#emc-devel | Logs for 2010-06-13

[01:47:19] <jmkasunich> alex_joni: yes, the compile farm is dead
[03:40:19] <cradek> I worked all day with 2.4.1 and never spotted even a little buglet
[03:42:19] <cradek> using touchy, lots of mdi, probing, canned cycles, a bit done in polar coordinates
[03:45:33] <KimK> Glad to hear it. I ran it again too, but same parts, doing same thing, no new info gathered.
[03:48:31] <cradek> do you have the kind of jobs where you can set up and just leave it for a long time?
[03:55:00] <KimK> Not on this job, it's trimming down and squaring up castings for grinding. First op is on a manual mill to get one flat starting surface. Then into the BP2 Chick vise two at at a time. Do 2nd side, flip, do 3rd side. About 1-2 minutes a side maybe? It's always "done", lol.
[03:56:26] <cradek> sounds a bit finicky but not too bad
[03:59:05] <KimK> Yes, not bad, just hard to walk away from. In the building across the street there's a Haas horizontal mill with a 6-pallet pallet pool. You can load that thing up and it'll be awhile before it's done.
[04:00:06] <KimK> Typically one 4-sided tombstone per pallet, so, lots of parts.
[04:01:23] <KimK> What are you working on tonight?
[04:01:32] <KimK> with 2.4.1
[04:02:06] <cradek> I made a mount for a high speed spindle on my little 5 axis desktop mill
[04:02:28] <KimK> Cool, how high is high speed?
[04:02:46] <cradek> I think it's about 20k, 1/8" shank tools only
[04:03:06] <cradek> it's a good little engraving machine so 1/8 only is no big loss
[04:03:22] <KimK> Well, 1/8 shank is no problem at those kinds of RPMs.
[04:03:46] <KimK> Pretty soon you're going to need a spindle chiller, lol.
[04:04:38] <cradek> the motor I had on it would run the original spindle nearly that fast, but after being on for a couple hours it was quite ready to melt
[04:04:52] <KimK> Am I bid 30,000 RPM? 40,000 RPM? 50,000 RPM going once... going twice...
[04:05:08] <cradek> we've been doing a lot of tiny detailed stuff, so many hours of runtime
[04:05:12] <cradek> hopefully this will work better
[04:05:40] <KimK> Stuff for yourself or paying jobs, lol?
[04:05:56] <cradek> with any luck, paying eventually
[04:06:26] <KimK> Ha, good, that's the kind to have. Do you know where you're staying in Ann Arbor?
[04:06:39] <cradek> yes, we're staying at the mumble mumble motel
[04:06:58] <cradek> jeff and I have reservations already but I forget the name :-/
[04:07:53] <KimK> Ha, OK. I guess I'm trying the dorms to save money, if they're not full yet.
[04:08:09] <cradek> you didn't call yet? I bet you should...
[04:08:38] <cradek> that was definitely tempting - it's a good price
[04:08:56] <KimK> Yes, you're right, that's first thing Monday's job.
[04:09:45] <KimK> Hopefully their summer enrollment is low and they'll be mostly empty.
[04:10:10] <cradek> I got the impression they'd be completely empty except for workshop folks - not sure though
[04:11:23] <KimK> Yeah, we'll see, I hope so. Maybe I'll be camping somewhere instead?
[04:13:04] <KimK> Do you have any idea how many EMC folks (users or devels) are coming? I'll send a reminder to the mailing list on Sunday or Monday to edit the 2010 wiki if you're coming.
[04:13:44] <cradek> I know jeff, john k and I are going
[04:14:00] <cradek> swp is going to be there a few days at the beginning
[04:14:21] <KimK> OK. And Moses couldn't make it, right?
[04:14:22] <cradek> seb yes, alex no, ray no, moses no
[04:16:06] <KimK> Yes, I just discovered Ray couldn't make it, too bad for us. You can add me to the pool if there's a list of people trying to cover for Ray.
[04:16:06] <cradek> I'm struggling to figure out what to bring
[04:16:52] <cradek> I have a feeling you'd be very good at giving beginner classes, especially for people retrofitting real machines
[04:17:12] <cradek> I'm not sure what interest there is in that, compared to desktop machines, compared to total homebrew toys
[04:18:01] <cradek> I don't think stuart knows for sure yet if he's coming
[04:19:53] <KimK> I'd be happy to do that, but I think most users are running Nema 23 tabletops with the StepConfig Wizard(?), and I haven't done much with that. I should find time to fool with that before the workshop.
[04:21:37] <cradek> realistically I don't think I have what it takes to do a class. I have enough trouble helping people one-on-one.
[04:22:53] <KimK> Yes, well, I wouldn't refuse help myself. It could easily take several people to cover for Ray.
[04:23:51] <KimK> Bring your (jepler's?) Etch-A-Sketch?
[04:24:06] <cradek> I wonder what he was planning to cover. I haven't heard from him for so long so I don't know how up to date he was on the latest stuff.
[04:24:27] <cradek> that was jepler's - I don't know if he still has it.
[04:24:57] <cradek> I could bring my little lathe - but why bother? everyone's seen a lathe.
[04:25:15] <cradek> lathe support is not new and cool anymore :-)
[04:25:23] <KimK> The Etch-A-Sketch is small and doesn't need 220V, in case that's an issue.
[04:26:28] <KimK> (Something about having one cord/socket and looking for some matching plugs on the Digital Machinist official pages.
[04:26:54] <cradek> ?
[04:27:14] <KimK> So it looks like lots of 120V extension cords and power strips will be the order of the day.
[04:27:31] <cradek> ah ok, no problem
[04:28:02] <KimK> I mean we should bring our own, they might not have anything?
[04:28:17] <cradek> sure, it never hurts to bring power and network cabling
[04:28:37] <KimK> the cord/socket business above was for 220V, as I understood it.
[04:28:53] <cradek> nothing that requires 220 is going in our car, that's for sure
[04:29:51] <KimK> And I don't know whether he meant 240V or 208V, but I suspect 208V at a college.
[04:31:02] <KimK> Back to what you said earlier: So what *is* new and cool?
[04:31:39] <cradek> good question
[04:31:42] <KimK> (Then I'd better go, I need food and sleep, lol)
[04:32:08] <cradek> touchy is new and cool
[04:32:22] <cradek> polar coordinates are very cool if you do any circular stuff
[04:33:04] <KimK> I guess everyone is expecting a new install CD for 10.04? That would be *really* new, lol.
[04:33:27] <cradek> a lot of gcode-generating free software is coming out of the woodwork - like how inkscape is becoming useful
[04:33:51] <cradek> I don't know if moses is going to get that done. his packages sure seem to work, fwiw
[04:34:16] <KimK> I haven't tried polar yet, I'll look forward to it. And (your?) coordinate rotation system?
[04:34:37] <cradek> yeah, that works, but I never use it
[04:35:02] <KimK> Who do you know (besides Stuart) who knows about how to get started using APT?
[04:35:03] <cradek> people seem to drill hole circles a lot, and now you can do it in polar with one line of gcode
[04:35:30] <cradek> there are some guys in #cam maybe?
[04:36:52] <KimK> Really, a bolt circle in one line? Yes, I'm in #cam a lot too, I thought maybe you'd know someone (and that they were coming, lol). I got a couple of used books on APT, we'll see if they do me any good.
[04:37:15] <cradek> I did write a program in apt once and I got gcode out of it
[04:37:19] <cradek> that was enough for me, I guess
[04:38:30] <cradek> yes one line: like G91 G81 R.1 ^60 Z-1 L6 (repeat 6 times, incrementing 60 degrees each repeat)
[04:39:35] <cradek> ... after you position over the first hole first, like G0 X3 Y0 or G0 @3 ^0
[04:40:01] <KimK> John's brother used to program in APT, he assures me it's very powerful. And now that there's a viewer (almost) available, I thought it might be worth looking into.
[04:40:15] <KimK> So L6, is that a 12" bolt circle then?
[04:40:30] <cradek> L is the number of holes
[04:40:37] <KimK> Oh, no the repeat, I see.
[04:40:46] <KimK> Where's the vector length?
[04:40:50] <cradek> if you are at X3 when you start you'll get 6" dia
[04:41:03] <cradek> it's incremental in angle only - radius is unspecified in that G81 line
[04:42:07] <KimK> And you start at zero or the first hole? (Wait, I should rtfm, sorry.)
[04:42:08] <cradek> just like you can repeat like G91 X... L... with Y unspecified to make a horizontal row from wherever
[04:42:37] <cradek> if you are at R0, incremental angle doesn't mean anything, it's an error
[04:42:48] <cradek> so you have to start at the radius you want
[04:42:59] <KimK> how does it know where the center is?
[04:43:03] <cradek> er not R0, I mean radius 0 = @0
[04:43:17] <cradek> center of the current system
[04:43:36] <KimK> Oh, OK, G54-G59.3?
[04:43:40] <cradek> yes
[04:44:18] <cradek> if you're working on something round, there's one obvious way to set your work coordinate system...
[04:44:34] <KimK> Very cool, it should eliminate a lot of extra calculation.
[04:46:14] <cradek> http://git.linuxcnc.org/gitweb?p=emc2.git;a=blob;f=nc_files/spiral.ngc;h=7a3a4a31f25a3e14925a9a41e56b4409e5a82668;hb=master
[04:46:42] <cradek> this is kind of cool: spiral done in incremental polar
[04:48:25] <cradek> nurbs/splines are cool and new (but really not useful for handwritten gcode)
[04:51:10] <KimK> Speaking of fixture offsets, I discovered the XYZ coordinate zero icon moving around on the axis display as I changed fixtures, I never noticed that before, or maybe never ran in show relative mode? Anyway, nice!
[04:51:34] <cradek> ah, cool
[04:51:45] <cradek> it rotates too, if you have some of those systems rotated
[04:54:36] <cradek> oh one more for roll call: I think Jon E is coming
[04:56:31] <KimK> Is the axis program window slated for any improvements? It didn't seem to step down the program line-by-line but skipped to certain lines(?) And it didn't rewind at the end, I think. Just a little different than most CNC controls I've used.
[04:57:03] <KimK> I think I saw Steve Stallings on the exhibitors list too.
[04:57:35] <cradek> emc's task controller only lets the gui know what line is currently *causing motion* so you'll never see an M3 or G90 line highlighted
[04:57:43] <cradek> I'm guessing that's what you noticed
[04:57:54] <KimK> Oh, OK.
[04:58:23] <cradek> it's not gui-specific and when we (?) eventually rewrite the task controller all the guis will work better.
[04:59:29] <KimK> OK, I'll look forward to it. (Maybe by that time I'll be able to help a little?)
[05:01:20] <KimK> I'm glad you and seb are both coming, you make good tutors. (And there's two of you to keep handing me off, so I don't wear you out, lol).
[05:01:31] <cradek> haha
[05:02:01] <KimK> Do you have any specific goals this trip?
[05:02:10] <cradek> it'll be fun to catch up with seb. he hasn't been around much since he got his mill. (COINCIDENCE?)
[05:02:36] <cradek> yes - take lots of time off work and have some fun
[05:02:45] <KimK> "To a man with a hammer, every problem looks like a nail"
[05:03:31] <cradek> if moses was going, I'd take my banjo or guitar...
[05:03:57] <cradek> I don't have any big specific emc goals, I guess. I have a few things in mind I might like to work on though.
[05:06:33] <KimK> There's always that "Webcam" line on the EMC fest pages, I copied it again when I started the 2010 page. Was there ever one? I don't recall anything but a time-lapse cam. It would be nice if there was a good way to telepresence in people like mozmck , alex_joni , maybe even Ray.
[05:07:08] <KimK> I did move "webcam" down because I wasn't sure if it ever happened, really.
[05:07:09] <cradek> we did have a live cam in ... uh '07?
[05:07:40] <KimK> Oh, OK, that's why I don't remember it. 2008 was my first visit.
[05:11:26] <KimK> Well, I think I've bothered you enough for now. Food and sleep are calling. Thanks for the chat. I'll be around again tomorrow, off and on at least. John and I still need to work on those castings.
[05:11:39] <cradek> goodnight!
[05:11:57] <KimK> bye
[05:28:16] <ries_> ries_ is now known as ries
[12:19:18] <jthornton> any place online to view 2.3 docs?
[12:25:49] <jthornton> or was calling named files with o call allowed in 2.3?
[13:26:23] <jepler> jthornton: http://linuxcnc.org/docs/2.3/html/
[13:26:25] <jepler> jthornton: beats me
[13:38:34] <jthornton> seems like it is in the 2.3 docs so it must be true
[13:52:36] <jthornton> looks like o<MyHumpyFileName> call fails to load the file due to converting it to lower case to seek the file or something like that. While o<mylowercasefile> call works fine. Should I document the lowercase requirement?
[13:55:32] <jthornton> bbl
[14:58:33] <steves_logging> steves_logging is now known as SteveStallings
[14:59:21] <SteveStallings> Hi KimK, and other potential workshop volunteers. I have been in contact with Ray about his lab/workshop.
[15:00:40] <SteveStallings> PMDX will be furnishing 5 of the switch/LED/test boards like he used last time he gave the lessons. Ray has promised to send me notes and/or handouts, but nothing has arrived yet.
[15:01:21] <SteveStallings> The school is furnishing computers and we are free to load Linux and EMC on them. The get wiped at the end of the show.
[15:02:05] <SteveStallings> What we need is people to run the sessions. I am unfortunately still to ignorant of EMC to try to teach others. Matt has offered to help.
[15:03:00] <SteveStallings> In discussions with George from HSM Digital Machinist, it has been decided to scratch the Tuesday session to give us time to try and be organized for sessions on Wed, Thurs, and Fri.
[15:04:17] <SteveStallings> The school has confirmed that the computers that we will be using do have old fashioned DB-25 parallel ports.
[15:06:33] <SteveStallings> Matt and I will be arriving around lunch on Monday, and will be staying at the dorm.
[15:08:13] <SteveStallings> The EMC classes are scheduled for 10 AM untill noon and will be in a classroom. The EMC lab/demo area may also be in that room, or maybe just outside it where PICO and PMDX have been assigned vendor tables.
[15:16:37] <cradek> hi steve
[15:16:45] <cradek> glad to hear you'll both make it
[15:16:53] <cradek> it's close to you, for once
[15:18:01] <jthornton> cradek: do you know if o<myfilename> call is supposed to allow caps in the name?
[15:18:35] <cradek> if I were to guess, I'd say no
[15:18:40] <SteveStallings> close is a relative term, still 9-10 hour drive
[15:18:58] <skunkworks_> cradek: http://www.youtube.com/watch?v=MucNO3Dij9o
[15:19:00] <jthornton> ok it seems like it converts the name to lower case
[15:19:09] <cradek> that doesn't surprise me at all
[15:19:15] <jthornton> I'll add that to the docs for now
[15:19:44] <cradek> skunkworks_: !!!
[15:20:05] <cradek> what are those blinks? barcode reads?
[15:20:34] <skunkworks_> yes
[15:20:34] <cradek> then it has low speed until locking in place?
[15:20:38] <skunkworks_> yes
[15:20:42] <cradek> sweet!
[15:20:58] <skunkworks_> Now I need to test my ladder. maybe this week
[15:21:01] <cradek> do you have it prepping right already?
[15:21:13] <skunkworks_> no - that was just running it manually
[15:21:18] <cradek> ah
[15:21:31] <cradek> you're 90% done if you can do that
[15:21:43] <skunkworks_> I have the ladder - but I didn;t have it with me. (didn't actually think we would get that far)
[15:21:48] <cradek> heh
[15:23:18] <skunkworks_> the only casualty was a audio tape that fell between the pallet and the hardstop... when pressure came up - it pulled the pallet to its resting postion and smashed the tape ;)
[15:23:45] <skunkworks_> counter balance seems to still work also :)
[15:24:27] <cradek> is this the first time you've had hydraulic pressure for a while?
[15:24:36] <skunkworks_> yes - few years
[15:24:47] <cradek> wow, and it works?
[15:25:00] <cradek> that's a relief I bet
[15:25:02] <skunkworks_> scary, huh?
[15:27:19] <micges> jthornton: in interpreter code there is close_and_downcase() function executed on all gcode lines, so for sure it isn't allowed to use caps for named interp variables
[15:27:41] <micges> you should mention this in docs
[15:28:07] <jthornton> micges: thanks for checking that I'll update the docs as quick as I get off this temp dialup connection
[15:28:43] <micges> cool
[15:29:52] <skunkworks_> We even got the computer hooked to the network and updated to 2.4.1 - only issue was deleting the emc.nlm or whatever from the ini.
[15:33:19] <cradek> cool. I'm definitely happy with 2.4.1.
[15:33:50] <skunkworks_> I got the extop loop working using the sample also.
[15:33:58] <skunkworks_> worked great
[15:34:10] <cradek> great
[15:35:03] <skunkworks_> the thing sounds like a waterfall with all the oil running ;) (we still have some access plates off )
[15:35:33] <cradek> does it recirculate lube or something? or does it just use a lot?
[15:35:58] <skunkworks_> lots of lube oiling
[15:36:14] <skunkworks_> it sprays all over the place within the machine.
[15:37:11] <cradek> the jr is so cool - it has a trough all the way around, everywhere lube drips. it collects it in a little cup for you to dump out.
[15:37:17] <skunkworks_> (there is one pump that runs almost everything (lube, counter balance, shifting, everything except pallet and collet clamp.)
[15:38:40] <skunkworks_> the hydralic fluid is used for lube and it all gets returned thru huge filters to the supply. (actually has a scavenger pump that pulls the oil out of the bottom of the table, upright)
[15:38:52] <cradek> ah, neat
[15:39:05] <cradek> I thought you were saying it just dumps it
[15:39:20] <skunkworks_> not on purpose ;)
[15:39:22] <cradek> have you moved an axis yet?
[15:39:30] <skunkworks_> no - getting closer. :)
[15:39:54] <cradek> how many hours from workshop is your place?
[15:39:59] <skunkworks_> maybe dad will get that going this week ;) Hi Dad!
[15:40:13] <skunkworks_> I think dad calculated it at 7 hours or so
[15:40:20] <skunkworks_> about the same to wichita
[15:40:27] <skunkworks_> or was it 8
[15:40:37] <cradek> darn, that's too far to run up and see it
[15:40:53] <skunkworks_> are you going thru chicago?
[15:41:18] <cradek> at least in one direction, probably both
[15:41:57] <cradek> I suspect we'll be in chicago thursday sometime
[15:42:50] <cradek> how close are you to chicago? (I still have no idea where you are)
[15:44:07] <skunkworks_> we would be 3hours north of davenport IA
[15:44:40] <skunkworks_> (figured you would be going through there....)
[15:44:57] <skunkworks_> Near Lacrosse WI
[15:45:28] <skunkworks_> I forget - are you bringing the bus?
[15:45:41] <cradek> no, probably jeff's car
[15:45:45] <skunkworks_> ah
[15:46:29] <cradek> going through lacrosse adds 4.5h to our drive to cleveland
[15:46:33] <cradek> (ouch)
[15:46:38] <cradek> maybe another time
[15:47:35] <skunkworks_> heh - that is ok. It would be nicer if we where a lot further along :)
[15:49:08] <skunkworks_> we may be out your way the 4th of july week. Camping trip to see shirleys dad. (we have not figured out the exact route yet though)
[15:49:49] <cradek> you better stop if you're close. :-)
[15:54:09] <skunkworks_> :)
[19:00:26] <micges> cradek: there are at least two bugs in emc
[19:00:45] <cradek> I'm not at all surprised to learn that
[19:01:38] <micges> sorry I don't understand that sentence :P
[19:03:17] <micges> one is that emccalib is interpreting commented lines
[19:03:35] <micges> in hal files
[19:06:09] <micges> second is that stepconf is generating hal lines like: loadrt hal_parport cfg="0x378 out "
[19:06:55] <micges> emccalib don't interpret that line also but it's valid hal scripts syntax
[20:20:11] <jepler> micges: please test: http://emergent.unpy.net/files/sandbox/0001-emccalib-fix-crash-unmatched-open-quote-in-list.patch
[20:29:41] <micges> jepler: works fine here but I have only sim, tomorrow I can test on rt machine with parport
[20:30:21] <micges> jepler: emccalib also interprets comments which is bad
[20:36:15] <kgartner> so I discovered a bug when you try to pass ctrl_type to stepgen, noticed there was some chat on the mailing list about it a while back, there was a patch suggested and it seems to work, did it just never get applied to the git tree?
[20:38:31] <kgartner> or at least it wasn't in with the 2.5.0~pre version I have
[22:18:59] <jepler> kgartner: I'd forgotten all about that
[22:19:18] <jepler> I never heard back positively whether the patch I'd proposed in this message fixed the problem for the original reporter: http://mid.gmane.org/20100128230457.GG22644@unpythonic.net
[22:20:15] <jepler> acemi's proposed patch, changing the type of array from array-of-charp to array-of-int, wasn't suitable for inclusion because it breaks backwards compatibility.
[22:39:09] <kgartner> gotchya, makes sense. Well I can load the module without crashing the kernel anymore, but I haven't run it on hardware yet.
[22:42:58] <kgartner> I'll have it tested on steppers within the hour
[23:12:36] <CIA-2> EMC: 03cradek 07v2.4_branch * r61605f1d860c 10/nc_files/lathe_pawn.ngc: this code assumes radius mode
[23:49:20] <voxadam> How complete is EMC's Modbus implementation?