#emc | Logs for 2008-07-01

[01:17:13] <Gamma-X> hello all
[01:36:46] <chr0n1c> hello
[01:38:10] <gezr> howdy yall, hope that fest was a good time
[01:39:42] <chr0n1c> "congradulations your email have one uno billion dollar" <-- spam i got a few mins ago
[01:45:42] <gavin> hi,may I take the EMC2 to control 5 step motors simutanously and output max. 20kHZ pulses for
[01:45:43] <gavin> each motor?
[01:48:38] <SWPadnos> yes
[01:54:31] <gavin> SWPadnos,thank you
[02:06:09] <Gamma-X> SWPadnos, wanna come to li??? lol
[02:24:24] <chr0n1c> http://www2.fwi.com/~kimble/scispec/scispec.htm <- the awesome bmp to 3d gcode program i have been hunting for two years
[02:33:13] <SWPadnos> gavin, you're welcome
[02:33:13] <SWPadnos> Gamma-X, not at the moment
[02:33:14] <SWPadnos> chr0n1c, why not use image-to-gcode, which seems to do all that program does, but for any file type that imagemagick can convert
[02:35:41] <chr0n1c> swp, well, i used that program to do some awesome stuff in the past, i just wanted to find it... i never got around to researching image-to-gcode
[02:35:51] <chr0n1c> *yet
[02:36:03] <SWPadnos> it's on any relatively recent EMC2 install
[02:36:14] <SWPadnos> I think it is anyway
[02:36:18] <chr0n1c> i know how to tweak that program, it does beautiful loooong code ;)
[02:37:03] <chr0n1c> it can do simple things, but you can make people drool with that program/the right cutter/the right machine
[02:37:28] <chr0n1c> * chr0n1c is going to look into the emc version
[02:37:33] <SWPadnos> if you have an emc install (or just a liveCD to boot from), look at torus.png and torus.ngc
[02:38:25] <SWPadnos> hmmm. actually, just torus.png - load it up in EMC and image-to-gcode should run as a filter
[02:38:35] <chr0n1c> i saw the tux 3d demo... looks nice i just need to figure it out
[02:38:57] <chr0n1c> do i still have to edit my ini file to be able to use it?
[02:39:13] <chr0n1c> i am kinda lazy is why i never did that
[02:39:22] <chr0n1c> oh, i didn't say that
[02:39:27] <chr0n1c> -lazy
[02:39:42] <SWPadnos> that was generated with something else, not image-to-gcode
[02:39:49] <chr0n1c> oh..
[02:40:00] <SWPadnos> you should be able to load it on many of the default configs
[02:40:14] <chr0n1c> well, you can do that with this acme profiler thing
[02:40:15] <SWPadnos> tux isn't really machinable - there are no roughing passes
[02:40:20] <chr0n1c> th elink i posted
[02:40:39] <SWPadnos> image-to-gcode will generate roughing passes as well
[02:41:24] <SWPadnos> the thing it doesn't do is show a cool triangle mesh generated from the image
[02:41:55] <gezr> oh oh, so far, ive got An A in cal 2, struggling though, tough stuff
[02:42:46] <chr0n1c> woot!
[02:42:51] <gezr> and it looks like im going to end up owing someone a beer, I wont be able to finish the mini mill, I have zero time
[02:43:01] <chr0n1c> you can do it!
[02:43:27] <gezr> calculus 2 suxzors hard core
[02:45:03] <gezr> but im close to knowing the math for emc
[02:47:02] <renesis_> ur mom knows the math for emc
[02:47:03] <renesis_> er, hi
[02:49:43] <fenn> would be nice if there were some way for an axis filter to show the path preview before actually finishing loading the program
[02:51:02] <SWPadnos> SMOP
[02:51:27] <fenn> axis-remote?
[02:51:30] <SWPadnos> python filters could import emcmodule and togl, and maybe do some other stuff to be able to generate previews
[02:54:00] <jepler> axis-remote probably won't work properly while a load is in progress (it'll enter something recursively that it shouldn't)
[02:54:38] <jepler> either changing (wildly improving) integration between axis and gcode generators, or making a full-blown standalone preview would be very welcome features
[02:55:27] <jepler> goodnight all
[02:56:08] <renesis_> bye
[03:02:59] <LawrenceG> * LawrenceG putting 8.04 on 500mhz P3 takes forever to do the updates
[03:03:28] <chr0n1c> cya jepler
[03:03:47] <chr0n1c> * chr0n1c likes the 8.04
[03:04:25] <chr0n1c> .. i think it finally hit the xp level of competition
[03:05:15] <chr0n1c> ...newish ubuntu in general
[03:06:03] <LawrenceG> yep.... runs well... although it seems to have dropped support for late 90's video cards like 1mb or 2mb S3 cards
[03:22:28] <alex_mobile> morning
[12:32:38] <ChanServ> [#emc] "This is the #emc channel - talk related to the Enhanced Machine Controller and general machining. Website: http://www.linuxcnc.org/, wiki at http://wiki.linuxcnc.org/"
[12:32:40] <SWPadnos> yes
[12:35:36] <als> thanks
[12:35:45] <SWPadnos> sure - thanks for reminding me
[14:54:49] <jepler> roguish: OK, we are trying to determine whether it is something you are running in the realtime code that contributing to the 'realtime delay' message.
[14:55:13] <Roguish> ok.
[14:55:53] <Roguish> as i have said. it occurs on 3 different installs. all running 'trunk'
[14:56:05] <jepler> do you mean 3 different machines, or just 3 different copies of the source code?
[14:56:15] <Roguish> 3 different computers
[14:57:20] <jepler> does it happen while running a distributed configuration file like sim/tkemc.ini or sim/axis.ini? or only when running your own configuration files?
[14:57:45] <Roguish> my own. i have 5i20's
[14:58:32] <Roguish> and don't mess too much with the realtime stuff, other that hal files to get the hardware to run.
[14:58:47] <jepler> OK, so back to the diagnostic we were trying to get before you had to reboot
[14:59:30] <Roguish> this box has a really simple 3 axis configuration with 3 small dc motors and simple pwm amplifiers (18200's)
[14:59:42] <cradek> have you posted your hal file somewhere?
[15:00:03] <jepler> . scripts/emc-environment, then run emc. in another terminal where you have also run . scripts/emc-environment, wait until you get the error, then pastebin the output of 'halcmd show thread' and 'halcmd show param *tmax'
[15:00:14] <Roguish> you want all of them?
[15:04:40] <jepler> bbiab
[17:41:05] <tomp> is using a breadboard for a pic masochistic/futile? is it smarter to build a pcb for prototypes? ( project is 170kHz to 2khz programmable osc)
[17:41:53] <tomp> er 200kHz to 2kHz
[17:43:37] <skunkworks_> I would say yes to pcb.. Playing with audio circuits was a pain with breadboards..
[17:44:10] <skunkworks_> *<50khz
[17:44:20] <tomp> ooh, good input, thx
[17:46:02] <archivist> I say yes as it debugs for production, if thats intended
[17:46:22] <archivist> else stripboard
[18:03:00] <tomp> archivist: so far the debugging just showed the bad connections in the breadboard :P
[18:03:01] <tomp> i'd like to find good stripboard in US
[18:03:31] <archivist> vero board
[18:03:45] <tomp> it's like tying baby brother to the antenna to watch your favorite show
[18:04:03] <tomp> vero is expensive, but available
[18:37:11] <LawrenceG> BigJohnT, yeo... you at the keyboard?
[18:42:45] <BigJohnT> just got back
[18:43:16] <LawrenceG> did you see the new gcode generator I added? http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Simple_EMC_G-Code_Generators
[18:43:27] <BigJohnT> not yep
[18:43:40] <BigJohnT> yet
[18:43:43] <BigJohnT> lol
[18:44:32] <LawrenceG> it started life as your arc buddy.... my python was pretty rusty, but your examples really helped
[18:44:50] <BigJohnT> cool
[18:45:34] <LawrenceG> the canvas is nice..... easy to draw a quick rendering of what the user has entered as params
[18:46:07] <BigJohnT> canvas changes with input?
[18:46:38] <LawrenceG> I just realized the example params would be a great LED flashlight disk..... 5mm leds are a press fit in 3/16 holes
[18:46:49] <LawrenceG> yes.. after pressing recalculate
[18:47:00] <BigJohnT> SWEET
[18:47:27] <BigJohnT> typing with one finger bbq on rest
[18:47:42] <LawrenceG> whats cookin?
[18:47:55] <BigJohnT> bbq takeout
[18:48:42] <LawrenceG> cool... I miss southern bbq.... noone does that up here
[18:49:47] <BigJohnT> mmmmm gooood bbq
[18:51:46] <fenn_> fenn_ is now known as fenn
[18:51:49] <BigJohnT> I can't wait to get home and look at the code. I might try and update my counterbore to show the hole locations
[18:52:40] <LawrenceG> I have already had suggestions to modify it for rectangular and oval arrays
[18:52:58] <BigJohnT> that would be cool
[18:53:59] <LawrenceG> not too hard... the only thing I would have to think about would be the test to see if a hole is inside or outside of an ellipse
[18:56:01] <BigJohnT> hmmm
[18:56:20] <LawrenceG> I could do it graphically on the canvas by drawing a filled object and reading the pixel color at x,y!
[19:05:41] <BigJohnT> cool
[20:40:18] <kirk_wallace> Hello, I want to try rigid tapping. I looked up G33.1 and wonder if I need the m3 (in), m4 (out), m3 (back to start), m5 (done) implied or just m3, m4, m5?
[20:42:49] <cradek> you run the spindle forward like normal and g33.1 does all that for you
[20:43:21] <kirk_wallace> G33.1 does the spidle commands for me?
[20:43:36] <cradek> yes
[20:43:51] <cradek> look at the end of http://cvs.linuxcnc.org/cvs/emc2/nc_files/mmount.ngc?rev=1.2
[20:44:11] <cradek> that's the code for the motor mounts we made at cnc workshop. At the end it taps some 1/4-20 holes.
[20:45:46] <kirk_wallace> Okay, I'll study the file. Anything else I should look into before I try the threads (4-40)?
[20:46:42] <cradek> for good tracking, make sure your traj and servo cycle times match and are as fast as possible
[20:47:16] <cradek> on the mazak we run at 2kHz (SERVO_PERIOD = TRAJ_PERIOD = 500000)
[20:49:11] <skunkworks_> I think JonE was running 4-40's.. At the workshop - the threads looked like crap until he fixed the traj_period also
[20:50:23] <cradek> yes he was really happy with them after that. he even took the spring out of his tap holder :-)
[20:50:35] <kirk_wallace> Oops, Mine are : SERVO_PERIOD = 1000000 , TRAJ_PERIOD = 10000000
[20:51:15] <cradek> yep that big TRAJ would be a killer. it has to run a lot more often than that to maintain the position nicely while the spindle reverses
[20:51:25] <cradek> at least remove one of the zeroes
[20:52:38] <cradek> the best description of the rigid tap cycle is here: http://www.linuxcnc.org/docview/html//gcode_main.html#sec:G33,-G33.1:-Spindle-Synchronized
[20:53:17] <kirk_wallace> No broken taps? It seems that it will either work or go badly in some way and there is no way to tell until you try.
[20:53:38] <skunkworks_> tap wood first.. :)
[20:53:50] <cradek> we were using an expensive pulley tap. it worked absolutely perfectly.
[20:54:27] <cradek> roland bought us two, but gave us one and locked the other in his desk :-)
[20:55:01] <skunkworks_> heh
[20:55:20] <kirk_wallace> The html doc link is where I got the impression that I needed the m codes, but it's just a list of what will happen automatically.
[20:55:49] <cradek> please suggest improvements in wording if you think the docs are unclear
[20:56:56] <cradek> I just read it again and think it's clear, but I wrote it, so I probably can't tell.
[20:59:55] <Dallur> BigJohnT: If you are still interested about building a THC, I just had a thought, how about getting a benchtop multimeter/ampmeter with a serial interface, using the new realtime serial driver and analysing the voltage and amps in realtime with HAL ? Total cost roughly $100 for a really nice very configurable THC :)
[21:00:22] <alex_joni> with proper separation :)
[21:00:29] <alex_joni> night all
[21:00:36] <BigJohnT> Dallur: sounds neat
[21:00:40] <cradek> goodnight alex
[21:01:02] <BigJohnT> night alex
[21:01:20] <fenn> Dallur: there's no realtime serial interface yet, and besides, aren't multimeters really slow?
[21:01:39] <archivist> some of the dvms have a slow conversion time
[21:01:51] <archivist> plus serail delay
[21:02:05] <fenn> and i dont see how it solves the noise problem either
[21:02:15] <kirk_wallace> I thought that my g-code may need to produce the actions listed as opposed to "When g33.1 is invoked, g33.2 will do x, y z"
[21:02:42] <alex_joni> fenn: rt-ser has been around a long time
[21:02:47] <kirk_wallace> opps g33.1 will doo...
[21:02:53] <alex_joni> but I bet RT is not really needed for THC
[21:03:21] <BigJohnT> I wondered about that
[21:03:30] <fenn> um i guess i meant "in hal"
[21:03:52] <kirk_wallace> I bit bang SPI in RT, does that count?
[21:04:32] <fenn> good enough for me
[21:06:03] <kirk_wallace> I came in late on the THC discussion. I guess SPI won't work if you need traditional serial.
[21:06:43] <kirk_wallace> Although, I suppose it could be bit banged also.
[21:06:55] <SWPadnos> if you have a sufficiently fast base thread, and sufficiently slow serial comms, you could make a HAL UART component that bit-bangs
[21:07:12] <SWPadnos> right :)
[21:07:48] <SWPadnos> you need a BASE_PERIOD that's ideally 4x or more * the serial bit rate
[21:08:05] <SWPadnos> err - base frequency - you know, 4 samples per bit or more
[21:08:32] <[1]Gamma-X> [1]Gamma-X is now known as Gamma-X
[21:08:43] <kirk_wallace> One issue with my SPI is that it takes a servo period for each bit out.
[21:09:03] <SWPadnos> yep
[21:09:11] <SWPadnos> oh - wait, that's slow :)
[21:09:31] <SWPadnos> you should be able to do that work in the base thread
[21:10:10] <kirk_wallace> EMC crashes when I try SPI in the base.
[21:10:36] <SWPadnos> you mean your SPI driver crashes EMC when you put it in the base thread? :)
[21:11:32] <SWPadnos> is that driver in CVS?
[21:11:44] <kirk_wallace> I am trying to recall. I think EMC won't start if I put SPI in the base period.
[21:11:56] <SWPadnos> hmmm
[21:12:10] <SWPadnos> that sounds more like a hal file error than anything else
[21:12:25] <SWPadnos> unless you mean the computer locks up hard instead of starting emc :)
[21:12:51] <SWPadnos> it could be taking too much time for the base period, but I think it was pretty short code
[21:12:52] <kirk_wallace> I wrote an HAL component: http://www.wallacecompany.com/machine_shop/EMC2/serial_dac/serial_dac.comp
[21:14:04] <kirk_wallace> I assumed that it took too much time, and it's fast enough in the servo period, so I didn't pursue it.
[21:14:46] <kirk_wallace> It would be nice to be able the see how much time is available in each period.
[21:15:11] <SWPadnos> you can look at the thread time and tmax, and subtract
[21:15:30] <SWPadnos> I think you can halscope thread time also, but I'm not positive
[21:16:29] <kirk_wallace> I'll have to note that, thanks.
[21:17:31] <Dallur2> I'm not to worried about the isolation, any multimeter with an rs232 interface most likely has sufficient isolation, probably better than the THC out there anyways
[21:19:45] <kirk_wallace> Dallur2: You might consider this: http://www.wallacecompany.com/machine_shop/EMC2/serial_adc/
[21:33:37] <BigJohnT> * BigJohnT is off to the house
[23:55:07] <dmess> hi all