <cradek> that sounds like a bug if tolernace changed from .1mm to .1in
[01:39:50] <jepler> (preview never shows the results of the naive cam detector)
[01:40:13] <jmkasunich> not a bug
[01:40:35] <jmkasunich> I change the G21 to G20, and didn't change the G64P0.1 to G64P0.004
[01:40:43] <jmkasunich> so it did what I told it to do
[01:40:59] <cradek> oh you put g20 above g64p0.1
[01:41:04] <jmkasunich> yeah
[01:41:08] <cradek> oops
[01:41:20] <cradek> bbl
[02:32:36] <jmkasunich> #1 done
[02:34:45] <jmkasunich> starting #2
[02:35:58] <skunkworks> looks like I can run 15a thru this all day. I cannot tell of the devices are getting hot or the circuit board traces.
[02:36:24] <skunkworks> jmkasunich: It only took a few years - but I think I have it. Or close enough for what I need. ;)
[02:37:01] <skunkworks> And Hot - I mean barely warm to the touch.
[02:38:59] <SWPLinux> skunkworks: if you have avideo camera, you may be able to get an IR filter for it
[02:39:16] <SWPLinux> and take slightly informative thermal videos
[02:39:39] <SWPLinux> plus you can see under certain kinds of clothing :)
[02:40:23] <skunkworks> interesting
[02:40:51] <SWPLinux> they won't be "scientific grade", but should give you an idea of hotspots
[02:41:08] <skunkworks> heh
[02:42:08] <skunkworks> really - the whole thing is just getting warm. without a fan. The heatsink doesn't seem to get warm enough to tell the difference between abient
[02:42:38] <SWPLinux> well, you can test a little bit by sticking a smaller heatsink on it
[02:42:50] <SWPLinux> see how much hotter it gets
[02:45:09] <skunkworks> * skunkworks is excited
[02:45:12] <SWPLinux> oh, if you have a video camera with a "night shot" mode (that's Sony's name I think), then that should be all you need
[02:45:29] <SWPLinux> yeah, it sounds like you've got it working pretty darned well
[02:45:31] <skunkworks> I do have one.. but I don't think it works anymore :(
[02:45:34] <SWPLinux> heh
[02:46:15] <skunkworks> the 12inch 14gage wires going to the motor get warm
[02:47:10] <SWPLinux> cool :)
[02:53:21] <jmkasunich> I'd be astonished if a video camera can measure _thermal_ infrared
[02:53:39] <SWPLinux> yeah, dunno how far down they go
[02:53:41] <jmkasunich> that is long-wave IR, as opposed to short wave IR that is emitted by IR LEDs, etc
[02:54:10] <SWPLinux> simple to try if you have a suitable camera though
[02:54:21] <jmkasunich> long wave IR cameras have lenses made of germanium or some such
[02:54:28] <jmkasunich> they look like mirrors in visible light
[02:54:44] <SWPLinux> yeah, that does help with transmission, but it's not necessary
[02:55:06] <SWPLinux> greenish mirrors :)
[02:55:13] <SWPLinux> red-green actually, IIRC
[02:56:39] <jmkasunich> we have this: http://www.electrophysics.com/Browse/Brw_ProductLineCategory.asp?CategoryId=184&Area=TG
[02:56:47] <jmkasunich> or at least, its very similar predecessor
[02:57:25] <jmkasunich> $20K+ for 320x240
[02:57:37] <SWPLinux> does it have an internal Dewar cooler?
[02:57:40] <SWPLinux> seems small for that
[02:57:50] <jmkasunich> if ordinary camera technology could do thermal IR, these folks would be out of business
[02:58:02] <jmkasunich> no, the last couple generations use room temp sensors
[02:58:07] <SWPLinux> huh
[02:58:21] <SWPLinux> yeah, that's probably true
[02:58:47] <SWPLinux> I'm sure you can't get anything close to real data, but I figured it might be possible to see something
[02:59:16] <SWPLinux> the military systems surprised me. we could see mice in the field outside, through a window
[02:59:22] <jmkasunich> http://jmkasunich.com/pics/IR-outside-northeast.gif
[02:59:52] <jmkasunich> http://jmkasunich.com/pics/back-door.jpg
[02:59:54] <SWPLinux> yep, I remember those :)
[02:59:57] <jmkasunich> shitty weatherstripping
[03:00:26] <jmkasunich> http://jmkasunich.com/pics/hot-dog.jpg
[03:00:29] <jmkasunich> cold nose
[03:00:39] <SWPLinux> heh
[03:00:48] <SWPLinux> and lazzor eyes
[03:00:55] <jmkasunich> #2 done (a minute or so ago)
[03:01:10] <jmkasunich> about 24 mins of machine time
[03:01:14] <jmkasunich> for #2
[03:01:17] <SWPLinux> slits?
[03:01:24] <jmkasunich> 2-1/2 hours or more for #1, with programming, etc
[03:01:50] <jmkasunich> dunno how far you read back - I modified the tops of some IGBT modules so the IR camera can see inside
[03:02:32] <SWPLinux> oh, not far. I don't have a screen session set up yet, so I only see stuff since I logged on (when I'm on the road anyway)
[03:02:49] <jmkasunich> these: http://www.infineon.com/cms/en/product/promopages/primepack/index.html
[03:02:56] <skunkworks> wow - that is cool
[03:05:01] <jmkasunich> skunkworks: what is? the IR cam?
[03:05:07] <skunkworks> I got these really cheap. http://www.irf.com/product-info/datasheets/data/irg4pf50wd.pdf
[03:05:15] <skunkworks> jmkasunich: the ir cam
[03:05:38] <jmkasunich> skunkworks: what are you gonna use the IGBTs for?
[03:06:14] <skunkworks> replace my mosfets that had way too high of Rds. The mosfets will work for the puma though. <10a
[03:06:36] <jmkasunich> the IGBTs aren't neccessarily gonna be better
[03:06:36] <skunkworks> that is what I have in the h-bridge at the moment. (they don't seem to get warm)
[03:06:48] <jmkasunich> IGBTs in there now?
[03:06:51] <skunkworks> yes
[03:06:58] <jmkasunich> what switching frequency?
[03:07:04] <skunkworks> 20khz
[03:07:13] <jmkasunich> I'm somewhat impressed
[03:07:49] <skunkworks> Looking at the graph - so am I. but they seem to perform very well so far.
[03:07:56] <jmkasunich> 2.25V / 20A = equivalent Rds-on of 0.112 ohms
[03:08:01] <jmkasunich> which isn't all that impressive
[03:08:04] <jmkasunich> its good for a high voltage part
[03:08:17] <jmkasunich> but I thought you could get much better in a modest voltage mosfet
[03:08:35] <jmkasunich> at low voltages, mosfets kick IGBTs ass
[03:10:09] <jmkasunich> what is your bus voltage again?
[03:10:24] <skunkworks> That is how I understand it. But the mosfets I had really had issues with higher current. I had actually liquified a pair at around 15a.
[03:10:30] <skunkworks> around 150v
[03:10:35] <skunkworks> 150v
[03:10:52] <jmkasunich> you are running 150V now, or that is your design target?
[03:11:03] <skunkworks> design target - right now 90v
[03:11:11] <jmkasunich> still, not bad at all
[03:11:29] <jmkasunich> switching loss will increase with voltage, but if you are already at 90 and things are cool, you should be OK
[03:11:51] <jmkasunich> the big stuff I work with has prohibitive switching losses well before 20KHz
[03:11:55] <skunkworks> I am very happy with it. I did add the blanking circuit (blanks current sense for 6us) which really helped stablize the current limit
[03:12:11] <jmkasunich> like, derate by 25% or more at 4KHz ;-(
[03:12:54] <skunkworks> the graph on the pdf shows that FIG1
[03:13:03] <skunkworks> similar
[03:14:09] <skunkworks> at 20khz they show maybe 12a.
[03:14:26] <jmkasunich> just for grins, compare with http://www.infineon.com/cms/services/dgdl?folderId=db3a3043133ffd30011344bfcfd902eb&fileId=db3a30431689f4420116d402010b0c9f
[03:16:10] <jmkasunich> input C: 3.3nF vs 82nF
[03:16:46] <skunkworks> heh
[03:16:59] <jmkasunich> actually, that is less difference than I expected
[03:17:05] <jmkasunich> those are some beefy IGBTs you have
[03:17:17] <jmkasunich> at least as discretes go
[03:17:48] <jmkasunich> these modules have 12 IGBT and 12 diode chips in parallel (per switch, two switches per module)
[03:19:24] <skunkworks> very neat. I cannot imagine getting the gate drivers close enough ;)
[03:19:59] <jmkasunich> we mount them direct to the module
[03:20:53] <jmkasunich> and our cap bank makes your cap look puny
[03:21:14] <skunkworks> heh
[03:22:04] <skunkworks> http://cgi.ebay.com/INTERNATIONAL-RECTIFIER-G4PF50WD-LOT-OF-4-NEW_W0QQitemZ250303186210QQcmdZViewItemQQptZLH_DefaultDomain_0?hash=item250303186210&_trksid=p3286.c0.m14&_trkparms=72%3A1205%7C66%3A2%7C65%3A12%7C39%3A1%7C240%3A1318%7C301%3A1%7C293%3A1%7C294%3A50
[03:22:14] <skunkworks> 4 for 3.49
[03:22:23] <jmkasunich> that's what you got?
[03:22:30] <skunkworks> yes
[03:22:35] <jmkasunich> nice
[03:22:43] <skunkworks> he sent the wrong ones first. but he made it right
[03:22:57] <eric_unterhause1> what's with the three knifes instead of leads?
[03:22:58] <jmkasunich> that's something like $0.02 per amp
[03:23:19] <jmkasunich> eric_unterhause1: that one pic does look weird doesn't it
[03:23:39] <skunkworks> heh - I could always lower the switching frequency if need be
[03:23:50] <skunkworks> but so far - I think it may be a keeper
[03:24:03] <jmkasunich> sounds that way
[03:24:26] <skunkworks> the current sense resistor is only rated at 22a anyways.
[03:24:30] <jmkasunich> well, I'm going to try to go to bed at a sane time
[03:24:34] <jmkasunich> goodnight
[03:24:39] <skunkworks> same here - night
[04:58:05] <LawrenceG> skunkworks, are you around still?
[05:30:29] <fenn_> fenn_ is now known as fenn
[07:33:29] <alex_joni> http://www.google.com.au/intl/en/gball/
[07:37:34] <JymmmEMC> damnit, I wanted one!
[07:37:50] <fenn> beats the hell out of CADIE
[07:39:32] <alex_joni> fenn: although 3D browser sounds nice too :D
[07:39:39] <alex_joni> http://www.google.com/intl/en/landing/chrome/cadie/
[07:40:30] <JymmmEMC> Boob-e-vision
[07:44:25] <alex_joni> heh: http://www.google.com/mail/help/autopilot/index.html
[08:09:59] <micges> good morning
[08:11:20] <micges> I have question: when gcode program enabled flood and mist and run to the end of file which is % not M2 will flood and mist should turns off or not ?
[08:28:28] <alex_joni> micges: I don't think the spec specifies this
[08:28:37] <alex_joni> it does say that M2 turns it off..
[08:28:43] <alex_joni> not sure about %%
[08:36:48] <fenn> cradek: another clock idea http://filer.case.edu/jec24/uploaded_images/panel-meter-clock-758892.jpg
[08:37:20] <archivist> good one fenn
[09:06:46] <piasdom> good mornin all
[09:17:27] <alex_joni> hi
[09:28:12] <archivist> hehe http://www.youtube.com/watch?v=C740zS9R9kk&flip=1
[09:28:39] <archivist> pjm_, see above
[11:09:22] <tomp> wow flip=1, cool
[11:12:09] <tomp> and the tapping is cool of course ( carried away by the trick)
[12:50:48] <pjm_> archivist heh i like the mirrored look!
[12:51:11] <archivist> good one today !
[12:52:56] <pjm_> ah yes april fools day
[12:53:22] <pjm_> i liked the UK governments 'gag' with their 2p price increase on petrol, that made me laugh
[12:53:41] <archivist> http://www.google.com/tisp/
[12:53:55] <archivist> cry!
[12:54:08] <archivist> bar stewards
[12:54:44] <pjm_> hehh
[12:55:09] <archivist> my wages are static!
[13:00:29] <pjm_> LOL: http://mail.google.com/mail/help/autopilot/index.html
[13:00:34] <pjm_> i need that ^^^^
[13:03:58] <archivist> just had to retrain a human because he was talking to a bot :)
[13:09:53] <skunkworks_> archivist: mesuring the current at the motor - rock solid.
[13:10:19] <fenn> so your electronic space heater works?
[13:10:27] <archivist> :)
[13:10:35] <skunkworks_> suprisingly really cool.
[13:11:00] <fenn> 15A*90V = 1350W or slightly less than a space heater
[13:11:26] <fenn> oh, but it's not actually running full duty cycle
[13:11:26] <skunkworks_> * skunkworks_ never really touched the motor..
[13:11:34] <archivist> average power though
[13:15:13] <skunkworks_> http://imagebin.ca/img/4Gy9os2i.png
[13:17:03] <skunkworks_> I still have to put all the correct values back in the circuit
[13:20:24] <skunkworks_> Couple things I am going to live with... 1.) A little hard to test the current sense voltage with the cap on (not impossible). 2.) the heatsink needs clearence holes thru it to mount the cap.
[13:22:28] <skunkworks_> But I packed the power traces as close togather as I could.
[13:36:38] <fenn> what's the max current?
[13:40:10] <fenn> * fenn guesses 20A from the 7W resistor
[13:48:58] <skunkworks_> fenn: yes - the resistor has a max 22a rating
[13:52:21] <skunkworks_> I could put one on the back of the board.... :)
[13:52:47] <skunkworks_> I have to mod the headsink anyways - just mill a slot for it.
[13:55:17] <jepler> skunkworks_: it looks like you could route a red trace from the top left terminal of the sense resistor up to the edge of the board for a test point
[14:02:13] <skunkworks_> jepler: I need to measure the voltage at pin 2 of the comparator (ic7) which goes to the left side of r15/16
[14:03:10] <skunkworks_> I am sure I could sneak it out somewhere
[14:04:08] <skunkworks_> I could sneak it out to the top side of the trimmer pot (on the top of the circuit board)
[14:04:44] <jepler> * jepler looks for ic2
[14:05:01] <jepler> oh you said ic7
[14:05:02] <jepler> pin 2
[14:05:34] <jepler> you want to measure the filtered current sense, not raw
[14:05:35] <jepler> ?
[14:06:04] <skunkworks_> I want to measure the reference voltage the comparator is using to compare to the raw.
[14:06:57] <skunkworks_> I probably am not making sense - I need to measure the reference voltage that the comparator is using to set the trip point.
[14:07:09] <jepler> oh, hm, I should read what you say
[14:09:00] <skunkworks_> (which in turn is the current limit)
[14:18:10] <jepler> have a schematic link handy?
[14:21:13] <skunkworks_> do you want a picture or the actual eagle file?
[14:21:26] <jepler> a picture is fine
[14:23:08] <skunkworks_> http://imagebin.ca/img/LbMnIL.png
[14:23:14] <skunkworks_> some values are not in yet.
[14:23:23] <skunkworks_> and it is really messy - sorry
[14:25:58] <skunkworks_> The flipflop one-shot is set to about 6us
[14:26:37] <skunkworks_> that activates the transistor and 'blanks' the current sence for that amount of time (begining of each pwm cycle)
[14:31:02] <jepler> 0 Results
[14:31:03] <jepler> Sorry the amount of criteria chosen does not exactly match IR product.
[14:31:51] <jepler> (just making fun of the terrible language)
[14:31:58] <skunkworks_> * skunkworks_ really should go over the circuit..
[14:32:35] <skunkworks_> jepler: I am using these http://www.irf.com/product-info/datasheets/data/irg4pf50wd.pdf
[14:40:31] <jepler> skunkworks_: let me know when you run them at 900V
[14:40:39] <skunkworks_> heh
[14:40:46] <tom2> how big is a to-247a pkg? that sounds like a ot of power and looks like a 220
[14:41:14] <jepler> Vce(on) = 2.25V - doesn't that mean you get 45W in the transistor when it's conducting 20A?
[14:41:51] <skunkworks_> jepler: yes - but don't ask me why it doesn't seem to get warm.
[14:42:23] <jepler> "HEXFRED(TM) diodes"
[14:42:33] <skunkworks_> It is hooked to a nice heatsink at the moment
[14:43:39] <skunkworks_> tom2: it is about .6"X.8"
[14:44:17] <jepler> there's a dimensioned drawing on page 10 of that pdf
[14:44:24] <tom2> thx, small , good idea to have your nice heat sink
[14:45:43] <skunkworks_> I have not acutally measured the Vce though
[14:53:36] <tom2> freshmeat sez the new gEDA uses Cairo to render, so no jaggies on schematics
[17:16:00] <LawrenceG> skunkworks, you about?
[17:30:31] <jensor> Does halui.joint.0.is-selected function. Using the hal meter on it continues to show false even though I select in axis.
[17:31:18] <SWPLinux> axis should have its own "selected" output
[17:31:37] <SWPLinux> doing something in one UI (Axis) doesn't have any direct effect on any other UI (halui)
[17:32:26] <jensor> I thought if i select an axis in the axis ui that I could have a pin available to me to indicate so
[17:32:42] <jensor> I could illuminat a lamp on a control panel
[17:32:48] <SWPLinux> I think that's true, but it would come from axis - not halui
[17:33:24] <jensor> I haven't been able to find such a signal aywhere
[17:33:26] <BJT-Work> are you selecting axis 0
[17:33:32] <jensor> yes
[17:33:44] <jensor> joint 0
[17:34:10] <BJT-Work> sounds like it should work then
[17:34:18] <jensor> I thought so too
[17:35:06] <BJT-Work> I don't have any linux machines here so I can't test it
[17:35:13] <SWPLinux> one sec - I'm checking
[17:35:33] <SWPLinux> I just have to fix up some merge errors first :)
[17:39:47] <SWPLinux> ok, it looks like the pins are called "axisui.jog.x" (for any of xyzabcuvw)
[17:44:28] <BJT-Work> jensor: are you using HAL meter or HAL Configuration Window to view the pin?
[17:45:43] <SWPLinux> it doesn't matter, that pin won't change based on anything you do in AXIS
[17:45:51] <jensor> halmater
[17:46:01] <jensor> Hal meter
[17:46:34] <SWPLinux> AXIS has its outputs, telling which axis is selected for jogging in AXIS, and halui has its outputs telling which joint(s) is(are) selected for jogging with halui
[17:47:21] <SWPLinux> those can be different joints/axes - conceptually, you can hold a key in AXIS and jog X while turning a handwheel and jogging Y with halui
[17:47:55] <SWPLinux> in which case, axisui.jog.x and halui.joint.1.is-selected will both be true
[17:48:48] <jensor> I can change halui.estop.is-activated in axis
[17:49:07] <SWPLinux> estop state is a machine-wide status
[17:49:23] <Optic> has anyone used one of those griffin iknob things for jogging? :)
[17:49:36] <SWPLinux> if you run a program or turn coolant on with one UI, that affects machine state
[17:50:01] <SWPLinux> if you make a selection in one program as to what that program will do when you pess the "+" key, that has no effect on machine state
[17:50:17] <SWPLinux> Optic: no, but I've thought about it for other things
[17:50:22] <jensor> I see
[17:50:26] <SWPLinux> in general, userspace jogging sucks
[17:50:48] <SWPLinux> you're much better off with a real encoder connected to a couple of GPIO pins
[17:50:59] <jensor> What do you mean user space?
[17:51:11] <SWPLinux> non-realtime
[17:51:15] <jensor> oh
[17:54:27] <BJT-Work> I didn't catch on that jensor was selecting an axis in AXIS good catch SWPLinux
[17:54:46] <SWPLinux> I've had my coffee today :)
[17:54:59] <BJT-Work> I'm ready for a beer
[17:55:13] <skunkworks_> LawrenceG: yes?
[18:08:47] <jensor> I tried "net xis selected axisui,jog.x parport.1.pin-02-out" and get debug msg "axisui.jog.x does not exist.
[18:09:22] <cradek> you'd need to connect that in a postgui hal file, since the axisui pins don't exist until AXIS is running
[18:09:49] <jensor> I am using Axis
[18:10:15] <jensor> It is in the ini file
[18:11:00] <cradek> http://www.linuxcnc.org/docs/devel/html/gui_axis.html#sub:Physical-jog-wheels
[18:20:48] <jensor> cradek The way I understand the reference is - In the HAL section put statement "POST_HALFILE=name.hal" then createfile name.hal with the statement: xisselected axisui.jog.x parport....etc
[18:20:54] <jensor> Is that correct?
[18:21:23] <SWPLinux> yes (assuming you correct the typos :) )
[18:22:07] <jensor> only thing approaching a typo i see is parport...etc
[18:22:24] <SWPLinux> POSTGUI_HALFILE
[18:22:32] <jensor> oh thanks
[18:22:35] <SWPLinux> :)
[18:22:42] <LawrenceG> skunkworks, I am interested in what you ended up with for the current monitoring/blanking... is the schematic available?
[18:44:35] <jensor> cradek and SWPLinux thanks very much - Works fine now!
[18:44:40] <cradek> yay
[18:49:12] <skunkworks_> LawrenceG: http://imagebin.ca/img/LbMnIL.png
[18:49:23] <skunkworks_> Not all the componants are correct
[18:55:56] <skunkworks_> gives you the big picture though ;)
[18:59:18] <skunkworks_> The blanking should be long enough to mask the mosfet charge spike.
[18:59:30] <LawrenceG> skunkworks, thankyou... I am still playing with a smaller drive here ... looking for really cheap options!
[18:59:45] <skunkworks_> for this circuit (after the on-delay of the driver ic - is around 6us)
[19:00:08] <skunkworks_> the h-bridge is really designed for 20khz.
[19:00:58] <skunkworks_> LawrenceG: I am using igbt's
[19:01:49] <LawrenceG> cool.. I have 2 different requirements in my drive.... a current fault detection which throws the drive offline and an analog 0-5v current feedback.... I dont have the cycle by cycle logic you have in your drive
[19:03:18] <skunkworks_> Cool
[19:03:53] <skunkworks_> there is on extra opto-isolator on mine that I have not decided what to do with.
[19:04:14] <LawrenceG> it is a little tricky ( as you found out) getting real current readings while ignoring all the noise
[19:04:24] <skunkworks_> I bet.
[19:04:52] <skunkworks_> the circuit really seems to work well with the blanking - you don't have to worry about filtering the turn-on spike.
[19:05:49] <LawrenceG> I am trying to keep the whole power stage at < $15 for a 15A 70V drive
[19:09:01] <LawrenceG> the shunt resistor is one of the big costs.... looking at using a loop of brass shim stock
[19:09:10] <skunkworks_> cool
[19:09:42] <skunkworks_> trace on the circuit board?
[19:09:58] <skunkworks_> there is a fine line between fuse and resistor then.. ;)
[19:10:23] <LawrenceG> tolerance is not really much of an issue..... its either "holy shit the smoke is leaking out" or "I am just warm and cozy"
[19:11:31] <skunkworks_> the guy on cnczone took your code and ran with it. Now he is making a step/dir interface for servodrives.
[19:11:42] <skunkworks_> or did you see that? I don't remember
[19:11:42] <LawrenceG> yea, high current like 15A and greater is hard to do on a PCB
[19:11:42] <cradek> git version
[19:11:53] <cradek> what
[19:11:57] <skunkworks_> heh
[19:12:43] <skunkworks_> LawrenceG: the picture I sent yesterday: the meter read 19.9A.
[19:13:10] <LawrenceG> yes... there are several guys that have made mods.... one of them in France has done a nice job as a uni project. It looks like step dir or quadrature can be supported as a configuration option
[19:13:30] <LawrenceG> There is only a few lines of code difference
[19:13:47] <LawrenceG> nice
[19:14:35] <skunkworks_> LawrenceG: layout at the moment. http://imagebin.ca/img/4Gy9os2i.png
[19:15:19] <LawrenceG> I priced the real cost of the OPA549 I was using and decided a discrete output stage could be cheaper.... problem when building with samples!
[19:16:53] <skunkworks_> I got a good deal on the igbts - bought a bunch for now.
[19:17:19] <LawrenceG> cute.... the big cap gives it a very industrial look.... that will drive some serious servos (or small Honda's)
[19:17:53] <skunkworks_> I am hoping. If I put one resistor on the top and one on the bottom - I could pus 40a ;) Poof.
[19:18:17] <skunkworks_> push
[19:18:41] <skunkworks_> *sense resistorr
[19:18:56] <LawrenceG> understood
[19:21:06] <LawrenceG> the sense resistor could probably just become the supply lead ( monitor drop across the negative return lead to the power supply)
[19:21:59] <LawrenceG> oops.. power fail here... not sure how long the UPS will last.... attempted restart also tripped out power, so it may be down for a while
[19:22:16] <skunkworks_> yeck
[19:22:40] <LawrenceG> its been very windy here, so there is likely a tree across the primary line
[19:23:36] <LawrenceG> oh well a good test cycle for the UPS batteries
[19:24:37] <LawrenceG> I dont think my DVD rip will complete before the batteries die :{
[19:26:23] <BJT-Work> * BJT-Work heads out to work on the Massey Harris Ferguson F40
[19:35:23] <issy> any idea about open source dxf library? thanks
[19:35:40] <cradek> the one from qcad
[19:36:26] <issy> yep , is thill ver.14 and is not free
[19:37:39] <cradek> it is dual license with gpl being one of the options
[19:38:05] <issy> another?
[19:38:40] <cradek> http://www.qcad.org/dxflib_doc_faq.html
[19:39:47] <issy> i want to make convertor from dxf to gcode inside the emc , with visualisation of the dxf
[19:40:39] <archivist> issy have you seen HeeksCNC
[19:40:48] <issy> no
[19:40:54] <issy> what is this?
[19:41:24] <archivist> http://code.google.com/p/heekscnc/
[19:41:36] <archivist> see also heekscad
[19:44:14] <alex_joni> issy:|en&u=http://www.christian-kohloeffel.homepage.t-online.de/dxf2gocde.html
[19:44:57] <issy> thanks
[19:45:19] <alex_joni> or here (in german) http://www.christian-kohloeffel.homepage.t-online.de/dxf2gocde.html
[19:46:44] <issy> look at dakeng.com the soft name is ace , liiks perfect (not the last one)
[19:46:54] <issy> www.dakeng.com
[19:55:33] <issy> well is borland c , but is writen with api , so it is easy to be used with any other compiler , it will take me 15 minutes
[20:02:31] <issy> well , it is ok for msvs , working
[20:09:03] <skunkworks_> LawrenceG: power back on?
[20:09:04] <skunkworks_> ;)
[20:15:34] <LawrenceG> yup...
[20:16:11] <LawrenceG> ate a can of beans for lunch while waiting to see the light
[20:24:30] <LawrenceG> skunkworks, any idea if I could use the second half of the lm393 as a crappy op amp if I added feedback?.. I need 1 OA and 1 comparator
[20:25:06] <skunkworks_> LawrenceG: I have seen people do that.. I would breadboard it ;)
[20:25:26] <LawrenceG> yea... not sure what to expect!
[20:33:44] <skunkworks_> LawrenceG: http://encon.fke.utm.my/nikd/latest/sloa067.pdf
[20:33:45] <skunkworks_> ;)
[20:35:43] <JymmmEMC> http://codepad.org/ is awesome!!! Will run code too
[21:06:36] <skunkworks_> LawrenceG: http://focus.ti.com/lit/ds/symlink/tlv2302.pdf
[21:06:41] <skunkworks_> one of each ;)
[21:10:39] <eric_unterhause1> that's interesting
[21:22:27] <JymmmEMC> Ya know, it really annoys me that sites FORCE a pdf download instead of letting a user select view/download.
[21:39:02] <jepler> skunkworks: hm page 8 of the ti pdf reminds me of your schematic -- you do the option labeled "NEVER!" with your unused op-amp
[22:07:06] <skunkworks> jepler: the datasheet for the 393 tells you to ground the unused inputs
[22:07:23] <skunkworks> I saw that too
[22:28:20] <skunkworks> jepler: All input pins of any unused comparators should be tied to
[22:28:34] <skunkworks> the negative supply
[22:34:58] <eric_unterhausen> wow, I downloaded the parts list for one of my vfds from Baldor, the fan costs more than I paid for the drive
[22:36:25] <skunkworks> heh
[22:36:29] <skunkworks> not suprising
[22:47:35] <eric_unterhausen> I should stay away from the internet on april 1st, my sense of humor doesn't extend to stuff that isn't funny
[23:07:55] <eric_unterhausen> is a 2 jaw lathe chuck some kind of an ongoing April fools joke?
[23:08:49] <jepler> skunkworks: I guess that was the advice of the op-amp section, not the comparator section..
[23:10:11] <eric_unterhausen> seem like either one would have problems with that?
[23:11:00] <eric_unterhausen> it seems that if you want something cheap from enco, look for it on ebay and they give you a much better price