#emc | Logs for 2008-05-04

[00:10:20] <lerman> jmkasunich: Are you still here?
[00:15:05] <jmkasunich> yeah
[00:15:13] <jmkasunich> (on and off)
[01:04:21] <lerman> jmkasunich: Is now "on"?
[01:12:30] <jmkasunich> YES
[01:12:46] <lerman> I was just looking at your web page...
[01:13:15] <jmkasunich> which part?
[01:13:28] <lerman> and was amazed at what you could do with a few operators and owords. Just wanted to tell you what a nice job that fusee code was.
[01:13:31] <jmkasunich> fun with fire, or fun with o-words
[01:13:52] <jmkasunich> thanks for writing o-words
[01:13:58] <jmkasunich> g-code is a real language now
[01:14:11] <lerman> It's great to see them used.
[01:14:30] <jmkasunich> I use them for probably half the programs I write
[01:14:33] <lerman> With all of the potential for hidden bugs.
[01:15:25] <jmkasunich> so far I haven't seen g-code do something I'd consider a bug
[01:16:01] <lerman> You might consider putting that code up on the wiki. On don't think people realize how powerful it is.
[01:16:12] <lerman> On -> I
[01:16:37] <jmkasunich> feel free to post it (or a link to it) on the wiki
[01:16:48] <lerman> I find that Axis is a big help in debugging code.
[01:17:04] <jmkasunich> there are permalinks for each post in my blog, so you can get to that one
[01:17:12] <lerman> I'll do that when I get a chance. Gotta run.
[01:17:17] <jmkasunich> yep - I'd be terrified to use o-words without the preview
[01:17:54] <jmkasunich> I said I've never seen them do something wrong - thats not the same as seeing them do something I didn't expect
[01:18:02] <jmkasunich> but like most programming, they were doing what I told them to do
[01:20:03] <lerman> One of the lisp environments had a feature called DWIM. (do what I mean). That's what it attempted to do.
[01:27:10] <tom1> mayor richard j daley of chicago's most famous quote. "Write what I mean, not what I say."
[01:43:39] <tom1> is the test Hardy Heron EMC using this rt? https://wiki.ubuntu.com/RealTime/Hardy
[01:44:32] <tom1> the developers are using it to do scada & pick&place
[01:53:36] <jepler> tom1: no. that is a different realtime system which doesn't seem to be good enough to do step waveform production. here's a blog entry I wrote about it: http://axis.unpy.net/01190912545
[01:55:26] <tom1> thx, this is very interesting, that the servo loop control is ok but the step generation is not.
[01:56:09] <tom1> maybe the faster thread is not ok and the slower thread is ok?
[01:56:42] <tom1> and servo is less demanding than stepper. ( gotta study)
[02:06:03] <cradek> more diesel questions - looks like the output of the turbo goes to a kind of "air radiator" (?) in front of the real radiator. why?
[02:18:38] <jtr> The intake air heats up when it's compressed (and probably picks up some heat just passing through the turbo). Extracting the heat increases the density of the intake air charge.
[02:19:01] <jtr> So you get more from the turbo...
[02:19:21] <cradek> ok thanks
[02:19:43] <cradek> I suspect my turbo might not be working. I will try to put a boost gauge on the output. there is an obvious place to hook it up.
[02:20:24] <jtr> is this the generator engine you were working on?
[02:20:33] <cradek> no it's the bus's main engine
[02:20:37] <cradek> no turbo on the generator
[02:23:49] <jtr> How did the generator turn out? I haven't had the opportunity to work on a diesel.
[02:23:58] <cradek> it runs great
[02:24:26] <cradek> I took apart the injector pump, fixed some obvious problems, and put it back together and reinstalled
[02:24:43] <cradek> sounds simple but it was a lot of finicky work
[02:27:20] <jmkasunich> hi cradek
[02:27:24] <cradek> hi
[02:27:29] <jtr> Are you referring to the tight clearances or picky adjustments?
[02:27:40] <jmkasunich> "intercooler" would the term for that air radiator
[02:28:21] <cradek> jmkasunich: ah I figured that was it (google told me so)
[03:36:47] <lewin1> lewin1 is now known as lewing_
[03:48:02] <lewin1> lewin1 is now known as lewing_
[04:32:48] <K`zan> Not sure that ethernet cable is working out. Just got it in and there seems to a lot of buzzing and not quite made steps... Will look into it more tomorrow.
[04:38:07] <tom1> has anyone looked at http://www.ar-cad.com/products/freecad/linux it's not cad as much as a kinematics simulator.
[05:30:35] <SkinnYPup> tom1: Got gcnccam working well. Have had problems with qcad arcs being off by .0006 and emc don't like that
[05:32:28] <SkinnYPup> tom1: I was making some "teardrops" for a small hub and belive it is the qcad trim function that is flubbing the arc start end and centers have any experience with qcad and these type errors ?
[05:51:08] <K`zan> Wondering if adding that 6 feet of 24ga cable to the steppers changed something. When I had the steppers connected directly to the HCNC drive I don't remember chattering or them seeming to miss steps. Just rechecked the wiring and it is correct. Strange.
[05:54:57] <JymmmEMC> did you check the cable itself?
[05:56:11] <K`zan> Just rang it out for continuity.
[06:02:07] <Sweeper> does the cable get warm when you run the machine?
[06:13:38] <K`zan> Wasn't on that long but the steppers get warm (not hot) quickly.
[06:25:11] <K`zan> Running flowsnake and then tux and will see what happens.
[06:43:52] <K`zan> Nope no heat on the cables, but I got some allignment issues to resolve and the Z-Axis is running backward. Will putz with that tomorrow.
[06:48:02] <Gamma-X> Hello everyone,
[07:23:31] <gfixler> anyone know what the reasoning behind large-machine websites not listing the prices for anything, but asking you to call for a quote?
[07:24:14] <gfixler> I can understand it for something like a custom job, where there are so many variable factors, but when you're simply buying a standard machine, like a laser cutter with no options, what's the point?
[07:25:32] <OoBIGeye> gfixler: i think there is a couple of factors, maby currency nost machines run on us dollars
[07:25:51] <OoBIGeye> and the fact that there is no such thing as s standard machine :)
[07:25:59] <gfixler> :)
[07:26:19] <gfixler> well, when you're ordering a particular model of laser cutter, it's going to be the same thing anyone gets when they order it
[07:26:23] <gfixler> it'll be part #something
[07:26:59] <gfixler> my guesses have been that they want to feel you out, and haggle a price, or that they just want to start a conversation with you, so they can talk you into buying from them, instead of having you shop around
[07:28:06] <OoBIGeye> yes ofcourse...
[07:28:29] <OoBIGeye> there are few people that buy machines at stock prces...
[07:28:42] <OoBIGeye> and those are idiots...
[08:01:43] <K`zan> Night folks
[11:16:09] <awallin> hi, I'm trying the rtai_smi.ko fix, but I get an error from insmod "no such device" ??
[11:17:49] <awallin> latency-test shows the delay jumping to 156 ms every now and then...
[11:28:07] <awallin> I need to run now but if anyone has some ideas on why insmod won't install rtai_smi.ko on my system please write here and I will check the logs
[12:50:13] <tom1> gfixler: if the price is not fixed and published, the sellers know they can get more from some people or cut thier price to underbid a competitor.
[12:50:22] <tom1> you must bear this in mind and be ready to fight for a good price.
[12:50:23] <tom1> the trick is to know the basic price which is usually old-boy networking.
[13:24:37] <fenn> just say 'well competitor XYZ is offering theirs for $(bid * 0.75) why is yours so high?'
[13:25:38] <rayh> 'cuz it's a better machine!
[13:37:38] <alex_joni> awallin: you need to have an intel chipset to load the rtai_smi module
[13:38:21] <alex_joni> awallin: check dmesg, but if it says no such device, then probably you either don't have one, or you have one which isn't recognized by the rtai_smi module (check the source of the module to see a list of things it knows about)
[14:12:13] <Assargadon> hey there
[14:12:45] <Assargadon> in approx. 1.5 hours I'll begin my HAL modules coding...first time.
[14:13:03] <Assargadon> Is any additional comments or things I should to know?
[14:14:02] <awallin> alex_joni: yep, dmesg says 'intel chipset not found' so my realtime-delay must be caused by something else...
[14:24:36] <alex_joni> awallin: lspci -vvv?
[14:24:52] <awalli1> what does that do?
[14:25:19] <awalli1> now I installed NVIDIA drivers through the envy installer but that seems to have locked up the computer
[14:25:34] <cradek> you don't want the nvidia drivers for a realtime system.
[14:25:45] <alex_joni> * alex_joni agrees
[14:25:57] <cradek> unfortunately you have to use vesa. even nv seems to cause some problems.
[14:25:58] <alex_joni> even the open source driver on NVIDIA hardware sometimes causes problems with RT
[14:26:30] <cradek> I like the old matrox cards the best.
[14:26:37] <alex_joni> intel seems quite fine too
[14:26:50] <awalli1> oh. how do I get rid of the installed NVIDIA driver?
[14:26:51] <alex_joni> especially because they do 3D stuff
[14:26:53] <cradek> yes even integrated i810 seems ok on the few I've tried
[14:27:05] <alex_joni> awalli1: change /etc/X11/xorg.conf
[14:27:11] <alex_joni> look for driver, and put vesa there
[14:27:13] <cradek> what's envy?
[14:27:18] <awalli1> was there some utility for doing that
[14:27:31] <cradek> alex_joni: it's worse than that, since the nvidia crap replaces the GL libraries
[14:27:51] <awalli1> envy was something I found with google, it automates the NVIDIA / ATI driver install
[14:27:51] <alex_joni> bugger.. I usually apt-get remove --purge *nvidia* stuff
[14:28:27] <cradek> awalli1: can you determine whether it used apt?
[14:29:08] <awalli1> I think it might have used apt
[14:29:26] <cradek> then I think you can just remove the apt packages
[14:29:34] <cradek> (but I'm not sure)
[14:32:00] <awalli1> could the realtime-delay be related to networking? previosly I had it connected to the network and did not get any warnings. now it's without a net connection and I get the occasional 150-160 ms delay with latency-test
[14:32:46] <alex_joni> awalli1: doesn't sound very plausible
[14:32:59] <alex_joni> but if you installed nvidia in the meantime, then it's more likely that's the case
[14:33:34] <alex_joni> awalli1: USB sticks & such, sometimes also cause RT delays
[14:36:19] <awalli1> I installed the nvidia drivers only after the problem appeared. now I reconfigured X and hopefully got away with nvidia, using the nv driver
[14:38:18] <jepler> awalli1: I recall at least one other person reporting that networking caused rela-time delays, but I don't have any idea what nic chipset they were using
[14:39:30] <alex_joni> jepler: but I think it was the other way around
[14:39:48] <alex_joni> (when the network was connected, and packages arrived.. maybe unmasked interrupts)
[14:40:05] <awalli1> well now that I reconfigured with sudo dpkg-reconfigure xserver-xorg the machine doesn't freeze anymore but I get an error "GLX extension missing"
[14:40:17] <awalli1> so glxgears and AXIS won't start :(
[14:40:29] <alex_joni> awalli1: you gotta thank NVIDIA for that
[14:40:38] <alex_joni> they replaced the GL libraries
[14:40:59] <jepler> alex_joni: oh, hm -- I didn't read closely enough
[14:41:50] <alex_joni> jepler: maybe I misremember.. so anyways awalli1 should try both ways (network connected and disconnected)
[14:42:17] <alex_joni> awalli1: try reinstalling the original libglu mesa (forgot the name)
[14:42:23] <awalli1> has realtime-delay reporting changed between 2.0.x and 2.2.4 ? I didn't get those with the old 2.0.x controlling a stepper machine. This is now 2.2.4 and an m5i20 / servo-machine
[14:42:40] <alex_joni> yes, quite a bit
[14:42:57] <alex_joni> 2.2.x is much more eager to report realtime errors
[14:43:06] <alex_joni> in 2.0.x tests were lots more limited
[14:53:25] <awalli1> hmh, now with the nv driver I don't have glx but do get th 156 ms delay after looking at latency-test for a while
[14:54:03] <awalli1> with the nvidia drivers the whole machine hangs quite a bit but sometime I get glxgears running in fullscreen with much higher update rate and smoother than with nv drivers
[14:54:04] <awalli1> go figure
[14:55:49] <tom1> rayh fenn that 75% or less is why japan outsells us & mid-europe machine. prepare for the indian machine tool invasion
[14:56:07] <tom1> and cars
[14:56:27] <rayh> I saw their $2500 car.
[14:56:57] <JymmmEMC> rayh: in person?
[14:57:11] <rayh> nah just a bunch of pics.
[14:57:24] <JymmmEMC> k
[14:57:40] <alex_joni> awalli1: that is expected
[14:57:53] <alex_joni> you get crappy 3D performance with the nv driver or with vesa
[14:58:43] <awalli1> ok, trying vesa next.
[14:58:49] <awalli1> or even the onboard video card
[14:58:52] <alex_joni> vesa will be crappier
[14:58:58] <alex_joni> performance wise
[14:59:03] <alex_joni> but better RT performance
[15:07:50] <awalli1> now with vesa I don't have GLX, so can't run glxgears or AXIS, but I can run latency-test and it again shows the 156 ms delay. it seems to come whenever starting big apps like OpenOffice or similar
[15:08:03] <awalli1> is there something similar like glxgears that I could load up the machine with?
[15:08:38] <JymmmEMC> SHA1 every file on your machine =)
[15:08:52] <JymmmEMC> or SHA256 even
[15:13:17] <awalli1> well, it doesn't seem to be related to cpu-load directly. running some filters in GIMP does put the cpu load at 100% for several 10s of seconds, but doesnt affect latency-test
[15:13:58] <gefink> awalli1>: have you /usr/lib/Mesa/demos/gears
[15:15:20] <alex_joni> gefink: I doubt that works .. if glxgears doesn't
[15:16:08] <gefink> awalli1: try hdparm -t /dev/sda to test ide-interrupts and dma
[15:16:30] <gefink> alex_joni i think it dont use glx
[15:17:03] <alex_joni> hmm.. I don't have that here..
[15:18:07] <gefink> alex_joni some other "gears" ?
[15:18:21] <alex_joni> no /usr/lib/Mesa at all
[15:18:31] <gefink> oh
[15:19:53] <gefink> have you x11perf ?
[15:19:57] <awalli1> gefink: hdparm -t gives around 72 Mb/s this is a SATA drive
[15:20:18] <gefink> awalli1: affekting latency?
[15:20:19] <awalli1> * awalli1 has around 20min on laptop power left...
[15:20:31] <awalli1> gefink: no, didn't cause latency to rise
[15:20:46] <awalli1> at least not the 156ms delays I see once in a while
[15:21:08] <awalli1> the jitter does go up depending on how much I load the cpu but that is normal I guess
[15:21:57] <gefink> aw: you say notebook. goes your harddisk sleeping?
[15:24:02] <awalli1> gefink: this irc connection is from the laptop and via 3G phone, the mill and emc2 runs on a desktop pc
[15:24:23] <gefink> awalli1: ok
[15:24:39] <Assargadon> Is EMC take part in GSoC this year?
[15:24:44] <alex_joni> Assargadon: no
[15:25:16] <Assargadon> alex_joni, why?
[15:25:52] <alex_joni> Assargadon: because nobody enrolled emc to GSoC
[15:26:23] <awalli1> now I timed the delay between 156ms latency, seems to be around 10min. something in cron that runs every 10min?
[15:27:52] <gefink> have you checked SMI ? Iintels service management interrupt?
[15:27:57] <gefink> ist it a intel CPU?
[15:28:37] <Assargadon> alex_joni, if I'll make application, is it will be ok?
[15:28:56] <alex_joni> Assargadon: this years GSoC is long closed for applications
[15:30:19] <Assargadon> ou
[15:30:45] <Assargadon> I miss "March" and "May"
[15:37:51] <alex_joni> Assargadon: there are a couple of things needed for emc to participate in GSoC
[15:38:12] <alex_joni> including people willing to act as mentors
[15:40:14] <Assargadon> there are lot of people mentored me :) why not mentor some developer?
[16:36:54] <awallin> so any ideas on the 156ms real-time delay which repeats every 10 minutes?
[16:37:40] <gefink> awallin: intels System management interrupt (SMI) ??
[16:38:44] <awallin> gefink: the rtai_smi.ko module would not load as it didn't recognize an intel chipset
[16:38:52] <awallin> insmod: device not found
[16:39:15] <gefink> awallin: no more ideas
[16:39:21] <awallin> although this is an intel MB with a celeron cpu
[16:39:33] <JymmmEMC> What's a decent s/ftp damon ?
[16:39:40] <alex_joni> awallin: what chipset?
[16:39:55] <alex_joni> JymmmEMC: I used proftpd in the past
[16:40:02] <JymmmEMC> alex_joni: sftp?
[16:40:06] <JymmmEMC> diable ftp
[16:40:20] <awallin> alex_joni: don't know, it says Intel desktop board D102ggc2
[16:40:28] <alex_joni> awallin: do a lspci -vvv
[16:40:30] <gefink> disable anacron and syslog ?
[16:40:32] <alex_joni> and you'll know
[16:40:49] <gefink> cron?
[16:42:21] <gefink> awallin lspci gives info about host and PCI bridge
[16:44:38] <awallin> I will do a re-install now as the OpenGL went away with the NVIDIA drivers
[16:45:01] <awallin> is the new Ubuntu OK for a 2.4GHz machine or should I stick with 6.06?
[16:52:17] <awallin> is there a simple install-script for Ubuntu 8.04?
[16:52:33] <JymmmEMC> ?
[16:52:48] <JymmmEMC> you want yo upgrade instead of doing a clean install?
[16:53:02] <awallin> I want to do a clean install, something like http://linuxcnc.org/dapper/emc2-install.sh
[16:53:19] <awallin> or is the liveCD better?
[16:53:37] <gefink> www.linuxcnc.org/hardy/emc2-install.sh
[16:53:45] <JymmmEMC> dont know if better, but simplist it seems
[16:54:14] <gefink> you need internet on the box
[16:54:38] <awallin> gefink: yeah, I got everytihng to the lab again so I have a net connection.
[16:59:24] <alex_joni> awallin: both should have the same outcome
[16:59:37] <alex_joni> with the difference that the stock hardy CD will have lrm in it, and might install it
[16:59:51] <alex_joni> (lrm = linux-restricted-modules)
[17:00:16] <awallin> alex_joni: ok, I'm burning a 8.04 cd right now. also checking if the 10min 'ghost' shows up when on a network connection
[17:29:57] <kirk_wallace> OT: I am installing the LiveCD to a Netvista. The AGP slot had a filler plug in the slot and a card in the slot doesn't get recognized. Are there some motherboards with AGP slots that where not intended to function?
[17:30:24] <eric_U> could be the bios doesn't set it up
[17:31:56] <kirk_wallace> The menu has a defalt video setting for "internal" or "PCI". I wonder if a BIOS update could fix it?
[17:35:09] <rayh> I wonder why it had a plug in it.
[17:36:51] <kirk_wallace> I was thinking, to keep someone from using it.
[17:40:31] <rayh> Yea probably.
[17:40:47] <eric_U> 99.9999999999% of the time, the entire connector would be missing in that case
[17:41:03] <eric_U> your system is the only thing keeping that from being 100%
[17:42:37] <awallin> hm, I only get a "Starting up ..." when booting the 8.04 rtai kernel for the first time??
[17:42:45] <kirk_wallace> I looked on Lenovo for a BIOS and the revision history mentions an AGP fix for something. I'll try the update. Wish me luck.
[17:42:59] <alex_joni> awallin: it hangs there forever?
[17:43:13] <awallin> alex_joni: if forever==1-2 minutes
[17:43:27] <alex_joni> hmm.. try passing "irqpoll" as a boot param
[17:43:35] <alex_joni> is this using the livecd or installed?
[17:43:39] <awallin> can I do that via grub?
[17:43:42] <alex_joni> yes
[17:43:43] <awallin> it's installed
[17:43:53] <alex_joni> F6 to change boot params
[17:43:57] <awallin> ok
[17:44:04] <alex_joni> or edit /boot/grub/menu.lst
[17:44:37] <awallin> f6 doesn't do anything?, e to edit commands before booting?
[17:44:47] <alex_joni> F6 is for the LiveCD to change.
[17:44:48] <alex_joni> use e
[17:45:09] <awallin> there are four lines: root, kernle, initrd, quiet
[17:45:14] <awallin> kernel
[17:45:31] <alex_joni> remove quiet
[17:45:34] <alex_joni> and put irqpoll
[17:46:01] <awallin> would that be 'irqpoll' as a new line, or as the last item on the kernel line?
[17:46:17] <alex_joni> awallin: I don't know.. just try
[17:46:34] <alex_joni> I usually do it on the LiveCD, and there everything is on one line
[17:46:51] <awallin> ok, here we go...
[17:48:47] <awallin> OHCI: BIOS handoff failed (BIOS bug?) fffffff
[17:49:42] <gefink> works USB?
[17:50:11] <awallin> gefink: don't know, but the installer booted fine. this is after installing 8.04 from CD and then running the hardy/emc2-install.sh
[17:51:18] <awallin> so the rtai kernel used with 8.04 is different from the magma one used with 6.06 ?
[17:51:56] <gefink> its an other kernelversion
[17:52:08] <gefink> its a big difference
[17:52:11] <awallin> 0000:00:13.0 OHCI: BIOS handoff failed (BIOS bug ?) ffffffff that's the whole error msg and then it just hangs
[17:52:40] <gefink> oh
[17:53:12] <awallin> the -generic kernel seems to boot OK I think
[17:53:21] <gefink> worked the grub-install?
[17:53:45] <awallin> yes, the 8.04 CD installed grub for me
[17:54:08] <gefink> try switching USB legacy support in the BIOS
[17:54:27] <awallin> a USB memory stick does work with the -generic kernel
[17:55:23] <gefink> you cann look for the correct rootdevice while booting
[17:56:20] <awallin> usb has three settings in the BIOS: All, 2.0 and Legacy. They are all enabled. Was your suggestion to only enable Legacy and disable the rest?
[17:56:33] <SWPadnos> opposite
[17:56:45] <awallin> disable Legacy?
[17:56:48] <SWPadnos> leave 2.0 on (not sure what "all" does), and disable legacy
[17:56:58] <SWPadnos> unless you have a USB keyboard only (no PS2 ports)
[17:57:18] <gefink> disable Legacy if you dont have an usb-keybord
[17:58:05] <awallin> ok, now the rtai kernel boots!
[18:01:34] <awallin> what other than glxgears should I be running while latency-test is running?
[18:01:54] <awallin> could the USB Legacy support have been the source of a real-time delay repeating every 10 min ?
[18:02:43] <awallin> there it is again, 155ms of delay!
[18:03:16] <gefink> x11perf
[18:03:20] <gefink> hmmm
[18:04:03] <gefink> no idea but SMI
[18:05:17] <awallin> well the rtai_smi.ko won't install at least on 6.06, is there one for this rtai kernel?
[18:07:12] <gefink> but in internet i read SMI duration is ~ 236277 not 155000000
[18:07:17] <awallin> the wiki page on SMI says it's only for dapper
[18:08:07] <gefink> have you 155 ms or 155 usec
[18:08:32] <awallin> milliseconds
[18:08:49] <gefink> i think that cant be SMI
[18:09:16] <gefink> that is a real 0ong time 0.15 sec
[18:10:03] <gefink> is a CD in the drive or have you an internal USB-Cardreader
[18:10:27] <gefink> try disable all USB
[18:10:31] <awallin> there's no cd in the drive. There are two usb ports at the front and more at the back.
[18:10:51] <gefink> no internal cardreader
[18:11:02] <awallin> Now I'm trying the kernel latency test in at realtime-xxx-rta/testsuite/kern/latency/run
[18:11:10] <awallin> but it doesn't produce any output for me?
[18:11:39] <awallin> when I hit CTRL-C I do get some output which is negative or very large
[18:11:55] <gefink> strabge
[18:11:59] <gefink> strange
[18:12:05] <awallin> I'll reboot and disable all usb
[18:15:54] <awallin> still no output from the testsuite latency test...
[18:16:11] <awallin> will have to chek if latency-test still shows the delay every 10min
[18:19:58] <kirk_wallace> The BIOS update didn't fix my missing AGP. Anyone have ideas on how to procede next?
[18:21:03] <alex_joni> awallin: the rtai latency test doesn't work for you?
[18:21:17] <alex_joni> cd /usr/realtime-*/testsuite/kern/latency
[18:21:20] <alex_joni> sudo ./run
[18:21:21] <alex_joni> ?
[18:21:30] <gefink> awallin: try to check syslog for the moment the problem happens
[18:21:36] <alex_joni> (don't do that while emc2 or latency-test is running)
[18:22:19] <awallin> alex_joni: I get some stars and 'type CTRL-C' then it just hangs. When I hit ctrl-C I get a few more lines of output but with negative, zero or very large values
[18:22:27] <gefink> alex_joni: ever seen a 0.15 sec delay?
[18:22:46] <alex_joni> gefink: yes, even larger
[18:22:51] <alex_joni> usually from nvidia
[18:22:57] <awallin> gefink: where would I look for syslog?
[18:23:07] <alex_joni> there is a CPU instruction which is something like a couple million cycles
[18:23:10] <Gamma-X> hello!
[18:23:20] <alex_joni> I know nvidia uses it to clear out CPU caches
[18:23:37] <gefink> in /var/log/messages oer /var/log/syslog
[18:24:44] <gefink> alex_joni: dont clean vesa this problem?
[18:24:56] <awallin> there is something about cron just after the RTAPI: ERROR
[18:25:49] <gefink> awallin: shut down cron
[18:26:22] <alex_joni> awallin: what about cron?
[18:26:49] <awallin> service settings shows I have anacron and atd running
[18:27:25] <awallin> in syslog it says cd / && run-parts --report /etc/cron.hourly on the line after RTAPI: ERROR
[18:27:46] <Gamma-X> alex_joni, havnt been on in a few months anything new come out in emc?
[18:27:56] <alex_joni> Gamma-X: 2.2.5
[18:28:32] <Gamma-X> any hopes of an auto pid?
[18:28:39] <alex_joni> same as last time..
[18:28:45] <gefink> awallin: you can use "sudo /etc/sbin/init.d/cron stop"
[18:28:51] <alex_joni> there is a component, but not sure if it brings usefull results
[18:32:12] <gefink> so what cron starts anacron
[18:32:23] <awallin> gefink: I tried sudo /usr/sbin/cron stop but that says Resource temporarily unavailable (can't llock crond.pid)
[18:33:18] <gefink> sudo /etc/sbin/init.d/anacron stop
[18:33:33] <gefink> and killall cron
[18:34:14] <awallin> there's no /etc/sbin on my system
[18:34:36] <gefink> sorry thats my path
[18:35:03] <gefink> sudo /etc/init.d/anacron stop
[18:36:21] <awallin> ok so anacron and cron are stopped
[18:36:45] <gefink> now wait 10 minutes :-)
[18:37:37] <awallin> it should coma at around 41min past, 4 mins to that...
[18:38:56] <awallin> well there was one, but only 50 us
[18:42:48] <awallin> there. 21:42
[18:43:00] <awallin> 155.7 ms
[18:43:26] <awallin> nothing new in syslog, does syslog only log the first RTAPI ERROR that occurs, similar to the warning in EMC?
[18:43:52] <gefink> syslog shold log all errors
[18:44:17] <awallin> it warned me when the 25us task took 50us a while ago
[18:44:31] <awallin> but no new line in syslog when the 155ms delay came
[18:45:15] <gefink> i dont know . maybe only the first is shown
[18:46:24] <awallin> ok, I will try again at 21:52 and see if I get the error in the log. anyway now there was nothing else in syslog at the time of the 155ms delay
[18:47:15] <alex_joni> awallin: using vesa now?
[18:47:40] <awallin> alex_joni: this is with nv (I think) as suggested by the 8.04 installer
[18:48:07] <alex_joni> awallin: humor me, and try vesa once
[18:48:13] <alex_joni> and see if it still does it
[18:48:24] <awallin> ok, will do.
[18:48:28] <gefink> one last idea : try "option noAccel" in /etc/X11/xorg.conf in the section Device
[18:48:45] <alex_joni> gefink: that's for nv
[18:48:52] <alex_joni> but then it's not much different than vesa
[18:49:20] <gefink> alex i tried with success for vesa
[18:49:46] <awallin> I'm running dpkg-reconfigure and it asks about framebuffer? should I use it?
[18:49:46] <alex_joni> gefink: using noAccel with vesa?
[18:49:53] <gefink> yes
[18:49:56] <alex_joni> awallin: don't use dpkg-reconfigure
[18:50:15] <alex_joni> it's just one line to replace (and it's easier to create a backup of the file.. )
[18:50:27] <awallin> alex_joni: so just manual editing of xorg.conf I wonder where it is?
[18:50:30] <alex_joni> gefink: ok, will try to rember
[18:50:35] <alex_joni> awallin: /etc/X11/xorg.conf
[18:50:50] <alex_joni> awallin: btw, the rtai_smi module exists for 8.04 too
[18:51:00] <alex_joni> maybe it's a bit newer and knows about your hardware
[18:51:14] <alex_joni> sudo insmod /usr/realtime-*/modules/rtai_smi.ko
[18:51:20] <alex_joni> sudo rmmod rtai_smi
[18:51:25] <alex_joni> then to unload it:
[18:51:32] <alex_joni> sudo insmod /usr/realtime-*/modules/rtai_smi.ko smiReset=1
[18:51:34] <alex_joni> sudo rmmod rtai_smi
[18:51:40] <alex_joni> * alex_joni adds it to the wiki
[18:52:16] <awallin> insmod: error inserting '/usr/realtime...': -1 No such device
[18:52:34] <awallin> dmesg says RTAI: Intel chipset not found
[18:53:06] <alex_joni> ok, so it's still not known
[18:54:07] <awallin> I just have "configured video device" in xorg.conf so the details must be in some other file
[18:54:51] <gefink> awallin: you are right. gi to the section "device"
[18:55:24] <gefink> and insert the line after device
[18:55:50] <awallin> just: Driver "vesa" ?
[18:56:02] <gefink> Driver "radeon"
[18:56:05] <gefink> and
[18:56:16] <gefink> Option noAccel
[18:56:30] <alex_joni> awallin: yes, make it Driver "vesa"
[18:56:44] <gefink> sorry i past my file
[18:57:16] <awallin> is there a way to restart X without rebooting?
[18:57:31] <gefink> ctrl alt and backspace
[18:58:17] <awallin> ha, that is a reboot with Ubuntu
[18:58:28] <alex_joni> no.. ctrl alt del is a reboot
[18:58:30] <gefink> what?
[18:58:31] <SWPadnos> it should onl ybe a re-login
[18:58:36] <SWPadnos> only be
[18:58:46] <alex_joni> ctrl alt backspace will kill X, and gdm will restart it
[18:58:47] <SWPadnos> ctrl-alt-del kills X, which gdm notices
[18:58:53] <awallin> ok, maybe I hit the wrong keys
[18:58:56] <SWPadnos> err - right, bksp ;)
[18:59:04] <awallin> now I get X with a low resolution, so something did change
[18:59:19] <alex_joni> ctrl-alt-del brings up the shutdown dialog here
[18:59:48] <gefink> wait 10 minits :-)
[19:00:06] <awallin> yep, now I have latency-test running again
[19:00:33] <awallin> started at 22:00
[19:02:42] <awallin> there. 155ms
[19:02:51] <gefink> sh...
[19:03:05] <alex_joni> it
[19:03:13] <awallin> syslog says this: Clocksource tsc unstable (delta= 154621535 ns)
[19:03:17] <alex_joni> aha
[19:03:20] <awallin> Time: pit clocksource has been installed
[19:03:25] <alex_joni> try using a different clocksource
[19:03:31] <alex_joni> let me grab you a bootoption
[19:04:19] <awallin> what's tsc?
[19:05:42] <awallin> something with a dynamically changing clock speed of the cpu?
[19:05:44] <jepler> emc uses the tsc to deliver the "motion controller" overrun messages and also for all time values reported in 'halcmd show', so if the tsc doesn't work these alerts may be spurious.
[19:06:01] <jepler> tsc is the time stamp clock. back when it was introduced on pentiums it was a simple counter that increased at exactly the clock rate
[19:06:34] <jepler> (which was of course fixed)
[19:06:41] <awallin> google finds someone using: clocksource=acpi_pm
[19:06:52] <alex_joni> that won't work
[19:06:56] <alex_joni> acpi is disabled
[19:07:01] <alex_joni> try clocksource=hpet
[19:07:10] <awallin> on the kernel line when booting?
[19:07:13] <alex_joni> yes
[19:07:21] <awallin> ok, will reboot...
[19:07:30] <alex_joni> but if TSC is unreliable then probably the latency-test results are unreliable
[19:07:48] <jepler> yes, rtapi/emc will be using the tsc no matter what kernel boot options you use
[19:07:49] <alex_joni> so the *glitch* you're seeing might only be wrong reports from TSC
[19:08:11] <eric_U> wonder if that is happening on my system
[19:08:12] <awallin> well, it does result in the emc warning message and following errors
[19:08:29] <alex_joni> awallin: check dmesg and look for odd times
[19:08:39] <alex_joni> things that are out of order
[19:09:57] <gefink> alex_joni: could be enabling acpi an option?
[19:10:13] <alex_joni> gefink: on a by-case basis maybe
[19:10:25] <awallin> there was something about acpi in the BIOS but I left all the default values as is
[19:10:32] <gefink> alex_joni: only in this one case
[19:10:39] <alex_joni> gefink: so if you decide to try to enable it on your machine, then you might find configuration where it works
[19:10:54] <alex_joni> but I don't think a kernel with ACPI enabled will work ok for most people
[19:11:33] <gefink> alex_joni: i dont think this is my problem- but a option for failing TSC
[19:11:58] <alex_joni> gefink: I saw reports about the TSC even with stock ubuntu kernels (with full-ACPI and all)
[19:13:09] <gefink> alex_joni: i also think ACPI makes most people problems, but awallin could use it as timer
[19:13:29] <gefink> ti replace TSC
[19:13:42] <awallin> the boot option doesn't seem to do anything, I still get 155ms of delay and at the same time the Clocksource tsc unstable (delta=154ms) followed by Time: pit clocksource has been installed
[19:14:00] <alex_joni> ok, then I guess you need clocksource=pit
[19:14:10] <alex_joni> but that doesn't mean you won't get the latency error
[19:14:31] <alex_joni> awallin: was this a machine that worked ok at some point?
[19:14:47] <alex_joni> * alex_joni is running out of ideas to try
[19:14:59] <gefink> also
[19:15:06] <awallin> it has been working OK with stepper motors through the parallel port with emc2.0.x
[19:15:23] <jepler> and the same version of ubuntu?
[19:15:25] <awallin> but since that emc version doesn't report errors (I think)
[19:15:32] <alex_joni> jepler: initially on 6.06
[19:15:46] <alex_joni> but updating to 2.2.5 showed him the overrun
[19:15:57] <alex_joni> and switching to 8.04 isn't any better than 6.06
[19:16:01] <awallin> yes, I had 6.06 with steppers and saw this 155ms thing also on 6.06. Now I have 8.04 and 2.2.5
[19:16:38] <alex_joni> awallin: might be that you just didn't notice it with steppers
[19:16:48] <alex_joni> but 155ms is way too much even for servos
[19:16:54] <gefink> awallin: lost steps? in 0.15 sec ther must a big "stop"
[19:17:27] <awallin> gefink: nothing that we would notice as really bad. did 1-2h runs with the steppers sometime
[19:17:29] <alex_joni> awallin: maybe something else changed too? BIOS setting?
[19:17:35] <awallin> not every 10min for sure
[19:17:48] <awallin> alex_joni: plugged in a m5i20 card, updated to emc2.2.5
[19:18:48] <alex_joni> hmm.. I doubt the m5i20 is the latency killer.. but maybe try another PCI slot?
[19:19:28] <alex_joni> if it's on a shared IRQ ..
[19:19:28] <awallin> only thing I find in dmesg is: rtai_sched: no version for "rtai_linux context" found: kernel tainted
[19:19:28] <gefink> if ther are no stopping motors i tend to a mesurementerror, not a real stop
[19:19:32] <alex_joni> awallin: did you try to run 2.2.5 ?
[19:19:43] <awallin> no, I can try that also
[19:19:48] <alex_joni> you said you get an error (Unexpected realtime delay..)
[19:19:49] <gefink> such a stop with a stepper should be fatal
[19:19:56] <alex_joni> but you also mentioned an ferror at some point
[19:20:07] <alex_joni> or did I misread?
[19:20:56] <awallin> yes, at the shop with the servos I get a real-time error message pop-up which is usually followed by a following error pop up
[19:21:01] <jepler> 2.0.5 had the servo-thread overrun check
[19:21:09] <awallin> now I'm in the lab so can't run the motors
[19:21:24] <jepler> 2.0.3 and earlier did not
[19:21:25] <alex_joni> awallin: same machine?
[19:21:38] <alex_joni> jepler: ok, but 2.1.x has an additional check.. right?
[19:21:43] <awallin> yes, same PC, the m5i20 is still there but unplugged the cables
[19:21:46] <alex_joni> the one with task 1..
[19:21:53] <jepler> alex_joni: yes
[19:22:11] <awallin> now I'm running the sim config and we will see if I get a real-time delay message within 10min
[19:22:27] <alex_joni> sounds like you should..
[19:22:41] <awallin> yes, now I got it.
[19:23:25] <gefink> awallin: try the sim machine who sounds the PC-Speaker
[19:23:34] <awallin> 2.66 2.64 2.67 and so on between calls normally, and now 413
[19:23:52] <awallin> as jepler pointed out these may not be nanoseconds but cpu ticks
[19:24:12] <awallin> gefink: which config is that?
[19:26:25] <gefink> sim_inch.ini - inch setup with steps routed to the PC speaker instead of parport
[19:27:10] <gefink> that is readable in the selectdialog
[19:27:55] <awallin> on 2.2.5 ?
[19:28:07] <gefink> yes
[19:29:09] <gefink> etc/emc2/sample-configs/stepper
[19:31:12] <awallin> found it
[19:31:20] <awallin> so you thing a 155ms delay would be audible?
[19:31:56] <gefink> yes. verry verry clear. we speak about 0.15 seconds
[19:32:29] <awallin> hmm something like g1x110f10 should take 11 minutes?
[19:32:54] <gefink> yes so i think
[19:34:29] <gefink> i hope that gives enough steps for nice sound
[19:35:22] <awallin> I don't have absloute pitch :)
[19:35:31] <awallin> but it does beep quite nicely at f10
[19:42:35] <awallin> yes!
[19:42:53] <gefink> ??
[19:42:53] <awallin> there's a clear gap in the tone, and emc throws the real-time delay dialog at me.
[19:43:01] <alex_joni> bugger
[19:43:21] <awallin> I am done listening to the pc speaker for today
[19:43:27] <alex_joni> haha
[19:44:00] <gefink> awallin: now im 100% out of idears.
[19:44:18] <awallin> I have one or two other machines I could try
[19:44:27] <alex_joni> awallin: last think I can think of is the 5i20 and PCI slot
[19:44:32] <awallin> but it seems the graphics card or the driver is not the problem?
[19:44:34] <alex_joni> thing
[19:44:42] <alex_joni> awallin: seems like it
[19:44:55] <gefink> awallin: disasemble and crossmount (just joking)
[19:45:05] <awallin> ok, I can do a test tomorrow without the m5i20, since we didn't get lost steps every 10min with the steppers
[19:45:59] <gefink> awallin: can you use a Pata harddisk or boot the Pata cdrom?
[19:46:23] <alex_joni> gefink: if he had 6.06 and 2.0.x, then I guess it's pata
[19:46:28] <awallin> gefink: this is a sata hdd, I could maybe find a pata one to try with
[19:46:39] <awallin> the 6.06 was on a pata drive but that drive died
[19:47:26] <gefink> awallin: im sure it happens if you work on a verry expensive thing
[19:47:42] <gefink> I dont trust sata with linux
[19:48:47] <awallin> ok so first test tomorrow is without m5i20, then with a pata drive. I need to go now but I will report how it goes. Thanks for all the help and suggestions
[19:49:17] <alex_joni> awallin: sorry it didn't work out better
[19:49:48] <alex_joni> see you
[19:49:52] <awallin> gnight.
[19:49:57] <gefink> awallin: by
[19:53:07] <gefink> the 10 minute beep gave the rest
[20:20:16] <gefink> gn8
[20:57:25] <alex_joni> good night all
[20:58:46] <tom1> gnite alex
[21:29:42] <tom1> skunkworks: could i get a copy of the english hardware docs on the pic_8255 card?
[21:29:49] <tom1> pci_
[21:31:06] <skunkworks> tom1: http://www.electronicsam.com/images/KandT/conversion/pci8255.pdf
[21:31:17] <tom1> thx :)
[21:31:42] <skunkworks> I would download it - as I don't know how long I will keep it up there. (I am sure someone will get upset ;))
[21:32:16] <tom1> i grabbed it! thx ( i only found my old Thai version )
[21:32:58] <tom1> and printed it! ( not to be lost again i hope _
[21:33:11] <skunkworks> heh :)
[21:35:35] <tom1> i wonder if they'd release the src for their library. I'll ask.
[21:46:07] <K`zan> Left the steppers just sitting for a while now, running about 140-150F, perhaps I should heat sink them.
[21:57:16] <tom1> if you weren't told this before..., some drivers will sense 'idle' after a time and reduce the current to the stepper ( and of those, some may have a bit of delay waking up )
[22:08:44] <tom1> has anyone run their limit switches to their amps and not back to emc? (just run 'near home' switches back to emc, so the next encoder Z phase is watched for)
[22:09:09] <K`zan> Yes, this had it, I jumpered it out and didn't unjumper it after setting the motor current. Was concerned about the start delay so I wasn't sure if I wanted it.
[22:09:39] <tom1> lotsa amps have limit switch inputs ("fwd enable, reverse enable" )
[22:10:01] <tom1> K`zan: ask around about that temp rise, sounds enuf for cooked pork
[22:10:35] <tom1> I'd use it if it gained 20/30 degrees
[22:10:48] <tom1> err, make that lost 20/30 degrees
[22:16:49] <K`zan> Ok enabled the idle current reduction.
[22:17:05] <K`zan> Letting it sit for a half hour or so and will check the temp again.
[22:20:02] <tom1> on an L298 the limit switch could just interrrupt the enA enB pins, maybe a pull down on the opened side
[22:20:10] <tom1> K`zan: cool ( i hope )
[22:20:21] <K`zan> :-)
[22:20:49] <K`zan> Gotta redo the z axis, now that I put it together I can see what / where the problems are.
[22:50:23] <K`zan> Engaging the idle reduction jumper take the temp down to 97-104F!
[22:55:41] <BigJohnT> http://www.youtube.com/watch?v=qlXCdfPsQUU
[22:56:18] <jmkasunich> BigJohnT: cool
[22:56:27] <BigJohnT> thanks
[22:56:28] <eric_U> did the fan survive?
[22:56:28] <jmkasunich> it sure cuts a lot faster than it starts
[22:56:51] <BigJohnT> just got done cutting all the slats to support the material
[22:57:01] <BigJohnT> just being careful
[22:57:17] <BigJohnT> now the saw horses are out from underneath
[22:57:28] <BigJohnT> yes it di eric
[22:57:38] <K`zan> 3Very good!
[22:57:39] <BigJohnT> did
[22:57:56] <BigJohnT> thanks
[22:58:18] <K`zan> Will be glad when I get my silly little project doing something :).
[22:58:48] <BigJohnT> mine just got finished making itself
[23:07:30] <K`zan> Must be a nice feeling :). I'm creating a prototype from the looks of it :). Hacking the original idea to what I have / want. Design on the fly makes my brain hurt fairly often :).
[23:10:45] <BigJohnT> been a long road but fun
[23:10:59] <K`zan> Understand that, but worth it in the end.
[23:13:38] <K`zan> It will be, it will be, it will be, it will be, it will be, it will be, it will be... LOL
[23:26:54] <skunkworks> let it be?