#emc | Logs for 2008-05-08

[00:32:20] <dmess> newer laptop???
[00:32:36] <dmess> in sol
[00:32:43] <dmess> i'm
[00:33:58] <dmess> what constitute SUCKS at RT.... this guy wanted to fight ove an RT thread running on linux in the 1st 5 minutes
[00:34:41] <dmess> i wanna place it it IN HIS FACE...
[00:34:44] <SWPadnos> if you need to generate steps with a 20000 BASE_PERIOD, and the computer has 25000 latency, that constitutes "sucks for RT"
[00:34:51] <dmess> even if in simul
[00:35:32] <dmess> i run real hardware ito this..
[00:35:36] <SWPadnos> laptops suck for running real machines, but will be great for running sim or RT demos, you just may get realtime errors
[00:35:43] <SWPadnos> do you have a newer laptop with a parallel port?
[00:35:46] <SWPadnos> or a PCI slot?
[00:36:12] <dmess> no he wont be hooked uppp.. just makin the sale...
[00:37:11] <dmess> of the EMC2 system... this guys a hardware developer on MY scale...
[00:37:27] <SWPadnos> ok, so he knows nothing about computers then? :)
[00:38:06] <dmess> oh he does... had ideas for a PIC withe 72 i/o foe 200 bucks
[00:38:20] <SWPadnos> I'd just get a mesa card
[00:38:24] <SWPadnos> (and a computer, of course)
[00:39:17] <dmess> what els does a MESA have... he's old school.. and just do it type of guy..
[00:39:31] <SWPadnos> FPGA
[00:40:14] <dmess> no he's saying have hal talk to the oi over the pci bus
[00:41:19] <SWPadnos> Mesa 5i20: PCI card with 200k gate FPGA and 72 I/Os, reprogrammable FPGA, configurations can be built with free Linux-based tools, driver already exists for this card, and new ones with more features are in development
[00:41:48] <dmess> fpga??? video format for the commadore 64???
[00:42:41] <dmess> Fu@#$%'n primitave graphic adaper
[00:43:05] <dmess> but i did learn assembler on it
[00:43:16] <SWPadnos> 6502, baby
[00:43:32] <dmess> you too??
[00:43:44] <SWPadnos> I learned on the Atari, but it had the same CPU
[00:43:55] <dmess> your too young arent ya??
[00:44:09] <SWPadnos> did a fair amount of C64 and VIC20 programming, and we even built a memory expander card in our computer club
[00:44:25] <ds2> Mmmmmm GTIA/ANTIC video :)~~~~
[00:44:26] <SWPadnos> depends on what you mean by "too young"
[00:44:31] <SWPadnos> I'm old enough to have kids who can vote
[00:44:34] <SWPadnos> (but I don't)
[00:44:50] <rayh> I wrote my first elliptical gcode generator program on a vic20.
[00:44:53] <dmess> we hacked the crap out of all the cartridge games in our place...
[00:45:11] <dmess> really RayH
[00:45:37] <SWPadnos> I probably still have a printout of my integer-only drawing program, which could do rubber-banding of ellipses (and rectangles and lines, of course)
[00:45:45] <rayh> Yep. I don't even remember when it was.
[00:46:23] <rayh> Then they upgraded the mazak control from m1 to m2 and I had to rewrite the program for 4 decimal places.
[00:46:32] <SWPadnos> multiple graphics modes, undo, joystick or dual paddle input (like an etch-a-sketch), save/load from disk, pattern fills - the whole 6.2 yards
[00:46:40] <dmess> 86 or so... paid 1600 dollars for a c64 drive and monitor
[00:46:43] <SWPadnos> (we didn't have 9 yards when I was a kid)
[00:46:53] <SWPadnos> good old 1541
[00:48:20] <dmess> my mom still has it in the shed.. all in original boxes..... but been cycling between 122F and -44f.. since then
[00:48:56] <dmess> im guessing pretty plastic
[00:50:14] <dmess> did you ever work with squirts??
[00:50:43] <SWPadnos> no
[00:53:02] <dmess> was the begining of the .gif animation thing we used them in games ONLY commadore used them
[00:54:43] <dmess> you could make a squirt and iterat that per line as opposed to animating a line at a time.. yes... it was always the same line
[00:57:04] <SWPadnos> oh, do you mean sprites?
[00:57:43] <ds2> Mmm player missle graphics
[00:58:00] <SWPadnos> the Atari, Commodore, and TI-99/4A all had sprite graphics
[00:58:33] <SWPadnos> and apparently others: http://en.wikipedia.org/wiki/Texas_Instruments_TMS9918
[00:59:06] <SWPadnos> (with that chip alone, which we also put onto an Apple plug-in card in high school)
[01:00:32] <dmess> sorry YES sprite... brain fart
[01:00:49] <SWPadnos> shit (almost) happens
[01:01:23] <dmess> 33 yrs too late
[01:02:41] <dmess> we had a super pet at the high school when the university student that liked us was home... ; (
[01:03:19] <dmess> that learned me some fortran early
[01:03:20] <SWPadnos> he
[01:03:24] <SWPadnos> h
[01:03:41] <SWPadnos> I wrote a turtle graphics-like program on a superPET for my geometry project in high school
[01:03:53] <dmess> i was 12-14 then...
[01:04:09] <dmess> cool
[01:04:37] <SWPadnos> too bad that teacher died - he was really good
[01:04:40] <dmess> how it is we all end up in same places and activities ehhh
[01:05:06] <dmess> that sucks
[01:05:41] <SWPadnos> yeah. it seemed we'd lose one teacher every year, and it was always the best/most popular ones
[01:06:14] <dmess> i lost a VERY good machine shop teacher 2 yrs into h/s... no one could replace HIM in my eyes
[01:06:48] <SWPadnos> yep, that's the kind of person I'm talking about
[01:08:17] <dmess> they become mentors postusley
[01:28:49] <lewin1> lewin1 is now known as lewing_
[01:49:02] <skunkworks> huhforgotaboutthespacebar
[01:49:07] <skunkworks> letsreboot
[03:15:38] <cradek> UPS, the UPS brandmark, and the color brown are trademarks of United Parcel Service of America, Inc. All rights reserved.
[03:18:00] <SWPadnos> shit was here before UPS, I'm sure there's prior "art"
[04:35:09] <Gamma-X> anyone been to ireland?
[04:35:16] <Gamma-X> i just did sumtin realy stupid....
[05:20:24] <lerman_> lerman_ is now known as lerman
[11:57:23] <fenn_> fenn_ is now known as fenn
[14:11:30] <BigJohnT_> OT, on a LM339 differential comparator which voltage goes to out IN- or IN+ or Vcc?
[14:12:39] <cradek> I don't understand the question
[14:13:12] <BigJohnT_> I'm trying to follow a schematic and it shows two inputs to each comparator and one output
[14:13:18] <archivist> yes
[14:13:21] <cradek> right
[14:13:28] <BigJohnT_> what is the output
[14:13:40] <cradek> high or low, depending on the comparison of the two inputs
[14:14:05] <BigJohnT_> is it switching the Vcc?
[14:14:30] <archivist> no
[14:14:35] <cradek> yes I think the output you get will be near vcc if it's high, near ground if low
[14:14:48] <cradek> if that's what you're asking?
[14:14:54] <BigJohnT_> ok that makes some sense now
[14:14:56] <BigJohnT_> yes
[14:15:36] <BigJohnT_> I found a circuit for a torch height control and I am just trying to understand it LOL
[14:16:08] <cradek> wellll I'm wrong I think. I think the output sinks current to ground only - open collector
[14:16:28] <cradek> "output compatible with ttl, mos, cmos"
[14:16:32] <jepler> all the examples seem to show pull-up resistors
[14:16:47] <cradek> so don't worry about it -- think of the output as digital if it's going into digital circuitry
[14:17:24] <BigJohnT_> it goes to a 2n3904 then to an opto then to the parallel port
[14:17:51] <jepler> Have a look at page 15 of http://www.national.com/ds.cgi/LM/LM139.pdf
[14:18:10] <jepler> I don't understand all of it, but I can see that the output will only sink current
[14:18:37] <BigJohnT_> * BigJohnT_ reading it now
[14:18:56] <cradek> BigJohnT_: if the other input of the opto goes to a supply voltage it will work fine (and I bet the 2n3904 is not needed)
[14:19:11] <jepler> so you'd want to hook V+ -- R --|>|-- (opto) LM339 output
[14:19:23] <cradek> I agree
[14:20:00] <archivist> can the LM source enough current for the opto
[14:20:02] <jepler> output sink current min 6.0mA
[14:20:12] <jepler> so it'll depend on the opto, but you could choose one that was OK
[14:20:28] <jepler> (page 3)
[14:20:31] <cradek> output current 20mA max - page 4
[14:20:36] <BigJohnT_> they have a 4n26n on this circuit
[14:24:14] <BigJohnT_> hmmm trying to figure out where to post the pdf so you guys can see it
[14:32:57] <rayh> Upload to wiki.linuxcnc.org
[14:33:31] <rayh> If you want it to remain fairly obscure don't put a link to it in the wiki.
[14:34:14] <BigJohnT_> ok thanks ray
[14:35:14] <jepler> units
[14:35:36] <jepler> er, oops
[14:35:50] <BigJohnT_> upload:THC.pdf
[14:35:55] <BigJohnT_> at wiki
[14:38:53] <BigJohnT_> ok here it is http://wiki.linuxcnc.org/uploads/THC.pdf
[14:40:24] <BigJohnT_> I assume that this design is using an AC torch voltage as it goes to a bridge rectifier, mine is a DC so can I pass up BR2?
[14:41:15] <archivist> some of the 339 outputs are wire ored
[14:41:42] <BigJohnT_> c and d?
[14:42:47] <archivist> hmm sore neck reading sideways....
[14:43:04] <BigJohnT_> me too trying to figure out how to rotate it
[14:43:40] <archivist> pins 2 14 13 are joined
[14:44:18] <archivist> er except 2 to 13,14 has a diode
[14:45:50] <BigJohnT_> right click on the drawing and you can rotate it...
[14:46:13] <archivist> not having a working printer here atm is a pain
[14:46:28] <archivist> not in gpdf
[14:46:36] <fenn> paper addicts!
[14:46:41] <BigJohnT_> darn
[14:47:37] <archivist> but there is some inverse logic there
[14:48:10] <BigJohnT_> a rotated one http://wiki.linuxcnc.org/uploads/THCr.pdf
[14:49:14] <BigJohnT_> is the AC input in the upper right to the BR1 just a DC power supply?
[14:50:17] <archivist> yes then regulated to 5 v in the 7805
[14:50:42] <BigJohnT_> so if I have 5v power supply I can skip that part of the circuit
[14:51:06] <archivist> yessum
[14:51:35] <archivist> the 3904 inverts the lm339 output
[14:51:51] <BigJohnT_> ok
[14:52:32] <archivist> and there is or'ing of the LM339 outputs
[14:54:24] <BigJohnT_> is the BR2 only there to convert the torch voltage from ac to dc?
[14:55:06] <archivist> yes
[14:55:31] <BigJohnT_> ok I have dc torch voltage signal
[14:55:52] <archivist> and if dc allows reverse connection
[14:56:06] <BigJohnT_> can you tell what range the torch voltage signal would be? 0-5
[14:56:19] <BigJohnT_> what do you mean by reverse connection?
[14:56:37] <archivist> doesnt matter +-
[14:56:48] <BigJohnT_> ok
[15:02:17] <lewin1> lewin1 is now known as lewing_
[15:05:11] <archivist> do you mean the up/down outputs
[15:06:02] <BigJohnT_> no the torch voltage input
[15:09:14] <archivist> well 47 k in series and 4k7 to 15k to ground, so its adjustable also there is an adjuster of the trigger point in the middle of the circuit
[15:12:51] <BigJohnT_> I saw the 10 k pot to adjust the set point.
[15:13:41] <BigJohnT_> as simple as it is I think I'll build one without the br's and use my bench power supply to simulate the torch voltage for testing.
[15:14:04] <archivist> needs calculating to check all switching points (but sort of working at work atm)
[15:14:23] <BigJohnT_> archivist: thanks for the help
[15:14:39] <archivist> yup you will soon see what changes and when
[15:54:28] <ChanServ> [#emc] "This is the #emc channel - talk related to the Enhanced Machine Controller and general machining. Website: http://www.linuxcnc.org/, wiki at http://wiki.linuxcnc.org/"
[16:15:34] <cradek> hahaha http://www.sears.com/shc/s/p_10153_12605_009H0259000B?sid=I0084400010000100600&aff=Y
[16:18:31] <fenn> * fenn wants a bigger picture for a desktop wallpaper
[16:19:23] <alex_joni> cradek: cool.. how much is it?
[16:19:29] <cradek> $8600
[16:19:38] <alex_joni> ouch.. but still :)
[16:20:37] <archivist> hmm needs a lot of garage wall to tidy
[16:21:56] <fenn> one might think it would come with a tool chest to put everything in
[16:22:34] <archivist> "a" tool chest /me would not be able to lift
[16:22:56] <fenn> not that kind, the ones with roller drawers
[16:23:43] <BigJohnT> Lowes has better tool chests
[16:26:51] <alex_joni> fenn: they have one
[16:26:56] <alex_joni> look at http://s7-apps.sears.com/searsZoom/searsDHTMLZoom.js
[16:27:10] <alex_joni> here's the pic URL: http://s7.sears.com/is/image/Sears/009H0259000?qlt=100,0&resMode=sharp&op_usm=0.9,0.5,0,0,0
[16:27:29] <fenn> ya not quite big enough
[16:27:45] <skunkworks> wow - how many ratchets do you need?
[16:27:57] <fenn> fourteen!
[16:28:20] <fenn> i mean, seventeen
[16:28:33] <alex_joni> fenn: check out the js
[16:28:44] <alex_joni> it can deliver higher res images.. but I have no time hacking it atm
[16:29:11] <alex_joni> fenn: function ts_zoom(baseImage,width,height,isSecure,baseImage2,newWidth, newHeight, imgVersion){
[16:29:15] <alex_joni> gotta run now..
[16:29:17] <fenn> seems like an awful lot of code for simply zooming into pictures
[16:29:23] <fenn> thanks alex
[16:30:33] <cradek> aha http://s7.sears.com/is/image/Sears/009H0259000?qlt=100,0&resMode=sharp&scl=+2
[16:31:55] <skunkworks> dad loves the boxend-open end wrenches that are ratchets.
[16:32:13] <cradek> ha, a timing light
[16:32:15] <cradek> that's funny
[16:32:46] <skunkworks> do cars need a timing light anymore? I think that about mid 90's there are no adjustments anymore.
[16:32:48] <cradek> I have a dwell meter too, and they both run on 6v OR 12v
[16:33:03] <cradek> I agree
[16:33:13] <cradek> but I guess mechanics still work on older cars.
[16:33:18] <skunkworks> I suppose
[16:34:12] <cradek> skunkworks: which wrenches are you talking about? can you find a picture?
[16:36:00] <archivist> * archivist imagines all the grown men dribbling over tool pictures
[16:37:25] <skunkworks> http://www.sears.com/shc/s/p_10153_12605_00942434000P
[16:37:25] <skunkworks> similar to that..
[16:37:34] <skunkworks> I have not actually used his yet
[16:37:36] <jepler> haha extreme
[16:38:00] <skunkworks> ;)
[16:38:01] <cradek> oh cool, the closed end ratchets
[16:38:05] <skunkworks> yes
[16:38:46] <cradek> hmm, only down to 7/16 though
[16:55:04] <fenn> cradek: how did you come up with that url?
[16:57:33] <cradek> half reading the js, half luck
[16:57:40] <archivist> must have been searching for mechanics porn
[16:58:48] <archivist> which reminds me I must get my adjustable site back up
[18:10:23] <LawrenceG> K`zan, how goes the machine construction?
[19:46:09] <dimas> hi all
[19:48:03] <dimas> first time online from ubuntu8.04-emc box
[19:48:30] <skunkworks> cool
[19:49:00] <assargadon> hi there
[19:53:50] <gefink> assargadon: hi
[19:55:07] <assargadon> we have holiday tommorow
[19:55:24] <assargadon> so I find a some portion of time for EMC and Robotics :)
[19:56:30] <assargadon> by the way, I use gedit for comp file editing in 6.06
[19:56:46] <assargadon> maybe some better tool possible?
[20:00:54] <gefink> assargadon: vim for putists
[20:01:09] <gefink> purists
[20:01:19] <skunkworks> * skunkworks uses gedit
[20:01:32] <K`zan> LawrenceG: Slowly :-). Now that it is basically together all the design flaws and misajustments are quite obvious - working at reducing them now, taking time and some afro-engineering, but it is coming along slowly but coming along.
[20:01:36] <skunkworks> * skunkworks doesn't know any better
[20:03:36] <BigJohnT> * BigJohnT uses gedit
[20:04:17] <K`zan> With the larger than original design steppers, the table is unbalanced and tilts up on the end opposite the stepper. A chunk of steel glued to the far end bottom of the table takes care of the problem rather nicely and the steppers are more than capable of dealing with it. Next comes the Y axis (too light and possibly too "uncontrolled" (slop) that an additional support bearing and more weight may well solve. Then the Z needs a bit more stif
[20:04:17] <K`zan> fining and that should make it about as well as it can be.
[20:05:06] <K`zan> What to do about EMCs lack of understanding of what goes on (distance moved not corresponding and backward movement in the live plot) will have to be resolved.
[20:05:12] <K`zan> Fun :-).
[20:05:18] <K`zan> gedit - rech...
[20:05:19] <BigJohnT> K'zan do you have any pictures?
[20:05:31] <K`zan> None perfectly recent but:
[20:05:50] <K`zan> http://wrlabs.shacknet.nu/~vw/MyMachineShop/PipeDreamMill/
[20:06:08] <K`zan> Got to get the latest out of the camera, will do that today if all goes well.
[20:06:49] <alex_joni> K`zan: what do you mean by "what to do about EMCs lack of understanding.." ?
[20:06:58] <skunkworks> Cool - youtube
[20:07:48] <BigJohnT> cool, a pipe mill
[20:07:56] <skunkworks> *saw the directions for making the pipemill on youtube
[20:08:09] <K`zan> alex_joni: What to do about EMCs lack of understanding of what goes on ----->>>>>(distance moved not corresponding and backward movement in the live plot) <<<<<-----will have to be resolved.
[20:08:20] <fenn> looks like a bad merge
[20:08:34] <K`zan> BigJohnT: Yep, sometime I can afford and I am learning a LOT from doing it :-).
[20:08:38] <alex_joni> K`zan: I still don't understand what you mean..
[20:08:53] <skunkworks> no position feedback
[20:08:54] <fenn> K`zan: i think you just need to set your scale properly
[20:08:58] <K`zan> Did CNC programming about 10 or so years ago, but they would never let me back in the shop.
[20:09:09] <K`zan> fenn: In theory, it is set correctly.
[20:09:20] <fenn> in theory, if it doesn't match reality, then it's not set correctly
[20:09:40] <gefink> K: mean you that the rubberhose ist elastic?
[20:09:46] <BigJohnT> did you measure the actual travel verses the commanded distance?
[20:09:52] <K`zan> Odd in that when I set microstepping to say 1/2 then the distance appears to be 1/2 of commanded, with 1/4 it seems that it is 1/4 of commanded,
[20:10:07] <fenn> that's not odd at all
[20:10:13] <K`zan> Yes, I did, tape measure, but that will resove the diff between 1" and 1/4" :-).
[20:10:20] <alex_joni> K`zan: if you set microstepping then you need to change scale accordingly
[20:10:28] <K`zan> fenn: Do not mistake me for ANY kind of expert :).
[20:10:33] <BigJohnT> yep
[20:10:38] <alex_joni> 1/2 stepping means you need twice as many steps for the same distance
[20:10:47] <fenn> emc only tells the drive the number of steps to take, not 'go one inch' or something
[20:10:56] <alex_joni> because on each commanded step (from emc2) it will go one half of a step
[20:10:58] <K`zan> alex_joni: Using stepconf, it is set correctly - I figure to fake it out with using more steps / rev and see what happens.
[20:11:08] <K`zan> 200 * 4 rather than 200.
[20:11:21] <alex_joni> K`zan: rewind..
[20:11:30] <K`zan> I though stepconf calculated that, but apparently not ?!?
[20:11:31] <alex_joni> what do you have your hardware set for?
[20:11:53] <gefink> K`zan: insert a 4 in the field for microstepping...
[20:11:56] <alex_joni> full steps, half steps, 1/4 steps.. etc
[20:12:02] <K`zan> 20 TPI (1/4-20) and steppers are 1.8 degree (200).
[20:12:16] <alex_joni> K`zan: ok, how about the stepper drivers?
[20:12:22] <alex_joni> I think yours can be jumpered
[20:12:27] <alex_joni> for microstepping..
[20:12:34] <K`zan> gefink: It is so set. The value that stepconf comes up with is correct from what I can figure.
[20:13:42] <K`zan> Checked the jumpers 4 times against the docs, either they are right or I am out in left field :). HobbyCNC 4 axis pro - 1/4 step: J1 OFF, J2 ON, J3 ON...
[20:13:56] <fenn> what is the value for scale that stepconf puts in the .ini file?
[20:14:33] <K`zan> fenn, don't know, gonna have to fire up that box and deal with (yuk, rech, barf :-) gnome :-)
[20:14:43] <fenn> you dont have to use gnome
[20:14:43] <K`zan> Bringing it up now...
[20:15:13] <fenn> in fact if it werent for stepconf, gnome wouldnt be needed at all
[20:15:16] <K`zan> I just haven't gotten around to installing KDE yet (same thing on all my other boxes :-). May get to that today too .
[20:15:26] <gefink> i think the axis sale shold be 16000
[20:15:31] <alex_joni> ok.. so 1/4 stepping means: 20 * 200 * 4 = 16000
[20:15:38] <K`zan> fenn: To be honest I don't know where the problem lies.
[20:15:55] <gefink> 200steps*4quatersteps*20TPI
[20:15:59] <K`zan> gefink: alex_joni that is the value that stepconf comes up with.
[20:16:09] <K`zan> But movement is still off.
[20:16:34] <gefink> are you sure yout rubbercoupling is fix
[20:16:43] <skunkworks> how much off?
[20:16:44] <alex_joni> K`zan: off how much?
[20:16:52] <K`zan> Wondering if perhaps coils are wired wrong, in that at small values of movement (1 IPS default) they sound like my garbage disposal :-)
[20:17:20] <alex_joni> K`zan: try to move one step at a time
[20:17:23] <alex_joni> and watch the motor
[20:17:25] <K`zan> alex_joni: 1" commanded - 1/4" movement. Was 1" commanded and 1/2" using 1/2 stepping.
[20:17:45] <alex_joni> if it turns in the same direction for each step
[20:18:52] <assargadon> Why integer pins are used rarely in EMC?
[20:19:39] <fenn> assargadon: convenience
[20:20:45] <K`zan_emc> Ok, up on the EMC box...
[20:21:09] <K`zan_emc> Had to reboot it since I was switched to the other box (KVM) and it came up in 640x480 :-)
[20:21:32] <assargadon> integer operations are much faster. In most cases some kind of "steps" used...
[20:22:35] <K`zan_emc> Grr, X on this box doesn't allign the screen, gotta move everything over to see it.
[20:23:19] <alex_joni> assargadon: if you can do a PID loop with integers.. be my guest
[20:23:33] <fenn> i've done PID with integers and it's no fun
[20:23:39] <K`zan_emc> Ah, perhaps the problem is that I have 2 configs and EMC is getting the wrong one?!?
[20:23:50] <alex_joni> K`zan_emc: :P
[20:23:59] <gefink> K`zan_emc: try to move 1/20" and check rtat the motor makes 1 full turn
[20:24:21] <K`zan_emc> Click on EMC-HAL-STEP-IN and have been editing PipeDream1 with stepconf.
[20:25:28] <K`zan_emc> WHich uses stepper_inch.ini - NOT the one I have been editing :).
[20:26:13] <K`zan_emc> Should be launching PipeDream1.ini, I think. Lemme make up another launcher for that and see what happens.
[20:26:51] <gefink> try using stepconf-wizard and make a new config
[20:27:26] <alex_joni> well.. good night all
[20:27:31] <alex_joni> gotta travel tomorrow :)
[20:27:52] <K`zan_emc> gefink: That is how I got PipeDream1 :).
[20:28:07] <K`zan_emc> Got a launcher for EMC to use that, lemme fire up the power supply and EMC.
[20:28:48] <BigJohnT> night alex
[20:28:58] <K`zan_emc> Oops, gotta put the Y axis back in, glue is dry on the weight on the X axis.
[20:29:02] <K`zan_emc> Night alex_joni !
[20:29:08] <K`zan_emc> alex_joni: Thanks for the help!
[20:36:39] <K`zan> Sorry, minor crises here, bbiab.
[20:42:37] <K`zan> Well, that is just weird as hell....
[20:43:10] <K`zan> My Y axis lead screw now doesn't even come close to aligning with the coupler.
[20:43:58] <K`zan> As far off as it is, I could not have had it attached before I took it apart. Utterly bizarre....
[20:44:50] <K`zan> Nothing has changed other than adding the weight to end of the X platform...
[20:45:05] <gefink> hmmm
[20:45:24] <K`zan> Too weird, I need to move it back to center which is about 3/8" from where it was. No slippage of holes.
[20:45:58] <K`zan> Even though the table is wood, no amount of humidity could shift it that far.
[20:46:10] <K`zan> Odd...
[20:46:25] <K`zan> Gremlins !
[20:46:29] <K`zan> Demons!
[20:46:34] <K`zan> Liberals ;-)!
[20:46:40] <K`zan> LOL
[20:48:33] <gefink> sometime its wired
[20:49:26] <fenn> K`zan_emc: iirc in that design the stepper motor bears the thrust loads, so that would go through the hose coupler, yes?
[20:49:59] <fenn> so if you ran the table up against the end-stops it could have pulled the screw out of the coupler
[20:55:19] <K`zan> Using hombrew aluminum couplers rather than the hose.
[20:55:34] <K`zan> With set screws :-)
[20:59:25] <gefink> yes
[21:00:08] <assargadon> I figured out an init problem.
[21:04:10] <gefink> linux init or emc init
[21:08:03] <BigJohnT> time to start drinking!
[21:08:18] <BigJohnT> opps, I already started...
[21:13:32] <K`zan> Heh, there is a thought, where is that Youkon Jack :-)?
[21:19:52] <gefink> good night
[21:27:58] <assargadon> If I starting encoder, it has some kind of initial state
[21:28:14] <assargadon> if this initial state differ from actual state
[21:28:33] <assargadon> it can make 1 additional step just after starting
[21:29:30] <assargadon> I can solve this problem with some kind of flag variable
[21:29:48] <assargadon> to detect first circle, I mean
[21:29:54] <assargadon> but it looks weird
[21:32:11] <skunkworks> you're trying to make a single pulse encoder work for position?
[21:32:38] <assargadon> yes.
[21:32:51] <assargadon> But it not really depends on encoder type
[21:33:11] <assargadon> quite same problem possible for quadruple encoder
[21:33:20] <skunkworks> I can't see how that will work.
[21:33:53] <assargadon> for example, your beginning state of input pins are 0 0
[21:34:11] <assargadon> and system's beggining state is 0 1
[21:34:31] <assargadon> when update() function will be calculated first time
[21:35:09] <assargadon> it should compare current state with remebered one, isn'tit?
[21:35:43] <assargadon> if it differ, it should increase/decrease coordinate and remeber the new state
[21:36:21] <assargadon> so initial values of remebered state will be 0 0
[21:36:39] <assargadon> anf first time update() will be calculated
[21:37:03] <assargadon> function will compare input pins 0 1 with remebered state 0 0
[21:37:14] <assargadon> and it assumes step happend
[21:37:20] <assargadon> isn't it?
[21:39:13] <skunkworks> * skunkworks doesn't know how the encoder counter code works
[21:40:22] <skunkworks> I assume at startup - it makes the remembered state equal to the encoder state
[21:40:43] <skunkworks> (without moving anything)
[21:40:56] <assargadon> if so, it should detect this startup, isn't it?
[21:41:34] <assargadon> I mean, it needs some way to detect startup loop
[21:42:59] <jepler> it doesn't really matter if the "power-up state" is 0, 1, or -1 counts since you don't know the position of the axis until you home anyway
[21:43:58] <lewin1> lewin1 is now known as lewing_
[21:46:18] <assargadon> jepler, do you mean that some kind of reset should happend before using any encoder output?
[21:52:16] <skunkworks> with a single channel encoder - you're never going to be sure which direction the shaft is turning. What happens when you push the table? How do you know which way the screw is turning?
[21:52:22] <jepler> assargadon: at power-up the encoder is really free to give any feedback value
[21:53:01] <jepler> assargadon: in emc, "homing" is used to compute the offset between encoder position and axis position
[21:53:17] <assargadon> skunkworks, screw is rotating in direction motors turning it
[21:53:46] <jepler> skunkworks: assargadon doesn't want to hear about why his single-channel encoder scheme won't work, so don't bother. we've told him before now.
[21:53:53] <assargadon> it's really hard to push the carriage hard enought to force screws rotating
[21:54:21] <assargadon> and after all, I have no another encoders, so ....
[21:54:42] <assargadon> I should do my best with equipment I have, isn't it ? :0
[21:55:04] <skunkworks> oh
[22:06:48] <dave_1> Hmmmm pretty quiet.
[22:07:10] <dave_1> I need help with setting spindle speed.
[22:07:46] <dave_1> apparently there is a comp to do this ?????
[22:09:09] <jepler> what signal does your spindle need to set the speed?
[22:09:18] <jepler> PWM, analog, ??
[22:09:59] <dave_1> analog off dac
[22:10:15] <dave_1> 0.025 V /RPM
[22:10:33] <jepler> which interface card with dac are you using?
[22:10:36] <dave_1> stg
[22:11:42] <jepler> first pass is to hook motion.spindle-speed-out to stg.<channel>.dac-value and set .dac-offset and .dac-gain parameters appropriately
[22:12:33] <jepler> "setp stg.<channel>.dac-gain 0.025" should make the scale equal to 0.025 V / RPM
[22:13:06] <dave_1> let me try and get my hands on what I have ...
[22:13:27] <dave_1> it is printed out someplace ... probably down in the shop.
[22:13:30] <dave_1> brb
[22:13:43] <jepler> oh is this partially set up for your machine?
[22:15:39] <rayh> Dave had a 5.10 install this morning running the older config.
[22:16:41] <rayh> But it was picking up the spindle stuff by reading variables in his INI file from that install.
[22:17:11] <rayh> Maybe we can get him to pastbin his old ini and hal files.
[22:17:19] <dave_1> this may be another naming problem.
[22:17:41] <rayh> dave_1, You at the machine now?
[22:18:07] <dave_1> no ... gaim doesn't like me ... at least on FC5 ...
[22:18:19] <dave_1> so I use Xchat on the mac
[22:18:24] <jepler> http://cvs.linuxcnc.org/cgi-bin/cvsweb.cgi/emc2/configs/max/max.hal.diff?r1=1.13;r2=1.14;f=h
[22:18:36] <jepler> ^^ there was a change in the naming of spindle pins from emc 2.0.x to later versions
[22:18:58] <dave_1> ah ... I was beginning to suspect that might be the problem.
[22:19:02] <rayh> As you get time from the machine issue apt-get install xchat.
[22:19:18] <jepler> this is supposed to be an exhaustive list of incompatible hal and ini changes over the versions: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?UPDATING
[22:19:22] <dave_1> had to change names for the index pins
[22:19:48] <jepler> spindle was down there at item 2.3.
[22:23:29] <dave_1> so I can change to the stuff in the link you posted and things should work?
[22:25:35] <dave_1> think I can handle that
[22:26:41] <dave_1> Actually the update has gone pretty well; slowest part was upgrading from the Live CD to present. Source site must have been loaded heavily
[22:27:00] <dave_1> as download speeds were as slow a 35 Kb/sec at times
[22:27:42] <dave_1> after getting the names for the index pulses correct(ed) homing on index works which is great.
[22:28:05] <dave_1> going to go make things work ... thanks
[22:31:11] <fenn> assargadon: if i were you, i'd rip out all the electronics and just use two opto-interrupter modules (with hand-made wiring to connect to parport)
[22:31:42] <toastydeath> i would use magic
[22:35:32] <assargadon> fenn, I just have no skills for it
[22:40:44] <dimas> * dimas has installed qcad
[22:49:32] <rayh> Hi dimas. What you think of it so far?
[22:50:43] <dimas> just overviewing..
[22:54:19] <dimas> seems complex enough
[23:01:59] <dave_1> Hey, jepler .... IT SPINS!!!! THANKS ...
[23:13:41] <assargadon> ok, encoder seems works :)
[23:14:12] <assargadon> my first realtime code, my first HAL module :)
[23:28:48] <dmess> you the man... :)
[23:41:49] <assargadon> aggr
[23:42:30] <assargadon> is it possible combine in 1 .hal pyvcp's and rt components?
[23:42:58] <assargadon> when control panel appeared, it's too late
[23:43:13] <assargadon> halrun do not detects pins already
[23:43:29] <SWPadnos> you can' t make a single component that has both RT and user interface parts
[23:43:39] <SWPadnos> you can make a UI and a RT component, then connect them together
[23:43:48] <jepler> you must use the "-W" or "-Wn" flag to loadusr, so that halcmd waits for the userspace component to create all its pins before executing the subsequent commands.
[23:43:50] <SWPadnos> you may need to use the -W option for halcmd loadusr
[23:44:12] <jepler> sim/pyvcp_demo.hal:loadusr -W pyvcp -c pyvcp pyvcp_demo.xml
[23:44:19] <jepler> as in the sample configuration files: ^^
[23:44:34] <SWPadnos> shouldn't that use the Wn variant?
[23:44:55] <SWPadnos> oh, no. nevermind
[23:45:07] <SWPadnos> time for me to run - see you later
[23:45:08] <jepler> I'm not sure, maybe it should
[23:45:11] <SWPadnos> nope
[23:45:19] <SWPadnos> you'd need one more pyvcp for that ;)
[23:45:38] <SWPadnos> loadusr -Wn pyvcp pyvcp -c pyvcp pyvcp_demo.xml :)
[23:47:45] <assargadon> jepler, I trying use -W
[23:48:15] <assargadon> but pyvcp do not send the "ready" message
[23:48:24] <assargadon> so this loop becomes infinite
[23:49:17] <assargadon> ok, I'll dig more about UI and rt combining tommorow :)
[23:49:44] <assargadon> I just dissapointed a little, as far as I can do it manually quite easy
[23:53:11] <jepler> you should read the pyvcp and halcmd manpages to understand whether to give the -n argument to loadusr and/or the -c argument to pyvcp
[23:53:39] <jepler> halcmd and pyvcp must agree on the component name, or as you say it will wait forever
[23:54:10] <jepler> you have three ways of doing it: use loadusr -Wn ..., use ... pyvcp -c ..., or use loadusr -Wn ... pyvcp -c ...
[23:55:48] <assargadon> there are no -Wn description in "help loadusr" output
[23:55:58] <jmkasunich> help halcmd
[23:56:04] <jmkasunich> then scroll down to the loadusr command
[23:56:18] <assargadon> -W wait for HAL component to become ready
[23:56:20] <assargadon> -w wait for program to finish
[23:56:21] <assargadon> -i ignore program return value (use with -w)
[23:56:23] <assargadon> ah ok
[23:57:19] <jepler> or type "man halcmd" to view the manpage
[23:57:29] <jepler> the great thing about the documentation is how many different pieces of documentation there are :-P
[23:57:42] <jepler> [flags] may be one or more
[23:57:42] <jepler> of:
[23:57:45] <jepler> ยท -Wn name to wait for the component, which will have the
[23:57:45] <jepler> given name.
[23:58:57] <assargadon> yes, I see