#emc | Logs for 2007-08-29

[00:26:44] <JymmmEMC> jepler: I was, yes. But from everything I've read BEOFRE I bought the xylotex board, it's prone to heat issues. So if you're not running at the upper limit, maybe just one fan should do nicely
[00:27:14] <SWPadnos> tomp, what kind of DAQ stuff?
[00:27:46] <cradek> JymmmEMC: you should run the 12v signal, not the 5v, the distance
[00:28:18] <SWPadnos> yeah - between the max232 and the computer. that's what serial lines are for
[00:28:42] <SWPadnos> note that a 40' cable may reduce the max data rate the chips can drive (shouldn't be a problem at 9600 baud though)
[00:29:56] <JymmmEMC> thanks cradek, SWPadnos, not what I wanted to hear, but that's what I'll do.
[00:30:17] <JymmmEMC> SWPadnos: Mostly 4800 baud, but it might go to 56K
[00:30:44] <SWPadnos> it may not with 40' of cable and a low power transceiver chip - check the specs before you build
[00:31:07] <JymmmEMC> SWPadnos: you mean the max232 itself?
[00:31:22] <SWPadnos> yes, there are many flavors of chip that are all called something like MAX232
[00:31:27] <SWPadnos> they add lots of suffixes to them
[00:31:48] <JymmmEMC> SWPadnos: What should I be looking for/at (for long hauls)?
[00:32:01] <SWPadnos> (low power, high speed, single supply, +5/+12 supply, triple supply, no external caps needed)
[00:32:09] <SWPadnos> dunno - good drive strength?
[00:32:18] <SWPadnos> you'll have to read datasheets to find this stuff
[00:32:32] <SWPadnos> also full voltage swing
[00:32:59] <SWPadnos> 40' of cable isn't a lot, but I think it isn't short enough to ignore completely either
[00:34:04] <JymmmEMC> so this is bad? http://pdfserv.maxim-ic.com/en/ds/MAX220-MAX249.pdf
[00:34:09] <JymmmEMC> voltage swing wise?
[00:34:15] <SWPadnos> I don't really have time to look at the moment
[00:34:18] <JymmmEMC> k
[00:46:18] <SWPadnos> .nz - cool
[00:46:37] <helen> Hey guys, was wondering if someone cud help me out with a problem
[00:46:58] <SWPadnos> ask away - people will answer if they can help
[00:47:59] <helen> ok, i am running kernel 2.6.19 on ubuntu6.06, i have compiled rtai3.5 on it and latency results seem fine
[00:48:24] <helen> then i compiled emc2 from source, but when i tried to launch it, it kept freezing on me
[00:48:30] <helen> i looked into the ini file
[00:48:38] <helen> and tried changing the base period around
[00:48:54] <SWPadnos> what CPU speed/type, and what was the original BASE_PERIOD?
[00:49:12] <helen> it is now at 1000ns and it doesnt freeze anymore, but i get an error:
[00:49:21] <SWPadnos> 1000?
[00:49:29] <skunkworks> yikes
[00:49:33] <SWPadnos> that's extremely low
[00:49:36] <cradek> have you tried the prebuilt kernel and rtai that we distribute? it's very easy to build them wrong
[00:49:51] <cradek> no need to build your own if using ubuntu 6.06
[00:50:13] <SWPadnos> you do need to build your own if you want kernel 2.6.19 for some reason though (?)
[00:50:17] <helen> the reason why i rebuilt it is because i need wacom support in the kernel
[00:51:06] <helen> the kernel version from the live cd doesnt seem to recognise the wacom device even after the warcom driver is installed
[00:51:29] <cradek> using that kernel, though, does emc run on your computer?
[00:51:35] <helen> yeah kernel2.6.19 can recognise warcom
[00:52:09] <helen> emc runs fine if i install it straight from the cd
[00:52:14] <cradek> (I'm trying to understand what troubleshooting you've done already)
[00:52:25] <SWPadnos> rtai latency test works
[00:52:25] <cradek> ok so it's your kernel or rtai, not the emc configuration, right?
[00:53:00] <cradek> SWPadnos: I've had immediate lockups when running emc, even though the latency test works, on some unfortunate kernel builds
[00:53:06] <helen> i dunt quite understand...
[00:53:07] <SWPadnos> heh
[00:53:18] <cradek> this is basic troubleshooting
[00:53:29] <helen> what shall i do?
[00:53:28] <cradek> if an emc config runs on the distrib kernel, but not yours, yours is bad, right?
[00:53:41] <cradek> narrow down the problem until you find it
[00:53:43] <helen> so its my kernel problem?
[00:54:27] <helen> are there tests i can do to figure this out?
[00:54:30] <cradek> I can only guess, I'm trying to explain how you might narrow down the problem space
[00:54:40] <cradek> yes of course
[00:54:51] <cradek> don't change the emc config that works on our kernel. try it on yours.
[00:55:16] <cradek> changing numbers willy-nilly just confuses the issue
[00:55:22] <skunkworks> btw - what was the error your where getting at 1000ns?
[00:55:37] <helen> ok here is the error
[00:55:57] <helen> insmod: error inserting '/usr/realtime/modules/emc2/motmod.ko': -1 Operation not
[00:56:03] <helen> permitted
[00:56:14] <cradek> that's probably because the period is bogus
[00:56:28] <cradek> -1 is a generic error, the real error message is above it in dmesg
[00:57:13] <helen> ok the default configs on kernel2.6.19 doesnt run, freeze on startup
[00:57:23] <cradek> ok
[00:57:28] <cradek> then changing the configs will not help you troubleshoot
[00:57:47] <helen> then what should i do?
[00:57:58] <cradek> you have to figure out what is wrong with your kernel
[00:58:03] <cradek> maybe compare the .config settings to ours
[00:58:13] <cradek> there are lots of things that have to be set right
[00:58:32] <cradek> acpi options and memory size spring to mind
[00:58:41] <cradek> also modversions
[00:59:04] <helen> i compiled this kernel using the .config from the live cd
[00:59:16] <SWPadnos> did you do `make oldconfig' ?
[00:59:18] <helen> it was the only way i could get rtai working
[00:59:25] <helen> make menuconfig
[01:00:14] <cradek> did rtai 3.5 even have a patch made for 2.6.19?
[01:00:31] <helen> yep it was the latest one it had
[01:00:34] <cradek> ok
[01:00:45] <cradek> I don't think I have used this combination
[01:00:55] <cradek> I know 2.6.18 was bad for rtai and not supported
[01:01:05] <cradek> I tried 17 and 19 but don't recall which combinations worked
[01:01:34] <cradek> unfortunately this is a huge pain and very much a shot in the dark
[01:01:33] <helen> what about CVS magma
[01:01:41] <helen> shall i try that?
[01:01:47] <cradek> (that's why we distribute a tested setup for folks)
[01:01:56] <cradek> yes 2.6.20.x with magma is a good bet
[01:02:02] <cradek> or whatever the newest it supports is
[01:02:21] <helen> that means i should use kernel2.6.20x?
[01:02:26] <cradek> are you going to run a machine with this, or just test/simulate?
[01:02:45] <helen> i am planning to run it
[01:02:56] <cradek> ok
[01:03:31] <cradek> on one machine I am using with a cvs magma checkout
[01:03:51] <helen> will the .config file from live cd work for
[01:03:54] <cradek> but that checkout was a while ago
[01:04:07] <cradek> not without some changes
[01:04:19] <cradek> but you may be able to start with it to get most of the settings
[01:04:49] <helen> uh..what other settings shall i change? is it documented somewhere?
[01:04:48] <cradek> it's probably unwanted advice, but in your place I'd reconsider whether I really needed the wacom on the same computer that controls the machinery
[01:05:26] <cradek> there is rtai documentation for building everything, but it is really not very complete or good
[01:05:29] <helen> is this usb plugin goin to interfere with the controller?
[01:06:00] <SWPadnos> possibly. many memory devices do
[01:06:06] <SWPadnos> (USB memory sticks)
[01:06:20] <cradek> plugging them in while emc is running the machine is bad on some computers
[01:06:39] <helen> ok
[01:09:13] <helen> what about emc1? is there anyway i could get the source for it?
[01:09:36] <cradek> of course but it will probably not run any better on your bad kernel
[01:09:56] <cradek> http://cvs.linuxcnc.org/cvs/emc/
[01:10:01] <cradek> ^ emc1 source
[01:10:59] <helen> how can i find out which kernels will be good?
[01:11:21] <cradek> I don't know a good answer to that
[01:11:25] <cradek> there may be some reports on the rtai mailing list
[01:11:38] <cradek> or, like I said works for me on one machine
[01:11:38] <helen> ok thank you!
[01:12:07] <cradek> good luck, hope I helped more than hurt
[01:12:09] <helen> mayb i should try
[01:12:15] <cradek> it's a real pain sometimes
[01:12:49] <cradek> another idea: you could try rebuilding our kernel with whatever changed driver parameter you need
[01:12:55] <cradek> (if the driver in that version worked)
[01:36:17] <skunkworks> jepler: well?
[01:37:22] <helen> hey sorry i was just researchin on my options
[01:38:05] <jepler> skunkworks: well it runs but it's noisy and doesn't achieve the speed claimed by the seller
[01:38:27] <jepler> the screws and rails are dry and I don't have anything to lubricate them with .. so anything more is postponed until tomorrow.
[01:39:00] <jepler> unless you think I should try canola oil
[01:39:05] <SWPadnos> heh
[01:41:05] <skunkworks> heh
[01:46:47] <skunkworks> jepler: what do you think about the quality so far?
[01:48:39] <jepler> skunkworks: I haven't cut anything, or even mounted the dremel yet
[01:48:43] <jepler> so it's too soon to say
[01:49:14] <jepler> if after lubricating the screws and rails it's still so screechy and stall-prone at 50% of the claimed speed, I will fall firmly into the "disappointed" camp
[01:49:56] <skunkworks> jepler: getting 7.5ips out of 30000 base_thread. Again - nice work.
[01:49:59] <jepler> though I have to admit I doubted the claim of 70ipm even as I was completing the auction...
[01:50:19] <skunkworks> I am sure it is tweekable..
[01:50:50] <skunkworks> but for most things even 40ipm is good. Well - if the accelleration is fast enough
[01:51:12] <jepler> I'm still using the default 20 in/s/s acceleration from stepper_inch
[01:51:37] <jepler> the stalls I have been seeing were during cruise, I don't think it's acceleration set too high
[02:11:08] <jepler> 'night all
[02:13:07] <skunkworks> night jepler
[02:13:18] <Jymmm> jepler:
[02:13:31] <Jymmm> really, dont run xylotex w/o at least ONE fan.
[04:25:08] <fenn> bleh.. why am i wasting my time on cnczone
[04:25:41] <SWPadnos> because someone has to educate the unwashed masses
[04:25:53] <fenn> at least now i have a long thorough explanation i can point the unwashed masses to so they will stop bothering me
[04:25:58] <SWPadnos> heh
[04:26:03] <SWPadnos> and tll them to take a bath too
[04:26:58] <fenn> more or less correct? http://cnczone.com/forums/showthread.php?t=37309&page=2
[04:29:02] <SWPadnos> heh - looks kinda familiar :)
[04:29:03] <fenn> ya
[04:29:22] <fenn> i wonder why this question comes up so much lately
[04:29:40] <fenn> like was there some sale at radioshack for stepper with encoders on them or something?
[04:30:12] <SWPadnos> Mariss has mentioned the closed loop stepper thing several times over the last couple of months, because he's made a lot of progress on the new drive
[04:30:22] <SWPadnos> I'm sure that has helped
[04:30:42] <SWPadnos> also, there are people selling cards that make Mach "closed loop"
[04:30:56] <fenn> yeah galil :P
[04:31:00] <SWPadnos> (ie, the DRO becomes "actual" and there's a traveling following error)
[04:31:13] <SWPadnos> well, that works of course
[04:31:44] <SWPadnos> so people talk about closed loop, and they think the steppers will be just like servos once they stick an encoder on the back
[04:31:53] <SWPadnos> and they don't know what they're talking about
[04:32:42] <fenn> hey why does a stepper's torque "corner"?
[04:32:47] <fenn> why is it flat at all?
[04:34:50] <cradek> hi cmorley, did you get your cvs working?
[04:34:57] <SWPadnos> read the whitepaper - it explains that very well
[04:34:57] <SWPadnos> (in short, it's because theory and practice are rarely equal)
[04:34:59] <SWPadnos> that explanation is pretty good. I don't think a motor engineer would agree with all points, but it's close enough for the masses
[04:35:18] <C_Morley> no i'm here for help. I did do a developers check out using cmorley
[04:35:42] <SWPadnos> diff and patch are your friends :)
[04:35:47] <cradek> cool, then you're most of the way there
[04:35:58] <C_Morley> it's commiting thats the proplem
[04:37:24] <C_Morley> could not open lock file `/cvs/emc2/src/hal/classicladder/,classicladder.c,': Permission denied
[04:37:48] <C_Morley> should that be emc2-trunk/src/hal etc??
[04:37:58] <cradek> are you sure that's in your devel checkout?
[04:38:02] <SWPadnos> did you copy your modified files into the new checkout dir?
[04:38:09] <C_Morley> yes
[04:38:16] <cradek> maybe you accidentally copied CVS/* too
[04:38:21] <cradek> ?
[04:38:20] <SWPadnos> when you did that, did you use a recursive copy?
[04:38:31] <C_Morley> ??
[04:38:37] <SWPadnos> cp -r
[04:39:03] <SWPadnos> or is it cp -R ?
[04:39:05] <cradek> ack, I have to run, thanks for helping SWPadnos
[04:39:12] <SWPadnos> sure, though I need to run as well :)
[04:39:19] <C_Morley> no used file browser
[04:39:43] <SWPadnos> ok, did you drag your classicladder dir over to the new checkout, or did you select files and drag them?
[04:40:00] <C_Morley> i'm over my head a little here so bear with me please
[04:40:25] <C_Morley> right click copy. then paste
[04:40:37] <Jymmm> * Jymmm hands C_Morley a Picinic Basket
[04:40:50] <fenn> so a stepper's torque curve is flat because of current limiting in the drive. bah
[04:40:51] <SWPadnos> if you selected the CVS folder from your modified source directory, then you overwrote the CVS folder in the new checkout with the anonymous checkout information
[04:41:18] <SWPadnos> CVS keeps information in the CVS/ directory about who did the checkout and what files are in the repository
[04:41:25] <C_Morley> oh i see . i think i under stand!
[04:41:39] <C_Morley> let me try again. thanks for help!
[04:41:42] <SWPadnos> sure
[04:43:35] <Jymmm> fenn: Very nice writeup.
[04:46:52] <fenn> it was a nice braindump session :)
[04:49:23] <Jymmm> fenn: I think you need to add 'whatsits' to M-W dictionary though.
[04:49:42] <fenn> its actually a planck unit of information
[04:49:57] <Jymmm> Ah, so it's already there =)
[04:50:09] <fenn> no i'm just yanking your chain
[04:50:40] <Jymmm> fenn: I was being serious, the word fits in the style of writing; That's usually how I write up white papers.
[04:50:53] <fenn> i hate unitless numbers, i dont believe in them
[04:51:19] <Jymmm> fenn: Well, I don't usually make up words along the way, but in this case, it fits.
[04:51:25] <Jymmm> =)
[04:53:19] <Jymmm> fenn: Anyhow.... nice job.
[05:06:58] <CIA-24> 03cmorley 07TRUNK * 10emc2/src/hal/classicladder/classicladder.c: Removed windows code and options not available for HAL version -easier to understand, maintain
[05:10:38] <CIA-24> 03cmorley 07TRUNK * 10emc2/src/hal/classicladder/classicladder.h: changes for creation of signed hal pins
[05:12:27] <CIA-24> 03cmorley 07TRUNK * 10emc2/src/hal/classicladder/classicladder_gtk.c: changes for creation of seperate windows for variables
[05:12:42] <CIA-24> 03cmorley 07TRUNK * 10emc2/src/hal/classicladder/classicladder_gtk.h: changes for creation of seperate windows for variables
[05:14:16] <CIA-24> 03cmorley 07TRUNK * 10emc2/src/hal/classicladder/config_gtk.c: changes for display of number of signed hal pins
[05:16:11] <CIA-24> 03cmorley 07TRUNK * 10emc2/src/hal/classicladder/edit.c: changes for handling the use of symbols for variables
[05:16:56] <CIA-24> 03cmorley 07TRUNK * 10emc2/src/hal/classicladder/edit_gtk.c: changes for tool tips to describe buttons
[05:18:39] <CIA-24> 03cmorley 07TRUNK * 10emc2/src/hal/classicladder/files.c: changes for auto assignment of symbol names
[05:18:55] <CIA-24> 03cmorley 07TRUNK * 10emc2/src/hal/classicladder/files.h: changes for auto assignment of symbol names
[05:21:59] <CIA-24> 03cmorley 07TRUNK * 10emc2/src/hal/classicladder/module_hal.c: changes for adding signed integer pins
[05:23:52] <CIA-24> 03cmorley 07TRUNK * 10emc2/src/hal/classicladder/symbols_gtk.c: changes for window that toggles show/hide
[05:24:51] <CIA-24> 03cmorley 07TRUNK * 10emc2/src/hal/classicladder/arrays.c: changes for adding signed integer pins
[05:58:10] <CIA-24> 03cmorley 07TRUNK * 10emc2/src/hal/classicladder/files.c: changed auto assignment S32 pin name from HALinS to CLinS
[07:10:18] <skinnypuppy1334_> Anyone around, I'm trying to install GCAM and get an error
[07:56:47] <skinnypuppy1334_> I'm trying to install GCAM and get an error checking for GTK+ - version >= 2.8.0... no
[07:56:48] <skinnypuppy1334_> *** Could not run GTK+ test program, checking why...
[07:56:49] <skinnypuppy1334_> *** The test program failed to compile or link. See the file config.log for the
[07:56:49] <skinnypuppy1334_> *** exact error that occured. This usually means GTK+ is incorrectly installed.
[07:56:49] <skinnypuppy1334_> configure: error: GTK+ 2.8 or higher is required
[07:57:55] <skinnypuppy1334_> I don't see GTK + in synaptic
[08:04:50] <Vq^> it has to be there
[08:04:59] <Vq^> synaptic uses gtk+ itself
[08:05:50] <Vq^> my best guess is that you lack the -dev packages
[08:09:42] <skinnypuppy1334_> I see gtk engines and gtkhtml
[08:11:21] <skinnypuppy1334_> there is a gtk-engines-gtk-qt that is unchecked
[08:24:17] <Vq^> it seems to be called libgtk2.0-dev on debian
[11:40:50] <cradek_> cradek_ is now known as cradek
[12:05:21] <jepler> cradek: there's no "xwd" on yggdrasil? an xwd of root should fit on a floppy..
[12:06:06] <JymmmEMC> on yggdrasil?
[12:06:07] <jepler> (re http://timeguy.com/cradek/01188356148)
[12:10:33] <skunkworks> do the little eyes follow the mouse?
[12:11:33] <jepler> skunkworks: you can install it on your modern system -- package 'xeyes'
[12:11:38] <jepler> yes they follow the mouse
[12:11:42] <skunkworks> heh
[12:12:43] <jepler> * jepler runs it just for fun
[12:12:52] <skunkworks> No
[12:12:55] <skunkworks> oops
[12:12:58] <skunkworks> I mean
[12:13:06] <skunkworks> No 'task bar'?
[12:13:33] <skunkworks> * skunkworks wonders what happems when you minimize a window
[12:13:40] <jepler> the thing at the upper left in back seems to be the "twm icon menu" and has entries for each window
[12:13:56] <skunkworks> ok - I get it.
[12:14:24] <skunkworks> jepler: did they recomend a lubricant for your router?
[12:14:54] <jepler> skunkworks: yes.
[12:14:55] <jepler> Yeah, I usually put lube on the rails and screws before shipping, but
[12:14:56] <jepler> I may have forgotten on yours. A light oil like tri-flow, or even
[12:14:56] <jepler> wd-40 works well. As does motor oil. The machine should run smooth
[12:14:55] <jepler> and quiet, let me know if oil does not solve the problem.
[12:15:34] <skunkworks> That is good - he seems to be responsive :)
[12:15:39] <jepler> yeah
[12:19:11] <skunkworks> stupid question - why is it called emacs?
[12:20:17] <JymmmEMC> macs was already taken
[12:22:55] <skunkworks> heh - but macs was short for machentosh
[12:23:00] <skunkworks> oops
[12:23:12] <skunkworks> how ever you spell that.
[12:23:40] <JymmmEMC> macintosh
[12:26:32] <skunkworks> morning JymmmEMC
[12:32:53] <cradek> jepler: I also took pictures of the deluxe tui installer, but I didn't post those
[12:33:56] <cradek> jepler: "now is when you ought to make your filesystems with mkefs:"
[12:34:06] <cradek> bbl
[12:42:05] <jepler> skunkworks: "The original EMACS was a set of Editor MACroS for the TECO editor" according to wikipedia
[12:44:27] <jepler> huh, rms's early ideas on what we now call "software licensing" had the same bad ideas people still have today.
[12:44:29] <jepler> Stallman realized the danger of too much customization and de-facto forking and set certain conditions for usage. He later wrote:
[12:44:34] <jepler> "EMACS was distributed on a basis of communal sharing, which means all improvements must be given back to me to be incorporated and distributed."
[12:49:31] <skunkworks> jepler: thanks
[13:08:26] <cradek> jepler: interesting
[13:17:18] <skunkworks> fenn: nice reply on cnczone.
[14:27:40] <CIA-24> 03jepler 07v2_1_branch * 10emc2/debian/changelog: note bugs fixed
[14:45:11] <tomp> cradek: i got a few versions of the 7+ cd yggsadril systems, want 'em?
[14:45:25] <cradek> tomp: nope :-)
[14:45:32] <SWPadnos> CDs - I want floppies dammit!
[14:45:39] <tomp> 8"
[14:45:47] <SWPadnos> yep
[14:46:02] <tomp> wish i had some of those steeper/screw assy's
[14:46:02] <SWPadnos> "this 4000 disk set has all the tools you need"
[14:46:26] <SWPadnos> hmmm. 4000 of those 8" floppies wouldn't even fill a CD
[14:46:40] <tomp> cpm format
[14:46:57] <SWPadnos> did those ever get to 1.2M?
[14:47:02] <SWPadnos> th e8" disks
[14:47:05] <tomp> no
[14:47:19] <SWPadnos> ok - I thought they topped out at 180k or 360k or thereabouts
[14:56:39] <skunkworks> we had a surplus outlet store that had some boxes of 8 inch disks. I was always going to buy one.
[15:01:52] <cradek> what's a good basic lube for plastic (delrin? nylon?) against drill rod?
[15:02:53] <SWPadnos> delrin shouldn't need much lube - it's meant to be a bearing surface
[15:03:02] <cradek> right
[15:03:22] <cradek> I don't know if it's delrin...
[15:03:28] <cradek> but even if so, a drop of oil seems like it couldn't hurt
[15:03:36] <SWPadnos> probably not
[15:04:03] <jepler> HDPE
[15:04:05] <SWPadnos> I'm sure there's someone who knows things like "never put WD-40 on deltin - it dissolves"
[15:04:10] <SWPadnos> delrin
[15:04:15] <SWPadnos> but I'm not that person ;)
[15:04:16] <archivist> use silicon based iirc
[15:06:11] <SWPadnos> hmmm. here's something you don't see in an IC datasheet every day: "Note 10: Exposing the micro SMD device to direct sunlight will cause misoperation. See Application Hints for additional information."
[15:06:52] <cradek> http://www.amazon.com/Slick-50-One-Lube-Spray/dp/B0002JN2IG
[15:07:05] <SWPadnos> 10W-30
[15:07:09] <cradek> I have a bottle of this stuff and it seems to work well, but I don't know what it is (or where it came from)
[15:07:25] <SWPadnos> I thought that was one of the silicon additives for engine oil
[15:07:36] <SWPadnos> or teflon or something
[15:07:36] <cradek> you might be joking, but I use 10W-30 for all sorts of stuff
[15:07:48] <cradek> no, it's a spray bottle meant for general purpose lubricating
[15:08:04] <SWPadnos> not really joking, but it depends on the application and the expected lifespan
[15:08:09] <SWPadnos> ah, ok
[15:08:36] <cradek> (I even use motor oil for tapping sometimes)
[15:08:45] <SWPadnos> http://www.slick50.com/Product_Pages/Products.html
[15:08:54] <skunkworks> be to :)
[15:09:00] <skunkworks> Me too that is.
[15:28:35] <jepler> I had better not search for silicone lubricant at work... those text ads are not what I want on my screen at the office
[15:28:42] <jepler> (thanks for that google)
[15:29:01] <cradek> hmm
[16:02:18] <skunkworks> jepler: maybe? http://www.amsoil.com/storefront/als.aspx
[16:02:46] <jepler> sounds promising
[19:02:02] <skunkworks> I would think any farm and barn store would have a ton of different lubricants.
[19:02:58] <skunkworks> or hardware
[19:03:16] <SWPadnos> or auto
[19:03:26] <SWPadnos> auto parts anyway
[19:05:23] <skunkworks> so - with the new doublefreq - a decently fast computer should approch 100khz step rate.
[19:05:44] <SWPadnos> it would still have to be a pretty fast computer - that needs a 10000 BASE_PERIOD
[19:05:51] <skunkworks> right
[19:05:59] <SWPadnos> but 60-80 KHz would be pretty common, I think
[19:06:06] <skunkworks> yes
[19:06:10] <SWPadnos> ~16000 to 12000 BASE_PERIOD
[19:06:28] <skunkworks> approach sounds better.. ;)
[19:06:37] <SWPadnos> heh
[19:06:56] <SWPadnos> like the price is right - get the closest without going over (or the computer hard locks :) )
[19:07:45] <jepler> or it starts printing "Unexpected realtime delay on task 1"
[19:07:56] <skunkworks> heh
[19:08:04] <SWPadnos> does RTAPI do that or is it motion (or whichever)?
[19:08:21] <jepler> there's a newish one in rtai_rtapi.c
[19:08:26] <SWPadnos> ah
[19:08:34] <skunkworks> it is on you tube now - http://www.youtube.com/watch?v=a_OUNLruaVQ in case you want to show your significant others clergy or such
[19:08:49] <SWPadnos> I wonder if it's a good idea to do that in a fast thread (kprint anything)
[19:08:50] <jepler> rt_task_wait_period() can return RT_TMROVRN
[19:09:07] <jepler> it's rtapi_print_msg()'d at most 10 times per session, total for all threads
[19:09:15] <SWPadnos> ok
[19:09:38] <jepler> it shows up *much* more readily than the motion 'realtime delay' message
[19:10:05] <SWPadnos> sure - it's got a threshold of one period
[19:10:18] <cradek> what does it mean?
[19:10:38] <skunkworks> If I really screw up - I get both..
[19:10:46] <SWPadnos> err - maybe not. without looking at the code, it seems like it would trip any time the period is exceeded
[19:10:58] <SWPadnos> rather than +0.1 or +0.2x the period
[19:12:11] <jepler> I don't know what it means, except that an error return from rt_task_wait_period is what the rtai kern/latency test calls an "overrun"
[19:13:23] <SWPadnos> someone remind me next week that I want to add some statistical stuff to the RTAI kernel latency test
[19:24:56] <skunkworks> SWPadnos: I put a reminder in my phone for you.
[19:25:18] <SWPadnos> ok, thanks
[19:26:06] <jepler> SWPadnos: with the HAL 'timedelta' component plus halsampler you should be able to gather the information you want and perform userspace analysis on it
[19:26:17] <jepler> (timedelta is a recentish addition on trunk)
[19:26:28] <skunkworks> * skunkworks was actually just being a smart ass
[19:26:35] <skunkworks> like usual
[19:26:35] <SWPadnos> jepler, that's probably true. I noticed it when you posted a screenshot of the GUI front end for it
[19:26:48] <SWPadnos> skunkworks, bad skunkworks
[19:26:54] <skinnypuppy1334_> Good day all, I'm trying to get GCAM installed on my garrage machine I put together last night. This is the error I get, configure: error: GTK GL Extension is required
[19:27:30] <SWPadnos> skinnypuppy1334_, twingy is the autho of gcam, he's in here a lot, you could ask him about it
[19:27:34] <SWPadnos> author
[19:27:51] <jepler> skinnypuppy1334_: did you try using synaptic ("package manager") to search for the package you need to install?
[19:27:53] <skinnypuppy1334_> Twingy, cool I'll be on the lookout
[19:28:04] <skinnypuppy1334_> jepler didn't see it in synaptic
[19:28:38] <skinnypuppy1334_> I remember GL being in some other package from playing with beryl months ago..
[19:29:42] <cradek> % apt-cache search GTK GL Extension
[19:29:47] <cradek> libgtkglext1-dev - OpenGL Extension to GTK (development files)
[19:32:00] <jepler> that package is in the "universe" repository; you may need to follow these instructions to enable it if you haven't already: https://help.ubuntu.com/community/Repositories/Ubuntu#head-5bbef89639d9a7d93fe38f6356dc17847d373096
[19:33:18] <cradek> hi jmk
[19:33:18] <skinnypuppy1334_> cradek jepler, thanks I'll poke at it and see what I can get acomplished
[19:33:23] <cradek> good luck
[19:56:15] <skinnypuppy1334_> Seems to have fixed the dependency, "/configure --prefix=/your/installation/path" finished and make; make install was very busy and seems to have installed. How do you launch gcam? I typed gcam in terminal , command not found.
[19:57:10] <jepler> did you literally write "--prefix=/your/installation/path"?
[19:57:20] <skinnypuppy1334_> no I used /home/user/gcam
[19:57:35] <jepler> OK
[19:57:44] <skinnypuppy1334_> The source files were in /home/user/tmp/gcam2007
[19:58:23] <jepler> that probably installed the executable file in /home/user/gcam/bin; you either have to change the PATH environment variable so that directory is listed, or give a full path when starting it: /home/user/gcam/bin/gcam
[19:58:49] <skinnypuppy1334_> I'll hop in the bin folder quick and see
[20:00:34] <skinnypuppy1334_> Strange, I see gcam in the bin directory in terminal but still command not found
[20:01:11] <skinnypuppy1334_> Terminal in the /home/user/gcam/bin directory
[20:03:03] <skinnypuppy1334_> WOO HOO full path at terminal works" /home/user/gcam/bin/gcam" how do I add this into the path? Remembering config.sys
[20:03:29] <jepler> I encourage you to find the answer to that yourself
[20:04:03] <jepler> some promising links come up when I search for 'ubuntu path' on google...
[20:04:10] <skinnypuppy1334_> jep, sure thing... thank all you guys again. I'm learning.
[20:27:06] <jepler> cradek: what is the name of that kind of 4-pin connector you were suggesting?
[20:28:09] <cradek> jepler: I don't know that it has a name, but it's found on CB microphones
[20:35:45] <SWPadnos> 4-pin XLR, or the screw-on type?
[20:47:50] <jepler> the screw type
[20:48:16] <SWPadnos> I haven't seen any current specs on them, but they look like very nice connetors for small to medium steppers
[20:48:24] <SWPadnos> and inexpensive - $3-$4 per end
[20:49:09] <jepler> indeed
[21:01:31] <skunkworks> this is in the intagrator manual linkpp parport.0.pin-01-in iocontrol.0.emc-enable-in
[21:02:08] <skunkworks> pin 1 in the normal setup of the printer port is an output - correct? I think that may confuse people.
[21:09:41] <SWPadnos> plus linkpp is deprecated
[21:10:35] <skunkworks> well - yes :)
[23:08:04] <fenn> egad. 18kW servos's