#emc-devel | Logs for 2009-02-23

[01:36:27] <CIA-2> EMC: 03jepler 07TRUNK * 10emc2/src/po/ru_axis.po: new translation
[04:05:06] <steve_stallings> steve_stallings is now known as steves_logging
[11:55:00] <christel> [Server Notice] Hi all, the server you are currently connected to (lindbohm.freenode.net) will go down for scheduled maintenance tomorrow (February 24th 2009) at 0900GMT. You may wish to re-connect to a server in main rotation by changing to irc.freenode.net prior to this time. Thank you for using freenode and apologies for any inconvenience this may cause.
[18:08:35] <alex_joni> anyone with a mesa around?
[18:09:00] <alex_joni> guess I'll wait for seb
[18:15:21] <BJT-Work> hi alex_joni
[18:16:18] <alex_joni> hi BJT-Work
[18:23:06] <buildmaster> build #23 of lenny-x86-trunk-realtime-rip is complete: Failure [failed configuring autotools] Build details are at http://emc2-buildbot.colorado.edu/buildbot/builders/lenny-x86-trunk-realtime-rip/builds/23
[18:23:21] <seb_kuzminsky> ^^^^ my bad, i broke the farm, please ignore
[18:23:43] <BJT-Work> you gotta put your name in the Blamelist: so I'm not the only one :)
[18:23:50] <seb_kuzminsky> heh
[19:26:58] <seb_kuzminsky> i renamed it emc2-buildmaster, better than the default name of buildmaster, which kept colliding with someone else's buildmaster...
[19:27:48] <seb_kuzminsky> emc2-buildmaster: status lenny-x86-trunk-realtime-rip
[19:27:48] <emc2-buildmaster> lenny-x86-trunk-realtime-rip: idle, last build 3162 secs ago: build successful
[20:08:59] <alex_joni> seb_kuzminsky: short question
[20:09:14] <alex_joni> does the hm2_pci export hm2_5i20.* pins for 5i20 ?
[20:09:27] <alex_joni> and hm2_5i22.* for 5i22, etc?
[20:13:52] <BJT-Work> http://www.linuxcnc.org/docview/devel/html//drivers_hostmot2.html#r1_4 :)
[21:08:56] <seb_kuzminsky> alex_joni: yes
[21:09:11] <alex_joni> ok
[21:09:16] <seb_kuzminsky> BJT-Work: thanks :-)
[21:11:38] <alex_joni> seb_kuzminsky: ok, cool
[21:21:05] <seb_kuzminsky> alex_joni: http://highlab.com/~seb/lots-of-pins
[21:21:48] <seb_kuzminsky> (that's from 2.2, as you can tell because the connector-name is in the gpio-name in hal)
[21:38:21] <alex_joni> hm2-stepper.hal:52: parameter or pin 'hm2_5i20.0.watchdog.timeout_ns' not found
[21:38:40] <alex_joni> that's on 2.2.x running the config hm2-stepper (from CVS TRUNK)
[21:38:42] <seb_kuzminsky> context pls
[21:39:15] <alex_joni> I asked for a dmesg.. still waiting on it
[21:39:28] <alex_joni> (german forum for emc2 I'm supporting once in a while..)
[21:39:48] <alex_joni> he commented the watchdog line, and got:
[21:39:54] <alex_joni> HAL: ERROR: function 'hm2_5i20.0.read' not found
[21:39:55] <alex_joni> hm2-stepper.hal:59: addf failed
[21:39:59] <seb_kuzminsky> the TRUNK configs are totally untested on 2.2
[21:40:08] <seb_kuzminsky> there's an addf to put the pet-watchdog function in the servo thread
[21:40:17] <alex_joni> so I suspect something doesn't get loaded right
[21:40:21] <seb_kuzminsky> sounds like maybe he's got a firmware without a watchdog?!
[21:40:27] <seb_kuzminsky> i didnt think we had such a thing
[21:40:55] <alex_joni> he said he took the firmware from the 5i20.zip from mesanet
[21:41:01] <seb_kuzminsky> hmm
[21:41:13] <alex_joni> (after the one in 2.2 didn't work..)
[21:41:41] <cradek> doctor doctor!
[21:42:09] <alex_joni> seb_kuzminsky: I'll wait for the dmesg before speculating further :)
[21:42:17] <alex_joni> cradek: feeling ill?
[21:42:17] <seb_kuzminsky> in hm2-stepper.hal, the first thing after setting the watchdog.timeout_ns is an addf of an hm2 function
[21:42:50] <seb_kuzminsky> will the .hal file abort if the loadrt fails?
[21:42:56] <alex_joni> yeah, the hm2_*.0.read.. which fails
[21:43:02] <alex_joni> seb_kuzminsky: it should
[21:43:20] <seb_kuzminsky> i would'a thought so
[21:43:25] <seb_kuzminsky> yeah, let's wait for the dmesg ;-)
[21:43:31] <alex_joni> my guess would be that it misdetects the board somehow
[21:43:49] <alex_joni> or doesn't detect it properly (e.g. doesn't export the proper foos)
[21:43:54] <seb_kuzminsky> there's quite a lot of sanity checks and cross-checking of the data it reads from the board
[21:44:17] <seb_kuzminsky> i think it's unlikely that a board without a consistent hostmot2 firmware would be detected as hostmot2
[21:44:31] <seb_kuzminsky> maybe there's a hostmot2 firmware .bit without a watchdog on mesa's site?
[21:44:38] <seb_kuzminsky> that seems unlikely too but not impossible i suppose
[21:44:53] <seb_kuzminsky> afaik we have the latest .bits from mesa's .zips
[21:44:57] <alex_joni> dunno.. but even without a watchdog, it shouldn't barf on 0.read
[21:45:16] <seb_kuzminsky> it might be the later addf ...pet_watchdog that fails
[21:45:30] <alex_joni> I think the error is pretty clear
[21:45:50] <alex_joni> 23:37 < alex_joni> HAL: ERROR: function 'hm2_5i20.0.read' not found
[21:45:56] <seb_kuzminsky> oh oops you're right, that *is* pretty clear :-/
[21:46:10] <seb_kuzminsky> now *there* is a useful error message ;-)
[21:46:17] <seb_kuzminsky> if only you read it ;-)
[21:46:32] <alex_joni> well.. maybe if it would blink :P
[21:46:49] <seb_kuzminsky> we need an <BlinkingEchoReverb> tag
[21:46:50] <alex_joni> 4 or use colour
[21:47:11] <alex_joni> one that survives going through a forum (html), then on irc
[21:48:11] <alex_joni> 4,1test
[21:48:37] <alex_joni> (that's supposed to blink on some strange terminals :)
[21:48:52] <alex_joni> with /set mirc_blink_fix OFF
[21:51:05] <alex_joni> seb_kuzminsky: btw, how's lenny working for you?
[21:51:15] <seb_kuzminsky> no problems at all
[21:51:23] <seb_kuzminsky> though it's also not as cool as i had hoped
[21:51:28] <alex_joni> heh
[21:51:38] <seb_kuzminsky> i didnt realize the gcc in Ubuntu had a bunch of optional stuff turned on by default
[21:51:41] <seb_kuzminsky> the gcc in lenny doesnt
[21:51:49] <seb_kuzminsky> so it doesnt produce as many warnings as i had hoped
[21:51:51] <seb_kuzminsky> :-(
[21:51:57] <alex_joni> ah, so maybe jaunty would be more edgy
[21:52:00] <seb_kuzminsky> heh
[21:52:19] <seb_kuzminsky> jaunty would be most kuddly
[21:53:06] <seb_kuzminsky> why, do you have a jaunty emc2 system squirreled away somewhere?
[21:53:16] <alex_joni> SWPadnos might have ;)
[21:53:26] <alex_joni> I only installed 8.10
[21:53:37] <alex_joni> what was that again? intrepid..
[21:53:51] <alex_joni> but I'll install a koala when it's out :D
[21:54:22] <alex_joni> https://lists.ubuntu.com/archives/ubuntu-devel-announce/2009-February/000536.html
[21:54:38] <alex_joni> "the *Karmic Koala*, the
[21:54:41] <alex_joni> newest member of our alliterative menagerie."
[21:55:03] <seb_kuzminsky> oh right, karmic
[21:55:30] <seb_kuzminsky> grr, one of my grad students requires an application of the clue bat, brb
[21:55:37] <alex_joni> heh
[21:56:59] <seb_kuzminsky> the "-c" flag to netcat means "crackpipe"
[21:58:03] <alex_joni> broken pipe?
[21:58:12] <seb_kuzminsky> broken gradstudent
[21:58:48] <alex_joni> heh
[22:06:24] <tomaw> [Global Notice] Hi all, as you probably noticed our services took a short break there. They're back online now and we're investigating the cause of the crash. Thanks for using freenode and sorry for the inconvenience!
[22:35:02] <cradek> alex_joni: I read that kibbitzing koala marketingspeak and I have no idea what it's talking about
[22:35:18] <seb_kuzminsky> i agree, "head in the clouds", no joke
[22:35:46] <alex_joni> cradek: you probably missed the "cloud computing" wave
[22:35:55] <alex_joni> I must admit I did that oo
[22:35:58] <alex_joni> too
[22:36:04] <cradek> apparently it has acronyms in it
[22:36:20] <cradek> "we'll make it boot faster" is something I can understand
[22:37:06] <seb_kuzminsky> back in the day they thought that "computation" would be a public utility like electricity (think mainframes and teletypes)... "cloud computing" is the modern take on that, renting cycles on other peoples server farms
[22:39:02] <alex_joni> apparently some vendors already make money from renting software
[22:39:23] <alex_joni> (which you can use online..)
[22:40:33] <seb_kuzminsky> stuff where you run other people's software on other people's machines and use your browser as the ui aren't interesting to me
[22:41:04] <alex_joni> you forgot that you have to pay for it
[22:41:08] <seb_kuzminsky> heh
[22:41:17] <seb_kuzminsky> decentralized (server-less) webapps are cool
[22:42:40] <seb_kuzminsky> http://testgrid.allmydata.org:3567/uri/URI:DIR2-RO:j74uhg25nwdpjpacl6rkat2yhm:kav7ijeft5h7r7rxdp5bgtlt3viv32yabqajkrdykozia5544jqa/wiki.html
[22:42:46] <seb_kuzminsky> that's a wiki without a server
[22:43:04] <seb_kuzminsky> the storage backend is an ad-hoc p2p filesystem
[22:43:16] <seb_kuzminsky> the frontend is a JS wiki editor that runs in the browser
[22:43:26] <seb_kuzminsky> *that* is the cool opensource part of cloud apps
[22:43:47] <seb_kuzminsky> ok i'm done being offtopic, i'll shut up now ;-)
[22:43:50] <cradek> where is the filesystem?
[22:44:09] <seb_kuzminsky> it's in the cloud
[22:44:09] <seb_kuzminsky> heh
[22:44:19] <cradek> in web browsers?
[22:44:24] <seb_kuzminsky> it's like raid-over-nbd
[22:45:00] <seb_kuzminsky> you run a webserver that does put and get to this p2p system called tahoe
[22:45:31] <seb_kuzminsky> it uses FEC to split your files across 10 peers, any 3 of which are sufficient to recover your file
[22:45:47] <seb_kuzminsky> it's cap-based, with the fec keys in the resulting url
[22:46:46] <seb_kuzminsky> that's why the url is so gross...
[22:48:08] <cradek> I'm glad other people work on stuff like this - I haven't learned any new tricks since around 1998.
[22:48:55] <seb_kuzminsky> yeah it's good we all obsess over different things, on the whole we're a pretty well-rounded bunch ;-)
[22:49:01] <cradek> yes
[22:49:22] <cradek> interesting, I've had the same email address for 9 years, today
[22:50:14] <alex_joni> cool
[22:51:38] <seb_kuzminsky> i was wishing the other day that someone in the rtai camp would provide a repo of rtai debs for us...
[22:52:36] <cradek> last I looked, they were ambivalent about packaging in general.
[22:52:57] <cradek> (anything but building everything custom for your machine using a kernel from kernel.org)
[22:53:47] <seb_kuzminsky> that's too bad
[22:54:10] <cradek> yeah.
[22:54:15] <cradek> pain for us.
[22:56:36] <seb_kuzminsky> there are rtai packages in all the ubuntu distros, but they're all broken and dont work
[23:07:17] <alex_joni> one of these days I'll set it up in ppa
[23:11:51] <seb_kuzminsky> alex_joni: it == rtai? or emc2?
[23:13:42] <alex_joni> rtai patched kernel + rtai
[23:13:50] <seb_kuzminsky> cool
[23:14:14] <alex_joni> maybe even emc2
[23:14:17] <seb_kuzminsky> i'm still planning to have the emc2-buildbot produce .debs for all our supported platforms
[23:14:33] <alex_joni> cool
[23:15:04] <cradek> what would be uber-cool is to have it do that based on the cvs tags
[23:15:20] <seb_kuzminsky> you mean notice new tags and build those?
[23:15:21] <cradek> it would automate the release-making procedure
[23:15:24] <cradek> yes
[23:15:44] <cradek> then the release procedure is simply to tag what we want in it
[23:16:09] <cradek> not sure how you'd get them into a signed apt repo though
[23:16:13] <seb_kuzminsky> does CVS support post-facto tagging?
[23:16:28] <alex_joni> even webtriggered would be ok
[23:16:36] <seb_kuzminsky> ie "tag the tree the way it was last night, when we all tested it so well, before we checked in all this sketchy stuff"
[23:16:44] <alex_joni> cradek: with ppa you can
[23:16:45] <cradek> yes
[23:16:48] <seb_kuzminsky> cool
[23:17:04] <cradek> but we always release from a stable branch anyway
[23:17:12] <seb_kuzminsky> does tag-creation generate CIA emails?
[23:17:17] <cradek> we never put sketchy stuff there :-)
[23:17:22] <alex_joni> basicly you only upload sources, and the signing, etc gets done automatically after compiling
[23:17:23] <cradek> no, I don't think so
[23:17:26] <seb_kuzminsky> heh cough cough
[23:17:35] <alex_joni> * alex_joni looks the other way too
[23:17:54] <cradek> actually it's pretty true
[23:18:24] <alex_joni> yeah, just kidding
[23:18:25] <seb_kuzminsky> i wish i had the resources to set up an always-on test system with a bunch of anyio boards in it
[23:18:36] <seb_kuzminsky> it's hard to test that stuff without a machine with a testing wireharness on it
[23:19:08] <cradek> thanks to you guys who do all the meta-work. I don't miss doing it. I get to concentrate on coding features when the mood strikes, and not worry about all the details.
[23:19:30] <seb_kuzminsky> heh
[23:19:36] <alex_joni> * alex_joni is happy hardy is working so well
[23:19:45] <alex_joni> no real complaints about the kernel
[23:20:13] <cradek> yep, wfm
[23:24:04] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/config/ini_config.lyx: add a few missing tid bits
[23:24:39] <seb_kuzminsky> and then there's BigJohnT :-)
[23:24:44] <BigJohnT> LOL
[23:24:51] <BigJohnT> what's up seb?
[23:24:55] <seb_kuzminsky> hey man
[23:25:32] <seb_kuzminsky> quick, commit the new pics that your previous checking references!
[23:25:48] <BigJohnT> LOL not this time
[23:27:05] <BigJohnT> seb_kuzminsky: if I want to connect my 5i20 to my gecko drives do I need the I/O card or just a breakout thingy?
[23:27:20] <seb_kuzminsky> bob should be enough i think
[23:27:35] <BigJohnT> cool
[23:27:58] <seb_kuzminsky> might want to check on 3.3V vs 5V level compatibility
[23:28:05] <seb_kuzminsky> i dont know what the geckos need
[23:28:22] <BigJohnT> ok thanks for the heads up... I think off the top of my head it is 5v
[23:28:46] <seb_kuzminsky> gecko input current levels should be low, compared to the sourcing capacity of the 5i20, but that'd be another thing to check
[23:28:58] <seb_kuzminsky> other than that, the less junk in the signal path, the better, i think
[23:29:09] <seb_kuzminsky> but i'm just a coder, so dont believe anything i say about electronics ;-)
[23:29:46] <BigJohnT> LOL ok I was just wondering for when Peter sends me the A-D thingy to test for a THC
[23:30:45] <seb_kuzminsky> tell me about that ADC thingy one more time? how does it work?
[23:31:35] <BigJohnT> it will "plug" into one of the ports of the 5i20 and I can use the digital in EMC to make a THC using EMC
[23:31:50] <BigJohnT> in EMC/ in to EMC
[23:32:18] <BigJohnT> Peter said you could add it easy :)
[23:32:22] <seb_kuzminsky> lol
[23:33:05] <seb_kuzminsky> the 12 bits (or however big it is) adc value shows up on a bunch of gpio pins? then you smush them together with a custom HAL component?
[23:33:20] <BigJohnT> what I need from it is a float pin or a U32 pin with the converted analog value of the tip voltage
[23:33:33] <BigJohnT> it can work that way
[23:33:51] <BigJohnT> using binary from the inputs
[23:34:44] <seb_kuzminsky> that's pretty wild
[23:35:01] <BigJohnT> 0 to 300 vdc => voltage divider => 0 to 10 vdc into the A-D into the 5i20
[23:35:02] <seb_kuzminsky> does the adc need to be triggered from hal, or does it constantly convert, all by itself?
[23:35:24] <BigJohnT> I'm thinking constantly
[23:35:34] <seb_kuzminsky> that would be very convenient :-)
[23:36:28] <BigJohnT> if the adc is "just there all the time" and you read it when you need it that would be best I guess
[23:37:03] <seb_kuzminsky> i wonder if that guy geo01005 from the reprap group knows about this
[23:37:23] <BigJohnT> I don't think so
[23:37:50] <seb_kuzminsky> he was looking for an ADC for emc2 to build a temperature controller for the print head on his 3d printer
[23:38:19] <BigJohnT> single channel?
[23:38:27] <seb_kuzminsky> for now yeah
[23:38:39] <seb_kuzminsky> he's pursuing SPI, for multi-channel adc/dac/gpio in the future
[23:38:53] <BigJohnT> that would be cool
[23:39:07] <seb_kuzminsky> THC is a big electric spark, no?
[23:39:10] <seb_kuzminsky> sketchy
[23:39:52] <BigJohnT> no, the THC just monitors the voltage from the plasma cutter and tells the Z to raise or lower based on the preset
[23:40:23] <seb_kuzminsky> isnt the plasma torch electrical? or is it some kind of burning gas?
[23:40:44] <BigJohnT> IT's plasma :)
[23:40:54] <jmkasunich> electrically burning gas
[23:40:56] <BigJohnT> yes electrical with a gas
[23:41:03] <seb_kuzminsky> ok
[23:41:15] <BigJohnT> low cost ones use air fine cut use other gasses
[23:41:38] <seb_kuzminsky> noble gases that dont react much, i guess? argon and stuff?
[23:42:07] <BigJohnT> they use O2 and a couple of others I forget
[23:42:16] <seb_kuzminsky> wild
[23:42:23] <BigJohnT> but that is HIGH priced equipment
[23:43:29] <BigJohnT> mine was like 4 or 5k if I recall
[23:43:47] <seb_kuzminsky> what can it cut?
[23:44:16] <BigJohnT> any metal or other object you get inbetween the plasma and the metal :)
[23:44:26] <seb_kuzminsky> heh
[23:44:35] <BigJohnT> I cut mostly steel and stainless steel
[23:44:55] <seb_kuzminsky> what's your kerf like?
[23:45:31] <seb_kuzminsky> or beam width or whatever it's called ;-)
[23:45:37] <BigJohnT> depends on the thickness but for thin material about 0.050" for thicker stuff it tapers from about 0.100" down
[23:45:52] <BigJohnT> to 0.050" or so
[23:46:24] <BigJohnT> and that depends on the nozzle I use too...
[23:47:44] <BigJohnT> well, I need to wander out to the shop and finish cutting up material to wittle on tomorrow at the machine shop
[23:47:56] <BigJohnT> talk with you later
[23:48:03] <seb_kuzminsky> ok, see you
[23:48:17] <seb_kuzminsky> heh wittling metal
[23:57:24] <seb_kuzminsky> buy guys
[23:57:27] <seb_kuzminsky> bye