#emc-devel | Logs for 2010-08-15

Back
[00:17:09] <mozmck> I think I have a candidate for a final live CD now.
[00:18:40] <cradek> yay!
[00:20:48] <mozmck> It now gets the emc2 repositories in there, and the theme has been changed to Clearlooks with the buttons on the right side.
[00:21:28] <mozmck> I fixed the md5sum problem so it verifies correctly, and I fixed it so the installer won't download and install another kernel
[00:22:12] <mozmck> also added mesa-utils, and I think that's about it.
[00:23:21] <mozmck> For the kernel fix, I simply commented out the call to remove_unusable_kernels() in /usr/share/ubiquity/install.py
[00:24:17] <mozmck> If anyone can verify that that should not break something, I'd appreciate it. I installed on one computer with no problems, and ran the livecd as well.
[00:41:22] <mozmck> I renamed the iso from ubuntu-10.04-desktop-i386-linuxcnc.iso to ubuntu-10.04-linuxcnc-i386.iso
[00:43:33] <mozmck> bbl. if there are no questions or objections about the way I disabled kernel updating as mentioned above when I get back, I'll upload the iso later tonight.
[01:15:25] <Dave911> SW: I wasn't interrupting the "checking the disk for errors" process, just the boot process itself, when interrupted was causing the disk to corrupt .. the later updates fix that .. I can't kill it after doing the updates.. I think I have reinstalled Linux on this flash card about 10 times now.
[01:33:03] <cradek> mozmck: hm, looks like I don't have that file - probably would have to boot live to see it. I did not hear what the original problem was...?
[02:41:32] <mozmck> cradek: the original (possible) problem is that ubuntu would download and install a -pae kernel for systems with more than 3 gig ram
[02:42:26] <cradek> oh it would do that at install time, and it would end up booting instead of yours?
[02:42:33] <mozmck> That function looked like it could also install an updated kernel of the same one if available.
[02:42:34] <mozmck> yes.
[02:42:43] <cradek> gotcha
[02:43:13] <mozmck> so by commenting out that line in install.py, it will not check for different kernels but just install what's on the CD.
[02:43:18] <mozmck> as far as I can tell.
[02:43:24] <cradek> sounds perfect
[02:44:50] <mozmck> ok, I'm going to start this iso uploading in a little bit. It is named differently as I mentioned, so that should help with knowing that you have the right one. Should I add a version number to the name? like -linuxcnc1
[02:44:58] <cradek> aha, I can get source for ubiquity, looking at it now
[02:47:15] <cradek> hm, a lot of stuff in that function...
[02:47:16] <mozmck> It's a little contorted. There's a file called check-kernels in that same directory, and /usr/lib/ubiquity/base-installer/kernel.sh
[02:47:31] <mozmck> yep.
[02:48:01] <cradek> what is /var/lib/ubiquity/install-kernels?
[02:48:55] <cradek> (it's not in my source package)
[02:48:58] <mozmck> that is written in one of those two scripts I just mentioned, with kernel package names.
[02:49:51] <cradek> ugh, I'm not going to be able to be any help without burning a cd and booting it somewhere - and I trust your testing more than my python-reading
[02:50:29] <mozmck> heh, it's kind of a tangled mess it seems to me.
[02:50:51] <cradek> yes
[02:51:24] <cradek> did you test the cd's media check too?
[02:51:46] <mozmck> what's that? I did the disk verify on bootup.
[02:51:48] <cradek> (the md5sum.txt problem)
[02:51:50] <mozmck> yes
[02:51:55] <cradek> right that's what I meant
[02:51:58] <cradek> cool.
[02:52:06] <cradek> I bet you've nailed it :-)
[02:52:36] <mozmck> I hope so. It's about time :)
[02:53:05] <cradek> I can't wait for everyone to start using it. Lucid is nice.
[02:54:48] <cradek> I really like how fast my mill boots now.
[02:55:10] <mozmck> yeah, the fast boot is very nice.
[05:32:11] <CIA-5> EMC: 03cmorley 07v2.4_branch * r16b99eb3439a 10/src/emc/usr_intf/pncconf/pncconf.py: fix tool table to work with random tool changer
[06:57:21] <CIA-5> EMC: 03cmorley 07v2.4_branch * r44bcdf232c1c 10/src/emc/usr_intf/pncconf/pncconf.py: fix sensitivity of invert encoder
[06:57:22] <CIA-5> EMC: 03cmorley 07v2.4_branch * r441b0d8478e1 10/ (9 files in 6 dirs): Merge branch 'v2.4_branch' of ssh://[email protected]/git/emc2 into v2.4_branch
[06:57:23] <CIA-5> EMC: 03cmorley 07v2.4_branch * r573a7ffd9547 10/src/emc/usr_intf/pncconf/pncconf.py: fix typo from last commit to pncconf
[06:57:23] <CIA-5> EMC: 03cmorley 07v2.4_branch * r7b056ead747a 10/ (9 files in 7 dirs): Merge branch 'v2.4_branch' of ssh://[email protected]/git/emc2 into v2.4_branch
[08:12:57] <alex_joni> * alex_joni watches bits dribbling in from micges
[08:13:01] <alex_joni> * alex_joni watches bits dribbling in from mozmck
[08:13:04] <alex_joni> sorr micges ;)
[08:13:17] <micges> heh
[13:08:03] <mozmck> alex_joni: fast upload huh? I put in on and went to bed. Looks like it uploaded fine now.
[13:50:32] <mozmck> Theres a new lucid iso at http://www.linuxcnc.org/mozmck/ I think this will be the final one.
[15:06:26] <cradek> mozmck: yay thanks! I downloaded it and the md5sum matches. I will try to find something weird to install it on later!
[16:49:41] <Jymmm> alex_joni: bits?
[16:50:08] <Jymmm> alex_joni: Care for a flash flood ? LOL
[18:26:05] <Dave911> If I setup a RIP EMC installation of the master and and then remove an "installed" version of EMC2 the menus go away.. What is the command to reinstall them?? Was it "make menus"? Is that written anywhere? I did a search on the wiki and couldn't find it...
[18:26:39] <SWPadnos> make <tab> <tab> might list all the available targets
[18:27:29] <Dave911> excellent .. I knew there was something else I should know.. :-) I'll try that..
[18:31:13] <Dave911> That works .. thanks..
[18:32:17] <SWPadnos> sure
[18:32:50] <SWPadnos> hey, did morfic get in touch with you recently (like yesterday or today)?
[18:47:10] <Dave911> I've had an open IRC chat window with morfic for a couple of days now.. The last exchange was Sat morning I think..??
[18:47:33] <Dave911> I'm a little confused ... about what he is confused about.. :-)
[18:49:30] <SWPadnos> heh
[18:49:50] <Dave911> I've got my RIP setup ok.. now the Mesa board driver can't find my bit file.. is there an official package that can be installed to load the bit files and link them? I seem to run into problems with these bit files everytime I setup a new PC with a RIP installation
[18:49:59] <SWPadnos> he thought you and I disagree on what is necessary in a spindle drive, if you want to use it both as a spindle and as an indexer/C axis
[18:50:20] <SWPadnos> yes, there are various emc2-firmware-* packages
[18:51:56] <SWPadnos> do you agree that a drive that can servo at zero speed can do both, with EMC2 closing the position or speed loop, as appropriate?
[18:52:20] <Dave911> I doubt that is true.. my question is how do they do (the spindle vendor) and what do they consider positioning ... I think morfic wants full servo positioning on the spindle.
[18:52:49] <SWPadnos> yes, I think that indexing and "contouring" are both goals
[18:53:06] <Dave911> Right .. that is what I thought..
[18:53:07] <SWPadnos> it seems that a servo drive/motor should be able to hold position with full torque, no?
[18:53:57] <Dave911> Yaskawa gave him a quote with a servo motor for the spindle drive that has no rating beyond 2000 rpm.. and they want about 4500 rpm max I believe..
[18:54:10] <SWPadnos> yeah, that was a mistake
[18:54:19] <SWPadnos> it seems the rep is a bit of a bonehead or something
[18:54:58] <Dave911> That's what I thought also ... I think he just needs to revisit that with Yaskawa if he wants to continue considering them
[18:56:10] <Dave911> There was another Yaskawa spindle drive, the M5 which can do positioning but it was not quoted.. I don't know why..
[18:56:12] <SWPadnos> I guess his confusion was that I had said that a drive that can hold position (if given a zero velocity command) can be used for indexing/contouring, and that EMC2 can close the loop - the drive doesn't need a "positioning" mode
[18:57:14] <Dave911> I know what you mean.. some of the spindle drives have positioning modes to overcome their own limitations since they are not true servo drives ....
[18:57:23] <SWPadnos> ah, ok
[18:57:59] <Dave911> I think Morfic is truly looking for C axis functionality from what he has told me..
[18:58:07] <SWPadnos> also, positioning mode requires a "smart" drive (AFAIK), which needs to be commanded to a position with either a serial command of some sort or step/dir (since PWM/analog aren't so great for position commands)
[18:58:10] <SWPadnos> yes
[19:00:07] <SWPadnos> in order from slow to fast, I think he wants all of indexing, C axis, and spindle
[19:00:27] <Dave911> Yes ... I never got into that with Morfic.. but yes I know the "positioning modes" of some of those drives use non-standard interfaces which might or might not be supportable by EMC2.. I agree...
[19:00:28] <SWPadnos> (indexing can be done with a brake, which is why I list it separately from C)
[19:01:18] <SWPadnos> so, you seem to have a concern that a velocity-mode drive that doesn't explicitly say that it can do positioning/orient might not work well at slow/zero speed - is that true?
[19:10:42] <Dave911> Sorry.. I was called away... Yes.. I think that Morfic needs to make sure that the drive can actually do servo positioning.. torque at zero speed..
[19:10:49] <robh__> id look at Yaskawa G7 VFD
[19:11:21] <SWPadnos> he's been getting some help from a Yaskawa rep, but it hasn't been the most helpful help there is, if you know what I mean
[19:11:37] <robh__> he needs to speak to a CNC side guy
[19:11:52] <Dave911> I know exactly what you mean... some salemen know what they sell and some do not ...
[19:11:55] <robh__> we spoke to a rep he was up on the CNC as they do supply alot of CNC builders with spindle setups
[19:12:39] <Dave911> I suspect that not a lot of thought was put into Morfic's Yaskawa quote..
[19:12:51] <robh__> controltechniques also do some smart and cleaver drivers,
[19:13:37] <robh__> i found when talking of CNC side use, you need someone that knows there products well, we have been there done that, you also need to drill out of them what it can truly do not what the sales sheet says
[19:13:54] <Dave911> I think that several suppliers can meet his needs... he just needs to be careful with his choices.. and he is..
[19:14:04] <Dave911> robh_: I totally agree..
[19:14:18] <robh__> yaskawa servo like drives switch mode in lower speeds to give u full torque to do servo like performance, then switch back in high speed to do 6K or more rpm
[19:14:54] <Dave911> SW: Do you know where the Mesa bit file packages reside?
[19:15:13] <SWPadnos> in the emc2 repository
[19:15:21] <SWPadnos> maybe not for lucid
[19:15:37] <SWPadnos> but you might be able to install the ones for hardy instead, I think they're the same
[19:16:55] <Dave911> OK... I thought they were somewhere else .. I can find those.. Thanks..
[19:18:11] <SWPadnos> sure
[20:39:10] <alex_joni> SWPadnos: for lucid too
[20:39:28] <SWPadnos> ecellent
[20:39:32] <SWPadnos> +x
[20:39:35] <alex_joni> they are called: hostmot2-firmware-* or hostmot2-firmware-all
[20:40:42] <alex_joni> whee DH is fast :D
[20:40:52] <alex_joni> started a download for the LiveCD
[20:41:06] <alex_joni> 40% done
[20:41:16] <alex_joni> 60% done
[20:41:30] <alex_joni> 80% done
[20:41:52] <alex_joni> fully done
[20:42:00] <alex_joni> Downloaded 675.7 megabytes in 1:20 seconds. (8569.22 KB/s)
[20:42:22] <SWPadnos> damn!
[20:42:25] <SWPadnos> you :)
[20:42:35] <alex_joni> md5sum matches
[20:42:41] <alex_joni> that's on the EU mirror
[20:42:49] <alex_joni> I have more decent speeds at home :P
[20:43:04] <SWPadnos> I have 2 Mbit/s max :(
[20:43:40] <alex_joni> this is only limited by the eth interface, if we added a Gbit NIC, it would work faster
[23:38:18] <Dave911> Hostmot2 files ......I found them Alex .. thanks..
[23:39:16] <Dave911> Alex... that is a blazing internet connection.. What is DH?
[23:39:17] <Dave911> It took me 33 minutes to download that iso..