#emc | Logs for 2006-04-19

[02:07:32] <fenn> the scrapyard giveth and the scrapyard taketh away: http://fenn.dyndns.org/pub/camera/DCP_0616.JPG
[02:10:52] <jmkasunich> what got given and what got taken?
[02:11:24] <skunkworks> I was wondering the same thing.
[02:12:20] <skunkworks> the scrapyard givith and fenn taketh away?
[02:12:59] <fenn> i got 20 feet of thick polished stainless tubing and a 1 horse dc motor with worm drive gearbox
[02:13:17] <fenn> last week we threw out 2 tons of misc. junk that had accumulated in the driveway
[02:13:26] <skunkworks> jmk - it was the seconds to nanosecond issue with my emc2. fixed it
[02:16:35] <jmkasunich> skunkworks: good to hear that
[02:16:46] <fenn> gonna try to run EMC on these: http://fenn.dyndns.org/pub/camera/DCP_0615.JPG
[02:17:27] <SWPadnos> is that a strange power supply connector on the left?
[02:17:35] <fenn> yes
[02:17:38] <SWPadnos> bummer
[02:18:00] <fenn> all the colored wires go to a 6 pin header like you see on old motherboards
[02:18:02] <SWPadnos> good thing you got the supplies too
[02:18:07] <fenn> yeah definitely
[02:18:33] <SWPadnos> too bad it's not ATX - that would be ideal for a picoPSU
[02:18:37] <Jymmm> fenn what are those?
[02:18:42] <fenn> the p/s says atx on it
[02:18:50] <SWPadnos> hmmm
[02:19:07] <fenn> jymmm they are netier xl1000's
[02:19:20] <fenn> fairly cheap on ebay
[02:19:54] <SWPadnos> the thing that may kill it for EMC is the integrated video
[02:20:10] <SWPadnos> (plus 64M RAM and k6-2/300)
[02:20:24] <SWPadnos> but it should be an interesting experiment
[02:20:24] <fenn> you can add more ram, its standard low profile sdram
[02:20:42] <SWPadnos> ok, that's cool
[02:21:00] <fenn> i figured it wouldn't have all the laptop-ish issues with apm at least
[02:21:28] <SWPadnos> probably not, but it's possible. it has power management (due to wake on LAN, I think)
[02:21:56] <fenn> netier (now wyse) provides practically zero information on it
[02:22:10] <skunkworks> I think I ran emc on a pcchips motherboard with sis video
[02:22:56] <skunkworks> that was a k6-2 /300 if i recall correctly
[02:23:10] <SWPadnos> I'd like to try compiling for the Excalibur FPGA development kit, which has a PowerPC core in it (and runs Linux)
[02:23:30] <fenn> * fenn just shakes his head
[02:23:40] <SWPadnos> heh
[02:23:57] <fenn> the whole fpga scene pisses me off
[02:24:12] <SWPadnos> because it;s too expensive?
[02:24:21] <fenn> we would see so much cool stuff if the manufacturers were more open about the designs
[02:24:31] <fenn> but instead they go and have a patent hissy fit
[02:25:00] <SWPadnos> yeah - the whole IP thing is frustrating
[02:25:10] <SWPadnos> it's pretty amazing how expensive some of this development software is
[02:25:37] <SWPadnos> there was a C-to-FPGA compiler at ESC - they wanted $45k for the base package
[02:25:49] <SWPadnos> add $15k for the image processing toolkit
[02:26:19] <SWPadnos> and I think you still need the vendor tools to place and route for a particular cihp
[02:26:47] <Jymmm> atmel doens't have anything like a fpga?
[02:27:06] <SWPadnos> they sure do
[02:27:21] <SWPadnos> there's even a series of AVR chips with some FGA fabric around them (FPSLIC)
[02:27:23] <Jymmm> same "toolkit" expenses?
[02:27:24] <SWPadnos> FPGA
[02:27:35] <SWPadnos> sort of.
[02:27:55] <SWPadnos> all of the chip vendors have a free version of their tools, which are limited in some way
[02:28:18] <fenn> how hard could it be to make the bitstream, really?
[02:28:22] <SWPadnos> either they won';t target the biggest (or oldest) chips, or they're time limited (otr they don't have all the optimizations, etc)
[02:28:44] <fenn> or they just revoke your "subscription" because you arent a certified business or educational institution
[02:28:49] <SWPadnos> it's easy to make "a" bitstream, but to make a nicely optimized one, that meets timing constraints, is a bitch
[02:29:08] <SWPadnos> no - the ones that are free are actually free (Xilinx, I think)
[02:29:24] <fenn> achiest dragon was bitching about altera the other day
[02:29:25] <SWPadnos> but they won't target the Virtex chips, which are the really fast/big ones
[02:29:31] <SWPadnos> altera is annoying
[02:29:50] <SWPadnos> their development kits bomb out after 1 year - then you have to get the free version, and lose some functionality
[02:30:10] <SWPadnos> or just keep buying development kits every year (which is what I seem to do)
[02:30:12] <fenn> that is so dumb
[02:30:21] <jepler> is it because their business is really the development tools, not the hardware?
[02:30:31] <fenn> how much does a dev kit cost?
[02:30:32] <SWPadnos> I'm not sure
[02:30:43] <SWPadnos> I asked every vendor about this at ESC
[02:31:09] <fenn> did you carry a ruler, and rap their knuckles when they answered wrong?
[02:31:12] <SWPadnos> it seemed to me that the chip vendors would want to provide you with any software that makes their chips perform better than the competition
[02:31:15] <jepler> this is why I stick with microcontrollers that gcc can target.
[02:31:25] <SWPadnos> I carried my laptop, and whacked them when they got it wrong ;)
[02:32:28] <jepler> are there any good IRC channels for electronics/microcontrollers/embedded stuff?
[02:32:49] <fenn> ##electronics ##microcontrollers #edev
[02:32:52] <SWPadnos> probably, but there are only 26 hours in a day ;)
[02:33:03] <Jymmm> jepler: irc://freenode.net/#emc/
[02:33:13] <jepler> Jymmm: ones where it's on-topic even
[02:33:22] <fenn> #robotics has a lot of avr-talk too
[02:33:30] <fenn> oh there's even #avr but its always dead
[02:33:46] <jepler> join ##electronics
[02:33:52] <jepler> * jepler blushes
[02:34:00] <fenn> bit rusty at the ol' command line eh
[02:34:06] <SWPadnos> heh
[02:34:50] <fenn> SWPadnos: you dont actually stay up 26 hours a day though, unlike some of us
[02:35:08] <SWPadnos> sometimes I stay up 36 hours a day, which really sucks
[02:35:17] <SWPadnos> I think that's why I have this bad cold
[02:35:31] <SWPadnos> (well, that and exposure to it)
[02:35:38] <fenn> mutant germs from outer space
[02:37:39] <SWPadnos> so - here's an interesting (sort of) question that came up last night
[02:37:47] <SWPadnos> "Why go with EMC?"
[02:38:15] <SWPadnos> I should probably write a nice long email, and post it to the dev or user list
[02:38:48] <cradek> because you want to control your cnc equipment with good open source software?
[02:39:02] <SWPadnos> well, that's one good reason
[02:39:18] <SWPadnos> there are two contexts for the question
[02:39:34] <SWPadnos> first, someone who has a BP that's currently controlled by DeskCNC
[02:39:51] <lilo> [Global Notice] Hi all. We're currently experiencing annoyancebot problems in a variety of different channels. You may want to consider temporarily using channel mode +R (only identified users can voice) for the time being.... sorry for the difficulties.
[02:40:00] <SWPadnos> the question there is "what features in EMC would cause him to switch?"
[02:40:09] <jepler> cradek: unlike you or me, most people won't answer "because it's the Free one"
[02:40:21] <cradek> dunno, I don't know squat about deskcnc
[02:40:24] <SWPadnos> (of course, this depends on what features he nees, and what DeskCNC already does, and what emc does that he doesn't need
[02:40:34] <cradek> jepler: I said open source, not free
[02:40:48] <SWPadnos> you can ask Fred of IMService at the workshop ;)
[02:40:56] <fenn> it's GNU/Linux dammit!!!!!11
[02:41:00] <jepler> cradek: I forget which one I have to say to use savannah.
[02:41:13] <fenn> ugh jymmm why did you do that?
[02:41:22] <cradek> maybe I don't care about advocacy, it's obvious to me that open source is best for me, and if someone else isn't like me, maybe that's his own problem
[02:41:28] <SWPadnos> This is the scenario that I've mentioned before - deskcnc and Mach can just load a DXF or BMP file, and start machining
[02:41:41] <Jymmm> fenn walops - annoyancebot
[02:41:52] <SWPadnos> there's a learning curve for most people, just because EMC runs on Linux
[02:42:04] <fenn> well there arent any here, so you're just stopping unregistered people from talking
[02:42:15] <jepler> Is "+r" different from "+R"?
[02:42:23] <SWPadnos> I think we're mostly registered
[02:42:33] <Jymmm> jepler "shouldn't" be.
[02:42:56] <cradek> I didn't see any problem on this channel
[02:43:12] <fenn> * fenn hunts for the annoyancebot
[02:43:22] <fenn> * fenn points at les_w -- get 'im!
[02:43:57] <jepler> wheeeeeee! hee hee heee </internet explorer icon>
[02:44:09] <cradek> ha
[02:45:07] <Jymmm> jepler slightly different
[02:45:33] <jepler> at least that doesn't keep anyone out of the channel
[02:45:43] <Jymmm> yeah
[02:56:22] <Jymmm> SWPadnos: any suggestions on a inexpensive LED driver?
[02:56:38] <SWPadnos> starting from what? (AC, a battery, etc)
[02:56:53] <Jymmm> SWPadnos uC controlled
[02:57:13] <SWPadnos> just for lighting, right?
[02:57:27] <SWPadnos> (not matrix displays or 7-segment)
[02:57:28] <Jymmm> SWPadnos Yeah (opposed to what?)
[02:57:34] <SWPadnos> matrix displays or 7-segment
[02:57:37] <SWPadnos> ;)
[02:57:39] <Jymmm> ah, plain old leds
[02:58:14] <SWPadnos> I just use a current limiting resistor and a transistor, with an AVR to control the brughtness using PWM
[02:58:19] <SWPadnos> brightness
[02:59:03] <Jymmm> SWPadnos I'm talking about 10 or 20 LED's
[02:59:18] <SWPadnos> I'm talking about hundreds, with 16 channels of control
[02:59:44] <Jymmm> how you gonna control 100's wiht a lil ole avr ?
[03:00:07] <Jymmm> nm you said 16ch
[03:00:20] <Jymmm> ok, what Q ?
[03:00:25] <SWPadnos> 16 channels, each bank of 4 can be attached to a separate voltage source, so using 24V on all 16 channels lets me have series strings of 6-12 LEDs each
[03:00:38] <SWPadnos> and I can put about 50 strings on a single channel
[03:00:56] <SWPadnos> (20-50 mA per string of LEDs, 2A transistors)
[03:01:22] <jmkasunich> you use "transistors" or mosfets?
[03:01:37] <SWPadnos> SOT-23 transistors - hold on one sec
[03:01:41] <SWPadnos> I'll get the part number
[03:01:49] <jmkasunich> 2A sot-23s?
[03:01:51] <jmkasunich> wow
[03:02:20] <jmkasunich> 2A * 0.2Vcesat = 0.4W, don't those tiny lil' thangs get hot
[03:02:22] <SWPadnos> yeah - they've got a very low voltage drop
[03:02:34] <SWPadnos> they're rated for 350mW
[03:03:07] <fenn> why not use fets? :)
[03:03:08] <SWPadnos> FMMT619
[03:03:20] <SWPadnos> more expensive, I think they were
[03:04:31] <SWPadnos> hmmm - I may have used the MMBT6427FS though
[03:04:50] <jmkasunich> either way, I'm impressed
[03:05:06] <jmkasunich> I think of SOT-23 transistors as being weak
[03:05:13] <SWPadnos> nope - the 6427 is a 1.2A, the 619 is 2A
[03:05:22] <jmkasunich> how much base current is needed to saturate them?
[03:05:32] <SWPadnos> they're good enough that the transistor still works even if it melts itself off the board
[03:05:32] <Jymmm> 625mW
[03:05:53] <Jymmm> http://www.ortodoxism.ro/datasheets/zetexsemiconductors/fmmt619.pdf
[03:05:56] <jmkasunich> thats impressive
[03:06:04] <jmkasunich> ls
[03:06:15] <SWPadnos> yeah - I like them
[03:06:39] <Jymmm> heh... 12A peak
[03:07:02] <SWPadnos> that's the 617
[03:07:11] <SWPadnos> the 619 is only 6A ;)
[03:07:31] <Jymmm> still
[03:08:17] <Jymmm> Just saw the disclaimer at the bottom
[03:08:21] <SWPadnos> jmkasunich, 500mA is the max base current, but since hfe is in the 200+ range, 10 mA should be close to saturation
[03:08:34] <jmkasunich> nice parts
[03:08:52] <SWPadnos> I used them on my pulse generator, so I could drive small stepper motors
[03:10:02] <SWPadnos> so Jymmm, does that answer your question?
[03:10:42] <Jymmm> SWPadnos pretty much, ty. Now how something else resolved too with that transistor
[03:10:49] <SWPadnos> cool
[03:11:10] <SWPadnos> there may be better ones now, that one was selected over a year ago
[03:11:58] <Jymmm> I wasn't even aware there was high power in such a small pkg available.
[03:12:19] <SWPadnos> ideally, there isn't a lot of power in that package ;)
[03:12:29] <SWPadnos> but high current, sure :)
[03:12:31] <Jymmm> well with the PWM yeah
[03:13:00] <Jymmm> $0.78/ea from mouser
[03:13:29] <Jymmm> ~~~~~
[03:13:49] <SWPadnos> how are they at 10 or 100 qty?
[03:14:15] <Jymmm> $0.39 @ 100, $0.20 @ 10K
[03:14:24] <Jymmm> $0.32 @ 250
[03:14:27] <SWPadnos> ok - about the same as DigiKey
[03:14:41] <SWPadnos> $0.3888 / 100
[03:16:07] <Jymmm> wow $0.44/ea from newark
[03:16:17] <SWPadnos> not bad
[03:16:44] <Jymmm> bastards.... 71 day leadtime, zero in stock
[03:16:57] <jmkasunich> no wonder they're so cheap
[03:17:01] <SWPadnos> heh - you can sell something you don't have at any price
[03:18:19] <SWPadnos> interesting. I think the relative prices of FETs and BJTs have done an inversion in the past year
[03:19:43] <SWPadnos> http://www.digikey.com/scripts/DkSearch/dksus.dll?Detail?Ref=3213&Row=191684&Site=US
[03:19:57] <SWPadnos> hmmm - I wonder if that will work
[03:23:17] <Jymmm> 1.2W in SOT23
[03:24:06] <SWPadnos> 1.9 for that last one
[03:24:30] <Jymmm> heh
[03:24:39] <jmkasunich> thats insane
[03:24:48] <jmkasunich> whats the Rth for a SOT-23?
[03:24:59] <jmkasunich> gotta be close to 100 deg C/watt
[03:25:01] <SWPadnos> I don't think the ground terminal would have withstood that much total current
[03:25:20] <SWPadnos> 65K/W
[03:25:31] <SWPadnos> junction to solder point
[03:25:48] <jmkasunich> so you need some honking big traces to heatsink the solder point
[03:26:02] <jmkasunich> Rth j-ambient is what I was asking about
[03:26:08] <SWPadnos> yep. that's a max rating as well
[03:26:17] <jmkasunich> (which obviously depends on the board, but...)
[03:27:30] <SWPadnos> I don't see any other thermal coefficients
[03:28:13] <SWPadnos> the Rds(on) is pretty low - 30-36 mOhm (or 48-57.6 mOhm at 150C)
[03:28:51] <jmkasunich> now you're looking at fets?
[03:29:09] <SWPadnos> yes - sorry
[03:29:19] <SWPadnos> that last link was for a 5.7A SOT23 FET
[03:29:26] <jmkasunich> those are very impressive numbers
[03:29:48] <SWPadnos> 1.9W dissipation, and $0.45 in single quantity
[03:30:25] <jmkasunich> 1.9W * 65K/W = 123.5K rise above solder point
[03:30:36] <jmkasunich> so if soldered to a 25C infinite heatsink, Tj = 150
[03:30:42] <SWPadnos> it's rated for that
[03:31:04] <jmkasunich> if soldered to a real board tho, Rth is gonna be at least double the stated value
[03:31:54] <SWPadnos> yep - though you can use a via for heatsinking
[03:32:17] <jmkasunich> if soldered to a real board with 50C air temp.... the real power rating is well under 1 watt unless you dedicate a lot of copper to heatsinking
[03:32:28] <jmkasunich> I hate specmanship
[03:32:33] <SWPadnos> yep
[03:32:48] <SWPadnos> though 1W at 30 mOhm gives many amps anyway
[03:32:58] <jmkasunich> yeah
[03:35:52] <|Magician|> SWPadnos: I seen where you had posted some comments on the Mesa 5i20 PCI Card...have you had much luck with it?
[03:36:08] <SWPadnos> I haven't tried to make any luck, so no ;)
[03:36:25] <SWPadnos> I'm not sure I'll have a chance before Fest
[03:38:40] <SWPadnos> |Magician|, were you talking about modifications to the FPGA, or just getting the card to work?
[03:39:06] <|Magician|> the mods for step & dir.
[03:39:16] <SWPadnos> ok - nope, haven't had any time for that
[03:39:40] <|Magician|> you using it for servos now?
[03:39:56] <SWPadnos> I'm not using it, but I have one installed in my test machine, and it appears to work fine
[03:40:47] <fenn> so swp you're not making rodeo videos, you're not setting up the bridgeport, what do you do all day?
[03:41:00] <SWPadnos> I wonder what I'll do the next day :)
[03:41:14] <|Magician|> I have been looking for something that could do PWM & step.....I thought maybe I had run up on something
[03:41:23] <jmkasunich> rodeo videos?
[03:41:27] <SWPadnos> the Mesa can, but doesn't
[03:41:33] <SWPadnos> jmkasunich, the camera array
[03:41:38] <jmkasunich> ah
[03:42:16] <|Magician|> yeah thats what Peter with Mesa was saying
[03:43:00] <SWPadnos> I didn't have a chance to talk to him when I was in San Jose, but he had said that he was willing to write the code for the step generators
[03:43:37] <|Magician|> Thats what he was saying when I was talking to him yesterday...he said he hasnt got around to it
[03:43:50] <SWPadnos> ok - it's not just me ;)
[03:44:50] <|Magician|> He also said that there was another guy named peter thst was messing with it...I seen where he had posted on the forum, but havent been able to make much of it
[03:45:03] <SWPadnos> Pete Vavaroutsos?
[03:45:12] <|Magician|> Yes
[03:45:41] <SWPadnos> I didn't gather that petev is working on it. He seemed pretty busy
[03:46:57] <|Magician|> Peter with Mesa just said that he might know some about it....he didnt say that he was working on it....I had just seen where he had posted about it a couple of times
[03:47:06] <SWPadnos> ah
[03:49:55] <|Magician|> Thats the closest thing that I have came to finding to something that would work for my setup so far
[03:51:35] <SWPadnos> I think the mesa card is a reat thing - I'd just like to be able to use it to its full potential, by being able to reprogram the FPGA
[03:52:14] <SWPadnos> err - great thing
[04:05:09] <SWPadnos> ok - the good news is that I was able to import the constraint file into my design software, so as soon as I re-register my stupid fscking FPGA design kit from Xilinx, I should be able to make new FPGA configs
[04:05:32] <SWPadnos> on that note, I think I'll head off to bed. night all
[04:05:39] <SWPadnos> SWPadnos is now known as SWP_Away
[04:09:30] <jmkasunich> logger_aj: bookmark?
[04:09:30] <jmkasunich> See
[04:09:48] <SWP_Away> (try it with no question mark)
[04:09:53] <jmkasunich> logger_aj: bookmark
[04:09:53] <jmkasunich> See
[04:10:05] <SWP_Away> logger_aj, bookmark
[04:10:05] <SWP_Away> See
[04:10:12] <jmkasunich> fsck
[04:10:12] <SWP_Away> I guess it's offline
[04:10:15] <jmkasunich> I know why
[04:10:19] <jmkasunich> Jymm set +R
[04:10:38] <SWP_Away>
[04:10:41] <jmkasunich> (that also explains why CIA messages that I was expecting didnt show up
[04:10:44] <SWP_Away> ah
[04:10:56] <jmkasunich> Jymmm: remove the +R please!!!
[04:10:59] <SWP_Away> hey Jymmm unset R, will ya?
[04:11:59] <jmkasunich> logger_aj: bookmark
[04:11:59] <jmkasunich> See
[04:12:04] <Jymmm> nag nag bitch bitch...
[04:12:04] <jmkasunich> yep, that was it
[04:12:24] <jmkasunich> thanks
[04:12:36] <SWP_Away> ok. NOW I can go to bed :)
[04:12:55] <Jymmm> SWP_Away and you couldn't before? =)
[04:13:15] <SWP_Away> err - well, I. um. good night
[04:13:29] <Jymmm> SWP_Away go visit your wife
[04:13:36] <SWP_Away> yeah, right
[04:13:42] <SWP_Away> I mean - I have a cold, dummy
[04:13:47] <Jymmm> is that such a bad thing?
[04:13:59] <Jymmm> visiting the wife, not the cold
[04:14:00] <SWP_Away> it is if I wake her up
[04:14:06] <CIA-4> 03compile-farm 07BDI-2.18 (2.2.18-rtl3.0) * 10emc2head/: build PASSED ; see http://linuxcnc.org/compile_farm/emc2head_slot2_log.txt
[04:14:07] <Jymmm> oh, heh
[04:14:08] <SWP_Away> she works, you know
[04:14:20] <jmkasunich> yay, it worked
[04:14:32] <Jymmm> I do now
[04:14:34] <jmkasunich> one slot at least
[04:14:49] <SWP_Away> I hope the TODO change doesn't fubar any of the compile farm slots ;)
[04:15:14] <jmkasunich> heh, actually I'm anxiously waiting for somebody to screw something up
[04:15:31] <jmkasunich> because reporting to the commit list is untested
[04:15:46] <SWP_Away> I'd help with that, but I'd have to boot the test machine
[04:15:52] <jmkasunich> I only report failure to the list (don't want to spam it with lots of unimportant messages)
[04:16:10] <SWP_Away> make sure you report the name of the last committer ;)
[04:16:22] <jmkasunich> list archives will have that
[04:18:40] <CIA-4> 03compile-farm 07BDI-TNG (2.4.18-rtai) * 10emc2head/: build PASSED ; see http://linuxcnc.org/compile_farm/emc2head_slot3_log.txt
[04:25:36] <CIA-4> 03compile-farm 07BDI-4.20 (2.6.10-adeos) * 10emc2head/: build PASSED ; see http://linuxcnc.org/compile_farm/emc2head_slot5_log.txt
[04:28:21] <CIA-4> 03compile-farm 07BDI-Live rc46 (2.4.25-adeos) * 10emc2head/: build PASSED ; see http://linuxcnc.org/compile_farm/emc2head_slot4_log.txt
[05:36:50] <lilo> [Global Notice] Hi all. Once again, we're still having problems with a kiddie attacking channels. You should definitely consider setting your channel to +R for now, if at all possible, until the problem is resolved. Apologies for the inconvenience.
[05:49:01] <lilo> [Global Notice] Hi all. We're currently having problems with trollbots on a number of channels, and as a result many channels are temporarily set to +R. If you can't talk on such a channel, please register with nickserv --- /msg nickserv help register .
[05:49:10] <lilo> [Global Notice] Apologies for the inconvenience, and thank you for using freenode!
[06:12:26] <lilo> [Global Notice] Hi all. You may want to consider reverting from channel mode +R at this point, though if our problem person comes back, we'd advise using it again promptly, and leaving it up for a while. Thanks to everyone involved for your patience!
[06:31:16] <lilo> [Global Notice] Hi folks. Our problem child is back. You'll probably want to use +R for a few more hours. Sorry for the inconvenience, and thanks.
[07:48:40] <chinamill> hello
[08:10:39] <alex_joni> hello
[11:06:45] <alex_joni> * alex_joni goes away till tuesday
[11:06:48] <alex_joni> alex_joni is now known as alex_joni_away
[11:54:45] <chinamill> Does anyone have the hal_show.tcl ??
[12:03:15] <chinamill> Hello SkunkWorks; Do you have the hal_show.tcl ?
[12:05:41] <SkunkWorks> HI - where would it be and what does it do :)
[12:08:44] <cradek> I looked and there is no such program in emc2
[12:08:46] <chinamill> I got it from rayh before, but I lost it somewhere, it is a light gui to view signals and pins etc.
[12:09:01] <cradek> that's in halcmd, just run halcmd from the menu
[12:09:05] <SkunkWorks> ah - I know he was working on some sort of gui
[12:09:10] <cradek> err halconfig
[12:10:22] <cradek> go to the watch tab
[12:10:35] <cradek> select your signals/pins in the tree on the left
[12:10:57] <chinamill> Eh, how do i start the gui?
[12:11:35] <cradek> when running emc, it's on the scripts menu (in tkemc) and on another menu in the other guis
[12:11:49] <chinamill> ok
[12:12:27] <chinamill> great, now it up :)
[12:15:16] <chinamill> what is ddt?
[12:22:42] <SkunkWorks> (I was thinking about the hal-whatever that allows you to connect the hal componants togather.
[12:22:45] <SkunkWorks> )\
[12:23:52] <SkunkWorks> cradek: - changed my periods from seconds to nanoseconds and it worked. thanks
[13:08:48] <chinamill> When I do: bin/halcmd linkps supply.0.variable Zpos-cmd I get: HAL: ERROR: signal 'Zpos-cmd' already has writer(s) ; HAL:0: link failed Is this too forbidden?
[13:09:08] <SWP_Away> SWP_Away is now known as SWPadnos
[13:09:29] <SWPadnos> you can only have one "write" pin attached to any signal
[13:15:57] <cradek> SkunkWorks: so you're using the 04-19 release now, have you had any problems?
[13:16:30] <cradek> there have been a few small bugfixes since then, I think we're getting ready for an official release
[13:17:56] <SWPadnos> woohoo!
[13:18:10] <SWPadnos> um - excuse me :)
[13:18:10] <ValarQ> woo
[13:19:27] <cradek> the last few fixes have been old, old bugs (not recent breakage), which I think is a good sign that we're gaining real stability
[13:19:38] <SkunkWorks> Cradek: I have not physically run the maching but brought up some files I made in eagle and ran them - no issues.
[13:19:45] <SkunkWorks> machine
[13:20:01] <cradek> ok, that's good
[13:20:13] <SkunkWorks> even found out g82 works. :)
[13:20:35] <cradek> that's one of the drill cycles? I use those all the time for pcbs
[13:20:44] <SkunkWorks> (think it was g82 - it was what pcb-gcode put out for drilling)
[13:20:50] <SkunkWorks> yes
[13:21:21] <SkunkWorks> I don't understand why they put a 10 second delay at the bottom of each hole though
[13:21:21] <cradek> they are sure an easy way to program a hole drill
[13:21:21] <cradek> 10 second???
[13:22:00] <cradek> I bet another system has the pause in msec or csec, I bet you should change it to .01
[13:22:51] <SkunkWorks> yes :) G82 X0.3500 Y2.3500 Z-0.1200 F5.00 R0.0200 P10.000000
[13:23:06] <SkunkWorks> it was supposidly the emc post proccessor\
[13:23:29] <cradek> close but not quite
[13:29:54] <chinamill> is there a map how the components interacts?
[13:30:04] <chinamill> (of how)
[13:30:33] <SWPadnos> HAL components, or "the components of the software"?
[13:30:36] <chinamill> Hal components
[13:30:53] <SWPadnos> hmmm - one sec
[13:32:19] <Jymmmmm> Jymmmmm is now known as Jymmm
[13:34:49] <SWPadnos> chinamill, there are some diagrams on the wiki, but I can't find them right now. I'm not sure they're what you need anyway
[13:35:05] <chinamill> ok
[13:35:10] <chinamill> thanks anyway
[13:35:11] <SWPadnos> ha: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?EMC_Components
[13:35:16] <SWPadnos> found them ;)
[13:36:38] <SWPadnos> I'm not sure if that's the kind of info you're looking for
[13:40:37] <SkunkWorks> cradek: not to many people must use the post proccessor for emc as there was still a few ';' in the program http://www.electronicsam.com/images/KandT/DRIVEtd.nc
[13:47:20] <jepler> SkunkWorks: http://unpy.net/cgi-bin/viewcvs.cgi/eagle/ulp/
[13:47:36] <jepler> SkunkWorks: here's what I use to turn eagle boards into millable files
[13:58:35] <Jymmm> http://www.tatjavanvark.nl/tvv1/pht10.html
[14:08:46] <SkunkWorks> logger_aj: bookmakr
[14:08:46] <SkunkWorks> I'm logging. I don't understand 'bookmakr', SkunkWorks. Try /msg logger_aj help
[14:08:52] <SkunkWorks> logger_aj: bookmark
[14:08:52] <SkunkWorks> See
[14:09:32] <SkunkWorks> Thanks jepler.
[14:15:32] <Jymmm> ROTFLMAO (dumbass) --> https://bugzilla.mozilla.org/show_bug.cgi?id=330884
[14:21:54] <Jymmm> VERY COOL --> http://www.hiddenpassageway.com/
[14:29:55] <chinamill> Hmm, I have some problems of moving Z with HAL. I want to be able to mave Z while interpreting gcode. My plan was to use the signal Zpos-cmd. But as soon as a give the pin that feeds Zpos-cmd a new value I get j.following err. Does anyone have any ideas?
[14:30:21] <cradek> make your FERROR for that axis big
[14:31:09] <chinamill> I'll try
[14:32:25] <jepler> MIN_FERROR
[14:32:42] <jepler> If MIN_FERROR is 1.0, then the axis can move 1 inch before emc will signal a following error.
[14:33:10] <chinamill> (I dont get any ferrors while jogging)
[14:33:30] <jepler> (or mm if your units are mm)
[14:37:19] <steves_logging> steves_logging is now known as steve_stallings
[14:41:06] <chinamill> it got better
[14:41:13] <chinamill> thanks
[15:04:31] <steve_stallings> steve_stallings is now known as steves_logging
[15:25:13] <giacus> logger_aj: bookmark
[15:25:13] <giacus> See
[15:36:17] <bill20r3> good morning.
[16:54:02] <chinamill> hmm, I have to admit my C knowledge is rusty... I try to alter the supply.c (for testing and learning). Now I try to change the first bit in the struct to be incoming instead of out. My test is to put the value of the incoming bit to the other outgoing _q. But I cant get _q to change value... Any ideas?
[16:54:34] <chinamill> I changed retval = hal_pin_bit_new(buf, HAL_RD, &(addr->q), comp_id);
[16:54:46] <chinamill> RD instead of WR
[16:57:12] <Lerneaen_Hydra> Is there anything difficult in using a PCI-parport controller?
[16:57:21] <Lerneaen_Hydra> and what changes are reuired?
[16:57:47] <cradek> Lerneaen_Hydra: usually just the IO address
[17:00:13] <Lerneaen_Hydra> usually?
[17:00:14] <rayh> Some of these PCI cards convert their ports to ordinary IO addresses.
[17:00:41] <rayh> So you use 0x378, 0x278, or 0x3BC
[17:01:10] <rayh> I've also used the real PCI addy in the EMC ini file and that works fine.
[17:01:39] <chinamill> I have put the sligtly modded supply.c in http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Test_New_Page
[17:05:54] <chinamill> What do I have to consider for reading a bit from hal pin? I tried to study the diferences between debounce.c and supply.c to change one of the stuct variables to a incoming.
[17:08:48] <rayh> Are you trying to set an output value based on an input signal?
[17:10:26] <chinamill> Yep
[17:11:10] <chinamill> I started to get too many grey hairs...
[17:12:19] <rayh> There are some "blocks" that can compare, add, set values.
[17:14:13] <chinamill> my first goal is to have a basic code skeleton for reding and writing to pins. I think it shouldent be very complicated
[17:15:23] <chinamill> more or less a hal component that do nothing but read and write
[17:15:33] <Lerneaen_Hydra> which tools do most of you use to get the refereance point for milling? (setting the 0 point for XY and Z)
[17:16:15] <chinamill> then I can start to implement the functionality
[17:24:01] <chinamill> LH: I have a flat metal piece partly emdedded in plasic that is conected to estop. I use it to align it to an edge and then i feed into it and it will do estop right att the edge - 1/2 diameter... Do you get it?
[17:27:42] <Lerneaen_Hydra> yes I understand, same procedure for X, Y and Z?
[17:27:57] <chinamill> yep
[17:28:24] <Lerneaen_Hydra> how long time does it take to measure in all three axes?
[17:28:55] <rayh> Why estop not home?
[17:29:12] <chinamill> home is better
[17:29:24] <rayh> These are limit switches?
[17:29:42] <chinamill> no, the mashine must stop
[17:31:20] <rayh> You'd like any one of these switches to stop the machine.
[17:33:25] <chinamill> it's just a thing I use kinf of like a temporary limit sw. to be able to measure in. Probobly not a great method, but low cost for sure.
[17:35:41] <rayh> You bet. Good idea.
[17:36:21] <rayh> What physical pin do you plan to use to get the switch into HAL?
[17:37:39] <chinamill> the measuring thing I have already working, but on what pin I put it I do not remenber
[17:38:26] <chinamill> My quest in the HAL components is for a torch height controller.
[17:42:12] <rayh> Ah. Would the height variable be binary or float?
[17:43:00] <rayh> I've seen them use air pressure near the torch.
[17:43:39] <rayh> That pressure is turned into an analog signal.
[17:45:01] <chinamill> I need to make a hal component that can read bit and float and write float to hal pins.
[17:45:46] <chinamill> I have to run for a while... will be back later to read back.
[17:51:27] <Lerneaen_Hydra> what would I do to set up EMC to use the other parport?
[17:52:39] <fenn> change the address in the loadrt hal_parport line
[17:53:47] <Lerneaen_Hydra> anything else?
[17:55:42] <rayh> You can add more than one addy there and see them in the HAL
[17:55:54] <Lerneaen_Hydra> how do I find the adress for the parport?
[17:56:54] <fenn> its usually 0x278 or 0x378 but you might find that in the bios
[17:57:45] <rayh> lspci might show a usefull address if it's not one of those above.
[17:57:47] <Lerneaen_Hydra> I find it hard to see where I would find adresses for PCI-based parports in the bios, in which menu were you thinking?
[17:58:14] <fenn> hmm.. i was thinking integrated peripherals maybe
[17:58:24] <Lerneaen_Hydra> ok
[17:58:24] <fenn> i''ve never had a parport card though
[17:58:55] <rayh> sudo lspci in a terminal should show the card name and addys.
[18:00:33] <Lerneaen_Hydra> that gave me an adress not at all similar to the one in standard_pinout.hal
[18:01:20] <rayh> What numbers did it show you?
[18:03:04] <Lerneaen_Hydra> 0000:01:0c.0 Serial controller: Timeda Technology Co Ltd PCI2S550 etcetc
[18:03:33] <Lerneaen_Hydra> it must be that one as the only other controller I have is a network card
[18:11:02] <rayh> oops try sudo lspci -vv
[18:11:37] <rayh> It should show 0x numbers
[18:12:01] <rayh> for that Timeda board
[18:20:53] <rayh> It ought to say something like xxx at e800 or b800 or some such.
[18:22:06] <Lerneaen_Hydra> this board has 2x RS232 and 1X parport, and the information shown from lspci is: size32, df80. size8, dff0, size8, dfe0
[18:22:25] <Lerneaen_Hydra> So i presume the parport has the adress df80?
[18:23:25] <rayh> Sounds like that would be a good place to try anyway.
[18:25:04] <rayh> loadrt hal_parport cfg="0x0378"
[18:25:04] <rayh> is the line in standard_pinout.hal
[18:25:27] <rayh> and just replace the 0378 with df80
[18:27:37] <rayh> or loadrt hal_parport cfg="0x0378 0x0278" if you want two parports in HAL.
[18:30:22] <SkunkWorks> ray: still around?
[18:30:49] <SkunkWorks> rayh:
[18:31:13] <rayh> You bet.
[18:32:05] <SkunkWorks> just wanted to let you know the wireless link is up and running. The only issue so far is every once in a great while the wireless connection will loose it for about 20 seconds.
[18:33:00] <SkunkWorks> maybe someone with a portable phone or something. going to keep an eye on it for a while and see. I can still add a better antanna on the other end.
[18:33:15] <SkunkWorks> (more directional)
[18:33:28] <rayh> Great to hear.
[18:34:06] <rayh> What happens when it looses the signal.
[18:34:17] <rayh> network drops out or ??
[18:36:50] <SkunkWorks> I can ping the AP connected to the computer but not the other side of the wireless link.
[18:37:08] <SkunkWorks> total disconnection/
[18:37:53] <SkunkWorks> so far they are pleased as punch going from there 46kbs connection ;)
[18:38:44] <Lerneaen_Hydra> rayh: What is it that defines which is parport0 and 1? is it the order of them loading or the physical position?
[18:38:46] <SkunkWorks> I have not looked to see if there is any kind of logging in the ap to get an idea.
[18:39:06] <rayh> order of loadrt command
[18:39:58] <Lerneaen_Hydra> and df80 => 0x0278 when going from hex to dec?
[18:40:13] <cradek> 0x278 is hex
[18:40:14] <rayh> SkunkWorks, I can see why. Imagine it from my 19k on a good day connection
[18:40:30] <cradek> the leading 0x means hex
[18:41:09] <Lerneaen_Hydra> so 0xdf80 is a valid adress for standard_pinout.hal?
[18:41:15] <cradek> sure
[18:49:43] <rayh> Cradek is right. I know it sounds goofy. I was pretty skeptical the first time I tried it.
[19:55:06] <giacus> hello craigski
[19:55:36] <giacus> where you phone from ?
[19:56:12] <giacus> helloo ??
[19:58:55] <Jymmm> you scared him off giacus
[19:59:09] <giacus> you think that ?
[19:59:12] <giacus> hmm
[19:59:17] <Jymmm> lol, nah =)
[19:59:21] <giacus> :)
[19:59:27] <Lerneaen_Hydra> I think Jymmm was joking ;)
[20:00:58] <chinamill> Does anyone know of any documenation on how to write a HAL component?
[20:01:52] <giacus> chinamill: the source code should be the best doc for that
[20:02:12] <Jymmm> chinamill or talk to jmkasunich
[20:02:22] <cradek> he'll tell you the same thing
[20:02:52] <cradek> supply.c and timedelay.c are very simple
[20:03:38] <cradek> what component are you going to write?
[20:03:55] <chinamill> I'll have a look at timedelay.c so far I'm having problems of reading and writing input/output
[20:03:58] <Lerneaen_Hydra> g'night all
[20:04:05] <chinamill> Night
[20:04:24] <chinamill> A component for a torch heigt controller
[20:06:16] <chinamill> timedelay.c looks like my best shot!
[20:10:34] <cradek> I thought you were going to be able to use existing hal components - did you find this is not the case?
[20:14:26] <chinamill> cradek: no I have to write one, a very simple one but still write one. Ineed to read pins with bit and float and write a pin with a float. (have to reboot, will be back soon)
[20:41:10] <K4ts> hello
[20:46:27] <chinamill> hello
[20:48:19] <rayh> K4ts, Will Prodi be better or worse for your country?
[20:48:29] <K4ts> yes
[20:48:38] <K4ts> :-)
[20:49:20] <rayh> Wait and see! Good plan.
[20:52:44] <chinamill> Better than power abusing Mr B anyway
[20:53:34] <K4ts> chinamill: sure!
[20:54:17] <giacus> http://www.giacus.org/photo/eventi/inde.png
[20:54:29] <giacus> hello rayh!
[20:54:45] <rayh> Hi giacus
[20:54:56] <giacus> howdy ?
[20:55:16] <rayh> I saw the www.giacus.org addy for you.
[20:55:16] <giacus> I'm back in naples for some week ..
[20:55:23] <giacus> :)
[20:59:06] <giacus> I finally found the time to subscribe myself to the emc mailing list
[20:59:11] <giacus> :(
[20:59:18] <giacus> haha
[21:00:58] <giacus> I'm also waiting the emc2 release to update the HAL it doc :P
[21:02:18] <rayh> I saw your link on that subscription.
[21:02:28] <giacus> oh, nice
[21:02:53] <rayh> As soon as you are ready we should make that hal doc available on www.linuxcnc.org.
[21:03:22] <giacus> yes, waiting it
[21:04:08] <giacus> the moment its a good moment ;P
[21:04:25] <giacus> for emc of course heh
[21:05:37] <giacus> I'm not following much it in these latest weeks
[21:05:48] <giacus> but its seems to be retty ready
[21:05:55] <giacus> pretty*
[21:06:33] <chinamill> Horay, reading and writing bit in hal component now :)
[21:06:52] <giacus> what I know for sure, there's alays a lot of work around a stable release :(
[21:07:02] <giacus> chinamill: nice
[21:09:15] <rayh> Fantastic chinamill.
[21:09:31] <chinamill> :)
[21:09:36] <rayh> I'm hearing of a lot of interest in emc2 these days.
[21:13:07] <giacus> rayh: yeah, you're right
[21:13:36] <giacus> peoples are interested on that
[21:14:24] <giacus> however, emc development is very difficult I guess
[21:14:55] <rayh> It is a rather complex set of files.
[21:14:56] <giacus> since there is not so much tester as in other projects
[21:15:10] <giacus> its also expensive
[21:15:41] <fenn> nah
[21:15:47] <giacus> but i'm sure emc2 could open some way for factory use
[21:15:55] <giacus> not just hobbist
[21:16:14] <giacus> since it is evoluted and stable yet
[21:17:10] <giacus> hey fenn
[21:17:17] <giacus> as expensive I meant the mechanical part
[21:17:29] <giacus> not the software of course :)
[21:17:43] <fenn> you can build really cheap stuff, it just wont be as fast as the expensive ones
[21:18:03] <giacus> I'm sure about that
[21:18:15] <giacus> I'm not sure all peoples know it
[21:18:40] <bill20r3> Mmmm cheap
[21:18:57] <giacus> what I really love, is that I can 100% take the control of anything
[21:19:03] <giacus> on my machine
[21:19:21] <giacus> but that, maybe is not so important for some peoples
[21:19:31] <giacus> who don't know enough ..
[21:19:41] <giacus> that's the brake I guess
[21:21:40] <rayh> The flexibility of EMC means more work to setup.
[21:22:30] <giacus> I agreed in all
[21:23:32] <fenn> more work to do sanity checks when writing the software
[21:23:43] <fenn> its easy to click on "stepper_inch" or whatever
[21:24:16] <giacus> fenn: yeah, true
[21:25:45] <giacus> however, I think politics of business is to sell the software to control the machine, togheter with the machine ..
[21:26:10] <giacus> that's a bit what was happening with computers with windowze preinstalled on
[21:26:17] <rayh> Sure.
[21:26:54] <fenn> if i were selling a machine i'd sell it with the software
[21:27:27] <giacus> dunno
[21:27:42] <giacus> I see 2 different things
[21:27:47] <giacus> machine its a material thing
[21:27:59] <giacus> software is unreal
[21:28:11] <giacus> but also in the software there's a lot of work
[21:28:19] <giacus> maybe more of machine
[21:28:35] <giacus> I'd sell the support
[21:29:13] <giacus> many peoples are living just with that
[21:29:30] <giacus> always when a business can be created around it
[21:29:41] <giacus> it depend on many things
[21:29:49] <giacus> but I think is possible
[21:30:41] <giacus> as well, emc2 could open a new way for that
[21:30:50] <giacus> right now
[21:31:47] <giacus> it is useful and it only depend on how many peoples will choose it
[21:31:52] <giacus> hi SWPadnos :)
[21:31:56] <SWPadnos> higia
[21:31:59] <SWPadnos> argh
[21:32:01] <SWPadnos> hi giacus
[21:32:02] <giacus> hehe
[21:32:08] <giacus> hi :)
[21:36:17] <giacus> I was reading abou Gnuplot
[21:36:31] <giacus> I suppose it has been used in emc development
[21:37:09] <giacus> I never looked at the source code yet :(
[21:38:04] <giacus> it also have Python support for what I seen
[21:39:56] <giacus> it seems to me what is really needed its an evoluted free cad/cam
[21:41:18] <giacus> I feel as years ago when was no a good browser around, (before mozilla)
[21:42:42] <bill20r3> I concur.
[21:42:56] <bill20r3> I'd even settle for "average"
[21:43:01] <giacus> so, wonder where's the secret
[21:43:38] <bill20r3> I think that nobody who's smart enough to code it has also needed it badly enough to code it.
[21:43:56] <bill20r3> that seems to be how most projects get started, someone scratching an itch.
[21:54:00] <giacus> mm we have to reinvent the wheel ..
[22:08:31] <Jymmm> giacus: http://images.google.com/images?q=tbn:ZnPYtiLq4tJ9_M:www.rhodes.edu/MathematicsandComputerScience/Students/images/Pi200-4b.gif
[22:09:42] <rayh> bbl
[22:12:20] <K4ts> night
[22:12:28] <giacus> Jymmm: cool
[22:13:09] <giacus> I was looking at open cascade
[22:13:15] <giacus> In short, Open CASCADE Technology Public License is LGPL-like with certain differences..
[22:13:57] <giacus> the differerence could make the difference
[22:14:08] <giacus> ]:)
[22:14:11] <bill20r3> the screenshots look nice.
[22:52:35] <giacus> night all
[23:13:06] <lilo> [Global Notice] Hi all. Our problem child continues to spam random channels. He is in no way affiliated with #fazlamesai, http://www.fazlamesai.net/ or teknolojitelevizyonu.com (Turkish Technology Television)
[23:14:02] <lilo> [Global Notice] You may want to consider setting channel mode +R for the time being (only registered users or +v users have voice) as a temporary measure. We apologize for the inconvenience!
[23:50:25] <dmessier> allo tous ;)