#emc | Logs for 2006-05-31

Back
[00:44:10] <rayh> logger_aj, bookmakr
[00:44:10] <rayh> I'm logging. I don't understand 'bookmakr', rayh. Try /msg logger_aj help
[00:44:19] <rayh> logger_aj, bookmark
[00:44:19] <rayh> See http://81.196.65.201/irc/irc.freenode.net:6667/emc/2006-05-31#T00-44-19
[00:44:28] <SWPadnos> calling the logger a bookmaker, eh? ;)
[00:44:41] <rayh> one does what one can.
[00:56:13] <jepler> SWPadnos: happen to know the best way to rotate an 8-bit register on AVR? right and left? add R, R; adc R, r0 seems to be a left-roll, and by using "T" there seems to be a 3-instruction right-roll.
[00:56:45] <fenn> are you doing spi?
[00:57:00] <SWPadnos> 1 sec - phone
[00:57:43] <jepler> fenn: no, I'm generating quadrature output
[00:58:12] <fenn> ah
[00:58:28] <jepler> put 0xcc in a register, and roll it one way or the other when you want to change the output. pick two adjacent bits and that's your quadrature output.
[01:00:22] <SkunkWorks> jmkasunich: hooked up the h-bridge today. 3 problems that took about an hour to figure out. 1 - I had inverted 2 pins on one of the ir2111. 2 and 3 once I got that figured out - I went to pulse the opto osolators to see what happens - when I would activate either of the high mosfets the current to the logic would go to .25 amps. Figured that wasn't right :). finally after checking with the ohm meter - I found out that both high side gates where grounded w
[01:00:25] <Jymmm> question... whats the big deal on encoders??? The number of slots (divisions) on the wheel? The phototransistor? soemthing else?
[01:01:20] <jepler> Jymmm: The number of counts per turn is a characteristic of the wheel.
[01:01:22] <fenn> the big deal?
[01:01:45] <fenn> also, a given sensor can only change so fast
[01:02:11] <fenn> so you cant spin a mouse wheel at 500krpm and expect it to keep count
[01:02:42] <Jymmm> is that a limitation of the sensor or the pic/uC ?
[01:02:48] <fenn> a limitation of the sensor
[01:03:05] <fenn> cheap optosensor rise/fall time is probably on the order of 10 us
[01:03:12] <fenn> *wild guess*
[01:03:22] <Jymmm> ok, so how fast switching does it need for say 1M rpm
[01:03:41] <SkunkWorks> 1000000rpm?
[01:03:42] <jepler> with a micro, you'll run into the processor speed limit before the rise- and fall-time limits of the sensor.
[01:04:07] <Jymmm> SkunkWorks fenn was exaduarting at 500,000 rpm, so I uped the ante =)
[01:04:46] <Jymmm> Atmel has a 4MHz or better clk
[01:04:47] <jepler> e.g., right now I'm struggling to read 4 quadrature inputs and do a tiny amount of processing, anticipating a <160kHz input rate
[01:05:00] <jepler> so that's 25 clocks per quadrature transition @ 16 MHz
[01:05:31] <SWPadnos> jepler, have you considered rol and ror? ;)
[01:05:36] <Jymmm> jepler whats that in rpm's ?
[01:05:53] <fenn> you'd be a fool not to use the 20mhz chips tho
[01:06:19] <SWPadnos> only certain ones are available in 20 MHz versions, and I'm not sure they're the high pin count ones
[01:06:51] <jmkasunich> jepler: what are you doing? is this the quadrature divider thing? or something else?
[01:07:03] <jepler> Jymmm: 1200 mm/min (50 in/min), 1mm leadscrew, 3:1 pulley, 500 cycles per revolution -> edges at 120kHz
[01:07:07] <jepler> jmkasunich: yeah, it's cradek's board
[01:07:18] <jmkasunich> you don't like the LUT approach?
[01:07:36] <fenn> SWPadnos: atmega48 is the new version of the atmega8
[01:07:37] <jepler> jmkasunich: perhaps I haven't heard of your LUT approach
[01:07:46] <SWPadnos> I think 3 cycles will be the minimum. lsr/ror, then brcc $+2 / sbr (0x01 or 0x80, depending on shift direction)
[01:07:54] <jmkasunich> I kinda explained it to cradek, but maybe not well
[01:08:07] <jmkasunich> I think the whole thing can be done in 10-12 clocks per channel
[01:08:27] <jmkasunich> (read quadrature, output slower quadrature)
[01:08:35] <jmkasunich> haven't thought about the index pulse yet tho
[01:08:50] <jepler> wouldn't you just pass the index pulse through?
[01:08:55] <fenn> oh, i guess they dont have the 40 pin dip package, damn
[01:09:23] <jmkasunich> if the regular pulses are too short for the PC to count, the index pulse is probably too short for it to see
[01:10:01] <SWPadnos> the mega644 is the only one I see (in DigiKey) that is 20Mz and has 32 I/Os
[01:10:03] <SWPadnos> MHzx
[01:10:09] <SWPadnos> MHz
[01:10:20] <jmkasunich> you'll have to stretch it, to at least base-period, 2x base-period would be better
[01:11:02] <jepler> jmkasunich: oh, I see what you mean now. Add a '555 and screw doing it on the microcontroller, I say
[01:11:33] <jmkasunich> what?!?
[01:11:36] <SWPadnos> just count full cycles instead of quadrature
[01:11:49] <jmkasunich> does the uC have two spare pins?
[01:14:05] <fenn> glad to see you guys getting interested in my project :)
[01:14:15] <cradek> haha
[01:14:21] <cradek> we're all on the same boat aren't we
[01:14:32] <SWPadnos> same handbasket, I'd say ;)
[01:14:37] <cradek> yep
[01:14:53] <Jymmm> fenn: Nah, just tired of hearing you whine about it is all =)
[01:15:05] <Jymmm> fenn (lol, couldn't resist)
[01:15:08] <fenn> i split it up into 3 jobs, one micro to count the encoders, one to talk to the first micro and do pwm and assorted sundries, and one to talk to the PC
[01:15:27] <SWPadnos> "where are we going, and what are we doing in this handbasket?" -- Rodney Dangerfield
[01:15:33] <jmkasunich> jeez, fenn, uC's grow on trees?
[01:15:36] <fenn> yep
[01:15:48] <fenn> there's only 2 per motor
[01:15:48] <jepler> jmkasunich: a yuppie food coupon will get you a dozen of the damned things
[01:15:57] <cradek> I've also used two in one project, but usually because the program space fills up
[01:16:00] <jepler> pretty soon they'll be sprinkled like sand
[01:16:03] <fenn> and at $0.88 each its not too bad
[01:16:11] <jepler> you'll brush them off the table like crumbs
[01:16:32] <fenn> but this way you have a 20mhz micro doing nothing but watching encoders
[01:16:39] <jmkasunich> yuppie food coupon? thats a new one for me
[01:16:49] <cradek> $20 bill
[01:17:21] <jmkasunich> fenn: I was brainstorming yesterday, trying to fit 4 encoders, 4 PWM, and a parport interface into one uC
[01:17:21] <SWPadnos> and talking on the SPI bus, which reduces the attainable count rate
[01:17:27] <fenn> also i wasnt looking forward to interleaving all the code
[01:17:44] <SWPadnos> with or without Z?
[01:17:47] <SWPadnos> (index)
[01:17:56] <fenn> jmkasunich: really the best solution would be to use a chip with hardware encoder counters, like dspic40f11
[01:18:06] <SWPadnos> DSP56805
[01:18:10] <SWPadnos> E
[01:18:14] <fenn> er, 30f4011? *checks*
[01:18:24] <jmkasunich> how cheap are those?
[01:18:35] <fenn> prolly between $10 and 20
[01:18:43] <jmkasunich> its real easy to go from $1 uCs to $5 uCs
[01:18:58] <jmkasunich> and when you start putting multiple $5 uCs on a board, it ain't cheap anymore
[01:19:01] <SWPadnos> 4 quadrature counters (I think), dual 6-channel PWM, 8xA/D with 2-channel simultaneous capture (1.2 uSec), etc
[01:19:14] <fenn> yes but you're pushing the limits by doing it in software already
[01:19:28] <jmkasunich> SWP: 1.2uS ADCs? really?
[01:19:36] <jmkasunich> now that gets interesting
[01:19:38] <SWPadnos> I think so
[01:19:43] <jmkasunich> * jmkasunich sees current loop in the uC
[01:19:45] <SWPadnos> can't remember if that's per sample or per pair
[01:19:56] <SWPadnos> 12 bit, too
[01:20:00] <SWPadnos> IIRC
[01:20:11] <jmkasunich> don't need the bits, need the speed
[01:20:17] <SWPadnos> $17.76 in singles at DK
[01:20:20] <jmkasunich> ouch
[01:20:23] <SWPadnos> heh
[01:20:32] <SWPadnos> but it is a dual 3-phase motor driver DSP
[01:20:38] <jmkasunich> that would definitely have to be a one-per-board thing, not a one-per-motor thing
[01:20:39] <SWPadnos> runs at 80 MHz
[01:20:58] <jmkasunich> yeah, and if I was building a 100HP (or even 1HP) AC drive it would be a good deal
[01:21:14] <jmkasunich> but not for a servo etch-a-sketch or sherline size driver
[01:21:15] <SWPadnos> the 56F801 is only $8.77
[01:21:36] <SWPadnos> inly has the cojones for 1 3-phase brushless motor though
[01:22:04] <fenn> hmm 48-lqfp packaging could be annoying
[01:22:11] <SWPadnos> interesting - there's an 802 now as well
[01:23:02] <cradek> I can solder lqfp but would rather not
[01:23:09] <SkunkWorks> * SkunkWorks still likes the idea of using the pc to do it.
[01:23:36] <cradek> I wish a simple divider could be made with a few generic logic chips
[01:23:46] <cradek> that's where I started, and it was a nice place
[01:24:09] <SkunkWorks> I only need 2000 edges per inch.
[01:24:22] <SWPadnos> ok - the 801 has 6 PWM outputs (3 phases * high/low drive, with deadtime), dual quadrature counters, 8xADC, and a bunch of other I/O
[01:24:54] <Jymmm> Um... if we're using all these uC's and whatnot, and goin into a computer anyway... why not make a PCI board that has everything one would need w/o the limitations of paraport, serial, usb, etc?
[01:25:05] <SWPadnos> oh shut up
[01:25:07] <SWPadnos> ;)
[01:25:07] <jmkasunich> cradek: define "generic logic chips" and "few"
[01:25:10] <fenn> Jymmm: because PCI is a big pain
[01:25:19] <SWPadnos> Jymmm, define "everything"
[01:25:25] <fenn> m5i20
[01:25:31] <SWPadnos> even better: define "everything one would need"
[01:25:43] <jmkasunich> gimme a 256x8 prom, and one 74HC374 per channel and I'll divide quadrature
[01:25:52] <fenn> m5i20 and free spartan3 programming software
[01:25:54] <SkunkWorks> * SkunkWorks is seriously thinking about lasercutting "mouse wheel" with the correct number of holes
[01:25:56] <SWPadnos> yep - the M5I20 is cool. and even configurable if you have Xilinx FPGA tools
[01:26:15] <jmkasunich> the LUT algorithm in the mega8 is exactly the same thing but only needs one chip
[01:26:20] <fenn> by free i mean libre
[01:26:48] <SWPadnos> you should only need a 16 element lookup per channel (unless you use index, which would need 64)
[01:27:00] <cradek> I meant something that you don't have to worry about being fast enough
[01:27:31] <jmkasunich> the 374 version could probably handle a megahurts (with 3 channels)
[01:27:42] <fenn> cradek: you can divide by 4 with a flipflop
[01:27:49] <A-L-P-H-A> GROUP Question... if I'm only getting 10ipm with EMC, after all the tweaking I can do... 10 uStepping (Gecko G201), with an AMD 1700XP, EMC, 5TPI ballscrew, 1.6 ration gear reduction... should I get an gecko G10X, or the step multiplier?
[01:27:50] <jmkasunich> but you're talking 5 chips probably
[01:28:16] <jmkasunich> A-L-P-H-A, thats not a 10 second question
[01:28:20] <jmkasunich> gotta do the math
[01:28:27] <jmkasunich> (did you do the math already?)
[01:28:35] <SkunkWorks> last time it was 150000 steps per inch.
[01:28:38] <fenn> that sound like something's wrong
[01:28:47] <jmkasunich> whats wrong is 150000 steps per inch
[01:28:49] <A-L-P-H-A> it was something retardedly high.
[01:28:53] <SWPadnos> 16000
[01:29:24] <jmkasunich> 5 tpi * 200 steps * 10usteps * 1.6 = 16000 all right
[01:29:25] <SWPadnos> you should very easily be able to get ~60 IPM with that setup
[01:29:40] <jmkasunich> so 60 IPM = 1 ips = 16000 steps/second
[01:29:51] <jmkasunich> base period of 20uS should be doable with that CPU
[01:29:57] <SWPadnos> right - should be easy for an Athlon XP 1700
[01:30:07] <jmkasunich> which gives a theoretical max of 25000 steps/sec
[01:30:17] <A-L-P-H-A> ODD... that's not happening though.
[01:30:27] <jmkasunich> well, what is your base period set at?
[01:30:34] <A-L-P-H-A> Let me grab my config file.
[01:30:35] <jmkasunich> what is your max vel set at?
[01:30:43] <jmkasunich> what is your accel set at?
[01:30:51] <cradek> put the ini on pastebin.com
[01:31:10] <SWPadnos> methinks there's an extra 0 in the base period
[01:31:15] <SWPadnos> 200 uS instead of 20
[01:31:26] <cradek> yeah I bet something silly is wrong
[01:31:26] <fenn> * fenn bets it needs headroom in stepgen
[01:31:27] <Jymmm> what is your PIN number and DOB?
[01:31:28] <jmkasunich> or maybe hes just using the default, which is 50uS
[01:31:41] <jepler> A-L-P-H-A: are you using emc2, or the one that comes with bdi4?
[01:31:42] <SWPadnos> need ~180 uS for htat low a rate
[01:31:43] <SWPadnos> tha
[01:31:45] <SWPadnos> that
[01:31:52] <A-L-P-H-A> bdi
[01:31:59] <cradek> oh
[01:31:59] <jmkasunich> bdi
[01:32:00] <A-L-P-H-A> brb... booting the emc machine
[01:32:07] <jmkasunich> * jmkasunich goes away
[01:32:22] <cradek> we're not able to be much help with bdi4emc
[01:32:31] <fenn> oh bollocks
[01:32:42] <fenn> it runs, its just a performance issue
[01:33:03] <jmkasunich> to be honest, tho I don't know the details, the "quickstep" algorithm that the newest bdi's have should be slightly better than emc2
[01:33:18] <jmkasunich> but what hes talking about is well within the capabilities of emc2
[01:33:21] <cradek> i've never run/tuned it, and it uses pid that's hairy
[01:33:34] <jmkasunich> fenn: bollocks yourself
[01:33:48] <fenn> jmkasunich: how many times?
[01:33:49] <jmkasunich> I don't use it, never have used it (or the emc1 that its based on)
[01:34:24] <cradek> I used to use emc1, but it's been a while, never used bdi4emc (or bdi)
[01:34:26] <jepler> cradek: I read in the bullet list that the use of pid is one of the things that was fixed in bdi4emc's quickstep, but having never used it, ...
[01:34:42] <cradek> jepler: ok
[01:35:08] <SWPadnos> right - it works like stepgen, but at a lower level than an RTAI app
[01:35:18] <SWPadnos> (from the docs - I've also never used it)
[01:35:34] <SkunkWorks> * SkunkWorks thinks if there wasn't enough headroom - there would be following errors.
[01:35:45] <cradek> SkunkWorks: it's not emc2
[01:35:45] <jmkasunich> so IOW, all of us understand in general how it works, but none of us know the details and none of us has used it
[01:35:46] <fenn> at a lower level than rtai?
[01:36:00] <jmkasunich> hence, we're not in a good position to help A-L-P-H-A
[01:36:15] <SWPadnos> kinda next to RTAI, as I understand it (or under it, executing as the actual timer interrupt vector)
[01:36:25] <SkunkWorks> emc1 wow
[01:36:26] <jmkasunich> something like that
[01:36:59] <jepler> I'm happy to try to help, but afraid of giving bad advice unintentionally
[01:37:01] <jmkasunich> it uses x86 assembly language, makes no pretense at portability, but works well on that platform
[01:37:29] <SkunkWorks> by the way - the 150k step per inch was my mistake. when we had this conversation before - I though 10 step microstepping made the 200 steps 20000 steps for some reason. sorry.
[01:38:16] <SWPadnos> bad SkunkWorks, bad SkunkWorks
[01:38:41] <SkunkWorks> made perfect sense at the time - someone else agreed with me :)
[01:38:54] <SWPadnos> heh
[01:39:01] <cradek> hmm, all this debate and A-L-P-H-A has disappeared
[01:39:07] <SWPadnos> booting
[01:39:17] <SWPadnos> in the shop, I guess
[01:39:37] <jmkasunich> ksweeny - thats a new name (I think)
[01:43:29] <giacus> G'night all
[01:51:41] <A-L-P-H-A> nonono
[01:51:47] <A-L-P-H-A> I'm here... well... in the other room
[01:51:55] <A-L-P-H-A> 1st i couldn't remember my password. duh.
[01:52:27] <A-L-P-H-A> 2nd, I can't mount a usb key, or a floopy drive... [I forget how to mount USB without X apps] [i can't find the floopy device in /dev
[01:52:36] <A-L-P-H-A> 3rd, I'm back here, at a lost. :(
[01:53:24] <A-L-P-H-A> I've got some ubuntu thing going...
[01:53:49] <A-L-P-H-A> should have been FD0... let me try again...
[01:59:09] <Jymmm> SWPadnos by "everything" I mean I/O
[01:59:13] <SWPadnos> heh
[01:59:36] <Jymmm> shitloads of I/O =)
[02:00:00] <SWPadnos> mesa: 72 I/O
[02:00:12] <SWPadnos> inlcuding special functions though - only 48 or so unencumbered
[02:00:55] <Jymmm> No, no, I mean like 128 or 256 I/O
[02:02:35] <A-L-P-H-A> ARGH
[02:02:37] <A-L-P-H-A> I hate that box.
[02:02:48] <A-L-P-H-A> it's going to die today... after I get the ini settings.
[02:02:52] <A-L-P-H-A> ARGH!
[02:02:57] <Jymmm> * Jymmm hands A-L-P-H-A a Jack-in-the-box coupon
[02:03:18] <A-L-P-H-A> I've gotta make a ethernet cable to connect that stupid box now too.
[02:03:19] <A-L-P-H-A> bbiab.
[02:03:42] <SWPadnos> heh - it might be faster to install ubuntu
[02:03:58] <Jymmm> * Jymmm hands A-L-P-H-A a box of cat5 and some RJ11 connectors
[02:05:49] <Jymmm> Okey, nobody got that joke...
[02:07:55] <SWPadnos> we use cat6 now, you insensitive clod!
[02:08:20] <Jymmm> lol
[02:08:38] <A-L-P-H-A> no... I'm still on cat5.
[02:08:45] <Jymmm> ROTF
[02:08:47] <A-L-P-H-A> cause I've wired the house already.
[02:09:01] <A-L-P-H-A> I just found a super long cat5 cable...
[02:09:04] <A-L-P-H-A> damn computer.
[02:09:08] <A-L-P-H-A> it's being laggy
[02:09:10] <A-L-P-H-A> Time: 22:12:07 -0500 GMT, Windows XP Professional, Service Pack 2 (5.1 - 2600), AMD XP 2800+ running at (1-AMD , 2123MHz, 512KB (100% Load)), DDR400 RAM Usage: 738/1024MB (72.07%), C: 12.14gb of 189.91gb free, E: 31.07gb of 372.62gb free, Current Uptime: 7hrs 31mins 1sec, Record Uptime: 3wks 12hrs 48mins 50secs, 3 Samsung 19" flat monitors (1 Trinitron, 2 CRT).
[02:09:17] <A-L-P-H-A> oh. I should update that thing.
[02:09:49] <A-L-P-H-A> Time: 22:12:46 -0500 GMT, Windows XP Professional, Service Pack 2 (5.1 - 2600), AMD XP 2800+ running at (1-AMD , 2123MHz, 512KB (100% Load)), DDR400 RAM Usage: 739/1024MB (72.17%), C: 12.14gb of 189.91gb free, E: 31.07gb of 372.62gb free, Current Uptime: 7hrs 31mins 40secs, Record Uptime: 3wks 12hrs 48mins 50secs, 3 Viewsonic 17" LCDs.
[02:16:40] <A-L-P-H-A> dang. there are 2gig SD cards now
[02:16:52] <Jymmm> 4gb I believe
[02:16:58] <A-L-P-H-A> yikes...
[02:17:18] <A-L-P-H-A> that's like 2000 photos before needing to remove the files, at 3megapixels.
[02:17:26] <A-L-P-H-A> with no low compression
[02:18:09] <A-L-P-H-A> bbiab
[02:29:42] <A-L-P-H-A> why's there no FTP client in kde?
[02:30:23] <SWPadnos> konquerer?
[02:30:30] <A-L-P-H-A> does that send?
[02:30:37] <SWPadnos> dunno
[02:30:47] <SWPadnos> how about scp?
[02:30:56] <A-L-P-H-A> no clue...
[02:31:02] <SWPadnos> or are you a GUI-only kind of guy?
[02:31:05] <A-L-P-H-A> I'll try and log onto here. and dcc it to myself
[02:31:08] <A-L-P-H-A> No.
[02:31:11] <A-L-P-H-A> I used to be DOS only.
[02:31:22] <A-L-P-H-A> but I don't remember the ftp commands.
[02:31:34] <A-L-P-H-A> brb
[02:31:58] <SWPadnos> scp user@host:/file/name .
[02:36:24] <joeKr> A-L-P-H-A: Would kGet work for you? (I've never used it, I use gFTP)
[02:36:26] <A-L-P-H-A> oh fuck wow.
[02:36:36] <A-L-P-H-A> joeKr, sorry... that wasn't directed to you
[02:36:43] <A-L-P-H-A> I finally got EMC and my gcode off that computer.
[02:36:46] <A-L-P-H-A> GAH!!!!
[02:36:58] <Jymmm> A-L-P-H-A ?
[02:36:59] <joeKr> A-L-P-H-A: no problem
[02:37:21] <Jymmm> * Jymmm hasn't been paying attention
[02:37:37] <A-L-P-H-A> http://pastebin.com/748320
[02:37:42] <A-L-P-H-A> that's my current EMC config.
[02:37:58] <A-L-P-H-A> which as last mod, MAY 9th, 2005.
[02:38:08] <A-L-P-H-A> So really, I haven't played with the machine much for over a year
[02:38:11] <A-L-P-H-A> sad.
[02:38:25] <cradek> this is the one that's bdi4emc and too slow?
[02:38:34] <A-L-P-H-A> yes
[02:38:36] <A-L-P-H-A> bdi4emc.
[02:38:44] <A-L-P-H-A> sec.
[02:39:06] <cradek> your scales are only 1000, I thought they were 16000
[02:39:13] <cradek> are you sure it's the right config?
[02:39:34] <A-L-P-H-A> NO
[02:39:53] <A-L-P-H-A> this is new. http://pastebin.com/748324
[02:39:54] <A-L-P-H-A> newer
[02:41:10] <A-L-P-H-A> what was the new ubuntu ISO image filename?
[02:42:10] <A-L-P-H-A> found the emc2-ubutnut-livecd.iso
[02:42:30] <SWPadnos> that should work
[02:43:01] <SWPadnos> the period is 24 microseconds, you should be able to get ~20 KHz out of that setup (or around 75 IPM)
[02:44:06] <cradek> so what's the symptom here?
[02:44:17] <A-L-P-H-A> cradek, emc4bdi is slow.
[02:44:21] <A-L-P-H-A> BTW... CORRECTION on the computer.
[02:44:29] <A-L-P-H-A> it's a *********duron 700!!!***********
[02:44:29] <A-L-P-H-A> :(
[02:44:39] <cradek> no I mean what goes wrong when you turn up your velocity
[02:44:44] <A-L-P-H-A> with 512 megs ram.
[02:44:48] <A-L-P-H-A> cradek, missing steps.
[02:45:01] <A-L-P-H-A> I'm going to rewire my steppers in parallel as well.
[02:45:06] <A-L-P-H-A> or.
[02:45:11] <cradek> what do you mean missing steps
[02:45:13] <A-L-P-H-A> I mean in series, to get faster stuff.
[02:45:15] <cradek> it ends up in the wrong place?
[02:48:06] <jmkasunich> wait just a cotten pickin minute here
[02:48:20] <jmkasunich> you just said ubuntu, and bdi4emc
[02:48:25] <jmkasunich> so which one is it?
[02:49:27] <cradek> this ini is not an emc2 ini
[02:51:06] <cradek> MAX_VELOCITY = 0.13
[02:51:12] <cradek> and its set really slow.
[02:51:22] <asdf23r> asdf23r is now known as A-L-P-H-A
[02:51:31] <A-L-P-H-A> compy crashed. lovely.
[02:51:47] <A-L-P-H-A> burning the emc ubuntu live CD now
[02:51:50] <SWPadnos> SWPadnos is now known as SWP_Away
[02:52:00] <A-L-P-H-A> does a duron have enough of a pulse train to work??
[02:52:10] <A-L-P-H-A> duron (amd) = celeron (intel)
[02:52:30] <cradek> I wonder: what makes you sure this is a software problem? missing steps (ending up in the wrong place) is probably not software
[02:52:45] <A-L-P-H-A> cradek, could be a combo of things.
[02:53:08] <A-L-P-H-A> could be that this is a duron 700 mhz computer, and I've wired my steppers in parallel (which ever method was used to get more torque)
[02:53:20] <cradek> A-L-P-H-A: I recommend you test to see what your problem is before you blindly start changing stuff
[02:54:17] <A-L-P-H-A> cradek, I think'll start with rewiring my steppers to series, and then work from there.
[02:54:20] <A-L-P-H-A> you maybe right.
[02:54:37] <cradek> so when you try to move too fast, the steppers stall (table stops moving)?
[02:54:48] <A-L-P-H-A> yes.
[02:54:50] <A-L-P-H-A> Or.
[02:54:51] <fenn> hmm i resemble that comment about a duron 700
[02:55:00] <fenn> (its my main dev machine)
[02:55:03] <A-L-P-H-A> resemble?
[02:55:04] <cradek> that's unlikely to be software.
[02:55:20] <cradek> unless it's really broken
[02:55:21] <A-L-P-H-A> resent.
[02:55:33] <fenn> seems like overkill for a milling machine
[02:55:38] <A-L-P-H-A> hmmmm.
[02:55:49] <fenn> i run 20 us on a p233
[02:55:59] <A-L-P-H-A> okay.
[02:56:02] <A-L-P-H-A> so it's probably my motors then.
[02:56:04] <cradek> my mill is on a PIII 666 at 20us
[02:56:09] <cradek> or drivers or power supply
[02:56:11] <A-L-P-H-A> let me see if i can pull any info off these motors.
[02:56:17] <cradek> or a mechanical problem (no lube)
[02:56:20] <A-L-P-H-A> drivers are gecko g201s
[02:56:35] <A-L-P-H-A> ball screws, spins well.
[02:57:01] <cradek> ok
[02:57:15] <cradek> motors and power supply appropriately sized?
[02:57:27] <A-L-P-H-A> should be.
[02:57:33] <A-L-P-H-A> let me pull all the info together.
[02:57:34] <cradek> you could put a scope on your step input to an axis and jog it
[02:57:42] <A-L-P-H-A> <-- owes no scope
[02:57:45] <cradek> oh
[02:58:08] <cradek> you should also run the latency tests on your machine
[02:58:12] <cradek> the rtai latency tests
[02:58:20] <cradek> maybe your machine is bogus for realtime which WILL cause stalls
[02:58:40] <alpha1125> okay at the emc machine
[02:58:44] <fenn> what are the symptoms again?
[02:58:49] <cradek> stalls
[02:59:16] <fenn> like, the computer stops responding? or the steppers lose steps?
[02:59:30] <cradek> the axis stops moving and ends up in the wrong place if he tries to go too fast
[02:59:33] <alpha1125> this is a 2amp 100VA transformer... 120VAC in.
[02:59:46] <alpha1125> 3 gecko G201s
[03:00:01] <fenn> does the problem get worse if you move more than one motor at once?
[03:00:57] <alpha1125> 4.2V 3.5A steppers for X and Y.
[03:01:26] <alpha1125> 3V 1ohm for the Z.
[03:01:34] <alpha1125> No... even if I move one axis.
[03:01:57] <fenn> ok so its probably not the power supply then
[03:02:03] <alpha1125> say I move the axis in the X at 1ipm for 1"... it moves 1.000 inches.
[03:02:22] <alpha1125> if I move that same axis 1" at 10ipm, it moves 0.854sdafasdf inches
[03:02:46] <cradek> how does that move sound/look?
[03:02:48] <alpha1125> I think if maybe I'm moving the motor TOO fast?
[03:03:08] <alpha1125> sounds and looks fine... smooth.
[03:03:21] <fenn> what if you move 2"
[03:03:26] <alpha1125> just that it's missing some 0.1" or something
[03:03:26] <fenn> at 10 ipm
[03:03:39] <alpha1125> same diff.
[03:03:48] <fenn> same number of steps lost?
[03:03:51] <alpha1125> it's late here now... 11pm... I'll work on it in the morning.
[03:04:01] <alpha1125> fenn, I'll try that tomorrow
[03:04:07] <alpha1125> going back to the workstation
[03:04:24] <jepler> goodnight all
[03:04:50] <Jymmm> g'night jepler
[03:05:02] <A-L-P-H-A> night jepler
[03:05:15] <fenn> pff all you diurnal people
[03:05:29] <A-L-P-H-A> quit using words that make me use the dictionary.
[03:06:32] <fenn> just when it started to get interesting "oh i'm tired"
[03:06:38] <A-L-P-H-A> I used to be a nocturnal ... till I needed to start looking for a job.
[03:06:56] <cradek> SPINDLE_REVERSE_INDEX = -3
[03:07:05] <A-L-P-H-A> sorry... that machine hasn't been turned on in ages.
[03:07:11] <cradek> strange, I wonder if bdi4emc adopted my old hackery
[03:08:38] <A-L-P-H-A> when was the last time paul_c was around? haven't seen him in ages.
[03:09:06] <fenn> few weeks ago
[03:09:15] <fenn> he's on #morphix a lot i think
[03:12:20] <A-L-P-H-A> k, I'm gonna go to sleep... night everyone
[05:25:19] <Jymmm> fenn wake up!
[05:25:33] <fenn> * fenn mumbles
[05:25:39] <fenn> i dont wanna go back to school
[05:25:49] <Jymmm> fenn: make me look up diurnal will ya!
[05:26:36] <Jymmm> fenn how ya doin?
[05:27:07] <Jymmm> I sent the fire dept on a wild goose chase!
[05:27:51] <fenn> oh i'm just learning that i dont know anything for the 50th time
[05:28:21] <Jymmm> what dont ya know now (and did you once know it) ?
[05:29:12] <fenn> how to make a motor driver
[05:29:27] <Jymmm> stepper or servo?
[05:29:37] <fenn> servo, should be easy right?
[05:30:05] <Jymmm> No clue... I would have thought stepper, but I know less than you =)
[05:30:18] <Jymmm> * Jymmm dont know nuttin, just lives here.
[05:30:28] <fenn> hehe sounds familiar
[05:33:24] <Jymmm> I was thinking about taking one of the common stepeper circuits off the web and maybe making a few boards just to be able to have soemthing to play around with.
[05:33:37] <fenn> that sounds like a good idea
[05:34:07] <Jymmm> I thought I'd route my oen, if they test good, them have soem made up.
[05:34:36] <fenn> you should try the toner transfer method if you have any interest in electronics
[05:35:02] <Jymmm> I have a gantry router
[05:35:10] <fenn> yes that works too
[05:35:16] <fenn> or, could work at least
[05:35:39] <Jymmm> especially with the hole drilling
[05:36:18] <fenn> wowo i've managed to nest myself 3 layers of chicken/egg
[05:36:29] <fenn> in a triangle no less
[05:37:00] <Jymmm> in eagle?
[05:37:05] <fenn> no, nevermind
[05:37:09] <Jymmm> lol
[05:38:30] <fenn> * fenn reads, and reads, and reads
[05:38:40] <fenn> my friends ask me what i do all day
[05:39:07] <fenn> i say "what? i've got 50 projects, can't you see?" pointing at the empty workbench
[05:40:07] <Jymmm> Well damnit.... throw a bunch of this on the bench then!
[05:40:18] <Jymmm> You know.... a clean bench is a sgin of a sick mind!
[05:40:22] <Jymmm> sign
[05:40:36] <fenn> well, i said "empty" not clean
[05:40:41] <fenn> there's a hole in the middle of all the junk
[05:40:52] <fenn> whereas if i was working on something there'd be junk there too
[05:40:58] <Jymmm> no holes! what you think this is a golf course?!
[05:43:14] <Jymmm> Oh btw... I REALLY did send the fire dept on a wild goose chase.... I could see heavy smoke due north of me. They couldn't find it but they did smell it, so they came back and looked off the balcony. Then about 15 minutes later they came back again to see if it was still there.
[05:43:56] <fenn> they were standing on your balcony?
[05:45:50] <Jymmm> Yeah... when they were lookign for it, they could smell it, but couldn't see it. So they came here (2nd story balcony) and could see the smoke.
[05:46:04] <Jymmm> What's really weird is nobody else called it in.
[05:46:24] <Jymmm> 1,000,000 damn ppl in this coty and nobody else saw it?! wth
[05:46:29] <Jymmm> city
[05:47:01] <Jymmm> what this NY or what?!
[05:47:11] <Jymmm> (the attitude I mean)
[05:47:54] <fenn> did they ever find anything?
[05:48:27] <Jymmm> Nope, that's why they came back a 2nd time.... to see if they could still see if there was smoke (there wasn't by this time).
[05:49:20] <Jymmm> Said to call back if I see it again.
[05:52:32] <Jymmm> I'm a pyro, but I got this thing abotu fire too.
[06:49:06] <alex_joni> http://tribewanted.com/index.htm
[06:52:02] <Jymmm> uh... oh hell no! I'm very fond of toilet paper and my toothbrush
[06:52:13] <alex_joni> :-)
[06:52:50] <alex_joni> get a stash of those with you
[06:53:02] <alex_joni> but I think 5000 people will be anarchy
[06:53:24] <Jymmm> Did I forget to mention I'm also extremely fond of hot and cold running water?
[06:53:37] <Jymmm> Not really, very much like Burning Man
[06:53:40] <alex_joni> there should be plenty of cold running water
[06:53:55] <Jymmm> fresh, not seasoned
[06:54:48] <Jymmm> oh, and biohazard free
[07:19:22] <alex_joni> Jymmm: how's your canny?
[07:19:42] <Jymmm> canny?
[07:19:51] <alex_joni> can based bipod :D
[07:19:56] <alex_joni> spray can
[07:20:32] <Jymmm> Oh, heh... was just doin some home work to see if it was easily doable.
[07:20:48] <Jymmm> I'll jsut start keeping my eye out for steppers and other parts now
[07:20:48] <alex_joni> it should..
[07:20:55] <alex_joni> depends on the motors mainly
[07:20:56] <alex_joni> right
[07:21:22] <Jymmm> NEMA23 should be big enough I'd think, if that.
[10:22:23] <anonimasu> h
[10:22:25] <anonimasu> hmm
[11:48:55] <giacus> hello
[11:49:01] <giacus> alex_joni: around ?
[12:22:42] <giacus> jepler: you familiar with pymedia ?
[12:23:02] <jepler> giacus: nope
[12:23:13] <giacus> ok
[12:23:32] <giacus> I'm just attempting to compile it from latest tarball
[12:23:54] <giacus> but it cand find FAAD and MP3LAME libs
[12:24:18] <fenn> maybe you should get them
[12:24:31] <giacus> where ?
[12:24:34] <giacus> http://rafb.net/paste/results/3tI7sJ79.html
[12:24:52] <giacus> have no more libs to install on debian testing
[12:25:16] <giacus> here's the issue: http://rafb.net/paste/results/nmEKJR18.html
[12:26:06] <jepler> why do you want to use non-free formats?
[12:26:16] <jepler> why not just continue with the ogg format that was detected.
[12:26:36] <giacus> jepler: I just need to convert a swf video to mpeg
[12:26:53] <giacus> and I'm following the pymedia requirements
[12:27:34] <jepler> oh
[12:28:02] <giacus> in order to get vnc2swf.py and edit.py converting the swf video
[12:28:39] <giacus> I'm going to record a desktop session with vnc2swf.py
[12:28:56] <giacus> and then convert them to use in cinelerra for a videoclip
[12:29:14] <giacus> emc2 videoclip :D
[12:29:54] <giacus> also tried instanbul, a desktop recorder for gnome
[12:30:15] <giacus> it doesnt worn under Ion3, because there's no tray bar
[12:30:54] <jepler> this package contains a "libfaad.a": http://packages.debian.org/stable/libdevel/libvlc0-dev
[12:30:55] <fenn> cant you run it as a background process?
[12:31:06] <giacus> jepler: stable ?
[12:31:12] <giacus> i'm using testing..
[12:31:24] <giacus> looking
[12:31:48] <giacus> fenn: cant get Istanbul running over here
[12:31:59] <giacus> just run in silent mode
[12:32:13] <giacus> cant see any window or icon ..
[12:32:42] <giacus> Ion3 support some kde tray bar apps, but not gnome
[12:34:16] <giacus> jepler: that package is not available on testing
[12:35:44] <giacus> fenn: btw. I succesful recorded the session using vnc2swf.py
[12:36:03] <giacus> the difficult is to convert it to mpeg
[12:37:00] <giacus> morning les_w :)
[12:37:39] <les_w> morning jacky
[12:38:25] <giacus> how are the plants ?
[12:38:35] <giacus> going up ?
[12:38:47] <les_w> pretty well
[12:38:53] <giacus> nicee
[12:39:03] <les_w> tomatoes, basil, cucumbers
[12:39:06] <les_w> peppers
[12:39:13] <giacus> wow hehe
[12:39:19] <giacus> I love peppers
[12:39:24] <giacus> the smallest
[12:39:34] <giacus> are evil peppers :P
[12:39:39] <giacus> very strong
[12:39:50] <les_w> they grow very hot here
[12:39:59] <giacus> yeah, here too
[12:40:27] <giacus> we use pepper on every food here ..
[12:40:53] <giacus> food withou pepper isnt good for my taste
[12:41:01] <les_w> I just have a few plants...sometimes I grow a lot though
[12:41:07] <les_w> I smoke them...
[12:41:21] <les_w> then make a peeper sauce
[12:41:31] <giacus> good
[12:41:34] <les_w> pepper, salt, vinegar, garlic
[12:41:43] <les_w> peeper haha
[12:42:02] <giacus> yes
[12:42:09] <giacus> I also like mexican food
[12:42:36] <giacus> very similar to the our
[12:42:38] <les_w> me too
[12:42:55] <giacus> maditerraneo :)
[12:43:00] <giacus> mediterraneo*
[12:43:13] <les_w> hmm
[12:43:26] <giacus> les_w: too cook is an art
[12:43:34] <giacus> are you clever to coock ?
[12:43:50] <jepler> that sounds like a great spam subject
[12:44:01] <giacus> hahaha sorry
[12:44:02] <jepler> Subject: are you clever to coock ?
[12:44:16] <giacus> missed some word ? :(
[12:44:24] <les_w> well...better get to work I guess...have many business visitors coming
[12:44:50] <les_w> the engineering staff of ITW finishing systems is coming
[12:44:53] <giacus> les_w: good, have fun ;)
[12:44:57] <les_w> I should have a cookout
[12:45:02] <les_w> later
[12:47:01] <giacus> jepler: I found to cook is a bit as to develop
[12:48:02] <giacus> you have to hack recipe in order to get a cool results :)
[12:48:47] <giacus> in that case, you no need to compile pepper_hot.so
[12:48:50] <giacus> :D
[12:51:11] <fenn> ./me configure --bed && goto sleep();
[12:51:14] <giacus> but anything should be done in the right order
[12:51:49] <giacus> fenn: have a emc dreams
[12:52:13] <giacus> or better.. girls dreams ;P
[12:52:58] <giacus> I'm away for a bit too, later
[13:14:04] <les_w> oh, I want to do a sell job on using real time linux for factory automation when my guests come
[13:14:58] <les_w> In particular I want to understand if HAL can be used for such things
[13:15:09] <SWP_Away> yes
[13:15:11] <SWP_Away> SWP_Away is now known as SWPadnos
[13:15:24] <SWPadnos> that's my answer anyway
[13:15:45] <les_w> I need trim to frequency for ceramic resonators
[13:16:13] <les_w> prob laser
[13:16:15] <SWPadnos> get a spectrum analyzer with a precision time reference, and (I shudder to suggest this) LabView
[13:16:53] <les_w> getting an agilent 4294 impedance analyzer today
[13:16:57] <les_w> ordering one
[13:17:01] <SWPadnos> cool
[13:17:42] <SWPadnos> actually, if you get an analyzer with ethernet, it's probably possible to write Linux "drivers" to read back the spectrum analysis plots / points
[13:17:49] <les_w> I don't know why thr things cost so much
[13:18:00] <SWPadnos> it's the NIST cert that costs so much ;)
[13:19:06] <les_w> well, I would need to do some FFTs of impedance, and use a control loop to ablate off mass with a YAG laser
[13:19:31] <SWPadnos> would you want to trim in an oscillator config, or from impedance?
[13:20:46] <les_w> well, it's at least 5th order....pole pair from piezo, pole pair from diaphragm, real pole from radiation reactance
[13:21:45] <SWPadnos> I guess the real question is whether it's better to trim in the expected use configuration, or based on a "theoretical" parameter
[13:22:02] <les_w> would use psuedo random noise sequence I think
[13:22:10] <les_w> lots of feedforward
[13:22:22] <les_w> i.e. a trim algo that learns
[13:24:38] <SWPadnos> how realtime is the trim process?
[13:24:50] <SWPadnos> ie, how many parts per (minute / hour / day)
[13:25:12] <les_w> well, fast
[13:25:20] <les_w> less than a second
[13:25:41] <SWPadnos> hmmm - that'll definitely need specialized equipment
[13:26:02] <les_w> I want it so fsat that the major issue is just robotically moving the parts in and out of the device
[13:26:08] <les_w> fast
[13:26:29] <SWPadnos> how many extra zeroes on the end of the budget? ;)
[13:26:40] <les_w> lots
[13:26:52] <SWPadnos> good. you probably need at least 6 (before the decimal)
[13:27:01] <les_w> heh
[13:27:11] <SWPadnos> it may be possible with less than 7 though
[13:27:34] <les_w> well, I set up such a manufacturing system for the mirror heaters long ago with dos/c
[13:27:41] <les_w> in pcs
[13:27:47] <les_w> it worked well
[13:28:13] <SWPadnos> I'm thinking of the speed at which you need to measure / trim / remeasure
[13:28:40] <SWPadnos> the PC end of it shouldn't be that demanding
[13:28:48] <les_w> it's heisenburg limited
[13:29:20] <les_w> shorter time window=less freq resolution
[13:29:26] <SWPadnos> right
[13:29:38] <SWPadnos> what frequency range are you tuning for?
[13:29:44] <SWPadnos> KHz, MHz, GHz ...
[13:29:53] <les_w> about 50 kHZ
[13:30:03] <SWPadnos> even worse for fast processing
[13:30:16] <les_w> say to 500 ppm
[13:30:20] <SWPadnos> how many trim cycles do you expect to need per part?
[13:30:27] <SWPadnos> oh - 500ppm is very wide
[13:30:34] <les_w> one!
[13:30:47] <SWPadnos> heh - that would be nice.
[13:30:59] <SWPadnos> 500ppm is terrible accuracy though, so it's probably possible
[13:31:41] <les_w> the trim will cause temperature changes that will affect readings , but that can be learned and fed forward in the control system
[13:33:01] <SWPadnos> I see two problems: 1) part placement in the machine (gotta contact it correctly, and in the proper orientation); 2) measurement speed
[13:33:10] <les_w> right
[13:33:50] <les_w> contact needs to be made at an antinode
[13:34:02] <les_w> as defined by a bessel function
[13:34:56] <les_w> measurement will be an autospectrum of the pseudorandom noise
[13:35:07] <SWPadnos> so you have to find the contact points first
[13:35:25] <les_w> well, it's a disk
[13:35:40] <les_w> so the antinodes are circles
[13:35:51] <SWPadnos> ko, just a radial arm motion then
[13:35:54] <les_w> a certain distance from the center
[13:35:57] <SWPadnos> disk drive, anyone?
[13:36:01] <les_w> haha
[13:36:27] <SWPadnos> ctually, an HD voice coil may be the best actuator you can get (if the radius is expected to change)
[13:36:41] <les_w> yes
[13:36:55] <les_w> I have made voice coil servo systems
[13:37:04] <SWPadnos> fast settling too
[13:37:10] <les_w> with 100g accel
[13:37:26] <SWPadnos> I wonder how many G's a hard drive head pulls
[13:37:37] <les_w> a bunch i'll bet
[13:38:35] <SWPadnos> yep
[13:39:27] <les_w> I have to get some stuff together for this
[13:39:50] <les_w> pretty much the entire engineering staff from Finishing will be here
[13:40:24] <les_w> I need to show them the capabilities of rt linux based systems
[13:40:56] <SWPadnos> show them your CNC
[13:41:00] <les_w> yeah
[13:43:14] <les_w> of course, I would not set up such a software system....
[13:43:21] <les_w> people like you would.
[13:43:22] <SWPadnos> ?
[13:43:25] <SWPadnos> heh - of course
[13:44:32] <les_w> i'll just do the math.
[13:44:38] <SWPadnos> that suits me fine ;)
[13:46:06] <les_w> I had better read up more on the HAL docs though
[13:46:46] <SWPadnos> hmmm
[13:47:09] <SWPadnos> the main difference between HAL and emc1 is that HAL is structured like components that get "wired" together
[13:47:17] <les_w> right
[13:47:28] <SWPadnos> rather than making monolithic apps that contain all functions and immutable "connections"
[13:48:12] <SWPadnos> I'm not sure how much more info the docs will give you (exept for some nice tutorials on using the command-line tools to manipulate HAL configurations)
[13:48:25] <SWPadnos> there won't be a lot that's applicable to an industrial process controller
[13:49:14] <les_w> I just need to sell these proprietary manufacturing systems as something modular and not too geeky
[13:49:48] <les_w> doesn't have to be a HAL of course
[13:50:14] <SWPadnos> right
[13:50:26] <SWPadnos> HAL will look pretty geeky, I think (unfortunately)
[13:50:50] <les_w> hmm
[13:51:15] <SWPadnos> it depends on the level of technical knowledge the people you demo it to have
[13:51:23] <les_w> I was hoping it could be presented as just a black box with inputs and outputs
[13:51:52] <SWPadnos> in that case, the geekiness doesn't matter, because the black box is easy to understand
[13:51:57] <SWPadnos> yes, you can present it that way
[13:52:19] <SWPadnos> Ray was teaching it like a patch panel
[13:52:30] <les_w> exactly
[13:52:42] <SWPadnos> each component has a bunch of jacks, you wire them together how you want, and the system does whatever you configured it to do
[13:53:47] <les_w> plug in FFT, pattern recognition, control loop, etc
[13:53:55] <SWPadnos> sure
[13:54:16] <SWPadnos> though the HAL stuff would be more suited to the controls only - the FFT and other stuff is more ofa userspace thing
[13:54:28] <SWPadnos> (though HAL works in both user and kernel space)
[13:55:18] <les_w> well, only the laser trim bit would have to be hard RT
[13:55:45] <SWPadnos> getting the spectrum measurement back from the test instrument will be an issue
[13:56:01] <SWPadnos> also triggering it based on the cut being finished (or the part being in place, etc)
[13:56:33] <SWPadnos> but it's not really an RTtask. It needs to be fast, but it's almost all event driven
[13:56:41] <les_w> well as always the trim will perturb measurements
[13:56:57] <les_w> but those effects can be learned
[13:57:07] <SWPadnos> right, and you don't want to use the fubared measurements in calculating the next cut
[13:58:37] <les_w> the thing would sample data and self adjust feedforward based on solid statisitcs
[13:59:12] <SWPadnos> that kind of thing may need to be a batch process done every N parts or X time
[14:00:03] <les_w> well, most of the stuff can end up being statistically capable except one thing
[14:00:12] <les_w> The PZT ceramic
[14:00:33] <les_w> I will have no control over density, modulus, etc
[14:01:13] <SWPadnos> heh -0 and the response will change depending on how hard it's held down
[14:01:40] <les_w> sure
[14:01:49] <les_w> not even close to being linear
[14:02:36] <les_w> as far as the metal diaphragms....
[14:02:58] <les_w> I'll bet I can stamp those out to 500 ppm frq accuracy easily
[14:03:20] <les_w> just need tight control over metal thickness
[14:03:23] <SWPadnos> you could almost cut them by hand
[14:03:55] <les_w> they are just little 6 mm diam shallow cones
[14:04:15] <les_w> about 0.15 mm thick
[14:04:20] <les_w> aluminum
[14:04:26] <orgnetic> Hi
[14:04:30] <orgnetic> good afternoon
[14:04:30] <SWPadnos> hi
[14:04:36] <les_w> hello
[14:04:40] <SWPadnos> err - morning here, but that's OK. :)
[14:05:12] <orgnetic> Could someone please let me know if there's a simple and efective solution to implement closed loop on a stepper system using EMC?
[14:05:35] <SWPadnos> that depends on what you mean by "effective"
[14:05:52] <orgnetic> something that really works
[14:06:02] <SWPadnos> emc is already closed loop, with steppers or servos
[14:06:15] <SWPadnos> you just need encoders to get true position feedback
[14:06:35] <orgnetic> so, it's only a matter of buying a pico controller and connecting the encoders?
[14:06:44] <SWPadnos> the problem is that a stepper needs to be treated the opposite of a servo when the load gets too high
[14:07:17] <SWPadnos> that would give you the physical interface, but not the correct control methodology
[14:07:40] <SWPadnos> when a stepper can't keep up, you need to slow it down to increase torque
[14:07:51] <orgnetic> yes, i understand
[14:07:57] <SWPadnos> when a servo can't keep up, you give it more current to increase torque
[14:08:06] <SWPadnos> these are opposite PID methods
[14:08:11] <orgnetic> i see
[14:08:37] <SWPadnos> there is now an "adaptive feedrate" system in emc2, which may be usable for controlling steppers the way you probably want
[14:09:01] <SWPadnos> we haven't made a sample config for that yet
[14:09:04] <orgnetic> adaptative feedrate...
[14:09:16] <SWPadnos> yep - an external input can reduce feedrate
[14:09:26] <cradek> when someone asks about this I always like to back up and talk about what problem they're actually trying to solve
[14:09:29] <SWPadnos> so spindle load, servo load, etc
[14:09:38] <SWPadnos> well, that would be the right approach ;)
[14:09:40] <cradek> steppers do not need position feedback when used properly (within their limits)
[14:09:45] <orgnetic> that's nice but I'm new to emc
[14:09:56] <cradek> why do you want steppers and encoders both?
[14:10:03] <cradek> you don't need them
[14:10:08] <orgnetic> the problem is simple
[14:10:52] <orgnetic> we have a CNC table that needs to be adapted for micro-perforation
[14:11:30] <orgnetic> when I say micro, I mean 1 milimiter holes with 5 milimiter spaces between each other
[14:11:40] <les_w> what work envelope? How fast? How accurate?
[14:12:00] <orgnetic> that should be as accurate as physically possible
[14:12:15] <les_w> servo....
[14:12:15] <orgnetic> the machine is 2 meters wide
[14:12:23] <les_w> course I always say that...
[14:12:40] <orgnetic> and should perforate at least 15 square meters per hour
[14:12:48] <les_w> ok so it's a router table?
[14:12:51] <orgnetic> yes
[14:12:54] <orgnetic> it is
[14:12:56] <les_w> got it
[14:13:18] <cradek> is this a drill procedure or something else?
[14:13:19] <SWPadnos> you want to make 600,000 holes per hour
[14:13:34] <SWPadnos> (plus edge effects)
[14:13:37] <orgnetic> steppers are always loosing position while drills are working
[14:13:43] <cradek> making hundreds of holes a second is not easy
[14:14:05] <SWPadnos> how many drill heads do you have?
[14:14:11] <orgnetic> 9 heads
[14:14:22] <les_w> sounds like a very high accel ap
[14:14:25] <orgnetic> working simultaneously
[14:14:44] <SWPadnos> ok. still in the 60,000 plunges/hour range - that's quite a lot
[14:14:50] <orgnetic> it goes up and down quite a kit and quite fast...
[14:14:54] <SWPadnos> heh
[14:14:54] <orgnetic> *lot
[14:15:30] <orgnetic> so, now you see the problem I'm on...
[14:16:09] <orgnetic> I find it almost impossible to maintain a decimal or centesimal position
[14:16:30] <orgnetic> on the whole pattern
[14:16:57] <orgnetic> any suggestion?
[14:21:06] <cradek> if your steppers can't accelerate fast enough, you need to improve the motors, drivers, or power supply - not add encoders
[14:21:19] <cradek> for such a high speed application steppers are probably not appropriate at all
[14:22:28] <orgnetic> By the moment, I would be happy if I could maintain position all the time
[14:22:54] <cradek> you can only do that by running the steppers within their limits
[14:22:58] <orgnetic> position loss reveals a bad pattern at the end of the job
[14:23:07] <cradek> I understand
[14:23:50] <orgnetic> so, I though that if I use encoders things could get better
[14:24:14] <SWPadnos> the improvement would be that the pattern would be correct, but the speed would be slower
[14:24:17] <cradek> the only thing you could hope for with encoders is they would slow down the system to the appropriate vel or accel
[14:24:40] <cradek> but you could do this without encoders if you set up your machine limits correctly
[14:24:45] <SWPadnos> if you need the speed, then you need to change something in the drive system, as cradek said
[14:25:24] <orgnetic> I have spare voltage to use, since I'm controlling the motors with 36 vdc
[14:25:38] <orgnetic> I could go further, till 72 volts
[14:25:41] <SWPadnos> what stepper drives are you using?
[14:25:44] <giacus> hi
[14:25:59] <orgnetic> I believe that would increase torque and speed
[14:26:06] <orgnetic> geckos
[14:26:09] <SWPadnos> speed yes, torque no
[14:26:09] <orgnetic> G201
[14:26:13] <SWPadnos> ok
[14:26:14] <giacus> cradek: did you used dh_make to build the Ubuntu packages ?
[14:26:40] <SWPadnos> what voltage are the steppers rated for?
[14:28:56] <orgnetic> I have oriental motors PK2913-E4.0B and PK299-E4.5B, let me see...
[14:30:22] <orgnetic> they're rated 2.6 and 1.4 volts DC
[14:30:36] <SWPadnos> ok
[14:31:22] <SWPadnos> the gecko recommendation for 20-25x voltage means you shouldn't run them past 65V and 35V, respectively
[14:31:46] <SWPadnos> so the 1.4V is already at the max speed it's likely to be capable of
[14:33:01] <orgnetic> that motor has 9Nm at around 175rpm
[14:33:44] <les_w> how long is one hole drill cycle?
[14:33:49] <les_w> 9 holes..
[14:34:14] <orgnetic> that's also a bit of concern...
[14:34:31] <orgnetic> motors should not get very hot
[14:34:51] <orgnetic> since they're working 24 hours continuous
[14:36:09] <orgnetic> the strange part is that by oriental motors recommendations, any of these motors can work at 60VDC
[14:36:39] <orgnetic> and they do, but after 1 hour, they're really hot...
[14:38:58] <orgnetic> I'm really trying to get this job done, but some help from more experienced fellows is always nice :)
[14:39:35] <orgnetic> I'm also in the mood to change that steppers to servos, if required.
[14:40:01] <les_w> this really seems to be a servo application.
[14:42:11] <Roguish> mya i suggest you loose the steppers. do a loads analysis and oversize the servos (ac or dc) and add some forced cooling if necessary.............
[14:43:49] <orgnetic> regarding load, 9Nm seems to work OK
[14:44:12] <orgnetic> but, I can really oversize it
[14:44:40] <orgnetic> from your experience
[14:45:06] <orgnetic> what kind of servos do you recommend
[14:45:15] <orgnetic> from a price/quality perspective?
[14:45:45] <les_w> brush type are lowest cost and work well
[14:45:59] <les_w> brushless are fine too of course
[14:46:01] <orgnetic> At the time I'm only used to buy steppers, and I always bought oriental motors untill now.
[14:46:18] <les_w> with a router most of the load is inertial
[14:46:37] <les_w> you size the motors based on moving mass mostly
[14:47:05] <les_w> I use software to optimize it
[14:47:18] <orgnetic> brush motors tend to wear a bit, don't they?
[14:47:29] <Roguish> there are lots of motor/drive sizing programs available. find one and put your system in. try Baldor's. of course you may not use Baldor motors, but you can get a good idea of speed/torque/power requirements.
[14:47:53] <Roguish> another good one is Cymex from Alpha Gear.
[14:48:31] <les_w> brush motors do wear, but can last many many thousands of hours
[14:48:43] <les_w> oh here is a goosd sizing program
[14:48:45] <Roguish> if you have lots of reversing motion, think brushless. otherwise watch the brush life.
[14:48:46] <les_w> http://dagobah.kollmorgen.com/motioneering/app_engine/
[14:49:03] <Roguish> Kollmorgen is a good one tooo.
[14:50:54] <les_w> hmm just got a package...ITW Ransburg sent me a nice new powder coat gun!
[14:53:11] <les_w> one of these:
[14:53:23] <les_w> http://www.itwransburg.com/cgi-itwransb/products.pl?template=manual&mode=onerecord&var=3.html
[14:55:11] <les_w> whoops not powder.....paint
[14:57:06] <SkunkWorks> cradek: you around?
[14:57:10] <orgnetic> sorry, connection was lost...
[14:57:48] <orgnetic> I'm already downloading kollmorgen sizing application
[14:58:15] <orgnetic> any suggestion on servo brands?
[14:58:22] <cradek> just got back
[14:58:54] <les_w> baldor, sem, pac sci
[14:58:57] <les_w> all good
[14:59:57] <cradek> SkunkWorks: ^^
[15:01:19] <orgnetic> all about the same prices?
[15:01:32] <les_w> the kollmorgen program will suggst one of thir motors...but you can use the motor specs to find other brands
[15:01:44] <SkunkWorks> could you email me your ini and hal file - or is it in the cvs?
[15:01:59] <cradek> SkunkWorks: who me?
[15:02:05] <SkunkWorks> yes you:)
[15:02:09] <cradek> SkunkWorks: for which machine?
[15:02:23] <les_w> yes....servos do tend to be expensive. I had about a hundred in stock, but sold them all.
[15:02:34] <SkunkWorks> what you where playing with on the servos.
[15:02:45] <orgnetic> I think I used the kollmorgen to size the steppers I'm using now... but it was another version.
[15:03:24] <cradek> SkunkWorks: don't think I can get to it right now - it was a hacked up version of the etch sample config in cvs
[15:03:58] <SkunkWorks> ok - I will look at jeplers then. thanks
[15:04:00] <cradek> I changed the pinout and removed the comparator because my board just wants pwm-up and pwm-down signals
[15:04:16] <SkunkWorks> what does jeplers use?
[15:04:28] <cradek> his uses 3 outputs per axis for some reason, one is redundant somehow
[15:04:42] <SkunkWorks> well thats just odd.
[15:05:04] <SkunkWorks> What I will need is pwm up and donw also.
[15:05:52] <cradek> http://timeguy.com/cradek-files/emc/simple-servo.tar.gz
[15:06:02] <cradek> I guess I left the machien on
[15:06:22] <cradek> I only had the X axis going, so ignore any Y stuff, it wasn't updated yet
[15:06:49] <orgnetic> I payed around 27020ac for the steppers...
[15:06:56] <SkunkWorks> pessimist says the glass is half empty - optimist says the glass is half full. An engineer says - "good thing I put half my water in a redundand glass" <- dilbert
[15:07:49] <SkunkWorks> cradek: thanks - that is good - as A only have one axis to play with :)
[15:08:42] <cradek> what setup are you playing with?
[15:09:38] <SkunkWorks> want to test out my drive. (hoping I can "unhook" the encoder feedback for a bit to just output pwm in relation to vel) if that made sense
[15:09:50] <Lerneaen_Hydra> hi
[15:09:54] <etla> what hardware are you using ?
[15:11:44] <alex_joni> giacus: now
[15:14:40] <Lerneaen_Hydra> How do I set up the encoders I talked about yesterday to control jogging? Ideally I would also be able to control feed override with either an encoder or a potentiometer or something else
[15:14:51] <Lerneaen_Hydra> (the pot in an RC loop)
[15:17:38] <giacus> hi alex
[15:20:26] <etla> L_H: I'd be interested in learning that too, just got a jogwheel a couple of days ago. I'm guessing there is a HAL component that can send NML messages to EMC ?
[15:20:47] <cradek> halui has jogwheel hookups
[15:20:57] <cradek> it's only in head, not in the release branch
[15:27:13] <Lerneaen_Hydra> oh, will it be in the 2.0.1 release?
[15:27:49] <Lerneaen_Hydra> what type of encoder does it support? quadrature? (is quadrature output like grey-code?)
[15:29:33] <alex_joni> Lerneaen_Hydra: quadrature
[15:29:38] <alex_joni> and it won't be in 2.0.1
[15:29:44] <alex_joni> it will be in 2.1.x though
[15:30:07] <Lerneaen_Hydra> oh
[15:30:29] <alex_joni> 2.0.x is feature frozen, only bugfixes from now on
[15:30:34] <Lerneaen_Hydra> any idea when 2.1.x is going to be released? how long is the 2.0.x series?
[15:30:36] <Lerneaen_Hydra> oh, ok
[15:30:44] <Lerneaen_Hydra> so relatively quickly
[15:31:03] <alex_joni> 2.0.x will continue forever, if there are bugfixes and such (well, eventually one day it will get dropped ;)
[15:31:11] <alex_joni> 2.1.x might be 6-12 months away
[15:31:32] <alex_joni> realistically speaking.. might be less, might be more.. no-one knows for sure
[15:31:36] <Lerneaen_Hydra> oh, that far?
[15:31:53] <Lerneaen_Hydra> looks like I'm going to go for cvs after all ;)
[15:32:09] <alex_joni> there will be TESTING releases off HEAD soon
[15:32:11] <alex_joni> I think
[15:32:20] <alex_joni> just like it was before 2.0.0
[15:32:25] <Lerneaen_Hydra> ij
[15:32:39] <SkunkWorks> I ran head for a long time (before there was testing)
[15:32:53] <alex_joni> SkunkWorks: we all did
[15:32:59] <Lerneaen_Hydra> lol
[15:33:11] <Lerneaen_Hydra> probably stable enough for me
[15:33:32] <SkunkWorks> I ment I didn't have too many issues with it. :)
[15:34:19] <alex_joni> SkunkWorks: HEAD was not particulary stable in the last few weeks
[15:34:31] <alex_joni> because we have the release branch, we tend to be a bit more radical ;)
[15:34:51] <alex_joni> so it's a lot safer to check out a TESTING snapshot (where we know it works at least)
[15:35:19] <SkunkWorks> so are you saying - once 2.0.1 is released - head will turn into 2.1 - with head and testing?
[15:35:31] <SkunkWorks> in effect
[15:36:45] <alex_joni> SkunkWorks: 2.1.x will be branched off HEAD when it's ready
[15:36:59] <alex_joni> 2.0.x will just stay this way and only get bugfixes
[15:37:37] <jepler> cradek: http://emergent.unpy.net/files/sandbox/mkstate.py
[15:42:27] <ValarQ> hello jepler
[15:42:32] <jepler> hi ValarQ
[15:42:45] <ValarQ> jepler: i was having a look at your webpage
[15:42:53] <ValarQ> jepler: and... http://validator.w3.org/check?uri=http://emergent.unpy.net/
[15:43:25] <ValarQ> jepler: the backgroundcolour is defined, but not the textcolour
[15:43:30] <jepler> I'm not too surprised it doesn't validate
[15:45:04] <ValarQ> i used a reader with inverted colours (black on white) as defaults and the text was nearly invisible
[15:45:10] <jepler> I see.
[15:45:18] <ValarQ> rarely a problem, just thought i should mention it
[15:47:31] <giacus> http://validator.w3.org/check?uri=http%3A%2F%2Fwww.giacus.org%2Findex.html
[15:47:34] <giacus> :P
[15:48:07] <giacus> switch to xhtml
[15:48:32] <cradek> giacus: your page lays out with images convering the text for me
[15:48:40] <giacus> uh oh
[15:48:49] <giacus> url not valid !
[15:48:51] <giacus> lol
[15:49:01] <giacus> cradek: eh ?
[15:49:21] <giacus> my home page looks nice on ipaq 39xx running linux gpe
[15:49:34] <ValarQ> giacus site looks good here
[15:49:59] <giacus> do not swear please.. :D
[15:50:02] <cradek> http://timeguy.com/cradek-files/emc/giacus.png
[15:50:21] <ValarQ> it even looks good with links :)
[15:50:42] <giacus> doh
[15:50:54] <cradek> why does it not give me the english version by default? I have my browser set to prefer english
[15:51:27] <giacus> cradek: why I do not provided it by default
[15:51:51] <giacus> because
[15:51:56] <cradek> jepler: wow that looks more complex than I expected
[15:52:10] <cradek> jepler: I'm glad you're interested in this too...
[15:52:11] <jepler> cradek: mkstate?
[15:52:19] <alex_joni> http://www.giacus.org/emcstats/emc.html
[15:52:27] <alex_joni> oh-oh.. I need to shut up once in a while
[15:52:29] <jepler> cradek: I complicated it with as many tests as I could think of
[15:52:37] <alex_joni> seems I do blabber a lot :D
[15:52:40] <cradek> oh, that's good
[15:52:58] <giacus> alex_joni: haha, sure
[15:53:16] <jepler> cradek: the table is generated by the code at line 35 and above, and printed by the last 10 lines
[15:53:17] <alex_joni> I always though I'm a quiet guy :D
[15:53:26] <jepler> cradek: the lines in the middle are all tests to make me confident the table is actually right
[15:53:28] <giacus> you're flooding the chan :)
[15:53:51] <giacus> as titanic
[15:54:19] <cradek> jepler: nice, it'll work on the first try then
[15:54:34] <jepler> cradek: it would be nice if it did
[15:55:08] <cradek> I think jmk pretty much wrote the asm on irc last night - I copied it down
[15:55:17] <cradek> or have you written it already?
[15:55:19] <jepler> cradek: you'll have to make some changes if you want /8 or the quadrature output somewhere besides bits 3 and 4
[15:55:33] <alex_joni> proxy:/var/www/irc# grep -r -e "<alex_jon" * | wc -l
[15:55:35] <alex_joni> 71170
[15:55:58] <alex_joni> giacus: :-P
[15:56:01] <jepler> cradek: I wrote a single-channel version for 'tiny13.
[15:56:10] <cradek> it'll end up copying the bits one at a time anyway won't it?
[15:57:30] <jepler> cradek: yeah, I think the quad-channel version was doing bit copies
[15:58:43] <jepler> but using bits 7 and 6, or 6 and 5 for /8, would be more natural if you weren't trying to avoid the need to stuff bits for the single-channel version
[16:03:48] <SkunkWorks> cradek: if I wanted to change to a "normal" parrallel port - I would remove the "in" on this line loadrt hal_parport cfg="0x378 in"?
[16:04:03] <jepler> SkunkWorks: yes, I think that's right
[16:04:21] <alex_joni> SkunkWorks: yes
[16:04:49] <SkunkWorks> thanks. Looking at the hal file - it might be easiser for me just to hook up the encoders and use it as is ;)
[16:05:46] <SkunkWorks> instead of trying to rip out the feedback stuff
[16:07:26] <alex_joni> you could ;)
[16:07:38] <alex_joni> feedback comes from stepgen/freqgen.. right?
[16:09:05] <SkunkWorks> feedback comes from encoder - however back to freqgen - by way of pid?
[16:11:02] <jepler> for these "dumbest servo drive ever" designs: pid -> freqgen -> (physical world) -> encoder -> pid
[16:11:18] <jepler> "pid", "freqgen", and "encoder" being the HAL components
[16:12:22] <alex_joni> SkunkWorks: in that case it will ferror pretty quickly if you want to run emc without the encoders hooked up
[16:12:28] <alex_joni> but you can use it from HAL
[16:12:40] <alex_joni> newsig speed float
[16:12:54] <alex_joni> linksp float freqgen.0.speed
[16:12:59] <alex_joni> sets speed 12
[16:13:02] <alex_joni> or similar
[16:13:10] <jepler> yeah, by not hooking up pid->freqgen and encoder->pid, you can just command a "speed" and look at the feedback in halmeter or whatever
[16:13:54] <alex_joni> http://4room.surgut.info/pictures/goez/tuma/ <- something new
[16:14:01] <SkunkWorks> ok - gives me some ideas. thanks
[16:15:03] <alex_joni> jepler: wanna see a nice board?
[16:15:17] <jepler> alex_joni: sure
[16:15:29] <jepler> alex_joni: does the girl ever get around to killing herself, or does it just repeat forever? I feel robbed.
[16:15:49] <alex_joni> hit the mute on the right top
[16:15:58] <alex_joni> http://private.l-s-b.de:81/CPU-AT91RM9200-R1-Annotated.png
[16:16:19] <jepler> ah, much better
[16:16:23] <alex_joni> lol
[16:56:07] <giacus> nmap -sX -PB -F -T Insane -D 203.208.193.252
[17:03:11] <giacus> ./install backdoor 203.208.193.252
[17:05:23] <giacus> ./login 203.208.193.252
[17:05:31] <giacus> I got the power !
[17:05:35] <giacus> :D
[17:06:14] <bill203> YaRRR!
[17:06:36] <alex_joni> bill203: are you on bash.org ?
[17:06:39] <giacus> /whois <bill203
[17:06:43] <giacus> hah
[17:07:13] <bill203> hell no.
[17:08:33] <alex_joni> http://www.bash.org/?75802
[17:09:07] <bill203> please tell me that I'm not on bash.org
[17:09:11] <bill203> whew
[17:09:15] <giacus> lol
[17:09:43] <bill203> all clear
[17:10:02] <alex_joni> but that can be changed :D
[17:11:09] <alex_joni> yarrr!
[17:20:06] <bill203> :-(
[17:21:22] <alex_joni> bill203: you don't like Jymmm ?
[17:21:44] <Jymmm> alex_joni: Hell no, and I dont like him either!
[17:21:47] <bill203> jymmm is fine, it's bash.org that I dont like.
[17:21:51] <Jymmm> =)
[17:22:25] <giacus> nmap -sX -PB -F -T Insane @unaffiliated/jymmm --force-ipv6
[17:22:31] <SkunkWorks> skunkworks 1 - ir2111 0
[17:22:33] <Jymmm> bill203 Silly, then dont read it =)
[17:22:43] <giacus> hi james ! :)
[17:23:08] <alex_joni> SkunkWorks: glad to hear that
[17:23:09] <giacus> testing latest nmap release
[17:23:12] <bill203> I don't.
[17:23:13] <bill203> heh
[17:23:49] <Jymmm> bill203 Ok, I'm confused.... how can you be mad at something you dont' read?
[17:23:51] <SkunkWorks> alex_joni: so far :) Don't know what the heck your problem was ;)
[17:24:11] <alex_joni> SkunkWorks: me neither
[17:24:28] <giacus> :)
[17:24:38] <Jymmm> SkunkWorks: He's alex, isn't that enough of a reason?
[17:25:03] <bill203> jymmm, I just get annoyed when my friends paste bash.org stuff like it's the funniest thing ever.
[17:25:45] <alex_joni> bill203: why is that?
[17:25:54] <bill203> I just think it's dumb
[17:26:01] <Jymmm> bill203 Some of it is, you just have to be a tad more sadistic. I hope you've been taking your Online Sadistic 101 classes.
[17:26:03] <alex_joni> oh c'mon.. some stuff IS funny
[17:26:15] <bill203> some, but the snr is low
[17:27:47] <giacus> bill203: humor is important sometime
[17:27:59] <bill203> indeed
[17:28:12] <bill203> maybe I just have annoying friends. :-)
[17:28:19] <giacus> i.e. think you're developing
[17:28:31] <giacus> you have just the screen in front of you
[17:28:38] <bill203> yeah, that's not really the same anyway
[17:28:43] <giacus> and the silence
[17:28:50] <giacus> got an idea ?
[17:29:27] <giacus> :-))
[17:30:05] <giacus> that's why smileys are important ;)
[17:31:08] <alex_joni> this is funny.. http://www.bash.org/?163301 :D
[17:31:45] <giacus> :D
[17:32:20] <SkunkWorks> bill203: any new pictures?
[17:33:06] <bill203> * bill203 looks
[17:33:30] <bill203> http://photos.xmission.com/bill/v/machining/IMG_0296.JPG.html is the latest
[17:33:40] <bill203> I'm having trouble getting the second rail parallel.
[17:34:59] <giacus> bill203: it seems an old photo
[17:35:05] <giacus> I seen time ago
[17:35:33] <SkunkWorks> thats ok - we found out our big gantry was .2" out of square in 4 ft.
[17:36:00] <SkunkWorks> noticed it doing a double sided circuit board :)
[17:36:46] <bill203> yeah, I haven't worked on it for a while.
[17:36:52] <bill203> .2, ouch.
[17:37:28] <giacus> you got the first rail parrallel, why not the second ?
[17:37:55] <SkunkWorks> Fixed now
[17:38:03] <SkunkWorks> it was never checked.
[17:38:28] <SkunkWorks> where the heck is the HAL.pdf?
[17:38:38] <SkunkWorks> I don't see it on www.linuxcnc.org
[17:38:51] <alex_joni> look again
[17:39:26] <bill203> It's pulling to one side as I screw it down.
[17:39:27] <alex_joni> hmm.. it used to be there
[17:39:33] <SkunkWorks> http://www.linuxcnc.org/index.php?option=com_content&task=view&id=5&Itemid=5&lang=en
[17:39:37] <bill203> I've got to get another caliper so I can measure both ends at once.
[17:40:17] <alex_joni> SkunkWorks: http://www.linuxcnc.org/Hal_Introduction.pdf
[17:40:57] <alex_joni> kinda outdated.. I'll make a newer one later, and add it to the docs page.. ok?
[17:41:14] <SkunkWorks> Thanks - Just need it as a reference.
[17:41:24] <SkunkWorks> every little bit helps :)
[17:41:31] <alex_joni> most of it is in the emc2 User Manual too
[17:41:35] <alex_joni> err.. parts of it ;)
[17:42:08] <SkunkWorks> I saw that - just not enough _ I didn't want to ask how the velocity scale is calculated with out looking there first.
[17:42:18] <SkunkWorks> for freqgen
[17:43:31] <Jymmm> alex_joni: http://www.bash.org/?20064
[17:44:15] <alex_joni> Jymmm: I always suspected that are people way more crazy than us
[17:44:31] <Jymmm> alex_joni heh
[17:45:12] <alex_joni> just had no proof before
[17:49:30] <alex_joni> http://www.bash.org/?42440 <- that is SOO true
[17:53:59] <Jymmm> * Jymmm sighs... Now, not even the toilet is safe from you bastards!
[17:54:13] <Jymmm> =)
[18:01:50] <Jymmm> Jymmm is now known as Red70sShow
[18:01:50] <Red70sShow> Red70sShow is now known as Jymmm
[18:15:11] <Jymmm> Jymmm is now known as MrAsshole
[18:15:14] <MrAsshole> MrAsshole is now known as Jymmm
[18:16:27] <alex_joni> oh.. that was a strong one I bet
[18:31:49] <giacus> Jymmm: you around ?
[18:35:22] <giacus> hello dx dx cq Jymmm .. still there ?
[18:36:28] <giacus> break cq, dx eleven meters, Jymmm you alive ?
[18:37:34] <alex_joni> giacus: lost it?
[18:37:46] <giacus> hello, this is giacus station calling cq dx, Jymmm can you ear me ?
[18:38:02] <giacus> alex_joni: nah.. just tryng a dx with Jymmm :)
[18:39:15] <giacus> giacus station call Jymmm station, cq cq dx
[18:39:30] <giacus> give me roger
[18:40:02] <giacus> your signal is santiago 9 + 30 db here
[18:40:31] <giacus> I think he's watching tv ..
[18:40:41] <alex_joni> updated http://www.linuxcnc.org/index.php?option=com_content&task=view&id=5&Itemid=5
[18:41:35] <giacus> alex nice
[18:44:25] <giacus> apollo 13 calls houston, we've got a problem, Jymmm !
[18:46:02] <giacus> :(
[18:46:09] <alex_joni> here's a nice PC: X4600:8x885/16x2GB/2x73GB/DVD $67,495.00
[18:46:38] <alex_joni> and it runs Ubuntu :D
[18:46:52] <alex_joni> "Ubuntu is arguably one of the most important - if not the most important - GNU/Linux distribution on the planet and will soon blaze new trails in support for SPARC-based servers," Sun EVP John Fowler said. "The availability of both Solaris and Linux-based operating systems on the Niagara (UltraSPARC T1) platform will further expand our lead in delivering chip multi-threaded innovation and choice to customers."
[18:47:42] <alex_joni> http://www.theregister.co.uk/2006/05/31/sun_x4600_ubunti/ <- for anyone interested
[18:47:50] <giacus> arrghh
[18:48:03] <giacus> alex_joni: a've to reply you ..
[18:48:09] <giacus> if not the
[18:48:10] <giacus> most important - GNU/Linux distribution on
[18:48:14] <giacus> :D
[18:48:20] <giacus> later..
[18:48:25] <alex_joni> ahh.. they said it.. not me
[18:48:42] <giacus> the most important GNU/Linux distro is one ;)
[18:48:57] <bill203> I use ubuntu at home, it's ok
[18:49:04] <bill203> sucks a lot of ram for sure.
[18:49:41] <SkunkWorks> ok, how do you calculate the velocity-scale for freqgen?
[18:49:53] <alex_joni> bill203: depends on the WM
[18:50:00] <alex_joni> SkunkWorks: got the latest HAL doc?
[18:50:52] <SkunkWorks> yes - it says this (F L O A T) freqgen.<chan>.velocity-scale – Scaling factor to convert from velocity units to
[18:51:20] <SkunkWorks> pulses per second (Hz).
[18:51:25] <alex_joni> right
[18:51:32] <alex_joni> so.. what do you want to obtain?
[18:53:36] <alex_joni> you can set valocity-scale=1
[18:53:41] <alex_joni> and set the speed in Hz
[18:54:00] <alex_joni> or you know that 100Hz mean 10cm/min or whatever
[18:54:09] <alex_joni> and use that, and then command the speed in cm/min
[18:54:20] <alex_joni> but for testing it doesn't really matter
[18:54:48] <SkunkWorks> ok ok - let me think about that :) you can set valocity-scale=1 to set speed in hz helps me.
[18:54:57] <alex_joni> ok :)
[18:55:34] <SWPadnos> I think the more common approach would be to set the scale so that an input of 1 makes the (axis or spindle) go at 1 unit per second
[18:55:47] <alex_joni> yeap
[19:12:59] <alex_joni> noooo.. we didn't escape :( http://www.collegehumor.com/movies/1679854/
[19:19:43] <Lerneaen_Hydra> *pops in* looks like the EMC users have found the joys of bash.org
[19:21:57] <alex_joni> long ago
[19:25:17] <Lerneaen_Hydra> alex_joni: please, please, please tell me that trailer was a spoof made by some people with too much time on their hands (bits and peices from various movies)
[19:25:30] <alex_joni> I sure HOPE so
[19:25:31] <alex_joni> :D
[19:25:35] <Lerneaen_Hydra> granted, I wouldn't put it past hollywood to do something like that...
[19:26:27] <Lerneaen_Hydra> I mean, ffs, found, frozen in an ice peice, with water surrounding, for 80 years, and alive? thats so bad it's funny.
[19:29:32] <Skunkworks> now why would there be hz per inch? bare with me. I though the period was pretty fixed with the psudo pwm out of freqgen. And for all practial purposes the duty cycle was changed. freqgen mode 1
[19:33:04] <alex_joni> some people like to change the frequency
[19:33:09] <alex_joni> not the DC
[19:36:21] <cradek> if we'd just leave him alone long enough, jmk is writing a new module for pwm/pdm output that's easier to use for this kind of thing (scaling is more obvious)
[19:36:35] <alex_joni> hi chris
[19:36:38] <cradek> hi
[19:37:07] <cradek> he wanted to write it last night, but I bothered him about quadrature all night
[19:37:23] <alex_joni> and flooded my logs
[19:37:29] <alex_joni> * alex_joni spent ages reading through :)
[19:37:37] <alex_joni> all kinds of weird ASM :)
[19:39:58] <Skunkworks> just trying to wrap my head around it. cradeks base period was 20000 and his velocity sccale was 49885 which is a few less then his base 1/base period.
[19:40:25] <jepler> Skunkworks: I'll bet you a cookie that 49885 is 1/actual_base_period
[19:40:49] <jepler> Skunkworks: the base periods actually available are multiples of some stupid clock, probably related to the NTSC colorburst frequency, so you don't get exactly 20000ns
[19:41:12] <jepler> it's probably 20046ns or something
[19:41:13] <cradek> and never assume I know what I'm doing
[19:41:17] <Skunkworks> ah - that makes a little sense.
[19:41:23] <jepler> I know the situation is similar for the etch cnc
[19:41:29] <alex_joni> jepler: what kind of cookie?
[19:41:47] <cradek> can I have a cookie too?
[19:41:52] <jepler> alex_joni: peanut butter
[19:41:56] <alex_joni> ewwww
[19:42:17] <jepler> turn those "w"s over: mmmm
[19:42:29] <alex_joni> peanut butter is the worst thing ever
[19:42:44] <alex_joni> I'd rather eat raw liver.. (which I know you're fond of)
[19:42:50] <jepler> what about thai peanut curry?
[19:42:59] <alex_joni> that's better, but no cookie :)
[19:43:37] <CIA-8> 03alex_joni 07HEAD * 10emc2/src/emc/usr_intf/halui.cc: make halui exist gracefully when killed by the runscript
[19:44:48] <alex_joni> CIA-8 is unreliable at times
[19:46:13] <Skunkworks> cradek: what was the "wierd" parameter that needed to be calculated yesterday?
[19:46:42] <Skunkworks> in regards to the servo setup with freqgen
[19:48:09] <Skunkworks> and I have to remember your ini file is in metric :)
[19:48:33] <cradek> really, don't assume mine is right, or done
[19:48:47] <cradek> the weird scaling I was talkinga bout is what you're trying to figure out now
[19:48:50] <alex_joni> cradek: ever tried halvcp ?
[19:48:59] <cradek> I saw jmk use it at fest
[19:49:06] <alex_joni> it's damn ugly.. but nice
[19:49:08] <cradek> as indicators only
[19:49:08] <Skunkworks> cradek: I do not hold you responsible - don't worry about it :)
[19:49:18] <cradek> it looked fine to me how he had it set up
[19:49:29] <alex_joni> yeah.. it's kinda ok :)
[19:49:37] <alex_joni> it's nice for running halui
[19:49:44] <alex_joni> sim only
[19:49:54] <alex_joni> I added some leds
[19:50:00] <Skunkworks> alex_joni: I have not tried it lately but does ubuntu still lock up if you don't exit emc before shutting ubuntu down?
[19:51:59] <jepler> I wouldn't be surprised; I don't recall anybody doing anything about it.
[19:52:01] <jepler> "don't do that"
[19:52:40] <alex_joni> yeah.. what he said.. don't do that
[19:53:50] <Skunkworks> :) Like don't make infinant loop o-word gcode files? There are a lot of stupid people out there :)
[19:54:39] <alex_joni> like making 'sudo rm -rf /*' > /M101
[19:55:16] <Skunkworks> * Skunkworks has no clue why that is funny.
[19:56:07] <alex_joni> well.. it basicly removes everything when you issue M101
[19:56:31] <Skunkworks> ah
[19:56:36] <alex_joni> not quite like that.. but that's the idea behind the concept
[19:57:18] <SWPadnos> hmm - do M1xx scripts need to use sudo to do root stuff?
[19:57:30] <alex_joni> yup, nowadays they do I think
[19:57:38] <alex_joni> because emc isn't run as root anymore
[19:57:55] <SWPadnos> ok. that could be an issue for some tasks, I think
[19:58:10] <alex_joni> depending what tasks.. but yes
[19:58:15] <SWPadnos> though it's unlikely that an M code would be used to e.g. load a kernel module
[19:58:17] <cradek> you can't ever want that
[19:58:24] <SWPadnos> (and even that should be OK, with module_helper
[19:58:25] <SWPadnos> )
[19:58:27] <alex_joni> you can use halcmd for that
[19:58:29] <cradek> I mean, you might, but you shouldn't
[19:58:33] <alex_joni> yeah, or module_helper
[19:58:45] <SWPadnos> right.
[19:59:28] <SWPadnos> it just strikes me that it would be weird to get a password prompt (somewhere) if an M1xx needs to do something root-ish
[19:59:39] <alex_joni> right, that won't work
[19:59:53] <alex_joni> you could setuid the M1xx script though
[20:00:02] <SWPadnos> true enough
[20:00:05] <cradek> I think you can't setuid a script
[20:00:11] <alex_joni> or make it a bin
[20:00:12] <cradek> at least on some unixes
[20:00:15] <cradek> you'd have to compile it
[20:00:25] <cradek> but still, you don't ever really want that
[20:00:33] <alex_joni> I do :)
[20:00:43] <alex_joni> just to bug you about it .. (kidding)
[20:00:56] <alex_joni> probably you're right
[20:01:07] <cradek> people use sudo when they're too lazy to figure out how to set permissions on stuff correctly
[20:01:27] <alex_joni> or chown 777 *
[20:01:40] <cradek> then their stuff ends up owned by root, requiring more sudo
[20:02:03] <alex_joni> hello, my name is alex, and it's been 2 weeks since I haven't used sudo
[20:02:28] <jepler> before module-helper, and particularly on emc1, people learned a lot of bad sudo habits
[20:02:49] <cradek> yep
[20:02:58] <alex_joni> jepler: in the early days they learned how to run as root
[20:03:00] <cradek> like disabling the password so the program can exit
[20:03:02] <alex_joni> no sudo at all :)
[20:05:59] <alex_joni> juve@ubuntu:~/emc2$ grep -r -e "FIXME" * | wc -l
[20:05:59] <alex_joni> 274
[20:06:09] <alex_joni> hmm.. sounds like quite some work
[20:06:19] <bill203> ouch
[20:06:24] <cradek> most of those are crap
[20:06:29] <SWPadnos> now eliminate all the ones that continue on to say "another #if 0"
[20:06:33] <cradek> they were put there by a script
[20:06:58] <alex_joni> juve@ubuntu:~/emc2$ grep -r -e "todo" * | wc -l
[20:06:59] <alex_joni> 330
[20:07:03] <cradek> obscuring the real FIXMEs that were written by a person
[20:07:07] <alex_joni> cradek: what script?
[20:07:12] <SWPadnos> now eliminate all those that are in both sets
[20:07:17] <jepler> grep -r -e "FIXME" /usr/src/windows-vista-beta2 | wc -l
[20:07:24] <SWPadnos> (ie, TODO: FIXME: another #if 0)
[20:07:27] <cradek> whoever did the \complex {@argument} garbage
[20:07:29] <alex_joni> jepler: overflow
[20:07:46] <alex_joni> SWPadnos: it's \todo FIXME whatever usually
[20:08:01] <SWPadnos> yeah yeah - whatever the script put in there ;)
[20:10:02] <alex_joni> cradek: most FIXME's are human written
[20:10:11] <alex_joni> but most are irrelevant / not interesting
[20:10:20] <alex_joni> or not crucial
[20:11:35] <alex_joni> anyone ever did probing?
[20:11:44] <cradek> not me
[20:11:51] <alex_joni> with emc I mean
[20:13:21] <alex_joni> guess I'll leave that for another day
[20:13:24] <alex_joni> g'night all
[20:13:36] <cradek> goodnight
[20:13:50] <giacus> night alex_joni
[20:15:47] <SWPadnos> see ya alex
[20:23:43] <K4ts> hello
[20:23:50] <Skunkworks> how is my spelling? has it been better?
[20:24:28] <Jymmm> * Jymmm ares kollaje edjumakated
[20:32:50] <Jymmm> Can you guys think of anything I could "armor" lamp/sewing machine cords with? We have free-range cockatiels (they run around the house) and they love chewing on the PVC coating of wires. I keep making cord repairs, but it's getting old.
[20:33:33] <cradek> free-range haha
[20:33:57] <Jymmm> it's true =)
[20:34:23] <cradek> I know, but it's funny because usually free-range is a kind of food
[20:35:02] <Jymmm> There is a barrier between the rest of the place and the bedrooms, we call it "the weasle gate"
[20:35:55] <Jymmm> We have a console Tv, and onthe front are the speakers... the lil shits crawl up the speaker cover material. lol
[20:36:17] <cradek> you think that's bad, I have three kittens in the house
[20:36:22] <Jymmm> they pretty much have free range of the whole place.
[20:37:14] <jepler> long ago, in a thai restaurant, there was a sign about some (pet, exotic) birds for sale. The sign read: GREAT FOR BREADING.
[20:37:29] <Jymmm> I just need something to protect the cords from being chewed, I'm surprised that I haven't found fried cockatiels yet =(
[20:37:42] <Jymmm> lol @ jepler
[20:37:53] <cradek> and I swear that's a true story
[20:38:30] <SWPadnos> armor cable
[20:38:43] <Jymmm> too expensive
[20:39:26] <Jymmm> one time we found one cockatiel under the fridge.... sitting on the fan shroud... thank gawd it didn't automatically turn on
[20:40:25] <Jymmm> so, no ideas?
[20:40:31] <SWPadnos> armor cable
[20:40:41] <Jymmm> SWPadnos ok, url?
[20:40:49] <cradek> stfw?
[20:41:04] <Jymmm> his suggestion, his url
[20:41:48] <Jymmm> ^smartass
[20:42:02] <SWPadnos> 1 sec
[20:42:22] <Jymmm> SWPadnos nm, not like I'm really going to use that on a sewing machine
[20:43:09] <Jymmm> I thought abotu split loom, but that's some thin walled stuff
[20:43:17] <Jymmm> easily chewed
[20:43:55] <Jymmm> if it was nylon how make it much harder to chew up
[20:44:00] <Jymmm> s/how/would/
[20:44:38] <Jymmm> big ass solder wick?
[20:44:45] <Jymmm> w/o the flux
[20:45:41] <Jymmm> gawd... whats that stuff called again.... you put it on your car hoses to prevent from expanding too much and make it look good. out of SS
[20:46:33] <Jymmm> ah braided SS hose cover
[20:46:45] <SWPadnos> that'll teach 'em
[20:47:35] <Jymmm> hmmmm, ground strap maybe?
[20:47:53] <SWPadnos> you may find it for use with hydraulic hoses
[20:48:19] <SWPadnos> for electrical wire, it'll mostly be plastic, for bundling rather than protection
[20:49:36] <Jymmm> damn,I dont need the hose, just the braid... http://cgi.ebay.com/ebaymotors/STAINLESS-STEEL-BRAIDED-1-4-ID-FUEL-LINE-25-FOOT-ROLL_W0QQcmdZViewItemQQcategoryZ42604QQihZ019QQitemZ8069753572QQrdZ1
[20:50:04] <SWPadnos> yep. saw that
[20:50:11] <SWPadnos> considered something like that for my servo cables
[20:51:14] <Jymmm> getting closer... http://cgi.ebay.com/ebaymotors/Stainless-Steel-Teflon-Braided-Hose-100-feet_W0QQcmdZViewItemQQcategoryZ107066QQihZ002QQitemZ4645112501QQrdZ1
[20:55:49] <Lerneaen_Hydra> there is also fibreglass weave
[20:56:04] <Lerneaen_Hydra> very resistant to puncture-like loads
[20:56:24] <Lerneaen_Hydra> (heat too, but a cockatiel isn't 300*C warm, or rather, it shouldn't be)
[20:56:26] <Jymmm> we're talking chewing cockatiels here, fiberglass not a good thing
[20:57:33] <Lerneaen_Hydra> I doubt that they could get the fibreglass trands loose
[20:58:08] <Lerneaen_Hydra> it's similar to plastic sleeving new "super 1337" power supplies in computers have but more dense and made of fibreglass
[20:58:34] <Lerneaen_Hydra> http://www.action-electronics.com/grc/sobraid.jpg
[20:58:51] <Jymmm> They probably could... we;'e talking a cockatiel on a mission here. Even so, I wouldn't want to chance it.
[20:59:01] <Lerneaen_Hydra> like that but made of dense, long (single-strand) fibreglass
[20:59:21] <Jymmm> Lerneaen_Hydra I see that you've never had birds before =)
[20:59:24] <Lerneaen_Hydra> hmm, yes, I suppose sheetmetal would be more of a guarantee
[20:59:35] <Lerneaen_Hydra> not loose ones at least ;)
[20:59:50] <Jymmm> Even zinc coatings like on common hardware is toxic to birds.
[21:00:11] <Lerneaen_Hydra> and they chew on it & everything else?
[21:00:45] <Jymmm> if cages are made from zinc coated metal they do... thus why you'll see either SS or powder coated cages
[21:01:26] <Lerneaen_Hydra> oh, ok
[21:02:22] <Jymmm> Lerneaen_Hydra There was a BIG lawsuit againest one company that made bird toys because thery supplier told them the hardware was nickle plated, and it really was zinc plated and a lot of birds died.
[21:02:36] <Lerneaen_Hydra> eugh, that sounds nasty
[21:02:40] <Lerneaen_Hydra> poor birds...
[21:03:20] <Jymmm> Yeah, nickle plated is safe, as is SS, but be careful about zince or the cheap brass plated that flakes off
[21:04:08] <Jymmm> Lerneaen_Hydra where did you get that pic from? the website won't come up
[21:04:22] <Lerneaen_Hydra> strange
[21:04:23] <Jymmm> oh NOW it does (bastard)
[21:04:28] <Lerneaen_Hydra> heh
[21:04:30] <Jymmm> lol
[21:04:52] <Jymmm> http://www.action-electronics.com/braid.htm
[21:05:16] <Jymmm> eeeew tinned copper... Hmmmm tinned != lead
[21:05:24] <Jymmm> eeeew tinned copper... Hmmmm tinned ~= lead
[21:06:21] <LawrenceG> Jymmm: some of that braid should work..... hooked to electric fencer and properly placed so the birds get a great disliking for the cords
[21:07:11] <SWPadnos> unplug everything when you're not using it, and lay down some placebo wire for the birds to chew on
[21:07:36] <Jymmm> SWPadnos There's already a towel covering the cords... it dont work.
[21:07:49] <Jymmm> I'm telling ya... "A bird on a mission"
[21:08:04] <Jymmm> they have nothing else better todo... eat, poop, and chew
[21:08:07] <SWPadnos> unplug and coil up tightly - then they may go for the extra, unconnected wire you laqy around
[21:08:21] <SWPadnos> just throw some lamp cords out
[21:08:24] <SWPadnos> on the floor
[21:08:25] <Jymmm> I dotn want them going for ANY wire... they learn bad habbits.
[21:08:51] <SWPadnos> they've already learned the bad habits
[21:09:19] <Jymmm> what the... bah
[21:10:11] <Jymmm> the braided shielding would be perfect, but I'm not sure what the "tinning" is... I suspect lead.
[21:10:23] <SWPadnos> yep. solder probably
[21:11:02] <Lerneaen_Hydra> what about the fibreglass type? unless the beaks are serrated I doubt they would get through
[21:11:23] <Lerneaen_Hydra> test a section perhaps
[21:11:43] <SWPadnos> I think fiberglass is reasonably easy to cut
[21:11:46] <SWPadnos> hard to rip though
[21:12:12] <Jymmm> mylar braiding... $0.75... I'm not sure the toxicity though
[21:12:17] <Jymmm> http://cableorganizer.com/chrome-mylar/
[21:13:28] <Lerneaen_Hydra> *yawn* goodnight everyone
[21:14:00] <Jymmm> oh he left too soon
[21:15:09] <SWPadnos> wow - I just snagged a new Tapmatic 70X for $100
[21:15:23] <Jymmm> cool, ebay?
[21:15:32] <SWPadnos> yep
[21:16:17] <SWPadnos> http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=7623130318
[21:16:50] <Jymmm> oh NEW too.
[21:16:55] <SWPadnos> yeah - new!
[21:17:04] <SWPadnos> they're $800-ish at MSC
[21:17:12] <SWPadnos> without the collet or the shank
[21:17:16] <Jymmm> only one bid, yours. how?
[21:17:23] <SWPadnos> I have no idea
[21:17:33] <SWPadnos> is there some really bat thing I missed?
[21:17:35] <SWPadnos> bad
[21:17:50] <Jymmm> things happen for a reason
[21:18:05] <SWPadnos> maybe it's payback for the accident
[21:18:14] <Jymmm> in your favor =)
[21:18:14] <SWPadnos> karma and all
[21:18:17] <SWPadnos> indeed
[21:18:32] <Jymmm> and he has 100%
[21:18:39] <SWPadnos> hmmm - maybe I should search for motor mounts again ;)
[21:18:43] <SWPadnos> yeah - I noticed that
[21:18:54] <jepler> you've apparently learned one thing: gloat about your ebay finds on irc after the bidding is closed.
[21:19:04] <cradek> always wise
[21:19:08] <Jymmm> what he said
[21:19:17] <SWPadnos> of course ;)
[21:19:40] <Jymmm> We may be friends on irc, but alls fair in love and ebay!
[21:20:11] <SWPadnos> heh
[21:21:09] <cradek> does this use normal hand taps with the square on the end?
[21:21:21] <Jymmm> i dont need 100 feet of this stuff =(
[21:21:45] <SWPadnos> I haven't been able to figure that out yet ;)
[21:22:02] <SWPadnos> I think they're supposed to work, but I've had some trouble with my 30 TC/DC
[21:22:19] <SWPadnos> though I don't have the wrenches to tighten the collet, so it may have been loose
[21:22:28] <Jymmm> you do now
[21:22:40] <SWPadnos> for this one, yes. they're different sizes
[21:22:46] <Jymmm> oh.
[21:22:55] <cradek> yeah the 3/8 collet part worries me
[21:24:15] <SWPadnos> this one is for big taps
[21:31:16] <Jymmm> I need to come up with a spindle so I can use my tapping head on my router
[21:54:12] <Jymmm> Hmmm, the RoHS regulations is kinda interesting. Anyone got their non-lead solder yet? http://www.action-electronics.com/rohs.htm
[21:55:20] <Jymmm> or are stock pilying the lead solder =)
[21:56:35] <fenn_> you can even get rohs compliant lead solder from jameco
[21:58:20] <Jymmm> Um, isn't the purpose behind the RoHS to not HAVE any lead/mercury, etc?
[22:01:34] <Jymmm> In red... Jameco bases RoHS compliance on information solely provided by the product's manufacturer.
[22:02:34] <Jymmm> “Member States shall ensure that, from 1 July 2006, new electrical and electronic equipment put on the market does not contain lead, mercury, cadmium, hexavalent chromium, polybrominated biphenyls (PBB) or polybrominated diphenyl ethers (PBDE).
[22:03:53] <K4ts> night I go to play a poker byeeeee
[22:04:13] <Jymmm> I think hexavalent chromium is what PG&E into BIG trouble
[22:04:18] <K4ts> hi Jymmm
[22:04:21] <Jymmm> G'Night K4ts, enjoy!
[22:04:33] <Jymmm> K4ts wear lots/little clothes =)