#emc-devel | Logs for 2010-07-06

[05:05:58] <ries_> ries_ is now known as ries
[10:26:19] <SteveStallings> SteveStallings is now known as steves_logging
[15:12:44] <SteveStallings> SteveStallings is now known as steves_logging
[18:01:19] <cradek> hi seb
[18:02:36] <seb_kuzminsky> heya
[18:03:05] <jepler> ooh it's seb
[18:03:13] <seb_kuzminsky> hi guys :-)
[18:03:37] <jepler> seb_kuzminsky: if you put together that care package for chris, I would take one of those integrated-driver stepper motors.
[18:03:49] <seb_kuzminsky> i
[18:03:53] <cradek> what care package?
[18:03:59] <seb_kuzminsky> i'd be happy to send you one, but what care package are you talking about?
[18:04:07] <jepler> cradek: seb told me he had stuff that he needed to send you
[18:04:08] <SWPadnos> heh
[18:04:10] <jepler> maybe I misunderstood?
[18:04:11] <seb_kuzminsky> the one with 100 lbs of cargo chains?
[18:04:17] <cradek> haha
[18:04:23] <SWPadnos> the other Chris?
[18:04:29] <cradek> don't you dare ship those
[18:04:33] <seb_kuzminsky> heh
[18:04:50] <seb_kuzminsky> maybe i'll sell them on craigslist and snail-mail you the cash ;-)
[18:05:23] <cradek> that would definitely be smarter.
[18:05:28] <seb_kuzminsky> i think we should try to get together in meat-space sometime this summer or fall and do care package transfer that way
[18:05:52] <jepler> that's fine, my project will languish for months anyway
[18:06:14] <seb_kuzminsky> shipping chains makes no sense, but shipping nema-17 steppers does
[18:06:25] <jepler> I agree
[18:06:34] <seb_kuzminsky> i'll see what remains in the box when i get home today and parcel-post some your way later this week
[18:06:41] <cradek> I was going to say a similar thing
[18:06:42] <cradek> my place is open, as is stuart's
[18:07:04] <cradek> seb_kuzminsky: I asked google, and it says my place and stuart's are about the same distance from you (1hr difference)
[18:10:18] <seb_kuzminsky> it was a drive-by micges
[18:28:41] <cradek> jepler: I forgot what you said the motor was for. was it something fun?
[18:29:07] <jepler> cradek: it's for the macro focus project
[18:29:14] <cradek> right
[18:36:28] <KimK> seb_kuzminsky: Hi Seb. Thanks again for all your help at the workshop. I got the encoder board driver (demo?) installed, but so far it seems to be showing only encoder 0. I had planned to connect 1, 2, and 3 anyway so I'll do that and then look for a declaration like "MAX_ENCODERS=1" or whatever it has. Once I've got all my ducks in a row, I'd like to send you some links to review and ask your advice.
[18:52:10] <cradek> KimK: do you have it reading an encoder already?
[19:05:51] <KimK> cradek: Well, yes, but I can't take any credit for that, it's in the supplied non-real-time driver (or demo?). But I was surprised when I moved the encoder cable around and only one encoder socket seemed to have any response. I just want to get it working "as-is" as a first step so I can verify it's good and so forth before I start fooling around with it.
[19:08:47] <KimK> Then I hope to get some good advice from Seb about how best to proceed to make a real-time-driver. Oh, and the existing driver is GPL or LGPL, I don't recall the specifics right now, though.
[19:09:02] <SWPadnos> what's the hardware?
[19:21:56] <KimK> Hi SWP. It's a USDigital PCI card with four differential quadrature encoder inputs. It is said to have 33MHz time-stamps, and 24-bit counters. It appears to use an FPGA a la Mesa, Pico, etc. Kind of pricey ($378), so I don't know how many people are going to rush out and buy one for EMC, but I would like to get it on the supported hardware list. (John threw $15 at it on eBay and it stuck! Were we surprised!) http://www.usdigital.com/products/interfac
[19:21:56] <KimK> es/pc/internal-card/pci-4e/
[19:21:59] <KimK> bah
[19:22:14] <KimK> http://www.usdigital.com/products/interfaces/pc/internal-card/pci-4e/
[19:22:58] <SWPadnos> ah, ok
[19:23:56] <micges> hello al
[19:23:57] <micges> l
[19:25:01] <micges> today I had weird problem with atemga88, external emf noise hang up ucontroller even if it has watchdog enabled
[19:25:41] <micges> anyone know how could it happen?
[19:26:11] <seb_kuzminsky> KimK: when you're ready, ping me on or off the emc-developers list and I'll take a look
[19:26:37] <micges> sorry wrong channel
[19:29:24] <KimK> SWPadnos: There is also a similar model with 3 encoder inputs where the "fourth" socket is used for a small amount of DC I/O (four-in, four-out), maybe good for a fast probe input? If I succeed with their 4-encoder model, I might ask USDigital if they will lend me their very similar 3-encoder model.
[19:30:07] <KimK> seb_kuzminsky: Thanks, I sure will.
[20:04:01] <cradek> isa version of the same thing? http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=320551945256
[20:06:07] <SWPadnos> there's no FPGA on that one, and of course the addressing is different vs. PCI
[20:39:56] <KimK> It does look kind of familiar, doesn't it? From the 5-pin sockets, it must be the S version (single-ended inputs). I'll watch it for amusement, but $150 for an *ISA* card? They'd better hope for an offer.
[20:40:08] <KimK> Hi Matt
[20:41:38] <KimK> Relatively huge breadboard area, though.
[20:41:48] <mshaver> Hey Kim!
[20:42:31] <KimK> Things getting back to normal after the workshop and 4th holiday?
[20:43:43] <mshaver> Question: If I have an encoder input and the "counts" or "position" shows up as a hal pin, is there any way to use this position in a gcode program?
[20:44:10] <mshaver> Yes, sort of. Tomorrow I have to drive back to Toledo to work on a machine.
[20:44:52] <KimK> Position, but not as an axis distance in the usual manner?
[20:45:38] <SWPadnos> "position" should be a float, which could be tied to an analog input of the motion controller, which can be read with M6<something>
[20:47:15] <KimK> Wow, M6 (toolchange?) has a "read analog input" option? I'll have to look that one up.
[20:47:29] <SWPadnos> M64/M65/something :)
[20:47:36] <KimK> Ah, OK
[20:47:39] <SWPadnos> (I just don't know what follows the 6)
[20:48:08] <mshaver> SWPadnos: I'll look into this!
[20:49:34] <mshaver> KimK: It's actually a digital micrometer which is supposed to measure a part, them the machine takes off "X" amount of material.
[20:49:51] <mshaver> then, not them
[20:50:23] <cradek> sounds interesting
[20:50:25] <KimK> Well, them better get it done, I say.
[20:50:32] <SWPadnos> you could use probing then
[20:50:47] <SWPadnos> hmmm, well maybe not
[20:50:49] <cradek> yeah a probe comes to mind right away
[20:51:07] <SWPadnos> but it's not an axis
[20:51:13] <SWPadnos> (though I guess it could be ...)
[20:52:46] <cradek> that's true too
[20:53:02] <mshaver> The way it works is that an M code actuates an air cylinder that pushed the mic body down depressing it's measurement plunger. The output is quadrature which goes to an encoder input. I could make it an axis if that would help.
[20:53:32] <cradek> the perversity of that makes me smile
[20:53:45] <KimK> That's interesting, so you're closing the "actually produced part loop"(?) with a micrometer measurement of the most-recently-produced part and then adjusting the cut of the next part to suit? Without adjusting the tool offset table? Or maybe you could do that too?
[20:53:46] <mshaver> (the plunger hits the top of a part in the fixture, thus measuring it)
[20:54:40] <cradek> I'd love to see a photo of this setup. I don't understand how you move the part from the machine to the fixture and back
[20:55:10] <mshaver> It's measuring the protrusion of a spark plug electrode. The machine then mills off another part establishing the spark gap.
[20:56:31] <cradek> http://www.linuxcnc.org/docs/devel/html/gcode_main.html#sec:M66-Input-Control
[20:56:38] <cradek> M66 E0 L0?
[20:56:52] <mshaver> The fixture is on the bed of the machine. You stick the spark plug in the fixture and run the gcode program (after I write it that is...) and it is moved to a position under the micrometer head. The mic is pushed down and the "height" of the center electrode shows up on an encoder input.
[20:57:12] <KimK> cradek: Ah. Spark plugs = produced in volume + competition = we're not likely to see a picture, lol
[20:58:04] <mshaver> Then it goes under the cutter where the height of the cut is proportional to the previously measured height of the center electrode.
[20:58:23] <cradek> sounds clever, but overly complicated when you could use a probe switch and the machine's own accuracy
[20:58:54] <cradek> probing vs measuring with a separate tool
[20:58:56] <mshaver> Then it goes in front of a camera for inspection, and then....
[20:59:16] <cradek> neato
[20:59:46] <mshaver> I didn't think this up. I'm just the "integrator".
[21:00:02] <cradek> can you use an extremely fast delta robot somehow too? then the youtube video would be even better
[21:00:08] <mshaver> Looking that M code up now...
[21:00:33] <mshaver> I'm thinking I can take pictures, etc.
[21:01:12] <mshaver> No ones mentioned anything about this being a big secret...yet.
[21:02:01] <KimK> Ha, "I'm just the integrator"... (it's that "tire swing" cartoon again, "What marketing sold", "What engineering designed", etc... finally---> "What the customer wanted")
[21:03:11] <KimK> Somebody handed Matt a mess and now he's got to make it work and come in under budget too.
[21:09:38] <mshaver> cradek: I think M66 will work. I just connect encoder.position to motion.analog-in-00 and then the M66E0L0 block latches the value in #5399, which I can use in the program. I hope all this stuff actually works :)
[21:10:02] <cradek> yeah me too
[21:11:53] <mshaver> KimK: It's not really like that. They've been really nice to work with, but they're used to spending money to solve problems, not seeking out the cheapest way to go, like I would.
[21:12:25] <cradek> you always have to wonder when you're about to use a feature used by approximately 0% of the user base
[21:12:47] <cradek> bbl
[21:13:51] <mshaver> It was going to be even worse. The micrometer interface modules they were looking at had either serial output, or BCD as a group of like 20 parallel bits!
[21:14:30] <mshaver> I looked through the catalog and found the quadrature output interface, much to my relief.
[21:16:46] <mshaver> I was beginning to think that I'd have to convert BCD as HAL bits, plus maybe a decimal point or something to a floating point number I could actually use.
[21:17:14] <mshaver> cradek: Thanks!
[21:42:03] <seb_kuzminsky> the buildbot just compiled emc2 master on lucid for the first time :-)
[21:42:10] <seb_kuzminsky> no packages yet though :-(
[21:43:03] <seb_kuzminsky> the lucid buildslave is using mozmck's rtai kernel on a 32-bit vm
[21:43:50] <seb_kuzminsky> KimK: it's that lucid vm i built while we were setting up your development machine at the cnc workshop
[21:50:56] <seb_kuzminsky> oops, it fails on realtime :-(
[22:19:07] <seb_kuzminsky> uhm, what's the secret location of the experimental lucid kernel packages, again?
[22:22:32] <JT-Hardinge> on the wiki
[22:22:34] <JT-Hardinge> ?
[22:22:46] <JT-Hardinge> the mozmck one?
[22:31:21] <KimK> JT-Hardinge: I can't tell you, it's a secret. Oh, wait, yes I can, lol. http://www.linuxcnc.org/mozmck/
[22:31:45] <JT-Hardinge> LOL, that is what I was hinting at :)
[22:35:30] <KimK> Hey, JT, are you doing EMC2 docs? I'm trying to learn git, and developing basics, and maybe help with docs at the same time; so I'd like to send some doc typo and doc errata patches your way (or someone's way). Shall I send it/them to you when I have any?
[22:36:22] <JT-Hardinge> yes, you can send them to me
[22:37:09] <JT-Hardinge> one second or three and I'll get you the email address
[22:38:58] <KimK> JT-Hardinge: And do you have the latest advice for me on which version of LyX is recommended? The last time I looked into this I recall seeing a note (where?) about using an older version of LyX to edit, is that still in effect? I'm hoping the current Synaptic one is OK, but I can downgrade if I must.
[22:40:05] <JT-Hardinge> the one on 8.04 is the one you have to use
[22:40:15] <JT-Hardinge> let me wander down to the beer cave
[22:44:15] <jthornton> KimK, did you get the pm?
[22:45:53] <KimK> Yes I did, thanks. So will there eventually be a big "project" to go through and "freshen" *all* of the docs once 10.04 is fully operational? (So that we're then using the LyX that goes with 10.04?) Just wondering how best to approach this, since 10.04 seems a lot better.
[22:50:42] <seb_kuzminsky> thanks KimK, that was the url i needed
[22:51:03] <seb_kuzminsky> btw i took the buildbot down because they're going to powercycle the building it's in tonight
[22:52:32] <jthornton> KimK, my thoughts at the moment is to split the G-Code (M and O too) to a separate manual.
[22:52:55] <jthornton> it just seems to be a separate subject from the GUI's etc.
[22:53:45] <jthornton> other than that I don't have any big changes in mind atm
[22:54:10] <jthornton> * jthornton wanders back out to the shop to finish setting up this job with only 4 tools :/
[22:54:13] <KimK> seb_kuzminsky: You're welcome. Ha, power, it's always something.
[23:11:35] <JT-Hardinge> KimK: I think that only dev will be able to use 10.04 LYX at some point and time but not now.
[23:13:33] <KimK> JT-Hardinge: OK. And as far as the 8.04 LyX version, we will continue to use that until 8.04 is no longer supported? So that at some point (like for 6.06 now? soon?) there will be a "final" set of manuals that will remain there for whoever wants them, written in whatever they were written in, but they will no longer be updated?
[23:15:20] <JT-Hardinge> I "think" (and jepler would know) that when it makes sense to do so we can start using the LYX version in 10.04 with DEV but would still have to use 8.04 with EMC 2.4.x
[23:17:13] <KimK> OK. Well, I'll go trim a few castings and wait to see if jepler has any advice for us later. I'll check back.
[23:17:40] <JT-Hardinge> .223?
[23:18:21] <KimK> Ha, a reloader. No, castings this time, not casings.
[23:19:14] <JT-Hardinge> heh, I read between the letters
[23:20:22] <KimK> These castings are about the size of a big (thick) paperback novel or dictionary, and have sprues or runners on two sides that I have to trim off with a bandsaw before I can even do anything with them.