Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2008-03-13.txt
fenn: (oh keeper of the emc wiki cad list) some cad functions in python for blender http://utenti.lycos.it/paoloprovenda/PROCAD_17.py
and instructions (in Italian) http://utenti.lycos.it/paoloprovenda/PROCAD17.pdf
hrm. if it were html i could googolate it
anyway, i'm still not convinced that blender will ever make a good cad program
is there even a way of drawing an arc or line in Blender? Just a simple line would be nice.
hey people. I made some documentation of a custom card that interfaces with the M5i20. its at http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?M5i20_Laser
nuffin much to it.
though it seems like it might have been easier (maybe) to do a custom FPGA config instead of using an AVR on the daughtercard
thought it might be useful to other people who are more at the elementary level.
yah that'd be great. I thought only EMC only loaded MOT-4 or whatever its called.
well, that is mostly true :)
note emphasis on 'elementary' :-)
when I have the chance (probably in a month or two), I'll clean up and post the FPGA code to drive my analog boards, plus the HAL driver source for that 5i22 config
that'd be nice.
I probably need to take out some customer-specific stuff before I can post them, which is why I haven't yet
actually i thought of blender not for cad, but for machine visulatization (like vismach)
fenn_ is now known as fenn
fly cuttin some beaznuss
alex_joni: how is it going with 2.2.4 release ? are you working with it too ?
micges: I must admit that not really
I've been way too busy lately, to find time for it
what you are doing in job ? (If I may ask)
oh, a lot of things..
but usually I work with welding robots
programming or using ?
programming and servicing
and selling :P
* archivist offers £1 for a robot :)) whats my chances of getting one?
archivist: maybe for a kitchen-robot
I think the only way I would get one is make it myself, with my current wages
how is motion transfered to most robots arms?
is it planetary gears?
(going to work)
Hi guys; its nice out today, 63F, lots of sunshine for a change
I just made a discovery that reduces the utility of the charge_pump signal
Running the axis logo, and hitting the keyboard esc key for an e-stop doesn't kill the charge_pump, on;ly the F1 off does.
Is there a command I can add to the halfile that will effect that?
the charge pump is meant to show the hardware that the software is alive and ready to run
which it is, if it's out of estop
unforch, the esc key doesn't stop it
no, it just stops the running program
it's "abort", not "estop"
(even leaving out the question of whether a PC key should ever be called Estop :) )
Oh, I thought it did stop everything at one time back up the log...
I don't think so. it maybe should stop the spindle, but not the charge pump
realize that on a servo system, the charge pump would likely control servo drive power, which may also mean it controls encoder power, and you don't want to lose that when you stop a program
I agree, and before I'm done with this, there will be a couple of those e-stop switches they use on the mini lathe hooked up here. But they havn't arrived yet from LMS
Now, I have another question, I not that with the ratios set to .2 and .8 in the stepconf's last setup page for spindle, I can hit the plus key until it locks pwmgen up at 100%, and the PMDX-106 of course drops out at that pointt, leaving it here a sing;lle tap of the - key would restart it at full speeed
have it miss-configured that some how?
that may be a Steve S question
I don't know if the PMDX-106 can handle 100% duty cycle on the PWM
It is set for ignoring the charge_pump ATM, and without pulses, that is a stop signal.
charge pump and spindle PWM are separate though ...
it doesn't care about the logic level, just the presence of the pulses.
for the spindle PWM, or for the charge pump?
treu, I'm just trying to figure out the interactions here.
Let se set it to use the charge_pump and see if it still stops at 100%
the charge pump should actually control all of the outputs, not just the spindle drive
sh--, can't, taint hooked up at the xylotex output yet.
(though I guess I should look at the PMDX-106 manual before saying that :) )
I have, and its not described quite as clewarly as I'd like. Could be my grokking it though.
argh - phone
invention of the devil, particularly when the politicians are 'in season'
worse - mother
ok, the manual does say that the duty cycle should be between 5% and 99%
so we need a limit to preevent 100% somehow...
resarting the stopped motor will probably clear the fuse
also, from the reading I'm doing (while somewhat listening to my mother), it looks like the PMDX-106 doesn't actually havea charge pump on it
no, it detects it from emc
um - one second here :)
EMC outputs a square-wave (or something like it) using its "charge-pump" component
presumably, there's an external charge pump safety circuit which will shut things down if that signal is lost
if there is no such circuit, then the software component and squarewave output are unnecessary
and that is what is in the PMDX-106
the PMDX-106 manual says that it has a charge-pump interface for use with the PMDX-131 and others
The wqveform is pretty fast, almost too fast for the halscopes rise & fall time
its a flat topped triangle
I'd ask Steve S, but it looks to me as though the 106 has no charge pump
gene: do you have a PMDX-131?
SWPadnos: otherwise it's hard to connect the 106 to anything
it will pass a software signal through to another board, and can disable with the CPOK input (which would come back from a 131 or 122)
looks to be at base thread speed for its rep rate. 100khz in other words, no, hooking direct to the parport breakouut in the xylotex
alex_joni, it looks like it has screw terminals for a lot of it
It does, and that is what I'd using
gene: it only has a charge pump ok input
ok, in that case I'm pretty sure ypu have no charge pump
so it has to rely on another charge pump component\
I don't think the xylotex has one?
its setup and connected to parport 17, which is on the breakout terms of the xylotex
gene: the PMDX-106 doesn't hava a charge pump component on it
if you're feeding a chargepump signal to it (square wave), all you accomplish is that you confuse it
it doesn't generate it, it detects it
no, it does not
gene: please try to read what I wrote
forget the name chargepump for a second
then why is the dip switch even there for that?
because you can connect the PMDX-106 to a PMDX-131 or PMDX-122
the 131 and 122 have logic on it for a chargepump detector
and they also have some logic which can notify the 106 that the chargepump is ok
and the 106 can disable its outputs if that input tells it that something is wrong
think of it like a daisy chain.. once the 131 notices that the CP signal is not ok, it tells all connected boards (like the 106) to switch off
it would be better if we called the signal from the PC something like "heartbeat"
so it doesn't detect the pulses, only the logic level?
which is useless if coming directly from the PC
because the PC can get stuck on either level..
dump that idea then unless I make it part of the e-stop I have yet to build
a charge pump is a circuit that will detect the loss of a heartbeat signal, and the output from that is generally used in an estop chain
Now, how do I prevent the pwm from going to 100%?
well, the simple way is to add a line to custom.hal:
setp pwmgen.0.max-dc 0.98
ok hang a few
gene: I agree that the 106 manual is a bit confusing
page 16 at the bottom is not that clear about the charge-pump
it seems pretty clear in the table
it says that the 131 or 122 is used, and it outputs an "OK" signal to the 106
or set the 106 to ignore it :)
SWPadnos: shouldn't have said parport pin 17
it's assuming that the user has a 131 or 122 quite a bit
table 4 is *way* clearer on this
That works, now I can dance on the + button forever
Yes, I see that now
I'd assume I could setup a net in hal to duplicate that?
I think even max-dc 0.99 should work :)
without the dither, maybe...
you gain 1% of speed with that..
they say 99.9% pwm is ok
only 100% is considered off
Thats adjustable, both in the VSR, and in the 106 as it has range pots for that, and I can give the motor 100% at .8 from pwmgen
so I don't see that as a problem.
I thought it said 99%, somwhere toward the beginning
I haven't been watching the whole conversation, but I think that in terms of emc's software pwmgen the important thing is that at least one BASE_PERIOD is off out of every output cycle of the pwmgen. If there are only 25 BASE_PERIODs, then values above (24.5/25)*100% might give all 25 periods on, and zero off, dependingdepending how rounding is done inside pwmgen.
I do see that the first click looks like a bit of a speed jump though. It should maybe be less of an increment, but auto-repeating as long as the click is held down?
gene: I think that's a feature of the 106
the PMDX-106 manual also says that 5% is the minimum, so it may do nothing until you get to 5%
and it discards a couple of toggles, until at least 3-4 are consistent
then jump to whatever you set it to, like 8% (if you had two, 4% increments)
And I think I set it at the default this time, .2 and .8
Nother silly Q: if i put, in the custom.hal, the pwggen type=1, will that override what in the maqin hal?
the step type needs to be known at load time
but you probably don't need the type 1 anyway
so I have to put that in everytime I rerun stepconf..
I think all of the direction stuff can be found in HAL - there's a spindle-fwd and spindle-rev signal
it takes the output_type = 1 to give me the reverseable spindle controls
you can use the absolute value of the spindle speed forthe PWM input
gene: then I guess you're better off by not using stepconf anymore :)
it would take some extra stuff in custom.hal, but I'm pretty sure it can be done
in gcode, na MS-100 would run reverse without that if I hace pin 14 set for CCW?
damn, not looking at the screen while I type grrrr
in gcode, an "MS-100" would run reverse without that if I have pin 14 set for CCW?
what is an MS-100?
oh, nevermind :)
I don't think so
whatever the spndle commands are, not fam since I've not used them till now.
reversible spindle isn't supported in stepconf, despite the ability to chose both CW and CCW outputs.
M3 and M4 are spindle forward vs. reverse codes
gene: applications/cnc/gcode quick reference
if you want a reversible spindle, remove the spindle outputs in stepconf and put the setup that is correct for your machine in custom.hal
humm, the one page I can find says its an error if the S number is negative, and I haven't found the M3-4 stuff yet, still looking
that's why I like to start with the gcode quick ref. you can click any gcode for more details
found it, so I don't really need the type=1 then?
I don't think so
let me check
you'll need to connect spindle-fwd (or spindle-rev, depending on the polarity of the PMDX-106 input) to the direction output
I am pretty sure that the hal value spindle-speed-out will be negative in one direction and positive in the other.
if you want a speed-controlled reversible spindle, remove the spindle outputs in stepconf and put the setup that is correct for your machine in custom.hal
I have that, spindle-ccw on parport 14, and it looks like I still have both buttons?????
that's probably the best bet :)
reversible, speed-controlled spindle isn't supported in stepconf, despite the ability to chose both CW and CCW outputs.
18:41 < alex_joni> gene: then I guess you're better off by not using stepconf
I know, as you indicated in the email :(
I'm not saying you can't use stepconf at all -- but you're going beyond stepconf's abilities with respect to the spindle control.
you can take all the spindle stuff out of stepconf, and just do that part in custom.hal
that's why custom.hal exists -- it usually lets you use stepconf for the portions it can do, and then add the rest
stepconf is not and will never be able to configure all possible setups. that's not its goal.
That and fix the one cable I have that needs a - sign in front of the screws tpi to make it run the right direction... :)
I just checked and with output_type=0, I still have both buttons (I didn't have, only the fwd button before)
Is that because I asked for the CCW output signal that I get both buttons now?
buttons in axis appear when you have certain pins connected to signals. it's immaterial if those signals actually do anything
ahh, so asking for the CCW gets me the other button.
connecting spindle-rev to anything gets you the button :)
gotcha. I'll get this figured out yet if you don't loose patience with me ;)
too late. err, I mean, I think I'll get more coffee :)
but sub spindle_ccw for spndle_rev I think
oh, I don't know. use halcmd to see the actual motion.whatever pin names :)
Yeah, I ned something to eat too, bbl
any opera fans in here?
no, I hate opera
too bad ;)
[17:11:26] <alex_joni> http://www.youtube.com/watch?v=NGsYrfqV9SQ
quite a decent voice, without majour training :)
That is my white font.
skunkworks_: you are assuming things again
*(my background is white.. )
not everyone has a white background :)
Heh - just thought about it.
maybe you could say it's your transparent font :P
the font is auto-learning - it matches what ever your background is..
That would have been better ;)
like a cameleon
I was up until 2:00am grouting the countertop tiles.
[17:52:35] <SkinnYPuppY> http://www.cnn.com/video/#/video/us/2008/03/12/woman.scalped.by.gocart.ksl
is this something i want to watch?
a kid at my high school stole a go-kart and went on a joy ride at night. they took the roll cage off so it would go faster. he never saw the wire cable across the path...
[17:55:27] <skunkworks_> http://www.cnczone.com/forums/showthread.php?t=54410
This is a don't get yer hair in the lathe video
oh well, apparently firefox isn't cool enough for CNN
Anyone have any comments on my email to the developer list about when to commit changes?
any time if you make a branch
any time it "seems to work and not break stuff" on trunk, I'd say
and only for bugfixes on a release branch
lerman: I would say (after reading your email) go ahead and commit it
it will only prove if hidden bugs still lurk
Thanks for the feedback. I'll run a few more tests and then commit it.
lerman: if you have some gcode test programs, it's easy to add regression tests
then you could run them before and after your changes.
I don't have any real good tests (although I generally run flowsnake). Is there some easy way to run them without using a gui?
yes that's what the test infrastructure does
I didn't know we had one. Tell me more, please.
if flowsnake is currently generating the right output, make a flowsnake test
look at the tests directory in the toplevel
% runtests .
Running test: ./lathe-comp
Running test: ./mill-g90g91g92
Running test: ./mill-line-arc-entry
Running test: ./mill-zchanges
Runtest: 4 tests run, 4 successful, 0 failed + 0 expected
these are gcode programs that exercise cutter compensation for mill and lathe
do you still have ed nisley's program that helped us find problems introduced in (iirc) 2.1?
that was a pretty extensive program
Not likely that I would have it.
I have ed.ngc which is a minimal test case he sent, but I don't have the one I'm thinking about
aha! I do have ed's complex program
Are you going to keep it a secret?
I'm adding it as a test
Way to go.
How do I get it into my tree? cvs update in the proper directory?
yes a regular update will get it
I'm sure it doesn't test everything, but it's fairly complex.
I just ran tests and most failed. I assume that's because I'm running in a simulator environment.
no, they should work
What is the new test called?
lerman: Runtest: 23 tests run, 22 successful, 0 failed + 1 expected
the tests all work right on my trunk/sim build
Runtest: 23 tests run, 22 successful, 0 failed + 1 expected
(the one that "fails" is actually testing that part of the testing script works properly)
lerman: the test only works if you used configure --enable-run-in-place
I am running in place.
Runtest: 22 tests run, 5 successful, 16 failed + 1 expected
and presumably sourced emc-environment ...
SWPadnos: actually it looks like you can get away without doing that
lerman: what's one of the failures? maybe there is one shared cause
I suppose I could try it on my sim /RIP "install" on 7.10
look at the files "stderr" and "result" in the directory with the failure
I did my update and didn't get the new file. I did 'cvs update'
cvs update -dP
you need the new directory
[18:55:07] <jepler> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?CVS#cvsrc
Still get a flock of failures but the new one passes.
let's figure out why those tests are failing, then. look at tests/abs.0/stderr or put it on pastebin
maybe there's a smoking gun, like an error that it can't start the realtime environment at all
before chris's addition, there were 5 gcode tests and the rest were tests of realtime components -- curiously, that's the same split you got between successes and failures
Precisely. stderr was empty.
what's in 'result'?
Realtime already running. Use 'halrun -U' to stop existing realtime session.
there you go -- do you have emc running? if so, shut it down before running the tests.
I guess I should stop the one that's running upstairs. I'll be right back.
if it's executing on the same machine you're using for tests, yes.
or you can just remember this for the future. you can run a single test by specifying it on the runtests commandline -- e.g., runtests tests/ccomp/mill-zchanges
Runtest: 23 tests run, 22 successful, 0 failed + 1 expected
that's much better
this is with your new interpreter changes?
Yes, it is.
cool, that's a good sign
I'm very impressed. By the test infrastructure. I'll commit those changes with confidence.
I doubt that program tests everything, but it's somewhat complex
If we find any bugs, I'll blame the test program. :-)
haha, we'll blame you for not making a test that finds it
I see that sub.0 passed. So, I can mark that bug as fixed.
what as wrong with sub?
Something about parameters being messed up when a second level subroutine was called.
oh does sub.0 test that?
It claims to.
I don't think I ever saw the bug.
N..... STRAIGHT_TRAVERSE(1.0000, 2.0000, 3.0000, 0.0000, 0.0000, 0.0000)
N..... COMMENT("#1 #2 and #3 are now corrupted upon return from dummy subroutine")
N..... STRAIGHT_TRAVERSE(1.0000, 2.0000, 3.0000, 0.0000, 0.0000, 0.0000)
looks right (I think?)
The changes have been committed.
Hey, wondering if anyone can help a new guy with what I hope is a simple question?
we'll only know if you ask your question ...
Sure. The answer is 42. Now tell me the question. :-)
I think that was already the simple question
asking to ask a question is useless due to recursion problems
ok...I got Ubuntu installed, got EMC installed and was playing around with axis before I hooked up my machine
Now...I cannot use any other profile other than axis
when I try to load anything else it just loads right into axis again
please explain exactly what you mean by "try to load anything else"
ahhh - runtests sources emc-environment for itself
all that stuff works on my sim/7.10/trunk
Ok..I'll try as best I can, new to EMC so not sure if I'm using proper terminology here
no problem, just say what you are doing, and what you expect vs. what you see
I loaded EMC up with Axis Simulator to test it out...I am trying to load EMC up with a profile that is not a simulator
I've got it hooked up to my mill now and I'd like to actually control the machine as opposed to just running a simulator
axis is not itself a simulator, it is a front end to EMC2, regardless of whether it is running in simulation mode or in "realtime" mode
Ok, well I've been using it as sim...It will not control my mill
I tried using one of the different "front ends" and it goes right back to axis and doesn't load the other ones
ok, sorry for the interruption. can you tell us how you are running EMC, and what you do once you have run it?
on the config chooser, you can pick (for example) sim/axis, or stepper/stepper_inch
sim means it does not control hardware
I tried using stepper_inch
the rest of the sample configs are for various kinds of hardware
it won't load up
ok we're getting somewhere
what does it do?
loads right into axis
ok we've stopped getting somewhere :-)
are you running from a desktop icon?
have tried both
AXIS will run in sim mode, or control hardware, either way
I think for some reason I am stuck in sim mode is what I'm saying
I understand but we need more information
if you are running emc from the menu, and then picking stepper/stepper_inch from the config chooser, you are NOT in sim mode
if your machine isn't set up exactly like the stepper_inch configuration, it may not do anything. that would look a lot like sim mode even if it is in realtime mode
please say what is not working, or what you see wrong, instead of saying what your conclusion is ("stuck in sim mode")
for instance, if your stepper drivers require an "enable" before they will turn on and move the motors, stepper_inch probably isn't providing that enable output, so nothing happens when you press the keys to jog.
then maybe we can help
ok lemme see if I can phrase it better
before you do that ...
can you give us a step-by step account of what you do and what you see?
sure...will take a minute, I have to go to the controller computer and do it...
(ie, click emc menu item, config pickjer appears, double-click stepper/stepper_inch, axis screen appears ...)
err - ok
this is why all computers should use command line interface :)
Ok, sorry bout the wait guys, I'm on the controller computer now
heh - that should make some things easier :)
from the main Ubuntu desktop I click the applications tab in the upper left...click cnc -> emc2.. this brings up the config selector
from here I use stepper -> stepper/inch, emc2 loads up and the config in use is axis
sounds good so far
ok, there's the problem :)
the user interface is AXIS, the config is stepper/inch
there's also the sim/axis config, which also uses the AXIS interface
Ok, I gotcha on that...
basically, you can choose your user interface separately from the type of machine (or simulator mode)
so then I have some other problem that is not letting my machine work...
many of the configs use AXIS, because it's pretty and very functional
yep, that's the likely issue :)
do you have all the information about what connects where on your machine?
Ok, well...I know the machine, steppers, controllers, and power supply work and are functional. They work fine under Mach, trying to get them to work in EMC though
emtffke1: if you want to verify that you've loaded a configuration that is intended to control hardware through the parallel port, you can choose Machine > Show Hal Configuration in the menu in AXIS. Then click the "+" next to "Components". You should see hal_parport shown in the list of Components.
now hang on, you've skipped information again, jumping right to "not letting my machine work"
sorry...machine powers up, motors lock, all the good power up stuff, but...i get no control input from EMC
what are you doing once the AXIS screen appears?
Well I toggle the E-Stop off, Toggle the machine power on, and then from there I'm a bit lost...I tried jogging and no movement, I tried running the EMC logo g-code that loads up automatically and nothing there either
you should put a meter on the parport pins
This might be a step worth trying: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?TroubleShooting#Parallel_port_no_longer_works_in_EMC_2_0_1_or_later_emc_starts_but_motors_don_t_turn
do you know the pinouts of your hardware? like what parallel port pins are step/direction for the various axes?
EMC appears to start, but no signals ever appear on the parallel port (the motors don't turn).
To fix this problem, upgrade to EMC 2.0.3 or newer. Then, add the command
before the line loading hal_parport in your hal file.
if your step and direction pins are reversed it won't work
(but cradek is very right that it's also important to verify that your pinout hal file is correct)
yeah, I configured the pins
I have not seen any sign that you made sure the signals/pinouts are configured right
how did you do that?
emc stepconf wizard, updated existing conf
a config generated by stepconf will not be called stepper/inch
it will have the name you gave it
it will show up under "My Configurations" in the chooser
maybe thats the issue, when it said update existing config I assumed it updated the one you selected
I'm not sure what config you would update, since stepper/stepper_inch is not a config that was generated by stepconf
stepconf actually can't read the sample configs - the "update" option is to update something you had previously generated with stepconf
maybe you should try creating a new config with stepconf.
the default values may look like a valid configuration, but they aren't taken from any of the samples
yeah I'll try that then and see if it works...I thought I updated the config, guess not though
(actually, the default alues are valid, but may not match your machine :) )
lemme try that out
you probably did - just run emc again and select the first one at the top of the list
"My-Configurations" / "My-Mill" (unless you changed the name to something else)
You guys rock
I was using the wrong config
it moves now?
then when I attempted to use the custom config, I had left the limit switches with pins...i don't have any yet so they were erroring...removed them and it works like a charm
great. I bet you have some fine-tuning to do (velocity, acceleration) but it's good to hear that it's moving.
Yeah the fine tuning shouldn't be bad...Now I just need to learn what I'm doing lol
I'm getting anxious to cut my first part
throw some wood on there and cut away! :)
Need to figure out how to get from design to wood now lol
just jog around, it'll be satisfying ;)
any suggestions on a good cam program? preferably free or with a trial, my budget is just about gone for the month on this machine
either or...my cad computer is windows
using autocad 2006 for cad
I guess a better question would be what is the source of the designs?
full 3D contouring or more like 2.5D stuff?
cradek, does REALIZE work in ACAD 2006?
well, for now I'd 3d or 2.5 d would work
the stuff I am working on now is 2.5d, but I'd like to be able to do some contouring later
hmmm. there was a link posted in the last day (here or in the #emc-devel channel) that had a long list of CAD-related programs for Linux
there aren't a lot of good free options, there are some "not too expensive compared to CAD programs" options
the free alibre is fairly good
unless you need a heap of parts in one assembly
emtffke1: gcam.js.cx or ... (dammit, bookmark on different computer)
I don't know which are really good, and that will depend somewhat on what you're doing, I imagine
[20:01:59] <SWPadnos> http://gcam.js.cx/index.php/Main_Page
well on the contouring aspect I eventually hope to be able to make rotors for my rc heli so I'd like to be able to do that
also some signs and stuff
ouch. you'll probably need/want 5-axis for the propellers
but the blunt of my work is 2.5d just cutting out parts
I saw a video of a guy cutting them on a 3 axis and it worked out real nice
took a bit of extra work, but turned out great
I suppose a ball end mill will do a good job for that
I was thinking impellers/screws, but a propeller doesn't have any "overhang"
just need to do a flip of the stock once
yeah. a fourth axis (A) would be really good though
do you guys know of any good "new guy" tutorials...Getting started kind of stuff
SWPadnos: heli rotors are usually the same pitch from root to tip with some changes at the tip and root...
BigJohnT, they have curvature on both sides though, no?
although flat bottoms will auto better...
go to page 2 or 3, there's a long list of CAD-related stuff there
I guess I think of military helicopters when people say heli - not models and that stuff :)
emtffke1: you ever talk to HelicopterJohn over on CNC zone?
well. the heli I have has collective pitch, if I am just doing normal flying, I can fly with flat bottom rotors, if I am doing acrobatic flight I need the same pitch on both sides
no I hadn't
the high performance helicopters have some wild blades!
(the last time I talked about helicopters, it was regarding some wear monitoring software for the Apache and similar :) )
he has a HAAS CNC and is real friendly and likes to talk about machining helicopter parts and stuff
alright where'd the german dxf2gcode go
here's the post: http://ubuntuforums.org/showpost.php?p=3001600&postcount=22
I'll have to look him up
SWPadnos: if someone tries it and lets me know, I'll know
he just made an adapter for the 2.5 ghz transmitters so the neck strap was on the cg... it's cool
"REALIZE is now confirmed to run on Autocad 2005."
so, chances are probably good
SWPadnos: should we add the list on the wiki site?
SWPadnos that is a great list...I'll have to check some of those out
I just bought a copy of SolidWorks 2008 today...
my ass hurts now LOL
there already is a list on the wiki: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Cam
I'm gonna try a simple part first, just to get the hang of it...A bearing spacer that is 1/2 OD, 5/16ID, and 3/8 high out of HDPE...Sound easy enough?
Ouch...5k on solidworks...I've got a student edition of autocad 2006 and I fear the day I decide to upgrade to solidworks
emtffke1: I would cut some air first then move on to wood then do the harder things once you get to know your machine
well I was gonna run a couple tests with the Z axis zero'd about 2 inches above the table and the router not running so I can test some stuff out
ok i found it.. http://www.christian-kohloeffel.homepage.t-online.de/dxf2gocde.html
that will let you cut outlines drawn in qcad
forget about full 3d cad software if you arent willing to pay for it
will it cut profiles from a dxf file
better brush up on my german I guess
cool it is python too
well eventually I'm willing to pay for it
just have to build the budget back up lol
emtffke1 I missed some of the chatter but what kind of machine do you have?
well this is my first machine I'm working with right now....It's a small desktop mini mill that I built of my own design..
cutting area about 8"x8"
with about 4" of z movement
cool, did you post some pictures on the wiki site?
no, had some pics posted on CNC zone
plan on getting some more and posting them and when I get it to cut I plan to put a video up also
That does sound pretty cool. Nice work.
the whole machine is built from HDPE cutting boards that I scrapped... All 1/2" thickness
post some pictures here
[20:24:03] <BigJohnT> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Case_Studies
Ok, I'll throw some up there in a lil bit
I've got a website I'm working on for basically a journal of my new machine I'm working on.
this one will have a cutting area of 24x48"
then just post a link
are you going to use the little one to build the big one?
some of the parts
motor mounts and stuff like that
don't have a big working envelope, but for some of the smaller detailed parts it will work out real nice
well I have to design and build a torque wrench adapter for my vaccum wrench... bbl
I did trace the axis following error I had last week. It turned out to be an MRP.
Mop Related Problem.
Well i'm gonna check back in a bit, got to go design that bearing spacer real quick and try to convert it to gcode
this ought to be fun
just write the gcode
it's very easy for simple parts
imagine how you would cut the part if machining manually
it's easy for simple parts if you know g-code
(except you're really good cranking out diagonals and circles)
and by 'simple' you mean 'easy to program in g-code'
<- g-code hater
fenn: I don't care to argue about that, but people who think they can program a cnc machine to make parts without knowing ANY gcode are probably going to have trouble.
learning the very basic gcodes (line, arc) is not hard at all. anyone who has turned cranks can do it.
I programmed my CNC mill for years without knowing any g code
cradek: i'd argue that that situation is due to the g-code language itself and lack of standardization
cradek: Do you know of any good places that have a good g-code tutorial?
but my mill is conversational...
emtffkev: just start looking at the g-code quick reference: http://linuxcnc.org/docs/html/gcode.html
"Coordinated helical motion" means circles
and take a look at the output from the g code generators on the wiki site
better yet, look at the sample files in the nc_files directory
Well, I was having a problem, but an obvious solution presented itself.
I guess being in the room make you smarter.
ok cool I'll check that stuff out
MASEngr: it's true
generally you get dragged up/down to the average intelligence of the group
In case it comes up later, if you try to start a manual (joystick-controlled) configuration without the joystick being plugged in, you get a "pin does not exist" error.
Now if you'll excuse me, the machinist would like to get back to work.
makes sense. hal doesn't know if the joystick is "mission critical" or not
I wish Ed Nisley's speech/tutorial/presentation was available. Seems like it was probably pretty good.
I think it would be tricky at best to run in "joystick mode" without the joystick being plugged in.
[20:39:30] <cradek> http://email@example.com/msg05348.html
anyone in here really familiar with ubuntu? Does it come with a samba server or do i need to install that seperate
I'm not sure if the server is included (I think it is), but you should be able to connect to a Windows machine easily
SWPadnos: just for the fun of it, the smb server is not included, but if you rightclick on a folder to share it, it asks you if you want to share it for doze PC's, and if so it'll get installed/configured mostly automatically
[23:23:17] <micges> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Micges_Laser
Hey, there is linux drivers for this... Think it could be used within EMC for jogging? I think it can do 4 or 5 axis... http://www.3dconnexion.com/3dmouse/spacenavigator.php#
it should work with hal_input
micges, nice looking machine there
micges: is that your laser?
yes that is machine for my ideas testing
micges: as in you own and/or built it?
micges: BASTARD! =(
I builded and programmed the thing :)
not alone ofcourse :P
micges: You dont' have any idea how long I've wanted a laser; especially that size... That's how I got into CNC in the first place. We need to talk sometime =)
micges: Can you give me a rough idea at the laser power and the $ (USD) to get it at where it's at?
JymmmEMC: For first I'm from Poland
I know where to get in Europe
so tell him the price in PLN ;)
easy enough to convert (only 2.5:1 now though)
I can't remember the name for the parts of a zloty (sorry for the wrong character set)
(like cents in a dollar)
ah, thank you
table + optic + software + electronics: 130.000 - 150.000 PLN
micges: very nice work
if poland is part of EU why do you use PLN?
source Co2 1000W - 1500W: 200.000 PLN
smaller better source: 250.000 PLN
( JymmmEMC that's $80k)
fenn, maybe for the same reason the UK still uses the pound
micges, your moving the work under the laser?
yes, in the case of the UK ;)
fenn: we will have euro in near future, now it will kill our all bisnes :)
heh - you're about at the point the DM was when they switched to the Euro
basically prices doubled overnight, because a lot of stores just didn't bother changing the numbers, and the Euro was worth about 2 marks
BigJohnT: laser working 24h/day if that your asking
euro = 3.5 PLN
no I was asking if you moved the piece to cut while the laser stays still?
oh right, I went the wrong way in my head :)
yes moving table
you can see the (very nice) way covers for the long direction in one of the images
yes i see them
is the laser critical to Z position?
yes It has som sort of THC
did you build the THC?
external logic on stepper
what are you using to sense the material height?
THC is all our idea
no I searched the words :)
ok it's cool
sensor -> analog -> PWM -> stepper
some sort of that
ok, it's totally out of the EMC control?
I still want to integrate the THC with EMC
I'm making some progress in the corner lock area
I want it too but have no idea how