#emc | Logs for 2010-11-01

Back
[00:03:25] <theorb> theorb is now known as theorbtwo
[00:08:42] <andypugh> UncleG: You need to add an encoder function to HAL, then connect the counts output to axis.N.jog-counts
[00:08:43] <andypugh> (s32, in) Connect to the "counts" pin of an external encoder to use a physical jog wheel.
[00:08:43] <andypugh> axis.N.jog-enable
[00:08:43] <andypugh> (bit, in) When TRUE (and in manual mode), any change in "jog-counts" will result in motion. When false, "jog-counts" is ignored.
[00:18:20] <ries_> ries_ is now known as ries
[00:22:20] <UncleG> andypugh do you have an example halfile I can attempt to decode and rework for my system, I have tried to make sense of the hal manual. To no avail.
[00:22:31] <andypugh> Sorry, no.
[00:22:57] <andypugh> Do you have any other encoders in your current hal file?
[00:23:01] <UncleG> yeah
[00:23:12] <UncleG> Im using the 5i20
[00:23:25] <UncleG> w/ 7i33
[00:23:41] <andypugh> Increase the num_encoders by one.
[00:23:56] <UncleG> okay.
[00:24:22] <andypugh> You should see where the pins need to go in dmesg
[00:24:33] <andypugh> (as in the physical pins)
[00:25:22] <andypugh> You will need an enable button of some sort (probably to a hm2_5i20.0.gpio.NNN.in pin
[00:25:56] <andypugh> (Again, see what is available as GPIO in the dmesg list of hm2 pins)
[00:27:29] <andypugh> Then "net" the enable button (possibly inverted) to the pin axis.0.jog.enable and the counts to .....
[00:27:32] <UncleG> I increased the encoders to 4, I do not know how to figure out what pins to use
[00:27:53] <andypugh> You are going to want one wire one encoder to jog any axis aren't you?
[00:28:06] <UncleG> Yes.
[00:28:51] <andypugh> If you start emc with the 4 encoders then type dmesg in the terminal you will get a listing of what every pin on the 5i20 is configured as
[00:29:41] <UncleG> well, there they are.
[00:30:00] <andypugh> You will need to choose a GPIO pin for each jog-enable, then physically wire to either a pushbutton or roatary selection switch.
[00:30:32] <andypugh> the magic incantation to connect the encoder to the axis jogs is:
[00:32:05] <UncleG> alright, I am way behind right now.
[00:32:15] <UncleG> Im trying to comprehend what I am looking at in dmesg.
[00:32:24] <andypugh> net jogwheel hm2_5i20.0.encoder.03.count => axis.0.jog-counts axis.1.jog-counts axis.3.jog-counts
[00:32:59] <andypugh> Look at the very end of the dmesg output
[00:33:21] <UncleG> for?
[00:34:13] <andypugh> Somthing like:
[00:34:14] <andypugh> [ 1177.539422] hm2/hm2_5i23.0: IO Pin 000 (P2-01): Encoder #1, pin B (Input)
[00:34:15] <andypugh> [ 1177.539430] hm2/hm2_5i23.0: IO Pin 001 (P2-03): Encoder #1, pin A (Input)
[00:34:25] <andypugh> And so on.
[00:34:38] <andypugh> There ought to be one for every pin on the 5i20
[00:35:39] <andypugh> It is only there if you have started emc2 successfully though.
[00:36:26] <andypugh> Alternatively:
[00:36:47] <UncleG> [ 5990.954510] hm2/hm2_5i20.0: IO Pin 012 (P2-25): Encoder #3, pin B (Input) [ 5990.954513] hm2/hm2_5i20.0: IO Pin 013 (P2-27): Encoder #3, pin A (Input) [ 5990.954516] hm2/hm2_5i20.0: IO Pin 014 (P2-29): Encoder #2, pin B (Input) [ 5990.954519] hm2/hm2_5i20.0: IO Pin 015 (P2-31): Encoder #2, pin A (Input) [ 5990.954522] hm2/hm2_5i20.0: IO Pin 016 (P2-33): Encoder #3, pin Index (In
[00:37:03] <UncleG> encoder #3 = my 4th encoder correct?
[00:38:05] <andypugh> Indeed. And the pins are on P2 (labelled on the silkscreen on the card) pins 25, 27, 33
[00:38:40] <andypugh> All the even pins are 0v / gnd / earth
[00:39:57] <andypugh> The odd pins (the ones you need) are the ones on the side of the header which has the alignment notch. (I lost a few hours on saturday to getting that part wrong)
[00:40:50] <UncleG> so, on 50pin terminal #2 of the 5i20, pin 25=A 27=B 33=Z
[00:41:35] <andypugh> That same section will list unalloctated pins as GPIO, you can define those as inputs and optionally invert them. More details here: http://www.linuxcnc.org/docview/html//man/man9/hostmot2.9.html
[00:41:45] <andypugh> In the General Purpose IO section
[00:42:13] <andypugh> The HAL command to, for example, set GPIO 56 to input, and inverted would be
[00:43:58] <andypugh> Ah, forget that last part. They default to inputs, and have both a an "in" and "in_not" pin in HAL.
[00:45:09] <andypugh> so there would be two input pins for GPIO 56 called hm2_5i20.0.gpip.056.in and hm2_5i20.0.gpip.056.in_not which you can use in "net" statements in HAL.
[00:45:29] <UncleG> See, I can understand that, this 10 minute conversation completely gave me some understanding of hal.
[00:45:44] <UncleG> I have a goal for you, rewrite all the documentation
[00:46:17] <andypugh> I wrote down a summary in a forum posting a few days back, rather than retype it, have a look here: http://www.linuxcnc.org/component/option,com_kunena/Itemid,20/func,view/catid,16/id,3535/lang,english/#4973
[00:46:47] <andypugh> It is about e-stops, but explains how to connect signals to each other in HAL.
[00:47:56] <andypugh> note the type above, hm2_5i20.0.gpip.056.in should be hm2_5i20.0.gpio.056.in
[00:48:28] <andypugh> Right, time to log off, pushing 1am here. (feels like 2am(
[00:48:40] <UncleG> Okay, I need to go about poking around with said pins and dmesg's and what not
[00:48:54] <UncleG> I thank you again.
[00:48:56] <UncleG> or
[00:49:09] <UncleG> for the first time
[00:49:10] <UncleG> :)
[00:50:12] <geo01005_> geo01005_ is now known as geo01005
[00:57:32] <Mendelbuild> Mendelbuild is now known as self_replicating
[07:47:58] <Guest710> hello everybody
[07:49:42] <micges> hi
[07:50:48] <Guest710> i want to know how to enter a jpg pic to AXIS , i need the procedure please if you can help me
[07:51:24] <micges> you mean process jpg to gcode?
[07:51:46] <Guest710> yes
[07:53:14] <micges> you must ad there lines to your ini file:
[07:53:18] <micges> [FILTER]
[07:53:20] <micges> PROGRAM_EXTENSION = .png,.gif,.jpg Grayscale Depth Image
[07:53:28] <micges> jpg = image-to-gcode
[07:53:42] <micges> thats all
[07:54:14] <Guest710> and then any picture will be interpreted? or i need something else?
[07:55:21] <micges> yes then if oyu open jpg file in Axis, image2gcode convrter will appear and it will generate gcode
[08:00:58] <Guest710> im new to all that well im working on a graduation project using this software is there an example of that procedure in the software that would help me giving the command right to the servo motors?
[08:02:29] <Guest710> by the way i will register an account for further working on this software
[08:02:52] <micges> cool
[08:03:04] <mrsunshine__> frecking stupid to do grinding without goggles :/
[08:03:17] <Guest710> yeah
[08:03:25] <mrsunshine__> "oh well its a fast job" .. well stuff in my eye and now it itches and stuff :/
[08:04:01] <micges> Guest710: I use mesanet pci card to out voltage to servos
[08:04:39] <Guest710> is it simpler than the pwm controller?
[08:04:45] <micges> sample config for this: http://git.linuxcnc.org/gitweb?p=emc2.git;a=blob;f=configs/hm2-servo/hm2-servo.hal;h=9a0b5903a9bccc78fd52a522d2997c66b65488a2;hb=6bd1cd502b8cdfdc97d3ece9e094c405b52468e6
[08:05:07] <micges> via parport?
[08:05:21] <Guest710> yeah
[08:05:39] <micges> yes it is simpler, but expensive
[08:05:54] <micges> servo via partport will do
[08:06:12] <micges> but I think only one will fit one lpt port
[08:08:28] <micges> you must at least connect: pwm, pwm dir, enc a, enc b, enable
[08:09:42] <Guest710> my work will be engraving on a piece of awood a simple jpg to enter, process to gcode using 3 axis machine
[08:10:25] <micges> what electronics/motors?
[08:11:57] <Guest710> 3 servo motors for axis and 1 dc for the engraving tool
[08:12:35] <Guest710> im adi sorry i didnt introduce myself
[08:14:35] <micges> hi adi
[08:15:03] <micges> it will be complicated but possible to use servos with lpt ports
[08:15:34] <micges> better solution are some servo cards but they aren't cheap
[08:16:41] <micges> http://www.linuxcnc.org/docview/html/
[08:16:58] <micges> there is section hardware drivers
[08:17:26] <micges> with few servo cards with drivers for emc
[08:17:39] <micges> but I have only experience with mesa/hostmot2
[08:18:03] <micges> ok
[08:18:43] <micges> what control signal your servo have?
[08:18:48] <micges> +-10V ?
[08:29:34] <Guest710> i didnt check that
[08:29:38] <Guest710> it might be
[08:35:33] <micges> bbl
[08:57:26] <Guest710> micges are u there?
[16:24:18] <kb8wmc> g'day to all
[16:24:31] <atmega> whuttup G
[16:25:03] <kb8wmc> how u atmega?
[16:25:42] <atmega> jus chillin with my emc homiez
[16:26:19] <kb8wmc> rgr that, I feel great about getting the Axis program to now load properly
[16:26:44] <atmega> was that a problem before? local display?
[16:27:08] <kb8wmc> it had to do with the graphics
[16:27:27] <atmega> and weird proprietary video chipsets?
[16:27:29] <kb8wmc> I changed the lib and all went well
[16:28:20] <kb8wmc> no, don't think it was necessarily a proprietary chip problem, more to do with the openglx and the ati card
[16:28:32] <kb8wmc> just would not play well together
[16:29:25] <kb8wmc> on startup of Axis, it locked up the entire system, excepting mouse, which was changed to a black round circle for pointer
[16:30:15] <atmega> that's a feature. The lock up was added to make windows users feel more comfortable
[16:30:31] <kb8wmc> I was hoping to see cradek in here, I ran across a "lisp" code for ACAD that it appears he is more than familiar with
[16:32:09] <atmega> realize?
[16:32:25] <kb8wmc> hey there you are, yes
[16:33:04] <atmega> then he's the guy
[16:33:05] <kb8wmc> I just ran across it and am doing some reading of it right now
[16:33:10] <atmega> lisp makes my eyes bleed
[16:33:17] <kb8wmc> how so?
[16:34:49] <atmega> functional programming is not my favorite
[16:35:29] <kb8wmc> ah
[16:36:03] <kb8wmc> I wanted something that I could possibly use with acad directly
[16:36:47] <atmega> I looked at it, seems nifty... then I bought a copy of cut-2d
[16:37:01] <kb8wmc> rgr that
[16:40:01] <cradek> you aren't one of those people who keeps emailing me are you?
[16:40:13] <atmega> heh
[16:40:15] <kb8wmc> hey cradek, no sir
[16:40:38] <cradek> "I saw your prgm. I wondr if it does the same thing as [commercial product I don't care about]. Any advice is apreciatted."
[16:41:09] <cradek> (that's a paraphrase)
[16:41:39] <kb8wmc> hmmm
[16:42:00] <cradek> (it's free - it's right there - instructions are one page long - why not just try it?)
[16:42:46] <kb8wmc> yep, I just finished copying the code to my puter and will put it in my cad files
[16:45:34] <cradek> great
[16:45:47] <kb8wmc> tnx for your endeavors
[16:46:44] <kb8wmc> I am an old guy that still does many things the old-fashioned way, kind of old dog, new tricks...lol
[16:48:03] <kb8wmc> I figure if I can generate g-code right from acad rather than another go-between program, I would prefer it
[16:48:27] <cradek> yes for simple stuff it's awfully easy that way
[16:49:25] <kb8wmc> I don't have too much time left to cram one more program into my feeble mind, learning curve ya know...
[16:50:31] <kb8wmc> and tnx again for your help on getting EMC2 problems resolved the other day
[16:51:07] <atmega> for tooling offsets, you just offset the original poly and put that on the cut layer?
[16:51:38] <cradek> yeah
[16:51:45] <cradek> autocad does offsetting fine
[16:56:19] <atmega> yep.. my autocad class has been quite useful, if not painfully slow
[16:57:42] <kb8wmc> I have used autocad since it was all command line
[16:57:47] <kb8wmc> lol
[16:57:57] <atmega> we are starting the 'advanced' part tonight. It will probably be painful also.
[16:58:20] <kb8wmc> hope not atmega
[17:02:26] <atmega> I'm old... everyone in the class is older
[17:02:37] <kb8wmc> what age are ya
[17:02:51] <atmega> 0x2f
[17:03:45] <kb8wmc> I am 63
[17:03:51] <atmega> damn, you are ancient.
[17:03:54] <atmega> :)
[17:04:05] <kb8wmc> duh, I beg your pardon....lol
[17:04:17] <atmega> I SAID... YOU ARE...
[17:04:58] <kb8wmc> yeah, I heard ya....
[17:05:27] <atmega> last autocad I did was 1989? v2 maybe? I wasn't any good at it then
[17:06:23] <atmega> I've been programming for 25+ years and can generally pick up any SW package easily... but cad stuff didn't go so well
[17:06:53] <kb8wmc> in 89 it might have been around version 10 or so, if it was contemporary
[17:07:04] <kb8wmc> rgr that
[17:07:16] <archivist_emc> acad was the worst, you can pick up and use solidworks in minutes
[17:08:23] <atmega> after my edu acad license expires, I won't be using it anyway
[17:08:40] <kb8wmc> I wanted to be a systems analyst when I got out of service so I went back to school in 1969, I did learn to do some programming in certain languages of the day, COBOL, BAL, RPG
[17:09:45] <kb8wmc> I have been on the fringes of programming over the years, not a programmer though
[17:11:37] <kb8wmc> I can appreciate the efforts of other's programming skills
[18:17:28] <skunkworks> http://www.machsupport.com/forum/index.php/topic,16365.msg110755.html#msg110755
[19:00:24] <skunkworks> Kobbled in a jog wheel yesterday. (hooked into the 7i48 on the 2nd 5i20 card) Pretty cool jogging the machine around. :)
[19:06:39] <skunkworks> dad almost has the b axis servo finished. at 24v it will take about 15 seconds to go full circle.
[19:08:53] <skunkworks> looks like the axisui outputs jog select pins for each axis - so I do not need an external axis select switch for now.
[19:10:17] <skunkworks> wow - emc is awesome.
[19:10:23] <skunkworks> sorry.
[19:22:02] <skunkworks> everything takes a bit of adjusting http://www.electronicsam.com/images/KandT/conversion/spindle/toolarmadjustment.jpg
[19:24:17] <atmega> that's a rather large machine
[19:24:21] <JT-Work> I don't see a BFH anywhere around to do the adjusting with
[19:24:35] <skunkworks> heh - it was probably pretty close... (rubber hammer)
[19:25:09] <skunkworks> the previous owners (trane company) only seemed to have a chisel and hammer for tools.
[19:25:36] <skunkworks> atmega: yes - it has a nice work qube.
[19:25:41] <skunkworks> cube.
[19:26:41] <JT-Work> don't you hate to see how some people abuse equipment like that...
[19:27:24] <skunkworks> yes. all castings are pinned and have nice threaded pull holes. I don't think they were ever used.
[19:28:16] <skunkworks> even the metal covers had threaded pull holes.
[20:09:25] <n2diy> I'm exploring making my own PCBs, from schematic capture, to gerber file generation I have covered with gEDA, now I need a CAM program to talk to EMC2 correct? Recomendations?
[20:09:55] <skunkworks> we are actually thinking about machining some parts... (we have to mill a hex in the center of a timing pulley for rigid tapping encoder) (1-1/2"hex)
[20:10:54] <cradek> how sharp do the corners have to be? can you just drill holes there first?
[20:11:29] <skunkworks> they are pretty rounded. I think 1/8 or 1/4 mill will work
[20:13:04] <skunkworks> http://www.electronicsam.com/images/KandT/conversion/spindle/belvel.JPG
[20:14:10] <skunkworks> the hex that is just below the edge of the horizontal casting
[20:14:36] <cradek> the smallest one?
[20:14:58] <skunkworks> yes - thank you. I was wondering how I was going to explain which one)
[20:15:03] <cradek> ha
[20:15:12] <cradek> nice, looks like the corners are quite flat
[20:15:18] <skunkworks> yes
[20:16:04] <skunkworks> then we can do a video rigid tapping a 1 inch tap.. ;P
[20:16:23] <cradek> sweet
[20:16:49] <cradek> (you have a 1" tap??)
[20:17:11] <skunkworks> heh - yes
[20:17:51] <skunkworks> 1 inch studs are what holds the k&t into the concrete.
[20:18:20] <skunkworks> the studs go about 2 feet down into metal plates that are tapped.
[20:18:53] <skunkworks> I wonder if dad has any pictures of the pour.
[20:47:52] <n2diy> I'm exploring making my own PCBs, from schematic capture, to gerber file generation I have covered with gEDA, now I need a CAM program to talk to EMC2 correct? Recomendations?
[20:49:37] <atmega> pcb2gcode
[20:49:38] <skunkworks> I have only done eagle -> pcb-gcode or gcode script ->emc2. You need to convert whatever file format you have to gcode using some sort of cam.
[20:50:46] <skunkworks> n2diy: have you looked at the wiki? http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Cam
[20:51:24] <n2diy> skunkworks: well, if pcb can generate the gcode, I think that covers it? No, I hadn't seen that wiki, let me go take a look, thanks.
[20:53:25] <kb8wmc> cradek: if I am understanding correctly Realize requires "Heavy" polylines, am I correct in this understanding?
[20:54:02] <cradek> yes, explained here: http://timeguy.com/cradek/autocad
[20:54:08] <n2diy> skunkworks: yes, I have seen that wiki, I have it open in another tab!? So, there is a lot to choose from, I was hoping there might be a consensus on which one is preferred, has good support, etc..
[20:54:49] <kb8wmc> cradek: thanks...
[22:03:24] <self_replicating> self_replicating is now known as mendelbuild
[22:48:31] <skunkworks> http://www.youtube.com/watch?v=p7DTZ8vxqZU
[23:17:21] <mikegg> any of you guys use a haimer taster?
[23:18:00] <fenn> a what??
[23:18:47] <mikegg> one of these http://www.haimer.de/usa/taster-universal.php
[23:18:59] <mikegg> was reading on the 'zone
[23:19:09] <mikegg> some guys swear by em apparently
[23:20:26] <fenn> sounds like a non-issue to me
[23:20:29] <skunkworks> fenn: !
[23:20:39] <fenn> hi skunkworks
[23:20:52] <skunkworks> how is it going?
[23:21:21] <fenn> i may be getting an X3 soon
[23:21:36] <skunkworks> nice! - good size.
[23:21:44] <fenn> building repraps, moved to san francisco
[23:21:51] <Connor> hey guys. What CAM software do you all use? I have Inventor 2008..
[23:24:37] <skunkworks> fenn: wow - quite the change. how are the rep-raps doing?
[23:26:13] <fenn> self-destructing, mostly
[23:27:01] <fenn> these robots, they don't build themselves ya know
[23:27:59] <skunkworks> heh
[23:28:19] <fenn> mikegg: you know about emc's probing move feature right?
[23:28:55] <fenn> used with an electronic probe it would "eliminate math errors" like the taster, with a few less button presses