#emc | Logs for 2008-01-31

[00:00:17] <gezar> oh, sorry, and I got interested in what you were asking about cause I didnt know either, it was interesting to me
[00:00:49] <ds2> yeah, I been wanting to do more stuff in the shop
[00:01:12] <ds2> didn't expect it either, moved 500lbs of bricks friday night then felt it that saturday :(
[00:01:17] <jepler> it's in this PDF: Applications > CNC > EMC2 Integrator Manual, section EMC related HAL > Hardware Drivers > Mesa Electronics m5i20
[00:01:28] <owhite> oh cool I can try to light up the on board LEDs.
[00:02:02] <owhite> hmm.. I dont have the integrator manual. but that's alright.
[00:04:40] <dmess> ds2 needs yoga.... HOT YOGA... ;)
[00:05:04] <ds2> uh err...
[00:07:03] <dmess> http://asanabodymind.com/ its working on my OLD aches
[00:33:28] <GNieport1> I'm still working on PID, any thoughts on what gain might need adjusting for a following error when jogging at max velocity once the axis has travelled 4-5 inches?
[01:19:51] <GNieport1> I'm still working on PID, any thoughts on what gain might need adjusting for a following error when jogging at max velocity once the axis has travelled 4-5 inches?
[01:20:09] <eric_U> proportional is always good
[01:20:21] <eric_U> maybe a little integral?
[01:20:26] <GNieport1> okay, thanks
[01:20:35] <eric_U> I hate integral though
[01:20:55] <GNieport1> I have P of about 1000, I of 30, D of 6.2
[01:21:58] <GNieport1> if you'll remember though, this is a smart servo that has its own internal loops, so comparisons might not mean much ;-)
[01:22:12] <eric_U> yeah, numbers are meaningless
[01:23:39] <Dr0nE> hello
[01:24:35] <Dr0nE> I am wondering if there is any script out there to convert DXF (2d) image into a _Plasma_ CNC cut path?
[01:27:43] <GNieport1> How do I clear the "homed" flag for an axis?
[01:28:46] <cradek> you can't unhome without restarting emc
[01:32:45] <twice2> GNieport1: is problem with only one axes?
[01:34:25] <twice2> if so, it's possible some mechanical issue with the axis
[01:34:53] <cradek> without seeing a plot of what goes wrong it's hard to say
[01:35:41] <cradek> maybe the error is building up throughout the move, or maybe it comes all of a sudden
[01:37:04] <twice2> sometimes it's hard to put a new hat on
[01:38:31] <owhite> hey folks. I'm running the halvcp test panel, and applied the board's on-board five volts to an input and it doesnt seem to be changing the LEDs on the pyVCP panel. Any suggestions?
[01:39:30] <owhite> oh...with my M5i20 board, that is.
[01:40:06] <Dr0nE> Is there a means to use both serial and parallel communications with EMC2 to execute a tool change?
[01:42:53] <cradek> Dr0nE: anything EMC's HAL can do can be used for changing tools - without more information it's hard to answer your question
[01:44:23] <Dr0nE> my thoughts are to build a tool holder that is connected via serial, but the axis/feedback is using parallel.
[01:44:45] <Dr0nE> basicly, EMC can drive both serial and parallel at the same time?
[01:46:51] <Ziegler> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Cam
[01:47:39] <Ziegler> http://timeguy.com/cradek/autocad << cradek did this
[01:48:15] <Ziegler> Dr0nE: ?
[01:48:21] <jepler> Dr0nE: emc's io controller produces some values known as iocontrol.0.tool-change and iocontrol.0.tool-prep-number. You can write a component which is hooked to those values and does whatever actions you need. that component, in turn, can communicate with any hardware that has HAL drivers to perform the necessary actions.
[01:48:52] <Dr0nE> ok
[01:49:12] <Ziegler> do you just want a dxf -> gocode converter?
[01:49:23] <Ziegler> err gcode
[01:49:31] <Dr0nE> no
[01:49:33] <jepler> emc doesn't include any hardware drivers that talk to serial devices, because there is simply too much variation among serial devices for us to standardize. but you can write your own drivers (for instance, a userspace driver that uses linux's serial drivers to send bytes back and forth)
[01:49:45] <Ziegler> Reason why drone?
[01:50:18] <Dr0nE> i'm using model drawings that arent set as cut paths, just outlines to the objects i want
[01:50:49] <Ziegler> well there are some dxf -> gcode that can do offsets
[01:51:24] <Ziegler> Might be able to edit REALIZE to do offsets if it already doesnt
[01:51:24] <Dr0nE> thanks, jepler. i'll have to look into that when i start to write the serial interface
[01:51:26] <Ziegler> or...
[01:51:37] <Ziegler> parallel out your outlines
[01:52:13] <Dr0nE> parallel out your outlines?
[01:52:41] <cradek> depending on your outlines you may be able to use tool radius offset to cut 'outside' the path
[01:52:53] <Ziegler> in autocad use the "parallel" command to move your partlines out the desired distance
[01:53:02] <cradek> Ziegler: "offset"
[01:53:15] <Ziegler> danke
[01:53:35] <Ziegler> I am a MicroStation CAD admin...
[01:53:46] <Ziegler> dont know autocad as much as I used to
[01:54:22] <Dr0nE> this is the sort of thing i would want it to accept - or its offset version: http://dans-workshop.info/cnc2/index.htm
[01:55:14] <cradek> Ziegler is right that you could generate tool paths from those with autocad's "offset" command and then use REALIZE
[01:55:48] <Dr0nE> REALIZE would complete this?
[01:56:04] <cradek> http://timeguy.com/cradek/autocad
[01:56:06] <Ziegler> http://timeguy.com/cradek/autocad << REALIZE
[01:56:11] <Dr0nE> i'll have to look it over
[01:56:12] <cradek> you can read about exactly what it does here
[02:04:57] <Dr0nE> cradek, minor question. i found were to load this script into autocad, but not sure to actualy run it, any idea? acad '08
[02:08:22] <cradek> Dr0nE: just type realize at the prompt
[02:09:48] <Dr0nE> ah
[02:13:56] <tomp> Dr0nE: I think i see mortise & tenon like features on those parts. Do they assemble that way?
[02:14:11] <Dr0nE> i'm getting this "; User warning: assignment to protected symbol: TYPE <- "LINE"" and it doesnt create an output file.
[02:14:48] <GNieport1> cradek: Hi, plot of following error during fast jog http://imagebin.ca/view/YR_LXqny.html
[02:15:18] <Dr0nE> tomp, these parts are planed to be put together as a puzzle, not quite mortise/tenon.
[02:15:35] <cradek> GNieport1: output is saturating. probably it just can't go that fast
[02:15:43] <GNieport1> hmm
[02:15:49] <Ziegler> hehe... giant plasma cut steel puzzle
[02:16:00] <jepler> GNieport1: "saturating" is the portion of the graph where the green gets to 'the top' and can't increase further
[02:16:50] <GNieport1> jepler, why would the axis reaching its max velocity cause a following error
[02:17:30] <GNieport1> oh, are you saying that the drive can't go 'faster' to catch up the error
[02:17:35] <jepler> GNieport1: because emc wants it to go (e.g.) 60in/s but with the DAC output at +10v the motor only gives 50in/s
[02:17:44] <GNieport1> hrm
[02:17:57] <cradek> Dr0nE: sorry I don't understand that error - I wonder if autocad broke backward compatibility somehow
[02:18:25] <Dr0nE> cradek, what version of autocad was this designed for?
[02:18:41] <GNieport1> would reducing the axis max velocity help, or do i need to modify my output_scale
[02:18:59] <cradek> I have R12, but I have reports of it working through autocad 2005
[02:19:14] <Dr0nE> ok, then its probly an issue with my version.
[02:19:15] <cradek> GNieport1: just reduce the max velocity
[02:20:29] <GNieport1> :-P
[02:20:42] <cradek> Dr0nE: I'd be happy to incorporate fixes if you can figure out what's wrong and we can fix it so it works on all versions
[02:21:34] <Dr0nE> i'll see what i can find.
[02:22:42] <jepler> http://www.kxcad.net/autodesk/autocad/AutoLISP_Developers_Guide/ws73099cc142f4875516d84be10ebc87a53f-7c65.htm
[02:22:59] <jepler> maybe you have to use a different name than 'TYPE' somewhere in REALIZE?
[02:23:05] <Dr0nE> lol, that will help, jepler. as i dont know autocad's programming
[02:23:14] <jepler> though that page says it is only a warning...
[02:23:38] <Dr0nE> it also asks a question during run, let me get the notes on it
[02:23:44] <cradek> (setq type (cdr (assoc 0 ent)))
[02:24:10] <Dr0nE> "Assignment to protected symbol: TYPE \nEnter break loop?"
[02:24:32] <Dr0nE> is that what you are talking about jepler?
[02:25:42] <cradek> Dr0nE: would you try this new version please http://timeguy.com/cradek-files/emc/realize.lsp
[02:25:51] <tomp> Dr0nE: i think so, the web page sez "f you choose No, the symbol's value is modified, and processing continues normally. If you choose Yes, processing is interrupted, and you enter a Visual LISP break loop."
[02:26:47] <Dr0nE> ah
[02:27:17] <Dr0nE> lol, so i'll have to correct that variable name then. just noticed it generated an output when i kept hitting no
[02:27:59] <jepler> Dr0nE: cradek did that in the url he just posted ^^
[02:28:33] <tomp> GNieport1: maybe your acc/decel is a bit strong, see the jerk off the line and just before max vel? something to try.
[02:29:32] <Dr0nE> missed that line.sry cradek
[02:29:52] <GNieport1> okay, and thanks for the motor jump thoughts the other day
[02:30:51] <GNieport1> tomp: I lowered the axis max velocity by 0.3 inch/sec and teh problem went away
[02:30:58] <Dr0nE> cradek, that version of REALIZE generated the same TYPE errors
[02:31:18] <cradek> let me make sure I uploaded the right one
[02:31:38] <cradek> yep
[02:31:44] <cradek> are you sure you reloaded the right one?
[02:31:46] <Dr0nE> actualy, let me download that again. seems it didnt take ur changes
[02:32:18] <GNieport1> tomp: my max velocity was set to the maximum possible motor speed, so the servo loop could never 'catch up'
[02:32:40] <tomp> ah, very good catch, now give it a bit of headroom
[02:33:23] <GNieport1> tomp, what percentage would be a good starting point
[02:34:02] <GNieport1> it has around 5% velocity headroom right now
[02:34:44] <tomp> now torture it with fast reversals. I'd be happy loosing 10% capability for more realiability, so i'd never ask more than 80% of max anything.
[02:35:13] <GNieport1> heh, in the past the machine did not like tapping 'jog' quickly
[02:35:14] <tomp> at fast reversal, the dac might overshoot, test and see
[02:35:51] <GNieport1> I have a file that does g0 +X; g0 -X
[02:36:44] <GNieport1> on second thought, does g1 have a steeper accel/decel curve?
[02:36:51] <cradek> no
[02:37:22] <GNieport1> okay, thanks
[02:37:31] <Dr0nE> cradek, ok. your edit cleared up the errors and generated an output file without any warnings. but the output file has no lines, just header/footer code
[02:38:31] <cradek> Dr0nE: I bet $1 that you did not read the instructions then: http://timeguy.com/cradek/autocad
[02:39:07] <Dr0nE> that nothing on layer 0 will be used?... that needs to change lol
[02:39:22] <tomp> i'd do this outside of any control, then report the cautious&adjusted&tempered max vel to the control's parameters (emc's ini), but if you wanna try inside emc , then g0 is your tool.
[02:41:49] <Dr0nE> hm, type error again.
[02:42:16] <cradek> what's the exact error this time?
[02:42:29] <Dr0nE> same as before "TYPE"
[02:42:53] <Dr0nE> but this time it has a "LINE entity at z>=0. Aborting..." message
[02:43:23] <cradek> ok, that's an intended error, maybe the TYPE is old or something
[02:43:50] <Dr0nE> omw, Visual LISP is screwing with the lsp file. it reverted the file back again
[02:54:18] <fenn> did i mention lately how happy i am with linux?
[02:55:46] <fenn> Dr0nE: is it "drone" or "Dr. One"?
[03:01:35] <Dr0nE> drone
[03:01:50] <Dr0nE> as in: lazy male bee in the hive
[03:19:24] <GNieport1> I managed to catch one of the 'jog-induced' following errors on the scope. This is two successive clicks on the jog button: http://imagebin.ca/view/Ua6xt9M0.html
[03:19:53] <GNieport1> It ended in oscillation :)
[03:20:30] <GNieport1> two independent servo loops seem to make quite a mess of thiongs
[03:20:39] <GNieport1> things, too
[03:21:16] <GNieport1> I can see why jmkasunich said 'I hate smart drives'
[03:26:16] <toastydeath> has anyone ever used a birmingham/victor/sharp/precision matthews lathe
[03:26:31] <toastydeath> 16" swing or larger, not 14" and under
[03:36:26] <tomp> borg equipment! a toy accessory is a hud lcd built by kopin. lores & b&w. if they ever come back in stock, it's a 25 buck headsup display. http://www.shopwildplanet.com/prod/WPT70133-HEAD.html
[03:36:29] <Gamma-X2> i hate overtime! AHHH
[04:08:04] <cradek> GNieport1: that oscillation doesn't look good does it
[04:09:34] <skunkworks> yikes
[04:10:01] <eric_U> there really is no issue in cascading pid controllers
[04:11:08] <LawrenceG> .001 at 300ipm is pretty good..... wonder what error looks like at normal cutting speed?
[04:11:13] <eric_U> probably has too much integral gain between the two, integral is destabilizing, pole on the origin and all
[04:13:22] <LawrenceG> at 300ipm... machine is moving 0.005" every servo update
[04:14:07] <toastydeath> is that contouring or straight line
[04:14:19] <skunkworks> GNieport1: didn't you say something about 60 cycle hum on something yesterday?
[04:43:27] <GNieport1> cradek: I've only had two or three thousand oscillations like that while tuning this
[04:44:16] <GNieport1> skunkworks: that was some hum on a oscope monitor port on my amplifiers
[04:45:16] <GNieport1> LawrenceG so you suggest to increase my servo thread by a factor of 10?
[04:45:26] <GNieport1> I'll give it a try tomorrow
[04:45:33] <GNieport1> good night,all
[04:53:39] <LawrenceG> GNieport1: I am not sure if you can go by 10, but you see what effect doubling it has (ie 500us instead of 1ms)
[04:55:02] <LawrenceG> GNieport1: I have played with faster, but encoder resolution becomes an issue.... at slow speeds, control degrades because of only a few encoder pulses/servo cycle
[07:03:56] <kdemetter> hi
[07:04:05] <kdemetter> i have a question
[07:05:40] <kdemetter> i'm looking for a Linux OS , capable of realtime , in order to control a small robot
[07:06:00] <kdemetter> it's controlled via the parallel port
[07:06:13] <kdemetter> is EMC capable of this ?
[07:07:29] <fenn> depends what you want the robot to do
[07:07:56] <fenn> emc is more for controlled environments, not much "AI" stuff
[07:08:32] <fenn> but you can extend it quite easily
[07:09:10] <kdemetter> well , it's mostly for basic movement and some sensors
[07:09:23] <kdemetter> i think i'll give it a try :-)
[07:09:29] <kdemetter> thx for your help
[10:58:27] <alex_joni> gezar: ping me when you're around
[13:03:06] <micges> hello\
[13:16:33] <alex_joni> hi micges
[13:20:15] <micges> alex_joni: always after release new emc is so quiet on irc?
[13:24:41] <alex_joni> micges: if it's a good release :P
[13:24:50] <alex_joni> if not.. we get lots of users in here complaining :D
[13:25:00] <alex_joni> (or do you mean developers?)
[13:28:50] <micges> developers
[13:32:53] <alex_joni> hmm.. dunno, it depends
[13:33:06] <alex_joni> developers are usually working on things they like to
[13:34:57] <alex_joni> so it's not that tied to releases
[13:35:15] <alex_joni> but usually before a release there's extra energy involved.. so after that it's quiet for a while
[13:57:11] <skunkworks_> logger_emc: bookmark
[13:57:11] <skunkworks_> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2008-01-31.txt
[13:58:27] <micges> alex_joni: me understand
[14:22:21] <archivist> oo /me has nearly got gerber reading, understood www.archivist.info/cnc/pcb/readgerber.php?gerberfile=5370%5C5370-Copper.pho
[14:24:43] <skunkworks_> if you got a gerber input filter for emc working - I think tons of people would be very very happy
[14:25:46] <archivist> well first part on the way
[14:26:28] <archivist> I need to join segments as kicad gives a nice borked but correct output
[14:26:55] <archivist> sample gerbers needed to test
[14:27:22] <skunkworks_> tom posted this the other day - thought it deserved another look http://www.ozbricks.com/bobfay/om4.htm
[14:27:36] <skunkworks_> also http://youtube.com/watch?v=6AptvePhMwE
[14:27:51] <archivist> did you notice how kicad routed outide the border through the connecter
[14:30:23] <skunkworks_> ? I don't know what your talking about.. I don't know kicad
[14:32:14] <archivist> see above link
[14:32:47] <archivist> the row of rectangles at bottom of board
[14:36:41] <skunkworks_> oh - look - pictures
[14:37:01] <archivist> it will probably end up standalone so one can decide a few things
[14:38:03] <archivist> skunkworks, thats live running code each time you click only the kicad pic is a screen grab
[14:38:18] <alex_joni> whee .. nice
[14:38:37] <skunkworks_> have you looked at the image import filter in emc2? A dialog box pops up and lets you adjust a bunch of parameters..
[14:40:12] <archivist> no, trouble is for double sided one needs to not cut border one side and drill register holes, then turn over and do reverse
[14:41:48] <skunkworks_> I get a 404 error when I click on the picture.
[14:41:55] <archivist> it currently draws a bitmap and stores as a .jpg, just getting my understanding correct first
[14:42:22] <archivist> skunkworks, it wasnt meant to be a link
[14:44:12] <archivist> fixed
[14:57:37] <alex_joni> bbl
[14:57:41] <[1]a-l-p-h-a> [1]a-l-p-h-a is now known as a-l-p-h-a
[16:11:39] <skunkworks_> cradek: do you still have the video of your max tapping?
[16:11:52] <skunkworks_> sorry - I mean threading..
[16:12:29] <skunkworks_> (when you got the lathe threading code running) (didn't you have a video?)
[16:14:00] <skunkworks_> or was it just pictures?
[16:15:03] <cradek> I don't think I had a video camera then
[16:16:26] <skunkworks_> that could be..
[16:16:33] <skunkworks_> (it was a long time ago ;))
[16:16:54] <cradek> it's a crappy lathe anyway
[16:17:13] <skunkworks_> I bet
[16:18:22] <archivist> I cured my crappy lathe, its no longet a lathe
[16:18:24] <skunkworks_> http://www.youtube.com/watch?v=wbfXXeiSHsw&eurl=http://www.cnczone.com/forums/showthread.php?t=51075
[16:19:29] <cradek> heh, cool tool changer
[16:20:41] <skunkworks_> (:
[16:20:48] <skunkworks_> odd :)
[16:27:59] <archivist> nice, I had a similar idea here and intend the lathe tools be on the rotary table
[16:28:28] <archivist> hadnt thought of the stationary drill
[16:29:35] <archivist> as my mill headstock is the old lathe headstock
[16:50:22] <tomp2> skunkworks_: do you know what the lego stuff used for ways?
[16:50:29] <tomp2> I imagine they used some linear rail or drill rod, but cant see it in any photo.
[16:50:30] <tomp2> i thought i'd use lego machines & pport steppers when i travel.
[16:50:52] <skunkworks_> I would doubt it - from my lego days - they had pieces that where smooth on one side..
[16:51:15] <tomp2> (portable mini cnc just for testing, not cutting)
[16:51:29] <tomp2> oh! cool
[16:51:29] <skunkworks_> the new mindstorm kits - the motors are servos.. iirc
[16:51:42] <skunkworks_> (feedback)
[16:51:51] <tomp2> yep, even some encoder in the module NXT
[16:51:54] <alex_joni> yeah, but rotaries
[16:52:01] <alex_joni> so you need to convert to rack/pinion
[16:52:08] <skunkworks_> * skunkworks_ bet alex knows more..
[16:52:17] <alex_joni> I played with one once
[16:52:36] <tomp2> yeh, i bet they use rack, i got haydon 5V steppers with 100mm of leadscrew captive
[16:52:44] <alex_joni> I think the protocol is available somewhere online, so you could just talk to them from emc2
[16:53:28] <tomp2> oh, i dont plan on using thier controller, just brix for a liteweight mock machine
[16:55:00] <tomp2> http://shop.lego.com/Product/?p=9842 thats a thought tho
[16:55:59] <alex_joni> that's the one
[16:57:27] <alex_joni> tomp2: http://jwstolk.xs4all.nl/millp.htm
[17:00:31] <tomp2> hah! thx :) looks a lot like my motors, but i hope my portable would be simpler & smaller
[17:01:12] <tomp2> lookin for the protocol, but found the disassembly http://www.philohome.com/nxtmotor/nxtmotor.htm
[17:01:55] <tomp2> i2c
[17:05:23] <alex_joni> heh: http://www.theregister.co.uk/2007/10/19/bofh_poll_20071019/
[17:07:06] <alex_joni> tomp2: nice
[17:18:18] <tomp2> :-)\
[17:49:27] <gezar> alex_joni: you there?
[19:09:25] <owhite> hey folks
[19:11:23] <owhite> would anyone have a suggested approach to allow me to make a gcode call that passes a 6bit value to hal, and those bits are used to set 6 bins to high or low?
[19:11:34] <owhite> s/bins/pins/
[19:23:33] <jepler> owhite: a M1xx script can execute 'halcmd sets' to set the value on a signal that doesn't have a writer. http://pastebin.ca/886557
[19:24:22] <jepler> owhite: I created bit signals 'sig1' ... 'sig6' (which could be linked to whereever the value needs to go) and then I can manipulate the signals with that script. e.g., if I execute M102 P9, then sig1 and sig4 are set TRUE, and the rest are FALSE
[19:26:28] <owhite> jepler: thank you very much. that seems pretty self explainatory. I'm away from my emc computer right now but I'll take a look at the code tonight and give it a shot.
[19:27:04] <jepler> owhite: the two tricky parts were these:
[19:27:13] <jepler> val=${1%.*} # strip off the fractional part of the "P-number"
[19:27:43] <jepler> and the "sets" function which has to turn any nonzero value into "1", because 'halcmd sets whatever 37' doesn't work if 'whatever' is a bit signal (only 0 or 1 are accepted)
[19:28:14] <owhite> got it. and what language am I looking at -- its a shell script?
[19:28:49] <BigJohnT> I was just going to ask that question...
[19:31:23] <jepler> yes it's a shell script for the bash shell -- that's what the first line (#!/bin/bash) means
[19:32:22] <owhite> well yah I figured as much but isnt that 'sets' a halcmd?
[19:33:23] <jepler> sets () { if [ $2 -eq 0 ]; then halcmd sets $1 0; else halcmd sets $1 1; fi }
[19:33:40] <jepler> this line creates a new shell command called "sets". The part inside the {} is what actually happens when bash encounters a "sets" command
[19:34:06] <owhite> okay.
[19:34:19] <jepler> if it's "sets mumble 0" then the shell will actually run "halcmd sets mumble 0"; if it's "sets mumble 39" then the shell will actually run "halcmd sets mumble 1"
[19:34:23] <jepler> that was one of the tricky parts :-P
[19:35:07] <skunkworks_> what about using another axis? b100? then 100 would be available to hal... ?
[19:35:47] <skunkworks_> 6 bit though
[19:36:11] <skunkworks_> 1-63
[19:38:24] <skunkworks_> then send it thru a mux? (don't know if it is the right hal componant)
[19:39:13] <skunkworks_> * skunkworks_ is just winging it again.
[19:40:51] <micges> +
[19:40:57] <owhite> I quite like jeplers suggestion -- just what I needed. pretty soon I'll be able to change the power settings on my laser through gcode.
[19:41:18] <owhite> gcode --> hal --> M5i20 --> pwm board --> laser.
[19:42:47] <gezar> alex_joni: Ill be here for about 20 more minutes then I have to go to my compsci lab class, then ill be back in 2 hours
[19:47:26] <renesis> hi
[19:47:43] <renesis> backlash compensation code is fine, yesno?
[19:52:02] <Unit41> lawlzorzoro
[21:23:36] <tomp2> jepler: thanks for the m102 stuff. the wiki talks about how a gear shift code would be done for the mazak using same, but i havent found it. Has it been published?
[21:25:41] <jepler> tomp2: I *think* that at present the gear shift is not done automatically, but manually
[21:25:44] <jepler> don't quote me on that
[21:26:14] <jepler> I also think that the demo_mazak configuration in emc 2.2.3 is the same one that we were using at fest in June 2006 (the last time I was there)
[21:27:40] <skunkworks_> I think the gear shift is automated.. I remember something about a custom hal module that would decide when to shift automatically...
[21:27:47] <skunkworks_> but I have been wrong before
[21:34:46] <skunkworks_> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?MazakSpindleGearChange
[21:55:26] <skunkworks_> from mazak hal file
[21:55:29] <skunkworks_> # select either high or low speed based on current gear ratio,
[21:56:33] <skunkworks_> # connect signals to gearbox controls
[21:56:50] <skunkworks_> linksp sp-engage-high-gear ax5214h.0.out-21
[21:57:00] <skunkworks_> linksp sp-engage-low-gear ax5214h.0.out-20
[21:57:21] <skunkworks_> linksp sp-in-high-gear ax5214h.0.in-34
[21:57:29] <skunkworks_> linksp sp-in-neutral ax5214h.0.in-35
[21:57:37] <skunkworks_> linksp sp-in-low-gear ax5214h.0.in-36
[21:57:48] <skunkworks_> not in that order..
[22:14:00] <alex_joni> darn, owhite left
[22:14:15] <alex_joni> wonder why the spindle-speed wouldn't be good to use as the laser power
[22:14:32] <alex_joni> (then things like spindle-override work too..)
[22:17:52] <skunkworks_> duh
[22:18:00] <skunkworks_> that seems like a really good idea ;)
[22:18:55] <alex_joni> he would just need to remember to clamp the pwm output to his needed max
[22:22:14] <skunkworks_> our fanuc controller uses t and s to set the 'power'
[22:24:18] <skunkworks_> t sets how long the laser is 'on' and s sets how distance between 'on'
[22:24:45] <skunkworks_> * s sets the distance between 'on'
[22:25:12] <alex_joni> you're talking about a pulsed laser
[22:25:19] <skunkworks_> yes
[22:25:25] <alex_joni> ok, nice
[22:25:35] <alex_joni> so on-time and off-time
[22:25:41] <skunkworks_> but it in effect sets duty cycle
[22:25:43] <alex_joni> are those in msec usually?
[22:25:49] <skunkworks_> ms and inches
[22:25:57] <alex_joni> I expect 3-4 msec pulses..
[22:26:02] <skunkworks_> yes
[22:26:10] <alex_joni> and the power is fixed?
[22:26:11] <skunkworks_> (it has been a while)
[22:26:20] <alex_joni> as in max power for the pulse each time/
[22:26:21] <alex_joni> ?
[22:26:46] <gezar> i r home
[22:27:06] <skunkworks_> well - we have a power 'knob' that sets max power. but you can (by setting the s and t's) get any power you want.
[22:27:38] <skunkworks_> (up to the max power set)
[22:27:46] <alex_joni> right.. that was my point
[22:27:48] <gezar> alex_joni: well, I get user and dev that is
[22:36:52] <tomp2> thx for the mazak gear info, what was the 'in high gear' sensor ( tomp imagines spring loaded gear axles and prox switches )
[22:37:12] <tomp2> (step 9)
[22:38:41] <alex_joni> tomp2: look around fennetic.org? he has some mazak docs & schematics
[22:38:55] <tomp2> thx (and thx fenn:)
[22:40:39] <tomp2> fenn.freeshell.org today
[22:43:33] <tomp2> oh yeah ( i was taking it apart & Ben was photographing ) a vert hyd cylinder with stop rings and proxes reading the vert posn.
[22:45:02] <gezar> watts is volts/amps?
[22:45:09] <tomp2> v*a
[22:45:32] <tomp2> in dc
[22:45:55] <ds2> or A if V and A are allowed to be complex
[22:45:57] <ds2> AC
[22:47:15] <gezar> I dont know
[22:49:47] <tomp2> wikipedia is my friend 1W = 1V * 1A "that is, if 1 volt of potential difference is applied to a resistive load, and a current of 1 ampere flows, then 1 watt of power is dissipated"
[22:50:24] <tomp2> and "i dont know' is a what , not a watt :)
[22:50:43] <gezar> well, I was answering another guys question in another channel, and it landed here
[22:54:06] <ds2> as long as resistive, everything is real... when it gets reactive, the values are complex
[22:54:29] <ds2> or you can reduce to the polar form and use V*A*cos(phase diff)
[23:40:56] <Hugomatic> Hi guys...
[23:41:06] <alex_joni> 'lo
[23:47:04] <Hugomatic> I use python scripts for circles, rectangles and stuff... is there a 'best practice' for pockets? I mean should I start from the outside and then remove the middle? Should I plunge straight in the corner? Does any of this matter at all? thanks
[23:51:23] <Hugomatic> I figure its best to start inside and work your way out...
[23:52:09] <alex_joni> Hugomatic: I believe that is most common
[23:54:28] <Hugomatic> thank
[23:55:35] <alex_joni> Hugomatic: keep asking in here, there are other people with more experience on different controls..
[23:56:11] <alex_joni> good night all
[23:56:22] <Hugomatic> bye