#emc | Logs for 2007-05-13

Back
[00:16:39] <CIA-8> 03petev 07TRUNK * 10emc2/src/hal/components/biquad.comp: -Initial revision.
[00:17:02] <a-l-p-h-a> anonimasu sup?
[00:17:04] <a-l-p-h-a> still kicking?
[00:17:50] <anonimasu> yeah
[00:17:50] <anonimasu> working
[00:17:58] <a-l-p-h-a> on a sunday?
[00:18:01] <a-l-p-h-a> sunday morning?
[00:18:14] <anonimasu> yeah
[00:18:17] <anonimasu> everytday
[00:18:48] <a-l-p-h-a> :(
[00:18:49] <a-l-p-h-a> sucky
[00:19:04] <a-l-p-h-a> I got 9-5... turns out to be 8:30-5... but I get the bank hours
[00:19:25] <anonimasu> hm, I've got 9-9
[00:19:25] <anonimasu> :/
[00:19:27] <a-l-p-h-a> bbiab...
[00:19:39] <anonimasu> with 9-4 hours of pay ;)
[00:41:14] <petev> e
[00:41:27] <Skullworks_PGAB> f
[00:41:41] <petev> sorry, my doze box is having issues
[00:43:33] <Skullworks_PGAB> Windows require frequent washing for clear functionality.
[00:43:42] <skinnypuppy1334> anyone here installed on 7.04 feisty yet?
[00:44:20] <Skullworks_PGAB> there was a thread on CNCZone by a few who have done it.
[00:45:00] <skinnypuppy1334> Thanks for the tip, I haven't been over there in a few months .
[00:45:17] <Skullworks_PGAB> * Skullworks_PGAB is dragging it out, still running Breazy.
[00:46:15] <skinnypuppy1334> I liked dapper just fine, but I can finally use my .flv files from youtube on feisty with sound finally .
[00:47:04] <Skullworks_PGAB> I'm just trying to keep the hardware req down - I'm running on a pIII-550
[00:47:11] <skinnypuppy1334> This is only a sim mode machine I wouldn't go against stable 6.06 otherwise
[00:49:28] <Skullworks_PGAB> I'm surprised how many people use the sim.
[00:56:31] <Skullworks_PGAB> * Skullworks_PGAB wonders what the next stable ubuntu kernal will be named... Howling Hurracane?
[00:56:55] <skinnypuppy1334> I haven't spent the cheese to get controlers and motors for my vertical mill.... I like the sim b/c it is giving me the confidence I'll know how to run it.
[00:57:21] <Skullworks_PGAB> what V mill?
[00:57:58] <skinnypuppy1334> I've got an Enco 10x54.. like a 2j style bridgeport.
[00:58:38] <Skullworks_PGAB> doing the Knee or quill for Z?
[00:59:35] <skinnypuppy1334> I'd prefer knee, I've been too busy with school until recently... just got my cnc diploma
[00:59:52] <Skullworks_PGAB> good for you...
[01:00:59] <Skullworks_PGAB> I still have my paper tapes from when i took CNC in college in 84
[01:01:15] <skinnypuppy1334> I've got two lathes, mill and a surface grinder im my garrage ..... paper tapes yeha!
[01:02:12] <Skullworks_PGAB> want to talk scary... Try a $4000 CAD/CAM program than ran on an Apple ][ with 48K mem.
[01:03:10] <skinnypuppy1334> yikes! I remember when motherboards looked like a 3xl pizza box
[01:03:46] <Skullworks_PGAB> I need a garage... - People are going to think I'm "Strange" when they see my IH mill in my kitchen.
[01:03:57] <skinnypuppy1334> ;o)
[01:05:23] <Skullworks_PGAB> Well - I was buying a nice ranch with a great place to put my machines... But the owners job tranfer didn't pan out and the deal fell thru.
[01:06:34] <Skullworks_PGAB> Atleast I had not ordered a lathe yet...
[01:09:51] <skinnypuppy1334> I guess I haven't plunged the money yet b/c I keep hoping to see a ser1 or boss BP reasonable near by.
[01:10:23] <Skullworks_PGAB> I know the feeling
[01:11:38] <Skullworks_PGAB> There is a dead Hurco KM3 (knee mill) that I want to buy, but the owner is content to use it as a place to store spare vises.
[01:11:58] <skinnypuppy1334> If only ballscrews were less expensive. It would still be a nice addition with acme leads running in single block and using the dro for final positioning for drilling reaming boring stuff.
[01:13:13] <skinnypuppy1334> I think I'd be afraid to climb mill anything with .007 backlash, or expect a good circle from a g2 g3 ...
[01:13:25] <Skullworks_PGAB> I spent the money and got a DRO with glass scales on the IH.
[01:14:55] <skinnypuppy1334> I got a deal on my mill, with dro, a 6 in vice, 12" rotary table, 12x36 lathe tooled, surface plate, height gage and some other tooling under 4k.
[01:14:59] <Skullworks_PGAB> make some backlash adjusting nuts - or self tensioning nuts (uses a die spring to prelaod)
[01:15:18] <Skullworks_PGAB> nice theft!
[01:15:42] <skinnypuppy1334> I had thought of making some backlash nuts using a fine pitch to adjust.
[01:16:54] <Skullworks_PGAB> spring preload works best - it allows differences in wear areas to work without binding.
[01:18:52] <skinnypuppy1334> Yeah, last one was a colchester 17" about 10ftbed , collet chuck, collets, 3 and 4 jaw burnerd chucks taper attach, $1000 on fleebay. Now it is a little old and could probably use a new cross feed screw and the bed is a touch worn near the chuck... but atleast I can do some larger pieces now...
[01:19:25] <Skullworks_PGAB> thats big for a garage
[01:20:02] <skinnypuppy1334> takes up the full wall to the garrage door from the kitchen door in the garrage....
[01:22:21] <skinnypuppy1334> I keep watching for surpluss steppers in my area but haven't found anything... I've got enough pc's to wire one to everything in the shop and the neighbors dog
[01:23:09] <Skullworks_PGAB> your going to use steppers for the mill?
[01:24:20] <skinnypuppy1334> Probably, I don't expect to do anything particularly heavy nor fast, but to speed along jigged replica parts.
[01:25:06] <skinnypuppy1334> I wouldn't expect to do die or mold making with a machine like this
[01:25:41] <Skullworks_PGAB> speed and steppers just don't fit on the same line....
[01:26:33] <skinnypuppy1334> Just curious how many ipm are people getting with .125 lead screws?
[01:26:35] <Skullworks_PGAB> your looking at some largish 34's for X/Y and a 42 for the knee...
[01:27:11] <Skullworks_PGAB> a air spring on the knee will help
[01:27:13] <skinnypuppy1334> Yeah I'd been directed towards some 1200 oz 34's
[01:27:36] <skinnypuppy1334> I'd read those help in supporting the knee
[01:27:40] <Skullworks_PGAB> I think that might be overkill
[01:28:33] <Skullworks_PGAB> but I'd say around 850/1000 would do the trick
[01:28:43] <Skullworks_PGAB> for X/Y
[01:28:49] <skinnypuppy1334> You can imagine I've had lots of opinions... some of those with opinions also complain of jerky action and other problems ... makes me wonder.
[01:29:24] <Skullworks_PGAB> the more the holding torque the lower your top speed will be
[01:29:28] <skinnypuppy1334> I was planning on cutting some belt cogs with 2:1 reduction on the x/y and 3 or 4:1 on the z
[01:30:04] <ds2> hmmmm
[01:30:16] <skinnypuppy1334> 850/1000 seems sufficient for 1:1 ?
[01:30:44] <Skullworks_PGAB> so ideally use only as big a motor as required
[01:30:59] <Skullworks_PGAB> yeah if your screw is .125"
[01:31:09] <Skullworks_PGAB> 1200 for a .2"
[01:31:26] <skinnypuppy1334> makes sense most of the ones complaining of jerky had 42's on like plazma gantrys
[01:32:31] <skinnypuppy1334> ok, makes sense now... the larger recomenders also had ballscrews... lead accounts for more torque then...
[01:32:37] <Skullworks_PGAB> well - big motors using a rack and pinion drive can work out if they can keep out of resonance zones
[01:33:08] <Skullworks_PGAB> they will not go over 200 rpm on the rack
[01:33:31] <Skullworks_PGAB> key is a good microstep drive for those applications
[01:34:48] <skinnypuppy1334> yeah I wanted to get geckos until I had enough time to build some boards and try out. I used to do audio amp repair and would like building some boards but don't want to risk the frustration the first time around
[01:35:47] <Skullworks_PGAB> if you use Geckos - I doudt you will ever "upgrade"
[01:36:17] <Skullworks_PGAB> what rocks with Geckos is the voltage range they support
[01:37:19] <skinnypuppy1334> I've only read good stuff about them. I wanted to build a 60v with transistor current limited leads to each gecko for protection
[01:37:26] <Skullworks_PGAB> using a 68VDC your torque curve stays up much higher
[01:37:41] <Skullworks_PGAB> no need
[01:38:14] <Skullworks_PGAB> Geckos have an adjustable current limiter built in
[01:41:05] <skinnypuppy1334> That's right. It was over the x-mas holiday I read about the geckos. I had tried out emc again and was really impressed with what it could do especially since I understood g code now.
[01:42:51] <Skullworks_PGAB> I just wish enough people would get interested in doing a version that would work for doing Hot wire foam cutting.
[01:43:31] <Skullworks_PGAB> but this would require a new gui and maybe a kinematics profile
[01:44:17] <Skullworks_PGAB> its kinda like doing Wire EDM Sideways....
[01:44:19] <skinnypuppy1334> that would be cool, kinematics b/c of the effects of tilting the wire to taper ?
[01:44:43] <Skullworks_PGAB> well a 2axis is a 4 axis
[01:45:00] <Skullworks_PGAB> LX LY RX RY
[01:45:19] <skinnypuppy1334> I haven't been around wire edm or wire cutters but run across a few before on diy sites
[01:46:11] <Skullworks_PGAB> right now there is no gui to support the LX LY RX RY.
[01:46:53] <skinnypuppy1334> gotcha my imagination has been unsteady since seeing one of those lathe/mill machines with a tilting lathe spindle
[01:46:58] <Skullworks_PGAB> also no real cad/cam to output code
[01:47:57] <Skullworks_PGAB> most Foam cutting warez includes the motor driver software so the code stays internal to the program.
[01:48:38] <skinnypuppy1334> I wish i had more exposure to source code programming
[01:48:40] <Skullworks_PGAB> common use is to cut model plane wings
[01:49:02] <skinnypuppy1334> model planes?
[01:49:32] <Skullworks_PGAB> I want to use foam cutting to make core components for lost foam Aluminum castings
[01:49:39] <skinnypuppy1334> never mind r/c planes
[01:50:37] <Skullworks_PGAB> like the castings to hold my servo motors...
[01:50:52] <skinnypuppy1334> I was thinking of maching something soft to use for some plaster molds I could blow great stuff into . My neighbor is a blacksmith with a foundry ;0) I know what you mean
[01:51:26] <skinnypuppy1334> we got a big industrial kitchen mixxer to use as a sand muller.
[01:52:13] <Skullworks_PGAB> yeah - cause the foam cutting cad/cam isn't available people buy the foam sheets and use CNC routers to cut out the patterns
[01:54:00] <Skullworks_PGAB> its a workaround - with hotwire I could get a much closer "near net shape"
[01:54:04] <skinnypuppy1334> I could see where nichrome wire would do better
[01:55:39] <Skullworks_PGAB> you slice your model (3d) into sections - cut each section then glue (or toothpick) the sections together before coating in plaster.
[01:57:02] <skinnypuppy1334> I wound some new ones for my kiln on a drill rod using the lathe feed. I took open length of the wire and made a spiral hangin on a stick . 120v and an amp meter in series I used an aligator clip to move up the wire length to get to the current I needed to get 600 watts out of each coil.. marked the two spots in length and wound.
[01:57:26] <Skullworks_PGAB> * Skullworks_PGAB really needs to go buy a ranch with room for shop and foundry...
[01:57:42] <tomp> rapid prototyping... candyfab: a hot air gun melts sugar spread in layers on a bed http://www.engadget.com/2007/05/10/diy-3d-printer-utilizes-hot-air-sugar-to-craft-random-objects/ uses "selective hot air sintering and melting (SHASAM)" :)
[01:58:06] <skinnypuppy1334> have you considered using a fuel oil burner for aluminum?
[01:58:26] <Skullworks_PGAB> yes
[01:58:58] <Skullworks_PGAB> and mixing in about 15% waste engine oil
[01:59:08] <skinnypuppy1334> that is what I want to do, the natural gas furnace we have is just too costly to run whenever....
[01:59:16] <Skullworks_PGAB> or less depending on results
[02:00:00] <skinnypuppy1334> I have done a lead pour in a crucible on a coal forge... would also work for al but not too neighborhood friendly
[02:00:08] <Skullworks_PGAB> the hobby casting forum reports those oil burners work great.
[02:00:43] <Skullworks_PGAB> I have plumbers furnaces for doing lead...
[02:01:09] <skinnypuppy1334> gr8... I have a box of 100 nozzles... errr 98 . I got some for the steam pressure washer
[02:01:17] <skinnypuppy1334> is that a blow lamp?
[02:01:42] <Skullworks_PGAB> ?
[02:02:22] <skinnypuppy1334> older plumbers torches, gun hobbiest I read use them for heat treating small parts
[02:02:41] <skinnypuppy1334> used a liquid fuel and you pumped much like a lantern
[02:03:21] <Skullworks_PGAB> no - this is like a single burner pump up coleman stove
[02:03:33] <Skullworks_PGAB> but bigger
[02:03:49] <skinnypuppy1334> neat, a plumbers furnace ha?
[02:03:53] <Skullworks_PGAB> and I'm lazy so I fit a compressed air fitting
[02:03:59] <Skullworks_PGAB> well
[02:04:50] <skinnypuppy1334> learn something every day .
[02:05:02] <Skullworks_PGAB> old cast iron sewer lines were joined by soaking a hemp rope in liquid lead then forcing it in the gap
[02:05:34] <Skullworks_PGAB> the leaded rope shrunk as it cooled forming a tight seal
[02:06:30] <Skullworks_PGAB> I bought the large cast iron "Dutch ovens" (pots) from Harbor freight.
[02:07:39] <Skullworks_PGAB> can melt about 80lbs of wheel weights in the large pot in one batch.
[02:07:51] <skinnypuppy1334> woah!
[02:08:03] <Skullworks_PGAB> takes about 20 minutes
[02:08:27] <Skullworks_PGAB> then I use a ladle to skim off the steel clips
[02:09:00] <skinnypuppy1334> reminds me the first time I saw steel bar float
[02:10:41] <Skullworks_PGAB> http://www.skullworks.net/include/DSCF0014.JPG
[02:11:08] <skinnypuppy1334> toxic waste ha!
[02:11:32] <Skullworks_PGAB> I cast up a pile of these ingots to feed my big bullet casting machine.
[02:12:05] <Skullworks_PGAB> yeah - if you have CNC toys... you can make jokes...
[02:12:12] <skinnypuppy1334> now that is cool!!!
[02:13:25] <skinnypuppy1334> whats that billet weigh?
[02:13:52] <Skullworks_PGAB> about 3 1/2 lbs
[02:14:23] <Skullworks_PGAB> need to take a better pic
[02:15:21] <Skullworks_PGAB> but I have about 1800lbs of those - and about 5 5gal buckets of wheel weights I still need to process.
[02:16:27] <skinnypuppy1334> I'm guessing that is the steel mold to the left?
[02:16:41] <Skullworks_PGAB> 7057 AL
[02:16:47] <Skullworks_PGAB> 7075
[02:17:17] <Skullworks_PGAB> was a scrap chunk that looked the perfect size...
[02:17:49] <skinnypuppy1334> I wish I were closer to a good scrap yard. I don't find much billet
[02:17:53] <Skullworks_PGAB> has an oak handle - I turned it from part of a broken pallet...
[02:18:54] <Skullworks_PGAB> I go to Denver to scope out the good scrap yards
[02:19:10] <Skullworks_PGAB> not much local either
[02:20:56] <skinnypuppy1334> I'm about 100 miles from the nearest decent one, I watch fleebay for the bargains sometimes.
[02:24:01] <skinnypuppy1334> I'm going to read the rest of the gcam manual....
[02:24:12] <skinnypuppy1334> are you using it ?
[02:24:37] <Skullworks_PGAB> no - I bought Bobcad V21
[02:25:23] <Skullworks_PGAB> I been using BCC since v10 in the DOS days circa 1989
[02:25:51] <skinnypuppy1334> I haven't had a chance, I used catia and mastercam x in class
[02:26:13] <Skullworks_PGAB> I'm sorry
[02:26:37] <Skullworks_PGAB> they spoil you then dump you out on your own... its rude.
[02:26:43] <skinnypuppy1334> ha I won't ever afford either of those.
[02:27:12] <Skullworks_PGAB> last time I priced Catia it was $19,900 base per seat.
[02:27:34] <Skullworks_PGAB> that was back about 1996
[02:28:07] <skinnypuppy1334> I've got a version of X from a book that will let you do everything except save and post code... I use it for quick prints and screen capture save image....
[02:30:17] <skinnypuppy1334> what did bobcad set you back?
[02:30:23] <Skullworks_PGAB> I have a complete version 9 with training CD set... and no time or need to learn it.
[02:30:54] <Skullworks_PGAB> not much BCC used to haggle
[02:31:17] <Skullworks_PGAB> but not as much now with the new v2007
[02:31:21] <skinnypuppy1334> i worked with 9 enough to realize the interface isn't intuitive and it would require a devotee to run it
[02:31:49] <Skullworks_PGAB> its now like MC - with a event tree on the side
[02:32:12] <Skullworks_PGAB> also uses Preditor editor and backploter
[02:32:42] <Skullworks_PGAB> so they are bundling alot of 3rd party wares into there package now
[02:33:11] <Skullworks_PGAB> this raises there cost so they have less marging to bargain with.
[02:33:43] <Skullworks_PGAB> think V21 cost $495
[02:33:59] <Skullworks_PGAB> as an upgrade
[02:36:51] <skinnypuppy1334> anyone out there offering jog handles for EMC?
[02:37:01] <Skullworks_PGAB> V21 was the version released in March 2006
[02:40:36] <Skullworks_PGAB> I need to go drop off lunch to a friend stuck at work...
[02:40:57] <skinnypuppy1334> good to chat with you.
[02:41:00] <Skullworks_PGAB> back tomorrow.
[02:41:06] <Skullworks_PGAB> yes you to.
[02:41:10] <Skullworks_PGAB> later
[02:41:16] <skinnypuppy1334> l8r
[02:56:52] <christel> [Global Notice] This is your captain speaking; in a little while our main US hub will be going down for some maintenance. As you may feel some turbulence as we jump I would ask you to keep your seat in a upright position, stow the tray table and keep your seatbelt fastened while the fasten seatbelt sign is illuminated. Thank you for flying freenode and have a great day!
[03:24:43] <tomp> i ran bin/halscope & got familiar msg "cant open thread blah" ( true there is no thread blah now ) so i rename the .scope.cfg file to .xxxscope.cfg... and try again, and get 'HAL: ERROR: duplicate component name 'halscope''... and 'halcmd show' agrees... how do i get kill it?
[03:28:43] <tomp> i ran bin/halscope & got familiar msg "cant open thread blah" ( true there is no thread blah now ) so i rename the .scope.cfg file to .xxxscope.cfg... and try again, and get 'HAL: ERROR: duplicate component name 'halscope''... and 'halcmd show' agrees... how do i get kill it?
[03:29:16] <jmkasunich> tomp: unloadrt scope_rt (or whatever the name of that component is)
[03:30:59] <jmkasunich> tomp: you should never have to mess with .scope.cfg
[03:31:08] <jmkasunich> that message is a warning, just go on with your life
[03:31:22] <jmkasunich> maybe I should remove the messages.....
[03:31:38] <tomp> Exported Functions:
[03:31:38] <tomp> 09 e0ce6160 00000000 NO 0 scope.sample
[03:31:38] <tomp> tomp@cncbox:~/emc2-head$ bin/halcmd unloadrt scope
[03:31:39] <tomp> HAL:0: ERROR: component 'scope' is not loaded
[03:31:49] <jmkasunich> show comp
[03:31:50] <tomp> and no scope appears for me to go on with
[03:32:20] <jmkasunich> you mean when the message first appeared you got no scope? or after you messed with the file?
[03:32:30] <tomp> show comp ---> 00008 User halscope 10222 initializing
[03:32:39] <tomp> yes, no scope appeared ever
[03:32:46] <jmkasunich> wtf
[03:32:51] <tomp> me2
[03:34:10] <tomp> scripts/realtime start ; bin/halcmd -f l298-jogger6.hal ; bin/halscope then what i listed
[03:34:41] <tomp> the hal fiel loads an xml for a pyvcp panel if that is of use
[03:35:29] <tomp> gui appears
[03:35:36] <tomp> the pyvcp gui
[03:36:44] <tomp> http://pastebin.ca/485633
[03:37:29] <jmkasunich> the .config file is used to try to restore the setup you had the last time you ran the scope - same thread, same sample rate, same channels, gains, offsets, etc
[03:37:48] <tomp> right, but i have different threads now
[03:37:57] <jmkasunich> when it finds something it can't duplicate, like you were probing a signal that doesn't exist now, it should print a warning and carry in
[03:38:03] <jmkasunich> if it hangs, thats a bug
[03:38:13] <jmkasunich> I'll try to reproduce it here
[03:38:17] <tomp> ok, trashing the file wont help
[03:38:37] <tomp> http://pastebin.ca/485634 is the xml
[03:42:07] <tomp> wait! bin/halcmd unloadrt scope_rt 'worked' , it got rid of the comp, but now get tomp@cncbox:~/emc2-head$ bin/halscope
[03:42:07] <tomp> halscope: config file '.scope.cfg' could not be opened
[03:42:07] <tomp> Floating point exception
[03:44:06] <tomp> i had done unloadrt scope_rt , unloadrt all & scripts/realtime stop then did scripts/realtime start ; bin/halcmd -f l298-jogger6.hal ; bin/halscope to get the new msg
[03:51:15] <jmkasunich> tomp: you've pastebin'd hal files and such, but not the output from halscope
[03:52:04] <jmkasunich> I started halcmd, did "loadrt threads name1=foo period1=1000000" to create a thread with a dumb name
[03:52:08] <jmkasunich> then I started halscope
[03:52:28] <tomp> out from the thing that wont run ?
[03:52:32] <tomp> output
[03:52:35] <jmkasunich> it read my .scope.cfg file, which wanted to use servo-thread (left over from the last time I ran scope)
[03:52:41] <jmkasunich> it printed this:
[03:52:41] <jmkasunich> jmkasunich@ke-main-1006:~/emcdev/emc2head$ bin/halscope
[03:52:42] <jmkasunich> halscope: could not find thread: 'THREAD servo-thread'
[03:52:42] <jmkasunich> halscope: could not set horizontal multiplier: 'HMULT 1'
[03:52:50] <jmkasunich> as well as about 10 more lines
[03:52:55] <jmkasunich> then it started up just fine
[03:53:03] <jmkasunich> I told it to use thread foo
[03:53:17] <tomp> ok, mine blows up and doesnt appear, my thread and time is in the .hal file posted
[03:53:23] <jmkasunich> and to observe parameter scope_sample.time
[03:53:41] <jmkasunich> I need the output from the run that blew up
[03:53:42] <tomp> show comp says i have 2 halcmds running :(
[03:53:53] <tomp> Loaded HAL Components:
[03:53:53] <tomp> ID Type Name PID State
[03:53:53] <tomp> 00002 User halcmd11422 11422 ready
[03:53:53] <tomp> 00008 User halscope 11103 initializing
[03:53:53] <tomp> 00007 User halcmd11088 11088 ready
[03:53:53] <tomp> 00006 User l298jogr6 11095 ready
[03:54:08] <tomp> and halscope is still trying to init
[03:54:30] <jmkasunich> so what do you actually have running?
[03:54:40] <jmkasunich> two halcmds is easy, just run it in two shells
[03:54:41] <tomp> nothing visible
[03:55:04] <jmkasunich> are you doing everything from the command line, or are you using icons or buttons in a GUI or something
[03:55:28] <tomp> i've been trying to stop it and have succeded in zombi-ing the pyvcp gui ... all from cmd line but yes from 2 terminals
[03:55:51] <tomp> no emc at all
[03:56:08] <jmkasunich> no emc is good
[03:56:35] <jmkasunich> you gotta get everything shut down - I'm sure its totally borked now, and the more you mess with it the worse its gonna get
[03:56:48] <jmkasunich> the only thing that would be remotely usefull is the very first error
[03:57:00] <tomp> right, i think i will reboot as all ive done hasnt stopped it
[03:57:05] <jmkasunich> don't do that
[03:57:16] <jmkasunich> rebooting is the absolute last resort
[03:57:16] <tomp> lookin for 1st error
[03:57:39] <jmkasunich> you have two open shells, right?
[03:57:49] <jmkasunich> are they running anything, or are they showing a bash prompt?
[03:58:15] <tomp> right, looking for life via prompts
[03:58:34] <tomp> alll $ ( had 3 )
[03:58:39] <jmkasunich> ok
[03:58:40] <tomp> 3 terminals
[03:58:48] <jmkasunich> if you do ps, do you see any hal stuff?
[03:59:19] <jmkasunich> ps -A actually, so you see stuff launched from all three shells
[03:59:22] <tomp> all 3 show bash & ps only
[03:59:40] <jmkasunich> are you running installed or in-place?
[03:59:44] <tomp> ps -A is LOADS
[03:59:50] <tomp> in-place
[03:59:54] <jmkasunich> loads of hal stuff?
[04:00:37] <tomp> no, only this looks hal/rt ish 11069 ? 00:00:00 RTAI_KTHRD_M:0
[04:00:51] <jmkasunich> that just means that you have the rt stuff loaded
[04:00:55] <jmkasunich> ok, next step
[04:01:09] <tomp> k
[04:01:13] <jmkasunich> you are in your top level CVS directory, right?
[04:01:27] <jmkasunich> from there, do "bin/halcmd show comp"
[04:01:28] <tomp> yes ~/emc2-head
[04:01:54] <tomp> tomp@cncbox:~/emc2-head$ bin/halcmd show comp
[04:01:54] <tomp> Loaded HAL Components:
[04:01:54] <tomp> ID Type Name PID State
[04:01:54] <tomp> 00002 User halcmd11671 11671 ready
[04:01:54] <tomp> 00008 User halscope 11103 initializing
[04:01:54] <tomp> 00007 User halcmd11088 11088 ready
[04:01:55] <tomp> 00006 User l298jogr6 11095 ready
[04:02:17] <jmkasunich> wtf is l298jogr6?
[04:02:22] <jmkasunich> is that a pyvcp thing?
[04:02:25] <tomp> yes
[04:02:38] <tomp> should i kill 2 of the 3 ternminals?
[04:02:39] <jmkasunich> and does that show up in ps -A?
[04:02:41] <jmkasunich> no
[04:03:06] <tomp> it does not show up in ps -A
[04:03:28] <jmkasunich> does pyvcp or anything remotely resembling that show up in ps -A?
[04:03:39] <jmkasunich> is there a process 11095 in ps -A?
[04:03:59] <tomp> ps -A output http://pastebin.ca/485653
[04:04:30] <tomp> no 11095 process in ps -A
[04:04:43] <jmkasunich> I thought you said that pyvcp was zombied
[04:04:51] <jmkasunich> if so, it should show up in ps -A
[04:05:00] <tomp> it is, the gui is listed and isnt visible
[04:05:08] <jmkasunich> what>?
[04:05:21] <jmkasunich> you just said (and I confirm) that its NOT listed in ps -A
[04:05:22] <tomp> i use zombie with poetic license
[04:05:34] <tomp> disregrard the zombie crack
[04:05:53] <tomp> and i posted the ps -A to avoid confusion
[04:06:17] <jmkasunich> (zombie has a very specific meaning, its a process that is borked up, and it shows up in ps as (zombie)
[04:06:27] <jmkasunich> so using that term confused me
[04:06:44] <jmkasunich> what you mean is that it shows up in "show comp" but is actually dead
[04:07:03] <tomp> yes i meant halcmd show said it existed,,, and it was not visible ( had zombie flavor )
[04:07:19] <jmkasunich> when you did "scripts/realtime stop" did it print any errors?
[04:07:32] <tomp> yes, lots...
[04:07:49] <tomp> tomp@cncbox:~/emc2-head$ scripts/realtime stop
[04:07:49] <tomp> ERROR: Module hal_lib is in use by scope_rt,threads,encoder,pwmgen,hal_parport
[04:07:49] <tomp> ERROR: Module rtapi is in use by scope_rt,threads,encoder,pwmgen,hal_parport,hal_lib
[04:07:49] <tomp> ERROR: Module rtai_sem is in use by rtapi
[04:07:49] <tomp> ERROR: Module rtai_shm is in use by rtapi
[04:07:50] <tomp> ERROR: Module rtai_fifos is in use by rtapi
[04:07:51] <tomp> ERROR: Module rtai_sched is in use by rtapi,rtai_sem,rtai_shm,rtai_fifos
[04:07:54] <tomp> ERROR: Module rtai_hal is in use by rtapi,rtai_sem,rtai_shm,rtai_fifos,rtai_sched
[04:07:58] <jmkasunich> ok, then doing realtime start wasn't a good idea
[04:08:04] <jmkasunich> because it never got shut down
[04:08:19] <jmkasunich> our first problem is to shut down and clean up your borked realtime stuff
[04:08:26] <tomp> k
[04:08:43] <jmkasunich> wow, thats really borked
[04:08:48] <tomp> ;)
[04:09:01] <jmkasunich> you have scope_rt, threads, encoder, pemgen, parport, etc all loaded
[04:09:09] <jmkasunich> but show comp doesn't list them
[04:09:28] <jmkasunich> ok, lets see what we can do
[04:09:29] <tomp> not now! that was the msg when i first tried.... lemme get current news for you
[04:09:54] <tomp> tomp@cncbox:~/emc2-head$ scripts/realtime stop
[04:09:54] <tomp> tomp@cncbox:~/emc2-head$
[04:10:03] <tomp> see, ok now
[04:10:08] <jmkasunich> thats good
[04:10:15] <jmkasunich> lets make sure nothing is borked
[04:10:17] <jmkasunich> lsmod | head
[04:10:20] <jmkasunich> and paste it here
[04:10:49] <tomp> Module Size Used by
[04:10:49] <tomp> rfcomm 32280 0
[04:10:49] <tomp> l2cap 18948 5 rfcomm
[04:10:49] <tomp> bluetooth 39012 4 rfcomm,l2cap
[04:10:49] <tomp> ppdev 6916 0
[04:10:50] <tomp> nls_iso8859_1 3968 4
[04:10:51] <tomp> nls_cp437 5632 4
[04:10:53] <tomp> vfat 10240 4
[04:10:56] <tomp> fat 44316 1 vfat
[04:10:57] <jmkasunich> ok, goot
[04:10:58] <tomp> dm_mod 48696 1
[04:11:04] <jmkasunich> nothing hal, rtapi, or emc related
[04:11:09] <jmkasunich> you have a clean shutdown
[04:11:23] <jmkasunich> now we can finally start to test things with a known starting point
[04:11:54] <jmkasunich> now, tell me exactly what you did when it got messed up
[04:12:31] <tomp> (tomp is laughing.....) sorry, i might be able to get it from the terminals .....
[04:12:44] <tomp> lemmee see
[04:13:35] <tomp> tomp@cncbox:~$ cd emc2-head/
[04:13:35] <tomp> tomp@cncbox:~/emc2-head$ scripts/realtime start
[04:13:38] <tomp> then...
[04:14:14] <tomp> tomp@cncbox:~/emc2-head$ bin/halcmd -f l298-jogger6.hal
[04:14:14] <tomp> Waiting for component 'l298jogr6' to become ready...........................................pyVCP: Creating widgets from l298jogrgui6.xml ... Done.
[04:14:14] <tomp> then...
[04:14:44] <tomp> tomp@cncbox:~/emc2-head$ bin/halscope
[04:14:44] <tomp> halscope: could not find thread: 'THREAD ???'
[04:14:44] <tomp> halscope: could not set record length: 'MAXCHAN 0'
[04:14:44] <tomp> .......snipped.....
[04:14:57] <tomp> halscope: config file '.scope.cfg' caused 28 warnings
[04:14:57] <tomp> Floating point exception
[04:14:57] <tomp> then
[04:15:07] <jmkasunich> it really said "THREAD ???"
[04:16:11] <jmkasunich> tomp: did it really say "THREAD ???" with the question marks?
[04:16:21] <tomp> it said thread space square box questionmarkinellipse questionmarkinellipse
[04:16:28] <tomp> unprintable apparently
[04:16:38] <jmkasunich> ok, thats important
[04:16:49] <tomp> k...
[04:16:49] <jmkasunich> it somehow read junk instead of the thread name
[04:17:05] <jmkasunich> or somehow junk was saved instead of the thread name the last time
[04:17:20] <jmkasunich> I wish you hadn't messed with the .scope.cfg file
[04:17:59] <jmkasunich> that unprintable junk is the first error, and the natural question is "was that junk in the file, or was the file read incorrectly"
[04:18:06] <tomp> the text file says (i just renamed them to hide them ) THREAD .....x...%
[04:18:20] <jmkasunich> oh, you saved the file...
[04:18:23] <jmkasunich> good
[04:18:23] <tomp> viewed as text
[04:18:27] <jmkasunich> and it has junk in it
[04:18:49] <tomp> yeah, on THREAD, rest looks reasonable, sci notastion for some, but ok
[04:19:06] <jmkasunich> that file is supposed to be human readable
[04:19:27] <jmkasunich> if there is binary junk in it, either the code that writes the file is busted, or the file got corrupted somehow
[04:20:11] <jmkasunich> can you pastebin the .scope.cfg file?
[04:21:18] <jmkasunich> the next important error message is the floating point exception
[04:21:29] <jmkasunich> thats what killed the scope before it had a chance to appear on-screen
[04:21:56] <jmkasunich> unfortunately it will be hard to figure that one out I think
[04:22:06] <jmkasunich> unless it happens every time
[04:22:13] <tomp> yes http://pastebin.ca/485665 but the chars wont copy,,, show as ??? AND thety are different again..... now show as questioninellipse sqr sqr questioninellipse questioninellipse bad spot on hard drive???
[04:22:43] <jmkasunich> this is the same file, and it looks different every time you read it?
[04:23:49] <tomp> garbage in garbage out.... it looks differnt twice... in a row so thats not good statitistics, but i will try again , i used kedit that time to view so i could copy & pastebin for you
[04:23:58] <jmkasunich> just cat it
[04:24:39] <jmkasunich> better yet, hd it:
[04:24:47] <jmkasunich> jmkasunich@ke-main-1006:~/emcdev/emc2head$ hd .scope.cfg
[04:24:47] <jmkasunich> 00000000 54 48 52 45 41 44 20 66 6f 6f 0a 4d 41 58 43 48 |THREAD foo.MAXCH|
[04:24:47] <jmkasunich> 00000010 41 4e 20 34 0a 48 4d 55 4c 54 20 31 0a 48 5a 4f |AN 4.HMULT 1.HZO|
[04:24:47] <jmkasunich> 00000020 4f 4d 20 31 0a 48 50 4f 53 20 35 2e 31 38 32 32 |OM 1.HPOS 5.1822|
[04:24:57] <skinnypuppy1334> what's hd ?
[04:25:03] <jmkasunich> hexdump
[04:25:14] <jmkasunich> shades of msdos's debug view of memory
[04:25:17] <skinnypuppy1334> aah
[04:25:44] <tomp> tomp@cncbox:~/emc2-head$ hd .xxxscope.cfg
[04:25:44] <tomp> 00000000 54 48 52 45 41 44 20 d9 14 f5 b7 a0 b7 f7 b7 78 |THREAD ........x|
[04:25:44] <tomp> 00000010 b7 f7 b7 25 0a 4d 41 58 43 48 41 4e 20 30 0a 48 |...%.MAXCHAN 0.H|
[04:26:01] <tomp> tomp@cncbox:~/emc2-head$ hd .xxxscope.cfg
[04:26:01] <tomp> 00000000 54 48 52 45 41 44 20 d9 14 f5 b7 a0 b7 f7 b7 78 |THREAD ........x|
[04:26:01] <tomp> 00000010 b7 f7 b7 25 0a 4d 41 58 43 48 41 4e 20 30 0a 48 |...%.MAXCHAN 0.H|
[04:26:23] <jmkasunich> so its not changing every time you look at it
[04:26:37] <tomp> hd is consistant a dozen times,
[04:26:39] <jmkasunich> thats good - that kind of gremlin would be truly insane
[04:27:33] <jmkasunich> lotta 7's in there
[04:29:19] <tomp> the thread names it could have used are in the hal file
[04:29:20] <tomp> loadrt threads name1=pdmthrd period1=500000 name2=slow period2=1000000 name3=fast period3=20000
[04:29:58] <jmkasunich> did that hal file load without errors?
[04:30:13] <jmkasunich> I thought the fastest thread had to be first
[04:30:42] <tomp> yes, and it ran once, didnt know 1st/fast rule
[04:32:37] <tomp> its got a knob to adjust duty cycle and an enable button and led for the pdmgenerator and a dro for an encoder
[04:32:39] <jmkasunich> there are several fields in that .scope.cfg that are borked
[04:33:05] <jmkasunich> THREAD, MAXCHAN, and I'm pretty sure HMULT
[04:33:17] <jmkasunich> MAXCHAN should be 1, 2, 4, 8, or 16, not zero
[04:33:44] <jmkasunich> I'm about 99.9% sure HMULT isn't supposed to be zero either
[04:34:31] <jmkasunich> I'm beginning to think the real problem was in the previous run of halscope
[04:34:50] <jmkasunich> it wrote an invalid .scope.cfg file
[04:34:54] <tomp> makes sense, but i'd have to look at doc/code to say fer sure
[04:35:05] <tomp> ok, prev run wrote bad file
[04:35:12] <jmkasunich> although the next run should reject the bad settings
[04:35:16] <skinnypuppy1334> Is anyone offering a jog handle for emc yet? I'd seen some people mention it on the mail list a few months ago.
[04:35:26] <jmkasunich> what do you mean offering?
[04:35:38] <jmkasunich> selling the actual handwheel/encoder thing?
[04:35:45] <skinnypuppy1334> yeah
[04:36:54] <tomp> no commercial product for emc that i know of, but you can buy encoders and buttons and lighta and use hal to make one
[04:36:55] <jmkasunich> there are companies that make nice ones, with knobs and detents and numbered marks on the dial
[04:36:59] <jmkasunich> aren't cheap tho
[04:37:16] <skinnypuppy1334> I've just recently finished my cnc diploma, I haven't had time to follow the developments on the forum.
[04:37:20] <jmkasunich> you could also put a homemade knob on any old encoder
[04:37:41] <jmkasunich> doesn't matter what the actual hardware looks like, as long as it makes a quadrature encoder signal when you turn it
[04:38:33] <tomp> hint ds3486 to convert differential encoder to single ended ( what i was trying to test just now )
[04:40:34] <jmkasunich> tomp: rm .scope.cfg
[04:40:38] <jmkasunich> and lets try again
[04:40:46] <jmkasunich> scripts/realtime start
[04:41:12] <skinnypuppy1334> I need to re read my digital electronics book from back in the day.
[04:41:32] <jmkasunich> bin/halcmd -vf <your-hal-file>
[04:41:41] <jmkasunich> note the -v, that stands for verbose
[04:41:46] <jmkasunich> it will echo a lot of crap
[04:42:21] <jmkasunich> bin/halscope
[04:42:26] <jmkasunich> and tell me what happens
[04:43:59] <tomp> ok, 1 terminal ran fine thru the bin/halcmd -vf l298.... and the pyvcp appeared
[04:44:21] <tomp> then in a 2nd terminal ( 1st is occupied now )
[04:44:21] <tomp> tomp@cncbox:~$ cd emc2-head/
[04:44:22] <tomp> tomp@cncbox:~/emc2-head$ bin/halscope
[04:44:21] <tomp> halscope: config file '.scope.cfg' could not be opened
[04:44:21] <tomp> Floating point exception
[04:44:57] <tomp> no halscope, prompt returned to $
[04:45:39] <tomp> pyvcp panel still present
[04:46:06] <jmkasunich> ok, three things to pastebin
[04:46:14] <jmkasunich> four
[04:46:43] <jmkasunich> everything in the first terminal (commands and results) from "rm .scope.cfg" on
[04:47:10] <jmkasunich> the result of "bin/halcmd show" (do it in the second terminal now)
[04:47:18] <jmkasunich> the .hal file
[04:47:24] <jmkasunich> the pyvcp file
[04:47:33] <jmkasunich> (I think you might have already posted those?)
[04:48:41] <tomp> everything in the first terminal (commands and results) from "rm .scope.cfg" on http://pastebin.ca/485687
[04:49:55] <jmkasunich> that directory listing starting at line 5 is because you used tab-completion to find the file?
[04:50:39] <tomp> the result of "bin/halcmd show" (do it in the second terminal now) http://pastebin.ca/485691
[04:51:22] <cradek> _petev: can you write some simple instructions about how to use at-pid? I'd like to try it tomorrow
[04:51:38] <_petev> ok, where should I put them?
[04:51:49] <jmkasunich> in a man page?
[04:52:08] <_petev> is there a good example to follow?
[04:52:12] <jmkasunich> emc2/docs/man/man9/at_pid.9 I believe is where it should go
[04:52:21] <jmkasunich> you can look at other man pages in that directory
[04:52:21] <_petev> ok
[04:52:22] <cradek> just start with any of the other manpages
[04:52:23] <tomp> reading your q now.... line 5??? wait till i look, yes, i used tab completion
[04:53:14] <tomp> the hal file http://pastebin.ca/485694
[04:53:31] <_petev> sheesh, I should have used comp, I see it made a man page for the biquad
[04:53:45] <cradek> yep that's pretty nice
[04:53:52] <jmkasunich> the problem is that it only gives you a rudimentary page
[04:54:04] <jmkasunich> you can embed more man stuff in the comp file though
[04:54:09] <_petev> after talking it over with SWPadnos, we are going to make a compromise on the biquad implmentation
[04:54:13] <_petev> I'll work on that next
[04:54:38] <jmkasunich> I think very few comps use the "embed more man stuff", the clarke transform ones do though, if you want to see the syntax
[04:54:59] <_petev> ok
[04:55:45] <jmkasunich> tomp:
[04:55:50] <jmkasunich> this isn't good:
[04:55:51] <jmkasunich> #
[04:55:51] <jmkasunich> 0 YES fast ( 3492, 27280 )
[04:55:51] <jmkasunich> #
[04:55:51] <jmkasunich> 1 parport.0.read
[04:55:51] <jmkasunich> #
[04:55:55] <jmkasunich> 2 encoder.update-counters
[04:56:02] <jmkasunich> zero is not a valid thread period
[04:56:07] <jmkasunich> congradulations, you just found a bug
[04:56:27] <tomp> do i get a prize :)
[04:56:38] <Jymmmm> meadow muffin
[04:56:47] <jmkasunich> as I mentioned, fast needs to go first, but if you have them out of order, threads needs to complain, not do the wrong thing
[04:57:05] <tomp> ok, now how do i get out of here?
[04:57:17] <jmkasunich> "bin/halcmd unload all"
[04:57:21] <jmkasunich> "scripts/realtime stop"
[04:57:33] <jmkasunich> hopefully both of those will run without errors
[04:57:39] <tomp> ok,
[04:57:44] <jmkasunich> if you get errors, speak up before you do anything else
[04:57:59] <jmkasunich> once you are out, reorder your threads command
[04:58:21] <jmkasunich> name1/period1 needs to be fastest, then name2/period2, and finally name3/period3
[04:58:41] <tomp> no errs on either cmd
[04:58:51] <jmkasunich> good, that means you got a clean shutdown
[04:59:03] <tomp> whew!
[04:59:11] <jmkasunich> the last time, when your realtime stop had a lot of errors, was because you didn't do "unload" first
[04:59:23] <tomp> ok, edit the order from fast to slow
[04:59:27] <tomp> thanks man!
[04:59:37] <jmkasunich> I'm gonna investigate the threads bug
[04:59:56] <jmkasunich> I can't guarantee that will fix your problem, if it still happens say something
[05:00:40] <_petev> jmkasunich, I have noticed several shutdowns where EMC complains it is already running on the next startup, however this startup seems to clean things up when you say not to re-start and just have it exit
[05:01:40] <jmkasunich> the next time it says that, before you answer the question, do "ps -A" and "lsmod" and save the results
[05:01:49] <_petev> ok
[05:01:52] <tomp> ok
[05:02:09] <jmkasunich> its hard to say what, if anything, its seeing that makes it think stuff is still running, but those commands should be enlightening
[05:02:56] <jmkasunich> once we know what is getting left behind we have a chance at figuring out why (and how to stop it)
[05:07:37] <CIA-8> 03cradek 07TRUNK * 10emc2/docs/man/man9/.cvsignore: new autogen page
[05:08:35] <tomp> jmkasunich: system locked up, cap lock & num lock would respond, mouse froze... right after i said 'ok' and entered 'kate' to edit the hal file. i cant report any info, system was locked tighter than a nuns buns
[05:08:53] <ohiopctechDOTcom> mmm nuns buns...
[05:08:54] <jmkasunich> wtf
[05:08:58] <tomp> wouldn't
[05:09:00] <jmkasunich> this was after a clean shutdown
[05:09:23] <jmkasunich> wait - what did you say OK to?
[05:09:35] <ohiopctechDOTcom> lol, anyone interested in a brand new 256mb usb flash drive on the cheap? i jsut found about 20 of them in a dumpster.. still in the packaging!
[05:10:02] <ohiopctechDOTcom> they are 256mb kingstons, good for transferring g-code to the machines instead of a floppy!
[05:10:12] <tomp> i didnt say ok to anything, the commands just worked, and it froze in the editor
[05:10:12] <jmkasunich> tomp: what did you say OK to?
[05:10:25] <cradek> ohiopctechDOTcom: party favors!
[05:10:39] <jmkasunich> you wrote " right after i said 'ok' and entered 'kate' to edit the hal file"
[05:10:52] <ohiopctechDOTcom> cradek: good idea
[05:11:04] <ohiopctechDOTcom> they are good for the connectors if nothing else
[05:11:19] <tomp> yes, i did say that, wtf did i mean? oh,,, i AND petev said OK, back track thru your terminal
[05:11:44] <tomp> you suggested soemthing
[05:11:52] <tomp> we both said ok
[05:11:55] <petev> ps -A, lsmod
[05:12:03] <ohiopctechDOTcom> really though, they have the 2006? version of taxcut and are still in the plastic.. never touched!
[05:12:04] <tomp> right
[05:12:18] <jmkasunich> oh, ok... "right after I said OK in the channel, I entered kate and it locked up"
[05:12:29] <ohiopctechDOTcom> i actually screamed "woohoo!" from inside the dumpster
[05:13:42] <ohiopctechDOTcom> ... so i posted an email in the Dayton Linux Usergroup mailing list today looking for some other people in dayton who have done things with emc...
[05:14:16] <ohiopctechDOTcom> oh! and i found another dead inkjet printer in the same dumpster to strip for parts...
[05:14:53] <tomp> will try again AFTER i edit the thread speed order
[05:15:00] <jmkasunich> ok
[05:15:34] <ohiopctechDOTcom> me will sell a us
[05:15:46] <ohiopctechDOTcom> me will sell a usb stick to the boss to back up our database on
[05:16:16] <ohiopctechDOTcom> ;) he'll go for that.. better than using 6 floppies!
[05:17:14] <ohiopctechDOTcom> c'mon 5 bux? nobody? hmmm...
[05:17:54] <Jymmmm> I buy 1Gb for $15
[05:18:08] <tomp> new hal file sez: loadrt threads name1=fast period3=20000 name2=pdmthrd period1=500000 name3=slow period2=1000000
[05:18:23] <jmkasunich> lol
[05:18:30] <ohiopctechDOTcom> ok, anybody can have one for postage +15%
[05:18:32] <ohiopctechDOTcom> lol
[05:18:50] <jmkasunich> tomp: why do you have name1 next to period3? thats very confusing to anybody who reads the file
[05:19:12] <jmkasunich> name1 and period1 go together, and should be the fastest thread
[05:19:21] <jmkasunich> name2 and period2 go together, and are the middle one
[05:19:31] <jmkasunich> name3 and period3 go together and are the slowest one
[05:20:15] <tomp> argh! bad cut & paste
[05:20:40] <ohiopctechDOTcom> i wonder if i could make an interface for 20 256mb flash drives and use them as one big drive
[05:20:51] <jmkasunich> RAID!
[05:21:09] <ohiopctechDOTcom> that would be 5120 mb
[05:21:17] <jmkasunich> a whopping 5G
[05:21:27] <ohiopctechDOTcom> or a 2.5g raid!
[05:21:43] <jmkasunich> you can probably get a regular 5G drive from the dumpster
[05:21:47] <tomp> ok... loadrt threads name1=fast period1=20000 name2=pdmthrd period2=500000 name3=slow period3=1000000
[05:21:55] <ohiopctechDOTcom> yeah
[05:22:00] <ohiopctechDOTcom> i find dead HD's all the time
[05:22:17] <jmkasunich> tomp: I thought the first time around that the pdmthread was the fastest one
[05:22:22] <ohiopctechDOTcom> i have been engraving stuff on the platters and putting them back together except for the top
[05:22:44] <ohiopctechDOTcom> i haven't done any in a while though
[05:22:52] <jmkasunich> tomp: sorry, my mistake
[05:22:54] <jmkasunich> you have it right
[05:23:15] <jmkasunich> fast is 50KHz, pdm is 2KHz, slow is 1KHz
[05:23:29] <tomp> but i already munged it by running halscope after loading bad threads... halscope blew up and dissappeared immediately
[05:23:37] <ohiopctechDOTcom> 4 copies of zone-alarm antivirus/firewall for windows.. anybody need that?
[05:23:49] <jmkasunich> sigh
[05:24:02] <tomp> tomp@cncbox:~/emc2-head$ bin/halscope
[05:24:02] <tomp> halscope: config file '.scope.cfg' could not be opened
[05:24:02] <tomp> Floating point exception
[05:24:02] <tomp> tomp@cncbox:~/emc2-head$
[05:24:40] <jmkasunich> please explain.... you posted it here so I could check it, I did, and pointed out some typos, you corrected them and posted again, and I said OK
[05:24:49] <jmkasunich> so, what version did you run?
[05:25:31] <ds2> Hmm
[05:26:06] <ohiopctechDOTcom> http://cgi.ebay.com/Tax-Cut-USB-2006-Tax-Cut-Federal-State-2006-USB-Flash_W0QQitemZ250113395969QQihZ015QQcategoryZ99328QQssPageNameZWDVWQQrdZ1QQcmdZViewItem
[05:26:10] <ohiopctechDOTcom> 20 of those***
[05:26:16] <ohiopctechDOTcom> free, lol!
[05:26:20] <ohiopctechDOTcom> muahahaha
[05:26:21] <tomp> i ran the version you found the errs in before you reported it, THEN i corrected them THEN i told you "i already munged it by running....
[05:27:17] <jmkasunich> why post if you ain't gonna wait for feedback?
[05:27:38] <jmkasunich> you probalby have a zero period thread again
[05:27:55] <tomp> it's asynchronous when i leave here, go to the terminal , enter commands, return to irc and see your message
[05:27:56] <jmkasunich> I'm trying to investigate that, between talking to people
[05:28:27] <tomp> right i appreciate it, not trying to confuse you
[05:30:12] <tomp> and yes, i got a zero thread again
[05:30:15] <tomp> Realtime Threads:
[05:30:15] <tomp> Period FP Name (Time, Max-Time)
[05:30:15] <tomp> 0 YES fast ( 3740, 28148 )
[05:30:54] <jmkasunich> can you do one more thing before you shut it down to try again?
[05:30:58] <tomp> yes
[05:32:00] <ds2> OT q - any model engine builders/designers?
[05:32:10] <jmkasunich> "cat /proc/rtapi/*"
[05:32:13] <jmkasunich> and pastebin the result
[05:33:51] <tomp> jmkasunich: http://pastebin.ca/485732
[05:34:14] <jmkasunich> thanks
[05:34:37] <jmkasunich> see line 43?
[05:34:56] <jmkasunich> that means that no thread can run faster than ~500000nS
[05:34:56] <tomp> yeh, what that?
[05:35:22] <jmkasunich> period1 is the period of the first thread, and that period is the one that sets the hardware timer
[05:35:41] <jmkasunich> any thread that comes after that has to be slower than the hardware timer
[05:36:53] <tomp> ok, and i asked for faster than 500uS ( 500000nS )
[05:37:06] <jmkasunich> right, you asked for 20000
[05:37:33] <jmkasunich> you gotta ask for that first, then the timer will get set to 20000, and all the others will be OK, because they're slower
[05:37:54] <jmkasunich> you might not get exactly 20000, there are hardware limitations, so it rounds to the nearest value the hardware can do
[05:38:32] <tomp> oh, it'll be ok if i can manage to type it in correctly (order and pairing )and execute it correctly
[05:38:56] <jmkasunich> I noticed something else in your hal file that you probably should fix
[05:39:01] <tomp> ?
[05:39:10] <jmkasunich> in the addf section
[05:39:25] <jmkasunich> you read the parport in the fast thread, then do encoder counting in that thread - good
[05:39:58] <jmkasunich> you generate PWM in the medium thread (2KHz), and write the parport in the slow thread (1KHz)
[05:40:01] <jmkasunich> thats not so good
[05:40:36] <jmkasunich> in fact, unless you are using very slow PWM, you should put the pwmgen and the parport write in the fast thread
[05:40:43] <jmkasunich> and eliminate the middle one completely
[05:41:25] <jmkasunich> if you are using very slow pwm, you can run the pwmgen in the middle thread, but you should run parport_write in the same thread (and after pwmgen) so the outputs of pwmgen get written to the hardware
[05:44:01] <tomp> i was using the PDM variant does that make any difference? I will try pwmgen AND parport_write (after pwmgen in sequence) in medium thread, and remove slow thread
[05:44:32] <jmkasunich> pdm won't be a problem
[05:44:52] <tomp> ok, thanks jmkasunich!
[05:44:53] <jmkasunich> I dunno why you want to get rid of the slow thread
[05:45:41] <jmkasunich> pwmgen.update tells the pdm generator "here's what you should produce for the next millisecond (or some other slow time period)"
[05:46:06] <jmkasunich> pwmgen.make-pulses is what actually produces the PWM, and should run as fast as possible
[05:46:14] <jmkasunich> what are you controlling with this PWM?
[05:46:27] <tomp> ? i misread your suggestion, ok, pwmgen.makepulses stays in slow thread
[05:46:35] <jmkasunich> no no no
[05:46:50] <jmkasunich> pwmgen.make-pulses is what actually produces the PWM, and should run as FAST as possible
[05:48:53] <jmkasunich> like so:
[05:48:59] <jmkasunich> # read inputs and count encoder
[05:48:59] <jmkasunich> addf parport.0.read fast
[05:48:59] <jmkasunich> addf encoder.update-counters fast
[05:48:59] <jmkasunich> # make pwm and write it to outputs
[05:48:59] <jmkasunich> addf pwmgen.make-pulses fast
[05:49:00] <jmkasunich> addf parport.0.write fast
[05:49:01] <jmkasunich> # do main control calculations
[05:49:03] <jmkasunich> addf pwmgen.update slow
[05:49:06] <jmkasunich> addf encoder.capture-position slow
[05:50:14] <tomp> ok, so MEDIUM thread is not wanted, right?
[05:54:09] <jmkasunich> right
[05:54:37] <jmkasunich> since I know nothing about your application, I can't be sure what period you need for your various threads
[05:54:46] <CIA-8> 03petev 07TRUNK * 10emc2/docs/man/man9/at_pid.9: -Initial revision.
[05:54:52] <tomp> yet, i think reassigning the PDM to the 'slow' thread changes the pulse ( the width of the pulse not the frquency)
[05:55:06] <jmkasunich> I am sure that you don't want the parport write to be slower than the pwmgen.make-pulses
[05:55:25] <jmkasunich> of course it does
[05:55:26] <jmkasunich> thats the point
[05:55:53] <tomp> ok, i think i played a while to get that period correct, now it will be different and is tied to other constraints
[05:55:58] <jmkasunich> if you run it at 2KHz, then the output pulse width is limited to 500uS increments
[05:56:04] <jmkasunich> stop everything
[05:56:07] <tomp> k
[05:56:16] <jmkasunich> please tell me what the heck you are doing with the PDM
[05:56:22] <jmkasunich> running a motor?
[05:56:27] <jmkasunich> blinking a light?
[05:56:29] <tomp> move a motor very slow, for edm
[05:56:40] <tomp> like 1mm/day
[05:57:27] <jmkasunich> so are you trying to create 500uS pulses into the motor every once in a while, or are you trying to create a very low average voltage?
[05:57:28] <tomp> (exxaggeration, but tells the story )
[05:58:33] <jmkasunich> with a 2KHz thread, when you command 0.1% duty cycle, you will get one 500uS long pulse every 500mS (one pulse every 1/2 second)
[05:58:41] <tomp> i'm not sure if i wanted a Nsec pulse,, this was a test, but it looked like a sizable pulse occasionally was better than PWM
[05:59:14] <jmkasunich> with a 50KHz thread, when you command 0.1% duty cycle, you will get one 20uS long pulse every 20mS (50 pulses per second, but each pulse much weaker)
[05:59:40] <jmkasunich> it could be that at very low speeds, one biggish pulse is better than a bunch of tiny ones
[05:59:49] <jmkasunich> so you probably should keep the medium thread
[06:00:20] <tomp> right, i was able to get lower duty cycle with PDM than with PWM.... PDM gave me response at 0.00n % , PWM failed at .1%
[06:00:31] <tomp> failed = no actual motion
[06:00:48] <tomp> so the big pulse was better than the average cvoltage
[06:01:37] <jmkasunich> keep the middle thread, and put both "pwmgen.make-pulses" and "parport.0.write" in it, in that order
[06:01:57] <jmkasunich> then you can mess with the period of the middle thread to see what works best
[06:02:02] <tomp> k, this really is helpful, i wouldnt have seen that
[06:03:32] <tomp> # read inputs and count encoder
[06:03:32] <tomp> addf parport.0.read fast
[06:03:32] <tomp> addf encoder.update-counters fast
[06:03:32] <tomp> # make pwm and write it to outputs
[06:03:32] <tomp> addf pwmgen.make-pulses pdmthrd
[06:03:32] <tomp> addf parport.0.write pdmthrd
[06:03:34] <tomp> # do main control calculations
[06:03:36] <tomp> pwmgen.update slow
[06:03:38] <tomp> addf encoder.capture-position slow
[06:03:41] <tomp> ?
[06:04:07] <jmkasunich> you need an addf at the beginning of the pwmgen.update line (next to last line)
[06:04:11] <jmkasunich> but the rest is exactly right
[06:04:33] <petev> jmkasunich, is there a hal function to return the period of a thread a function is being run by?
[06:04:48] <jmkasunich> its passed to the function
[06:04:50] <tomp> jmkasunich: got it, thanks
[06:04:59] <petev> what's passed is a jittery actual period
[06:05:14] <petev> I want the period that the thread is supposed to be running at
[06:05:24] <petev> ie, what motion is using for calcs, etc.
[06:05:35] <jmkasunich> I'm about 99.9% sure what's passed is the actual period
[06:05:38] <jmkasunich> lemme check the code
[06:05:42] <petev> ok
[06:06:00] <jmkasunich> actual meaning programmed meaning exactly what you want
[06:06:08] <petev> ok, great
[06:06:27] <petev> so it's not the actual time between runs? the same value is always passed?
[06:07:51] <jmkasunich> right, same value is always passed
[06:08:03] <petev> ok
[06:08:21] <jmkasunich> there doesn't seem to be a clean and fast interface for getting actual time in nS or other reliable time units
[06:08:34] <petev> ok
[06:08:35] <jmkasunich> so it would be very hard to pass the actual time
[06:09:06] <jmkasunich> and as you say, you usually want the programmed period, not the jittery actual one
[06:09:31] <petev> right, except for timer stuff, but even then it's debatable
[06:10:01] <jmkasunich> we found about a year back that RTAI's get_time_ns function took many microseconds to run
[06:10:12] <petev> nice ;-)
[06:10:23] <jmkasunich> so now the time logging stuff is done in clocks instead of seconds
[06:18:12] <Jymmmm> lol
[06:23:29] <tomp> jmkasunich: it runs fine now, and i can exit clean with "bin/halcmd unload all; scripts/realtime stop", and then run again w/o problems. THANKS!
[06:23:40] <jmkasunich> you're welcome
[06:24:11] <jmkasunich> sorry about the troubles - I'm changing things now to detect that error and refuse to run
[06:24:39] <tomp> stupid user, trix are for kids!
[06:27:41] <tomp> i'll begin looking for a big gear reduction, and replace the 1 'kick' every 12 seconds with PWM and a constant feed
[06:28:53] <Jymmmm> * Jymmmm scratches his head... how to blink an LED in 10,000 easy steps!
[06:29:11] <jmkasunich> step 1: get LED
[06:29:15] <Jymmmm> got it
[06:29:16] <jmkasunich> step 2: get battery
[06:29:20] <Jymmmm> got it
[06:29:24] <jmkasunich> step 3: connect LED to battery
[06:29:31] <jmkasunich> step 4: disconnect LED from battery
[06:29:36] <jmkasunich> step 5: goto step 3
[06:29:38] <Jymmmm> Did I forget to mention this is via a ATMEL uC?
[06:29:55] <jmkasunich> yes, you forgot to mention that
[06:30:06] <Jymmmm> jmkasunich: Actually, you missed a step.
[06:30:36] <Jymmmm> Step 3.5: Get another LEd cause you just blew that one up, goto step 3
[06:30:45] <jmkasunich> nah
[06:30:49] <jmkasunich> just use a weak battery
[06:30:52] <Jymmmm> 12V battery =)
[06:31:39] <Jymmmm> AND.... I'm programming th uC via BASIC! LOL
[06:31:40] <jmkasunich> step 2: get suitable battery (and don't get it from that troublemaker Jymmmm)
[06:31:47] <Jymmmm> rotf
[06:32:07] <Jymmmm> too bad the chip I'm using is no longer made!
[06:33:34] <Jymmmm> I GUESS I should see what the replacement is
[06:34:07] <Jymmmm> ah... ATtiny2313
[06:34:55] <Jymmmm> Slower than shit, but ya gotta love an internal RC osc =)
[06:38:11] <Jymmmm> Now they're getting it... "Novice's Guide to AVR Development"
[06:40:15] <tomp> woohoo! this ds3486 fixed up my encoder! i was getting real bad readings, now 1 rev is 1mm ( 1um resolution!) i was worried the reader head was munged
[06:40:49] <tomp> i was using just A and B, now using A /A B & /B with the differential decoder
[06:45:01] <tomp> thanks all, g'nite!
[06:52:48] <CIA-8> 03jmkasunich 07TRUNK * 10emc2/src/hal/hal_lib.c: enforce the existing requirement that new threads may not be faster than existing ones
[07:05:57] <CIA-8> 03jmkasunich 07TRUNK * 10emc2/docs/man/man3/hal_create_thread.3hal: document requirement that threads be created fastest first
[07:05:57] <CIA-8> 03jmkasunich 07TRUNK * 10emc2/docs/man/man9/threads.9: document requirement that threads be created fastest first
[07:06:41] <jmkasunich> the man page for threads actually says that you have to create them fastest first, but the code didn't check to see if you actually obey that rule
[07:06:43] <jmkasunich> now it does
[07:06:54] <Jymmmm> heh
[07:07:23] <Jymmmm> Gawd how I HATE plastic radiators!
[07:07:42] <jmkasunich> goodnight all
[07:07:46] <Jymmmm> nite!
[07:11:54] <petev> gn
[07:12:11] <Jymmmm> nite
[07:43:13] <CIA-8> 03petev 07TRUNK * 10emc2/src/hal/components/biquad.comp:
[07:43:13] <CIA-8> -Added enable and valid pins.
[07:43:13] <CIA-8> -Added type, f0, and Q params.
[07:43:13] <CIA-8> -Added coefficient calculation for basic low pass and notch filters.
[07:46:24] <ds2> hmm
[07:52:15] <CIA-8> 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
[07:53:28] <petev> hmm, what is that, errors in math.h on the farm?
[13:36:08] <CIA-8> 03jepler 07TRUNK * 10emc2/src/hal/components/biquad.comp: the identifier 'type' is used (probably incorrectly) by breezy's math.h, which caused an error when the comp convenience define for the parameter was present
[13:44:50] <CIA-8> 03compile-farm 07Ubuntu 5.10 (breezy) non-realtime (2.6.12-10-386) * 10emc2head/: build PASSED
[16:12:09] <xemet> hi
[16:12:31] <ohiopctechDOTcom> yo!
[16:12:36] <xemet> simple idea for pyVCP...would be nice to have a widget to put images
[16:12:55] <xemet> so one could put small icons near to buttons or labels
[16:19:28] <Guest884> well, my router disconnected me...
[16:20:01] <Guest884> oh, xemet again.
[16:20:05] <Guest884> noooo
[16:20:45] <Guest710> so...
[16:20:49] <Guest710> ok, I'm xemet
[16:21:00] <Guest710> I don't know why he call me Guest710
[16:21:10] <tomp> Guest710: got one.. (images for pyvcp) but pastebin.ca and pastebin.com both are down... it's not standard but an easy patch, maybe 12 lines
[16:21:51] <Guest710> so you've a patch to put images using pyVCP?
[16:21:55] <tomp> yes
[16:22:16] <Guest710> but couldn't be added to the official version?
[16:22:26] <tomp> the python name for the widget is PhotoImage, no not official
[16:22:58] <Guest710> ok, I don't need it right now, but in the future I will ask to you
[16:23:07] <tomp> k
[16:23:57] <Guest710> so, second question, is it possible to launch halscope from halcmd with channels configured, samples etc. so that will be opened a window ready to be used?
[16:24:08] <tomp> anyone know a service besides pastebin.ca and pastebin.com? bytecolor? you use something else?
[16:24:53] <Guest710> I'm thinking about halmeter...you can launch it like halmeter pin ddt.out (if I remember well)
[16:25:13] <Guest710> and halmeter will open with the pin you specified.
[16:25:13] <ohiopctechDOTcom> just post it on your myspace or yahoo profile or something
[16:25:14] <bytecolor> tomp, umm I really dont have a specific I use, deadbeefbabe.org, (#python) or slexy.org
[16:25:30] <Guest710> is it possible to do a similar thing with halscope?
[16:26:01] <tomp> Guest710: halscope will remember the last scope settings, so passing params is not needed, , and you can experiment with running halscope from the .hal file ( i dont know from experience)
[16:26:18] <tomp> bytecolor: ohiopctechDOTcom: thanks
[16:27:21] <tomp> xemet: http://deadbeefbabe.org/paste/4745
[16:27:29] <Guest710> yes...but I will need it for few little hal programs in order to test things, so would be nice if I open a program I've an halscope with configured channels for that program, if I open another I will have halscope with other settings
[16:27:58] <ohiopctechDOTcom> does hal have .ini files?
[16:28:10] <ohiopctechDOTcom> use different .ini's for each prog
[16:28:22] <ohiopctechDOTcom> or two different home directories?
[16:28:28] <Guest710> hal has .ini files??
[16:28:34] <ohiopctechDOTcom> i was asking
[16:28:39] <Guest710> ah ok
[16:28:42] <ohiopctechDOTcom> note the "?" ;)
[16:29:00] <Guest710> yes I toght you know that hal has ini files,
[16:29:14] <Guest710> I'm not very expert
[16:29:27] <tomp> Guest710: the info is saved in a file ".scope.cfg", saved in the (for me...) /home/tomp/emc2-head/.scope.cfg , so if you want 2 configs, you'll have to swap the files
[16:29:28] <Guest710> and english is not my first language so forgive me
[16:30:06] <Guest710> tomp, so to use that patch...what should I do?
[16:30:09] <tomp> so if you work in a different directory for each .scope.cfg, then the saved info may be good for you
[16:30:52] <Guest710> ...uhm so If I have a directory for every program and in every directory a different scope.cfg...it should work
[16:30:55] <Guest710> right?
[16:31:47] <tomp> Guest710: read 1st line: # pyvcp_widgets.py patch to allow Photimage you add it to pyvcp_widgets.py
[16:32:32] <Guest710> ...add where in the file?
[16:32:40] <tomp> Guest710: regarding scope... if you have a separate directory for each EMC then the .scope.cfg may suffice. I dont knwo your setup
[16:32:43] <Guest710> at the beginning, at the end, indiffirent?
[16:32:49] <ohiopctechDOTcom> you could write a script to place to correct config files in the right dir at the time you want to run each respective config
[16:32:50] <tomp> Guest710: indifferent
[16:33:21] <Guest710> ok...the file is python so it doesn't need compiling right?
[16:34:29] <tomp> Guest710: right, python needs no compile, and, you will see similar 'class' statements in the file , so put the new code in that area
[16:34:35] <Guest710> wow, the ghost xemet has quit...
[16:34:45] <tomp> doppelgangers!
[16:35:05] <Guest710> tomp, and what is the xml code to insert an image?
[16:35:38] <ohiopctechDOTcom> my boss just called me from work with the MOST stupid computer question ever...
[16:35:54] <ohiopctechDOTcom> "how do i logg the other ebay name off and log the company name in?"
[16:36:14] <tomp> Guest710: here's example http://deadbeefbabe.org/paste/4746
[16:36:16] <ohiopctechDOTcom> ITS FSCKING SUNDAY! MY DAY OFF!
[16:36:22] <ohiopctechDOTcom> geez.. :)
[16:36:30] <barum> :)
[16:36:31] <Guest710> nice, he pays you for replies to those questions?
[16:36:53] <barum> Hi to all
[16:36:59] <ohiopctechDOTcom> yeah.. i'll just sit outside for 20 mins tomorrow after i clock in... ;)
[16:37:19] <ohiopctechDOTcom> hi barum
[16:37:58] <Guest710> great, thank you tomp
[16:39:26] <barum> My question: Is it there some implemented protocol to send data other than bits on port pins per motor?
[16:39:47] <ohiopctechDOTcom> what other data? like a mp3?
[16:39:56] <barum> Buut not Gcode itself couse complex
[16:40:01] <ohiopctechDOTcom> or a pic of cindy crawford ina swimsuit?
[16:40:08] <tomp> barum: yes some pins are floats, suited to analog values
[16:40:19] <ohiopctechDOTcom> oh.. like digits..
[16:40:27] <barum> data like X Y Z fixed point values
[16:40:40] <barum> max speed
[16:40:58] <ohiopctechDOTcom> hm... i'm not sure on that, i'd guess no.
[16:41:06] <tomp> barum: yes, all your example dtat is implemented,no special protocol, just like any other pin
[16:41:25] <tomp> barum: the pin 'knows' the data type
[16:41:56] <tomp> barum: read the hal manual, you will see the same things you asked for
[16:42:25] <ohiopctechDOTcom> all data is bits... the thing on the other end would decdide how to treat the data.. right?
[16:42:43] <ohiopctechDOTcom> or bytes*
[16:43:09] <tomp> yes, the manual explains it
[16:43:59] <jepler> barum: every time the motion-controller function runs, the axis.N.motor-pos-cmd is updated to hold the requested motor position.
[16:44:42] <jepler> barum: many people hook that to the step generator to create step and drection bit outputs, but you can do other things (such as send it to a PID loop that controls a servo motor to go to the correct position)
[16:44:54] <Guest710> tomp, why this patch is not in the official version? I think would be a nice feature!
[16:46:06] <tomp> Guest710: http://imagebin.org/8579 i dont know, i offered it. use with caution... it wont be in the next update, so you have to keep track
[16:46:08] <barum> I make controler based on PIC that uses serial port to get vectors
[16:46:48] <tomp> barum: vectors of new destination for motors?
[16:47:05] <barum> So thinking to make parser of Gcode to calc and send
[16:47:09] <Guest710> tomp, very nice...I hope it will be included in the official version one day!
[16:47:22] <tomp> Guest710: thanks
[16:47:52] <barum> vectors are kx, ky, kz, delta_t, acceleration
[16:47:58] <barum> for ramping
[16:48:00] <Guest710> and would be nice to have a way to put images in the buttons too
[16:48:42] <tomp> barum: you should read about the 'hal-streamer'. i am not knowledgeable about it, but it may be suited to what you want. the concept is NOT inside EMC, but IS inside HAL. EMC would want to do those calculations for you.
[16:48:53] <tomp> Guest710: yes, i hacked btns as well
[16:49:31] <barum> OK, do I have to write hal module?
[16:49:33] <jepler> barum: emc does acceleration and blending. It assumes that whatever it is controlling takes a position request and returns an actual position, typically every 1ms.
[16:49:52] <jepler> barum: if your system is radically different from this, it will be a poor match for emc.
[16:50:20] <barum> It returns position when finished.
[16:50:45] <Guest710> tomp, where do I find the pyvcp_widgets.py ?
[16:50:49] <barum> I may change that
[16:51:03] <Guest710> (in the installed version, not run in place)
[16:51:06] <tomp> barum: sorry, you speak of motion control, of velocity profile, not just destination... you're past me :)
[16:52:18] <tomp> Guest710: FIRST: backup the original! then, for me, it's in ~/emc2-head/lib/python/
[16:52:53] <jepler> Guest710: in the installed version, pyvcp_widgets.py is "byte compiled" into pyvcp_widgets.pyc, and copied to somewhere in /usr/lib
[16:53:40] <Guest710> jepler....so it is compiled...and I imagine I cannot modify it, right?
[16:53:48] <CIA-8> 03jepler 07v2_1_branch * 10emc2/docs/src/config/emc2hal.lyx: update docs to reflect changing this pin to I/O type
[16:54:19] <CIA-8> 03jepler 07TRUNK * 10emc2/docs/src/config/emc2hal.lyx: merge rev 1.2.2.8: name and direction of spindle sync pin
[16:54:26] <Guest710> well... now I try to relog with my nick...
[16:54:34] <tomp> Guest710: jepler's right, the .pyc is automaticly created whn a new file is found ( the compili]ation is no obstruction) but my pyc are in the same directory
[16:54:48] <xemet> so
[16:54:59] <xemet> oooh, xemet returned
[16:55:05] <tomp> xemet: Guest710: jepler's right, the .pyc is automaticly created whn a new file is found ( the compili]ation is no obstruction) but my pyc are in the same directory
[16:55:35] <tomp> xemet: your system may be different, go look :)
[16:56:24] <tomp> xemet: if i want a new .pyc, i just remove the old one and edit the .py file, next runtime, it get reborn
[16:57:06] <barum> Where i can find some source for HAL to start with?
[16:57:09] <barum> Something more than just headers.
[16:57:10] <xemet> wait a moment, let me understand, so I've EMC2 installed, if I search the file pyvcp_widgets.py and I modify it, will the pyc be automatically recreated?
[16:57:36] <jepler> I don't guarantee what will happen if you put a new pyvcp_widgets.py in the middle of an installed package. do not cut corners -- either use RIP (where editing .py works without extra steps) or build a fresh package / "make install" with the desired modifications.
[16:57:37] <tomp> xemet: yes , but back it up 1st!
[16:58:16] <xemet> what's RIP?
[16:58:20] <tomp> xemet: listen to jepler: i dont know that stuff, but mine works as you saw
[16:58:29] <jepler> ./configure --enable-run-in-place (RIP = Run In Place)
[16:58:33] <xemet> ok
[16:59:00] <jepler> then you don't install anything outside the emc2 directory you got from CVS checkout or whatever
[16:59:35] <jepler> in that case, you can edit or patch lib/python/whatever.py to your hearts content with no need to worry about .pyc files
[16:59:48] <tomp> xemet: i never heard RIP either, thats what i do, i use run in place as in ~/emc2-head/...
[16:59:51] <jepler> (but if you edit py files in src/ you have to 'make' to have them copied to the right location within the emc2 directory)
[17:00:01] <xemet> yes, I know, but would be nice to have that feature in the installed version I've on my PC...it is not run in place
[17:00:38] <xemet> so maybe I will try the feature in my run-in-place version...and will wait an official release to use it in the installed version
[17:01:24] <CIA-8> 03jepler 07TRUNK * 10emc2/debian/ (configure rules.in): new bfload binary is setuid
[17:01:32] <ohiopctechDOTcom> sweet my buddies got another wrt54g for me to hack :)
[17:01:43] <Unit41> the 64 bit dev's are hiding ?
[17:01:50] <Unit41> :P
[17:02:52] <ds2> ohiopctechDOTcom: what are you putting on them?
[17:02:58] <ohiopctechDOTcom> ddwrt
[17:04:50] <ohiopctechDOTcom> i set them up to latch on to any signal they find and makes a subnet to distribute to the switch
[17:04:50] <ohiopctechDOTcom> and my buddies love it!
[17:04:50] <ohiopctechDOTcom> lol
[17:04:49] <ds2> oh that's not hacking...
[17:04:49] <ds2> ;)
[17:04:50] <ohiopctechDOTcom> well, kinda
[17:04:50] <ohiopctechDOTcom> it's hacking in the general sense of making something do something it wasn't made to do
[17:05:00] <ohiopctechDOTcom> aKa: tweaking?
[17:06:05] <ds2> thought you were doing something dramatic like putting asterisk on there
[17:06:27] <ohiopctechDOTcom> i did have one running a webserver
[17:06:37] <ds2> they are very nice and cheap dev boards for MIPS stuff
[17:06:54] <ohiopctechDOTcom> yeah, maybe the cheapest complete mips computer you can buy?
[17:07:10] <ds2> for $10-$20, yep
[17:07:27] <ohiopctechDOTcom> i sold the last ddwrt i did up on ebay for 45 bux
[17:07:31] <ds2> and you can run them off a battery (they burn 10W max)
[17:07:34] <ohiopctechDOTcom> and i got it free from a dumpster
[17:07:54] <ohiopctechDOTcom> :| i should stop giving away my secrets
[17:07:55] <ohiopctechDOTcom> lol
[17:07:56] <ds2> I prefer the ASUS WL500 though... comes with a parallel port + USB
[17:08:08] <ohiopctechDOTcom> i think you can add serial to the ddwrt
[17:08:11] <ds2> you are lucky, you don't have too much competition out where you are
[17:08:30] <ds2> plain WRT54 requires hardware mods for serial/usb
[17:08:38] <ohiopctechDOTcom> i run into people all the time... they get scared and leave when i pull up
[17:08:54] <ohiopctechDOTcom> the only real DD competition i have is people looking for books!
[17:09:26] <ds2> People do DD all the time around here so it is hard to find one with anything good in there
[17:09:40] <ohiopctechDOTcom> where at?
[17:09:50] <ds2> San Jose, CA
[17:10:05] <ohiopctechDOTcom> if i had a better vehicle i would expand my operation tri-state
[17:10:34] <ohiopctechDOTcom> i'm working with 170,000 miles on the OD
[17:10:38] <ohiopctechDOTcom> HAHAH!
[17:10:56] <ohiopctechDOTcom> paint still looks good though!
[17:11:49] <ohiopctechDOTcom> go to places people wouldn't normally go
[17:11:57] <ohiopctechDOTcom> that's where you find the cool shit..
[17:15:30] <ds2> model t ford?
[17:15:32] <ds2> ;)
[17:15:54] <ohiopctechDOTcom> close 1990
[17:21:09] <xemet> simple question...
[17:21:30] <xemet> I've a very simple bash script that launch a hal file by halcmd:
[17:21:44] <xemet> halcmd -f syncro-resolver.hal
[17:22:01] <jmkasunich> syncros and resolvers? sounds interesting ;-)
[17:22:29] <xemet> this .hal file has a pyVCP...when I run it it load the widgets and...after it loaded them it closes them...
[17:23:02] <xemet> have I to add something at the end of the script in order to wait so I do something before it closes the window?
[17:23:06] <jepler> man halcmd, search for "waitusr"
[17:23:34] <jepler> presumably you would want to "wait" for the panel to exit, so you would use the same component name as when you did the "loadusr" to create it.
[17:23:47] <xemet> ...is the problem with the bash script or with the hal file?
[17:23:58] <xemet> if I launch the hal file from halcmd everything works
[17:24:32] <jmkasunich> what does the bash script look like?
[17:24:47] <xemet> or...better, If I: 1) open a terminal 2) start realtime 3) launch halcmd -f syncro-resolver.hal
[17:24:55] <xemet> everything work..
[17:25:14] <xemet> the script is: #!/bin/bash
[17:25:28] <xemet> halcmd -f syncro-resolver.hal
[17:25:58] <xemet> and nothing else...I'm not very expert with bash...or I'm not expert at all...this is the first time I use a bash script
[17:27:15] <jmkasunich> this is hard to understand
[17:27:20] <jmkasunich> you say if you do:
[17:27:37] <jmkasunich> realtime start; halcmd -f syncro-resolver.hal
[17:27:41] <jmkasunich> it works fine
[17:27:55] <jmkasunich> (meaning, at the end of those two commands you have widgets on your screen)
[17:27:58] <jmkasunich> but if you do:
[17:28:09] <jmkasunich> realtime start; some_script
[17:28:11] <jmkasunich> it doesn't work
[17:28:25] <jmkasunich> where the only thing in "some_script" is "halcmd -f syncro-resolver.hal"
[17:28:29] <xemet> ...it load the widgets and closes them...
[17:29:04] <xemet> well...this is the reason I'm asking...I thought it was strange because the script has only one line that does the same thing I do from terminal
[17:29:46] <jmkasunich> instead of -f in the halcmd, use -vf
[17:29:57] <jmkasunich> that means "verbose" and it will print messages as it runs
[17:30:01] <jmkasunich> maybe that will tell you something
[17:30:09] <xemet> I try
[17:31:01] <xemet> ok, it wrote things and closed the terminal window...so I cannot read them
[17:31:21] <xemet> maybe I've to add something at the endof the script?
[17:31:44] <jmkasunich> why doesn't it just give you back your prompt?
[17:33:06] <xemet> wait...that is the problem...I was run it by double click and selecting "run from terminal"
[17:33:19] <jmkasunich> argh
[17:33:25] <xemet> now I've tried running it from terminal and it works
[17:33:29] <jmkasunich> clicking is evil
[17:33:53] <xemet> so, is there any way to run it in the evil mode (by clicking)?
[17:34:24] <jmkasunich> you need to prevent the script from ending until you are done using it
[17:34:29] <jmkasunich> that is where waitusr comes in
[17:34:35] <CIA-8> 03petev 07TRUNK * 10emc2/configs/boss/ (pid_test.hal pid_test.ini): -Initial revision.
[17:35:11] <jmkasunich> the last line of the hal file should be a waitusr command that waits for pyvcp, that will prevent the file from ending until you close the pyvcp window
[17:35:13] <CIA-8> 03petev 07TRUNK * 10emc2/configs/boss/ (boss.hal boss.ini): -Some updates after testing.
[17:35:50] <xemet> ok great, now another question...is it possible to do a clean up when I close the pyVCP window?
[17:36:10] <xemet> for example if I put "unload all" after the waitusr...will it work?
[17:36:22] <jmkasunich> yes
[17:36:32] <xemet> ok! now I try
[17:37:00] <jmkasunich> your bash script can start realtime, invoke halcmd -f, then when the hal script ends (with unload all), it can stop realtime
[17:37:19] <jmkasunich> or you can use halrun. which does that for you without making you write the script
[17:39:26] <xemet> ok, now everything work...I've done it without halrun, i hate halrun
[17:39:38] <ohiopctechDOTcom> so much anger
[17:39:38] <jmkasunich> ok ;-)
[17:39:58] <tomp> nice! i added unload all to my hal files, and scripts/realtime start ... stop to the bash script ... cool!
[17:42:46] <xemet> ok, now I've a nice icon and when I click it everything starts
[17:43:14] <xemet> if only could be avoid the "run in terminal" question...would be nice
[17:43:19] <xemet> any way to do that?
[17:43:30] <jmkasunich> I don't know - I never use icons
[17:43:33] <xemet> I would like to dounle click it and let it simple start
[17:43:58] <xemet> jmk I know, you programmers don't use clicking or icons or all that stuff...
[17:44:09] <tomp> about the halrun technique... just do halrun -kf blah.hal?
[17:44:29] <jmkasunich> tomp: I think just halrun blah.hal
[17:44:40] <tomp> trying now
[17:45:03] <xemet> but students who never used linux like to double click on something and see it working
[17:45:04] <jmkasunich> xemet: I
[17:45:06] <jmkasunich> oops
[17:45:28] <jmkasunich> xemet: I am sure it is possible to make an icon that will run a program or script without asking questions
[17:45:38] <jmkasunich> but I can't help you do that because I've never done it myself
[17:46:12] <jmkasunich> if you right click on the desktop, one menu item is "create launcher"
[17:46:26] <jmkasunich> the create launcher dialog has a button for "run in terminal"
[17:46:31] <xemet> I'm sure too...after chatting in this channel I discovered it is possible to make everything...the problem is always how to
[17:46:46] <tomp> jmkasunich: halrun blah.hal runs fine & exits clean
[17:46:53] <jmkasunich> tomp: good
[17:49:48] <xemet> jmk ok, create launcher works, now I've a nice icon that starts everything I need
[17:53:01] <SWPadnos_> SWPadnos_ is now known as SWPadnos
[18:28:20] <lerneaen_hydra> quiet here
[18:28:36] <alex_joni> * alex_joni drops a needle
[18:42:25] <anonimasu> * anonimasu swallows the needle
[18:42:32] <jmkasunich> ouch
[18:44:03] <anonimasu> my digestion rocks ;)
[18:44:55] <lerneaen_hydra> like the french dude?
[18:45:13] <anonimasu> yeah
[18:45:21] <Unit41> where is my RT dir located ?
[18:45:23] <lerneaen_hydra> he's, strange
[18:45:53] <Unit41> --with-realtime=<path>
[18:48:26] <Unit41> aha I had to run aclocal
[18:48:53] <alex_joni> Unit41: depends on your setup
[18:49:35] <Unit41> should I pastebin my error log ?
[18:49:39] <alex_joni> Unit41: sure
[18:51:13] <Unit41> www.pastebin.ca/486052
[18:51:28] <alex_joni> brb..
[18:58:02] <Unit41> 64 bit
[19:04:12] <alex_joni> back
[19:04:13] <alex_joni> Can't write 4320 bytes at offset 72 from buffer of size 4064.
[19:04:25] <alex_joni> it seems the buffer size is too small to fit something inside
[19:05:21] <alex_joni> http://cvs.linuxcnc.org/cgi-bin/cvsweb.cgi/emc2/configs/common/emc.nml.diff?r1=1.5;r2=1.6;only_with_tag=RELEASE_2_1_4
[19:05:27] <alex_joni> apply that change backwards
[19:53:36] <Unit41> how ?
[19:57:12] <SWPadnos> Unit41, patch -r -p1 < this/patch/file
[19:57:24] <SWPadnos> or similar (see man patch for the real options ;) )
[19:58:13] <alex_joni> Unit41: it's just one line
[19:58:26] <alex_joni> you need to open emc.nml (located in emc2/configs/common)
[20:03:19] <alex_joni> and edit the line with toolSts
[20:03:31] <alex_joni> change the size from 4096 (in your file) to 8192
[20:03:33] <JymmmEMC> Heh, thats kinda cool... http://www.google.com/search?q=la+time
[20:04:42] <JymmmEMC> http://www.google.com/search?q=singapore+time
[20:09:59] <tomp> yeh, no need for java and us gov time
[20:10:40] <SWPadnos> hmmm. should I upgrade my A64 box to Feisty, or just do the 197M of updates to Edgy?
[20:10:45] <SWPadnos> decisions decisions
[20:11:07] <JymmmEMC> SWPadnos: Got 10GigE ?
[20:11:27] <Unit41> www.pastebin.ca/486207
[20:11:28] <SWPadnos> hmmm - lemme check
[20:11:32] <SWPadnos> no, no I don'[t
[20:11:57] <SWPadnos> I have 2MegE, for the most part ;)
[20:12:25] <SWPadnos> what the heck. I'd love to see some of the XGL effects at 3840x2400 resolution
[20:12:26] <JymmmEMC> SWPadnos: Well, then you should just dl the internet-on-a-floppy then
[20:13:20] <JymmmEMC> SWPadnos: you still playing with atmel uC's?
[20:13:22] <SWPadnos> ok, only 1067 files to download :)
[20:13:30] <SWPadnos> playing, no way. that's work, man!
[20:13:54] <SWPadnos> why do you ask?
[20:14:43] <Unit41> gahhhh emc cant open of the rtlib so files
[20:15:41] <Unit41> is it because there compiled for 32 bit ?
[20:15:43] <JymmmEMC> SWPadnos: I was playing with a AT90S2313, but I noticed they've been replaced by Attiny23 and was wondering the basic diff?
[20:16:08] <SWPadnos> I'm not sure offhand
[20:17:10] <Unit41> oops nm
[20:17:21] <Unit41> I compiled them...
[20:17:35] <jepler> JymmmEMC: the closest new part to at90s2313 is attiny2313. here's atmel's doc on replacing: http://www.atmel.com/dyn/resources/prod_documents/doc4298.pdf
[20:18:52] <JymmmEMC> jepler: you play with them too?
[20:19:19] <jepler> JymmmEMC: no, not much, but I knew where to look for that particular document
[20:19:24] <SWPadnos> heh
[20:19:47] <SWPadnos> I suspect the mega2313 has the extended instruction set
[20:19:50] <JymmmEMC> jepler: TY
[20:20:04] <JymmmEMC> I didn't expect them to be that different
[20:20:10] <SWPadnos> plus the more flexible PWM generator
[20:20:53] <JymmmEMC> What's a USART ?
[20:21:02] <JymmmEMC> (S) ?
[20:21:18] <SWPadnos> universal synchronous/asynchronous receiver/transmitter
[20:21:27] <JymmmEMC> Oh, cool.
[20:21:28] <alex_joni> Universal Synchronous Asynchronous Receiver Transmitter
[20:21:30] <SWPadnos> ie, it can do SPI-like modes as well as "RS232-like" modes
[20:21:32] <alex_joni> so you can select
[20:21:39] <JymmmEMC> very cool.
[20:22:30] <JymmmEMC> Are "fuses" literally that.... fuses. As in when set, they can't be changed ever?
[20:22:48] <alex_joni> in a microcontroller?
[20:22:49] <SWPadnos> hey - there isn't a mega2313 at all :)
[20:22:50] <petev> no
[20:22:52] <alex_joni> usually you can refuse them
[20:22:59] <SWPadnos> no, they're flash in these devices
[20:23:06] <alex_joni> unless you fuse the wrong ones...
[20:23:21] <JymmmEMC> oh, ok. the naming was a bit ambigous, had to ask.
[20:23:26] <petev> but be sure not to set the ones that disable jtag and serial programming
[20:23:31] <SWPadnos> heh
[20:23:35] <JymmmEMC> petev: =)
[20:23:44] <SWPadnos> unless you don't want JTAG
[20:23:56] <SWPadnos> but disabling serial is definitely a bad thing
[20:24:12] <petev> the only other way to program is high voltage parallel, and it's a pain
[20:24:16] <JymmmEMC> In the STK200 kit I have (ancient) it has a ISP dongle that connects to the paraport
[20:24:53] <SWPadnos> spring for either the STK500 ($79), or the ATAVRISP2 ($35)
[20:24:58] <SWPadnos> you'll be happy you did
[20:25:09] <JymmmEMC> SWPadnos: main diff?
[20:25:13] <petev> JymmmEMC, what are you trying to build?
[20:25:30] <petev> the AT90USBKEY is pretty slick if it meets your needs and very cheap
[20:25:36] <JymmmEMC> petev: At this point, a blinking LED =)
[20:25:53] <SWPadnos> the STK500 is a full development board with sockets for most types of AVR, serial or parallel programming vis RS232 (or possibly USB now), and it comes with one or two processors
[20:26:08] <SWPadnos> the ATAVRISP2 is a USB programmer
[20:26:14] <petev> I have several SKT500 and they aren't that nice
[20:26:21] <petev> I much prefer a usb key
[20:26:23] <SWPadnos> either one removes the timing headaches of parport-based programmers
[20:26:52] <SWPadnos> the USB AVRISP2 is more sensitive to the target design than the older serial version
[20:27:07] <SWPadnos> I have boards that I can program with the serial version but not with the USB
[20:27:40] <SWPadnos> (even though you'd think the USB one, with its stable power source, would be better than the serial one that leeches off the target)
[20:28:09] <JymmmEMC> AT90USBKEY looks like it only supprts flatpacks
[20:28:24] <petev> it has an AT90USB part on it
[20:28:29] <petev> with load of flash
[20:28:47] <petev> the IO is a bit limiting as it's fine pitch headers
[20:28:57] <JymmmEMC> Anything I'd build would be DIP
[20:29:03] <petev> but it has user buttons, mouse poitner thing, etc.
[20:29:08] <petev> pretty good for dev work
[20:29:22] <SWPadnos> petev, I think we're talking about programming devices, not necessarily prototyping boards
[20:29:58] <petev> for programming, I like the JTAG, but it's a bit spendy
[20:30:05] <petev> but it's nice for debug
[20:30:11] <SWPadnos> plus you lose several pins from the target
[20:30:15] <petev> the USB serial programmer is ok
[20:30:31] <SWPadnos> I've had overheating problems when I forgot to disable JTAG, and I used the pins in the app
[20:30:33] <petev> but if you have a AT90USB, why not program from USB?
[20:30:42] <petev> that's what I do, and it's very convenient
[20:30:49] <SWPadnos> well, if you have a 2313, for example, you don't have USB ;)
[20:30:51] <petev> it has a built in boot loader
[20:31:05] <petev> but the AT90USB is only a few dollars
[20:31:14] <petev> so for one off experimenting, why not use it?
[20:31:58] <JymmmEMC> petev: I only have one machien in the house that has USB2 =)
[20:32:07] <petev> what?
[20:32:23] <JymmmEMC> the rest are USB1.1
[20:32:32] <petev> JymmmEMC, you still using 8" floppies ;-)
[20:32:51] <petev> the AT90USB is only USB 1.1 anyhow
[20:32:54] <JymmmEMC> petev: Of course not silly.... 5.25"
[20:33:04] <petev> and it doesn't take long to programm a few K of flash
[20:33:09] <JymmmEMC> notchied for DSDD
[20:33:38] <SWPadnos> you aren't pushing it with DSHD?
[20:33:55] <JymmmEMC> Nah
[20:34:07] <SWPadnos> wimp
[20:34:37] <JymmmEMC> * JymmmEMC waves around the Colorado Tape backup drive.... errrrrrrrEEEEEEEEEEEEEEEerrrrrrrWHEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEE
[20:34:46] <SWPadnos> heh
[20:34:52] <SWPadnos> TRAVAN, baby!
[20:35:16] <jepler> interesting, I didn't know about at90usbkey
[20:35:38] <JymmmEMC> * JymmmEMC slaps a 8bit ISA monochorme/paraport card on the table.... BEAT THAT!
[20:36:30] <JymmmEMC> Yes, manual included!
[20:36:37] <jepler> how about an rs232 level converter for the vic 20 expansion port?
[20:36:41] <SWPadnos> * SWPadnos calls with a 512k AST memory expansion card
[20:37:10] <JymmmEMC> jepler: I could fart and fill up the memory on a VIC20
[20:37:18] <SWPadnos> pleas, don't
[20:37:20] <jepler> JymmmEMC: indeed I am confident you could
[20:37:25] <SWPadnos> I'm only one continent away
[20:37:56] <JymmmEMC> It took me like 1.5 days wrote this lil proggy on the vic20 then got a OUT OF MEMORY error, I was so pissed! LOL
[20:39:16] <SWPadnos> and no cassette to save it on :(
[20:39:29] <JymmmEMC> No, I had the tape drive
[20:40:24] <JymmmEMC> But the very first time I ever sat down and wrote a whole program from scratch. I was bummed.
[20:41:34] <JymmmEMC> Any comments on this -vs gecko vampire? http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&ih=005&sspagename=STRK%3AMEWA%3AIT&viewitem=&item=150120823285&rd=1&rd=1
[20:42:21] <anonimasu> go google for some specs first..
[20:42:23] <JymmmEMC> I'm just playing with Atmel is all, even having to learn BASIC right now, till I get half a clue.
[20:42:42] <JymmmEMC> anonimasu: I got the manual for it
[20:43:11] <JymmmEMC> but exp trumps propiganda docs =)
[20:44:57] <JymmmEMC> http://www.parkermotion.com/manuals/oem/OEM_manuals.htm
[20:46:10] <anonimasu> looks fancy
[20:47:00] <JymmmEMC> You can TUNE it for the motors you're using, geckos dont seem to offer that or does it automagically.
[20:47:41] <anonimasu> JymmmEMC: you get what you pay for..
[20:48:06] <anonimasu> they are probably 10 times what a gecko costs..
[20:48:09] <JymmmEMC> I'm assuming geckos do it automagically, can't find anything that mentions that in their docs.
[20:48:18] <anonimasu> they have a tuning potentiometer..
[20:48:31] <JymmmEMC> the parker or geckos?
[20:48:38] <anonimasu> geckos..
[20:48:45] <JymmmEMC> ah
[20:48:47] <anonimasu> I cant remeberthe older ones..
[20:48:56] <anonimasu> but I cant remeber what it sets..
[20:50:32] <anonimasu> the parker drives looks nice
[20:50:57] <JymmmEMC> Parker manual 58 pages, gecko 3 pages. Just bugs me a little.
[20:51:25] <SWPadnos> From the Geckodrive manual: ADJUST: This trimpot adjusts the motor for the smoothest possible low-speed operation. Set the motor speed to about 1/4
[20:51:27] <SWPadnos> revolution per second, and then turn the trimpot until a distinct null is noted in the motor?s vibration. This will result in the most
[20:51:28] <SWPadnos> even microstep placement for a given motor and power supply voltage.
[20:52:33] <JymmmEMC> Are the geckos pulling logic pwoer from the main?
[20:52:53] <SWPadnos> yes, they have an internal regulator
[20:53:03] <JymmmEMC> k
[20:53:18] <anonimasu> I'd buy the parker drives for that price..
[20:53:28] <anonimasu> if you can find 2 of them ;)
[20:53:35] <anonimasu> for like $30 each
[20:53:47] <SWPadnos> the Parkers also have closed-loop operation and /or stall detection
[20:53:57] <JymmmEMC> I have an email into the guy to see if he has three of them, thus my earlier comment about singapore time =)
[20:54:22] <JymmmEMC> I emailed him about customs and he said that singapore and the US have a free trade agreement =)
[20:55:54] <anonimasu> did he have 3?
[20:55:57] <JymmmEMC> 16 jumper selectable motor resolutions (200 - 50,800 steps/rev)
[20:56:17] <JymmmEMC> anonimasu: It's Monday 4am in singapore, I'll let you know in 8 hours =)
[20:56:24] <skunkworks> we ran them at 1000step/rev most of the time
[20:56:59] <skunkworks> iirc - the 650's have 'tuning' pots
[20:57:08] <skunkworks> we never adjusted them
[20:57:20] <JymmmEMC> skunkworks: Are those the ones you were tlaking about that need to have the daughter board jumpered?
[20:58:59] <skunkworks> The oem650x can have the daughter boards removed to make them into oem650 with 5 or so small jumpers. (my old barracuda scsi hard drives had the right size jumpers)
[20:59:25] <JymmmEMC> skunkworks: what does the duaghter baord do?
[21:00:07] <skunkworks> It allows you to do higher level motion. I never played with it.
[21:00:13] <JymmmEMC> ah, ok.
[21:00:16] <skunkworks> We just used them as step and direction
[21:00:44] <JymmmEMC> I have a SCSI drive around here I can steal some jumpers from if needed.
[21:01:12] <JymmmEMC> cw/ccw ?!
[21:01:47] <toastydeath> lo
[21:01:52] <JymmmEMC> Ho
[21:01:52] <skunkworks> ?
[21:02:15] <alex_joni> good night all
[21:02:21] <skunkworks> night alex
[21:02:22] <JymmmEMC> skunkworks: I say mention fo cw/ccw and Rs-232.... thought you could drive these like servos
[21:02:25] <JymmmEMC> night alex_joni
[21:03:09] <skunkworks> Jymmm: like I said - we only used them as step and direction drives.. most likely how your xylex or what ever drives you using work.
[21:03:49] <JymmmEMC> skunkworks: That I understand, just thrying to "absorb" what they're capable of, even if I never use those features.
[21:05:06] <JymmmEMC> If I've learned anything over the years is that paraport has a 2ms (?) timing limitation for software stepping.... does serial have that limitation too?
[21:05:17] <alex_joni> 2 us
[21:05:22] <JymmmEMC> ty
[21:05:28] <alex_joni> serial should be the same
[21:05:34] <alex_joni> they both go over the ISA bus..
[21:05:44] <alex_joni> (even if your modern PC doesn't have one anymore)
[21:06:07] <JymmmEMC> hahahaha, I bought my laptop becasue it has serial and parallel ports
[21:06:34] <JymmmEMC> right next to the 5 USB2 and FW ports
[21:07:24] <JymmmEMC> and built in FDD above the DVDRW drive
[21:07:49] <JymmmEMC> Battery life 4.5 minutes =)
[21:11:55] <JymmmEMC> I like that these drives have limit switch inputs on them
[21:13:06] <skunkworks> JymmmEMC: isn't that if you use the 'x'?
[21:13:21] <JymmmEMC> skunkworks: I believe so, still RTFMing
[21:15:51] <skunkworks> In my opinion it is going to take quite a bit of software to make the x work with emc. For one thing there isn't at the moment a rt serial driver for emc. iirc
[21:16:45] <JymmmEMC> skunkworks: Like you said, I cna pull the dboard and toss in jumpers. But might be fun to play around with.
[21:19:13] <skunkworks> The gear shift input looks interesting. I wonder if you could change the scaling in emc on the fly - say anything over a certain speed - emc would throw the gear shift bit and divide the input scale by 8 (or whatever it would be with your gearing)
[21:20:02] <skunkworks> it would allow nice microstepping down low and higher top speed from the printer port.
[21:23:51] <JymmmEMC> The drive has a software guide?!
[21:24:59] <skunkworks> for the indexer - yes
[21:26:12] <JymmmEMC> you can even set backlash on the drive
[21:26:53] <SWPadnos> this drive has positioning mode(s), which is where all those features are used
[21:27:12] <SWPadnos> I suspect that if you use the step/dir mode, then most of those advanced settings aren't used
[21:27:44] <JymmmEMC> SWPadnos: Yeah, I understand... just foaming at the mouth is all - feature rich.
[21:28:03] <SWPadnos> feature rich for non-coordinated-motion systems ;)
[21:28:58] <JymmmEMC> At the price... I might get 4 (one spare), then play around with a uC for shit and giggles.
[21:29:58] <JymmmEMC> Then I can do this LOL http://www.geckodrive.com/photos/3900IPM.avi
[21:30:52] <skunkworks> SWPadnos: did you see the gearshift input?
[21:30:57] <SWPadnos> nope
[21:31:19] <JymmmEMC> SWPadnos: onboard microstepping multiplier that cna be change on the fly
[21:31:34] <skunkworks> when the bit is set the drives divides its ppr by 8
[21:31:35] <JymmmEMC> but recommended that motor be stopped when doing so
[21:31:53] <skunkworks> I wonder if it would work on the fly if the timing was correct.
[21:32:00] <skunkworks> something else to play with ;)
[21:37:23] <JymmmEMC> I don't know if I'm reading this right or not, but it almost sounds like they're saying that you could dump a program to the drives and it'll execute them directly.
[21:38:41] <JymmmEMC> Description
[21:38:41] <JymmmEMC> L Loops indefinitely
[21:38:41] <JymmmEMC> A1Ø Sets acceleration to 10 revs/sec2
[21:38:41] <JymmmEMC> V5 Sets velocity to 5 revs/sec
[21:38:41] <JymmmEMC> D25ØØØ Sets distance to 25,000 steps
[21:38:41] <JymmmEMC> T2 Waits 2 seconds
[21:38:43] <JymmmEMC> G Executes the move (Go)
[21:38:45] <JymmmEMC> N Ends loop
[21:38:48] <JymmmEMC> Y Stops loop
[21:39:05] <skunkworks> that is what I understand.
[21:39:08] <SWPadnos> yes, that's for something like a packaging machine that's expected to run the same pattern once turned on
[21:40:39] <JymmmEMC> Non volitile?
[21:41:00] <SWPadnos> could be. there may be a "store program" command somewhere
[21:42:37] <JymmmEMC> Set Power-up Sequence Mode
[21:49:58] <anonimasu> hmm
[21:50:15] <anonimasu> JymmmEMC: I hope you arent thinking you will be pushing that many pulses by software even if you scale it..
[21:50:24] <anonimasu> JymmmEMC: remember that resolution suffers too..
[22:03:07] <JymmmEMC> Nah
[22:03:49] <JymmmEMC> Just impressed that it's capable of such things at least hardware stepping
[22:04:20] <JymmmEMC> or internally via program