Back
[00:03:45] <cradek> I think halscope in trunk is busted. it just triggers randomly.
[00:06:22] <skunkworks> hmm - is that what Lerman was seeing?
[00:06:29] <skunkworks> a few days ago
[00:07:12] <skunkworks> * skunkworks doesn't know if he was using trunk
[00:08:07] <skunkworks> http://www.linuxcnc.org/irc/irc.freenode.net:6667/emcdevel/2008-08-25.txt
[00:08:22] <skunkworks> time stamp 17:19:34
[00:17:47] <cradek> I wouldn't be surprised
[00:17:52] <cradek> sounds like the same problem
[00:26:11] <jmkasunich> how'd that happen?
[00:26:29] <cradek> bit rot...
[00:26:43] <cradek> nobody is using trunk, so it happens when we're not looking
[00:26:53] <skunkworks> * skunkworks is glad he doesn't have commit access..
[00:26:57] <skunkworks> ;)
[00:27:05] <jmkasunich> no, somebody broke-ified it
[00:27:23] <jmkasunich> * jmkasunich knows who touched it last
[00:27:39] <cradek> uh-oh
[00:30:07] <jepler> whee, my two-pad touch sensor prototype works
[00:30:14] <jmkasunich> jepler: ?
[00:30:44] <jepler> jmkasunich: homemade capacative touch sensors
[00:31:06] <jmkasunich> finger touch, or tool length?
[00:31:09] <jepler> finger touch
[00:31:10] <skunkworks> what kind of repeatabillity?
[00:31:14] <skunkworks> oh
[00:31:36] <jepler> two .5x1" pads on the bottom side of a circuit board, slide a finger across the top and get an analog slider value out
[00:31:44] <jmkasunich> cradek: triggering works right in 2.2 ?
[00:31:54] <jmkasunich> jepler: slick
[00:32:34] <skunkworks> I think jepler is having fun with his pcb proto-typer
[00:32:39] <jepler> skunkworks: I sure am!
[00:32:43] <skunkworks> :)
[00:34:44] <jmkasunich> cradek: can you describe the random triggering in more detail? does it happen regardless of the type of the selected trigger channel?
[00:35:44] <jmkasunich> the code that makes the actual trigger decision is in scope_rt.c, and is unchanged for over a year
[00:37:19] <cradek> yes it triggers every few seconds, for no reason that I can see
[00:37:40] <jepler> is it the auto/manual button that is broken?
[00:37:51] <jmkasunich> every few seconds - is that the same interval as when in auto?
[00:38:00] <cradek> oh does this mean I'm an idiot?
[00:38:07] <cradek> sorry.
[00:38:09] <jmkasunich> I hope so
[00:38:37] <jmkasunich> ISTR ken saying that he checked auto/manual, and that wasn't it
[00:38:42] <jmkasunich> so you might not be an idiot
[00:38:57] <jmkasunich> auto/normal I mean
[00:40:03] <cradek> well for me the problem was being on auto
[00:41:16] <jmkasunich> heh
[00:41:25] <jmkasunich> glad to see it was an easy bugfix
[00:50:07] <cradek> guh.
[01:04:37] <fenn> i wish someone else would make docs into their own package so i dont have to
[01:04:54] <fenn> users shouldnt have to download 130MB of packages to compile
[01:05:29] <fenn> (most of the extra crap is texlive that is pulled in for lyx)
[01:34:48] <cradek> hm, I learned something: if your encoder seems to have 100.5 lines instead of 100, try tightening the setscrews on the encoder coupling
[01:38:35] <jmkasunich> heh
[01:38:46] <jmkasunich> you've had a lot of trouble with encoder couplings
[01:40:03] <cradek> I thought this was counting wrong - icky noise problem or something (it's not differential) but then after more study I noticed the index moving...
[01:43:09] <jepler> cradek: oh you got the spindle encoder "working"?
[01:43:38] <cradek> yes
[01:43:40] <jepler> cool
[01:43:47] <cradek> it's fine now, I ran it 10k revs and it was still on the right count
[01:44:48] <jepler> oh good
[01:51:04] <jepler> hah, I couldn't have gotten that starburst if I'd tried:
http://emergent.unpy.net/files/01219970985/img_7503-medium.jpg
[01:51:20] <jmkasunich> nice
[01:51:30] <jmkasunich> thats your touch swtich
[01:51:34] <jmkasunich> switch
[01:51:35] <jmkasunich> ?
[01:52:09] <jepler> yeah
[01:52:15] <jepler> http://emergent.unpy.net/index.cgi/01219970985
[01:52:21] <jepler> ^^ more info about it
[02:00:00] <skunkworks> that starburst is so good it looks photoshopped. :) Very neat
[02:03:29] <jepler> too bad about the blobs and smears of solder
[02:15:30] <cradek> the cutout of that board looks really excellent
[02:15:41] <cradek> you have really improved that machine.
[02:17:00] <cradek> I might have resistors > 1M. would that improve it?
[11:34:18] <CIA-40> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/gui/ (axis.lyx mini.lyx tkemc.lyx): shorten chapter names
[11:44:00] <CIA-40> EMC: 03bigjohnt 07v2_2_branch * 10emc2/docs/src/gui/ (axis.lyx mini.lyx tkemc.lyx): shorten chapter names
[12:43:34] <skunkworks_> good morning.
[12:50:20] <jepler> hi skunkworks_
[12:54:27] <jepler> this is just silly:
http://www.flickr.com/photos/oskay/sets/72157600744322220/
[12:56:37] <jepler> cradek: I assume a higher resistor value would increase the range of detected values.
[12:58:46] <jepler> but I'm not sure a greater range is needed
[13:02:34] <cradek> I like the ribbon cable
[13:15:23] <skunkworks_> they should have used real bread... ;)
[13:51:51] <cradek> jepler: surface mount resistors would have worked really nicely on that board
[13:54:12] <cradek> darn, I failed to convince alex there was a problem for him to find before bedtime...
[14:08:27] <jepler> cradek: and actually I put pads for surface-mount resistors on there .. but I don't have any appropriate-value smd resistors anyway
[14:08:52] <jepler> the profile of the connector is so much higher than those 1/8W resistors that it doesn't really matter, though
[14:32:52] <cradek> jepler: in mdi, is it easy to make control-space still work to clear the messages?
[14:34:09] <jepler> cradek: probably it's possible ..
[14:34:29] <cradek> cool, with alex's SPINDLE_MODE=1, you can get manual control + css
[14:35:24] <jepler> cradek: probably it's possible ..
[14:38:07] <jepler> alex_joni: aha, you didn't give me the *first* exception yesterday
[14:42:26] <CIA-40> EMC: 03jepler 07TRUNK * 10emc2/lib/python/rs274/glcanon.py: fix selecting motions (exceptions 'line() takes exactly 3 arguments' and subsequent 'invalid operation')
[14:42:27] <CIA-40> EMC: 03jepler 07TRUNK * 10emc2/src/emc/usr_intf/axis/scripts/axis.py: fix selecting motions (exceptions 'line() takes exactly 3 arguments' and subsequent 'invalid operation')
[14:43:00] <cradek> jepler: for the second thing I said "you can" - I meant it works
[14:44:14] <CIA-40> EMC: 03jepler 07TRUNK * 10emc2/src/emc/usr_intf/axis/scripts/axis.py: make ctrl-space work in mdi to clear all notifications
[14:45:11] <jepler> apparently I was just repeating myself
[14:45:26] <cradek> whee, it works with the jogwheels too
[14:49:05] <jepler> cradek: did you see that I hooked up that stepper jogwheel? It loses steps, just like the stepper-to-quadrature circuits I've read about online.
[14:49:19] <cradek> neat
[14:49:28] <jepler> but it does work
[14:50:09] <jepler> If I can keep it, I need to figure a way to mount it
[14:50:10] <cradek> there are real encoder jogwheels for $60 now - not sure if it would be worthwhile to make some boards.
[14:51:04] <jepler> if I was spending money I know I'd pay a bit more for a real wheel .. if it was half as much for the stepper-jogwheel I might consider it
[14:51:08] <jepler> I don't know how many people are like me
[14:53:56] <rayh> No question. You are unique, jepler.
[14:54:08] <jepler> thanks for the kind (?) words, ray
[14:54:23] <rayh> But in this I'd do the same out to 4-5x.
[14:54:38] <SWPadnos> heh
[14:55:06] <jepler> you mean you'd pay 4x as much for a proper encoder jogwheel, compared to my 2x as much?
[14:55:09] <SWPadnos> I think the possibility of losing position when doing slow moves (when you're likely trying for maximum precision) is a bad thing
[14:55:23] <rayh> You bet it is.
[14:56:14] <jepler> at $30 I think it would be hard to have a product, at $15 it would be impossible.
[14:56:26] <rayh> My handwheel experience is mostly testing the accuracy of an axis' motion. So getting a teenth when you ask for it is critical.
[14:56:59] <SWPadnos> jepler, you mean as a thing to sell to let people use some old stepper as an MPG?
[14:57:20] <rayh> But that is quite different from folk trying to drive an axis toward a part.
[14:57:25] <jepler> SWPadnos: for something ready to put on a panel and wire up
[14:58:41] <SWPadnos> hmmm. I think I didn't understand you then :)
[14:58:57] <jepler> what don't you understand about understanding me?
[14:59:15] <SWPadnos> <jepler>at $30 I think it would be hard to have a product, at $15 it would be impossible.
[14:59:29] <SWPadnos> have as in sell or have as in buy?
[14:59:36] <jepler> as in sell
[14:59:42] <SWPadnos> ok, that's what I thought
[15:00:24] <SWPadnos> one thing I've found is that selling things for too little is a great way to waste a lot of time
[15:00:46] <SWPadnos> there's a minimum you need to make on each item to have it be worthwhile
[15:01:21] <jepler> that's what I'm saying .. finished circuit board + stepper + knob + coupling/mounting assembly + margin is much greater than fifteen dollars
[15:01:24] <jepler> probably greater than $30
[15:01:48] <SWPadnos> ah, so it's not a great plan to compete against the $60 MPGs - I think I may be getting it now :)
[15:01:57] <jepler> right, that's what I'm saying
[15:02:08] <SWPadnos> woohoo!
[15:02:30] <SWPadnos> thank you for taking the time to explain it to me (or for letting the caffeine sink in)
[15:03:30] <jepler> bbl
[15:09:16] <rayh> /away
[15:09:23] <SWPadnos> run awaay!
[15:09:32] <rayh> That's me.
[15:53:34] <cradek> neat, thanks for the ctrl-space
[15:55:21] <skunkworks_> umm... chris - we are waiting for video.. ;)
[15:57:09] <cradek> I hope it's almost done, since I've used 15/16 of my mesa digital outs
[16:11:48] <skunkworks_> heh - you still have the printer port.. ;)
[16:33:48] <CIA-40> EMC: 03cradek 07TRUNK * 10emc2/src/emc/task/ (taskintf.cc emctaskmain.cc emctask.cc):
[16:33:48] <CIA-40> EMC: change the spindle mode option: estop, abort, following error, amplifier
[16:33:48] <CIA-40> EMC: fault, machine off should always stop the spindle. this makes the
[16:33:48] <CIA-40> EMC: SPINDLE_MODE option apply only to mode switches, which I think addresses
[16:33:48] <CIA-40> EMC: the original complaint.
[16:34:44] <CIA-40> EMC: 03compile-farm 07Ubuntu 6.06 LTS (dapper) realtime (2.6.15-magma) * 10emc2head/: build FAILED ; see
http://linuxcnc.org/compile_farm/emc2head_slot7_log.txt
[16:34:47] <CIA-40> EMC: 03compile-farm 07Ubuntu 5.10 (breezy) non-realtime (2.6.12-10-386) * 10emc2head/: build FAILED ; see
http://linuxcnc.org/compile_farm/emc2head_slot1_log.txt
[16:34:48] <CIA-40> EMC: 03compile-farm 07BDI-4.51 (2.6.16.20-rtai) * 10emc2head/: build FAILED ; see
http://linuxcnc.org/compile_farm/emc2head_slot6_log.txt
[16:35:19] <CIA-40> EMC: 03compile-farm 07Ubuntu 5.10 (breezy) realtime (2.6.12-magma) * 10emc2head/: build FAILED ; see
http://linuxcnc.org/compile_farm/emc2head_slot2_log.txt
[16:36:07] <cradek> * cradek hides from alex
[17:42:00] <cradek> uh, how did I do that?
[17:42:14] <jepler> cradek: you're a bad man
[17:42:58] <cradek> ohhhh
[17:43:56] <cradek> ... oops
[17:43:59] <CIA-40> EMC: 03cradek 07TRUNK * 10emc2/src/emc/nml_intf/emc.hh:
[17:43:59] <CIA-40> EMC: change the spindle mode option: estop, abort, following error, amplifier
[17:43:59] <CIA-40> EMC: fault, machine off should always stop the spindle. this makes the
[17:43:59] <CIA-40> EMC: SPINDLE_MODE option apply only to mode switches, which I think addresses
[17:43:59] <CIA-40> EMC: the original complaint.
[17:47:26] <CIA-40> EMC: 03compile-farm 07BDI-4.51 (2.6.16.20-rtai) * 10emc2head/: build PASSED
[17:47:42] <CIA-40> EMC: 03compile-farm 07Ubuntu 6.06 LTS (dapper) realtime (2.6.15-magma) * 10emc2head/: build PASSED
[17:48:11] <CIA-40> EMC: 03compile-farm 07Ubuntu 5.10 (breezy) realtime (2.6.12-magma) * 10emc2head/: build PASSED
[17:48:22] <cradek> duh
[17:49:36] <BigJohnT> when you run the compile farm does that rebuild the live cds?
[17:50:00] <cradek> no
[17:50:11] <cradek> that has to be done tediously by hand
[17:50:24] <cradek> the compile farm checks every cvs commit for stupid (which it found, in my case)
[17:51:01] <BigJohnT> ok, thanks
[17:51:09] <jepler> nothing but human intervention rebuilds the live CDs
[17:51:14] <jepler> rebuilding the live CD every 30 seconds is a dumb idea anyway
[17:53:11] <CIA-40> EMC: 03compile-farm 07Ubuntu 5.10 (breezy) non-realtime (2.6.12-10-386) * 10emc2head/: build PASSED
[17:53:32] <BigJohnT> what rebuilds the heml docs on the EMC web site?
[17:53:43] <BigJohnT> heml > html
[17:54:20] <cradek> jeff's computer (I think)
[17:54:33] <BigJohnT> ok
[17:54:34] <jepler> yes, that's right
[17:54:37] <cradek> the compile farm is on john k's computer
[17:55:49] <BigJohnT> is the EMC web page on your computer jeff?
[17:56:03] <cradek> no, it's on dreamhost
[17:56:11] <BigJohnT> ok
[17:56:53] <BigJohnT> who edits the web page mostly?
[17:57:20] <cradek> alex mostly
[17:57:25] <BigJohnT> have I run out of questions yet? :)
[17:57:38] <BigJohnT> ok
[18:00:16] <BigJohnT> well I have to get back to making plastic snow... talk to you guys later
