#emc | Logs for 2005-06-06

[00:05:53] <dan_falck> rayh hi
[00:06:01] <rayh> Hi Dan.
[00:06:22] <dan_falck> I just got my mill set up with the DRO card and an old version of EMC
[00:06:32] <dan_falck> just using it as a DRO at the moment
[00:06:40] <dan_falck> it's pretty slick
[00:07:20] <dan_falck> I'm still unpacking the servo drives
[00:07:47] <dan_falck> But, it's a good feeling to see it almost alive again
[00:07:54] <rayh> Great.
[00:08:16] <rayh> I've started working with emc2 of late and like some of it a lot.
[00:08:31] <dan_falck> what do you like about it?
[00:08:41] <dan_falck> (haven't set it up yet here)
[00:09:54] <rayh> It starts up much faster than emc and feels really solid.
[00:10:02] <rayh> That is just an impression
[00:10:23] <rayh> but I left it on overnight and there was none of the sluggishness getting it going again.
[00:11:28] <rayh> Paul's got some good ideas for slimming down debug down at the rt level.
[00:11:39] <dan_falck> do you have all your tools working with it now? (diagnostic stuff)
[00:12:24] <rayh> I can read pins with ioshow and the exercise programs but not set any.
[00:12:55] <rayh> Jmk and I are talking about a connector that works like halcmd but with
[00:13:05] <rayh> simple interactive commands and replies.
[00:13:31] <rayh> That way we can open this program from tickle and see and set some stuff.
[00:13:32] <dan_falck> Weyland said that it runs stepper motors very smoothly
[00:13:41] <rayh> Yes.
[00:13:59] <dan_falck> do you have it hooked up to a drive box?
[00:14:05] <rayh> John says that there is an issue with running right up at max vel
[00:14:12] <rayh> Not yet.
[00:14:35] <rayh> Remember back in the old days when we had problems with following error on very long moves.
[00:14:57] <rayh> It was like steps gradually accumulated and couldn't get fed back in.
[00:15:09] <dan_falck> was that with freqmod?
[00:15:24] <rayh> Yes I think so, when it first came out.
[00:15:48] <rayh> Maxine says hi.
[00:15:56] <dan_falck> Hi Maxine!
[00:16:05] <rayh> Tell your family hello from us.
[00:16:11] <dan_falck> sure will
[00:16:54] <dan_falck> Robin says hi too
[00:16:54] <rayh> Gotta run a lightning storm on the way.
[00:16:57] <dan_falck> ok
[01:26:20] <CIA-8> 03paul_c * 10emc2/src/ (doxconfig emc/kinematics/trajectory.h): Include the AMS math symbols in the LaTeX block. Correct a minor error in trajectory.h
[01:51:46] <anonimasu> hello alpha
[02:18:51] <SWP_Away> jmkasunich: are you around?
[02:18:56] <jmkasunich> yeah
[02:18:59] <SWP_Away> SWP_Away is now known as SWPadnos
[02:19:02] <SWPadnos> hi
[02:19:10] <jmkasunich> I think I fixed the prob with stepgen
[02:19:17] <SWPadnos> cool - what was it?
[02:19:20] <jmkasunich> need to test it some more
[02:19:46] <jmkasunich> an error in how I calculated some lookahead stuff... off by 1/2 sample period
[02:19:54] <SWPadnos> ah
[02:20:23] <SWPadnos> I've got a general question about how HAL hardware drivers operate
[02:20:24] <jmkasunich> (well actually, that's what was wrong with the version I was testing last night... which is completely different than the original code)
[02:20:33] <jmkasunich> shoot
[02:20:58] <SWPadnos> looking at parport, there's a read_all and a write_all function
[02:21:16] <SWPadnos> what causes these to be called?
[02:21:30] <jmkasunich> they get called if you add them to a thread
[02:21:33] <SWPadnos> can any HAL module ask for parport to write it's pins to the port?
[02:21:53] <jmkasunich> no, the system config determines what code runs when
[02:22:15] <SWPadnos> OK - so I should be looking at the .hal or the user of the parport for this info (like stepgen)
[02:22:16] <jmkasunich> hal modules never call other hal modules
[02:22:21] <jmkasunich> yes
[02:22:34] <jmkasunich> note that read_all is redundant
[02:22:44] <SWPadnos> That's what I was thinking - the pins are the only interface
[02:22:45] <jmkasunich> there is a numbered function for each port
[02:23:38] <SWPadnos> doesn't read_all actually load the hardware state into the halpin array, or do all read functions read the hardware?
[02:23:54] <SWPadnos> oh - or is the _all meant for all ports in the system?
[02:23:58] <jmkasunich> if you have two ports, you could put parport.0.read in a 50uS thread and use that port for step/dir or encoder sigs, and choose to put parport.1.read in a 1mS thread and use for generic I/O
[02:24:02] <SWPadnos> (ie, all 4 parallel ports)
[02:24:20] <SWPadnos> OK
[02:24:31] <jmkasunich> yeah, parport.read_all is equivalent to parport.0.read and parport.1.read, etc
[02:25:08] <SWPadnos> It was a little hard to see within the parport.c file, since nothing (other than setup / cleanup) actually _happens_ in that file :)
[02:25:09] <jmkasunich> need to define some conventions... the _all functions were provided to minimize the length of the hal file, but they aren't consistent
[02:25:25] <SWPadnos> it just exports functions to be added to threads
[02:25:37] <jmkasunich> right - callback type stuff
[02:25:50] <SWPadnos> OK - I was thinking in the context of a HAL classic ladder
[02:25:55] <SWPadnos> read all, think, write all
[02:26:19] <jmkasunich> no, parport.read_all simply reads all parports
[02:26:39] <jmkasunich> if you have other i/o, like stg or whatever, you need to add those read functs to a thread as well
[02:26:53] <SWPadnos> right - I wanted to see how to get the PLC read / operate / write model to work
[02:26:59] <SWPadnos> so this would be easy, if the ladder has exclusive access to a driver resource
[02:27:11] <SWPadnos> if the PLC shares pins with stepgen, there could be trouble
[02:27:20] <jmkasunich> so in the thread, you would have parport.X.read, stg.X.read, ladder.0.evaluate, parport.X.write
[02:27:22] <SWPadnos> not pins, ports - sorry
[02:27:28] <SWPadnos> exactly
[02:28:01] <SWPadnos> the issue would be defining which read functions need to be called - I guess this is a manual operation for the integrator, since it goes in the .hal file
[02:28:08] <jmkasunich> yes
[02:28:21] <SWPadnos> hmmm - that's not ideal :)
[02:28:36] <jmkasunich> actually I'm right in the middle of generating a really complex hal file right now, to test the stepgen
[02:28:44] <SWPadnos> how big is it?
[02:28:48] <jmkasunich> 18 hal modules
[02:28:59] <SWPadnos> hmmm - so about 6k or so?
[02:29:04] <jmkasunich> 30 signals
[02:29:22] <jmkasunich> the actual .hal file? I doubt it
[02:29:29] <fenn> you will look back on this and say "look at that primitive example file I wrote"
[02:29:40] <SWPadnos> by *hand*
[02:29:44] <jmkasunich> it's not an example, its a test
[02:29:52] <jmkasunich> I want to torture test the stepgen
[02:30:22] <jmkasunich> so I have two signal generators, muxes to select sine, triangle sawtooth or square waves from each
[02:30:26] <jmkasunich> summed together
[02:30:34] <SWPadnos> heh - that should be fun
[02:30:42] <jmkasunich> thru a differentiator and comparator to detect direction of motion
[02:30:59] <jmkasunich> the comparator output switches a mux to generate + or - backlash comp
[02:31:00] <SWPadnos> did you scale the outputs of the waveform generators to be out of bounds for stepgen?
[02:31:03] <jmkasunich> which is added in
[02:31:20] <jmkasunich> the result can be limited or not (set by a mux) and then goes to the stepgen
[02:31:40] <jmkasunich> and finally a sum block compares stepgen feedback to command to detect overshoot, etc
[02:31:57] <SWPadnos> heh - you know the drill: over range, under range, oscillating , step functions, impulse functions, etc. :)
[02:32:21] <jmkasunich> yeah... thats why the sine, triangle, square, etc
[02:32:42] <jmkasunich> hmmm, don't have anything that makes an impulse
[02:32:57] <anonimasu> * anonimasu is writing aux control stuff for a plc
[02:33:05] <SWPadnos> well - a single cycle square wave - can you change the duty cycle?
[02:33:17] <jmkasunich> tho adding two squares of slightly different frequency should do it
[02:33:30] <SWPadnos> (10000 unit period, 1 unit duty)
[02:33:39] <jmkasunich> nope... at least not in siggen v1.0, could add that later
[02:33:45] <SWPadnos> that'll give beats which will include impulses from time to time
[02:33:52] <jmkasunich> zactly
[02:33:57] <jmkasunich> that's why I did two
[02:34:01] <SWPadnos> heh
[02:34:16] <jmkasunich> like add a small fast square to a large slow sine or triangle
[02:34:46] <SWPadnos> yep
[02:34:52] <fenn> oh what fun
[02:35:01] <jmkasunich> shame that schematic based tool doesn't exist
[02:35:07] <SWPadnos> heh
[02:35:21] <jmkasunich> I sketched the diagram on paper, labeled everything, now I'm editing the hal file that builds it all
[02:35:29] <SWPadnos> This screams of a pared-down LabView
[02:35:36] <jmkasunich> yep
[02:35:46] <SWPadnos> One other question - about pins and signals
[02:35:54] <jmkasunich> hopefully I pared out the yucky parts of labview ;-)
[02:36:02] <SWPadnos> that would be most of it :)
[02:36:05] <fenn> jmkasunich: any thoughts on what to use to generate the schematic?
[02:36:18] <fenn> UML?
[02:36:30] <SWPadnos> a modified open source schematic editor (which doesn't exist yet)
[02:36:36] <jmkasunich> I was thinking gEDA (open source schematic editor)
[02:36:52] <jmkasunich> replace the ordinary parts libraries with special hal libraries
[02:37:06] <jmkasunich> generate a netlist using the geda netlister
[02:37:30] <jmkasunich> once you have a netlist, you are 50% of the way to a hal file
[02:37:37] <fenn> sounds like you've got a handle on it
[02:38:00] <jmkasunich> well I've been thinking about it for over a year, but haven't gotten past thinking
[02:38:09] <SWPadnos> handle != finished product - feel free to help ;)
[02:38:20] <SWPadnos> (and so will I)
[02:38:23] <jmkasunich> alex did look at geda, he wame away less than thrilled
[02:38:31] <jmkasunich> s/wame/came
[02:38:49] <SWPadnos> I recall that - there was talk of Eagle, but it's not open source
[02:38:51] <jmkasunich> if I was more GUI oriented I'd probably want to do an editor from scratch, but that is a TON of work
[02:39:07] <fenn> nah not really
[02:39:17] <fenn> because you know how it works by the time you're done with it :)
[02:39:25] <jmkasunich> in the dream world, you could click on signals in the picture and launch halmeters, and other such stuff
[02:39:35] <jmkasunich> that wouldn't work with geda
[02:39:46] <fenn> no, you would have to hook a halmeter into the circuit, right?
[02:40:04] <jmkasunich> halmeter isn't a normal component that you have to link in
[02:40:08] <fenn> oh
[02:40:20] <jmkasunich> it has a menu, and you pick the signal, pin, or param you want to viuew
[02:40:23] <jmkasunich> view
[02:40:34] <SWPadnos> is there a graphic front end to spice?
[02:40:39] <SWPadnos> (free, Linux)
[02:40:45] <jmkasunich> good question
[02:40:57] <jmkasunich> we have one a work, doze and very much not open
[02:41:15] <SWPadnos> yeah -I know those exist (like OrCad)
[02:41:16] <jmkasunich> that's a very interesting thought
[02:41:27] <anonimasu> you usually turn on coolant at the same time as you turn off the spindle isnt that right?
[02:41:29] <anonimasu> err turn on
[02:41:31] <anonimasu> the spindle
[02:41:47] <jmkasunich> depends on the g-code program, but I imagine so
[02:41:47] <fenn> sure
[02:42:09] <anonimasu> I am not hooking the spindle to the computer..
[02:42:25] <anonimasu> not until I am entirely sure emc dosent turn it on in a toolchange..
[02:42:28] <anonimasu> :D
[02:42:35] <SWPadnos> you also shouldn't hook the coolant control to the spindle control :)
[02:42:47] <SWPadnos> (unless there's a switch)
[02:42:58] <anonimasu> SWPadnos: yeah, ofcourse
[02:43:04] <SWPadnos> if that's what you were thinking
[02:43:14] <anonimasu> if ( coolant == 1 ) { }
[02:43:15] <anonimasu> ;)
[02:43:36] <SWPadnos> if (big manual switch is on) ...
[02:43:41] <anonimasu> heh
[02:43:45] <anonimasu> got a plc to handle it for me..
[02:43:59] <SWPadnos> OK - that's good
[02:44:10] <anonimasu> yeah, I trust it a bit more :)
[02:44:15] <jmkasunich> swp: if this sketch was in machine readable form, I'd send it to you along with the matching hal file, might be educational
[02:44:31] <jmkasunich> without the pic, the hal file is hard to read
[02:44:41] <SWPadnos> heh
[02:44:47] <SWPadnos> got a digital camera?
[02:45:05] <jmkasunich> yeah, but I can barely read my own scribbles
[02:45:10] <SWPadnos> look at ValarQ's software: http://arda.no-ip.org/crap
[02:45:17] <SWPadnos> it can't be as bad as my mother's writing
[02:45:22] <SWPadnos> (or mine, for that matter)
[02:46:06] <jmkasunich> craplication... I like it
[02:46:10] <SWPadnos> heh
[02:46:17] <jmkasunich> looks like board only tho, not schematic/netlist
[02:46:31] <SWPadnos> right, but it's a graphical editor written in python
[02:46:35] <SWPadnos> ...
[02:46:45] <SWPadnos> by sineone lurking on this very channel ;)
[02:46:48] <SWPadnos> someone
[02:46:52] <jmkasunich> right - if I did it, I'd have to use C, and I'd probably go insane
[02:47:09] <SWPadnos> well - perhaps ValarQ would be persuaded to help out a little :)
[02:47:12] <SWPadnos> could
[02:47:17] <SWPadnos> or would, I suppose
[02:47:22] <jmkasunich> halscope just about killed me, and it really needs to be revised and improved, parts are pretty crappy
[02:47:55] <SWPadnos> I wish I had a better handle on how these silly GUI toolkits work.
[02:48:05] <SWPadnos> and the designers
[02:48:12] <SWPadnos> like QT Designer and KDevelop
[02:48:16] <fenn> seems high level stuff is best left to high level languages, no?
[02:48:25] <jmkasunich> yeah
[02:48:31] <SWPadnos> It'as always a royal pain to add custom components to the libraries
[02:48:40] <jmkasunich> but when all you have is a hammer, every problem looks like a nail
[02:48:43] <jmkasunich> my hammer is C
[02:48:50] <SWPadnos> or "palettes" ans they're sometimes called
[02:48:54] <SWPadnos> as
[02:49:12] <SWPadnos> I've done that kind of user interface work on Windows, using Borland C++
[02:49:40] <SWPadnos> but I'm not sure I've ever gotten a custom component to install correctly on any system other athan my own
[02:50:29] <SWPadnos> Paul made a QT <-> emc library, which works great
[02:50:56] <SWPadnos> I was trying to make a generic "axis display" widget, that could be dropped into an application (or several, for multiple axes)
[02:51:09] <SWPadnos> it's a nightmare for a near-hardware guy like myself
[02:51:32] <jmkasunich> I want (wanted) to make some GTK widgets that would display the value of a hal pin
[02:51:38] <jmkasunich> (or manipulate a hal pin)
[02:51:45] <jmkasunich> need my round tuit
[02:51:52] <SWPadnos> too painful
[02:51:58] <jmkasunich> buttons, knobs, lights
[02:52:06] <SWPadnos> I have a friend with one - I'm sure you could borrow it :)
[02:53:10] <jmkasunich> linkps mux4.2.in0 compraw
[02:53:17] <jmkasunich> oops, wrong window
[02:53:25] <SWPadnos> I was going to say :)
[02:54:26] <SWPadnos> hmmm - so the fourth (or fifth - from 0) mux input 0 is getting connected to the comparator raw output, huh?
[02:54:30] <SWPadnos> what's the 2?
[02:54:41] <jmkasunich> close, but no cigar
[02:54:44] <SWPadnos> heh
[02:54:48] <jmkasunich> mux4 is a 4 input mux
[02:54:58] <SWPadnos> OK - the third 4-input mux, first input
[02:54:59] <jmkasunich> .2. is the second instance of a 4 input mux
[02:55:06] <SWPadnos> not from 0?
[02:55:13] <jmkasunich> right, third, not second, my bad
[02:55:39] <SWPadnos> OK - got it. missed the fact that the 2 is after a decimal, and the 4 before
[02:55:42] <jmkasunich> its first input (in0) is being connected to the raw backlash comp signal
[02:55:59] <jmkasunich> other inputs are connected to ramped and s-curved versions of the backlash
[02:56:11] <SWPadnos> well - the signal name was just a guess - that could be swahili (with the right locale)
[02:56:25] <jmkasunich> so the compensation can be applied as a step, ramp, or ramp with accel and decel
[02:56:27] <SWPadnos> be in Swahili ...
[02:56:31] <SWPadnos> cool
[02:57:03] <jmkasunich> of course in emc proper, all that crap is hardcoded, not a whole boatload of hal modules
[02:57:16] <SWPadnos> does the mux change "immediately", or after a "propagate" function?
[02:57:36] <jmkasunich> the mux has only one function
[02:57:41] <SWPadnos> nevermind - there can only be one writer to the selector bits
[02:57:54] <jmkasunich> it reads the select input, copies the appropriate input to output
[02:58:31] <jmkasunich> in this case there are no writers to the select bits, I'll tickle them with halcmd sets
[02:59:24] <SWPadnos> hmmm - so something connected to the mux output relies on the update() function being called befoe, to get the output signal in the correct state?
[03:00:20] <jmkasunich> the mux's function gets called before the functions associated with downstream blocks
[03:00:36] <jmkasunich> and after functions associated with upstream blocks
[03:00:54] <SWPadnos> automagically, or because it's configured to be that way?
[03:01:09] <jmkasunich> because I did "addf" commands in the proper order
[03:01:20] <SWPadnos> OK - configured for proper operation :)
[03:01:31] <jmkasunich> someday I'd like to have code that evaluates the netlist and works out the correct order
[03:01:51] <fenn> if you ever get the schematic gui to work, make sure it displays the values of each signal so you don't have to fire up a halmeter for each one
[03:01:58] <SWPadnos> that's a LabView "feature" that I find most annoying, but in this application, it would be a good thing
[03:02:17] <SWPadnos> fenn: that's not a good idea
[03:02:25] <jmkasunich> I wouldnt want it to cultter the screen with every signal value
[03:02:25] <fenn> you say that no matter what i say
[03:02:28] <SWPadnos> it's better to select which signals you want to look at
[03:02:30] <SWPadnos> heh
[03:02:37] <SWPadnos> sorry - I'm not trying to be an ass
[03:02:47] <jmkasunich> but click on a signal, and have a little box that displays the value
[03:03:04] <fenn> well, a preferences option at least then
[03:03:14] <SWPadnos> there are functions that run 10k times per second or faster - it would be a heck of a load to keep up with a lot of signals at a high rate
[03:03:27] <SWPadnos> and there's limited memory for that
[03:03:30] <fenn> i know, you update the signals at the gui's rate
[03:03:37] <jmkasunich> well any GUI isn't gonna be able to keep up with RT, guis are user space
[03:04:05] <SWPadnos> rght, but a scope should capture at full speed, and then draw the information later, if the GUI can't keep up
[03:04:07] <jmkasunich> halmeter updates either 10 or 5 times per sec (if linux lets it)
[03:04:08] <SWPadnos> right
[03:04:25] <jmkasunich> yeah, that's what scope does
[03:04:49] <jmkasunich> scope can capture 64K samples and display the waveforms several times per second
[03:05:15] <jmkasunich> I take that back, 16K samples, 64K bytes
[03:05:15] <SWPadnos> Modern PCs can do a lot - look at the new digital scopes and you'll see that.
[03:05:36] <jmkasunich> 1 chan 16K, 2 chan 8K, 4 chan 4K, 8 chan 2K, or 16 chan 1K samples
[03:05:58] <SWPadnos> (those new 6000 series scopes from Agilent can draw 30 million (!) vectors per second)
[03:06:07] <fenn> egad
[03:06:20] <SWPadnos> that's vectors too, in color
[03:10:19] <jmkasunich> most of this HAL GUI stuff is pretty far down the road
[03:10:37] <jmkasunich> one thing I'd like to do sooner rather than later is the ability to save setups from halscope
[03:11:13] <SWPadnos> that's a good idea
[03:11:30] <jmkasunich> you spend a couple minutes setting signal sources for 5 or 6 channels, setting gains, setting trigger source, level and position, and then if you want to repeat the test you have to do it all again
[03:12:06] <jmkasunich> I'd also like to be able to write the waveform data to a file for further processing
[03:13:25] <SWPadnos> the load / save stuff would go in scope.c, right? (vs. scope_rt, etc.)
[03:13:48] <jmkasunich> yes
[03:14:13] <jmkasunich> scope_rt is unlikely to need any significant changes... all it does is capture data
[03:14:50] <jmkasunich> might not go in scope.c, there are several files, scope_horiz, scope_vert, etc
[03:14:58] <SWPadnos> does scope_usr.h have the structs for gain, offset, etc., or are those in the RT header
[03:15:02] <jmkasunich> might add another one for general menus stuff like saging, etc
[03:15:12] <jmkasunich> * jmkasunich looks (it's been a while)
[03:15:18] <SWPadnos> usr - nevermind
[03:15:22] <SWPadnos> scope_chan_t
[03:16:03] <SWPadnos> and scope_vert_t and scope_horiz_t
[03:16:29] <SWPadnos> wait - no - those last are for the gtk widgets
[03:16:48] <jmkasunich> its a mix
[03:17:03] <jmkasunich> scope_horiz_t has stuff associated with horizintal
[03:17:22] <jmkasunich> some is wigits, some is things like scaling (which may be controlled by widgets)
[03:17:23] <SWPadnos> it looks like it's only display-related stuff
[03:17:29] <SWPadnos> pixels_per_sample and the like
[03:17:53] <SWPadnos> the scaling is screen scale - it should be recalculated if the window is resized (if it supports that)
[03:18:01] <jmkasunich> it does
[03:18:15] <SWPadnos> not periods_per_sample - that would be in the RT cod
[03:18:20] <SWPadnos> code
[03:18:24] <jmkasunich> right
[03:18:37] <jmkasunich> set from user space, used in realtime
[03:18:42] <SWPadnos> so you load the "real" info, and call "resize"
[03:18:47] <SWPadnos> right
[03:19:17] <jmkasunich> there are a lot of things in scope that are poorly designed
[03:19:20] <jmkasunich> I was learning as I went
[03:19:21] <SWPadnos> heh
[03:19:32] <SWPadnos> it works, and that's an achievement
[03:19:55] <SWPadnos> especially with a GUI widget set, and no RAD design tool
[03:20:19] <fenn> what's _t stand for?
[03:20:22] <SWPadnos> type
[03:20:31] <jmkasunich> typedef names
[03:20:43] <jmkasunich> I think I'm mis-using that syntax
[03:20:58] <jmkasunich> think of things like size_t
[03:21:09] <fenn> this is like bool or int?
[03:21:24] <jmkasunich> those are scalars though, and here I use the _t to indicate a typedef'ed name for a structure
[03:21:27] <SWPadnos> yes, but for more complex data structures, in this case
[03:21:30] <fenn> right
[03:21:32] <SWPadnos> (these cases)
[03:21:41] <SWPadnos> but the hal types are scalars
[03:21:47] <SWPadnos> hal_bit_t
[03:21:49] <SWPadnos> etc.
[03:22:11] <jmkasunich> the hal types are, but I used _t for a lot of structs, not just in halscope but in the emc2 motion controller too
[03:22:20] <SWPadnos> yep
[03:22:22] <jmkasunich> paul was very unhappy with me for that
[03:22:35] <SWPadnos> heh - so he was :)
[03:22:52] <jmkasunich> I hate using struct tags tho
[03:22:55] <SWPadnos> I wonder if the style guide will say "and no _t suffixes, ever!"
[03:23:22] <jmkasunich> I don't want to declare pointer to a struct as "struct foo *fooptr;"
[03:23:26] <SWPadnos> well - you can do a typedef, but just leave off the _t
[03:23:36] <jmkasunich> I just want to write "foo_t *fooptr;"
[03:24:02] <jmkasunich> but in many cases, I declare things like "foo_t *foo;"
[03:24:03] <SWPadnos> I think those go back to the days when a struct, typedef, function and variable might be named the same thing
[03:24:08] <jmkasunich> then use "foo->member"
[03:24:32] <SWPadnos> you can use the same name (but that might get you shot)
[03:24:55] <jmkasunich> you can say "typedef struct { stuff } foo;"
[03:24:57] <SWPadnos> there are separate namespaces in C - certain things can have the same name without a collosion
[03:25:06] <jmkasunich> and then "foo *foo;" ?
[03:25:17] <SWPadnos> (I'd have to look up which things)
[03:25:57] <jmkasunich> I think I'd rather use struct tags than overload the names
[03:26:06] <jmkasunich> but I prefer the _t convention
[03:26:07] <SWPadnos> I could also be thinking of C++, but I'm pretty sure that there were separate spaces for types and user defined identifiers (which includes function names)
[03:26:28] <SWPadnos> overloaded names suck, and I should be shot for mentioning them o_O
[03:27:25] <jmkasunich> I suppose I could call the type "foo_s", s for struct ;-)
[03:27:35] <jmkasunich> paul would like that about as much as he likes _t
[03:27:36] <SWPadnos> at least in C - in C++, operator overloading is the best thiing, ever.
[03:27:40] <SWPadnos> heh
[03:27:54] <SWPadnos> maybe he prefers the letter s - do a checkin to CVS and find out ;)
[03:28:04] <jmkasunich> or not
[03:28:14] <SWPadnos> (107 files modified, comment "changed _t to _s")
[03:29:10] <jmkasunich> next day: (107 files modified, comment "reverted stupid change, bloody yanks are daft")
[03:29:16] <SWPadnos> heh
[03:29:40] <SWPadnos> and jmk CVS account deleted :)
[03:37:07] <fenn> is g++ same thing as gcc-c++?
[03:38:22] <jmkasunich> I think so
[03:43:18] <fenn> just wondering, why do you suggest downloading gcc-2.95?
[03:43:23] <fenn> to compile emc2
[03:43:53] <jmkasunich> RT code (and kernels, and kernel modules in general) are sensitive to the compiler that is used
[03:44:12] <jmkasunich> you really don't want to compile modules with a compiler different than the one used to compile your kernel
[03:44:35] <fenn> but gcc is up to 4.0 now.. why is it still so old?
[03:46:25] <jmkasunich> kernels are important... don't want to compile them with a bleeding edge compiler
[03:47:39] <fenn> sorry.. it's just that it's conflicting with a bunch of stuff on my system
[03:48:22] <SWPadnos> what's conflicting?
[03:48:45] <fenn> gcc, glibc-devel
[03:49:05] <SWPadnos> ah - what distro?
[03:49:30] <fenn> fc3
[03:49:42] <SWPadnos> hmm
[03:49:50] <jmkasunich> do you have RT installed on that? (patches, etc)?
[03:49:57] <fenn> no not yet
[03:50:06] <jmkasunich> so you have a stock redhat kernel
[03:50:42] <jmkasunich> the RTAI folks recommend using virgin kernel source from kernel.org, not the RH kernel since it may have conflicting patches
[03:50:43] <fenn> i need to download kernel sources, patches, compile patched kernel, _then_ i can compile emc2
[03:50:53] <jmkasunich> yep
[03:50:59] <jmkasunich> that's why I like the BDI
[03:51:18] <fenn> well, i don't wanna trash my main computer just to compile emc2
[03:51:29] <fenn> and dont wanna do devel work on a crappy dumpster cmputer
[03:51:45] <fenn> * fenn whines
[03:51:51] <jmkasunich> you can build a kernel without replacing your original one
[03:51:56] <fenn> i know
[03:52:11] <jmkasunich> I'm curious... does your box have kgcc or egcs?
[03:52:26] <fenn> no.. never heard of them
[03:52:45] <fenn> assuming kgcc is kernel compiler
[03:52:55] <jmkasunich> they are usually symlinks to the compiler that was used to build your kernel
[03:53:13] <jmkasunich> so gcc might get you version 4.xx
[03:53:27] <jmkasunich> but kgcc will get you something older, that the kernel folks trust
[03:53:40] <fenn> if i compile the new kernel and emc2 with gcc 3.42 will it work?
[03:53:48] <jmkasunich> no idea
[03:53:56] <fenn> so there's a chance it will :)
[03:54:12] <jmkasunich> if you are compiling a kernel with rtai patches, I'd ask questions on the rtai list (or ask paul)
[03:54:21] <fenn> right
[03:54:31] <jmkasunich> the rtai folks probably know which gcc versions build a good realtime kernel
[03:55:08] <jmkasunich> then whatever compiler you use for rtai, use the same for emc2
[03:55:27] <jmkasunich> that part is probably automatic... the newer versions of rtai install a script called rtai-config
[03:55:39] <jmkasunich> you can "ask" it all kind of things about your rtai install
[03:55:48] <jmkasunich> compiler version, compile flags, lib paths, etc
[03:56:08] <jmkasunich> emc2's config script gets the compiler version from rtai-config, if present
[03:56:39] <jmkasunich> crap-- almost midnight again
[03:56:49] <jmkasunich> I talk too much
[03:56:55] <fenn> i eat too much :)
[03:57:48] <fenn> well i'm gonna try it out with gcc 3.4.2
[03:58:22] <A-L-P-H-A> fenn, I just had a full stack of ribs, garlic mash potatoes /w gravy, and grilled veggies.
[03:58:50] <fenn> hoo boy.. i've been eating nothing but raw food for the last two weeks
[03:59:14] <A-L-P-H-A> cool ribs are good.
[03:59:16] <A-L-P-H-A> raw... bah
[03:59:19] <A-L-P-H-A> what are you? Japanese?
[03:59:42] <fenn> i'm turning japanese i think i'm turning japanese i really think so
[04:00:05] <A-L-P-H-A> uhuh
[04:00:37] <A-L-P-H-A> heh... I was watching David Chapelle... know who he was?
[04:00:44] <A-L-P-H-A> was=is
[04:00:47] <fenn> some farmer from arkansas
[04:00:56] <A-L-P-H-A> Black Comedian.
[04:00:57] <fenn> and sidelines as a comedian
[04:01:03] <SWPadnos> see ya all later
[04:01:06] <A-L-P-H-A> later
[04:01:07] <fenn> g'nite
[04:01:10] <SWPadnos> SWPadnos is now known as SWP_Away
[04:01:25] <A-L-P-H-A> dehumidifier is going on now.
[04:01:28] <A-L-P-H-A> back in a bi
[04:01:30] <A-L-P-H-A> t
[04:02:28] <jmkasunich> time for me to crash
[04:14:35] <A-L-P-H-A> http://www.craftsmanshipmuseum.com/images/Harley%20Cylinders.JPG
[04:20:16] <fenn> whatever floats your boat
[04:21:03] <fenn> personally i like the miniature milling machines
[04:22:39] <fenn> http://www.craftsmanshipmuseum.com/images/Jordan06.JPG
[04:25:56] <Jymmm> fenn is the mill operational?
[04:26:15] <fenn> yeah it uses a mini dc motor
[04:26:43] <Jymmm> then that's cool - I'm not into models
[04:27:38] <fenn> working models are cool
[04:27:43] <fenn> knicknacks are not
[04:27:49] <Jymmm> yeah
[04:28:27] <Jymmm> doesn't have to be accurate, just operational
[04:28:41] <fenn> http://www.barryjordan.freeola.com/bbridge.htm
[04:29:30] <fenn> wonder what he uses for v-belts
[04:30:03] <Jymmm> braided hairs
[04:32:12] <A-L-P-H-A> heh
[04:32:16] <A-L-P-H-A> that's cool.
[04:34:46] <Jymmm> engraving http://www.craftsmanshipmuseum.com/images/Lindsay03.jpg
[04:35:06] <A-L-P-H-A> I'm not a huge fan of engraving. I like simple, clean stuff.
[04:43:20] <fenn> simple, clean, and totally ridiculous: http://homepage.eircom.net/~skycam/C-17A_Globemaster_III/
[04:47:27] <A-L-P-H-A> holy shit. and it works?
[04:48:08] <A-L-P-H-A> I'm off to bed. at least I think I will.
[06:07:03] <A-L-P-H-A> Time: 02:05:51 -0500 GMT, Windows XP Professional, Service Pack 2 (5.1 - 2600), AMD XP 2800+ running at (1-AMD , 2123MHz, 512KB (0% Load)), DDR400 RAM Usage: 684/1024MB (66.80%), C: 41.42gb of 75.42gb free, D: 0gb of 0gb free, N: 24gb of 372.62gb free, Current Uptime: 8hrs 34mins 13secs, Record Uptime: 3wks 12hrs 48mins 50secs, 3 Samsung 19" flat monitors (1 Trinitron, 2 CRT).
[06:10:12] <A-L-P-H-A> hey aj.
[06:10:25] <alex_joni> morning A-L-P-H-A
[06:11:06] <A-L-P-H-A> what time is it there?
[06:11:14] <alex_joni> morning
[06:11:14] <A-L-P-H-A> +3gmt?
[06:11:18] <A-L-P-H-A> 10am?
[06:11:19] <alex_joni> [09:09] <alex_joni> morning
[06:11:22] <alex_joni> +2 gmt
[06:11:24] <A-L-P-H-A> +2 gmt
[06:11:26] <A-L-P-H-A> k
[06:11:32] <alex_joni> yup
[06:11:53] <A-L-P-H-A> frack... gotta do so much this week.
[06:12:23] <A-L-P-H-A> repoint a satellite dish. car shop. finish machining those brackets. pay my tuition. register my other car.
[06:12:25] <A-L-P-H-A> :(
[06:16:31] <fenn> that's not very much for a whole week
[06:17:42] <fenn> alex_joni i'm patching a kernel with rtai, but rtai doesn't say which kernel version to use
[06:17:51] <fenn> i'd assume it matters a whole bunch
[06:17:51] <alex_joni> fenn: go with 2.6.10
[06:17:59] <alex_joni> if you want 2.6
[06:18:02] <alex_joni> or any 2.4.2x
[06:18:07] <fenn> crap i'm halfway through compiling 2.6.11
[06:18:08] <alex_joni> 25 maybe
[06:18:16] <alex_joni> 2.6.11 might be a bit much
[06:18:24] <alex_joni> not really sure rtai has a patch for 2.6.11
[06:18:36] <fenn> documentation on the website is non-existant
[06:18:48] <fenn> and the mail archives are down
[06:19:31] <alex_joni> what rtai did you get?
[06:19:45] <fenn> 3.2
[06:19:53] <fenn> "vulcano" or some such
[06:20:32] <alex_joni> magma?
[06:20:48] <alex_joni> or volcano ?
[06:20:53] <alex_joni> err vulcano
[06:21:09] <alex_joni> http://cvs.gna.org/viewcvs/rtai/vulcano/base/arch/i386/patches/
[06:21:20] <alex_joni> doesn't look like there's a 2.6.11 patch yet :(
[06:21:59] <alex_joni> you also need an adeos patch
[06:22:01] <fenn> oh, duh
[06:22:01] <alex_joni> http://download.gna.org/adeos/patches/v2.6/i386/
[06:22:14] <fenn> i forgot that rtai is its own OS, and linux runs on top of it
[06:22:24] <alex_joni> smthg like that
[06:22:31] <alex_joni> but.. beware of http://download.gna.org/adeos/patches/v2.6/i386/adeos-linux-2.6.11-i386-r9-dont-use.patch
[06:22:38] <fenn> heh
[06:22:40] <alex_joni> the dont-use is in there for a reason
[06:22:40] <alex_joni> :D
[06:23:17] <fenn> guess i'll kill this compile
[06:23:39] <alex_joni> fenn: you'll want to compile after the patching
[06:25:04] <fenn> i know, i was making sure it would compile alright before messing it up with patches
[06:25:19] <fenn> i'm using the redhat kernel right now
[06:25:33] <alex_joni> scrap that ;)
[06:25:40] <alex_joni> you can't patch a RH kernel
[06:25:48] <alex_joni> * alex_joni will be back later
[06:25:53] <alex_joni> alex_joni is now known as alex_joni_away
[06:26:03] <fenn> what i mean is, i'm _running_ the RH kernel, and am patching the one from kernel.org
[06:57:45] <Jymmm> G'Night Folks!
[07:36:55] <alex_joni_away> alex_joni_away is now known as alex_joni
[09:40:55] <jacky^> morning
[09:41:07] <jacky^> hi alex_joni
[10:56:39] <jacky^> hi
[11:01:45] <ValarQ> 'lo
[11:15:08] <jacky^> bye
[12:04:22] <cradek> hi jepler
[12:06:56] <ValarQ> hello cradek, jepler and you other guys :)
[12:11:01] <jepler> good morning
[13:57:11] <steves_logging> steves_logging is now known as steve_stallings
[13:58:08] <rayh> morning steve
[13:58:42] <steve_stallings> morning Ray
[14:00:17] <rayh> I couldn't find the big gecko stuff on their site. Is it ready for sale?
[14:01:59] <steve_stallings> The G202? There were 150 in the first batch and I did order one by phone. Think he decided not to list on regular web page until next batch was ready. He did announce on Geckodrive list.
[14:02:43] <steve_stallings> Got my G202 quickly, but have only used it as mechanical model to test my PMDX-131.
[14:03:18] <rayh> Okay. I spec'd your 131 and some gecko drives for a project
[14:03:36] <ValarQ> hello Padnos
[14:03:49] <rayh> but the motors are vexta PH296 NEMA34
[14:05:01] <rayh> I'm not certain how good the 210 are for 1.8v at 4.5 amp.
[14:10:31] <CIA-8> 03paul_c * 10emc2/src/doxconfig: Added some todo tags - Need to check that these are removed *when* the features are implemented.
[14:10:46] <CIA-8> 03paul_c * 10emc2/src/emc/rs274ngc/interp_convert.cc: Added some todo tags - Need to check that these are removed *when* the features are implemented.
[14:12:03] <steve_stallings> I would not expect any problems, but you will need to heatsink the G210.
[14:23:49] <jacky^> hi all
[14:25:06] <steve_stallings> morning
[14:25:26] <jacky^> hi steve_stallings !
[14:26:41] <jacky^> steve_stallings: my compliments for the great work you made :)
[14:26:54] <jacky^> cool
[14:27:29] <steve_stallings> thanks, I think, duh.. what did I do?
[14:28:03] <jacky^> i'm a very newbie to emc, from italy (sorry my english is too bad :\)
[14:28:38] <jacky^> i've seen some pictures on linuxcnc
[14:28:44] <jacky^> very nice
[14:29:47] <steve_stallings> OK, but I am just the webmaster, the machines were built by others
[14:30:04] <jacky^> i seen, great work
[14:35:30] <rayh> Thanks Steve.
[14:35:41] <jacky^> hi rayh
[14:36:20] <jacky^> thanks too you :)
[14:38:03] <jacky^> i'm playng around sherline mini, it seem to me very nice
[14:39:23] <jacky^> the only thing i don't like too much is synergy :\
[14:39:46] <jacky^> it seem to me very diffcult to use
[14:40:15] <jacky^> infact, i leave it ..
[14:40:19] <jacky^> :\
[14:42:00] <jacky^> hi Phydbleep , are you ?
[14:43:26] <jacky^> my home switches wont work
[14:51:16] <jepler> cradek: if you were cutting PCBs only, would you be better off with the original maxnc spindle or with your current sherline spindle setup? I assume that once you have the second pulley, the sherline will be better in all cases.
[14:51:36] <cradek> that's right
[14:51:52] <cradek> the other one had slightly higher speed (~ 10k vs ~ 8k) with the stock pulleys
[14:52:08] <cradek> for PCBs you want the spindle as fast as possible.
[14:52:28] <jepler> but you hope to attain >20k when you improve the pulley
[14:52:44] <cradek> yeah, I should get 18-20k
[14:53:14] <cradek> 25k would be ideal but you start to need special bearings.
[15:22:33] <websys> dan - yes I log this so I can catch up on things
[15:23:25] <websys> Amusing that you mentioned smart-cam - the guy that's taking it over used to work here
[15:29:59] <dan_falck> websys hi
[15:30:07] <websys> good morning
[15:30:26] <dan_falck> it looks like a lot of people try synergy, but get confused and give up
[15:30:37] <SWP_Away> SWP_Away is now known as SWPadnos
[15:30:59] <SWPadnos> re: synergy - that's me :)
[15:31:14] <SWPadnos> though I didn't try very hard
[15:31:23] <dan_falck> I've got to admit, I have a certain way of checking out CAD/CAM software- take it for a drive
[15:31:39] <websys> I know - it's hard to give 30 years of (necessary) options available to keep everyone happy and still keep it simple
[15:31:55] <dan_falck> and I couldn't get anywhere. But of course I was too lazy to read the manual
[15:32:04] <SWPadnos> heh - I think the key to simplicity is "discoverability"
[15:32:09] <dan_falck> so, I will give it another try soon
[15:32:21] <websys> I think simplicity is inversely proportional to capability
[15:32:26] <dan_falck> true
[15:32:26] <SWPadnos> you don't have to know how to use the software if you can find out from within the software
[15:32:40] <SWPadnos> often true, I agree
[15:33:10] <websys> We've tried simple interfaces but then someone always complains that it doesn't have a required "feature"
[15:33:57] <websys> Trying to set up a switchable interface that starts simple and allows user to go to different levels of expertise
[15:34:33] <SWPadnos> The CadMax interface has a good feature set, including a command line for advanced use
[15:34:46] <SWPadnos> use the menus if you like, type in values if you like
[15:34:57] <websys> On one end we have autocad type hole poking and on the other a parametric solid modeler that does 5-axis work
[15:35:10] <SWPadnos> (what would be really great would be tab-completion in that command line)
[15:35:41] <websys> Underneath the interface is a command line/script based language
[15:35:56] <SWPadnos> right - similar to Autocad
[15:36:05] <SWPadnos> (and every other 30-year old CAD program :) )
[15:36:16] <websys> Technically you could run Synergy from the command line/ no gui
[15:36:32] <SWPadnos> harder to visualize parts, I'd bet :)
[15:36:43] <websys> Try impossibler
[15:36:57] <SWPadnos> AsciiArt to the rescue
[15:37:04] <websys> But if you setup a family of parts....
[15:37:36] <SWPadnos> does Synergy have a macro capability?
[15:37:41] <websys> Yes
[15:38:00] <dan_falck> nice
[15:38:25] <SWPadnos> cool - so can you, for instance, make a macro that does shoulder bolt holes, and have parameters for the diameter and depth of the shoulder and hole?
[15:38:25] <websys> Actually runs a type of basic interpreter
[15:38:40] <SWPadnos> (and X,Y, and orientation)
[15:38:40] <websys> Alreay there
[15:38:44] <SWPadnos> cool
[15:38:51] <SWPadnos> I should look again
[15:38:58] <dan_falck> I will too
[15:39:24] <SWPadnos> I think the thing that gets me is the way the GUI works
[15:39:25] <SWPadnos> this is something that happens with a lot of UNIX programs
[15:39:26] <websys> Need feedback on how to simplify without removing capability
[15:39:39] <SWPadnos> especially ones that are designed as command line, and have an optional GUI
[15:39:54] <dan_falck> dumb question, but is the manual online?
[15:40:00] <websys> People don't want to learn languages anymore - just point and click everything
[15:40:02] <SWPadnos> (this isn't specififc to Synergy - I'll look at it again and send feedback)
[15:40:19] <SWPadnos> well - the right-click context menu is pretty ubiquitous these days
[15:40:26] <websys> and hopw the developer gives you what you need to point/click at
[15:40:41] <SWPadnos> also shift-click to extend selection, drag box to multi-select, etc.
[15:41:11] <SWPadnos> I know it's a royal PITA to make programs that act all touchy-feely Mac-ish
[15:41:26] <websys> I made the right-click context menu user-configurable so the user controls what it does
[15:41:29] <SWPadnos> but it is expected these days, especially for visual things, like 3-d modelling
[15:41:32] <SWPadnos> cool
[15:41:50] <SWPadnos> Hi Paul
[15:42:00] <websys> also allow the user to make their own interface but nobody does it
[15:42:05] <paul_c> Afternoon
[15:42:15] <SWPadnos> yeah - we're too lazy (at least I am :) )
[15:42:21] <paul_c> Ah... Bob...
[15:42:34] <websys> more like - don't want to take the time to learn
[15:42:41] <SWPadnos> well - that too
[15:42:47] <websys> which I can't blame anyone
[15:43:05] <websys> but from my end it's hard to know what people want
[15:43:14] <SWPadnos> for someone like me, where mechanical CAD is only a part of what I do, it isn't worth the time (unfortunately)
[15:43:46] <websys> wish I could get an open-source group to put together interfaces
[15:44:36] <dan_falck> Bob, what's base price of 3 axis CAM?
[15:44:45] <paul_c> Does Synergy support Autodesk Inventor imports ?
[15:45:20] <websys> 3 axis simultaneous or 2-axis with 3rd positioning?
[15:45:32] <dan_falck> how about 2 axis
[15:45:38] <websys> $295
[15:45:54] <dan_falck> how much for 3 axis simultaneous?
[15:46:08] <websys> Can't do Inventor because it's based on ACIS
[15:46:32] <websys> $795 or $1250 for complete solid modeling
[15:46:58] <dan_falck> that's pretty reasonable
[15:47:02] <websys> We're based in Parasolids ( i'e' SolidWorks, SolidEdge )
[15:47:24] <SWPadnos> cool - I can transfer from CadMax when I switch entirely to Linux
[15:48:48] <websys> CadMax = Solids?
[15:48:59] <SWPadnos> yep - parametric solids
[15:49:04] <SWPadnos> $295
[15:49:24] <SWPadnos> also based on the parasolid kernel
[15:49:42] <websys> May have to adjust the price then
[15:49:53] <SWPadnos> h - sorry to be the bringer of bad new :)
[15:49:57] <SWPadnos> heh...
[15:50:07] <SWPadnos> does Synergy import/export IGES?
[15:50:11] <websys> Yes
[15:50:18] <SWPadnos> ah - then don't worry about the price
[15:50:27] <SWPadnos> CadMax charges an extra $500 for that
[15:50:33] <websys> inport/export DXF/IGES/Parasolid
[15:50:57] <SWPadnos> plus there's no Linux version, and none planned (when I last spoke with them)
[15:51:39] <websys> Working on the 64 bit version of solids now
[15:51:43] <websys> in Linux
[15:52:03] <SWPadnos> *very* cool - is it multithreaded?
[15:52:53] <websys> Well sort of
[15:53:08] <SWPadnos> heh - modeller and GUI are separate, but otherwise not?
[15:54:05] <websys> It's uses worker threads but doesn't spin of the calcs when possible
[15:54:30] <SWPadnos> ah
[15:54:57] <SWPadnos> does it have sheet metal (bending/flattening) capabilities?
[15:56:21] <websys> Used to but we gave up on that when other packages started giving it away
[15:56:29] <SWPadnos> ah
[15:56:48] <SWPadnos> that's the main (capability) difference between CadMax and SolidWorks
[15:56:53] <paul_c> methinks I've found a file limit for Synergy...
[15:56:59] <websys> ?
[15:57:08] <paul_c> 116247K dxf
[15:57:12] <SWPadnos> there are a lot of UI differences as well, but I think that's the main feature that SW has
[15:57:24] <SWPadnos> which CM doesn't
[15:58:07] <SWPadnos> try exporting arcs as arcs instead of lines :)
[15:58:10] <websys> Paul - can I ftp that file or can you email it to me?
[15:59:21] <paul_c> I'll zip it up for you.
[16:01:19] <websys> Thanks
[16:05:38] <paul_c> Even with bzip2, it's nearly 15Megs..
[16:06:08] <websys> well it'll be interesting
[16:27:49] <fenn> yo, anyone know if there's any special config options for compiling an rtai kernel?
[16:29:30] <fenn> pipeline profiling?
[16:29:39] <fenn> ACPI/APM off>
[16:29:41] <paul_c> As long as you avoid APM and ACPI, turn off module versioning...
[16:31:14] <paul_c> CONFIG_REGPARM must be "not set"
[16:31:37] <paul_c> and CONFIG_ADEOS either m or y
[16:32:50] <fenn> hmmm.. where's CONFIG_REGPARM in xconfig?
[16:33:58] <paul_c> Firmware drivers
[16:34:32] <robin_sz> she was only the wicket-keepers daughter ...
[16:35:33] <fenn> well, i don't see it, so i'll assume it's off
[16:52:27] <stevestallings2> Hey Les... you mentioned the Kmotion web site having EMC derived software source code. Where?
[16:53:00] <stevestallings2> stevestallings2 is now known as steve_stallings
[17:19:45] <les> steve are you still there?
[17:21:06] <les> source code is at Http://dynomotion.com/Software/Download.html
[17:21:25] <les> I think it needs a doze box to install on
[17:21:39] <les> or I can shoot you some source code
[17:22:36] <les> the TP is interesting to look at.
[17:30:27] <SWPadnos> les: did you have a chance to read those papers from de Boor?
[17:34:48] <SWPadnos> Hiya, Martin
[17:34:55] <Imperator_> Hey Stephen
[17:35:01] <SWPadnos> How are things in Deutschland?
[17:35:18] <Imperator_> dark weather at teh moment
[17:35:40] <SWPadnos> hmmm - similar here - hot, but 60% chance of thunderstorms
[17:35:48] <Imperator_> and our PCB for the Servo card is nearly finished
[17:38:40] <SWPadnos> excellent!
[17:39:21] <SWPadnos> I may be involved in a high-speed FPGA project soon, so I will gain some experience that could be applied to a PCI interface
[17:40:40] <les> I read some of the deboor stuff
[17:40:50] <SWPadnos> heh - how'd you like it?
[17:40:57] <les> heh
[17:41:02] <les> ok I guess...
[17:41:45] <SWPadnos> it looked like some of the concepts could be applied to aggregating many line segments into a spline, and getting measures of how well the spline path follows the G-code path
[17:42:03] <SWPadnos> but the math is pretty hairy stuff, and most of it would have to be thrown away for EMC
[17:42:10] <les> some could yes
[17:46:40] <steve_stallings> Les, thanks, got it. Interesting that it is not linked on web page. It is not a compressed file, but rather an executable. Did you just run it to install?
[17:49:34] <Imperator_> SWPadnos: a PCI interface would be great !!! Our Card is based on a FPGA, so switching from ISA to PCI should be easy, if somebody has a ready to use PCI interface
[17:49:51] <SWPadnos> there are some on http://www.opencores.org
[17:50:02] <SWPadnos> but they interface to the wishbone bus, I believe
[17:50:24] <Imperator_> jep, but i havent seen a PCB example
[17:50:39] <SWPadnos> no - and there are pretty strick trace layout rules for PCI
[17:50:43] <SWPadnos> strict, that is
[17:51:06] <Imperator_> on www.fpga4fun.com is al easy interface
[17:51:33] <Imperator_> how do you mean that ??
[17:51:55] <SWPadnos> the traces have to be the same length, the impedance must be controlled, etc.
[17:52:01] <SWPadnos> ISA is much easier to lay out for
[17:52:46] <Imperator_> yes, thats why i wounder why they have no PCB example for there core
[17:53:07] <SWPadnos> well - it's open *cores* .org, not openPCBdesigns.org :)
[17:53:16] <Imperator_> :-)
[17:54:07] <Imperator_> i think they use a pci development board that you can buy somewhere, but that is normaly also not that cheap
[17:54:21] <SWPadnos> no - it's usually quite expensive
[17:54:36] <SWPadnos> there are analog and digital PCI I/O boards available for not much money though
[17:54:56] <Imperator_> have one from www.kolter.de they have a pci interface in a CPLD and it is very cheap (150 EUR)
[17:55:06] <SWPadnos> that's pretty good
[17:55:12] <Imperator_> hey Alex
[17:55:18] <SWPadnos> Hi, Alex.
[17:55:21] <alex_joni> hell´┐Ż
[17:55:32] <alex_joni> what's up?
[17:56:15] <Imperator_> the PCI interface cannot be that complicated if tey can do it in a smal CPLD
[17:56:36] <SWPadnos> no - the interface doesn't need to be too complex
[17:56:42] <Imperator_> talking with Stephen about the servocard
[17:56:57] <SWPadnos> it gets worse if you want to have busmastering and that kind of thing though
[17:57:03] <SWPadnos> (which you don't)
[17:57:07] <fenn> could you reverse engineer their cpld?
[17:57:20] <alex_joni> fenn: pretty hard
[17:57:23] <Imperator_> ok that one can only do single write and read
[17:57:24] <alex_joni> but it's doable
[17:57:53] <SWPadnos> there are PCI interfaces in the public domain, so reverse engineering isn't necessary here
[17:57:56] <Imperator_> want to have a "Free Seervo Card"
[17:57:59] <fenn> ok
[17:58:08] <fenn> me too :)
[17:58:18] <SWPadnos> free as in speech, not free as in beer :)
[17:58:19] <fenn> i hear mshaver is givin' em away
[17:58:32] <alex_joni> fenn: so you've heard?
[17:58:32] <Imperator_> what
[17:59:24] <fenn> this card would have amps built into it? or are they too big?
[17:59:39] <alex_joni> amps don't go on boards
[17:59:43] <alex_joni> that's bull
[17:59:49] <alex_joni> you could only have small ones
[17:59:55] <alex_joni> and those are not very helpfull
[18:00:13] <SWPadnos> well - a Sherline mill takes less power than an ATI X800, but then again ... :)
[18:01:09] <alex_joni> now there's 2 of 'em
[18:01:11] <alex_joni> :D
[18:01:20] <SWPadnos> uh-oh - les^2
[18:01:27] <les_away> aw storms knocking me off
[18:01:33] <alex_joni> as long as he's not cubic
[18:01:35] <alex_joni> =))
[18:01:40] <les_away> heh
[18:01:49] <alex_joni> god forbid quintic
[18:01:53] <alex_joni> LOL
[18:02:03] <alex_joni> so les^2 is pretty ok :)
[18:02:26] <SWPadnos> but there can be problems with the solution in some quadrants
[18:02:36] <les_away> The only math I am doing today is acoustics stuff..
[18:02:37] <SWPadnos> (he's imaginary - look at that)
[18:02:39] <jepler> jerk limits can't be met unless he's quintic
[18:02:50] <SWPadnos> then he's a jerk - oh, well ;)
[18:03:20] <alex_joni> jepler: welcome back
[18:03:24] <steve_stallings> Les, assume you missed my msg about KMotion.
[18:03:40] <steve_stallings> I got the download, funny it is not linked on their web site.
[18:03:47] <jepler> alex_joni: thanks
[18:04:04] <steve_stallings> It is not a typical compressed file. Did you just run the executable to install?
[18:04:11] <alex_joni> jepler: been working on axis & emc2
[18:04:18] <jepler> alex_joni: you have? How's it going?
[18:04:18] <alex_joni> bye les
[18:04:25] <alex_joni> jepler: got it to compile
[18:04:27] <alex_joni> and build
[18:04:29] <alex_joni> and link
[18:04:34] <jepler> alex_joni: we're happy to accept a patch as long as it doesn't break building on emc1
[18:04:37] <fenn> woohoo!
[18:04:38] <alex_joni> but.. I get some errors :(
[18:04:47] <les> grr
[18:04:50] <jepler> alex_joni: paste them in /msg if you like
[18:05:07] <steve_stallings> Les, quit hiding from us. 8-)
[18:05:10] <les> are you on a doze box steve?
[18:05:23] <steve_stallings> Got one around here somewhere...
[18:05:25] <alex_joni> jepler: lemme run my devel box
[18:05:41] <steve_stallings> Gee, this one, running NT...
[18:05:59] <les> read up and download and install that software...it has lots of source code
[18:06:14] <les> for both the host and the dsp
[18:06:26] <les> even has a c compiler
[18:06:27] <steve_stallings> So the executable is targeted at Windows?
[18:07:00] <les> The rs274 stand alone and the TP run on windows
[18:07:17] <les> servo loops and the rest are on the 600 mflop dsp
[18:07:32] <les> usb comms
[18:07:36] <steve_stallings> Development host == windows
[18:07:46] <les> but the dsp has a 40,000 segment buffer
[18:07:52] <les> right
[18:08:08] <steve_stallings> Did you spot any TP enhancements?
[18:08:08] <alex_joni> les: got sources from the TP ?
[18:08:12] <les> dev in windows and usb flash the card
[18:08:41] <les> yeah... I can send...
[18:08:48] <les> dcc?
[18:09:33] <alex_joni> sure
[18:10:09] <les> The whole thing is pretty impressive...but it is not a turnkey machine controller
[18:10:12] <les> k...
[18:12:26] <alex_joni> thx les
[18:12:30] <alex_joni> will look later
[18:12:33] <les> It needs to use more than a few g and m codes
[18:12:58] <alex_joni> heh
[18:13:05] <les> not hard though....I see it spitting out cannonical messages
[18:13:35] <les> and It can do inline IO in the gcode
[18:14:19] <les> I cannot run the hardware untill I hook some motors to it
[18:14:27] <SWPadnos> I note that the G-codes go up to 99 (.x) - is there some reason that they can't be extended above 100 (like for I.O)?
[18:14:48] <les> guess not
[18:15:11] <SWPadnos> I'm sure it was a paper tape limitation in the early days
[18:15:18] <les> this board uses only g0-4 ...a problem.
[18:15:49] <fenn> i've seen g-codes into the 200's
[18:15:49] <SWPadnos> aren't most of the other G-codes for modes that the higher level can deal with?
[18:16:01] <SWPadnos> OK - on what controls?
[18:16:06] <fenn> no idea
[18:16:09] <SWPadnos> heh
[18:16:21] <SWPadnos> were they for macros and other stuff like that?
[18:16:27] <fenn> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?gcode
[18:16:53] <SWPadnos> ah -thatk you
[18:16:54] <SWPadnos> thank
[18:17:27] <les> ah...just got hit by an outflow boundary...temp dropped 10 degrees in an instant
[18:17:34] <fenn> that was from cnczone.com - they didnt say where they got it from
[18:17:38] <SWPadnos> cool ;)
[18:17:50] <SWPadnos> which way is that weather headed?
[18:17:56] <les> here.
[18:18:06] <SWPadnos> brb - lunch
[18:18:09] <les> k
[18:18:19] <les> has some hail in it....
[18:18:52] <les> steve must have downloaded the kmotion stuff?
[18:19:52] <steve_stallings> yes
[18:20:30] <les> The code reads fairly easily
[18:21:08] <Jymmm> les : it's hailing there?
[18:21:21] <les> not here but...
[18:21:21] <steve_stallings> OT - Well I didn't win them, but they sure looked neat... http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&category=71394&item=7521121525&rd=1
[18:22:11] <les> couple miles away radar indicated hail
[18:22:15] <les> looking..
[18:22:16] <Jymmm> ah,ok.
[18:22:40] <Jymmm> Has anyone actually *built* a sound proof enclosure?
[18:23:26] <Phydbleep> Jymmm: Define 'sound proof',
[18:23:41] <Jymmm> Phydbleep at least 20 db drop
[18:24:48] <Phydbleep> 3/4" MDF shell with 1" styro-sheet inside.
[18:25:10] <Jymmm> Phydbleep have you actually built that? Or are you just guessing?
[18:25:20] <les> yes I have.
[18:25:40] <les> will have to again for this spray gunn
[18:25:43] <les> gun
[18:25:49] <Phydbleep> Jymmm: Built, used, sold to someone who wanted their printer as quiet as mine.
[18:26:06] <les> acoustic intensity is high enough to light fires if focused
[18:26:27] <Jymmm> les : what materials did you use and how did you sandwich the materials?
[18:26:29] <fenn> just don't set it to the "brown note"
[18:26:34] <Phydbleep> les: Cool! A Sonic Cigarette Lighter. :)
[18:26:35] <les> 50-100 kHz
[18:26:37] <jepler> how much hard drive space do I need to install bdi-4.20? is 800 megs enough?
[18:27:12] <les> Jymmm what do you want to do...it depends
[18:27:14] <jepler> is it known whether the install will run under QEMU or not? I don't want to run a mill, just look at the layout of files in it
[18:28:00] <Jymmm> les I need to build two... one to hold a shopvac that's not much louder than a household vac the other is to enclose a gantry router.
[18:28:24] <fenn> jepler, you can use dpkg to install the emc package, which contains most of the files you need
[18:28:36] <Jymmm> les I'd like to have a solid surface inside the gantry cse if at all posisble (coolant)
[18:28:46] <fenn> does that help at all?
[18:29:03] <jepler> fenn: I don't have a debian system
[18:29:17] <fenn> use alien to convert to rpm?
[18:30:57] <les> A heavy walled enclosure is effective for sound reduction
[18:30:58] <Phydbleep> jepler: Get 'apt-get'
[18:32:14] <alex_joni> jepler: why not get the source code from CVS?
[18:32:18] <Phydbleep> jepler: http://ayo.freshrpms.net
[18:32:29] <Jymmm> les : for the shopvac case, I was thinking mdf lined with EPS and an air chamber "zig-zagged" with EPS as well.
[18:33:54] <les> MDF works fairly well
[18:34:43] <Jymmm> les do you have a "acustics for dummies" refernce by chance? Things like what sound doens't like (right angles, curves, etc), what material/construction absorbs sound (carpet padding+bubblewrap+MDF)
[18:36:48] <les> ah there is a lot on the web I guess. In general resonant absorbers and such are used. Padding doesn't do much.
[18:37:35] <les> I have my old textbook out right now because I am having to design a 1/4 wave acoustic stub tuner to match impedance
[18:37:44] <les> this is ultrasonic work though.
[18:38:02] <Jymmm> les : yeah, wrong bandwidth =)
[18:39:13] <les> I have designed some silenced power tools for ITW but I cannot talk much about it
[18:39:25] <Jymmm> les : I was thinking the desinty of each material would absorb a certain portion of the specturm.
[18:39:38] <les> More like reflect
[18:40:18] <les> acoustic impedance is kind of like optical refractive index
[18:40:41] <les> it is measured in Rayls
[18:41:21] <les> =kg/ m^2 *s
[18:41:39] <Jymmm> les : Well I dont want todo something silly like toss in a sheet of plastic and actually turn it into a speaker cone =)
[18:41:39] <les> so big mass= big impedance
[18:42:14] <les> router noisre right?
[18:42:21] <les> noise
[18:42:30] <Jymmm> les primarily
[18:43:06] <les> fan and brush noise from the motor
[18:43:16] <les> and work noise from the tool
[18:43:22] <Jymmm> bingo
[18:43:43] <les> on PC fan noise is the biggest
[18:43:50] <les> by far
[18:43:53] <fenn> Jymmm try orthopaedic bed covering "eggcrate" foam
[18:44:14] <Jymmm> fenn I'll be running some collant as well
[18:44:17] <Jymmm> coolant
[18:44:31] <steve_stallings> steve_stallings is now known as steves_logging
[18:44:47] <les> the bed stuff is too light anyway
[18:44:58] <Jymmm> carpet foam? EPS?
[18:45:05] <les> put some over your ears and you will see
[18:45:14] <fenn> EPS will turn into a speaker cone
[18:45:25] <les> something heavy
[18:45:42] <fenn> oil bonded clay/sand
[18:45:54] <Jymmm> the carpet foam sounds good, as it's built from shredded pieces and they go every which direction.
[18:46:03] <Jymmm> But everything sound sgood in theory
[18:46:54] <les> air is 430 Rayls
[18:47:12] <les> steel is about 10^7
[18:47:40] <fenn> http://www.roushind.com/news_downloads/white_papers/Insight.pdf
[18:47:51] <Jymmmm> logger_aj, bookmark
[18:47:51] <Jymmmm> See
[18:49:04] <Jymmmm> 1/8" pvc (HDPE?) sheet + couple inches of carpet foam + 3/4" MDF all sandwiches together?
[18:50:40] <les> just MDF will do...you are trying to isolate, not absorb
[18:50:54] <fenn> good point les
[18:51:19] <Jymmmm> les I'm in an apartment, I'd rather go a lil extra if it'll help
[18:51:52] <les> router in apt...uh oh
[18:52:08] <fenn> jymmm set your MDF box on top of the carpet pad to isolate it from the table or floor
[18:52:15] <Jymmmm> les ya do what ya can with what ya got =)
[18:52:39] <Jymmmm> fenn good point
[18:52:54] <fenn> assuming your router frame is happy without any foundation
[18:53:07] <Jymmmm> s/assuming/hoping/
[18:53:38] <fenn> well, you can add mass/stiffness to the bottom of the box if it isn't happy
[18:54:21] <Jymmmm> oh you said router frame, sorry I thought you said the encloseure itself.
[18:56:37] <Jymmmm> Hmmmm, is drywall more dense than mdf ?
[18:56:57] <fenn> not even close to the density of mdf
[18:57:10] <fenn> and it sucks worse to work with
[18:57:13] <alex_joni> * alex_joni runs axis1.0 with emc2
[18:57:20] <fenn> * fenn cheers
[18:57:23] <alex_joni> even compiled on emc2 ,)
[18:57:41] <Jymmmm> fenn: ok, cool! was jsut thinking what are done in homes and such
[18:57:42] <alex_joni> great piece of SW :P
[18:57:49] <Jymmmm> axis?
[18:58:24] <alex_joni> Jymmmm: axis.unpy.net
[18:59:26] <Jymmmm> Phydbleep did you try your enclosure with something other than a dot matrix printer?
[18:59:50] <fenn> Jymmm drywall is for fire protection, nothing more
[19:00:42] <Jymmmm> fenn Heh, that could be a VERY GOOD THING in a router enclosure =)
[19:01:01] <fenn> nah, it's mostly airtight anyway
[19:01:20] <fenn> or, it should be
[19:01:37] <fenn> otherwise the sound will get out!
[19:01:42] <Jymmmm> fenn : Nope, that's what the shopvac is for, dust control
[19:02:19] <fenn> you should put the shopvac inside the enclosure
[19:02:36] <Jymmmm> fenn it'll have it's own enclosure
[19:02:49] <fenn> then have a return line for the "filtered" air
[19:03:01] <fenn> shop vacs aren't good enough at filtering to go blowing it into your bedroom
[19:03:19] <fenn> esp if you are machining circuit boards etc
[19:03:24] <Jymmmm> fenn better than nothing at all.
[19:03:53] <fenn> but for safety's sake return the shopvac air to the enclosure.. my shopvac has a "blower port" which is the same size as the intake port
[19:04:11] <Jymmmm> safety sake?
[19:04:12] <jepler> paul has a nice collection of photos in the bdi-4 installer
[19:04:13] <fenn> you don't want to be breathing wood/fiberglass/pvc fumes all day long
[19:04:38] <Jymmmm> I was planning on it anyway, but that makes sense.
[19:04:43] <alex_joni> jepler: doing an BDI-4 install?
[19:04:57] <jepler> alex_joni: yeah
[19:05:10] <Jymmmm> fenn this shopvac has screwlock fittings too.
[19:05:11] <alex_joni> nice
[19:05:21] <alex_joni> jepler: found it pretty nice (the install)
[19:05:36] <alex_joni> and debian is also pretty ok (if you got a decent net connection)
[19:05:43] <jepler> it annoyed me a bit that even after I deselected a bunch of packages it was still around 1 gigabyte
[19:05:58] <alex_joni> jepler: don't forget to do the steps mentioned in the wiki to have a build system
[19:06:00] <Phydbleep> Jymmmm: I was using mine with a printer and with a prototype pcboard mill.
[19:06:22] <Jymmmm> Phydbleep ah, ok.
[19:10:07] <Phydbleep> Jymmmm: Not dt-matrix though.. It was an antique hi-speed band printer.
[19:12:09] <alex_joni> fenn: still there?
[19:12:32] <fenn> yeah, about to reboot to test out my rtai kernel
[19:12:47] <alex_joni> what about wiki.linuxcnc.org ?
[19:12:56] <fenn> uh, still working on it :)
[19:12:57] <alex_joni> * alex_joni can't find shiet now :)
[19:13:36] <fenn> what's wrong?
[19:13:48] <alex_joni> the BDI4 stuff is not visible from the front page
[19:13:58] <alex_joni> or I don't know where to find it
[19:14:08] <fenn> compiling emc2
[19:14:24] <fenn> oops that's under installing emc2
[19:15:06] <alex_joni> why didn't you use the existing emc2 about?
[19:15:17] <alex_joni> emc2 description
[19:15:30] <alex_joni> EmcVersion2
[19:15:31] <fenn> like i said, i'm still working on it
[19:15:38] <alex_joni> ok..
[19:16:01] <fenn> and you can't rename pages.. annoying as hell
[19:16:17] <fenn> my goal this week is to get it on pukiwiki so i can rename all the pages
[19:18:10] <fenn> ok i'm rebooting
[19:20:43] <alex_joni> * alex_joni will be back later
[19:20:45] <alex_joni> bye
[19:20:51] <jacky^> ciao alex_joni
[19:20:57] <Jymmmm> bye bye alex_joni
[19:21:09] <alex_joni> ciao jacky^, Jymmmm & co
[19:21:11] <alex_joni> :P
[19:21:28] <fenn> well it appears to have worked
[19:23:30] <SWPadnos> your fc3 RT kernel?
[19:23:37] <fenn> yeah
[19:23:44] <SWPadnos> cool
[19:24:00] <fenn> i'm trying to use the rtai test tools but my sudoers file isn't working how i thought it should
[19:24:08] <fenn> so i get permission denied
[19:26:51] <fenn> i do sudo -l and it says Sorry, try again. ??
[19:27:54] <SWPadnos> does this work: sudo -c "command name and parameters"
[19:28:00] <SWPadnos> (or is that C?)
[19:28:20] <fenn> no, illegal option -c
[19:28:53] <SWPadnos> sorry - su -c
[19:29:58] <fenn> the problem is rtai uses sudo in a script and sudo isn't working
[19:30:08] <SWPadnos> ah - that would be a problem
[19:30:32] <SWPadnos> can you try going root with su, then executing the script from within the root shell?
[19:30:58] <SWPadnos> (nopefully sudo isn't so brain-dead that it will prevent root from doing things)
[19:31:08] <SWPadnos> hopefully
[19:32:04] <fenn> ah that worked.. well, at least it has a different error now :)
[19:32:10] <SWPadnos> heh
[19:32:47] <fenn> ERROR: Module rtai_calibrate does not exist in /proc/modules
[19:33:14] <fenn> everything in /usr/realtime segfaults
[19:33:39] <SWPadnos> what does uname -a return? (dod you add an extraversion for the ADEOS kernel?)
[19:33:41] <SWPadnos> did
[19:33:49] <fenn> i think maybe my vanilla kernel modules got mixed up with the 2.6.10 kernel
[19:34:12] <fenn> Linux snacker 2.6.10 #2 Mon Jun 6 12:45:44 EST 2005 i686 athlon i386 GNU/Linux
[19:34:20] <SWPadnos> did you modprobe rtai_sched (or whatever it's called)?
[19:34:36] <fenn> i just patched the vanilla sources and renamed the bzImage files
[19:34:53] <SWPadnos> ah - that was brave :)
[19:35:01] <fenn> or foolish :)
[19:35:08] <fenn> what other options are there?
[19:35:09] <SWPadnos> often the same thing
[19:35:13] <jepler> Anybody else wished they could compile & run axis on emc2? alex_joni has contributed a patch for it: http://axis.unpy.net/index.cgi/downloads/01118085613
[19:35:30] <cradek> yay whee!
[19:35:49] <SWPadnos> add a menu item to grub / lilo, and name one "known to work 2.6.10", and the other "2.6.10-experimental"
[19:36:01] <SWPadnos> and have two bzImages in /boot
[19:36:08] <fenn> yeah that's pretty much what i did
[19:36:21] <SWPadnos> OK - then you're not brave (or foolish) ;)
[19:37:15] <fenn> the BDI has a kernel-sources-2.6.9-adeos rpm
[19:37:22] <SWPadnos> fenn: what happens if you modprobe rtai_sched?
[19:37:38] <SWPadnos> (I think that's it - you may want to check the name)
[19:37:40] <fenn> rtai_ksched ?
[19:37:45] <SWPadnos> could be the one
[19:37:52] <fenn> not found
[19:38:04] <fenn> i'm looking at it.
[19:38:07] <SWPadnos> that should be a link to wither rtai_up or rtai_mp (or something like that)
[19:38:19] <fenn> it links to rtai_up
[19:38:36] <SWPadnos> I assume you did a make install (or modules_install) on RTAI?
[19:39:02] <fenn> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?FC3
[19:39:11] <fenn> i'm doing a writeup of what i did
[19:39:25] <fenn> the source of the wiki is easier to read
[19:40:03] <SWPadnos> did you compile RTAI on the patched kernel? (ie, while that kernel was running)
[19:40:12] <fenn> no
[19:40:16] <SWPadnos> you should do that
[19:40:22] <SWPadnos> make clean, etc.
[19:40:41] <fenn> why does it matter what kernel is running?
[19:40:44] <SWPadnos> I believe it uses uname -a to get some locations (and I know EMC does)
[19:40:50] <fenn> ugh
[19:40:59] <SWPadnos> heh
[19:42:34] <SWPadnos> I think you can manually configure that stuff, but the automatic configuration assumes that you're compiling on the kernel you want to run on
[19:42:41] <SWPadnos> which is reasonable
[19:44:03] <jacky^> anyone can tell me the command to use in emc to move of 50 units in a dirction ?
[19:44:21] <SWPadnos> g1 x50
[19:44:30] <SWPadnos> plus a feedrate
[19:44:31] <fenn> hrm aspell is hogging the cpu.. that's weird
[19:44:33] <SWPadnos> g1 x50 f10
[19:44:39] <SWPadnos> (would take 10 minutes)
[19:44:39] <jacky^> x mean x axes ?
[19:44:43] <SWPadnos> yes
[19:44:49] <jacky^> ok thanks
[19:44:50] <SWPadnos> sorry - 5 minutes
[19:45:07] <jacky^> very slow ?
[19:45:07] <SWPadnos> g0 will go at the maximum speed configured in the ini file (the rapid rate)
[19:45:17] <SWPadnos> the F specifies units per minute
[19:45:25] <jacky^> uhm..
[19:45:32] <jepler> that moves from the old position to x=50. It doesn't move x by +50 units.
[19:45:35] <SWPadnos> are you using mm?
[19:45:37] <jacky^> i need to know how much move
[19:45:44] <jacky^> yeah, mm
[19:45:45] <SWPadnos> ah - sorry - one second
[19:45:47] <jepler> G91 sets incremental distance mode
[19:46:17] <SWPadnos> right
[19:46:17] <jepler> so that g91 g0 x50 will move the X axis by 50 units
[19:46:22] <SWPadnos> G91
[19:46:30] <jacky^> also 50 mm ? right ?
[19:46:32] <SWPadnos> G1 X50 F100
[19:46:42] <jacky^> if.. input scale it's right
[19:46:44] <SWPadnos> G20 and G21 switch between inches and mm
[19:46:46] <jacky^> i suppose
[19:47:26] <jacky^> i'm going to try..
[19:47:45] <SWPadnos> so G21 G91 G1 X+50 F100 will move 50 mm away from the current position in 30 seconds
[19:48:07] <SWPadnos> (you can probably move a lot faster than that if you're impatient)
[19:48:35] <jacky^> i just need to know if input scale parameter is right
[19:48:51] <SWPadnos> steppers?
[19:49:03] <jacky^> yes, 200 step
[19:49:17] <jacky^> screw is 6 mm
[19:49:21] <SWPadnos> OK. make sure that input_scale and output_scale ar ethe same, or you'll get following errors
[19:49:27] <jacky^> pitch i don't know
[19:49:28] <SWPadnos> ... are the same ...
[19:49:46] <jacky^> SWPadnos: yeah, are the same
[19:49:59] <SWPadnos> OK - good
[19:50:09] <ValarQ> * ValarQ really has to learn those G-codes
[19:50:35] <SWPadnos> I don't know them - I just happen to have RS274NGC_3.pdf open :)
[19:50:42] <ValarQ> :)
[19:51:15] <SWPadnos> ValarQ: what would you think of basing a HAL configuration tool on the crapplication GUI?
[19:51:16] <ValarQ> i have a friend who is working with a yasta milling machine the entire days
[19:51:21] <ValarQ> and has been for some years
[19:51:46] <ValarQ> SWPadnos: shouldn't be that advanced
[19:52:06] <jacky^> ah.. SWPadnos , the value of output scale i insert was 100, after running sherline mini it change to 100.000 , it's normal ?
[19:52:32] <SWPadnos> in the .ini file?
[19:52:36] <jacky^> yes
[19:52:37] <ValarQ> SWPadnos: i'm a console junkie thought, someone has to tell my how the gui should be designed (or at least basic design)
[19:52:47] <SWPadnos> that seems odd to me, but I'm not an expert
[19:52:50] <jepler> jacky^: probably normal. That's a decimal point, not a thousands separator, so the two values are equal.
[19:52:57] <SWPadnos> ValarQ: Just thinking of being able to do schematic-like editing, and remembered your crap ;)
[19:53:05] <jacky^> ok, thanks
[19:53:20] <ValarQ> SWPadnos: i'm happy to help doing tools but i will always use vi for myself
[19:53:25] <SWPadnos> oh yeah - I forgot about the separator differences :)
[19:53:34] <SWPadnos> feel free - it's all about choice
[19:53:35] <ValarQ> SWPadnos: schematic-like?
[19:53:57] <SWPadnos> yes - drop HAL part on page, wire between HAL pins, output text .hal file
[19:54:07] <ValarQ> SWPadnos: hmm... intresting idea
[19:54:21] <ValarQ> SWPadnos: shouldn't be that hard with pygtk
[19:54:29] <SWPadnos> jmk has been thinking about it for a long time - I pointed out crapplication to him last night
[19:54:56] <ValarQ> heh, that crap ;)
[19:55:02] <Jymmmm> http://www.axyz.com/table_mill.html
[19:55:15] <SWPadnos> he seemed interested, but he does low-level stuff mostly, so a GUI in python isn't his thing :)
[19:55:19] <ValarQ> craphaldesigner :)
[19:55:31] <SWPadnos> halp_me
[19:56:01] <SWPadnos> or crapahalic :)
[19:56:02] <ValarQ> heh :)
[19:56:15] <jepler> what is this "crapplication" you're talking about?
[19:56:21] <SWPadnos> heh
[19:56:26] <jepler> I've done a GUI in Python before
[19:56:27] <ValarQ> jepler: my pcb-editor :)
[19:56:39] <ValarQ> jepler: http://arda.no-ip.org/crap
[19:56:58] <SWPadnos> jepler: heh - I thikn I've heard of it :)
[19:57:17] <ValarQ> oh, a famous gui
[19:57:49] <jacky^> after executed cG1 y50 F100 command moved by 10 mm it's right ?
[19:57:51] <jepler> ValarQ: looks interesting.
[19:58:08] <SWPadnos> off by a factor of 5
[19:58:32] <fenn> cool latency avg 655 ns
[19:58:39] <fenn> is that good?
[19:58:43] <ValarQ> SWPadnos: it needs a good way to figure out which pins a driver has
[19:58:47] <jacky^> actually input scale is = 100
[19:58:56] <ValarQ> SWPadnos: like the stepgen driver
[19:59:04] <SWPadnos> yep
[19:59:12] <SWPadnos> phone - one minute
[19:59:15] <ValarQ> SWPadnos: the pins are dependent on the cfg flag
[19:59:18] <ValarQ> SWPadnos: ok
[20:00:19] <fenn> there was a latency survey somewhere..
[20:00:52] <fenn> ah nm found it
[20:01:18] <jepler> when benchmarking a realtime kernel, what relevance does the average latency have? None, I'd say
[20:02:05] <fenn> um, isn't average latency the whole point??
[20:02:15] <jepler> surely the point would be *worst-case* latency
[20:02:52] <fenn> i'm outta my realm of knowledge here
[20:03:56] <jepler> You need to generate a 40MHz pulse-train by flipping a bit every 25uS. Can you do it on a system with average latency 655 ns, worst-case latency 1ms? No.
[20:04:05] <fenn> max is 13557 ns.. comparable to other duron results
[20:04:32] <jepler> er, 40kHz pulse-train
[20:04:52] <fenn> this box is mainly for development anyway, so the results dont really matter
[20:09:32] <fenn> 25000 ns on the shop box.. so theoretically 40khz
[20:09:57] <jepler> I just wanted to explain why worst-case latency was important. It's a pet peeve of mine.
[20:10:41] <fenn> if you have higher latency than your period, does it skip a cycle or is it just timed wrong?
[20:10:45] <jepler> I guess the latency need not be nearly so small for servo machines as for steppers
[20:13:11] <jepler> I don't know what actually happens. People refer to "losing steps", but whether it's because emc actually skips them or because the direction pulse is just asserted for too little time, I'm not sure.
[20:20:45] <rayh> I've never known emc to actually loose steps. If the pulse generator can not keep up with the commanded speed, it will toss those it can't produce in following error.
[20:21:26] <rayh> I have often known the connection (both wire and waveforms) to loose a lot of steps.
[20:23:55] <SWPadnos> latency is pretty important when it becomes a significant part of the interrupt period
[20:24:16] <jepler> too much jitter hurts stepper performance, I suppose
[20:24:39] <SWPadnos> right - I was thinking jitter there
[20:24:57] <SWPadnos> latency is also important, because it's time that can't be spent doing the RT task
[20:25:07] <SWPadnos> if you think about it, it's time spend processing other stuff
[20:25:10] <SWPadnos> spent
[20:25:33] <SWPadnos> so if you have 10 uS latency, and you want to run a 15 uS PERIOD, you only get 1/3 of the CPU for the step generator
[20:25:45] <SWPadnos> the other 10 uS are spent doing other stuff
[20:26:30] <SWPadnos> sorry to digress - I should have read the log first :)
[20:26:59] <SWPadnos> ValarQ: yes - there would be a problem with determining the pins exported
[20:27:12] <SWPadnos> especially since it can vary, even for the same module (parport, for example)
[20:27:19] <ValarQ> SWPadnos: yeah, there is no automatic way to get them :/
[20:27:40] <SWPadnos> unless you try loading hte module, and looking at the pins it exports ;)
[20:27:40] <ValarQ> SWPadnos: one could define that in some file thought
[20:27:51] <ValarQ> SWPadnos: hmm :/
[20:27:57] <SWPadnos> not necessarily - consider a PLC
[20:28:10] <SWPadnos> it would export only the pins needed by the specific ladder diagram it's running
[20:28:18] <ValarQ> it could be done in realtime mode thought
[20:28:35] <SWPadnos> and there could be 3 of them simultaneously, each with a different diagram
[20:28:48] <SWPadnos> yes - the hal library provides pin discovery functions
[20:29:01] <ValarQ> it's pretty nice to get the changes immediately :)
[20:29:03] <SWPadnos> but the module needs to be loaded, in the configuration in which it will be used
[20:29:56] <ValarQ> yeah
[20:34:13] <SWPadnos> bummer - I hadn't thought about the pulymorphism of modules before - it does complicate a HAL schematic editor
[20:34:17] <SWPadnos> polymorphism
[20:35:12] <fenn> how do you do polymorphism in C?
[20:35:47] <SWPadnos> not quite the same - I;m thinking of the fact that a single module can look different, depending on the configuration when it's loaded
[20:36:14] <jepler> fenn: with common initial fields in a structure, explicit use of casts
[20:36:55] <SWPadnos> the parport module exports the pins that are always inputs and outputs (on the control port), and the 8 data port pins as either inputs or outputs
[20:37:10] <SWPadnos> so it's two differetn modules, from a schematic point of view
[20:37:41] <SWPadnos> a PLC might export only the pins needed for the particular ladder it's executing, so that would vary all over the place
[20:37:56] <ValarQ> one could start with an halcmd interface only thought
[20:39:15] <SWPadnos> just make a GUI front end that operates halcmd in the background?
[20:39:21] <ValarQ> yeah
[20:39:31] <ValarQ> and be able to save the .hal file
[20:40:08] <SWPadnos> actually - that could work - there's a halcmd option to export the current setup in .hal file format
[20:40:28] <SWPadnos> d00d - u r0xx0rz :)
[20:41:03] <ValarQ> hmm, i would prefer to be able to throw in some designblockstuffpositions as comments in the .hal
[20:41:08] <SWPadnos> I was just thinking that it would be difficult to filter out what's necessary to duplicate a setup, but halcmd does that for you
[20:41:35] <SWPadnos> you can export the file (or to a pipe), and rewrite it as you see fit
[20:42:09] <ValarQ> yeah, that might work
[20:42:33] <SWPadnos> or have a separate file (which wouldn't be good for transferring setups)
[20:43:14] <ValarQ> how should the schemedesignmode look/work ?
[20:43:42] <SWPadnos> more or less like your pcb layout package
[20:43:52] <SWPadnos> a list of available modules that can be dropped on the canvas
[20:44:11] <SWPadnos> if there are configuration parameters for the module, then prompt the user for them
[20:44:12] <ValarQ> how about the pins?
[20:44:30] <ValarQ> one has to display their name in some clever way
[20:44:42] <SWPadnos> load the module to see what pins it exports, and then draw a rectangle with inputs on the left and outputs on the right, or something like that
[20:45:09] <ValarQ> could play around with that a bit...
[20:45:12] <SWPadnos> tooltips - hover over a pin, and the name, type, and direction show up
[20:45:28] <SWPadnos> or click on one, and a status area shows the information
[20:46:03] <SWPadnos> maybe put certain named pins on top or bottom, like "enable" and "error" or "status"
[20:46:33] <ValarQ> hmm
[20:46:53] <fenn> ooh and multi-colored signal lines :)
[20:47:01] <SWPadnos> actually, a config file could specify regexps and top / bottom / left / right placement for amtching pins
[20:47:07] <SWPadnos> matching
[20:47:28] <SWPadnos> OK - specify attributes like color and position
[20:47:57] <ValarQ> fenn: heh :)
[20:48:56] <SWPadnos> yeah - and different line types, like dot-dot-dot-dash-dash-dash-dot-dot-dot when you connect two outputs together, or different type pins :)
[20:49:32] <fenn> and multiple shiny buttons with different skins
[20:49:53] <SWPadnos> well - of course it has to be themable :)
[20:50:03] <ValarQ> gtk2...
[20:50:05] <fenn> i want one with big levers and a panic button
[20:50:05] <ValarQ> done and done
[20:50:09] <SWPadnos> heh
[20:50:18] <SWPadnos> ValarQ: are you familiar with LabView?
[20:50:25] <ValarQ> SWPadnos: nope
[20:50:29] <SWPadnos> OK.
[20:50:34] <ValarQ> SWPadnos: whats that?
[20:50:36] <SWPadnos> http://www.ni.com
[20:51:00] <SWPadnos> it's a graphical "programming" environment for control systems and operator interface panels
[20:51:14] <ValarQ> ew
[20:51:23] <SWPadnos> something that's a very scaled-back version of that is what the idea is :)
[20:51:41] <SWPadnos> yeah - it largely sucks for programmers, but it is pretty
[20:52:31] <ValarQ> those cute ui:s hurts my eyes
[20:52:51] <SWPadnos> heh - at least you've never had to *program* with it ;)
[20:53:37] <ValarQ> http://arda.no-ip.org/iceshot.png
[20:54:33] <SWPadnos> hmmm - looks familiar for some reason
[20:55:21] <ValarQ> icewm with a half stolen theme
[20:55:30] <SWPadnos> right
[20:55:51] <ValarQ> i painted it black :)
[20:55:57] <SWPadnos> actually - this is pretty cool: http://www.ni.com/fpga/
[20:56:17] <jepler> argh, my bdi install just blew chunks
[20:56:19] <ValarQ> * ValarQ runs X11 to be able to use more terminals
[20:56:26] <SWPadnos> heh
[20:56:56] <jepler> The package kudzu- cannot be opened. This is due to a missing file or perhaps a corrupt package.... (over and over again)
[20:57:18] <SWPadnos> during installation?
[20:57:21] <jepler> yeah
[20:57:31] <SWPadnos> hmmm - try burning another CD
[20:57:38] <jepler> I'm installing in qemu from the .iso
[20:57:44] <jepler> $ qemu -hda bdi.img -cdrom bdi-4.20.iso -boot d
[20:57:58] <SWPadnos> hmmm - can't help you there ;)
[20:58:03] <jepler> I never md5summed it, maybe it's a bad download
[20:58:34] <SWPadnos> could be - I'd check
[20:59:18] <jepler> nope, the md5sum is right
[20:59:25] <jepler> cfabb394dcb1478fae6410bdd19392e9 bdi-4.20.iso
[21:01:15] <jepler> oh -- the qemu disk is full. Despite the fact that it was a good 300 megs larger than what the installer told me it required.
[21:03:24] <jepler> * jepler starts over (ugh)
[21:07:59] <jepler> argh. I answered the wrong thing so now the installer is restarting. and it takes a good 5 minutes to start under qemu
[21:08:02] <jepler> sigh
[21:10:51] <robin_sz> meep?
[21:13:05] <jepler> you just missed me complaining about the bdi-4.20 installer. I'm now in my 4th run of it...
[21:13:20] <jepler> and I don't even want to run bdi, not really
[21:15:07] <jepler> (just want to see if I can get AXIS compiling for it)
[21:16:10] <jepler> well, see you guys around...
[21:22:45] <jacky^> umpf..
[21:23:00] <jacky^> finally i found the correct value for input scale: 400
[21:23:20] <jacky^> ghghghg :\
[21:24:24] <jacky^> G1 y+50 F100 move 50 mm
[21:28:07] <jacky^> emc can detect home switch ?
[21:28:11] <robin_sz> yes
[21:28:50] <jacky^> should i find parameters in .ini file ?
[21:29:13] <robin_sz> somewhere in there, yes
[21:29:17] <ValarQ> jepler: oh, now i think i know which pythonbased frontend you were famous for :)
[21:29:26] <jacky^> thanks robin_sz
[21:35:53] <jacky^> hi alex_joni
[21:35:57] <alex_joni> yo jacky^
[21:37:26] <ValarQ> 'lo alex_joni
[21:38:08] <alex_joni> hello ValarQ
[21:40:12] <ValarQ> enough hal-fe coding for today
[21:40:21] <ValarQ> * ValarQ is of to sleep for a couple of hours
[21:40:45] <alex_joni> nice
[21:40:48] <alex_joni> hal-fe?
[21:41:02] <ValarQ> hal frontend
[21:41:26] <ValarQ> alex_joni: mr Padnos had some funny idea which i'm trying to implement
[21:41:46] <alex_joni> what idea?
[21:42:05] <ValarQ> a hal design tool
[21:42:36] <SWPadnos> hey - it's not funny
[21:42:43] <SWPadnos> :)
[21:42:47] <ValarQ> SWPadnos: :P
[21:42:52] <SWPadnos> o_O
[21:42:59] <SWPadnos> :d
[21:43:00] <alex_joni> SWPadnos: what's not funny?
[21:43:00] <ValarQ> SWPadnos: maybe fun :)
[21:43:02] <SWPadnos> hmmm
[21:43:04] <SWPadnos> could be fun
[21:43:10] <SWPadnos> should be fun
[21:43:11] <alex_joni> * alex_joni thinks SWPadnos is funny
[21:43:20] <alex_joni> ValarQ: halgui?
[21:43:20] <SWPadnos> yeah - har har :)
[21:43:31] <SWPadnos> crapahalic, I think
[21:43:44] <alex_joni> crapaholic
[21:43:53] <ValarQ> alex_joni: i guess the native speakers interprets "funny" different than us
[21:43:57] <SWPadnos> no - crapa-HAL-ic
[21:44:02] <alex_joni> ValarQ: right
[21:44:18] <SWPadnos> funny "strange" or funny "ha ha" ?
[21:44:22] <alex_joni> SWPadnos: crap-a-HAL-ic
[21:44:24] <ValarQ> SWPadnos: yeah, thats the development name now :)
[21:44:29] <alex_joni> funny = strange
[21:44:34] <SWPadnos> cool - let's hope it sticks :)
[21:44:39] <SWPadnos> OK - then I' mfunny
[21:44:45] <SWPadnos> i'm funny, even
[21:44:52] <SWPadnos> odd, too
[21:45:03] <ValarQ> * ValarQ is confused
[21:45:07] <SWPadnos> heh
[21:45:09] <ValarQ> * ValarQ goes to sleep
[21:45:19] <alex_joni> ValarQ: hang on a moment
[21:45:22] <SWPadnos> night night
[21:45:23] <ValarQ> ok
[21:45:25] <alex_joni> tell me what you want to do
[21:45:30] <alex_joni> *g*
[21:45:39] <alex_joni> * alex_joni goes to bed soon too
[21:45:42] <SWPadnos> I can describe it, if you really want to go to sleep
[21:45:49] <alex_joni> it's almost 1am
[21:45:53] <ValarQ> alex_joni: i'm trying to make a graphical tool to create .hal files
[21:45:53] <SWPadnos> heh
[21:45:57] <alex_joni> SWPadnos: that's ok too
[21:46:05] <ValarQ> 23:44 CEST
[21:46:10] <alex_joni> sounds like the halgui that's missing
[21:46:14] <alex_joni> [00:44] <ValarQ> 23:44 CEST
[21:46:18] <SWPadnos> have you seen ValarQ's crapplication?
[21:46:32] <alex_joni> lol.. no?
[21:46:47] <SWPadnos> http://arda.no-ip.org/crap
[21:47:13] <SWPadnos> I was talking to jmk last night, and thought of it (we ended up chatting about a HAL gui)
[21:47:45] <ValarQ> SWPadnos: i have ripped the pcb stuff out of crapplication now
[21:47:54] <SWPadnos> cool - that was fast!
[21:48:05] <ValarQ> SWPadnos: yeah :)
[21:48:23] <ValarQ> SWPadnos: its python :P
[21:48:36] <SWPadnos> heh - maybe I'll learn me that someday :)
[21:49:00] <ValarQ> python code is fast to delete
[21:49:01] <alex_joni> oh ValarQ: that looks perfect for hal
[21:49:20] <alex_joni> SWPadnos: if the snake causes you trouble.. you just need to play a tune for it
[21:49:27] <SWPadnos> rm -f pythoncode.py
[21:49:44] <SWPadnos> see - even I can remove python code :)
[21:49:50] <alex_joni> rm -fR /*py*
[21:50:09] <alex_joni> * alex_joni too
[21:50:11] <ValarQ> find / -name '*.py' -exec rm \{\}\;
[21:50:25] <SWPadnos> yeah - I was just about to say that
[21:50:26] <alex_joni> right
[21:50:33] <ValarQ> :P
[21:50:39] <websys> rm -r / works too
[21:50:51] <alex_joni> ValarQ: looks simple enough to be hijacked for the hal purpose
[21:50:56] <alex_joni> yo Bob
[21:51:01] <websys> Hi
[21:51:03] <ValarQ> websys: multi-language support! :D
[21:51:06] <SWPadnos> http://www.bash.org/?400459
[21:51:18] <alex_joni> how's the 64 stuff?
[21:51:28] <websys> Giving me headaches
[21:51:30] <SWPadnos> 8^2
[21:51:32] <alex_joni> SWPadnos: you didn't bash it hard enough
[21:51:37] <SWPadnos> heh
[21:51:48] <websys> int = 32 bits, char * 64 bits
[21:51:55] <alex_joni> I can imagine
[21:52:00] <alex_joni> ValarQ: love the filename extension you used
[21:52:05] <alex_joni> *big grin*
[21:52:17] <ValarQ> alex_joni: .crap?
[21:52:32] <SWPadnos> who srites this .crap?
[21:52:34] <SWPadnos> writes
[21:52:40] <alex_joni> yeah
[21:52:50] <alex_joni> SWPadnos srites .crap
[21:52:56] <alex_joni> :)
[21:53:02] <SWPadnos> yes, he soed
[21:53:05] <SWPadnos> :)
[21:53:13] <SWPadnos> wy cnat I ytpe?
[21:53:14] <ValarQ> * ValarQ had enough of this crap
[21:53:19] <SWPadnos> good night
[21:53:20] <ValarQ> * ValarQ is finally of to bed
[21:53:28] <alex_joni> ValarQ: you plan on releasing halcrap in swedish?
[21:53:37] <alex_joni> night ValarQ
[21:53:39] <SWPadnos> bork, bork, bork!
[21:54:00] <alex_joni> don't forget to Spara
[21:54:05] <alex_joni> :D
[21:54:11] <alex_joni> * alex_joni wonders
[21:54:19] <SWPadnos> heh
[21:54:24] <alex_joni> why does ValarQ's parport have 36 pins?
[21:54:34] <websys> ValarQ - you want Synergy in another language?
[21:55:00] <SWPadnos> because it's the centronics end of the cable?
[21:55:21] <alex_joni> centronics sux
[21:55:28] <alex_joni> :D
[21:55:33] <SWPadnos> not if you want to use a standard printer cable
[21:55:34] <alex_joni> just like centriod does
[21:56:16] <websys> Was wondering how long computers will continue to have parports - all the printers are USB now
[21:56:34] <alex_joni> websys: as long as there's emc
[21:56:38] <SWPadnos> not long, for consumer PCs
[21:56:38] <alex_joni> ;)
[21:56:51] <alex_joni> or was that the other way around?
[21:56:58] <websys> alex - wishful thinking
[21:57:27] <alex_joni> websys: always ;)
[21:57:41] <alex_joni> lol.. I like this:
[21:57:47] <alex_joni> Welcome to Vq's homepage
[21:57:54] <alex_joni> This webpage is best viewed with a browser.
[21:57:59] <SWPadnos> yeah - I liked that as well
[22:00:14] <alex_joni> * alex_joni is getting ghostscript now
[22:00:21] <alex_joni> :(
[22:00:53] <alex_joni> first it was latex
[22:00:58] <alex_joni> nah..
[22:01:06] <alex_joni> first it was doxygen
[22:01:06] <alex_joni> then latex
[22:01:12] <alex_joni> now ghostscript
[22:01:12] <alex_joni> :)
[22:02:27] <websys> Well you could always write your own format
[22:02:48] <websys> that everyone would complain about
[22:03:01] <alex_joni> sure ...
[22:03:06] <alex_joni> :)
[22:03:29] <SWPadnos> why bother - it's been done ;)
[22:03:36] <websys> and while you're at it build something to replace Linux
[22:03:55] <SWPadnos> on that note - gotta run. Catch you later.
[22:03:57] <alex_joni> websys: why not replace M$ crap?
[22:04:00] <SWPadnos> SWPadnos is now known as SWP_Away
[22:04:12] <Phydbleep> And stop tracking mud across my nice clean kitchen floor! :)
[22:04:17] <alex_joni> laters
[22:04:32] <alex_joni> Phydbleep: who stepped on your toes?
[22:05:05] <Phydbleep> alex_joni: That's my standard finish for any bitching. :)
[22:05:34] <websys> Toes? I thought someone stepped on his ....
[22:06:07] <alex_joni> websys: lol
[22:06:34] <alex_joni> * alex_joni wouldn't wanna know how Fido's kitchen floor really looks
[22:07:44] <Phydbleep> Pretty bad at the moment.. Getting ready to replace part of it that was fire damaged and badly patched ~30 years ago.
[22:08:11] <alex_joni> fire damaged = experiment that got out of control ?
[22:08:28] <alex_joni> or was it young Phydbleep playing with matches ?
[22:09:18] <websys> Bizarre toothbrushing incident?
[22:09:19] <Phydbleep> alex_joni: I'm not sure.. The hole in the underlayment is at the right place for a grease fire from the stove or from a gas water heater falling over.
[22:10:52] <alex_joni> websys: they didn't have electric toothbrushes 30 years ago
[22:10:58] <alex_joni> at least not overcharged ones
[22:11:02] <websys> :-)
[22:11:11] <alex_joni> like those 50k RPM ones
[22:11:39] <websys> 50K rpm? that'd remove your gums
[22:12:06] <alex_joni> well it's geared down for torque
[22:12:13] <alex_joni> gotta be aggressive on plaque
[22:12:18] <alex_joni> lol
[22:12:45] <Phydbleep> Be a man, use a glass-bead blaster. :)
[22:12:45] <websys> enough torque to remove your molars?
[22:13:02] <Phydbleep> jacky^: :)
[22:13:19] <jacky^> hi Phydbleep :-)
[22:13:49] <websys> Glass beads? Why not gargle with drain cleaner?
[22:13:57] <jacky^> sometime my connection died :\
[22:14:35] <Phydbleep> websys: Give your teeth that funny blue tint?
[22:14:42] <Phydbleep> Gives
[22:15:31] <websys> Ultra white
[22:15:44] <websys> and minty freash breath
[22:16:22] <Phydbleep> websys: No, That's Tidy-Bowl not Draino.
[22:16:45] <websys> ;)
[22:17:02] <Phydbleep> * Phydbleep hand websys a suspisiously large pink mint..
[22:17:16] <Phydbleep> Here, Try one of these. :)
[22:17:31] <websys> Please don't eat the big pick mints
[22:25:25] <websys> mmmm - yummy but what's that strange ammonia like scent?
[22:33:28] <alex_joni> evening paul
[22:33:33] <alex_joni> pretty late :)
[22:33:36] <paul_c> Hi Alex
[22:34:03] <alex_joni> what's up?
[22:34:42] <paul_c> testing code....
[22:35:07] <alex_joni> that's good I guess... :)
[22:36:09] <websys> Paul - did you need those dxf files for something or was that just testing?
[22:36:31] <paul_c> Just stress tests ;)
[22:37:15] <paul_c> although it would be nice to view the large one.... Even Qcad barfs on that one !
[22:37:17] <websys> ok - I could send back the "revolve" part as a .syn file for you to read
[22:37:45] <websys> the big one didn't barf - it just didn't have enough memory to read
[22:37:52] <paul_c> you found the problem ?
[22:38:40] <websys> Read it on a machine with 750 Mb and after an hour it stopped to say "Unable to allocate memory"
[22:39:01] <websys> So with enough memory it should read
[22:39:56] <websys> We set an aritrary 1 GB database limit
[22:40:04] <websys> could increase it if needed
[22:40:35] <paul_c> guess my measly 768Meg isn't enough then...
[22:40:47] <websys> same here
[22:41:34] <websys> could prob. modify it to read a much as it can and then abort with whatever it read
[22:42:05] <websys> usually people want the whole file or nothing at all
[22:50:21] <paul_c> maybe a warning on screen that the full file was not read due to memory limits.
[23:02:27] <alex_joni> paul_c: any thoughts on running an emc2-sim ?
[23:03:15] <jepler> If there was an emc2-sim I'd use it
[23:03:22] <jepler> you might even see an official axis-on-emc2
[23:03:27] <alex_joni> heh
[23:03:31] <jacky^> Phydbleep: do you think homing vel= 1 is too high value for the optoswitches ?
[23:03:42] <alex_joni> mm?
[23:03:56] <jacky^> yeah, mm
[23:04:48] <alex_joni> not really
[23:05:04] <jacky^> the value was 0.1, too slow :\
[23:05:12] <jacky^> ok, thanks
[23:05:27] <paul_c> jacky^: 60mm/min - kinda slow...
[23:05:47] <jacky^> right
[23:06:12] <paul_c> alex_joni: Shouldn't be too difficult to write an rtapi_sim wrapper round standard posix calls.
[23:06:33] <alex_joni> I see...
[23:06:41] <alex_joni> well.. I might take a look at this
[23:06:55] <paul_c> jacky^: 50 to 100 mm/Sec would be more "fun"
[23:07:00] <alex_joni> yay.. finally producing png's of the formulas
[23:07:07] <jacky^> ok, paul_c tnx
[23:07:13] <alex_joni> paul_c: not for homing...:)
[23:07:25] <paul_c> alex_joni: cvs up - Did some more today...
[23:07:45] <alex_joni> I just finished downloading about 50 Megs to make it work
[23:07:46] <alex_joni> :(
[23:07:50] <paul_c> alex_joni: Sure, home at full speed..... Unless you're a wimp.
[23:08:36] <paul_c> * paul_c has been upgraded to a 2Meg line, so *big* downloads are nuffink now o/
[23:08:47] <alex_joni> paul_c: 2 MB?
[23:09:04] <alex_joni> or 2 Mbit?
[23:09:54] <paul_c> 2Mbit
[23:10:02] <alex_joni> ahh.. ok
[23:10:10] <alex_joni> pretty nice for home use ;)
[23:10:20] <alex_joni> I'm still waiting for my FO link
[23:10:27] <alex_joni> should be 1-2 months away
[23:10:47] <paul_c> still sucks when the other end can only do 56Kb upload.
[23:10:53] <alex_joni> right
[23:10:59] <alex_joni> I have 11 at work
[23:11:05] <alex_joni> and 155 at school
[23:11:17] <alex_joni> so I know those frustrations ;)
[23:13:09] <jacky^> here, in italy, work well the guaranteed bandwith
[23:13:14] <jacky^> not others ..
[23:14:04] <alex_joni> ok.. I'm going to bed
[23:14:08] <alex_joni> pretty late now
[23:14:20] <jacky^> night alex_joni ,
[23:14:36] <alex_joni> night
[23:15:44] <alex_joni> morning rayh
[23:15:48] <alex_joni> I was just leaving
[23:16:36] <rayh> Hi Alex. Have a good night.
[23:16:56] <jacky^> hi rayh
[23:17:58] <rayh> How you doing?
[23:18:15] <jacky^> oh, no.. i'm here yet
[23:18:22] <jacky^> for a little
[23:18:26] <jacky^> :)
[23:18:43] <Phydbleep> jacky^: I dont have a clue about anthing other than simulations with emc
[23:19:09] <rayh> jacky^: What are you looking for?
[23:19:32] <jacky^> i'm here.. just to try to learn something
[23:20:02] <rayh> I can try to answer questions.
[23:20:10] <jacky^> :D
[23:20:27] <jacky^> i think would be too many..
[23:20:45] <jacky^> i'm learning how to setup emc
[23:21:02] <rayh> EMC or EMC2?
[23:21:18] <jacky^> difficult question..
[23:21:26] <jacky^> BDI 4.20
[23:21:39] <jacky^> i think emc2, right ?
[23:21:47] <jacky^> or not ?
[23:22:06] <rayh> The EMC that ordinarily runs with 4.20 is EMC.
[23:22:25] <jacky^> ah, ok, then emc ..
[23:22:27] <rayh> EMC2 can be compiled with it now
[23:22:54] <rayh> INI file questions?
[23:23:31] <jacky^> there are meany variable i don't know how they work
[23:23:54] <jacky^> i can say you that my machine is a prototipe
[23:24:02] <rayh> That file can cause real problems even for me.
[23:24:08] <jacky^> a veri inexpensive machine
[23:24:28] <rayh> Okay. How is it constructed?
[23:24:39] <jacky^> rayh: wich documentation shoul i read about ini files ?
[23:24:51] <jacky^> i found interesting the handbook
[23:25:30] <jacky^> rayh: ah.. my cnc is very strange
[23:25:32] <rayh> Handbook is the best
[23:25:36] <jacky^> i can give you an url
[23:25:39] <jacky^> a moment..
[23:25:48] <rayh> Though you will still have many questions.
[23:26:15] <jacky^> not so much
[23:26:48] <jacky^> because, i repeat, i don't know how many variable work..
[23:27:01] <jacky^> so no make sense for me..
[23:27:10] <jacky^> i need before understand
[23:27:15] <jacky^> hard work :\
[23:27:35] <rayh> Yes.
[23:28:02] <jacky^> also my english is very bad.. :\
[23:28:06] <fenn> the only ones i touched were input_scale and units
[23:28:19] <rayh> You are doing very well with english.
[23:28:39] <rayh> UNITS = 1 for mm?
[23:28:44] <fenn> yeah
[23:28:50] <jacky^> well, here's the cnc i build:
[23:28:55] <jacky^> http://www.roboitalia.com/modules.php?name=coppermine&file=thumbnails&album=75
[23:29:06] <jacky^> don't say me nothing.. LOL
[23:29:09] <jacky^> :)
[23:29:59] <jacky^> i hope in the next future can buy a real machine
[23:30:10] <jacky^> after i learned something
[23:30:42] <rayh> I don't say much about some I built, except they move.
[23:31:04] <jacky^> hehe
[23:31:26] <jacky^> yeah, anywai work
[23:31:31] <rayh> Just getting the thumbnails.
[23:32:04] <jacky^> oh, can't you zoom the images clicking on ?
[23:32:08] <paul_c> in the meantime.....
[23:32:16] <Phydbleep> jacky^: Is that foam you're cutting?
[23:32:25] <paul_c> * paul_c has been doing some "stuff" with doxygen
[23:32:33] <jacky^> Phydbleep: yes, it is
[23:32:51] <Phydbleep> paul_c: Any incriminating pictures to share?
[23:33:01] <rayh> Yes I could zoom images but they take very long to download.
[23:33:22] <jacky^> rayh: i know.. this server is very slow
[23:33:34] <paul_c> checkout emc2 head & run it through doxygen - Loads of picyure for you to admire/puzzle over
[23:33:37] <rayh> No my telephone is much slower.
[23:33:43] <jacky^> rayh: don't worry
[23:33:45] <jacky^> np
[23:34:37] <rayh> You do very nice work, jacky^
[23:35:02] <jacky^> hehe.. tnx
[23:35:21] <paul_c> rayh: It may be possible to use doxygen to produce some usr docs from the source code...
[23:35:31] <rayh> I misread the motor connectors.
[23:35:45] <rayh> paul_c: That would be great. I'll try it here after a bit.
[23:37:48] <rayh> How round was the circle you cut.
[23:40:27] <Phydbleep> rayh: All the circles I cut are round.. Working on a lathe makes that easy. :)
[23:41:02] <rayh> Yes that does make it a bit easier.
[23:42:01] <rayh> paul_c: Tell me again how to run doxygen?
[23:45:19] <rayh> got it (doxygen) on the way
[23:48:09] <paul_c> cd emc2/src
[23:48:20] <paul_c> doxygen doxconfig
[23:48:46] <paul_c> browse the html files in e,c2/docs/html
[23:50:26] <rayh> Okay. Will do. Long strange pauses in the download this evening.
[23:50:50] <rayh> Was that board letter okay?
[23:53:47] <paul_c> on emc2 direction ?
[23:54:24] <rayh> The report on the vote on that motion.
[23:56:07] <paul_c> looked OK to me.
[23:56:39] <rayh> Good. I never thought of it that way. Thanks for pointing that out to me.
[23:57:52] <rayh> I was trying to build a stand alone interpreter today but linking failed.
[23:58:16] <paul_c> in which tree ?
[23:58:57] <rayh> I took it from emc/src/rs*new