#emc | Logs for 2006-10-13

[00:02:14] <gkamysz> Works now, thanks
[00:02:32] <gkamysz> bbl
[00:02:54] <jmkasunich> cool
[00:20:16] <jymmmm> jepler: You wouldn't know/remember pascal would you?
[00:25:17] <jmkasunich> cradek: I have absolutely no clue whether those kins are rightg
[00:25:21] <jmkasunich> right even
[00:27:14] <SkunkWorks_> jmkasunich: So far the circuit works. Trying to muddle my way thru the nist-laith setup to see if I can set it up as a spindle to try pwm
[00:27:26] <SkunkWorks_> if that made sense
[00:29:25] <jmkasunich> you want to use it as a spindle drive, and set the speed from within emc using S1000, etc?
[00:30:14] <SkunkWorks_> correct
[00:30:31] <SkunkWorks_> to play until I get some encoders
[00:37:42] <jymmmm> * jymmmm hands SkunkWorks_14 defective mice
[00:38:09] <jymmmm> never said they had to be GOOD ones =)
[00:42:41] <SkunkWorks_> don't tempt me :)
[00:43:46] <SkunkWorks_> so if I am reading the nist hal file right - his laith has a spindle on and a pin for pwm out?
[00:43:57] <SkunkWorks_> pin for spindle on and a pin for pwm out?
[00:44:16] <SkunkWorks_> # spindle enable
[00:44:16] <SkunkWorks_> linksp spindle_on=> parport.0.pin-01-out
[00:44:16] <SkunkWorks_> setp parport.0.pin-01-out-invert 1
[00:44:16] <SkunkWorks_> linkpp freqgen.0.up=> parport.0.pin-14-out
[00:44:35] <jmkasunich> looks like it
[00:44:36] <SWPadnos> yep
[00:44:45] <jmkasunich> I'd recommend using pwmgen instead of freqgen
[00:45:10] <SWPadnos> didn't you make pwmgen after fest?
[00:45:22] <jmkasunich> yeah, I think so
[00:45:27] <jmkasunich> (not sure if its in 2.0.3)
[00:45:28] <SWPadnos> (partly due to this lathe)
[00:45:32] <SWPadnos> right
[00:46:23] <SkunkWorks_> What I need really want is 2 pins - one for m03 and one for m04 that outputs some form of pwm :) one for forward and one for rev
[00:46:42] <SkunkWorks_> I think the nist laith only went one direction
[00:46:48] <SWPadnos> there should be spindle.0.forward-out or some such
[00:47:16] <SWPadnos> I think there are 3 outputs: on, forward, reverse (and brake - four, yes four outputs)
[00:48:10] <jmkasunich> 05 bit OUT FALSE motion.spindle-brake
[00:48:10] <jmkasunich> 05 bit OUT FALSE motion.spindle-forward
[00:48:10] <jmkasunich> 05 bit OUT FALSE motion.spindle-on
[00:48:10] <jmkasunich> 05 bit OUT FALSE motion.spindle-reverse
[00:48:10] <jmkasunich> 05 float IN 0.00000e+00 motion.spindle-revs
[00:48:11] <jmkasunich> 05 float OUT 0.00000e+00 motion.spindle-speed-out
[00:48:15] <jmkasunich> 05 bit OUT FALSE motion.spindle-sync
[00:49:09] <jmkasunich> all you really need is spindle-on and spindle-speed-out
[00:49:37] <jmkasunich> spindle-speed-out is a signed number, goes negative for reverse
[00:49:59] <SWPadnos> unless you want to enable forward/reverse output bridges, plus the PWM
[00:50:06] <jmkasunich> he has an H bridge
[00:50:21] <SkunkWorks_> h-bridge with 2 imputs
[00:50:26] <SWPadnos> right, but the PWM is unsigned ;)
[00:50:27] <SkunkWorks_> inputs
[00:51:14] <SWPadnos> so you could do an and block with PWM output and fwd for one direction, and PWM+reverse for the other direction
[00:51:25] <SWPadnos> but that seems kludgey to nme
[00:51:27] <SWPadnos> me
[00:51:31] <jmkasunich> I think the pwmgen handles signed stuff just fine
[00:51:33] <jmkasunich> checking now
[00:51:42] <SWPadnos> oh - it has up and down outputs?
[00:52:09] <jmkasunich> there are three output types
[00:52:18] <jmkasunich> type 0 = single pwm pin
[00:52:31] <jmkasunich> type 1 = single pwm pin, plus a direction pin
[00:52:44] <jmkasunich> type 2 = two pwm pins, up and down
[00:52:47] <SkunkWorks_> That is cool
[00:53:00] <jmkasunich> for type 2, one will pwm and the other remains off
[00:53:06] <SkunkWorks_> I might have to get head then- I don't think it is in the rel version
[00:53:09] <jmkasunich> which one pwm's depends on the sign
[00:53:15] <SWPadnos> are they like the up/down PWM that was discussed way back, or are they one output for positive sign, one for negative sign?
[00:53:19] <SWPadnos> ok - that's good
[00:53:41] <jmkasunich> type 2 can connect directly to the two sides of the H bridge
[00:54:13] <SWPadnos> kewl
[00:54:18] <SkunkWorks_> crap - I don't think I am going to have enough time to play with this tonight.. Damn
[00:54:29] <jmkasunich> I really need to make a man page for that component
[00:54:50] <SWPadnos> heh
[00:55:56] <SkunkWorks_> that sounds like exactly what I would want. doesn't it also do pure pwm? (I thought there was a discussion that for a computer the freqgen psudo pwm was better)
[00:56:14] <jmkasunich> it can do several flavors of pwm
[00:58:42] <jmkasunich> there are two params that determine the kind of PWM
[00:59:26] <jmkasunich> pwm-freq can be set to zero for pulse density modulation
[00:59:55] <SkunkWorks_> * SkunkWorks_ thinks jmk is reading his comments in the pwmgen source :)
[01:00:07] <jmkasunich> that gives you the lowest possible ripple current, but can result in a switching frequency as high as the thread rate divided by two
[01:00:12] <jmkasunich> skunkworks is right
[01:00:49] <jmkasunich> if pwm-freq is non-zero, it sets the maximum switching frequency (handy if you need to limit the frequency to control switching loss)
[01:01:24] <jmkasunich> in that case, the parameter dither-pwm comes into play
[01:01:45] <jmkasunich> if it is true, then the actual frequency is allowed to jitter around, to give a more accurate average output voltage
[01:02:12] <jmkasunich> if false, the duty cycle and frequency are fixed (and the resolution is usually pretty poor)
[01:02:42] <jmkasunich> for example, with a 20KHz thread and a 2KHz pwm-freq, there are only 10 possible output duty cycles
[01:03:05] <jmkasunich> if you ask for 16%, you are out of luck, you'll get 20%
[01:03:32] <jmkasunich> with dither turned on, it will switch back and forth between 20% and 10% to give an average of 16%
[01:04:22] <jmkasunich> I would recommend pulse density modulation
[01:04:43] <SkunkWorks_> does anyone know for sure if it is in 2.0.3? Could I run a command to check (I have emc running from terminal)
[01:05:09] <jmkasunich> sure
[01:05:38] <jmkasunich> halcmd loadrt pwmgen output_type=2
[01:07:01] <SkunkWorks_> :( HAL:0: ERROR: Can't find module 'pwmgen' in
[01:07:13] <jmkasunich> guess its not there then
[01:07:27] <SkunkWorks_> well I have a computer at home that I will put head on and use that to play.
[01:07:31] <jmkasunich> you can use freqgen, it can do pulse density modulation
[01:07:41] <SkunkWorks_> right
[01:07:44] <jmkasunich> (it won't do pure pwm or dithered, but thats ok)
[01:15:33] <gkamysz> How do I know how fast the computer can output steps for a stepper setup? TCNC told you what that was upon booting.
[01:17:34] <jmkasunich> its determined by the BASE_PERIOD setting in the ini file
[01:17:51] <jmkasunich> you have to experiment a little to see how small of a base period your computer can handle
[01:19:22] <jymmmm> There is a test available
[01:19:31] <jymmmm> rtai
[01:20:28] <jymmmm> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?action=browse&diff=1&id=TroubleShooting
[01:21:25] <jymmmm> anything near or below 20,000 ns lat_max is good (iirc)
[01:32:41] <SkunkWorks_> my cutting and pasting has gotten the best of me
[01:32:43] <SkunkWorks_> HAL: ERROR: pin 'motion.spindle-speed-out' not found
[01:33:03] <jmkasunich> bin/halcmd show pin motion.spindle
[01:33:09] <jmkasunich> bin/halcmd show pin iocontrol.spindle
[01:33:15] <jmkasunich> one of those will show you the pins
[01:33:35] <jmkasunich> (they used to be in iocontrol, and got moved to motion, maybe 2.0.3 still has them in iocontrol)
[01:34:05] <jmkasunich> oh, if you are running an installed copy (not run-in-place) drop the /bin/ and just say halcmd
[01:34:18] <SkunkWorks_> ah - I think I may just have to call it a night. Get head on the other computer to play with.
[01:37:11] <SkunkWorks_> ace@ubuntu:~$ halcmd show pin motion.spindle
[01:37:11] <SkunkWorks_> Component Pins:
[01:37:11] <SkunkWorks_> Owner Type Dir Value Name
[01:37:11] <SkunkWorks_> 03 float R- 0.00000e+00 motion.spindle-revs
[01:37:12] <SkunkWorks_> 03 bit -W FALSE motion.spindle-sync
[01:37:29] <jmkasunich> those are for threading
[01:37:50] <jmkasunich> if they are the only ones then the others are in iocontrol
[01:38:04] <jmkasunich> show pin iocontro.spindle (probably)
[01:38:15] <jmkasunich> if that doesn't work, maybe I'm mis-spelling something
[01:38:29] <jmkasunich> just do show pin and it will give you the entire list, then browse it
[01:52:01] <CIA-8> 03jepler 07HEAD * 10emc2/src/ (Makefile.inc.in configure configure.in):
[01:52:01] <CIA-8> if the mandb program is available, build index.db which allows 'apropos' to
[01:52:01] <CIA-8> work on manpages in the RIP directory
[01:52:02] <CIA-8> 03jepler 07HEAD * 10emc2/docs/src/Submakefile:
[01:52:02] <CIA-8> if the mandb program is available, build index.db which allows 'apropos' to
[01:52:02] <CIA-8> work on manpages in the RIP directory
[01:54:32] <jmkasunich> stray cats?
[01:55:25] <jepler> jmkasunich: your guess is as good as mine
[01:55:37] <jepler> a "cat" is the preformatted version of a "man" file, because you can just "cat" it
[01:55:41] <jepler> a stray cat would be one without a man?
[01:58:34] <jymmmm> jepler: You wouldn't know/remember pascal would you?
[01:59:01] <jepler> jymmmm: nope, not really
[01:59:51] <jymmmm> jepler: Ok. I'm trying to convert an OLD pascal program over to php. lots of trig in it.
[02:00:10] <SkunkWorks_> jmkasunich: it works for one direction :)
[02:00:37] <jmkasunich> cool
[02:00:37] <jepler> jymmmm: that's sure not my idea of fun
[02:01:09] <jymmmm> jepler: Necessity =)
[02:01:14] <SkunkWorks_> Now I am done for the night :)
[02:01:43] <jymmmm> jepler: it does lat/lon conversation and Great Circle calculations.
[02:01:50] <jepler> jymmmm: I don't know how standard pascal is, but maybe you can use the pascal directly with a modern compiler? like this one, which I have never heard of but found after a bit of searching: http://freshmeat.net/projects/fpc/
[02:01:51] <jymmmm> jepler: beats doing it my hand =)
[02:01:52] <CIA-8> 03jmkasunich 07HEAD * 10emc2/src/hal/drivers/hal_ppmc.c: fix some warnings
[02:02:21] <SkunkWorks_> m03
[02:02:29] <jmkasunich> M04
[02:02:31] <SkunkWorks_> oops
[02:02:32] <SkunkWorks_> :)
[02:03:23] <SkunkWorks_> side question - how would you set it so that you never got to 100%dc?
[02:03:50] <SkunkWorks_> s600
[02:03:53] <SkunkWorks_> oops
[02:08:07] <jepler> SkunkWorks_: by setting 'maxfreq' to a value that's dependent on the actual base period you get
[02:08:39] <jepler> for instance if your period corresponds to 1kHz and you set maxfreq = 950, you'd get at most 95% duty cycle .. I think
[02:09:15] <SkunkWorks_> makes sense
[02:09:56] <SkunkWorks_> the scope signal looks pretty cool :)
[02:11:12] <SkunkWorks_> I get to s550 and it is all but 99% or so s560 takes it to 100% :)
[02:11:34] <SkunkWorks_> * SkunkWorks_ is just playing :)
[02:11:57] <SkunkWorks_> ok - now I am done :)
[02:13:57] <SkunkWorks_> cool
[02:17:05] <jmkasunich> dang - missed him
[02:17:21] <jmkasunich> the pwmgen component has max and min duty cycle limts (parameters)
[02:19:08] <jepler> * jepler is playing with pwmgen+siggen+halscope now
[02:21:07] <cradek> jmkasunich: thanks for helping gkamysz
[02:21:44] <jmkasunich> cradek: np
[02:41:31] <CIA-8> 03jepler 07HEAD * 10emc2/docs/man/man3/hal_param_new.3hal: markup fix
[02:42:03] <CIA-8> 03jepler 07HEAD * 10emc2/docs/man/man3/ (32 files): make all the hal_p...new functions have entries
[02:44:20] <jepler> hm that didn't work like I expected...
[02:45:06] <jmkasunich> what didn't?
[02:45:29] <jepler> the use of .so to make one manpage include another
[02:46:00] <jepler> I guess I have to do it in a way that doesn't work for 'man ./whatever.3' but does for 'man 3 whatever'
[02:48:37] <A-L-P-H-A> jepler, or jmkasunich, which one of you changed the to twist lock belts?
[02:48:41] <jmkasunich> me
[02:48:54] <A-L-P-H-A> jmkasunich: I just did it today... much quieter.
[02:49:15] <CIA-8> 03jepler 07HEAD * 10emc2/docs/man/man3/ (36 files): fix up .so includes and add more of them
[02:50:01] <A-L-P-H-A> jmkasunich: know of any other upgrades to make it better? [besides a whole new machine]
[02:50:12] <jmkasunich> what is "it"?
[02:51:14] <A-L-P-H-A> jmkasunich: well... I would assume it being the topic we were just talking about, or that I enquired about. For the clear record, powertwist plus t-locking v-belts from FennerIndustrial.com
[02:51:36] <A-L-P-H-A> :)
[02:51:46] <A-L-P-H-A> I was trying to be a smartass, not snippy.
[02:51:56] <A-L-P-H-A> I think it was more snippy than smartass.
[02:52:02] <jmkasunich> I thought you were talking in general terms.... you know "now that I've replaced the belts, what else can I do to make my _machine_ better"
[02:52:07] <jmkasunich> "it" was the machine
[02:52:26] <jmkasunich> is it a shoptask like mine, a mini-mill, a lathe, a table saw, etc
[02:52:32] <A-L-P-H-A> I think I was more dumbass than smart at all then.
[02:52:37] <A-L-P-H-A> hahaha.
[02:53:06] <A-L-P-H-A> I upgrade the belts on my 6x4 bandsaw. And my chinese mill that's 28x7x3.5.
[02:53:26] <A-L-P-H-A> it's a King Machinery PD-25. I believe... met go read the model again.
[02:53:26] <jmkasunich> I haven't thought about the banksaw, I bet it would work nicer with those belts
[02:53:41] <A-L-P-H-A> runs much quieter.
[02:53:43] <jmkasunich> (I have a harbor fright one)
[02:54:41] <A-L-P-H-A> http://www.bellmachineryltd.com/new_kingpdm25.htm
[02:55:02] <A-L-P-H-A> exact machine, except mine's this pukey green.
[02:56:14] <skunkworks> jmkasunich: thanks agian for the help. I am really excited about the posibilities. :)
[02:56:26] <A-L-P-H-A> skunkworks: how's the h-bridge going?
[02:56:49] <jepler> 'night guys
[02:56:53] <A-L-P-H-A> bbiab... making tea
[02:56:56] <skunkworks> A-L-P-H-A: I ran it as a spindle.
[02:57:34] <skunkworks> jepler: - thanks also for the explaination of setting up the max duty cycle
[02:59:15] <skunkworks> used freqgen to output a pwm signal in relation to S number.
[02:59:44] <skunkworks> (pretty much just coppied what was in cradeks nist laith hal) :)
[03:00:47] <skunkworks> night again :)
[03:16:04] <A-L-P-H-A> * A-L-P-H-A pkes K`zan.
[03:16:06] <A-L-P-H-A> pokes
[03:20:51] <jmkasunich> * jmkasunich goes to sleep
[03:27:32] <A-L-P-H-A> I'm going to sleep too.
[03:27:37] <A-L-P-H-A> night every
[05:37:28] <LawrenceG> greetings all... who is working on the fet bridge?
[05:54:20] <Jymmm> is this a trick question?
[05:54:24] <Jymmm> * Jymmm chuckles *
[05:59:00] <LawrenceG> hey Jymmm
[05:59:25] <Jymmm> LawrenceG =)
[05:59:59] <LawrenceG> was just wondering how muchsmoke had leaked out so far :}.... I would like to build a 3 1/2 bridge version for some 3 phase servos motors
[06:00:24] <Jymmm> I haven't been keeping up lately, so no clue here
[06:00:26] <Jymmm> there
[06:00:45] <LawrenceG> it was jmk and ????
[06:00:50] <Jymmm> fenn?
[06:01:30] <LawrenceG> not sure... will have to ask when more people are awake....
[06:04:16] <LawrenceG> speaking of sleep.... I am off to bed.... cheers
[06:05:58] <Jymmm> G'Night LawrenceG
[08:18:32] <Bo^Dick> the HIP4081A mosfet driver has a voltage rating of 80 volts. so does the cheapest geckodrive, is this a coincidence?
[09:16:39] <Bo^Dick> has anyone of you guys ever looked inside a commersial high-performance stepper driver?
[09:19:25] <Bo^Dick> a schematic of such a product would be more than lovely
[11:52:20] <anonimasu> Bo^Dick: I have commercial stepper drivers..
[11:52:43] <anonimasu> Bo^Dick: the ones "www.aratron.se" sells are less advanced then the grckos..
[11:52:48] <anonimasu> geckos..
[12:02:25] <Bo^Dick> cool
[12:03:06] <anonimasu> I hate steppers.
[12:03:11] <Bo^Dick> three-phase stepper drivers
[12:03:14] <Bo^Dick> thats cool
[12:03:36] <anonimasu> :)
[12:03:47] <anonimasu> though you need optocouplers to use them with a pc...
[12:04:00] <Bo^Dick> why
[12:04:40] <anonimasu> well, because they have the inputs floating.. or something like that
[12:05:40] <Bo^Dick> a buffer could do the trick
[12:06:06] <anonimasu> I think they were pretty cheap.
[12:06:16] <anonimasu> though not as cheap as the geckos..
[12:06:58] <anonimasu> :D
[12:07:34] <Bo^Dick> so what did they cost then?
[12:07:58] <anonimasu> call them and ask..
[12:08:10] <anonimasu> it's 2 years since I bought them..
[12:08:17] <Bo^Dick> ok
[12:08:46] <anonimasu> I buy belts/pulleys from them..
[12:08:54] <anonimasu> err toothed belt/gears..
[12:12:35] <Bo^Dick> to me it would be pretty acctractive information to see at least what parts successful commersial stepper drivers use
[12:15:41] <anonimasu> well order a gecko..
[12:17:01] <Bo^Dick> thats a splendid idea
[12:17:36] <Bo^Dick> i could mention that i've seen a screecshot of a gecko on the web and forgot to save that image
[12:17:55] <Bo^Dick> so now i'm recommended to buy a gecko for that purpose
[12:18:11] <Bo^Dick> do i regret that i didn't save that image or what
[12:22:16] <anonimasu> Bo^Dick: seriously this is getting very tiring.
[12:23:25] <anonimasu> Bo^Dick: if you have enough time and money to sit on irc and whine about how much stuff costs you've got way too much money.
[12:25:42] <Bo^Dick> i'm an opportunist. i know at least someone has opened a gecko and perhaps taken a screenshot. i know it was no hallucination.
[12:26:05] <anonimasu> Bo^Dick: No, you are a cheap bastard.
[12:26:18] <Bo^Dick> my obtain isn't just to make a stepper driver is that was what you thought
[12:26:42] <anonimasu> are you going to sell stepper drivers cheaper then geckos?
[12:26:44] <Bo^Dick> this whole project wouldnt have been the least economic
[12:27:08] <Bo^Dick> do you seriously think i was going to make another gecko-replica?
[12:27:22] <anonimasu> I have no idea since all you do is whine about the price of the,
[12:27:24] <anonimasu> them..
[12:27:38] <Bo^Dick> i'm not saying they're too expensive
[12:27:57] <Bo^Dick> i'm learning advanced electronics thats all
[12:28:12] <anonimasu> ok ^_^
[12:28:23] <skunkworks> do I have to separate you two?
[12:28:50] <anonimasu> nah..
[12:28:53] <Bo^Dick> i'm maybe gonna publish my design when it gets fisished as an alternative to the gecko (DIY alternative)
[12:29:20] <anonimasu> Bo^Dick: http://www.altera.com/literature/hb/cyc2/cyc2_cii51003.pdf
[12:29:21] <Bo^Dick> the DIY alternative are currently very substandard and i'm trying to get that over with
[12:29:22] <anonimasu> err
[12:29:23] <anonimasu> wrog place..
[12:29:29] <anonimasu> http://flemingcnc.com/cnc/index.html
[12:29:31] <anonimasu> look at the bottom..
[12:29:46] <anonimasu> driving stepper motors with quadratic equations..
[12:29:48] <anonimasu> ^_^
[12:31:00] <Bo^Dick> cool
[12:31:14] <anonimasu> though it might not be of too much use
[12:32:30] <Bo^Dick> i think that is the math for the pulse generation not the controller for the phase drivinng
[12:32:35] <Bo^Dick> am i right?
[12:34:22] <anonimasu> hm way too stupid to understand that stuff
[12:34:41] <anonimasu> yeah.. it is :)
[12:35:35] <Bo^Dick> i've got a hexfet driver called HIP4081A and it's rated 80V just like the geckodrive. i was wondering if this is a pure coincidence or not
[12:37:18] <anonimasu> hm
[12:37:20] <anonimasu> have a look
[12:37:20] <anonimasu> http://www.toshiba-components.com/applications/HomeAppliances/asspmotorcont/StepperMotorController/TA84002F.html
[12:38:13] <Bo^Dick> i think allegro still outperforms it
[12:39:35] <anonimasu> did you see http://www.cnczone.com/forums/showthread.php?p=202868
[12:40:50] <anonimasu> awell
[12:40:50] <anonimasu> l
[12:40:52] <anonimasu> laters..
[12:40:55] <anonimasu> need to get some work done
[14:07:03] <CIA-8> 03jepler 07HEAD * 10emc2/docs/man/man3/ (11 files): new manpages
[14:08:25] <SWPadnos> Bo^Dick, the 80V gecko rating is pure coincidence
[14:08:38] <SWPadnos> they actually use FETs that are rated to ~115v
[14:08:50] <SWPadnos> and under-rate their drives
[14:09:17] <skunkworks> SWPadnos: I got the h-bridge to run using spindle speed. :) cool.
[14:09:24] <SWPadnos> cool stuff
[14:10:06] <skunkworks> emc seems kinda powerfull - Don't you think? :)
[14:10:10] <SWPadnos> Bo^Dick, if you want to see photos of Geckos, join the geckodrive yahoo group - there are photos there
[14:10:23] <SWPadnos> skunkworks, err - yeah, I think so ;)
[14:25:37] <A-L-P-H-A> Hi folks.
[14:27:11] <skunkworks> A-L-P-H-A: did you get 'it' working? :)
[14:28:03] <A-L-P-H-A> it was always working.
[14:28:14] <A-L-P-H-A> your sister can contest to that! :D
[14:28:52] <A-L-P-H-A> hahaha... just saw a baby elephant walking through a trail... and it stumbles... because it ran into a small tree. hahahaha
[14:29:02] <A-L-P-H-A> it didn't hurt itself... it just scared itself.
[14:29:08] <A-L-P-H-A> documentaries are funny.
[14:35:15] <CIA-8> 03jepler 07HEAD * 10emc2/docs/src/Submakefile:
[14:35:15] <CIA-8> there are enough manpages to warrant splitting them by section in the html
[14:35:15] <CIA-8> index
[14:45:52] <Bo^Dick> feel free to take part of a sophisticated discussion regarding a DIY project for a stepper driver! http://www.carmi.se/misterstarshine/Projects/stepper_design.htm
[14:49:33] <CIA-8> 03jepler 07HEAD * 10emc2/docs/src/Submakefile: because of included files, the directory where groff executes is important
[14:56:06] <jepler> sigh
[14:56:25] <jepler> now I have to get dependency generation correct for the manpages
[14:57:00] <jepler> otherwise, foo.3.html won't be updated if foo.3 does .so man3/bar.3 and bar.3 is revised
[14:57:30] <SWPadnos> I have no idea what you just saaid
[14:57:33] <SWPadnos> said
[14:57:38] <cradek> I was just thinking that
[14:57:54] <cradek> what's .so?
[14:58:08] <SWPadnos> maybe that should have been "so"
[14:58:17] <cradek> no, it still doesn't parse
[14:58:24] <SWPadnos> it's closer though ;)
[14:58:31] <cradek> maybe .so a nroff command
[14:58:33] <cradek> is
[14:58:39] <cradek> ?REDO FROM START
[14:58:50] <SWPadnos> yeah: -ENOCAFFEINE
[14:59:46] <jepler> ".so" is a roff directive to source (include) another file
[15:00:00] <SWPadnos> ah - that makes more sense then
[15:00:13] <cradek> gcc -MM foo.3
[15:00:28] <jepler> for instance, all but one of the hal_pin_new family of functions just include the page that actually describes all of them
[15:00:59] <cradek> and I bet the included name is on the line after the .so to make it harder
[15:07:14] <cradek> ok no it's not, that's good
[15:08:16] <jepler> yeah -- I'm almost done, assuming that there are no indirect .so
[15:08:59] <cradek> yeah indirect is much more complicated
[15:11:12] <skunkworks> is that how normal linux works - the man pages are built by taking the info from the source files?
[15:12:29] <jepler> skunkworks: no, that's a misunderstanding of what I said
[15:13:03] <jepler> skunkworks: when I said "source code", I meant the manpages themselves which are "source" in the sense that when it's not formatted by a browser, HTML is "source code"
[15:13:18] <skunkworks> ah
[15:13:32] <skunkworks> totally missed that one :) what else is new
[15:13:35] <jepler> there are systems for automatically making documentation from sourcecode (javadoc and its clones being one well-known example) but that's not what I meant
[15:14:42] <CIA-8> 03jepler 07HEAD * 10emc2/docs/man/man3/hal_get_lock.3hal: don't source yourself
[15:15:00] <CIA-8> 03jepler 07HEAD * 10emc2/docs/man/man3/rtapi_clock_set_period.3rtapi: note that set_period can return a failure code
[15:15:32] <Lerneaen_Hydra> 'l
[15:15:34] <Lerneaen_Hydra> 'lo
[15:15:51] <jepler> hi Lerneaen_Hydra
[15:15:57] <CIA-8> 03jepler 07HEAD * 10emc2/docs/src/Submakefile:
[15:15:57] <CIA-8> limited dependency generation for manpage source, so that the html pages
[15:15:57] <CIA-8> are rebuilt at the right times
[15:15:57] <CIA-8> quiet some commands
[15:15:57] <CIA-8> use 'mv -f' instead of 'mv'
[15:18:38] <jepler> yuck, look at the arguments section here: http://linuxcnc.org/docs/html/man/man3/hal_param_s16_new.3hal.html
[15:18:50] <jepler> at some moments I wonder if the .txt version would be better
[15:19:21] <cradek> it works, but it's not very pretty
[15:21:23] <SWPadnos> it's making a table with varying widths for each argument
[15:21:24] <SWPadnos> weird
[15:21:24] <cradek> I don't understand why it's so bad since all the tr are align left
[15:21:43] <cradek> I think there are four columns, whitespace - argument - whitespace - description
[15:21:45] <CIA-8> 03jepler 07HEAD * 10emc2/docs/man/man3/hal_signal_new.3hal: hal_link is on this page
[15:22:00] <skunkworks> another odd question - could the feedback from an ecoder be converted to rpm (or whatever) so that it could be used as feedback for a spindle. (to keep the spindle at what ever speed you set it to)
[15:22:03] <cradek> they are 11,5,4,77%
[15:22:18] <skunkworks> within emc
[15:22:27] <cradek> (which doesn't add up to 100)
[15:22:52] <cradek> skunkworks: yes you would just use a pid loop for the spindle
[15:23:17] <SWPadnos> skunkworks, yes, use a ddt to get the position difference (though I'm not sure what happens when the counter overflows)
[15:23:22] <jepler> there's a little wrinkle with the index pulse as well
[15:23:47] <cradek> hmm are we all answering different questions?
[15:24:16] <jepler> ... one of the men felt the elephant's trunk, and he said "it is like a snake" ...
[15:24:23] <SWPadnos> heh
[15:24:47] <SWPadnos> base question: can you get a velocity from an encoder
[15:24:58] <SWPadnos> second tier meaning: can you control spindle RPM with that
[15:25:12] <SWPadnos> third tier: what problems may occur with the current software set up in this manner
[15:25:20] <SWPadnos> yes, 3 people, 3 answers ;)
[15:25:31] <jepler> clearly you use PID to take commanded speed and actual speed and get a commanded PWM duty cycle
[15:25:41] <jepler> so the problem is in getting the actual speed from the encoder
[15:26:09] <SWPadnos> doesn't the encoder component have a delta/speed output pin?
[15:26:23] <cradek> jepler: I finally figured out it starts a new table with different width percentages
[15:26:35] <SWPadnos> (which shuold be calculated before the count is reset, to eliminate index issues)
[15:26:36] <cradek> jepler: do you know what causes the new table?
[15:26:44] <jepler> 32769 s32 OUT 0 encoder.0.counts
[15:26:47] <jepler> 32769 float OUT 0.00000e+00 encoder.0.position
[15:26:57] <jepler> cradek: that's what SWPadnos said earlier: 10:22:21 <SWPadnos> it's making a table with varying widths for each argument
[15:26:57] <SWPadnos> hmm
[15:27:00] <jepler> no, I don't know why
[15:27:22] <cradek> that's not quite true - name and dir are in the same table
[15:28:24] <jepler> hm maybe I can't have newlines between those
[15:28:33] <jepler> I am still far from an expert on the man language
[15:28:38] <skunkworks> Interesting. So it may be possible :)
[15:29:06] <cradek> skunkworks: is this for a mill?
[15:30:05] <skunkworks> I was thinking for a laith. for now. The k&t has an older vfd for the spindle.
[15:30:37] <skunkworks> I was thinking having it switchable between threading and normal spindle using a servo.
[15:31:02] <skunkworks> but it would be nice to have speed feedback.
[15:31:10] <cradek> sounds cool
[15:31:26] <cradek> sounds challenging
[15:31:53] <cradek> doing one or the other would not be too bad, but switching in and out of threading mode is more challenging
[15:32:07] <skunkworks> I am thinking big picture without really knowing if it would work :)
[15:32:17] <SWPadnos> the encoder component should be updated to have a speed output, since the speed should be calculated before any index-based position reset
[15:32:46] <skunkworks> I thought someone here was doing that already.
[15:33:06] <skunkworks> no that was switching between an axis and not on the spindle
[15:33:12] <skunkworks> now that I think about it.
[15:33:26] <SWPadnos> you can use a ddt component after the encoder, but then you'll either have index issues or rollover issues
[15:33:27] <cradek> that sounds easy to add to the encoder block
[15:33:47] <SWPadnos> * SWPadnos wishes he had a computer to develop on
[15:34:07] <skunkworks> so actually I would want 3 modes - spindle sync mode (threading), axis mode, and rpm mode. if that makes sense
[15:34:26] <SWPadnos> RPM can be there with either of the other modes
[15:34:28] <cradek> wonder if you could build emc-sim under cygwin/mingw
[15:34:37] <cradek> or you chould just get the nice free vmware
[15:34:41] <SWPadnos> I think it would be faster to order a new computer ;)
[15:34:46] <cradek> could + should = chould
[15:34:52] <SWPadnos> heh
[15:36:47] <jepler> speed is just (counts - old_counts) / delta_time, as long as enough counts pass each delta_time that it's not too granular a measure, and index isn't enabled
[15:37:08] <jepler> you could write in about 10 lines of 'comp' the thing that computes that, and outputs speed as a float suitable for sending to PID
[15:37:14] <SWPadnos> it's the same with index, as long as the calculation is before the index checking code
[15:37:29] <SWPadnos> hence the reason for putting it into the encoder compoonent
[15:37:32] <jepler> you'd use .counts rather than .position so you don't start to lose precision when you've done hundreds of thousands of revolutions
[15:38:04] <SWPadnos> counts gets reset with index, doesn't it?
[15:38:10] <jepler> both do
[15:38:17] <jepler> I think, anyway
[15:38:25] <SWPadnos> ok, so it really should be in the encoder component
[15:38:30] <jepler> I'm all for making the encoder do it
[15:38:35] <cradek> me too
[15:38:58] <SWPadnos> is there an encoder written in comp now?
[15:39:01] <skunkworks> I am just glad I can give you guys more thing to think about :)
[15:39:02] <cradek> with its own function so you can run it in a slower or faster thread
[15:39:20] <jepler> well, I may or may not be abusing .IP in the manpages (their example is for bullets, not the names of function parameters) but I don't get better results by using .TP
[15:40:29] <danfalck> cradek: in truetype-tracer, are there any options for changing the size of the font? I didn't see info in the README. thanks
[15:40:57] <jepler> danfalck: in the g-code ouput there is a scale factor you can change
[15:40:58] <cradek> you can scale in the gcode after
[15:41:11] <danfalck> ok thanks
[15:41:12] <cradek> I think by default it's about an inch tall
[15:41:37] <danfalck> in dxf out you would just use a CAD program to scale then
[15:41:53] <cradek> yes I guess so
[15:41:57] <danfalck> ok
[15:42:05] <cradek> the dxf was sort of an afterthought and I assumed it would go right into a cad program
[15:42:09] <jepler> the #1 .. #4 control stuff like the feed rate, cut depth, and scale
[15:42:28] <jepler> G00 X [.5+1536.00*#3] Y [2556.00*#3]
[15:42:30] <cradek> see my webpage where it talks about only doing 80% of each project :-)
[15:42:33] <jepler> all the distances are scaled by #3
[15:42:46] <danfalck> thanks
[15:43:56] <jepler> well, the XY distances are. The Z distances aren't
[15:44:22] <cradek> isn't this wrong? [.5+1536.00*#3]
[15:44:36] <cradek> or does emc do the order of operation incorrectly?
[15:44:52] <jepler> cradek: that's not actually TTT output, I massaged it to get the axis splash screen
[15:44:58] <jepler> I wanted it all translated over .5 inch post-scaling
[15:45:00] <cradek> oh ok
[15:46:46] <jepler> so that it wouldn't start under the tip of the cone
[16:17:40] <skunkworks> Eat it while listening to Garrison Keillor talk about the Lives of the Cowboys.
[16:17:42] <skunkworks> :)
[16:18:18] <skunkworks> I like the skits brought to you by the catsup advisory board.
[16:20:06] <skunkworks> can anyone see a problem with the low side fets being on all the time when no pwm signal is present?
[16:20:23] <skunkworks> I can't think of any off the top of my head.
[16:20:52] <jepler> it will act as a brake, right?
[16:20:56] <skunkworks> right
[16:21:08] <skunkworks> a little more so than the free wheeling diodes
[16:22:31] <skunkworks> it effectivly shorts out the motor
[16:24:13] <skunkworks> My thoughts is that the servos should not be allowed to coast if everything is set up correctly. only thing that should do it is an estop.
[16:24:47] <SWPadnos> actually, estop should apply any available brake
[16:25:28] <SWPadnos> err - I may have misunderstood what you meant by "it" there ;)
[16:25:59] <skunkworks> I am thinking an estop will shut the pwm to the drive - which will short the servo out.. I should also be cutting the servo power also I suppose :)
[16:26:46] <SWPadnos> power should be cut, but only if the machine is designed such that losing power will cause it to stop as rapidly as possible (ie, a brake is engaged while de-powered)
[16:27:03] <SWPadnos> if e-stop is really e-coast, you have a problem
[16:27:57] <skunkworks> as of right now - an e-stop will turn on the 2 low fets - which should brake the motor quite well
[16:28:17] <SWPadnos> yep - that's probably desirable
[16:28:49] <skunkworks> I was just wondering if that mode would cause other problems with emc as the servo loop.
[16:29:16] <SWPadnos> I haven't thought of any, but then again I haven't thought about it ;)
[16:29:18] <skunkworks> Just worrying a bit - no big deal. :)
[16:33:37] <skunkworks> * skunkworks wonders off to finally order some decent copper clad
[16:56:03] <anonimasu> :)
[16:58:49] <Jymmm> * Jymmm lol @ SWPadnos
[16:58:58] <SWPadnos> what fer?
[16:59:06] <Jymmm> if e-stop is really e-coast, you have a problem
[16:59:10] <SWPadnos> heh
[16:59:25] <Jymmm> I read that as "East Coast" not "Emergency Coast"
[16:59:43] <SWPadnos> I thought of that when re-reading it
[16:59:47] <Jymmm> =)
[17:00:22] <Jymmm> I was thinking... damn if the machien is THAT big that you refernce it's location by costlines...
[17:00:30] <Jymmm> coastlines
[17:01:04] <SWPadnos> "please step into New York to home the X axis"
[17:01:26] <Jymmm> and In LA to enable flood coolant
[17:01:34] <Jymmm> <rimshot>
[17:01:37] <SWPadnos> heh
[17:02:03] <Jymmm> so, what would gator country be?
[17:02:15] <Jymmm> oh clamping!
[17:02:28] <Jymmm> CHOMP!
[17:02:45] <SWPadnos> yah
[17:03:11] <Jymmm> Ok, Texas?
[17:04:05] <SWPadnos> overheating
[17:04:24] <skunkworks> thermal comp
[17:04:59] <Jymmm> I thought that was AZ
[17:05:06] <Jymmm> or NV
[17:30:48] <anonimasu> ^_^
[17:39:59] <skunkworks> I found some 2oz double sided copper clad for cheap on ebay.
[18:25:19] <ejholmgren> god hath smiled upon me
[18:25:44] <ejholmgren> one of the techs just threw out a pair of 180oz step-syns
[18:27:21] <jepler> nice find
[18:28:00] <ejholmgren> do steppers wear out?
[18:28:50] <ejholmgren> should have asked why he took them out of the printer
[18:32:43] <SWPadnos> I think that for the most part, steppers don't wear out
[18:32:51] <Lerneaen_Hydra> not really, unless they're burnt out or something
[18:33:13] <SWPadnos> the bearings can wear, like any motor, and if they're subject to overcurrent, then they ,ay get demagnetized, which will cause a power loss
[18:33:19] <SWPadnos> s/,ay/may/
[18:33:53] <SWPadnos> and of course, it's possible for a winding to short out or whatever, but there aren'y anu brushes or commutator rings to wear out
[18:36:18] <SWPadnos> I typo very well sometimes ;)
[18:40:30] <Lerneaen_Hydra> inded yuo od. ;)
[18:54:21] <SWPadnos> hay - im knot htat badd
[18:57:01] <ejholmgren> so ... a bipolar driver can drive bi and uni motors, and a unipolar driver can only drive uni?
[18:57:01] <Jymmm> SWPadnos: ewe arrrrrrs theee!
[18:57:08] <ejholmgren> asdf!
[18:57:29] <SWPadnos> hey - it's not talk like apirate day any more. wait for next year
[18:57:36] <SWPadnos> ... a pirate ...
[19:05:26] <skunkworks> I heard emc cures planters warts also... Is that true?
[19:08:04] <daniel_br> Hello from Brazil
[19:08:25] <skunkworks> Hello
[19:09:15] <ejholmgren> planning on milling an extremity?
[19:09:53] <daniel_br> I have a question about M48 and M49 gcode, someone can help me?
[19:11:11] <daniel_br> http://www.linuxcnc.org/docs/html/gcode/main/main.html
[19:12:38] <daniel_br> I need know if is possible use a external switch for on/off feedoverride
[19:25:13] <jepler> daniel_br: m48 and m48 refer to the feed override feature of the GUIs, which is also available in 2.1 through halui
[19:25:27] <jepler> so the short answer is: yes, but probably not with version 2.0.x of emc2
[19:27:55] <daniel_br> I running head now
[19:28:20] <daniel_br> can I create a new pin for that?
[19:30:17] <jepler> the halui_halvcp configuration has an example of this
[19:30:18] <jepler> linkpp button.fo.increase halui.feed-override.increase
[19:30:18] <jepler> linkpp button.fo.decrease halui.feed-override.decrease
[19:30:58] <jepler> you would connect your external "increase feed override" button to halui.feed-override.increase and the "decrease feed override" button to .decrease
[19:31:06] <jepler> bbl
[19:32:46] <daniel_br> I see that, but I need simple use one output from my VFD for on/off feed override
[19:36:21] <SWPadnos> you want to use some load sensor as a feed override? (to reduce cutting speed when the spindle load goes too high)
[19:36:26] <daniel_br> my vfd output is on when the spindle rpm is ok
[19:38:27] <daniel_br> I need stop machine feed override until the spindle rpm is ok
[19:39:39] <SWPadnos> ok, so you want to stop motion until the VFD signals that the spindle is up to speed?
[19:40:00] <daniel_br> exactly
[19:40:17] <SWPadnos> ah. you can do that using the adaptive feed override input to the motion controller
[19:40:30] <skunkworks> you guys did that on the mazak
[19:40:38] <jepler> SWPadnos: remember that's not enabled unles some g-code is issued
[19:40:49] <SWPadnos> true
[19:41:23] <SWPadnos> unfortunately, FO is hard to use this way, since it conforms to the NIST principle regarding multiple locus of control
[19:41:55] <daniel_br> but the m48 is for do this
[19:41:57] <SWPadnos> and anything that deals with FO should read the actual value and reset itself to that setting (e.g. a slider on a GUI)
[19:42:16] <SWPadnos> M48 either allows or doesn't allow feed override, it doesn't pause motion
[19:42:59] <SWPadnos> in other words, you can normally set feed override from 0% to 125% (or more) of the programmed rate. M48 and M49 tell the motion controller to use this override, or ignore it
[19:43:07] <daniel_br> The Interpreter will interpret RS274/NGC commands which enable (M48) or disable (M49) the feed and speed override switches. It is useful to be able to override these switches for some machining operations. The idea is that optimal settings have been included in the program, and the operator should not change them.
[19:43:07] <daniel_br> EMC2 reacts to the setting of the speed or feed override switches on the control panel, when these switches are enabled.
[19:43:09] <SWPadnos> but it doesn't set the feedrate to 0
[19:43:40] <SWPadnos> is there a motion.hold input (or the like)?
[19:43:52] <daniel_br> from http://www.linuxcnc.org/docs/html/gcode/main/main.html
[19:44:48] <daniel_br> I need wait until my spindle is ok for continue running my gcode after some pause
[19:47:44] <SWPadnos> I'm looking at options - hold on one sec
[19:49:38] <jepler> motion.adaptive-feed seems like the thing to use, except for the fact that it takes an enabling code
[19:50:16] <jepler> you can put that in the startup g-code, but a user can still turn off adaptive feed by issuing the opposite code
[19:50:31] <SWPadnos> yeah. I'm kinda surprised that there's no feed-hold input
[19:50:51] <jepler> or a "spindle at speed" input
[19:50:59] <SWPadnos> wait a sec - isn't there a spindle at speed input?
[19:51:01] <SWPadnos> heh
[19:51:13] <jepler> no, I don't see it
[19:52:36] <SWPadnos> I wonder what happens when a Sxxx command is sent. it can't be expected that the spindle will be at speed immediately
[19:53:32] <cradek> I'm pretty sure it doesn't do anything but set the speed output
[19:54:11] <cradek> there should be a loopback like tool change has: spindle speed changed -> spindle at speed
[19:54:14] <cradek> or something
[19:54:30] <SWPadnos> hmmm. I guess I'd call that a bug (though there are times when you may want to keep motion going while changing the spindle speed)
[19:54:39] <SWPadnos> yep - that's what I was thinking
[19:54:58] <cradek> or a feedhold and you could do whatever you want with it
[19:55:27] <cradek> we've seen adaptive-feed abused as feedhold more than once - it means we probably need feedhold
[19:55:33] <SWPadnos> I'd say that by default, you'd want to pause motion when changing spindle speed, unless the motion is a G0 move
[19:56:14] <cradek> that seems reasonable at first glance
[19:56:24] <daniel_br> i think this spindle feedback is basic feature for a good cnc control
[19:57:05] <daniel_br> and I think emc2 is the pcbased cnc control
[19:57:10] <SWPadnos> you can do almost what you want using the adaptive feed override input, but as jepler said, you need to have a code to enable it (and it can be disabled)
[19:57:27] <daniel_br> err the best cnc control
[19:57:59] <SWPadnos> you can just have a mux2 component, with a 1 in one input and a 0 in the other. use the VFD output as the selector for hte mux, and run the output of the mux to the adaptive-feed pin
[19:58:24] <SWPadnos> that'll give you 0 speed or programmed speed (including feed override)
[19:58:46] <CIA-8> 03jepler 07HEAD * 10emc2/docs/src/gcode/main.lyx: document G50,G51
[19:59:16] <daniel_br> mux2?
[19:59:50] <SWPadnos> somewhere in the HAL documentation, you'll see the blocks component, which has several things that it can do
[20:00:09] <SWPadnos> one of them is a multiplexer, or something that takes several inputs and has a selector to choose which input gets sent to its output
[20:01:17] <daniel_br> ok I'll find this information
[20:01:32] <SWPadnos> I'd point you closer to it, but I don't remember the links :)
[20:01:41] <cradek> daniel_br: I suggest filing a feature request in the sourceforge tracker
[20:01:54] <cradek> (that way it won't be forgotten)
[20:02:14] <daniel_br> ok chris I'll do it
[20:07:46] <daniel_br> and the option hols some time... how can I use this?
[20:08:08] <daniel_br> err hold some time
[20:09:06] <SWPadnos> I think that would be a bit input, so you'd just connect the spindle "at-speed" pin to the feedhold pin (possibly inverted, depending on your hardware)
[20:14:11] <daniel_br> spindle "at-speed" pin and feedhold pin , can I create these pins in the halcmd?
[20:14:31] <cradek> no
[20:14:41] <SWPadnos> nope. halcmd only allows you to connect pins. you have to do some programming to make new pins :)
[20:15:00] <cradek> and also halcmd is the wrong place for these pins because they need to be realtime (they will be motion.something)
[20:15:24] <daniel_br> so I have problems
[20:15:32] <cradek> err I meant halui
[20:15:44] <cradek> * cradek will read the screen
[20:17:12] <SWPadnos> I don't think there's a totally foolproof way to do what you want right now
[20:17:21] <Lerneaen_Hydra> 'night
[20:17:30] <SWPadnos> I could be wrong, since waiting for a spindle to get up to speed seems like a very basic function
[20:17:44] <SWPadnos> so it "should" be there :)
[20:25:48] <daniel_br> How many time for this feature be in head?
[20:26:22] <SWPadnos> heh - unknown at this point.
[20:26:51] <SWPadnos> it's a pretty important feature, so it may be soon-ish rather than late-ish
[20:27:00] <cradek> if you make the feature request all the developers will see it
[20:27:17] <cradek> that's the best thing to do
[20:27:30] <daniel_br> chris can you send the link for that
[20:27:40] <cradek> sourceforge.net/projects/emc
[20:27:49] <cradek> then click feature requests
[20:46:34] <skunkworks> I wonder what jacky has been up to..
[21:10:21] <CIA-8> 03jepler 07HEAD * 10emc2/docs/src/gcode/main.lyx: document M62..M65 (digital output control)
[21:14:45] <CIA-8> 03jepler 07HEAD * 10emc2/docs/src/ (Master_Integrator.lyx Submakefile index.tmpl): begin documenting emc's interface to HAL: motion, axis.N, iocontrol
[21:14:46] <CIA-8> 03jepler 07HEAD * 10emc2/docs/src/config/emc2hal.lyx: begin documenting emc's interface to HAL: motion, axis.N, iocontrol
[21:57:47] <SWPadnos> ugh. driver update requires reboot. bbiab
[22:08:14] <daniel_br> see: http://www.isd.mel.nist.gov/personnel/kramer/pubs/RS274NGC_3.web/RS274NGC_3TOC.html
[22:08:50] <daniel_br> Feed and Speed Override Switches
[22:08:50] <daniel_br> The feed and speed override switches may be enabled (so they work as expected) or disabled (so they have no effect on the feed rate or spindle speed). The RS274/NGC language has one command that enables both switches and one command that disables both (see Section 3.6.5). The canonical machining functions have separate commands for the two switches (see Section 4.3.9). See Section 2.2.1 for further details
[22:26:45] <skunkworksemc> Ok - so it looks like I need to also install gcc and egcs on dapper to compile head?
[22:42:39] <skunkworks> ok - so if I read the directions - it works :)
[23:13:00] <skunkworksemc> so what did I f*** up? http://pastebin.ca/201710
[23:14:17] <skunkworksemc> http://pastebin.ca/201715
[23:14:53] <dmess> bon soir tous...
[23:17:38] <skunkworks> pre-compiled emc still works
[23:18:31] <skunkworks> dmess: I seem to be talking to my self :)
[23:20:53] <K`zan> skunkworks: Don't knock it, I do that and it is some of the better quality conversations I have these days ;-) LOL
[23:21:42] <K`zan> First successful part designed in Synergy! Short of putting it on something with CNC on it :-).
[23:21:52] <skunkworks> Cool
[23:21:55] <dmess> me too.. ; }====````
[23:22:36] <K`zan> Although I do hate arguing with myself, still hard too loose :).
[23:22:45] <dmess> congrats.... it seemed like a robust system to me...
[23:23:33] <dmess> what does it post process>>???
[23:23:35] <K`zan> Indeed seems so, I still got LOOOOOOOOOOOOOOOTTTTTTTTTTTTTTSSSSSSSSSS to learn though. Next we try dovetailing the other side of the part and then the drilling and tapping.
[23:24:03] <K`zan> dmess: My ignorance is legion at this point - "default post" is all I know at this point :-(.
[23:24:41] <K`zan> I need to get EMC and Synergy configured for what I plan on doing - tried the gcode in EMC and it had several problems :-( that I don't understand yet.
[23:24:41] <dmess> but of what?? APT source is my guess and HOPE
[23:25:21] <K`zan> dmess: Sigh, just dunno at this point, whatever it is, my Z axis on EMC is out of range :-).
[23:25:26] <dmess> does your intermediate code have GOTO/... in it??
[23:25:52] <K`zan> Dunno lemme go look, that is still up on another desktop...
[23:25:58] <dmess> post is pouched..
[23:26:47] <K`zan> No gotos, 321 lines of stuff that looks like this:
[23:26:50] <K`zan> %
[23:26:50] <K`zan> ( ****DESCRIPTION? **** )
[23:26:50] <K`zan> N1G17G20G40G49
[23:26:50] <K`zan> N2G54G80G90G94
[23:26:50] <K`zan> N3G00X-10.Y-6.Z-2.
[23:27:01] <K`zan> Got to remove the first two lines to even get started.
[23:28:32] <dmess> why>>?? that is posted... i need the synergy pre post CL ( cutter Location ) file
[23:28:55] <K`zan> Sigh, I have not a CLUE what you are talking about at this point.
[23:29:06] <dmess> what controlEMC?? EMC2??
[23:29:19] <K`zan> Booting up the emc2 box and going to try it over there...
[23:29:21] <K`zan> emc2
[23:29:34] <K`zan> is what I have - off the ubuntu live cd.
[23:29:55] <K`zan> booting it now.
[23:30:07] <dmess> the cad /cam system (SYNERGY) works in something more ROBUST than g codes...
[23:30:34] <K`zan> Probably so, but I have had it installed going on the second day now :)
[23:31:02] <K`zan> Pretty pleased with getting the part designed, fully aware that is probably the easy part :).
[23:31:31] <dmess> be proud... cut it and be brave
[23:31:52] <dmess> version??
[23:32:01] <dmess> bdi
[23:35:11] <dmess> first lines should pose no problem
[23:36:27] <K`zan_EMC> No luck, everything out of range. I guess I stop this and figure out how to set up EMC, a 1"x2"x3/4" piece should not be out of range anywhere :-).
[23:37:06] <K`zan_EMC> I suspect that what Synergy thinks of as 0,0,0 is considerably different than what EMC thinks it is (obviously :-).
[23:37:18] <dmess> its default is in metric
[23:37:20] <K`zan_EMC> The uMill is not a bridgeport.
[23:37:40] <dmess> what is it??
[23:38:01] <K`zan_EMC> I'm not at all sure about this, but in theory EMC2 is set up in inches, assuming I got the config right - something I need to check again.
[23:38:19] <K`zan_EMC> Harbor Freight Micro-Mill
[23:38:42] <dmess> 3 axis??? 5?? 6??7 ??
[23:38:54] <K`zan_EMC> * K`zan_EMC sits back and scratches arse and thinks about where to start.
[23:39:09] <K`zan_EMC> will be 3 axis when I actually decide to hemorage $$$ :-).
[23:39:13] <K`zan_EMC> x,y,z
[23:39:18] <dmess> at you arse since your there
[23:40:38] <dmess> you screwed up you coordinate system setting if your code is correct but out in the rhubarb
[23:41:01] <K`zan_EMC> {~/emc2/configs} $ ll
[23:41:01] <K`zan_EMC> total 16
[23:41:01] <K`zan_EMC> -rw-r--r-- 1 vw vw 8455 2006-09-30 23:36 stepper_inch.ini
[23:41:01] <K`zan_EMC> -rw-r--r-- 1 vw vw 816 2006-06-06 14:35 xylotex_pinout.hal
[23:41:29] <K`zan_EMC> Possible... Got Lots to learn...
[23:41:35] <jmkasunich> are you running axis? the preview should show you where your toolpath is
[23:41:55] <K`zan_EMC> Yes, axis, it shows something that doesn't seem to make sense right off.
[23:42:11] <dmess> not currentl aval..
[23:43:51] <K`zan_EMC> looks like x0 is about a foot to the left of the part in axis. That could be a problem :-)
[23:43:59] <dmess> if the coord sys is fubar'd then even axis will error out on limits
[23:44:40] <K`zan_EMC> Hummm, all I need to do now is figure out how that applies to anything at this point (I will NOT get discouraged here...).
[23:44:48] <dmess> can you TRANSLATE yor geometry?
[23:45:09] <K`zan_EMC> Last time I was messing with this, everything was set up and all I had to do was design parts :-).
[23:45:37] <K`zan_EMC> Probably could if I knew from what and to what... As noted, my ignorance is LEGION at this point in time.
[23:45:42] <dmess> did you import this model??
[23:46:07] <K`zan_EMC> I copied the G code out and loaded it into emc2...
[23:46:34] <K`zan_EMC> Which I am sure is part of the problem until I figure out how to tell Synergy what it is working with.
[23:46:55] <dmess> no th model / figure you are cutting
[23:46:58] <K`zan_EMC> Might work on a bridgeport based system.
[23:47:40] <dmess> no 3 axis should be brianless
[23:47:43] <K`zan_EMC> dmess, I *really* am not being obtuse here, honest. No idea what that even is unless it is the file synergy saves, sigh.
[23:49:00] <K`zan_EMC> All I ever had to do working in a "real" machine shop was generate the Gcode and load that into the machine in question...
[23:49:31] <K`zan_EMC> All this was already set up, NOW I get this part of the education :)!
[23:51:00] <K`zan_EMC> Gonna go look at a few things, bbl.