Back
[03:05:03] <DJ9DJ> moin
[04:32:12] <JesusAlos> hi
[04:32:59] <Loetmichel> mornin'
[04:34:07] <JesusAlos> I change my PC and when put mesanet 5i25 PCI card, PC don't start
[04:35:26] <mrsun> hmm, 12tpi, 12.45mm bolt ... a 1/2 inch bolt does not fit ... wtf is it :/
[04:35:39] <mrsun> ofc 1/2 inch might not be 12tpi ?
[04:35:58] <Jymmm> My acme (iirc) is 5/8" 10TPI
[04:37:06] <Jymmm> no, that's not right, my acme is 1/2" 10TPI, my ballscrew is 5/8"
[04:37:24] <Jymmm> oh hell, I don't remember now and I'm too lazy to look =)
[04:37:48] <Jymmm> lol
[04:43:33] <Jymmm> mrsun: just grab a rule and count?
[04:44:33] <At|lantis> hello
[04:44:39] <mrsun> well the bolt existing is 12 tpi .. the question was about standard inch bolts if they were 12 tpi .. but aparently they are 13 tpi :P
[04:44:44] <mrsun> its an old machine
[04:44:50] <mrsun> trying to find a bolt for it :P
[04:45:01] <Jymmm> ah
[04:45:28] <mrsun> so might have to get the lathe in working condition again so i can cut a 13 tpi thread :P
[04:45:31] <mrsun> 12 ...
[04:46:03] * Jymmm can now Base64 en/decode by hand, YAY! (NOT)
[04:49:26] <At|lantis> hey i have an issue with a via pcm7e boad
[04:49:30] <At|lantis> anyone has the same?
[04:49:37] <At|lantis> cant get linuxcnc working on it
[04:49:44] <At|lantis> it refuses to boot ;(
[04:58:14] <At|lantis> anyone knows help?
[05:05:42] <Jymmm> http://www.idot.com.tw/en/downloads/
[06:05:43] <JesusAlos> problem with 5i25
[06:06:08] <JesusAlos> only can start PC when disable W5 swith
[06:06:35] <JesusAlos> W5 shitch disconnect FPGA from PCI
[06:07:45] <micges> JesusAlos: what motherboard?
[06:09:24] <JesusAlos> http://www.gigabyte.com/products/product-page.aspx?pid=3465#fq
[06:09:55] <JesusAlos> in this
http://es.msi.com/product/mb/760GM-P23--FX-.html#/?div=Basic motherboard run ok
[06:16:03] <JesusAlos> micges: do you think is for the motherboard?
[06:16:44] <micges> brobably some problem with bios pci configuration
[06:16:57] <micges> but you need ask PCW for details
[06:17:38] <JesusAlos> pcw_home ?
[06:18:06] <micges> yes, but there is night there now
[06:18:22] <JesusAlos> where is from?
[06:18:23] <micges> try in 6 hours
[06:18:28] <JesusAlos> ok
[07:39:00] <r00t4rd3d> http://i.imgur.com/6NZvzGY.jpg
[07:43:16] <r00t4rd3d> At|lantis, whats the error you get?
[07:43:48] <r00t4rd3d> did you install it from to the hard drive or are you trying to boot the livecd?
[07:43:57] <r00t4rd3d> -from
[07:45:17] <r00t4rd3d> and from looking at the specs of that board, its probably not going to be worth a shit as far as running linuxcnc
[07:51:01] <r00t4rd3d> i could be wrong but i doubt you will get a good latency reading from that
[07:58:51] <At|lantis> r00t4rd3d, its the only (available) board i have for the cnc
[07:58:53] <At|lantis> so ...
[07:58:56] <At|lantis> would love to use it
[07:59:04] <At|lantis> what u mean by not worth a shit...
[07:59:17] <At|lantis> u mean its gona be too slow?
[08:20:29] <jthornton> oh crap it's an hour later than I thought...
[08:36:57] <DJ9DJ> re
[09:40:36] <r00t4rd3d> At|lantis, are you just trying to run the livecd?
[09:41:37] <At|lantis> yeah
[09:41:39] <At|lantis> try
[09:41:44] <At|lantis> but ... not really wokng
[09:42:07] <At|lantis> u think i should leave it n look for anohter board on ebay?
[10:16:04] <JesusAlos> hm2/hm2_5i25.0: Unsupported Device ( ) found on sserial 0 channel 0
[10:16:27] <JesusAlos> this error appears when start my servomachine
[10:16:37] <JesusAlos> and this
[10:16:47] <JesusAlos> hm2/hm2_5i25.0: Unsupported Device ( ) found on sserial 0 channel 2
[10:20:42] <micges> JesusAlos: so 5i25 is working?
[10:21:27] <JesusAlos> but in other PC
[10:22:08] <micges> ah
[10:22:29] <micges> what is your mesa boards setup?
[10:23:28] <JesusAlos> you refer my Pncconf?
[10:23:45] <micges> no, what did you connected to 5i25
[10:23:58] <JesusAlos> 7i77 card
[10:24:04] <Jymmm> toaster
[10:24:30] <JesusAlos> my English is no good
[10:24:39] <micges> it's ok
[10:25:09] <pcw_home> the 'Unsupported Device' error can be somewhat misleading ,I just means there nothing there (no field power on 7I77?)
[10:26:19] <JesusAlos> I have power in 7i77. And card run ok
[10:26:36] <JesusAlos> I can force pins like setp hm2_5i25.0.7i77.0.0.output-00 1
[10:26:48] <JesusAlos> and it's ok
[10:27:05] <JesusAlos> but message appear every time i restart machine
[10:28:06] <JesusAlos> 5i25 led run too: setp hm2_5i25.0.led.CR01 1
[10:28:50] <pcw_home> When is field power turned on?
[10:29:18] <JesusAlos> yes
[10:29:52] <JesusAlos> If no power can't start machine
[10:30:15] <JesusAlos> can't find hal pins
[10:30:47] <pcw_home> I was asking _when_ it was turned on
[10:31:32] <pcw_home> there might be a race condition if turned on just before starting linuxcnc
[10:37:11] <JesusAlos> is a problem this condition?
[10:37:34] <pcw_home> JesusAlos: what version LinuxCNC? this could also be a driver bug in master
[10:37:36] <pcw_home> also if this is in dmesg, make sure you clear dmesg before you run linuxCNC
[10:37:38] <pcw_home> so you are not looking at an old message
[10:38:03] <JesusAlos> how clean dmesg
[10:38:05] <JesusAlos> ?
[10:38:39] <JT-Shop> sudo dmesg -c
[10:40:38] <JesusAlos> my version is 2.6
[10:40:46] <JesusAlos> http://pastebin.com/B18TyV9K
[10:40:56] <JesusAlos> is dmesg
[10:48:06] <pcw_home> you have two dmesgs stacked first probably had no field power second is correct
[10:49:18] <pcw_home> actually 4 stacked
[10:52:11] <pcw_home> sudo dmesg -c should fix this, or just look at the last linuxcnc messages (at the end of dmesg)
[10:52:28] <JesusAlos> I don't undestand. I clean my dmesg befor start linuxcnc
[10:53:21] <pcw_home> the dmesg in the pastebin is of 4 linuxcnc starts +PC startup so it was not cleared
[10:56:07] <JesusAlos> http://pastebin.com/cWwFnkan
[10:56:32] <pcw_home> And that is fine...
[10:59:32] <pcw_home> you can get rid of the:
[10:59:34] <pcw_home> hm2/hm2_5i25.0: Unsupported Device ( ) found on sserial 0 channel 2
[10:59:36] <pcw_home> complaint by disabling sserial channel 2 in your ini file
[11:02:02] <JesusAlos> What values is?
[11:02:51] <JesusAlos> [HOSTMOT2] # **** This is for info only **** # DRIVER0=hm2_pci # BOARD0=5i25
[11:03:02] <pcw_home> If you currently have
[11:03:04] <pcw_home> CONFIG="num_encoders=6 sserial_port_0=000xxx"
[11:03:05] <pcw_home> you change it to
[11:03:07] <pcw_home> CONFIG="num_encoders=6 sserial_port_0=00xxxx"
[11:04:01] <JesusAlos> I don't have CONFIG value in my .ini file
[11:04:08] <JesusAlos> Create it?
[11:04:24] <JesusAlos> in wich section?
[11:05:38] <pcw_home> Are you sure? thats quite strange, heres an example:
[11:05:40] <pcw_home> [HOSTMOT2]
[11:05:42] <pcw_home> DRIVER=hm2_pci
[11:05:43] <pcw_home> BOARD=5i25
[11:05:45] <pcw_home> CONFIG="num_encoders=6 sserial_port_0=00xxxx"
[11:07:46] <JesusAlos> no. The error still appears
[11:08:13] <JT-Shop> JesusAlos: can you pastebin your ini file?
[11:08:13] <pcw_home> Dont think so
[11:09:29] <JesusAlos> http://pastebin.com/jNydb2ji
[11:11:06] <pcw_home> AFAIK the config string needs quotes
[11:11:08] <pcw_home> and did you clear you dmesg
[11:14:12] <JesusAlos> http://pastebin.com/gWYkqR75
[11:14:31] <JesusAlos> now It's with quotes
[11:15:05] <JT-Shop> like this? CONFIG="num_encoders=6 sserial_port_0=000xxx"
[11:15:42] <JesusAlos> CONFIG = "num_encoders=6 sserial_port_0=00xxxx"
[11:15:49] <JT-Shop> yep
[11:16:18] <JT-Shop> run your configuration from a terminal and see what the error message is there
[11:18:03] <JesusAlos> http://imagebin.org/249735
[11:19:08] <JesusAlos> I attempt create a new machine and...
http://imagebin.org/249736
[11:19:26] <JesusAlos> is related?
[11:19:45] <pcw_home> looks like maybe a hal file error(config line not used), can you post your hal file
[11:19:55] <pcw_home> (config line ignored)
[11:20:32] <pcw_home> I know the sserial_port_0=whatever does work
[11:21:03] <micges> JesusAlos: pastebin your hal file please
[11:22:43] <JesusAlos> http://pastebin.com/dXCKT4fL
[11:24:14] <JesusAlos> look my serial port from my Bios if is enabled?
[11:24:36] <pcw_home> Thats not related
[11:24:47] <JesusAlos> or this serial port don't have sense?
[11:24:56] <JesusAlos> ok
[11:26:31] <pcw_home> looks like pncconfs hal file ignores the CONFIG variable so you need to change it in the HAL file
[11:27:48] <pcw_home> (change the sserial_port_0=000xxx)
[11:28:41] <JesusAlos> you refer this line?
[11:28:43] <JesusAlos> loadrt hm2_pci config=" num_encoders=6 num_pwmgens=0 num_3pwmgens=0 num_stepgens=0
[11:28:59] <pcw_home> or maybe you can disable the expansion connector sserial in pncconf, dont know
[11:29:13] <pcw_home> yes that line
[11:29:25] <JesusAlos> in this case Don't use pncconf
[11:31:22] <JesusAlos> loadrt hm2_pci config="num_encoders=6 sserial_port_0=00xxxx"
[11:31:24] <JesusAlos> yes
[11:31:29] <JesusAlos> is the solution
[11:31:42] <JesusAlos> Why don't read CONFIG line?
[11:33:24] <pcw_home> Dont know, theres no law that hal/ini files work a single way, does pncconf n ormally creat a ini file as well as a hal file?
[11:33:50] <JT-Shop> the entries in the ini file that can be read by hal lines only get read if the hal line addresses it
[11:34:04] <JT-Shop> http://www.linuxcnc.org/docs/html/config/ini_config.html#_custom_sections_and_variables_a_id_sub_custom_variables_a
[11:35:33] <pcw_home> Right so pncconf uses the ini file for some parameters but not for others
[11:36:11] <JT-Shop> I prefer to do my config in the hal file like this loadrt hm2_pci config="num_encoders=1 num_stepgens=5 sserial_port_0=0XXX"
[11:36:59] <JT-Shop> it's not something your going to change often and it is confusing to me at first to have hal lines spread over two files
[11:37:11] <pcw_home> thats what pncconf does
[11:38:32] <JT-Shop> but does it also put the [HOSTMOT2]
[11:38:32] <JT-Shop> section in?
[11:38:50] <pcw_home> I sort of prefer the other way where all parameters are in the ini file and all structure is in the hal file
[11:41:09] <pcw_home> pncconf does create a hostmot2 section in the ini file with a CONFIG= line so that confused me
[11:41:23] <JT-Shop> that is confusing for sure
[11:41:45] <pcw_home> so its there but ignored
[11:41:57] <JT-Shop> that needs to be fixed
[11:44:39] <JesusAlos> So, I ever write this line in a new configuration machine?
[11:46:31] <pcw_home> if the machines are similar just use the same file and make adjustmets in the ini/hal files as required
[11:47:34] <JesusAlos> ok
[11:48:21] <JesusAlos> other problem I have is, when install 5i25 card in this PC
http://es.gigabyte.com/products/page/mb/ga-h55m-d2h_13/
[11:48:34] <JesusAlos> the PC don't start
[11:49:11] <JesusAlos> neither bip tone
[11:49:38] <pcw_home> not very helpful :-(
[11:51:31] <JesusAlos> you refer the mother board is not helpful?
[11:52:34] <pcw_home> Yes an error message would be nice instead of a BIOS crash
[11:53:40] <ReadError> quad done
[11:53:46] <ReadError> err, mt
[11:56:38] <JesusAlos> I refear the PC don't start only when conect 5i25 card
[11:57:03] <JesusAlos> and when disconnect W5 switch from 5i25 PC can start ok
[11:57:26] <JesusAlos> W5 is the swithc that can disconnect FPGA from PCI
[11:59:39] <pcw_home> you might try updating the BIOS, and does the card work with other PCI cards?
[12:03:16] <JesusAlos> Go to update BIOS
[12:03:27] <JesusAlos> but I have F3
[12:03:35] <JesusAlos> F3 version
[12:07:14] <L84Supper> http://www.3ders.org/articles/20130310-diy-engines-these-makers-3d-printed-their-own.html
[12:08:28] <L84Supper> http://www.youtube.com/watch?feature=player_embedded&v=VhWugjxpueg UVa Mech Dept Jet Engine
[12:08:58] <pcw_home> If its a BIOS issue it may simply never work with that motherboard
[12:12:14] <pcw_home> Also if its an older motherboard thats been used (but with empty PCI slots)
[12:12:15] <pcw_home> you may have dirt in the PCI slots that needs to be cleaned
[12:13:52] <L84Supper> have you checked to see if other PCI cards work in that slot?
[12:14:11] <JesusAlos> the motherboard is new
[12:14:28] <L84Supper> it could be a bad slot (bad connector, solder connection, etc etc)
[12:14:41] <L84Supper> new boards ca also be bad boards
[12:15:30] <JesusAlos> I think is for any issue of configuration, because if connect 5i25 card without W5 switch, PC start ok
[12:15:32] <L84Supper> it's frustrating with some BIOS
[12:23:17] <pcw_home> It could be a incompatibility with the 5I25's PCI core but its fairly unlikely (only one verified incompatible motherboard found so far and it was to old/obscure to bother with)
[12:25:29] <pcw_home> GA-H55M-D2H doesn't seem available anywhere
[12:31:16] <pcw_home> ASUS E45M1-M PRO is my current favorite (no fan but runs circles around Atom MBs)
[12:31:54] <Loetmichel> pcw_home: atoms are P4s in a smaller process IIRC
[12:32:42] <pcw_home> I didnt think they were quite that bad (P4s are slower than P3s clock/clock)
[12:33:45] <JesusAlos> run ok intergated VGA card with linuxcnc?
[12:34:01] <pcw_home> P4: When Intel though consumers were so stupid that clock speed was the only criterion of speed
[12:34:14] <pcw_home> Yes integrated is fine
[12:34:25] <Loetmichel> pcw_home: the customers WERE
[12:34:33] <Loetmichel> and ARE, most of them ;-)
[12:35:04] <JesusAlos> I run linuxcnc in economic motherboard
http://www.msi.com/product/mb/760GM-P23--FX-.html
[12:35:14] <pcw_home> P4 has a 20 deep pipeline (corners like a barge)
[12:35:37] <JesusAlos> seems so slow
[12:36:45] <pcw_home> what is slow?
[12:38:32] <JesusAlos> when use programs like draftsight environment go slow
[12:39:03] <pcw_home> Normally I dont thin LinuxCNC is normally compute bound except maybe for complicated preview
[12:39:04] <JesusAlos> and linuxcnc graphican environment too
[12:40:06] <pcw_home> other than preview the Atom MBs seem ok (with Axis) and they are quite slow
[12:40:42] <pcw_home> pncconf is slow on any machine
[12:43:53] <pcw_home> apparent speed also depends on if your video hardware is well supported
[12:45:51] <L84Supper> http://www.youtube.com/watch?feature=player_embedded&v=bt2vQdgpZZg 3D Printed Rocket Engine - Test
[12:53:32] <tjtr33> pcw_home: hello, any latency numbers for the Asus E45M1? nice with 2 pci and a pport hdr
[12:55:42] <pcw_home> Yes, about 10 usec after turining off cool and quiet and C6 state in the BIOS
[12:56:31] <L84Supper> pcw_home: using the integrated graphics?
[12:57:22] <pcw_home> ~ 10 usec for both servo and base thread glxgears and flash video (integrated graphics)
[12:57:25] <tjtr33> pcw_home, thanks
[12:58:29] <pcw_home> heat sink runs hotter than Atoms but CPU performance and video performance are better
[12:58:55] <tjtr33> pcw_home: really good info, another q: is mesa working on an fpga cape/shield?
[12:59:27] <pcw_home> Yes for the cubieboard
[12:59:34] <tjtr33> again thanks
[13:00:15] <pcw_home> Someone already has xenomai running on the cubie
[13:00:31] <L84Supper> E-450 1650 MHz 0.04 micron tdp = 18 Watt
[13:00:49] <tjtr33> ( thats why i couldnt successfully search irc archives for beaglebone && mesa && fpga ;)
[13:00:54] <pcw_home> and cubie (A10) SPI runs up to 100 MHz
[13:01:29] <L84Supper> pcw_home: are you connecting to the cubie over SDIO?
[13:01:29] <pcw_home> so theres a nice minimal pin interface with decent bandwidth
[13:01:35] <pcw_home> SPI
[13:01:38] <Err> pcw_home: do you work for Mesa?
[13:02:11] <L84Supper> I'm glad the SPI controller actually works
[13:02:33] <Loetmichel> Err: he IS mesa ;-)
[13:02:40] <pcw_home> I'd rather not connect to pins that are tee'ed to others on the card (or get along with the SD card driver)
[13:02:49] <L84Supper> the SPI controller on x86 chipsets are typically crippled
[13:03:02] <L84Supper> makes sense
[13:03:17] <Err> in that case, you make some impressive boards for cheap :-)
[13:03:27] <pcw_home> Well not quite 'is' (that might make the other 6 people that work there a bit miffed)
[13:03:39] <Err> (I work with FPGA designs as a day-job, and I'm surprised at your product prices vs. what your BoM must be in relatively small quantity)
[13:04:20] <Err> do your boards target any non-LinuxCNC markets?
[13:04:48] <pcw_home> Sure most of our business in non LinuxCNC
[13:06:21] <pcw_home> the volume of our miniPCI adapters alone is > all the LInuxCNC stuff
[13:06:24] <pcw_home> most FPGA cards are used in Windows systems
[13:06:30] <Err> let me rephrase - is your market mostly CNC stuff?
[13:06:40] <pcw_home> No
[13:06:40] <JesusAlos> http://imagebin.org/249756
[13:06:44] <JesusAlos> http://imagebin.org/249757
[13:06:54] <JesusAlos> is my case 7i77
[13:07:23] <JesusAlos> proof tanks
[13:07:34] <Err> anyway, sorry to be nosy - just curious (as your products are interesting :-)
[13:09:01] <pcw_home> But it is a fun part compared to OEM products so I spend more time at it
[13:09:03] <pcw_home> (and since I'm old and crusty its about my speed)
[13:09:49] <Loetmichel> pcw_home: old and crusty? YOU? who believes THAT?
[13:12:30] <Err> heh, it's just not very often that I see a design with an FPGA on it and my initial reaction isn't "wow, I could make that board for 20% of what they're charging, including my labor" :-)
[13:15:09] <pcw_home> Loetmichel: creeps up on you
[13:15:27] <Loetmichel> pcw_home: hmm?
[13:15:46] <pcw_home> crustyness
[13:16:05] <Loetmichel> <- old enough to say that:
http://www.cyrom.org/palbum/main.php?g2_itemId=14070
[13:16:33] <archivist> that's a youngan
[13:17:01] <Loetmichel> "you should have pictured the WORKPIECE, NOT ME!"
[13:17:03] <Loetmichel> ;-)
[13:17:07] <pcw_home> Yeah Ive you you beat by a few years
[13:17:25] <Loetmichel> <- "only" 44 jears old ;-)
[13:17:25] <pcw_home> few decades maybe
[13:17:36] <pcw_home> 61
[13:17:36] <archivist> kids
[13:18:05] <archivist> ew pcw is just and only just older that me
[13:18:12] <Loetmichel> pcw_home: thats the best age, only a few jears left to retirement and still full of ideas and power ;-)
[13:18:27] <Loetmichel> archivist: harhar ;-)
[13:19:57] <Loetmichel> archivist: without the 4 yoars of working in a machine shop in the home basement of my employer, who has 4 kids... i wold still have blonde hair ;-)
[13:20:01] <Loetmichel> years
[13:21:00] <Loetmichel> not fun to have 5y, 7y, 11y, 13y kids running around moving CNC mills...
[13:21:15] <Loetmichel> selfbuild cnc routers that is...
[13:21:39] <archivist> I shall remain old free and single :)
[13:24:33] <pcw_home> My wife does nag at me about retiring, but work is still fun for me
[13:25:24] <Err> pcw_home: is your cubieboard-targeted interface going to be migratable to other embedded boards (like the beaglebone, for instance, or other cortex-A* family ARM parts) for availability
[13:25:28] <Err> ?
[13:29:23] <L84Supper> SPI is SPI, I wonder what the board size will be
[13:30:39] <L84Supper> http://www.hardkernel.com/renewal_2011/main.php or
http://www.wandboard.org/ could be options for other ARM boards
[13:32:01] <L84Supper> http://www.bddgroupllc.com/ is another a10 board, also Olimex has an a13 and a 10 board on the way
[13:32:13] <Err> yeah, I was thinking that (about SPI) - so I guess the real question is "is that the only bus connection"?
[13:33:18] <Err> The AllWinner parts are interesting to me, but I'm not sure that I'd want to build a machine controller around one (simply because their support will almost certainly disappear once the next latest-and-greatest part comes out)
[13:33:30] <IchGuckLive> hi all B)
[13:34:27] <Err> I have some non-trivial experience with the TI Sitara 335x family, although there are tons of long-term-support Cortex parts these days
[13:34:29] <L84Supper> Err: you just swap in the next ARM board with the later device, just like x86 but only slower
[13:35:29] <L84Supper> I just wouldn't design an add on FPGA or similar somehow locked into any given ARM board
[13:36:45] <L84Supper> couldn't the current EPP fpga card just offer SPI?
[13:36:54] <pcw_home> Yeah no standards (living in the shadows of PADs)
[13:36:56] <pcw_home> but the cubioboard one is just a varient of the generic
[13:36:58] <pcw_home> interface one
[13:38:44] <Err> L84Supper: well, I would actually consider building my own board, really (if the platform ended up being very stable)
[13:39:29] <pcw_home> Sure you can just chose 4 of the EPP signals on a 7I43 with interleaved
[13:39:31] <pcw_home> gnds (the data pins) once we have the SPI FPGA support ironed out
[13:39:42] <L84Supper> I'm back to trying LinuxCNC, UI, slicing, image processing and printhead control running at the same time on the same x86 cpu
[13:39:59] <pcw_home> (this will also wrok with our new low cost spartan 6 card)
[13:40:05] <L84Supper> great
[13:40:11] <L84Supper> less for me to make
[14:12:44] <L84Supper> pcw_home: if I'm controlling 5-axis with say using a 6i25 + 7i77, how much latency is there from when the encoder signal gets latched until it's sitting in a PCIe buffer ready for transfer over the bus?
[14:19:15] <Jymmm> L84Supper: Are you attempting the impossible again?
[14:20:29] <L84Supper> it's not impossible until you say so Jymmm :)
[14:20:35] <L84Supper> or proven wrong
[14:20:47] <Jymmm> L84Supper: Bullshit
[14:28:29] <Jymmm> L84Supper: But if you want to prove me wrong go catch a bullet between your teeth
[14:28:39] <L84Supper> http://www.renishaw.com/en/tonic-resm-rotary-angle-encoder-system--10243
[14:29:04] <L84Supper> what velocity? easy enough if hand tossed :)
[14:30:26] <Jymmm> => speed of sound
[14:32:25] <Jymmm> looks expensive and something one might see on a pharmaceutical class action lawsuit commercial =)
[14:33:01] <pcw_home> L84Supper: about 1-2 usec per 32 bit read on the PCIE cards PCI is faster
[14:33:17] <pcw_home> (say .5 usec)
[14:33:52] <Jymmm> pcw_home: oh gawd, you encouraged him...
[14:33:56] <L84Supper> great, plenty fast enough
[14:35:37] <L84Supper> Jymmm: i know that you're preparing for a return to life in the stone age, but are you now also trying to take others with you?
[14:35:57] <mrsun> damn im good! =)
[14:36:03] <mrsun> managed to grind a gear tooth cutter ;P
[14:36:14] <mrsun> cut the teeth, mounted gear and it feels totaly smooth =)
[14:36:26] <Jymmm> mrsun: cool
[14:36:31] <pcw_home> looks like thats a plain (though expensive) quadrature encoder (vs Resolute which is BISS or proprietary)
[14:36:47] <mrsun> if contacts etc are good or not i dont know :P
[14:37:01] <mrsun> but hopefully the lathe wont do "katchonk" every time it passes the broken teeth now :P
[14:37:20] <Jymmm> L84Supper: iirc, I think you are just asking too much from a mechanical operation is all.
[14:37:28] <mrsun> or it wil simply break the teeth of again .. hehe :P
[14:38:51] <Jymmm> L84Supper: If it was a galvanometer laser, far more reasonable
[14:39:51] <Jymmm> L84Supper: you can't break the laws of inertia... IT"S THE LAW!!! ;)
[14:41:23] <L84Supper> Jymmm: it's more about coordinating multiple systems, mechanical movements with galvos, lasers and micro-nozzles
[14:41:57] <Jymmm> L84Supper: your new sex toy?
[14:42:25] <L84Supper> it's a whole new industry
[14:43:52] <Jymmm> L84Supper: and what moves the nozzles?
[14:44:49] <L84Supper> piezo for something small
[14:45:27] <Jymmm> L84Supper: Ok, fair enough then... post a video when you get it going =)
[14:46:00] <mrsun> https://plus.google.com/photos/112857503660532884288/albums/5853798485566245441 some work left on the bullgear but =)
[14:46:05] <tjtr33> laser waffles & donuts, printable food, yummm crunchy elvis heads with maple syrup
http://www.grasshopper3d.com/forum/topics/a-solid-waffle-for-laser
[14:48:01] <Jymmm> Elvis is Dead
[14:48:18] <L84Supper> did he like waffles?
[14:48:37] <Jymmm> PB & banana sandwiches
[14:49:22] <L84Supper> i haven't seen a script for laser cutting those
[14:49:41] <Jymmm> L84Supper: cutting what, PB&B sandwiches?
[14:50:15] <mrsun> one of those machines that reduces the size of stuff what are those called? :)
[14:50:21] <mrsun> say you have a router and a "huge" pattern
[14:50:26] <mrsun> and the router cuts it alot smaller
[14:50:34] <mrsun> and you just follow the big pattern
[14:50:37] <Jymmm> mrsun: pantegraph
[14:50:41] <Jymmm> somethign like that
[14:51:14] <Jymmm> http://en.wikipedia.org/wiki/Pantograph
[14:52:55] <mrsun> thanks =)
[14:53:41] <Jymmm> L84Supper: Just toss on a piece of bread, add a handful of peanuts, hit the peanuts with the laser to desired creaminess. Toss on a peeled banana, hit with laser selecting desired thickness, top with another slice of bread.
[14:54:03] <Jymmm> mrsun: np
[14:54:26] * mrsun is thinking if its possible to make a hss bit grinder using the technique =)
[14:54:34] <Jymmm> mrsun: I'm just surprised I remembered the name of it =)
[14:54:35] <mrsun> for for example single point gear cutting tools
[14:54:50] <mrsun> could have a 10 - 20x the size "pattern" and grind the big very precise :P
[14:55:24] <mrsun> could be a neat little tool to have if it worked :P
[14:55:48] <mrsun> making grove cutters and everything to quite precise dimensions without the fiddly work :P
[14:56:56] <Jymmm> mrsun: There's this new technology called CNC too, you should check it out =)
[14:57:01] <mrsun> haha :P
[14:57:39] <mrsun> yeah i guess as you would have alot of torque at the point of the grinder it wouldnt mater .. i was thinking the "feel" of it so you do not break stuff etc :P
[15:05:02] <tjtr33> cornell prints food , piccard gets PBJ with Bananas from replicator
http://www.wired.com/video/design/design/1466790480001/3d-printer-makes-customized-cookies/1409776093001
[15:06:15] <Jymmm> liquid turkey.... Ewwwwwwwwwwwwwwwwwww
[15:06:46] <tjtr33> if you let that one run thru, its followed up by drone pizza delivery
[16:10:42] <L84Supper> gummy bear printer
[16:26:14] <tjtr33> L84Supper, well maybe a hot air rework head following a pancake mix nozzle, a big nozzle to get volumetric speed up the wait isnt so long
[16:27:02] <tjtr33> ordered a cubieboard from shanghai
[16:32:13] <pfred1> well that only took 2 days now to try to install Linuxcnc
[16:32:36] <L84Supper> tjtr33: did they say they were in stock?
[16:32:44] <tjtr33> yep
[16:32:59] <L84Supper> they have been selling out faster than they stock them
[16:33:21] <L84Supper> #cubieboard just FYI
[16:33:23] <tjtr33> wait lemme check on that, i recall they said 7days to process
[16:33:34] <tjtr33> thx
[16:33:39] <pcw_home> Where in they in stock? (I missed the last batch)
[16:34:05] <L84Supper> https://www.miniand.com/products/Cubieboard%20Developer%20Board has them as well
[16:34:30] <L84Supper> but out of stock in Australia as well
[16:34:50] <tjtr33> L84Supper, yep, om stock thru alibaba's aliexpress, from UpSea (ShangHai)
[16:35:01] <tjtr33> in
[16:35:04] <L84Supper> Switzerland: klingler.net
[16:35:20] <tjtr33> 2x the std price from the swiss
[16:35:22] <FinboySlick> Speaking of little systems... You guys seen the Minnow Board?
[16:35:26] <pcw_home> miniand says no stock (they made 3000 more saying mid-march availability)
[16:35:47] <pcw_home> maybe next week some time
[16:36:05] <L84Supper> http://www.minnowboard.org/ x86 ?
[16:36:17] <FinboySlick> L84Supper: Yeah... Atom. Atom targets it at makers.
[16:36:21] <FinboySlick> I mean Intel.
[16:36:31] <L84Supper> FinboySlick: whats the price?
[16:36:31] <FinboySlick> I was a bit underwhelmed.
[16:36:35] <FinboySlick> 200$
[16:36:39] <L84Supper> yikes
[16:37:03] <FinboySlick> Let's take the smallest atom we can find, stick it on a board for makers and sell it twice the price of our normal atom boards.
[16:37:38] <L84Supper> no mention of the actual size
[16:37:54] <FinboySlick> From the presentation it's about 4" square.
[16:38:50] <tjtr33> pcw i would have waited but didnt get a hint of times from their site
[16:39:13] <L84Supper> http://wwwd.amd.com/catalog/salescat.nsf/formfactor?openview&restricttocategory=COM%20Express%20Basic
[16:39:44] <L84Supper> http://wwwd.amd.com/catalog/salescat.nsf/formfactor?openview&restricttocategory=3.5-inch
[16:40:03] <L84Supper> http://wwwd.amd.com/catalog/salescat.nsf/formfactor?openview&restricttocategory=Pico-ITX
[16:40:12] <L84Supper> http://wwwd.amd.com/catalog/salescat.nsf/formfactor?openview&restricttocategory=Nano-ITX
[16:40:21] <L84Supper> ^^ all in that 4" range
[16:42:11] <pcw_home> if its not a high volume standard product expect it to cost more (PC104 express CPUs are in the ~$1500 range)
[16:43:43] <L84Supper> yeah, mini-itx is the lowest cost board after the various ATX size boards
[16:43:52] <t12> anyone here have much expirence with c-mount optics?
[16:44:25] <L84Supper> t12: what are you trying to do? c is one of the standards
[16:44:33] <t12> i picked up a light amp unit
[16:44:40] <t12> it has two female c-mount threads
[16:44:56] <t12> lens for imaging threads in fine on one side, but all the c-mount cameras seem to be female threaded also
[16:45:11] <t12> do i need a relay optic to focus on the amplified image plane
[16:45:44] <t12> if so are those relay optics a kinda standard thing?
[16:46:27] <L84Supper> http://www.edmundoptics.com/microscopy/relay-lenses-couplers/
[16:49:16] <t12> everything seems to be for adapting some other mount to c-mount
[16:50:43] <L84Supper> http://www.qioptiq.com/optem-c-mount-series.html
[16:53:33] <t12> hmmm
[16:54:46] <t12> sounds like the answer is
[16:54:49] <t12> obscure/random part
[16:57:10] <t12> maybe i'll get something like this
http://www.edmundoptics.com/microscopy/relay-lenses-couplers/relay-lenses/1937 and machine some housing/threading
[17:02:56] <tjtr33> Altera's home school FPGA course
http://www.altera.com/education/training/curriculum/fpga/trn-fpga.html
[17:10:27] <pcw_home> is that like a correspondence course?
[17:10:59] <tjtr33> dunno yet lots to read, and has links to 'find a local class [Afganistan]'
[17:11:15] <tjtr33> :)
[17:16:23] <pfred1> what is the latest kernel RTAI builds against now?
[17:21:26] <pcw_home> AFAIK 3.5.7 with some bugs...
[17:22:02] <pfred1> pcw_home I should say bugless then I guess
[17:24:09] <pcw_home> You need to consult a more involved guru
[17:24:42] <pfred1> last time I built LInuxcnc it was still called emc2 so I'm a bit out of date over here
[17:24:46] <pcw_home> I think that before this the latest was 2.6.38
[17:25:09] <pfred1> hmmm it patches against that late a kernel?
[17:25:40] <pcw_home> Have you visited the RTAI website?
[17:25:48] <pfred1> it took me forever to setup a Gentoo system that runs
[17:25:57] <pfred1> not recently no :)
[17:26:43] <pfred1> but I'm there now
[17:26:43] <pcw_home> Thats a fair amount of pain, you hate ubuntu that much?
[17:26:54] <pfred1> honestly yes
[17:27:26] <pfred1> last time I built Linuxcnc was on Debian it is OK but now I want to see it fully optimized
[17:32:16] <pfred1> oh look RTAI supports ARM
[17:32:16] <pfred1> does that mean we could run LinuxCNC on an ARM system?
[17:32:17] <pfred1> be the first kid on your block to run your CNC machine with your cell phone!
[17:32:17] <pcw_home> yes but I think most of the ARM people are going with Xenomai
[17:32:18] <pfred1> so is it a safe assumption to say that any version of RTAI will work with LinuxCNC?
[17:32:18] <pcw_home> running on Rasberry Pi and BeagleBone so far
[17:32:18] <pfred1> or does LinuxCNC need a specific one?
[17:32:18] <pcw_home> probably not safe
[17:32:18] <pfred1> ought oh
[17:32:18] <pfred1> K I'll check on LinuxCNC's site to see what the official word is there
[17:32:18] <pcw_home> But I dont really know. If you want to use the latest kernels checkout the Xenomai stuff
[17:32:18] <pfred1> well I have no specific preference i just want it all to work
[17:32:42] <pfred1> the hardware I'm using will run on 2.6.38 just fine
[17:32:55] <pfred1> I think
[17:33:15] <pcw_home> there are wiki instructions for Xenomai install and the developers would love more success/failure reports
[17:33:25] <pfred1> hmmm it has a funky controller chipset in it a nvidia
[17:33:45] <pfred1> I wonder how far back support for it goes?
[17:33:53] <pfred1> I'd better check that out
[17:35:11] <pfred1> NVIDIA Corporation MCP61
[17:37:32] <pfred1> I seem to be good there
[17:37:32] <pfred1> http://cateee.net/lkddb/web-lkddb/BLK_DEV_AMD74XX.html
[17:49:56] <L84Supper> pfred1: we use Gentoo all the time, we''ll post howto's for building with the latest RTAI and kernels
[17:50:04] <DJ9DJ> gn8
[17:50:22] <L84Supper> we hate Ubuntu that much :)
[17:50:49] <pfred1> L84Supper I'm looking at this now
http://wiki.linuxcnc.org/cgi-bin/wiki.pl?EmcOnGentoo
[17:51:22] <pfred1> L84Supper Ubuntu is great if all you want to do is drop a CD in and go but after a while it just rubs me the wrong way
[17:52:14] <pfred1> too many processes going too much stuff in the background it is like a plate of spagetti
[17:52:33] <L84Supper> pfred1: we just got tired of the not really wanted to be shared stuff like bullet proof x, the network manager, hodpodged back ported kernel patches etc etc
[17:52:59] <L84Supper> their attitude was why do you really want the source for that
[17:53:01] <pfred1> well as I understand it Ubuntu is Debian testing with a bunch of bad hacks going on
[17:53:46] <pfred1> so yeah for a production system it doesn't instill me with a whole lot of confidence
[17:53:47] <L84Supper> our thought was why did you break it and change the names and directory structures of what was working just fine
[17:54:16] <pfred1> Ubuntu's target audience as I understand it is casual desktop users
[17:54:31] <pfred1> if their system acts a little loopy like who cares?
[17:54:34] <L84Supper> it's great for that, but we have to support what we use
[17:55:11] <pfred1> I have never seen anything quite as demanding as LinuxCNC when it comes to computing thourh
[17:57:10] <pfred1> L84Supper do the instructions on
http://wiki.linuxcnc.org/cgi-bin/wiki.pl?EmcOnGentoo look like they would work to you?
[17:57:12] <L84Supper> we've been using it with gentoo for the past year or more
[17:57:18] <pfred1> I am really new at Gentoo yet
[17:57:28] <pfred1> yeah well I'm on day 3 over here
[17:57:39] <pfred1> so a lot of this Gentoo stuff is new to me
[17:57:51] <L84Supper> pfred1: I don't really touch it much myself. We have a full time dev here that keeps up with the kernel and gentoo stuff
[17:58:02] <pfred1> but hey I managed to get a box over here to tun X Window
[17:58:47] <pfred1> run X Window even
[18:09:52] <L84Supper> http://phys.org/news/2013-03-caltech-self-healing-electronic-chips.html
[18:10:11] <pfred1> I could use those when I'm breadboarding
[18:10:11] <pfred1> I melted a spot on my breadboard
[18:10:12] <pfred1> I was pretty proud of that :)
[18:10:12] <pcw_home> Yeah will those help when customer vaporise the bond wires?
[18:10:12] <pfred1> it is amazing how little milliwatts really are sometimes
[18:10:12] <pfred1> I have an issue with my BOB now where I didn't fully think out how much current it is going to have to supply to my optical isolators
[18:10:12] <pcw_home> we had a customer make his own PC power cord and swapped hot with GND
[18:10:12] <pfred1> always good to put blocking diodes into devices
[18:10:12] <pcw_home> that was interesting
[18:10:12] <pfred1> for those times when red looks a lot like black
[18:10:13] <L84Supper> lol
[18:10:13] <pfred1> it really is amazing how many amps you can get out of an every day wall outlet
[18:10:33] <pcw_home> we have reverse diodes on most new cards (but if the are 5V or 3.3V they are parallel reverse diodes so the power supply may still win)
[18:10:33] <pfred1> I had an electrician tell me once yo ucan get a 50,000 amp surge before wires explode and breakers trip
[18:10:34] <pfred1> I've blown a pair of pliers just about clean in half in an outlet box
[18:10:34] <pfred1> there is some serious power there
[18:10:34] <roh> well.. energy is power to time... so thats all relative
[18:10:34] <pcw_home> A friends father use to work in a a Alcoa aluminum plant shorts in the cells would make the bus bars whip around like snakes
[18:10:35] <pfred1> all I'm saying is if you're messing with the stuff wear good shoes
[18:10:59] <roh> 50kA seems much, but its usually gone in splitseconds since fuses trip/the supply cannot provide
[18:11:00] <pfred1> oh yes he was talking about a very shurt time duration event
[18:11:00] <pfred1> short
[18:11:10] <pfred1> that initial slash
[18:11:10] <pfred1> flash
[18:13:13] <roh> check youtube videos about protectors against such stuff... there is interresting stuff
[18:13:39] <roh> goes as far as explosive bolts shorting busbars with a crowbar in case of overcurrent. to 'make sure' the fuses trigger fully and not just smore
[18:14:21] <pfred1> I love electricity though
[18:14:54] <pfred1> stuff is like magic
[18:14:54] <roh> kills the inserts and some parts, but saves the connection box which could be in flames elsewise (and further loss of property and health)
[18:14:54] <pfred1> I was on a job once and a substation xformer blew up
[18:14:55] <pfred1> it was like an atomic bomb mushroom cloud the works
[18:14:55] <roh> naah. just be sure to remember that everything above a hundred volt and over a small 2 digit amps is bad idea to tamper with if you do not really know what youre doing.
[18:14:55] <pfred1> everyone shouted do it again!
[18:14:55] <roh> pfred1: thats what youtube is for. so we dont need to destroy stuff for people to learn about how much respect one should have
[18:14:55] <pfred1> actually with enough amps I think 36 volts though it could be 38 is entering dangerious to human health
[18:14:55] <pfred1> I had a meter once that would beep whenever it was hooked up to volts equal or above the OSHA rating
[18:14:55] <roh> true. thats why you shouldnt touch live wires at all
[18:14:55] <pfred1> well it all depends how well grounded int othe circuit you are
[18:15:03] <pfred1> I mean I've gotten 120VAC a lot
[18:15:30] <pfred1> but just so long as I'm wearing shoes keeping me insulated from the ground and it is only one contact point it is just annoying
[18:15:34] <roh> i live in 230V country... doesnt allow for stupid things too often
[18:15:51] <pfred1> yeah 230VAC now you're talking a big bite
[18:15:56] <roh> and even 60V ac is bad for your heart
[18:16:10] <pfred1> we were demoing this place out and some guys were up on a hi lo lopping cables
[18:16:24] <pfred1> and they broke open a conduit and were taking out all the cables and they were hot
[18:16:36] <pfred1> man that blew up like the 4th of July!
[18:16:50] <pfred1> it was 3 phase 460 1000 amps
[18:17:08] <L84Supper> warn them first, then get out your camera
[18:17:08] <pfred1> we had to get them loppers down with a broom handle
[18:17:33] <roh> hmm.. the english version of
http://de.wikipedia.org/wiki/Fünf_Sicherheitsregeln is really bad
[18:17:38] <pfred1> when the boss showed up later and he saw the loppers all he said is they were a brand new pair!
[18:18:20] <pfred1> we were pretty pissed at the electrician after that
[18:18:28] <roh> i hope you sued him.
[18:18:30] <pfred1> everything was supposed to be dead
[18:18:33] <pfred1> nah
[18:18:38] <roh> such people need to loose their jobs and license
[18:18:46] <roh> no excuses.
[18:19:00] <roh> if you adhere to standard safety protocol such things cannot happen.
[18:19:17] <pfred1> I'm just glad i gave up on trying to lever the sub panel those wires were going to
[18:19:38] <pfred1> hey when we wrecked places it was pretty nuts
[18:19:44] <roh> hm. evil that english translation of the 5 security rules.
[18:19:48] <pfred1> stuff happened
[18:20:21] <pfred1> I worked for another contractor when we hit some live BX cable he was like no sweat he took a sledgehammer and beat the stuff until it blew up
[18:20:39] <pfred1> then he wasl ike it's dead now!
[18:21:13] <roh> still stupid and very irresponsible. thats why people get hurt and outages happen.
[18:21:16] <pfred1> I got to admit his technique worked
[18:21:23] <roh> no wonder the us has a 3rd world infrastructure
[18:21:42] <pfred1> here is one you'll love i wasn't on this job a buddy of mine was
[18:22:01] <pfred1> he was piuring a concrete floor and one of the masons was bullfloating it
[18:22:15] <pfred1> well he ran the handle of the bullfloat into an open panel and hit the bus bars
[18:22:35] <pfred1> bullfloat handles are aluminum rod about 3 inches in diameter and he was standing in wet concrete
[18:23:01] <pfred1> he was toast
[18:23:25] <roh> sigh. yes. and why the hell is there an open panel?
[18:23:35] <pfred1> that was the question that was raised
[18:23:42] <roh> and why was there power on it?
[18:23:50] <pfred1> it was just a temp job panel and someone didn't put a cover on it
[18:23:59] <pfred1> well we need power on jobsites
[18:23:59] <roh> thats rule 1 of the 5 i posted the url to. switch off.
[18:24:29] <pfred1> but yeah that was a total doh someone got sued on that deal
[18:24:41] <pfred1> someone probably went to jail
[18:25:57] <roh> true. thats what distribution boards are for. on building sites. but they are usually waterproofed, mechanically hardened (steel or abs plastic) and to a certain degree idiot proofed
[18:26:26] <roh> dunno how that works in other countries
[18:26:37] <pfred1> well a lot of electrical contractors toss up used panels they pulled out of other jobs as temp site panels
[18:27:05] <pfred1> things usually end up getting ripped out by a backhoe or something by the end of the job
[18:27:32] <roh> idiot-proofed as in 'you cant touch metal without tools' like any sockets here and 'fuses/gpf for individual sockets if possible'
[18:27:34] <pfred1> that one was obviously in really poor shape
[18:27:51] <pfred1> GFIs are horrible
[18:28:12] <roh> why? have so much broken equipment that they mistrigger for you?
[18:28:15] <pfred1> guys always end up cutting the things off
[18:28:34] <pfred1> beats me they just trip too much
[18:28:55] <pfred1> when I was roofing we'd run extension cords clear across flat commercial roofs
[18:29:02] <pfred1> and we'd use the stuff when it was raining
[18:29:04] <roh> well.. fix your rf protector caps
[18:29:11] <pfred1> I mean if the weather was nice we wouldn't hang track
[18:29:20] <pfred1> we'd only do that in the rain
[18:29:40] <pfred1> but anyways we'd have the extension cords in puddles up on the roof
[18:29:55] <pfred1> like one extension cord plugged into another one and the ends under water
[18:30:24] <roh> dont get me wrong.. i also want to get stuff done.. but i had my fair share of accidents with power and have some permanent reminders on both hands from one special incident.
[18:30:42] <roh> especially workspaces need proper protection circuits
[18:31:51] <pfred1> man these instructions are for 2.6.34.5-r4
[18:31:52] <Valen> I've never had an issue with gfi
[18:32:10] <Valen> though half the time the trip is too sensitive for you to notice why it tripped
[18:32:20] <roh> Valen: i have. it was broken. also people never dare to test them regulary like ordered in the manuals
[18:32:37] <roh> also the caps in modern smpsu are wearing off.
[18:32:38] <Valen> we found out an air conditioning compressor was leaking
[18:32:51] <Valen> about 70v to the case ;->
[18:33:14] <Valen> we wanted to keep using the compressor for our mill air supply (it was silent)
[18:33:21] <Valen> so we mounted it on isolators ;->
[18:33:22] <roh> but there is measurement equipment for that. and its even needed in workshops to go around and check equipment regulary afaik
[18:33:58] <Valen> here no charity will accept any electronic good (even a toaster) because it needs to be test and tagged before they can touch it
[18:34:16] <Valen> as such no charities have computers or anything anymore
[18:34:41] <roh> well.. then you should find a certified electrician who donates 'plugging it into his test device and nod or shake head' ;)
[18:34:57] <Valen> $25 a device for a tag
[18:35:26] <roh> i know. thats why the certifier should donate it to the nonprofit.. could even help writing off tax
[18:35:33] <roh> (thats how it works here)
[18:36:01] <Valen> it doesn't appear to be worth it for them here
[18:36:05] <pfred1> Valen I just picked up a PC at the thrift store here
[18:36:44] <Valen> I give all my dead computers to a friend
[18:36:52] <Valen> he cuts them up and makes space ships out of them
[18:37:16] <roh> do you have recycling yards?
[18:37:18] <pfred1> sweet got any pics?
[18:37:33] <Valen> nah, they go through my dad
[18:37:39] <Valen> i havent seen any in ages
[18:37:48] <pfred1> Valen when i picked that PC up it was dead but I reseated the RAM and it runs now
[18:38:15] <pfred1> Linux acme 3.7.10-gentoo03 #6 SMP PREEMPT Fri Mar 8 14:04:48 EST 2013 i686 AMD Athlon(tm) 64 X2 Dual Core Processor 5000+ AuthenticAMD GNU/Linux
[18:38:30] <Valen> yup people toss them, when they are simpleish to fix
[18:38:39] <pfred1> I got it for $5
[18:38:49] <pfred1> big money!
[18:39:37] <pfred1> I'm setting it up to be ny new LinuxCNC box
[18:41:04] <pfred1> testing the latency off the live CD it was meh I'm hoping a fixed install with some tweaks works better
[18:48:15] <Err> heh, I have an old AMD64 first-gen that's 3800+ or some crap, and it has horrid latency as well (tried it last week)
[18:48:32] <Err> it seems like it should have way better performance, really, even though it's 6 years old
[18:49:16] <pfred1> well this box will clock like 5000ns but then it spikes now and again
[18:49:29] <roh> tried uniprocessor?
[18:49:37] <roh> which chipset?
[18:49:39] <pfred1> I just ran off the CD
[18:49:46] <Err> yeah, it's the instantaneous spikes that would be an issue, when I was testing
[18:49:47] <roh> avoid anything nvidia
[18:49:47] <pfred1> NVIDIA Corporation MCP61
[18:49:50] <tjb1> How can I view what errors linuxcnc has?
[18:49:52] <roh> eeek.
[18:50:02] <tjb1> I try to jog my Z axis and it just says "joint 2 following error"
[18:50:09] <pfred1> tjb1 /var/log/syslog ?
[18:50:21] <pcw_home> dmesg?
[18:50:38] <pfred1> pcw_home that should all be in syslog too
[18:51:10] <pcw_home> also running linuxcnc from a command line
[18:51:20] <pfred1> tjb1 you can watch errors in real time with this command tail -f /var/log/syslog
[18:51:23] <tjb1> what do I type for that
[18:51:45] <pfred1> then you can see them as they are happening
[18:51:57] <tjb1> "-f /var/log/syslog"?
[18:52:03] <pfred1> less /var/log/syslog to see them after the fact
[18:52:15] <pfred1> "tail -f /var/log/syslog"
[18:52:32] <pfred1> man tail
[18:52:43] <pfred1> tail - output the last part of files
[18:53:00] <pfred1> the -f switch tells tail to keep outputting
[18:53:03] <tjb1> Nothing came up when the error came up
[18:53:31] <Err> so why do nvidia chipsets suck?
[18:53:42] <pfred1> beats me
[18:53:54] <roh> well... check linus thorvalds on youtube for that
[18:54:06] <Err> I shall google
[18:54:08] <pfred1> he is talking about their graphics adapters
[18:54:14] <roh> in short: they dont give out specs -> bad, mostly reverseengineered only drivers
[18:54:21] <uminded> is their an easier way to compile new HAL drivers other than rebuilding the ENTIRE linuxcnc package?
[18:54:23] <roh> or binary which usually means no rtai
[18:54:28] <Err> hm, I wouldn't think that the graphics card would affect scheduling, but ... maybe
[18:54:45] <pfred1> I hear the nvidia binary driver is a total no go with LinuxCNC
[18:54:58] <pfred1> but I may just try ot to see for myself
[18:54:58] <roh> nvidia chipset != nvidia graphics
[18:55:14] <pcw_home> following error usually means that you have exceeded the maximum step velocity or
[18:55:15] <pcw_home> something else is funny with what you are asking the stepgen to do
[18:55:27] <Err> yeah, I don't care about running the nvidia binary driver for an embedded machine
[18:55:45] <pfred1> me either I want the maximum performance I can get though
[18:56:04] <pfred1> and the nosfratu drivers are scary
[18:56:07] <tjb1> pcw_home: Im guessing I didnt convert the names right that send position to thc.comp
[18:56:10] <roh> for cnc you dont need performance. you need reliability.
[18:56:28] <pfred1> roh I software step gen so I need some performance
[18:56:53] <Err> I don't ever do RT on a PC, but I do things that require RT constraints on an embedded uC all the time
[18:56:54] <roh> even really old radeon cards for free do fine. we use a radon9000 series something
[18:57:08] <pcw_home> Yeah a step input to the stepgen on position mode will do that
[18:57:10] <pfred1> my old machine in the highest microstepping mode can only go .8 IPS
[18:57:24] <roh> on a via board. either km400 or kt400 series dunno which anymore. one was good and one bad for rtai
[18:57:37] <Err> I actually think that, anymore, there are plenty of uCs that could do *all* of the RT parts in a remote board (either internal or external card), and the PC could just be a pretty interface, G-code interpreter/parser, etc
[18:58:05] <pfred1> Err that is great if you have that uC
[18:58:20] <pfred1> me I'm working with what i have now
[18:58:26] <roh> Err: maybe. but there is no solution as complete, tested and versatile as linuxcnc as far as i know.
[18:58:33] <Err> I'm not throwing stones here - just musing out-load
[18:58:34] <Err> loud
[18:58:50] <pfred1> I have heard about something that runs on arduino for CNC
[18:58:51] <pcw_home> also you really want to move all of HAL or its a mess
[18:58:54] <pfred1> starts with a G?
[18:58:57] <roh> n.p. but it stays a hypothetical or something much more specific than linuxcnc
[18:59:39] <pfred1> grbl
[18:59:54] <Err> I don't know linuxcnc's architecture at all (just started looking at it, preliminarily, for some machines that I don't know enough about ;-) - but in general, the less RT stuff that you make a PC do the better I think :-)
[19:00:21] <pcw_home> This is being worked on (machinekit) but will require a pretty decent micro (good floating point)
[19:00:41] <pfred1> strangely GRBL doesn't seem to have its own webpage beyond git
[19:00:41] <roh> Err: learn some more before sentencing it ;) i was wrong in the beginnning too. i also would like more modern platforms and some arms to be able to run emc2 on.
[19:00:42] <Err> there are a bunch of cortex-m4 parts now that I would think can keep up with the processing
[19:00:58] <pfred1> https://github.com/grbl/grbl
[19:01:02] <roh> grbl is childs toys... can only do stepper based machines open loop.
[19:01:12] <Err> ...but I don't do motor control programming :-)
[19:01:15] <pfred1> roh that is all I'm doing
[19:01:46] <pfred1> but I can get a PC for cheaper than I can get an arduino
[19:01:59] <Err> (however, a micro with an FPGA for the control loops ... that would be my preferred architecture, because that's what I know, and should be extremely capable)
[19:02:25] <roh> pfred1: there are many correct solutions to one questions. if you need more flexibility or ui than an avr can do... complexity kills you fast. avrs are ok for simple stuff.
[19:02:47] <pfred1> roh I'm OK for simple stuff too
[19:03:01] <roh> but as soon as you want a nice 3d gl view in your device, arms and avrs get annoying fast
[19:03:03] * pfred1 is all about the simple stuff
[19:04:10] <pfred1> hey look Grbl does have a website
http://dank.bengler.no/-/page/show/5470_grbl?ref=checkpoint
[19:04:19] <pcw_home> For floating point performance per $ its hard to beat a PC
[19:04:22] <roh> Err: about fpgas.. nice for you, but doing that is complex for most people.. writing some hal rules in emc is very easy compared to that
[19:04:44] <pfred1> pcw_home especially when I can get PCs for free or next to it
[19:04:52] <Err> roh: my thought here would be to integrate these pieces as 'accelerators' to linuxcnc - not to throw it away
[19:05:05] <Err> (so the HAL rules would still exist)
[19:05:34] <Err> just because I might be able to roll my own doesn't mean that I want to - maintaining one-offs sucks, especially for production
[19:06:18] <roh> Err: that would be nice.. but i dont believe that happening soon, sadly.. to many have tried and not delivered (yet) afaik.
[19:06:35] * Valen has wanted to get into fpgas
[19:06:56] <roh> Valen: me too. and then the toolchains drove me away.
[19:07:07] <Valen> arms have gotten dirt cheap these days though
[19:07:17] <Valen> the pi and its similars
[19:07:30] <Valen> $50 for a gigahz cpu with a dull stack
[19:07:34] <pfred1> dirty deeds done dirt cheap
[19:08:10] <Valen> full stack rather
[19:08:30] <Valen> I have a need to do lots and lots of fast PWM often
[19:08:34] <roh> Valen: yeah. too bad they usually have muxed gpio
[19:08:46] <Valen> (like 300 or so PWM channels sometimes)
[19:09:05] <pcw_home> machinekit aims to separate the real time portions of LinuxCNC to suitable hardware (so GUI/intepreter can run on anything)
[19:09:16] <Err> Valen: that would be a fantastic application of an FPGA
[19:09:25] <pfred1> pcw_home but where is the fun in that?
[19:09:30] <Valen> thats what i figured Err
[19:09:32] <Err> pcw_home: that's exactly the sort of arch I envision
[19:09:46] <Valen> pcw_home: link?
[19:10:07] <roh> pcw_home: if there are fresh developments to test.. holler.. we have stacks of different develboards of nearly all archs around at raumfahrtagentur
[19:10:20] <pcw_home> Proababl yfor fairly capable micros (M4 might be a little wanting)
[19:10:22] <Valen> roh now thats a name!
[19:10:34] <roh> Valen: ?
[19:10:37] <pfred1> I was just thinking try to say that 3 times fast
[19:10:40] <Valen> raumfahrtagentur
[19:10:55] <roh> hehe.. means 'spacetravelagency' in german
[19:11:13] <roh> its a hackerspace in berlin
[19:11:18] <Valen> to english language sensibilities it sounds much like klingon
[19:11:21] <pcw_home> we made a 96 channel 20 MHz scaler for someone on a 5I22
[19:11:43] <pcw_home> (its a hotmot2 module!)
[19:11:44] <Valen> nice pcw_home
[19:11:59] <Valen> whats this machinekit you speak of?
[19:12:12] <pfred1> those Germans with their compound words
[19:13:37] <roh> hehe.. you havent had the fun to learn metalworks in 2 languanges
[19:15:35] <roh> there seem to be _very_ few people knowing all the special words for things in that area in more than one language
[19:15:42] <pfred1> someday I have to take a pilgrimage to Solingen Germany
[19:15:50] <Valen> specialist translator? ;->
[19:16:22] <Valen> I have always wondered about putting the PID into the firmware of the mesa cards
[19:16:23] <pfred1> come back with as much steel as I can carry
[19:16:47] <Valen> pfred1: that probably wont be much, steel is heavy lol
[19:17:07] <pfred1> Valen oh I don't know I'm good for a couple of hundred pounds
[19:17:32] <Valen> still not a great deal of steel really
[19:17:53] <pfred1> you don't build bridges out of Solingen steel
[19:20:09] <Valen> theres not much mythology left in steel these days
[19:20:46] <pfred1> you wouldn't think so butthat stuff they make in Solingen still seems like a cut above to me
[19:21:12] <Valen> I think if you were to get mill certified steel of suitable grades you might be supprised
[19:21:51] <roh> i think one cannot even buy metals anymore without a proper datasheet
[19:21:54] <pcw_home> Valen: machinekit is mhaberlers plan for clean separation of RT and non RT portions of LinuxCNC
[19:21:55] <pcw_home> so the RT portion (motion/hal) can run on the most suitable RT hardware and the GUI/interpreter
[19:21:57] <pcw_home> can run on the most suitable GUI hardware/software
[19:22:02] <roh> last time i had real trouble selecting
[19:22:23] <roh> one step short of startrek and selecting an atomic composition ;)
[19:22:26] <Err> pcw_home: does it still run linux on the RT portion?
[19:23:05] <Valen> I thaught there already was that somehow, I seem to recall being able to run axis on a different machine to the controller over a network?
[19:23:18] <pfred1> metallurgy was still a black art up until the space age then humanity had to get serious about the stuff
[19:23:52] <pcw_home> Err: Not sure if that will be a requirement or not (for really high performance it would be better to dump linux)
[19:23:57] <Valen> roh: theres a difference between the datasheet and actually getting a cert from your supplier
[19:24:23] <roh> Valen: sure. but it never was an issue for me so far.
[19:24:27] <Err> pcw_home: yeah, that's my thought (although programming on RTOSes - ore bare metal - is often more difficult)
[19:24:29] <pfred1> Valen then there is NASA who will literally stand over your shoulder while you weld
[19:24:44] * Valen has a degree in space science ;->
[19:24:44] <pcw_home> Thats the tradeoff
[19:24:57] <Valen> hardware support would be rather minimal too I'd imagine
[19:26:20] <Valen> I wonder about something in the way of a hypervisor
[19:26:29] <Valen> so cnc can run in dom0
[19:27:00] <roh> well.. any ui interactions would have as much delay as the solution provides
[19:27:00] <Valen> then you can pass any hardware to linux
[19:27:46] <Valen> sorry a linux domU, so you should get linux level hardware support for video and network etc
[19:28:16] <Valen> and you could allocate it fixed time slices to run in
[19:28:42] <Err> heh, any time you're doing virtual memory management and a full-blown OS RT constraints are going to be relatively gross vs. a simple RTOS machine
[19:29:05] <Valen> err point is all that stuff can be happening inside the VM
[19:29:08] <Err> (assuming you allow for pages to swap out)
[19:29:22] <Valen> the RT stuff can allocate it a deterministic slice of CPU
[19:29:39] <Valen> https://sites.google.com/site/realtimexen/
[19:29:50] <Valen> there ya go, its already been invented lol
[19:29:51] <pcw_home> but a machinekit RT system even running linux should be a lot more manageable
[19:29:52] <pcw_home> as it may really only have one non real time hardware driver (Ethernet)
[19:30:18] <roh> pcw_home: well.. arent there ethernet based motor drivers?
[19:30:23] <Err> yeah, if *all* the linux box is doing is RT stuff, that should be doable
[19:30:51] <Err> run with no swap, just a handful of processes... you should be able to minimize latency considerably
[19:31:10] <Valen> latency has nothing to do with how much work the computer is doing
[19:31:15] <pcw_home> I was considering Ethernet for Machinkit as the (non real time) host interface
[19:31:32] <Valen> archivist: would often compile mysql with distcc whilst cutting,
[19:31:47] <Err> Valen: it has to do with how many tasks are competing (although with priorities, it doesn't have to be 1:1)
[19:32:35] <Valen> Err: no not even that is my understanding, there are certain operations undertaken by drivers that won't allow the RT interrupt to fire
[19:32:51] <Valen> those are what cause latency (on a system that isn't overloaded)
[19:33:03] <Valen> as RT stuff runs with a higher priority than anything else
[19:33:05] <Err> Valen: certainly busted drivers will also increase latency :-)
[19:33:29] <Err> but even with RT scheduling, task switching has a cost (to *get* to the higher-priority thread when it needs the CPU)
[19:33:56] <ReadError> is it too much to have an atom board running 2 machines at the same time?
[19:34:05] <ReadError> if one of them is going much slower, so less io ?
[19:34:06] <Valen> with quad core 3Ghz cpu's task switching isn't that expensice
[19:34:31] <Err> actually, multi-core might make it *more* expensive, if you have to synchronize cores at task-switch time
[19:34:33] <pcw_home> There basically is no RT scheduling since the only RT task is LinuxCNC
[19:34:43] <Err> (but I don't really know the rules of linux RT these days - haven't looked into it in years)
[19:34:46] <Valen> I use isocpus anwyay
[19:34:54] <Valen> so there is one CPU just for RT tasks
[19:35:03] <Valen> and one for everything else
[19:35:29] <Valen> I do wonder given multicore arch if rather than timers a spinloop would be better
[19:35:45] <Valen> just busywait until task fire time
[19:35:49] <andypugh> pcw_home: What does "Line Driver" mean in the context of an encoder, and what does it take to interface it to a bare 7i43 pin?
[19:36:00] <pcw_home> so there is no competition for RT (only bad drivers. power management etc that cause latency)
[19:36:06] <pfred1> once I saw a chart of different Linux kernel versions compared to each other based on performance and it isn't a linear progression
[19:36:34] <pcw_home> line driver means the encoder has differential outputs
[19:36:53] <Valen> so pcw_home, mesa gear got enough herbs to shift everything from the motion planner on down into firmware? ;->
[19:37:25] <Err> well, I have to bow out for dinner, but thanks for the interesting discussion (I've already fallen to modern jackassing at this point -
http://www.thisamericanlife.org/radio-archives/episode/293/a-little-bit-of-knowledge - as my experience is all old or theoretical on linux and RT)
[19:37:55] <pcw_home> Not in a FPGA, but eventually in a ZYNQ based controller
[19:38:06] <Err> ooh, are you goign to make zynq cards?
[19:38:15] <pcw_home> Yes
[19:38:19] <Err> sexy
[19:38:30] <Err> I might have to get some for work, just as toys
[19:38:43] * Valen hasn't heard of this thingie
[19:38:44] <Err> (we're an all-altera shop, but those look nice)
[19:38:46] <pcw_home> I like the fast low latency CPU/FPGA connection
[19:38:50] <Err> yes
[19:38:57] <Err> that's a fantastic arrangement
[19:39:18] <Err> Altera is supposed to have a setup like that coming out soon, too, but they suck for announcing products a hojillion years before they ship
[19:40:13] <Valen> fpga wrapped around a dual core arm?
[19:40:26] <Err> they're more side-by-side
[19:41:27] <Valen> finding it hard to find an overview
[19:41:57] <pcw_home> Yeah and unlike previous chips the CPU is independent (no FPGA code needed to boot)
[19:42:58] <Valen> ahh, so its a dual core arm SOC with close coupled FPGA
[19:43:13] <pcw_home> so its ideal for smart re programmable I/O (and maybe eventually machinekit)
[19:43:50] <Valen> I could hang my buttloads of PWM off it and have a linux stack in there to make life easy
[19:44:04] <pcw_home> Dual Gig E/ SDIO USB on proc
[19:44:15] <Valen> i saw that lol
[19:44:31] <Valen> could also make for a good packet switching engine
[19:45:08] <pcw_home> Software radio as well
[19:45:11] <Valen> full hardware crypto
[19:45:22] <Valen> IE put the whole algorithm into the FPGA
[19:45:43] <pcw_home> Yep lots of fast FPGA cells
[19:46:30] <Valen> though for that application shared ram would probably be nice
[19:47:08] <Valen> for a zero copy thing
[19:47:28] <pcw_home> CPU/FPGA bandwidth is quite good
[19:48:06] <pcw_home> and FPGA/SDRAM
[19:49:56] <Valen> I believe the state of the art these days with the fancy nic's is the packet hits the nic, nic informs OS, OS then reads the bits of the packet it wants and can then direct it as it chooses
[19:50:20] <Valen> the packet body never actually leaves the nic or if it does, it does a direct transfer to the outbound nic
[19:50:36] <Valen> the message body never goes through the cpu
[19:51:21] <Valen> i don't think linux itself has gotten to that stage, but there are extensions that do it
[19:52:07] <roh> Valen: it does. what you describe is scather-gather.
[19:52:14] <roh> and zerocopy.
[19:52:40] <pcw_home> There are also all hardware stacks (some pushed by the fast trading stuff) Theres one of these on Opnecores
[19:54:47] <pcw_home> http://opencores.org/project,udp_ip_stack
[19:55:22] <TekniQue_> mhmm
[19:55:31] <Valen> I hate all that rapid trading stuff
[19:55:32] <TekniQue_> I made something similar for one of my projects
[19:55:47] <TekniQue_> a UDP-only IP stack
[19:56:31] <TekniQue_> what's rapid trading?
[19:56:40] <Err> heh, ethernet in fpgas hasbeen popular for years in the video processing arena
[19:57:38] <pcw_home> TekniQue_, trading 100 ns ahead of your competition...
[19:57:44] <TekniQue_> stocks?
[19:57:45] <TekniQue_> ah
[20:30:30] <uminded> loadrt give the error cannot open shared object file: No such file or directory when loading the rasberry pi hal_gpio module
[22:04:34] <skunkworks> pcw_home: pete is very happy with the 5i25 - I think he may buy another - or at least the stepper interface.. He sold someone locally his pico stuff.
[22:06:18] <pcw_home> good to hear (and nice you helped him set it up)
[22:07:27] <skunkworks> no problem
[22:08:57] <skunkworks> he has 2 machines he needs controls for. He has been just moving computers back and forth.
[22:09:05] <skunkworks> *computer
