#linuxcnc-devel | Logs for 2015-01-11

[00:00:11] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 bfa19c5 06linuxcnc 10docs/src/getting-started/Getting-LinuxCNC.txt docs: clean up "Burning the CD" * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=bfa19c5
[00:00:12] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 f2ec7eb 06linuxcnc 10docs/src/getting-started/Getting-LinuxCNC.txt docs: fix up the "Testing LinuxCNC" section a bit * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=f2ec7eb
[00:00:12] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 d13ccb7 06linuxcnc 10docs/src/getting-started/Getting-LinuxCNC.txt docs: clean up Installing LinuxCNC a little bit * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=d13ccb7
[00:00:14] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 e139173 06linuxcnc 10docs/src/getting-started/Getting-LinuxCNC.txt docs: clean up Updates to LinuxCNC a bit * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=e139173
[07:59:46] <jthornton> in 2.7 Getting-LinuxCNC.txt the md5sum does not match the http://linuxcnc.org/index.php/english/download page
[08:01:14] <jthornton> I'm a little short on download bandwidth or I'd download the iso and see which one is correct
[09:39:04] <Tom_itx> jthornton, 2.7?
[09:39:08] <Tom_itx> i don't see that
[09:42:28] <Tom_itx> i'm checking 2.5 & 2.6 ...
[09:50:18] <JT-Shop> yea 2.7
[09:50:30] <Tom_itx> where did you find the iso for it?
[09:50:33] <Tom_itx> i'll check it
[09:50:44] <Tom_itx> i'm pulling the other 2 right now
[09:51:37] <JT-Shop> http://www.linuxcnc.org/docs/2.7/html/getting-started/index.html#_getting_linuxcnc
[09:52:07] <Tom_itx> ok that's where i'm at
[09:52:13] <JT-Shop> oh that is the 2.7 wheezy iso
[09:52:19] <JT-Shop> so it might be correct
[09:52:29] <Tom_itx> which one did you want?
[09:53:11] <JT-Shop> someone on the list was complaining about it but I think they are wrong now cause they were compairing 2.6 to 2.7
[09:54:39] <Tom_L> i'll check em anyway
[09:54:45] <JT-Shop> ok thanks
[09:54:50] <Tom_L> since they're about here
[09:55:10] <JT-Shop> I need to schedule a download after midnight if I can figure out how to do so
[09:56:06] <Tom_L> you've got limited bandwidth anyway don't you?
[09:56:33] <Tom_L> just the guy i was looking for :D
[09:56:53] <pcw_home> uh oh
[09:56:58] <JT-Shop> yea, 10G a month but after 12 and up to 6am unlimited
[09:57:19] <Tom_L> pcw_home do you know which .vhd is the default on the 7I90? it looks like a servo bit file to me..
[09:57:38] <Tom_L> the 'user' one shipped on it...
[09:58:26] <pcw_home> Not sure just some common EPP mode file
[09:58:53] <pcw_home> probably SVST8_4 or similar
[09:59:34] <Tom_L> i was wondering what the step/table1 dir/table2 entries were on the 2nd & 3rd headers
[10:00:04] <pcw_home> step/dir outputs
[10:01:16] <Tom_L> nothing special then...
[10:01:54] <pcw_home> No except they may have all the possible table outputs
[10:02:27] <Tom_L> i just wondered if you added new functions to hostmot2 for the newer versions of lcnc
[10:02:42] <pcw_home> since the config was started as a demo, and does not have sensible step/dir pinouts
[10:03:18] <pcw_home> no, the table stuff has been there since the first hostmot2 config
[10:03:29] <Tom_L> the first header shows pwm where the other 2 are all stepgen
[10:03:45] <pcw_home> though its only recently been supported
[10:03:57] <pcw_home> (in the driver)
[10:04:14] <pcw_home> so probably svst4_8
[10:04:17] <Tom_L> i'm not gonna use it, just wanted to know what file it was
[10:04:19] <Tom_L> ok
[10:04:46] <pcw_home> run mesaflash --readhmid on it
[10:05:04] <Tom_L> that's where i found out what it had
[10:05:19] <pcw_home> that will list the pins better (dmesg list only shows enabled pins)
[10:05:43] <Tom_L> ahh i didn't realize that
[10:06:00] <Tom_L> makes cents though
[10:06:33] <pcw_home> well if mesaflash --readhmid works there's a good chance you can re-flash it without incident
[10:07:06] <Tom_L> i don't need any special entries in the other 2 headers unless i'm gonna use them right?
[10:07:12] <Tom_L> just gpio
[10:07:36] <pcw_home> nope GPIO is fine
[10:08:19] <pcw_home> all GPIO is fine (the sserisl remote firmware uses all a GPIO config as do the fallback configs)
[10:08:19] <Tom_L> i copied the one from the 7I43 over to the first header and left the others alone so it should be ok
[10:09:36] <Tom_L> i tested the old 7I47 and it's good so i may put the new one on a header beside it and just use it for io
[10:10:54] <pcw_home> lesson: if you muck up a config and it does not work, the way to fix it is to _temporarily_ move the EEPROM jumper to select the other chip
[10:10:55] <pcw_home> power up with the working config and _immediately_ move the jumper back
[10:11:27] <pcw_home> if you bust both images, you are up the creek
[10:13:50] <Tom_L> i was reading that
[10:13:58] <pcw_home> fallback will save you form things like losing power while programming or programming with busted files,
[10:14:00] <pcw_home> but if you program with a correct bit file (correct size and CRCs) that does not work, you depend on the alternate EEPROM
[10:14:12] <Tom_L> you move the jumper back while powered so you can write back to the first config right?
[10:14:20] <pcw_home> yes
[10:14:52] <Tom_L> leaving W4 disconnected for this?
[10:14:58] <pcw_home> the jumper just routes the EEPROM CS signal to one or the other EEPROMS
[10:15:04] <Tom_L> that's just to write the backup config right?
[10:15:12] <Tom_L> ahh ok
[10:15:12] <pcw_home> W4 does not work yet
[10:15:16] <Tom_L> oh
[10:15:20] <Tom_L> heh
[10:15:35] <Tom_L> wired but not in firmware?
[10:15:47] <pcw_home> It needs some magic in mesaflash to setup the flash chip options
[10:16:41] <pcw_home> its data thats written to non volatile control registers in the EEPROMs to enable their WP inputs
[10:17:22] <pcw_home> not related to firmware at all
[10:17:57] <Tom_L> right
[10:19:28] <Tom_L> i was reading about the bitfile format a little.. what are the multiboot features it's talking about/
[10:19:29] <Tom_L> ?
[10:19:56] <Tom_L> is that firmware rather than user bit files?
[10:20:48] <pcw_home> Spartan 6 (and other FPGAs) have the ability to boot different images (at different starting addresses in the flash)
[10:21:41] <pcw_home> A common use is fallback (if the primary bitfile fails to load, load an alternate)
[10:23:20] <pcw_home> this is especially important on cards like the 5I25, say someone pulls the power cord when writing a bit image
[10:23:22] <pcw_home> without fall back this would be a bricked card
[10:23:39] <Tom_L> yeah i see why it's there
[10:23:53] <Tom_L> could you recover it with jtag?
[10:24:02] <Tom_L> that would probably be the only way
[10:24:13] <pcw_home> Sure, you can always recover with JTAG
[10:24:36] <Tom_L> probably not too many want to cross that bridge yet
[10:24:46] <pcw_home> but about .01% of our customers have that
[10:25:18] <Tom_L> i've got an old jtag for xilinx... i should get a better one
[10:25:37] <Tom_L> it's on a dev board i got years ago
[10:25:38] <pcw_home> we made a parallel one years ago and still use it
[10:26:00] <Tom_L> got a schematic for it?
[10:26:10] <Tom_L> it's just a single buffer chip probably
[10:26:10] <pcw_home> (improved from Xilinxes shoddy design)
[10:27:13] <pcw_home> yeah we added Schmitt triggers on the clock receiver and HC buffering so ~2V to 5V work OK
[10:28:00] <pcw_home> I think we have 2 chips on a ~1.5x1.5" board
[10:28:22] <pcw_home> i have a stack of boards if you want a couple
[10:28:46] <Tom_L> i wonder what they get for their fancy usb one
[10:29:10] <Tom_L> i'd love that..
[10:29:18] <Tom_L> been meaning to make a dongle just never have
[10:29:20] <pcw_home> Their are cheap USB clones on EBAY
[10:29:36] <pcw_home> there
[10:29:40] <Tom_L> i've got a jtag for some cypress? chips i think somewhere
[10:29:59] <pcw_home> but Ive never bothered
[10:30:39] <Tom_L> and an old 68332 board with fpga chips on it with the jtag built in one of them
[10:30:57] <pcw_home> the bad thing is AFAIK it has to look like a Xilinx JTAG adapter (impact only works with Xilinx hardware)
[10:31:18] <Tom_L> yeah jtag is bad about that
[10:31:29] <Tom_L> they're very simple but almost proprietary
[10:31:53] <pcw_home> I did manage to write a JTAG programmer that works on Xilinx CPLDs
[10:33:16] <pcw_home> FPGAs are trickier, the current Impact thing loads a JTAG bootstrap loader first to speed things up
[10:35:23] <pcw_home> so anyway the thing about multiboot is dont enable any multiboot
[10:35:24] <pcw_home> features in the bitfiles, this is all handled by mesaflash
[10:44:00] <Tom_L> i was able to salvage the 7I47 and the old one draws the same current as the new one so it's likely ok now. the 7i43 is toast (spare parts)
[10:46:56] <pcw_home> wonder what happened, 5V chips on 7I47 have 7VDC abs max so must have been higher (or backwards)
[10:49:39] <Tom_L> i have a suspicion but can't prove it
[10:49:52] <pcw_home> Ive had a few people fry 7I77s (a more expensive proposition)
[10:49:54] <pcw_home> with bad 5V so maybe I should replace the reverse protection diode with a zener
[10:49:55] <pcw_home> It will work for reverse and forward (but will fry with large current)
[10:50:10] <Tom_L> i may have got the 24v mixed with the 5v. that's the absolute only thing i can see so far
[10:50:20] <Tom_L> but i can't verify it
[10:50:53] <pcw_home> if you have a 24-5v regulator a flaky ground can do this
[10:50:57] <Tom_L> a bud of mine thought it may have been a ground failure of some sort but i have my doubts
[10:51:13] <Tom_L> i've gone thru and checked all those
[10:51:18] <Tom_L> back to the main pannel
[10:51:35] <Tom_L> either 0 ohms or real close to it
[10:53:17] <pcw_home> unfortunately on overvoltage, buck switching regulators may fail in such a way as to pass the overvoltage down the line
[10:53:44] <pcw_home> (bad for things like the 7I43)
[10:55:08] <pcw_home> you can see that the linear regulator on the 7I47 saved the 3.3v parts (it probably has a 30 or 40V max input)
[10:58:07] <Tom_L> the regulator to the boards is still good
[10:58:17] <Tom_L> i swapped it out anyway
[10:58:51] <Tom_L> it's likely the same reg i use
[11:00:22] <Tom_L> nope
[11:00:33] <Tom_L> LM2596 on mine
[11:34:32] <seb_kuzminsky> JT-Shop: at(1) is your friend for casual scheduling of future jobs
[11:34:59] <seb_kuzminsky> & thanks for looking at the md5sum issue
[11:36:31] <seb_kuzminsky> i'm md5summing the images on wlo and here's what i got:
[11:36:40] <seb_kuzminsky> b515c872335336ccfc96471d66b687d8 binary.hybrid.iso
[11:37:19] <seb_kuzminsky> 3aa031895ac74a81d0639b09cd812c22 linuxcnc-2.7-wheezy.iso
[11:37:27] <Tom_itx> i checked the current 3 listed and they match
[11:37:43] <Tom_itx> 2.5 2.6 and 2.7, i didn't check the 2nd 2.7 ver
[11:38:23] <seb_kuzminsky> there's only one 2.7 iso
[11:38:36] <Tom_itx> the sha256sum one
[11:38:40] <seb_kuzminsky> oh, right
[11:38:40] <Tom_itx> i didn't check
[11:38:51] <seb_kuzminsky> that might have been a mistake on my part
[11:38:58] <seb_kuzminsky> to add the sha256 i mean
[11:39:14] <seb_kuzminsky> there's also the issue cradek raised, where we have the checksums written down in several places
[11:39:34] <archivist> they should not be in the main docs methinks
[11:39:38] <seb_kuzminsky> there's the Download page on wlo, and the Getting LinuxCNC section of linuxcnc.org/docs
[11:39:45] <seb_kuzminsky> there may be a place on the wiki too
[11:40:54] <seb_kuzminsky> archivist: i put it in the "real" docs on purpose, thinking it'd be easier to keep up to date there than in the wiki or the joomla pages
[11:41:02] <seb_kuzminsky> i'm open to arguments against it
[11:41:33] <cradek> I/we just need to have a list of things to update when I/we change the images
[11:41:47] <archivist> the sum changes when the docs change, to understand recursion you need to understand recursion
[11:43:11] <cradek> that's an interesting twist
[11:44:08] <seb_kuzminsky> heh
[11:44:09] <Tom_itx> it shouldn't be on the wiki. anybody can change that
[11:44:44] <seb_kuzminsky> well, the sum changes when the image is updated, and that only happens when you poke cradek
[11:45:50] <seb_kuzminsky> Tom_itx: i think the wiki is a great place to jot down thought and sort of prototype documentation, but i'd like to see good finished docs migrate to the git repo and the asciidoc source where it's a little easier to edit & manage
[11:46:38] <seb_kuzminsky> cradek: good point, i'll add it to the Release Checklist wiki page
[11:46:48] <cradek> I think the md5sum should be shown right next to the download link, and I don't really care about anywhere else
[11:47:03] <cradek> I guess I also think it should be only in one place
[11:47:37] <cradek> so I guess my preference means poking joomla to update it, and everywhere else includes it only by reference
[11:48:33] <cradek> but I don't have strong feelings and especially if someone else is maintaining the other locations, whatevs
[11:48:40] <Tom_itx> seb_kuzminsky, with that said, i added the 'how to' for mesa bit files to the wiki if you think that should be refined and added to the docs somewhere you're welcome to it
[11:49:08] <Tom_itx> it was previously just on my webpage
[11:50:21] <Tom_itx> not sure it belongs in the official docs...
[11:54:37] <seb_kuzminsky> cradek: can you remind me of the url to the live/install git repo again?
[11:54:41] <seb_kuzminsky> i promise to write it down this time
[11:54:50] <seb_kuzminsky> Tom_itx: link?
[11:56:12] <Tom_itx> http://wiki.linuxcnc.org/cgi-bin/wiki.pl?Editing_MESA_Bitfiles
[11:56:14] <Tom_itx> and
[11:56:16] <seb_kuzminsky> my intent with all the work i've been doing on the Getting Started documentation was to make it easier to go from "what's this linuxcnc thing i've heard about" to "ah, here's the docs that show me how to run a sim config"
[11:56:16] <Tom_itx> http://wiki.linuxcnc.org/cgi-bin/wiki.pl?Editing_MESA_Bitfiles_Spartan2
[11:57:24] <Tom_itx> i've been meaning to resize those images...
[11:57:38] <seb_kuzminsky> cradek: oh, i see it, n/m
[12:03:53] <seb_kuzminsky> i think i'm just allergic to joomla
[12:04:19] <cradek> understandable
[12:05:26] <seb_kuzminsky> http://wiki.linuxcnc.org/cgi-bin/wiki.pl?ReleaseCheckList
[12:07:00] <seb_kuzminsky> Tom_itx: those wiki pages are great
[12:07:08] <Tom_itx> thanks
[12:07:14] <seb_kuzminsky> They might fit in the Developer's Handbook
[12:09:32] <Tom_itx> i put them under 'advanced configuration' because i didn't know where else they would fit in
[12:10:01] <seb_kuzminsky> is that a part of the wiki? i meant the developer's handbook part of the git docs
[12:10:03] <cradek> wow those ARE great
[12:10:14] <cradek> I would have no idea how to do any of that
[12:10:44] <Tom_itx> first time is a bit intimidating but then it's easy
[12:11:25] <Tom_itx> i should add a bit about loading the new cards with mesaflash since mine used the ini file
[12:11:42] <Tom_itx> that may be covered somewhere else
[12:13:37] <seb_kuzminsky> the mesaflash package does not come with a manpage, that's probably a bug
[12:14:52] <seb_kuzminsky> hrm, the 2.6 git docs show the live/install image md5sum in two places, and they're different and both wrong
[12:15:59] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.6 1218005 06linuxcnc 10docs/src/common/Getting_LinuxCNC.txt docs: update the md5sum of the Live/Install Image * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=1218005
[12:16:09] <seb_kuzminsky> in 2.7 i changed the docs to only have the md5sum in one place, at least
[12:17:20] <Tom_itx> those pages may only be used by a very small % of users
[12:18:03] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 dacf09f 06linuxcnc Merge remote-tracking branch 'origin/2.6' into 2.7 * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=dacf09f
[12:20:16] <KGB-linuxcnc> 03Sebastian Kuzminsky 05master 0aca6a1 06linuxcnc Merge remote-tracking branch 'origin/2.7' * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=0aca6a1
[12:21:15] <seb_kuzminsky> i guess a normal newbie would type linuxcnc into google, which brings you to www.linuxcnc.org, and they'd click around there looking for info
[12:22:10] <seb_kuzminsky> the docs we keep in git and build into html & pdf, the ones that i've been focusing on, are a couple of clicks away from that, and not very prominent
[12:22:44] <seb_kuzminsky> you have to click on the Documentation tab at the top, then "All Documentation in html" link at the top of that tab
[12:23:14] <Tom_itx> i have found it hard sometimes to find the 'good' stuff
[12:23:18] <archivist> I hate the chase to html, would love that to be more direct
[12:23:23] <seb_kuzminsky> that takes you to the docs for the current release, which is 2.6
[12:23:53] <seb_kuzminsky> Tom_itx: what's the good stuff, in your opinion?
[12:24:03] <Tom_itx> customization things
[12:24:09] <Tom_itx> nothing in particular
[12:24:32] <Tom_itx> but that's to be expected on an open source project
[12:24:33] <archivist> I think the version in the url is important for checking stuff so all versions should be able to work that way
[12:24:49] <Tom_itx> alot of mods are one off
[12:26:11] <seb_kuzminsky> i think the docs we keep in git are of higher quality than the docs that live purely in joomla
[12:26:25] <Tom_itx> absolutely
[12:26:31] <seb_kuzminsky> i'd like it if more of our website was managed with git+asciidoc
[12:27:07] <archivist> removing joomla from the docs would be good
[12:27:51] <seb_kuzminsky> archivist: above you said "the version in the url is important", do you mean like in "http://www.linuxcnc.org/docs/2.6"?
[12:28:18] <seb_kuzminsky> we're very much wedded to joomla at the moment, changing that up would be a huge task
[12:28:22] <archivist> yes, you can edit that and see what was in an earlier version
[12:28:29] <seb_kuzminsky> basically reimplemeting our whole website
[12:28:34] <seb_kuzminsky> archivist: yeah, i like that too
[12:29:05] <seb_kuzminsky> there's also the clicky-friendly version-selection page here: http://www.linuxcnc.org/docs/
[12:29:10] <seb_kuzminsky> (which you probably know about already)
[12:29:28] <seb_kuzminsky> (although there's no link from joomla to that "docs front page")
[12:29:47] <archivist> I used that feature for my bot in another chan, just updated one table entry with the current version and all just works
[12:30:27] <archivist> I usually avoid joomla and use google to find the page I want
[12:30:54] <seb_kuzminsky> i usually avoid joomla and go straight to docs/$VERSION/html, and from there i know where to look
[12:47:43] <jthornton> seb_kuzminsky, you could just mention in the docs to check the md5sum on the download page
[12:56:38] <jthornton> I found in the beginning that having the same info in two or more places a huge nightmare to keep current. One would get updated and one or more would go stale.
[14:45:25] <skunkworks> is tclx a requirement of master now or just robs branch?
[15:12:07] <pcw_home> I just rebuilt master (Friday) and I had to add it
[15:12:42] <pcw_home> or maybe it was 2.7
[15:13:22] <skunkworks> ok - good.
[15:13:23] <skunkworks> thanks
[17:12:33] <seb_kuzminsky> skunkworks, pcw_home: tclx is required by 2.6 and later, i added it about a month ago
[17:12:49] <seb_kuzminsky> it's required to make ^C work right in the latency-histogram script
[17:13:04] <skunkworks> great!
[18:38:54] <alex_joni> ha http://blog.wenzlaff.de/wp-content/uploads/2015/01/php-foto.png
[18:39:26] <alex_joni> hopefully our depends-on graph is smaller ;)
[19:56:45] <pcw_home> Hmm what I thought was a hm2_eth bug is in fact another sserial bug
[20:12:58] <pcw_home> well not sure if pathologically slow code is defined as a bug or not